Diagnosis, Treatment, and Prevention of Congenital Toxoplasmosis in the United States

Congenital toxoplasmosis (CT) is a parasitic disease that can cause significant fetal and neonatal harm. Coordinated efforts by pregnant women, researchers, physicians, and health policy makers regarding potential primary and secondary preventive measures for CT and their implementation may lead to a lower incidence of CT as well as lower morbidity and mortality rates associated with CT. In the United States, the age-adjusted seroprevalence of Toxoplasma gondii among women of childbearing age (15–44 years) has declined over time (15%, 11%, and 9% in 1988–1994, 1999–2004, and 2009–2010, respectively; among US-born women only, the seroprevalence rates during these time periods were 13%, 8%, and 6%, respectively). Thus, approximately 91% of women of childbearing age in the United States are susceptible to Toxoplasma infection. Should these women become infected during pregnancy and remain undiagnosed and untreated, they could deliver an infant with CT. However, the incidence of acute primary infection is likely very low in the current era and is probably much lower than the 1.1 in 1000 pregnant women originally reported in 1960s.

There are 3 ways CT can occur. First, CT can develop through transmission of T gondii to the fetus from a previously seronegative, immunocompetent mother who acquired acute primary infection during pregnancy or within 3 months before conception. Second, CT can occur through reactivation of toxoplasmosis in a previously T gondii–immune pregnant woman who was severely immunocompromised during pregnancy. Third, CT can result after reinfection of a previously immune pregnant mother with a new, more virulent strain (eg, after international travel or after eating undercooked meat from areas where more virulent atypical strains predominate).

In cohorts of women who have been screened routinely during pregnancy and treated accordingly once primary infection was diagnosed, the mother-to-child transmission (MTCT) rate was <5% after an acute primary maternal infection very early in pregnancy, but MTCT rates were much higher with acute maternal infections acquired later in pregnancy (15%, 44%, and 71% after maternal seroconversions [acute primary infections] at 13, 26, and 37 weeks of gestation, respectively).¹ The risk of MTCT in untreated women may be higher. Factors (any or a combination) associated with an increased risk of MTCT are as follows: (1) acute T gondii infection during pregnancy, (2) immunocompromising conditions, (3) lack of antepartum treatment, (4) high T gondii strain virulence, and (5) high parasite load.

The incidence of CT, according to early cumulative published data from the New England Newborn Screening Program over a 12-year period (1988–1999), was 0.91 cases per 10 000 live births, which would have translated to the birth of approximately 365 infants with CT in the United States each year. The incidence of CT decreased after 1999 and, over the past 9 years (2006–2014), was approximately 0.23 cases per 10 000 live births. The true incidence of CT in the United States might be higher, because the sensitivity of the newborn screening test (blot-spot immunoglobulin [Ig] M test) is approximately 50% to 75%, and fetal losses attributable to severe CT were not counted.

Recent data from the National Reference Laboratory for Toxoplasmosis in the United States showed that 85% of 164 infants with CT identified over a period of 15 years were severely affected: 92% had chorioretinitis, 80% had intracranial calcifications, 68% had hydrocephalus, and 62% had all of these manifestations. These data were based on cases referred to reference toxoplasmosis centers in the United States, and they were not population based. The generalizability of these findings at a population level may be limited. The rate of symptomatic CT among infants diagnosed via the New England Newborn Screening program (1986–1992) was 40% (25% with eye disease at birth or follow-up and 29% with central nervous system [CNS] disease). However, in neonatal screening programs, spontaneous abortion, fetal demise, and pregnancy terminations attributable to severe CT are not captured. The rates of symptomatic CT in European cohorts are much lower than in the United States. Possible reasons for those disparities include differences in T gondii strains and the absence of antepartum screening and treatment of CT in the United States. In addition, differences between population-based, prospectively identified CT cases versus more selected cases referred to reference centers may bias reporting estimates.

Evidence in favor of antepartum treatment benefits to decrease the risk of MTCT of CT is variable and accumulated from several observational studies. One randomized therapeutic trial has been conducted in pregnancy.² European data from the Systematic Review on Congenital Toxoplasmosis (SYROCOT) international consortium, which performed a meta-analysis of individual patient data published in 2007,¹ suggested that the odds of MTCT were 52% lower (odds ratio [OR]: 0.48; 95% confidence interval [CI]: 0.28–0.80) when antepartum treatment was promptly initiated within 3 weeks after maternal seroconversion as compared with ≥8 weeks. The overall risk of transmission among women with primary infection in France decreased from 29% to 24% (P = .022) after 1992, when monthly antepartum screening became mandatory. These rates do not represent rates of MTCT without compared with prenatal treatment, because the majority of the women in France diagnosed with acute Toxoplasma infection, even in the earlier period when the frequency of prenatal screening was less intense, were prenatally treated. These rates may call into question the efficacy of treatment in preventing MTCT, because women in the postimplementation era were more likely to be identified by serologic testing alone rather than by ultrasonographic abnormalities in the setting of established infection. An alternative explanation for this paradoxical phenomenon is that early prenatal screening and treatment led to a decrease in fetal deaths, and thus the effect on the risk of MTCT was not very prominent. In other words, with prompt initiation of prenatal treatment, children who would have otherwise died of CT survive, making the decrease in MTCT less impressive. Evidence that supports this hypothesis includes data from the same prospective cohort study from Lyon, France, by Wallon et al,³ published in 2013, which showed a significant reduction in symptomatic disease among infected pregnant women when comparing cases reported before 1995 with those after 1995, when amniotic fluid (AF) testing by polymerase chain reaction (PCR) was initiated (from 11% before 1995 to 4% after 1995; P < .001). In Austria, where there is a nationwide antepartum screening program (the Austrian Toxoplasmosis Register), Prusa et al⁴ reported a sixfold lower risk of MTCT in women who received antepartum treatment compared with untreated women (9% [87 of 1007] vs 51% [32 of 63]), but this finding may have also been related to the timing of identification during gestation. In a recently published retrospective cohort study by Hotop et al⁵ from Germany, where spiramycin is given until the 16th week of pregnancy, followed by at least 4 weeks of combination therapy with pyrimethamine, sulfadiazine, and folinic acid (independently of the infection status of the fetus, with subsequent treatment determined according to the infection status of the fetus), very low rates of MTCT of CT (4.8% [33 CT cases infants/685 pregnant women]) were reported. An early Cochrane systematic review by Peyron et al⁶ published in 2000 evaluated the effect of treatment in pregnancy and concluded that, despite more than 3200 articles, only 9 studies had evaluated the effectiveness of prenatal treatment on the risk of MTCT, and these results were conflicting, resulting in insufficient evidence to evaluate the efficacy of treatment.

Some observational studies have evaluated the association of antepartum treatment and symptomatic infant disease. Recent reanalysis of data from 14 European centers⁷ suggested that antepartum treatment was associated with lower odds of severe neurologic sequelae in infants with CT (OR: 0.24; 95% CI: 0.07–0.71) (the authors advised caution in interpretation because the study included only 23 such very severely affected cases of CT and, of these, there were 9 terminations). In the Wallon et al³ prospective cohort study mentioned previously, the risk of symptomatic CT among infected mothers in France decreased from 11% to 4% (P < .001) after 1995 when T gondii testing with AF PCR assay was initiated, but the increase in the mild maternal infections included in the latter group may have played a role as well. Recent data from Germany by Hotop et al⁵ also showed a lower risk of symptomatic disease with early versus late initiation of maternal treatment (19% with early therapy versus 70% with late antepartum therapy; P = .006).

One cost-effectiveness model was recently developed by Stillwaggon et al⁷ to evaluate the implementation of universal antepartum screening after the French protocol of monthly serologic screening during pregnancy (including confirmatory testing at a reference laboratory of any positive results) with antepartum treatment, fetal ultrasonography/AF PCR assay, and infant follow-up/treatment.⁸ This decision analysis made a number of assumptions, including a cost of $12 per test, an estimated cost of fetal death of over $6 million, and an incidence of acute primary maternal infection during pregnancy of 1 in 1000 (including additional sensitivity analyses). It also assumed that treatment was highly efficacious and inexpensive. Although the study concluded that screening in the United States would be cost-effective, it remains unclear whether these conclusions would be reached if data were used assuming higher costs of screening, lower costs of loss, and less efficacy of treatment. A previous decision analysis did not arrive at the same conclusion.⁹ An early Cochrane systematic review on treatments for toxoplasmosis during pregnancy published in 2000 suggested that in countries where screening or treatment is not routine, these technologies should not be introduced outside the context of a carefully controlled trial.⁶ 

Infants with suspected/proven CT may need to be managed in consultation with toxoplasmosis reference centers in the United States. The diagnostic criteria for CT include any of the following: (1) persistence of positive Toxoplasma IgG antibodies beyond 12 months of age (gold standard); (2) positive Toxoplasma IgG antibodies and positive Toxoplasma IgM antibodies and/or positive Toxoplasma IgA antibodies; (3) positive Toxoplasma PCR assay results from amniotic fluid, peripheral blood, cerebrospinal fluid (CSF), urine, or other body fluids; and (4) positive neonatal Toxoplasma IgG antibodies (but negative Toxoplasma IgM and IgA antibodies) and serologic evidence of acute maternal T gondii infection during pregnancy and evidence of clinical manifestations suggestive of CT.

At the National Reference Laboratory for Toxoplasmosis (Palo Alto, CA) and the Toxoplasmosis Center (Chicago, IL), clinical evaluation of infants with suspected CT includes detailed physical examination, neurologic evaluation, ophthalmologic examination (preferably by a retinal specialist), and brainstem auditory evoked responses. Imaging evaluation includes the following: (1) computed tomography of the head (or head ultrasonography) and (2) abdominal ultrasonography. If CT has not been confirmed but also has not been ruled out, an infant’s workup includes complete clinical evaluation and serial IgG antibody titers every 4 to 6 weeks after birth until complete disappearance of Toxoplasma IgG antibodies.

At the National Reference Laboratory for Toxoplasmosis and the Toxoplasmosis Center, treatment of infants with suspected CT is continued for 12 months and includes pyrimethamine plus sulfadiazine plus folinic acid.

1. Pyrimethamine: 2 mg/kg per day, orally, divided twice per day for the first 2 days; then from day 3 to 2 months (or to 6 months [considered for symptomatic CT]), 1 mg/kg per day, orally, every day; and after that, 1 mg/kg per day, orally, 3 times per week

2. Sulfadiazine: 100 mg/kg per day, orally, divided twice per day

3. Folinic acid (leucovorin): 10 mg, 3 times per week

CT is a parasitic disease that can cause significant fetal and neonatal harm. Coordinated efforts by pregnant women, researchers, physicians, and health policy makers regarding potential primary and secondary preventive measures for CT and their implementation may lead to a lower incidence of CT as well as lower morbidity and mortality rates associated with CT. The purpose of this technical report is to summarize available information regarding the diagnosis, treatment, and prevention of CT.

The following clinical areas were targeted:

• evidence regarding the risk of maternal infection and MTCT and the risk of symptomatic disease in the United States and Europe;

• important differences between the North American and European literature that may need to be taken into consideration by practicing physicians and health care professionals in the United States;

• significant differences in the clinical spectrum and severity of CT seen in North America and Europe;

• diagnostic considerations in the mother, fetus, and infant;

• evidence from observational studies regarding the effectiveness of antepartum treatment to decrease MTCT and to prevent severe CT;

• prenatal and postnatal treatment protocols; and

• feasibility of routine antepartum screening and treatment of T gondii infections in the United States.

First, a PubMed search for publications in English language with the use of the following search strategies was performed: (1) (congenital toxoplasmosis) AND (mother to child transmission OR mother to fetus transmission OR mother to infant transmission) (up to September 15, 2014), (2) (congenital toxoplasmosis [ti]) AND (outcome OR follow-up OR chorioretinitis OR eye disease OR ocular disease OR intracranial calcifications OR ventriculomegaly OR hydrocephalus) (up to August 31, 2014), and (3) (toxoplasmosis [TI] AND United States) (January 1, 2004, to September 15, 2014) (see Supplemental Table 14). Second, the reference lists of key publications were screened to identify additional pertinent articles. Third, the evidence from some early key publications from cohorts of infants with CT whose mothers had not received antepartum treatment was also reviewed, because such older cohorts are more similar to the cohorts of children with CT still seen in the United States, where the majority of infants with CT are born to mothers without antepartum treatment. Fourth, we perused the systematic reviews already performed by the European Toxoprevention Study (EUROTOXO) group, the European Initiative for the study of CT on the following topics: (1) burden of disease from CT,¹⁰ (2) evaluation of diagnostic performance of diagnostic tests for CT,¹¹ (3) assessment of postnatal treatment effects,¹² (4) strategies for the prevention of CT,¹³ and (5) evaluation of adverse effects from antepartum and postnatal treatment of CT.¹⁴ 

A total of 457 articles were screened; 403 articles were identified with PubMed searches (288 with the first 2 searches and 130 with the third search strategy), 50 additional articles were identified by hand-screening the reference lists of key publications, and 225 articles were finally considered to be pertinent to this technical report and were included.

The identified evidence consisted of observational studies. Currently, there are no randomized controlled trials (RCTs) performed for the evaluation of different therapeutic approaches for CT, according to an early systematic review from the Cochrane Pregnancy and Childbirth Group (up to date as of February 2006) by Peyron et al.⁶ 

The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system^15,16 was used to assess the quality of evidence. The GRADE system is a step toward recognition that evidence from observational studies might become the vehicle for generating evidence for policy makers, especially in situations in which RCTs are considered unethical and large treatment effects have been documented in observational studies. According to the GRADE system (Fig 1), the quality of evidence for the effectiveness of antepartum screening and treatment would be considered of high quality on the basis of the following 2 criteria: “exceptionally strong evidence from unbiased observational studies” and “further research is unlikely to change our confidence in our estimates.” The quality of evidence for the effectiveness of postnatal treatment would be considered of moderate quality on the basis of the criterion “exceptionally strong evidence from unbiased observational studies”; however, “further research, if performed, is likely to affect our confidence in the estimate of efficacy and may change our estimate.”

T gondii is an obligate intracellular parasite with worldwide distribution that infects approximately one-third of the human population and a wide range of animals and birds. Infection with T gondii in humans can have devastating consequences for fetuses, children, immunocompromised patients, and immunocompetent individuals infected with virulent strains. Although members of the Felidae family are the definitive hosts, cat ownership per se is not correlated with the prevalence of human infection in most studies.^17,–22 The parasite has 3 infectious stages: (1) tachyzoites, which are responsible for rapid spread of the parasite between cells and tissues and for the clinical manifestations of toxoplasmosis; (2) bradyzoites, which are within tissue cysts and stay dormant for the life of the host unless the individual becomes severely immunocompromised; and (3) sporozoites, which are within oocysts, are shed by members of the felid family, and widely disseminate the agent in the environment.²³ Approximately half of infected individuals do not exhibit the conventional risk factors for acute infection or report clinical symptoms suggestive of acute toxoplasmosis at the time of their primary infection.^24,25 A survey of 76 women who gave birth to infants with CT indicated that 61% of these women had no exposure to cat litter or raw meat, 52% had no acute toxoplasmosis-like febrile illness during pregnancy, and 52% had neither of the 2.^24,25 Serologic testing is the only method that can reliably establish whether an individual has ever been infected, is chronically infected, or is experiencing an acute infection.

Genotyping studies have allowed the identification of 3 main clonal lineages of T gondii (types 1, 2, and 3) in Europe, North America, and South America.^26,27 Atypical strains that do not fall within these 3 clonal lineages have been frequently reported in North America and South America. In Western Europe, the predominant T gondii strain implicated in human disease was type 2,²⁸ in North America all 3 main clonal lineages strains were implicated in human disease (types 1, 2, and 3), and an additional new fourth clonal lineage (type 12) was recently identified.²⁹ In South America, mainly types 1 and 3 and atypical strains were identified^28,30,31; and in Africa, mainly type 3 and recombinant types 1/3, 1/2, and 1 were detected.³² In North and South America, atypical and more virulent strains have been implicated in more aggressive clinical manifestations in immunocompetent individuals.^33,–36 

It has been suggested that differences in T gondii strains may at least partially explain the observed differences in the clinical spectrum of CT in different parts of the world and particularly in Europe, North America, and South America. In the United States, more severe disease, and even preterm birth, has been associated with infections with non–type 2 strains³⁷; and in South America, more severe ocular disease has been reported in children with CT as compared with that reported in Western European countries.³⁸ 

Atypical T gondii strains have been reported from several areas of the world,^39,–49 including, but not limited to, Central and South America, Australia, and Africa. This distribution of strains may need to be considered when clinical syndromes consistent with severe toxoplasmosis are encountered in individuals returning from these areas.

Humans are incidental hosts and become infected primarily by the oral route, that is, by ingestion of oocysts present in contaminated food, water, or soil or by ingestion of tissue cysts contained in infected meat²³ (Table 1). They also can become infected in utero by transplacental transmission of the parasite from an acutely infected mother to the fetus, from an infected organ donor in the setting of organ transplantation,^50,–53 and rarely, by blood transfusion or laboratory accidents.⁵⁴ Recent studies suggested that oocysts were the predominant route of transmission of T gondii infections in the United States; 78% (59 of 76) of pregnant women with acute primary T gondii infections during pregnancy who gave birth to infants with CT had serologic evidence suggestive of being infected by oocysts.^24,55 

Different types of locally produced meat from retail meat stores in the United States (eg, pork, lamb, goat, wild game meat) have been found to be contaminated with T gondii.^{40,47,57,–60} Moreover, with the globalization of the food market, imported meats from other countries (that could be infected even with atypical, more virulent T gondii strains) also can be found in the United States and Europe. Freezing (below –20°C [–4°F] for at least 48 hours)⁶¹ and thorough cooking to at least 63°C (145°F) for whole cut meat (excluding poultry), 71°C (160°F) for ground meat (excluding poultry), and 74°C (165°F) for poultry (whole cut or ground)^62,63 have been shown to inactivate T gondii tissue cysts. Neither microwave cooking nor chilling at 5°C for 5 days is sufficient to kill tissue cysts (microwave cooking may not generate a homogenous temperature of 67°C [153°F]).^61,64 

The following risk factors for acute T gondii infection were reported in a US study²⁰: eating raw ground beef, rare lamb, or locally produced cured, dried, or smoked meat; working with meat; eating raw oysters, clams, or mussels; drinking unpasteurized raw goat milk; or having ≥3 kittens in the household. Eating raw oysters, clams, or mussels is a novel risk factor that was recently identified; however, in selected populations in the United States or in other parts of the world,^65,66 additional/different risk factors (eg, contact with cat feces or drinking untreated water) also have been implicated. Untreated water has been found to be a source of major community outbreaks of acute toxoplasmosis in Canada and Brazil.^67,68 

In up to 50% of individuals in whom an acute T gondii infection was confirmed, it was not possible to identify a known risk factor for their infection.²⁴ Thus, toxoplasmosis should be considered in the differential diagnosis of ill patients with symptoms suggestive of toxoplasmosis, even in the absence of known risk factors for T gondii infection.

Community outbreaks of acute toxoplasmosis have been described in several parts of the world,^{34,49,69,–74} including North America.⁶⁷ Outbreaks of acute Toxoplasma infection within families (ie, with more than 1 family member being infected) have been documented in the United States,⁷⁴ and the prevalence of within-family clusters of acute Toxoplasma infections in the United States has also been studied.⁷⁵ Although a cost-benefit analysis of routine screening of additional family members of index cases with acute toxoplasmosis might be useful, until such a study is completed, it is important for physicians to be aware of this phenomenon and consider screening additional family members of index cases diagnosed with acute toxoplasmosis, especially if pregnant women, immunocompromised patients, or young children live in such households.⁷⁴ Young children living in such households who acquire acute T gondii infection might not be able to appropriately communicate problems with their vision if acute toxoplasmic chorioretinitis occurs.

A seasonal pattern for acute toxoplasmosis has been recently identified in Europe⁷⁶ and the United States.⁷⁷ In France, the first highest peak of acute Toxoplasma infections during pregnancy was observed between August and September, and the second highest peak was observed between October and December.⁷⁶ In the United States, a similar peak during December was observed for cases of acute toxoplasmic lymphadenopathy referred to the Palo Alto Medical Foundation Toxoplasma Serology Laboratory (PAMF-TSL),⁷⁷ the National Reference Laboratory for the Centers for Disease Control and Prevention and the US Food and Drug Administration (FDA).

In the United States, the rates of T gondii IgG-seropositive individuals has decreased over the past 2 decades. Data from the most recent NHANES for 2009–2010^78,79 showed an overall age-adjusted seroprevalence in people older than 6 years of 12.4% (95% CI: 11.1–13.7%) (The respective unadjusted seroprevalence rate was 13.2% [95% CI: 11.8–14.5%].) The age-adjusted seroprevalence among women of childbearing age (15–44 years) in the whole US population was 9.1% (95% CI: 7.2–11.1%) in 2009–2010 compared with 11% in 1999–2004 and 15% in 1988–1994. For US-born women of childbearing age, the respective seroprevalence rates were 6%, 8%, and 13% in 2009–2010, 1999–2004, and 1988–1994, respectively. Most US women of childbearing age are susceptible to Toxoplasma infection. If women become infected during pregnancy, they could give birth to an infant with CT. People born outside the United States were significantly more likely to be seropositive than people born in the United States (25.1% vs 9.6%, respectively).⁷⁸ Seroprevalence rates were also higher in persons with a Hispanic versus a non-Hispanic white racial background (15.8% vs 10.2%, respectively).⁷⁸ In addition, seroprevalence rates in children living on farms in Wisconsin have been reported to be fivefold higher than rates in children not living in farms (18% [29 of 159] vs 4% [8 of 184], respectively).⁸⁰ 

In the United States, toxoplasmosis was found to be the second leading cause of death and the fourth leading cause of hospitalizations attributable to foodborne illnesses.^81,82 Recent cumulative US data over an 11-year study period (2000–2010) identified 789 toxoplasmosis-associated deaths, with a cumulative productivity loss attributable to toxoplasmosis of $815 million.⁸³ Black and Hispanic persons had the highest toxoplasmosis-associated mortality. HIV infection, lymphoma, leukemia, and connective tissue diseases were associated with increased risks of toxoplasmosis-associated deaths.⁸³ Population-based data estimated that T gondii infects approximately 1.1 million people each year in the United States⁸⁴; toxoplasmic chorioretinitis is estimated to occur in approximately 2% of T gondii–infected individuals (approximately 21 000 people in the United States per year), and symptomatic chorioretinitis in approximately 0.2% to 0.7% of T gondii–infected individuals (approximately 4800 people in the United States per year).⁸⁴ 

According to the 2010 report by the Council of State and Territorial Epidemiologists, reporting of toxoplasmosis is not mandatory in most of the United States.⁸⁵ As of 2010, toxoplasmosis was a reportable disease only in 19 states; in an additional 9 states, toxoplasmosis was previously a reportable disease, but not as of 2010.

T gondii seroprevalence rates vary in different parts of the world⁸⁶ and can range from <10% in some northern European countries⁸⁷ to as high as 60% to 80% (eg, in Mexico⁸⁸ and Brazil^89,90).

The incidence of acute primary T gondii infections during pregnancy in the United States, according to the National Institutes of Health–sponsored Collaborative Perinatal Project in 22 845 women (1959–1966) who were screened every 2 months during pregnancy, at birth, and at 6 weeks postpartum, was estimated to be 1.1 cases per 1000 pregnant women.⁹¹ However, that is likely to be an overestimate of the true incidence of acute T gondii infections during pregnancy in the current era, with the overall decrease in T gondii seroprevalence. Extrapolating from recent data from the New England Newborn Screening Program of the incidence of CT of approximately 0.23 per 10 000 live births, and after taking into account that this value might underestimate the true incidence of CT by approximately 50%, the true incidence of CT could be as high as double that value (approximately 0.5 CT cases per 10 000 live births). The incidence of CT may also be higher in areas and in subpopulations in the United States with higher overall T gondii seroprevalence rates.^66,78 Then, assuming an overall MTCT rate of 25% (consistent with more recent data), the estimated incidence of acute primary infection during pregnancy in the United States would have been approximately 0.2 per 1000 pregnant women. If these numbers (0.2–1.1 acute cases per 1000 pregnant women) are extrapolated to the approximately 4 million pregnancies per year in the United States,⁹² approximately 800 to 4400 women per year in the United States acquire acute T gondii infections during pregnancy.

Of note, when France first began antepartum screening for toxoplasmosis in the early 1980s, the incidence rate of acute primary T gondii infections during pregnancy at that time was 4 to 5 per 1000 (the population seropositivity rate at that time was very high, and the percentage of nonimmune, susceptible pregnant women was low).⁹³ 

Recently reported rates of acute primary infections during pregnancy in other countries ranged from 0.5 in 1000 pregnancies in Sweden⁹⁴ to 2.1 in 1000 pregnancies in France (95% CI: 1.3–3.1).⁹⁵ However, chronologic and methodologic differences between such studies preclude accurate direct comparisons.

Data regarding the risk of MTCT of T gondii infection come from studies in which almost all women were routinely screened during pregnancy and therefore received antepartum treatment once primary infection was diagnosed.^1,3,96 The SYROCOT meta-analysis of individual patient data from an international consortium, provided important information from 26 countries participating in the consortium. However, data for the effect of antepartum treatment on the MTCT risk were based only on European cohorts that had antepartum screening/treatment programs; neonatal cohorts from the United States, Brazil, and Colombia were excluded. Across the 26 cohorts, the overall risk of MTCT was <5% after seroconversion (maternal acute primary infection) early in pregnancy, 15% (13%–17%) after seroconversion at 13 weeks, 44% (40%–47%) after seroconversion at 26 weeks, and 71% (66%–76%) after seroconversion at 37 weeks (Fig 2).¹ Consistent rates for the MTCT risk were published from the Lyon, France, cohort for the past 2 decades (1987–2008).³ For maternal seroconversion before 8 weeks of gestation, the MTCT risk was <8%. For maternal seroconversion between 8 and 10 weeks of gestation, the MTCT risk was <10%³ (Fig 3). Other factors associated with an increased risk of MTCT included acute T gondii infection during pregnancy, immunocompromising conditions, lack of antepartum treatment, and T gondii strain virulence, higher parasite load, and parasite source (oocyst/sporozoite-related infections from contact with cat feces versus tissue cyst/bradyzoite-related infections from ingestion of undercooked meat).

RCTs to assess the efficacy of toxoplasmosis antepartum screening and treatment programs for the reduction in MTCT risk have not been conducted, although data from prospective and retrospective observational studies provide information on the association of prenatal treatment and MTCT as well as symptomatic congenital disease. Observational studies have led to the uptake of routine screening in many countries for the purpose of providing treatment, making the conduct of needed RCTs difficult.⁷ 

The individual patient-data meta-analysis of the European data from the SYROCOT international consortium published in 2007¹ revealed that when antepartum treatment was promptly initiated within 3 weeks after maternal seroconversion, the odds of MTCT were 52% lower when compared with delayed treatment (≥8 weeks after seroconversion; adjusted OR: 0.48; 95% CI: 0.28–0.80). Decreases of similar magnitude in MTCT risk were observed when antepartum treatment was initiated between 3 and 5 weeks and between 6 and 8 weeks after seroconversion, when compared with delayed treatment (ORs [95% CIs]: 0.64 [0.40–1.02] and 0.60 [0.36–1.01], respectively). These analyses might have also underestimated the true effectiveness of antepartum treatment, because some of the women who received antepartum treatment who would have otherwise lost their fetuses because of T gondii infection might have eventually delivered infants with only mild CT. Of note, the overall risk of MTCT among women with primary infection in France decreased from 29% (125 CT cases per 424 infected pregnant women) before 1992 to 24% (388 CT cases per 1624 infected pregnant women) after 1992 (P = .022) when monthly antepartum screening became mandatory (as opposed to simply “recommended” without a specified frequency of screening before 1992). These rates may call into question the efficacy of prenatal treatment to prevent MTCT, because the women in the postimplementation era were more likely to be identified by serologic testing alone rather than by ultrasonographic abnormalities in the setting of established infection. An alternative explanation for this paradoxical phenomenon is that early prenatal screening and treatment led to a decrease in fetal deaths, and thus the effect on the risk of MTCT was not very prominent. In other words, with prompt initiation of prenatal treatment, children who would have otherwise died of CT survive, making the decrease in MTCT risk less impressive. Evidence that supports this hypothesis includes data from the same prospective cohort study from Lyon, France, by Wallon et al,³ published in 2013, which showed a reduction in symptomatic disease among infected pregnant women when comparing cases reported before 1995 with those after 1995 when AF testing with PCR was initiated (from 11% [87 symptomatic CT cases per 794 infected mothers] before 1995 to 4% [46 symptomatic CT cases per 1150 infected mothers] after 1995; P < .001). With the routine use of AF PCR, fetal infections were properly diagnosed, and the initiation of treatment with pyrimethamine/sulfadiazine (P/S) was promptly instituted. Poor adherence to monthly screening by pregnant women might have also contributed to this phenomenon. A recent study by Cornu et al reported that only 40% of seronegative pregnant women in France had all ≥7 screening tests according to the French scheme.⁹⁷ However, this study was based on only a small sample of pregnant women in France.

Prusa et al⁴ reported in 2015 that, in Austria, where there is a nationwide antepartum screening program (the Austrian Toxoplasmosis Register), a sixfold lower risk of MTCT was observed in women who received antepartum treatment compared with untreated women, after also taking into account the gestational age at maternal infection (9% [87 CT cases per 1007 infected pregnant women] versus 51% [32 CT cases per 63 infected pregnant women]). However, their findings may have been related to the timing of identification during pregnancy.

Hotop et al⁵ reported in 2012 that, in Germany, where spiramycin is given until the 16th week of pregnancy, followed by at least 4 weeks of combination therapy with pyrimethamine/sulfadiazine/folinic acid (independently of the infection status of the fetus, with subsequent treatment determined according to the infection status of the fetus), very low rates of MTCT of CT were seen (4.8% [33 CT cases per 685 infected pregnant women]).

An early Cochrane systematic review by Peyron et al⁶ published in 2000 evaluated prenatal treatment in pregnancy and concluded that, despite more than 3200 articles, only 9 studies were identified by that time that had evaluated prenatal treatment and the risk of MTCT, and these results were conflicting, resulting in insufficient evidence to evaluate the efficacy of treatment.

Some observational prospective and retrospective cohort studies have shown the association of antepartum treatment (and particularly of prompt initiation of prenatal treatment as soon as possible after acute maternal infection) with the prevention of symptomatic infant's disease. Cortina-Borja et al,⁷ in a recent analysis of data from 293 infants with CT from 14 European centers in 6 countries, suggested that antepartum treatment was associated with lower odds of severe neurologic sequelae or death (SNSD; OR adjusted for gestational age at maternal seroconversion: 0.24; 95% CI: 0.07–0.71; 10 cases of SNSD per 104 untreated CT cases versus 13 SNSD per 189 treated CT cases). (The authors advised caution in interpretation, because the study included only 23 such very severely affected cases of CT, and of these, there were 9 pregnancy terminations.) This composite outcome of SNSD included the following outcomes: pediatric report at any age of microcephaly; insertion of intraventricular shunt, an abnormal or suspicious neurodevelopmental examination that resulted in referral to a specialist, or seizures during infancy or at an older age that required anticonvulsant therapy; severe bilateral visual impairment in both eyes assessed after 3 years; cerebral palsy; or death from any cause before 2 years of age, including termination of pregnancy (Fig 4).

Wallon et al,³ in their prospective cohort study from France (published in 2013), showed that the risk of symptomatic CT decreased in France after 1995, when testing for T gondii in AF by AF PCR was initiated, from 11% (95% CI: 9–13% [87 symptomatic CT cases per 794 infected mothers]) before 1995 to 4% (95% CI: 3–5% [46 symptomatic CT cases per 1150 infected mothers]) after 1995 (P < .001), but the increase in the mild maternal infections included in the latter group may have played a role as well. AF PCR testing allowed for the prompt initiation of effective antepartum treatment (P/S) for a greater number of women with infected fetuses. Earlier reports from France also had shown that the severity of clinical manifestations dramatically decreased after the introduction of antepartum screening and treatment (29% [68 of 234 CT cases] of CT cases were symptomatic between 1984 and 1992 versus 99% of 147 CT cases between 1949 and the 1960s).⁹⁸ 

A recent retrospective cohort study in 685 infected pregnant women from Germany by Hotop et al,⁵ published in 2012, also showed a lower risk of symptomatic disease with early versus late initiation of maternal treatment (fewer than 4 weeks versus more than 8 weeks after seroconversion; 19% with early antepartum therapy versus 70% with late antepartum therapy; P = .006). In this study,⁵ after the diagnosis of acute primary maternal infection, spiramycin was initiated in all women up to week 16 of gestation, at which time treatment was changed in all women to P/S for at least 4 weeks. P/S was continued until further AF PCR and fetal ultrasonographic testing excluded or confirmed fetal infection; in the former situation, antepartum treatment was changed back to spiramycin until delivery, whereas in the latter situation, P/S was continued until delivery.

Kieffer et al,⁹⁹ in their prospective cohort study published in 2008 in 335 infants with CT from the Lyon, Paris, and Marseille cohorts, found that a delay of more than 8 weeks from the time of maternal seroconversion to initiation of maternal treatment was associated with an increased risk of chorioretinitis in the first 2 years of life (hazard ratio [HR]: 2.54; 95% CI: 1.14–5.65).

A previous report from the European Multicenter Study of Congenital Toxoplasmosis (EMSCOT) study by Gras et al¹⁰⁰ from 2005 also showed 72% lower odds of intracranial lesions in infants whose mothers initiated treatment less than 4 weeks after seroconversion (OR: 0.28; 95% CI: 0.08–0.75). However, the treatment benefit was lost when antepartum treatment was initiated more than 4 weeks after seroconversion.¹⁰⁰ 

It should be noted here that the analysis of data from the European cohorts participating in the SYROCOT international consortium (550 infants with CT per 1745 women), published in 2007,¹ failed to show that prenatal treatment reduced the risk of clinical manifestations (adjusted OR for treated versus not treated: 1.11; 95% CI: 0.61–2.02). However, a characteristic of the SYROCOT analysis could explain this finding. Specifically, the 4 cohorts with the highest burden of disease from North and South America were excluded from their analysis (2 cohorts from Brazil, 1 from Colombia, and 1 from the United States [Massachusetts]). The justifications for this exclusion were as follows: (1) differences in the burden of the disease compared with the remaining cohort (eye disease was more frequent and more severe in South American than in European cohorts), (2) differences in the T gondii strains, and (3) differences in the methods used for the diagnosis of intracranial calcifications (computed tomography versus head ultrasonography in Europe). It is not surprising that epidemiologic studies that excluded severe cases and attempted to estimate the risk of symptomatic disease in groups of treated infants with CT reported only a weak effect for antepartum treatment. Moreover, in the analyzed European cohorts, only 18% of the pregnant women (307 of 1745) across the 26 cohorts were untreated, and only one-third (164 of 550) of the infants with CT were born to mothers who had not received antepartum treatment. This finding might have limited the study power to detect statistically significant differences between prenatal treatment and no treatment on clinical manifestations of CT. Additional discussion on the relationship between prenatal treatment and the risk of symptomatic CT can be found in the Supplemental Information.

Data regarding the disease burden of CT in the United States are limited, although this information is crucial for policy makers and public health decisions for routine antepartum screening.¹⁰¹ At the time of this report, only Massachusetts and New Hampshire offer routine newborn screening for CT.¹⁰² The New England Regional Toxoplasma Working Group, during an early 6.5-year period of neonatal screening in Massachusetts (1986–1992) and a 4-year period in New Hampshire (1988–1992), identified 52 cases of CT among 635 000 screened infants, and the estimated incidence of CT was 0.82 cases per 10 000 live births.^103,–105 This estimated incidence translates (if extrapolated to the approximately 4 million infants born each year in the United States) to approximately 330 infants being born each year in the United States with CT. Subsequent cumulative published data from the New England Newborn Screening Program¹⁰⁶ over the 12-year period of 1988–1999 identified 93 cases of CT among 1 019 904 screened infants, and the estimated incidence of CT was 0.91 cases per 10 000 live births, which would translate to ∼365 children being born each year in the United States with CT. The respective incidence among US-born pregnant women was 0.82 cases per 10 000 live births (H. W. Hsu, MD, Massachusetts Department of Public Health, personal communication, 2015). The incidence of CT has decreased since 1999 and, over the past 9 years (2006–2014), was approximately 0.23 cases per 10 000 live births (H. W. Hsu, MD, Massachusetts Department of Public Health, personal communication, 2015). This decrease of approximately 50% appears to be in line with the decrease in seroprevalence in adults in the United States since early 2000.⁷⁸ Older prospective studies, even before the 1970s, had estimated an incidence of CT of up to 10 cases per 10 000 live births.¹⁰³ 

The true incidence of CT in the United States might be even higher, because the sensitivity of the newborn screening test (filter paper blot spot Toxoplasma-IgM) used in the New England Neonatal Screening Program is approximately 50% to 75%.^78,104,106 The incidence of CT may also be higher in areas of the United States where the seroprevalence of T gondii infections is higher. Moreover, neonatal screening programs might also underestimate the true incidence of CT, because fetal losses attributable to severe CT were not counted.

Neonatal screenings also exist in the United States for several inborn errors of metabolism, even though the numbers of affected infants are very small (annual incidence of phenylketonuria, which is the most common of these diseases, is 1 per 15 000 live births¹⁰⁸), because treatment markedly alters the outcome in those infants. The role of antepartum screening and treatment of CT in the United States may need reevaluation in the light of recently accumulated evidence from several observational studies that showed that prompt initiation of treatment can alter the clinical course of the disease.^109,–112 

In France, the first National Surveillance System for CT was established in 2007, approximately 30 years after the initiation of the national antepartum screening programs. The prevalence of CT according to this system was estimated to be 3.3 per 10 000 live births (95% CI: 2.9–3.7; 272 CT cases per 818 700 births),¹¹³ and the prevalence of symptomatic CT was 0.34 per 10 000 live births (95% CI: 0.2–0.5; 28 symptomatic CT cases per 818 700 births).¹¹³ Moreover, among the 272 CT cases, there were 11 pregnancy terminations (4%) (6 abortions and 5 cases of fetal demise). The overall seroprevalence rates in the French population also decreased by nearly 50% (from 84% to 44%) between the 1960s and 2003, which could have also explained, in part, the observed low rates of CT in France. However, over the same time period, strong evidence has accumulated indicating that the universal antepartum screening/treatment program dramatically affected the clinical severity of the disease.^1,3,99,114 

Cumulative data from a recent 13-year period (1996–2008) from the Rhones-Alpes area of France revealed that approximately 80% of infants with CT are now asymptomatic at birth (161 asymptomatic infants per 207 CT cases³). For comparison, only 60% (95% CI: 45–74% [29 of 48]) of US infants with CT were asymptomatic at birth, according to an early report from the New England Newborn Screening Program.¹⁰⁴ However, because the screening methods used in France (antepartum screening and treatment) and in the New England Newborn Screening Program (only postnatal screening and treatment as indicated) were different, the utility of such a comparison is limited.

According to the 2006 EUROTOXO study report,¹⁰ reported rates of CT vary across different European countries (ie, 0.7 in 10 000 in Sweden,⁹⁴ 0.8 in 10 000 in the United Kingdom,¹¹⁵ 1 in 10 000 in Austria,⁴ 1.6 in 10 000 in Denmark,¹¹⁶ 5 in 10 000 in Switzerland, 11 in 10 000 in Poland,¹¹⁷ and 13 in 10 000 in Germany¹⁰). However, the differences in those rates should be interpreted cautiously because of chronologic and methodologic differences.

A systematic review of the global burden of CT,^118,–122 which was part of a larger study of the global burden of foodborne toxoplasmosis coordinated by the World Health Organization, has been published.¹²² Nine major databases from published and unpublished sources were systematically reviewed, and their authors were also directly contacted. The global yearly number of CT cases was estimated to be 190 100 (95% CI: 179 300–206 300). Cumulative incidence rates of CT were reported according to geographic region: Africa, 20 to 24 in 10 000 births; North America, 6 in 10 000 births; South America, 18 to 34 in 10 000 births; European countries, 5 in 10 000 births; Eastern European countries, 15 to 16 in 10 000 births; Southeast Asia, 8 to 13 in 10 000 births; and West Pacific regions (Australia and New Zealand), 12 to 14 in 10 000 births.¹²² The highest incidence rates were reported in South America and were considered to be driven by the more virulent genotypes circulating in that region.^29,122 In addition, the global burden of disease from CT was estimated to be equivalent to 1.2 million disability-adjusted life-years.¹²² 

Significant differences in the prevalence and severity of major manifestations (eg, chorioretinitis, brain calcifications, and hydrocephalus) have been observed in the United States, France, other Western European countries, and South America. A compendium of ocular, neurologic, and other manifestations reported in the literature to be associated with CT is shown in Table 2.

In the United States, CT is notable for its severity in presentation,¹²⁴ but CT is not on the list of nationally notifiable diseases.¹²⁵ Over a period of 15 years (1991–2005) at the PAMF-TSL,¹²⁴ 84% of infants diagnosed with CT were severely affected (116 of 138 CT cases with clinical information, with their clinical findings confirmed at the NCCCT study center)¹²⁴; 92% had chorioretinitis (119 of 129 CT cases with this information), 80% had intracranial calcifications (94 of 118 CT cases with this information), 68% had hydrocephalus (67 of 99 CT cases with this information), and 62% had all 3 manifestations (53 of 86 CT cases with this information).¹²⁴ Accurate estimates of additional clinical manifestations, such as microcephaly, cerebral atrophy, hepatomegaly, splenomegaly, skin rashes, including purpura, jaundice, thrombocytopenia, elevated CSF protein, and CSF pleocytosis, were not possible, because information regarding whether all infants had been evaluated/tested for these outcomes was not available.

It is possible that children who were eventually referred to the PAMF-TSL for laboratory confirmation of their diagnosis were a select group of infants and not representative of the true spectrum of CT in the United States. On the other hand, these rates could also possibly underestimate the severity of CT in the United States, because they did not capture the CT cases that resulted in spontaneous abortion, fetal deaths, or pregnancy termination because of severe disease.

Similar high rates of severe CT were reported by the National Chicago-Based Collaborative Study on Congenital Toxoplasmosis (NCCCT) in its cumulative report from 1981 to 2004 (85% with chorioretinitis, 85% with intracranial calcifications, and 50% with hydrocephalus).¹¹² These reported rates pertained only to the 96 children with CT in the “severe disease stratum” and not all of the 120 children with CT in the cohort (including 24 CT cases who were asymptomatic or with only mild clinical manifestations). For the vast majority of infants in the NCCCT study, their mothers have not received antepartum treatment. These high rates are also similar to those reported in the 1950s by Eichenwald¹²⁶ from an early Danish cohort of untreated infants (94% of CT cases had chorioretinitis, 50% had intracranial calcifications, and 28% had hydrocephalus). Additional clinical manifestations reported in the NCCCT study¹¹² included microcephaly (15%), seizures (20%), microphthalmia (20%), hepatosplenomegaly (30%–35%), thrombocytopenia (40%), anemia (20%) skin rashes (25%), and jaundice (60%).

For newborn infants with CT diagnosed by the New England Newborn Screening Program (1986–1992),¹⁰⁴ 40% (19 of 48 CT cases) had either chorioretinitis or CNS abnormalities. More specifically, 19% (9 of 48) had eye disease diagnosed at birth (4% [2 of 48] had active chorioretinitis at birth and 15% [7 of 48 CT cases] had retinal scars at birth without active inflammation). Another 3 infants, among the 30 infants with CT without eye disease at birth, had eye disease diagnosed during the follow-up at ≥1 year. Overall, 25% (12 of 48) of infants with CT had eye disease at birth or during follow-up and 29% (14 of 48 CT cases) had CNS abnormalities (elevated CSF protein: 25% [8 of 32 CT cases]; intracranial calcifications: 20% [9 of 46 CT cases]; enlarged ventricles identified on computed tomography or ultrasonography: 2% [1 of 47 CT cases]). According to the New England Newborn Screening Program, only 60% of infants were asymptomatic at birth, compared with 80% of infants with CT who were reported as asymptomatic in Europe.³ As noted previously, spontaneous abortion, fetal demise, or pregnancy terminations because of severe CT were not captured in the New England Newborn Screening Program.

There are possible explanations for the observed differences in the reported rates of severe CT between the European Union and United States. The lower rates of severe CT in the European Union cohorts might reflect the following: (1) differences in the spectrum of CT observed in population-based, prospectively identified consecutive CT cases versus more selected cases referred to reference centers such as the PAMF-TSL and the University of Chicago Toxoplasmosis Center; (2) differences in the implicated T gondii strains; or (3) differences in the use of antepartum treatment of CT (mothers of infants in the US cohorts did not receive antepartum treatment). Approximately 61% of T gondii strains implicated in CT in the United States are non–type 2 strains,³⁷ whereas in France, 95% of cases are caused by type 2 strains.^29,127 It is possible that the non–type 2 strains found in the United States are more pathogenic than the type 2 strains commonly found in Europe.^128,129 

In 1 report, only 31% (34 of 108 CT cases; 95% CI: 23%–41%) of infants with CT in the United States who were treated during their first year of life developed new chorioretinal lesions during long-term follow-up.¹¹⁰ However, in another report, 72% (18 of 25 CT cases; 95% CI: 51%–89%) of US children with CT who were diagnosed after infancy (and did not receive any postnatal treatment during infancy) developed new chorioretinal lesions during follow-up.¹⁰⁹ When postnatal treatment was started at ≤2.5 months of age and continued for 12 months, recurrent eye disease developed in 9% of children without substantial neurologic disease at birth and in 36% of children with moderate to severe neurologic disease at birth.¹¹² 

Initial studies in untreated children with CT revealed unfavorable long-term neurologic outcomes in approximately ∼50% of children.⁹⁸ Subsequently, cumulative reports from the NCCCT study (1981–2004) with data on 120 children with CT (24 of whom were asymptomatic or had only mild findings and 96 of whom had severe clinical findings) reported relatively favorable neurologic outcomes after 1 year of postnatal treatment¹¹² (Table 3). However, long-term follow-up was not reported for all of the children in the 2 groups. More analytically, treatment of the 24 children without substantial neurologic disease at birth resulted in the following: (1) abnormal cognitive and neurologic outcomes in all but 1 child (4%), (2) interval decrease in IQ (≥15 points) in 4% of children (1 of 24), (3) abnormal auditory outcomes in no children, (4) vision impairment in 15% of children (2 of 13), and (5) recurrences of eye diseases in 9% of children (1 of 11). For comparison, treatment of the 96 children with severe CT at birth resulted in (1) abnormal neurologic and/or cognitive outcomes in 27% of children (18 of 66), (b) interval decrease in IQ (≥15 points) in 16% of children (9 of 55), (3) abnormal auditory outcomes in no children, (4) vision impairment in 85% of children (47 of 55 [with the retinal lesion causing the visual impairment already present at birth]), and (5) recurrence of eye disease in 36% of children (17 of 47). In the above estimates of clinical outcomes, 11 additional children who died between 3 months and 7 years of age were not considered. The majority of these children already had severe neurologic manifestations at birth. Although in this cohort there were children with severe and likely irreversible brain and eye disease, overall, these treated children had substantially better outcomes than did children from historical cohorts in the 1960s–1980s who were untreated or treated for only 1 month.

Earlier reports from the NCCCT study in 36 postnatally treated CT cases (1981–1991) also provided evidence for reversible disease or later neurologic adaptation.¹³⁰ Several of those infants had CNS abnormalities attributable to T gondii infection (including CSF pleocytosis, hypoglycorrhachia, and CSF protein elevation) that resolved during postnatal therapy. Some additional favorable neurologic outcomes reported in this earlier report included the following: (1) resolution of seizures in 50% (4 of 8 children) within the first months of life (however, there were 2 other children who developed new seizures between 3 and 5 years of age), (2) resolution of tone and motor abnormalities by 1 year of age in 60% (12 of 20 children), (3) IQ improvement in 79% (23 of 29 children [normal Mental Developmental Index] at 1 year of age; however, 21% [6 of 29 children] had a Mental Developmental Index <50 but without significant deterioration over time), and (4) normal/near-normal neurodevelopmental outcome in 94% (17 of 18 children) without hydrocephalus and in 75% (6 of 8 children) with obstructive hydrocephalus with shunt. However, because of the small numbers, there was large uncertainty around those estimates for the long-term prognosis of these infants. Moreover, in children with hydrocephalus ex vacuo and elevated CSF protein, the outcome was poor.¹³⁰ 

A small case series from the United States described resolution of intracranial calcifications in postnatally treated infants with CT.¹¹¹ In 75% (30 of 40) of cases, brain lesions resolved or diminished in size by 1 year of age, but in 25% (10 of 40) of cases, lesions remained the same. The diminution or resolution of intracranial calcifications was also associated with improved neurologic functioning.¹¹¹ However, an increase in the size and/or number of calcifications has also been reported during months to years after the initial infection, but this finding might reflect healing.⁹⁸ 

In Europe, the vast majority of pregnant women undergo routine Toxoplasma antepartum screening and treatment if needed. Thus, the rates of clinical manifestations and the severity of CT reported over the past 30 years from Europe pertain primarily to children whose mothers received antepartum treatment and who also received postnatal treatment. Data on the effect of postnatal treatment are more limited in European cohorts, because this effect is more difficult to separate from the effect of antepartum treatment.

The overall risk for “any” clinical manifestations in the European cohorts of the SYROCOT study group report was 19% (105 of 550 CT cases) in infants followed up to 1 year of age. (The SYROCOT study also included non-European cohorts from the United States, Brazil, and Colombia, but these were excluded from the analysis of the effect of prenatal treatment on clinical manifestations.) For comparison, the overall risk for “any” clinical manifestations across the 26 cohorts was 24%, and in the 2 Brazilian and the 1 Colombian cohorts the overall risk was 38% to 77% (3 of 8 CT cases, 17 of 22 CT cases, and 3 of 8 CT cases, respectively).¹ Reported rates in the SYROCOT study might have underestimated the true rates of clinical manifestations because of the relatively short period of follow-up (up to 1 year of age).

In Lyon, France, only approximately 20% of infants with CT were symptomatic at birth (46 of 207 CT cases) during the study period 1996–2008.³ These results were consistent with earlier reports from European cohorts. Berrebi et al,¹³¹ in their 20-year prospective study from Toulouse, France (1985–2005), reported that only 27% (29 of 107) of infants with CT were found to be symptomatic during a median follow-up of 8 years (range: 1–20 years).

The reported rates of eye disease and intracranial calcifications during the first year of life in 22 European cohorts of SYROCOT were low (14% and 9%, respectively).^1,124 (The eye disease and neurologic outcomes of CT are discussed separately.) Other severe manifestations, such as hydrocephalus, splenomegaly, and pneumonia, have rarely been reported.¹³² 

The odds of developing “any” clinical manifestation (chorioretinitis and/or intracranial lesions) in the SYROCOT study decreased by 4% per additional week of gestation at seroconversion (OR: 0.96; 95% CI: 0.93–0.99 per week). The decrease in the incidence of eye disease and intracranial calcifications according to gestational age at seroconversion is shown in Figs 5, 6, and 7.

In the Dunn et al⁹⁶ study from Lyon, France, the risk of symptomatic CT decreased with increasing gestational age at maternal seroconversion (at 13 weeks: 61%; 95% CI: 34%–85%; at 26 weeks: 25%; 95% CI: 18%–33%; at 36 weeks: 9%; 95% CI: 4%–17%). If the gestational age at the time of acute primary maternal infection was known, then the risk of delivering a symptomatic infant could be calculated by multiplying the risk of MTCT for that gestational age by the risk of symptomatic disease for that gestational age. For example, for acute primary maternal infections (seroconversion) at 26 weeks of gestation, there would be a 40% risk of MTCT and a 25% risk of symptomatic disease in the infant, which would translate to a 10% risk (0.40 × 0.25) of symptomatic CT in the infant.⁹⁶ In this study,⁹⁶ the maximum risk of delivering a symptomatic infant was considered to be 10% (8%–14%), which occurred at approximately 24 to 30 weeks of gestation. It is important to note that all of these estimates of MTCT risks and risk of symptomatic CT in the infant were based on a maternal cohort in which approximately 95% of women received antepartum anti-Toxoplasma treatment during pregnancy, so these numbers would presumably underestimate the true risks in the United States.

A recent analysis of earlier data from France focusing on infants whose mothers were infected in the first trimester revealed favorable outcomes for infants whose mothers had routine antepartum screening and treatment.¹³³ Among 36 such infants with normal fetal ultrasonography and with antepartum and postnatal treatment, 97% were asymptomatic or mildly symptomatic (without major vision loss) at up to 4 years of age (range of follow-up: 1–12 years).¹³³ 

For European cohorts participating in the SYROCOT international consortium, the risk of eye disease diagnosed in the first year of life was 14% (79 of 550 CT cases).¹ For comparison, across the 26 cohorts, the risk was 18% (125 of 524), and for the South American cohorts, the risk was 47% (18 of 38 CT cases).¹ 

Data from the EMSCOT¹³⁴ revealed that 12% of children with CT (33 of 281) had their eye disease diagnosed during infancy, and by 3 years of age 17% (49 of 281) of children with CT had at least 1 chorioretinal lesion. Two-thirds of children with eye disease had a lesion at the posterior pole, and almost 50% of children with posterior pole lesions had visual impairment in the affected eye; this condition occurred in only 17% of those with peripheral lesions.

Visual impairment in the most affected eye has been reported to occur in 29% (20 of 69) of children with CT who have chorioretinitis,³⁸ but severe bilateral impairment was rare and occurred in only 9% of children with CT.^134,135 However, according to the EMSCOT, more than 90% of children with chorioretinitis had normal vision in their best eye.¹³⁴ 

Data from the long-term follow-up cohort study from Toulouse¹³¹ indicated that chorioretinitis occurred in 26% of CT cases (28 of 107) when follow-up was extended to 20 years (median follow-up: 8 years; range: 1–20 years). The location of the lesions was unilateral peripheral in 71%, bilateral peripheral in 7%, and unilateral macular in 21%. Results were similar in an earlier report by the same team.¹³³ 

Another recent cumulative report from the Marseille cohort (1995–2010), with a median follow-up of 4 years (maximum follow-up: 12 years), revealed a cumulative prevalence of eye disease of 19% (24 of 127 CT cases); visual impairment occurred in only 6% (8 of 127) of CT cases.¹³⁶ 

A French CT cohort from Lyon and Marseille⁹⁸ with 335 CT cases (1996–2002) revealed that chorioretinitis was diagnosed at birth in approximately 2% of CT cases, at 6 months in approximately 5% of CT cases, at 12 months in approximately 10% of CT cases, and at 24 months in approximately 12% of CT cases (Fig 6). A delay of more than 8 weeks between maternal seroconversion and initiation of antepartum treatment and the presence of intracranial calcifications was significantly associated with an increased risk of a chorioretinitis before 2 years of age.⁹⁹ Freeman et al,¹³⁷ in the EMSCOT, found that nearly 50% of children with only eye disease (22 of 50) had their first eye lesion detected before 4 months of age, and nearly 80% of children with additional clinical manifestations of CT had their eye disease diagnosed before 4 months of age.

The risk of eye disease increases from 10% early in infancy, to 16% by 4 years of age,^135,137,138 to 20% by school age,¹³⁷ and to 30% by 12 years of age.^98,135 In a large French cohort study by Berrebi et al¹³¹ with up to 20 years of follow-up, 39% of children with chorioretinitis (11 of 28 chorioretinitis cases among 107 CT cases) had their eye disease diagnosed at birth, 85% (24 of 28) had eye disease diagnosed before 5 years of age, and 96% (27 of 28) had eye disease diagnosed before 10 years of age.¹³¹ 

Wallon et al¹³⁹ described the long-term follow-up of 477 infants with CT from the Lyon cohort (1987–2008), with 81.5% of infants having mothers who received antepartum treatment. Almost all infants (99.8%) received postnatal treatment for 12 months.

Over a period of 12 years, approximately 30% of treated children experienced at least 1 recurrence of eye disease or developed a new chorioretinal lesion.¹³⁹ This proportion increased to almost 50% when the follow-up was extended to 18 years of age. The eye lesions were first seen at a median age of 3 years (range: 0.0–20.7 years). Lesions were mainly unilateral in 69% of cases and did not cause visual loss in 81% of cases. This promising long-term prognosis for ocular disease in children with CT might not apply to children with CT in the United States, where different T gondii strains are implicated and there is no routine antepartum screening/treatment.

The probability of recurrent chorioretinitis by 4 years of age increased significantly when the first episode of chorioretinitis was diagnosed before 4 months of age (probability: 30%; 95% CI: 0%–66%) and when additional clinical manifestations also were present before 4 months of age (probability: 70%; 95% CI: 31%–98%).¹³⁷ 

Although chorioretinitis could occur at any time in the child’s life, children who did not have chorioretinitis before 4 months of age were at low risk of developing eye lesions by 4 years of age.¹³⁷ In children with CT without chorioretinitis before 4 months of age, the probability of developing chorioretinitis by age 4 years was low (8%) whether or not additional clinical manifestations were present during early infancy. Most children with CT (92% [230 of 249]) in the EMSCOT cohort did not have eye disease diagnosed before 4 months of age.¹³⁷ 

The risk of chorioretinitis decreased by 3% for any additional week of gestation when maternal primary infection occurred (OR: 0.97; 95% CI: 0.93–1.00 per week; P = .04).¹ One study estimated that the risk of chorioretinitis was 2.1 times higher if the maternal primary infection occurred before compared with after 20 weeks of gestation.¹³⁷ Delay in maternal treatment of more than 8 weeks after maternal seroconversion and the presence of intracranial calcifications at birth were independent predictors of chorioretinitis (HRs [95% CIs]: 2.54 [1.14–5.65] and 4.3 [1.9–10], respectively).⁹⁹ The risk of chorioretinitis was 3.6 times higher if the infant with CT had additional clinical manifestations of CT at birth (ie, head ultrasonographic abnormalities, serious neurologic findings, lymphadenopathy, or hepatosplenomegaly).¹³⁷ 

Long-term follow-up by survey questionnaire of 126 adults with CT from the Lyon cohort (follow-up range: 18–31 years) revealed that treated CT had little effect on quality of life and visual function.¹⁴⁰ 

There are limited data from European cohorts on the incremental benefit of postnatal treatment on the risk of eye disease. The EMSCOT study¹³⁷ did not find an increased risk of eye disease with delayed initiation of infant treatment. However, the study was underpowered to detect such a difference, and moreover, all infants had already received the benefit of maternal treatment during pregnancy. The early 2005 EUROTOXO systematic review of studies evaluating postnatal treatment effect did not identify good-quality published studies to address this question.¹² 

The overall risk of intracranial lesions diagnosed during the first year of life for the European cohorts in the SYROCOT study was only 9% (49 of 550).¹ For comparison, across the 26 cohorts in the SYROCOT study, the risk of intracranial lesions diagnosed during the first year of life was 13% (88 of 691), and for the Brazil/Colombian cohorts the risk was 53% (20 of 38). For the European SYROCOT cohort, the risk of intracranial lesions decreased by 9% for any additional week of gestation at maternal primary infection (OR: 0.91; 95% CI: 0.87–0.95¹; Fig 7).

First, the most important difference is likely the fact that there were no routine antepartum screening/treatment programs in the United States and South America, so infants did not receive the benefit of antepartum therapy. Initiation of postnatal therapy could also be delayed for weeks to months. The use of routine antepartum screening/treatment programs best explains the decrease in clinical severity in European cohorts over time. Second, in Western Europe, the predominant strain was type 2,²⁸ which seems to be the least virulent,^128,141,142 whereas in North America, all of the 3 main types (1, 2, and 3) were encountered, in addition to the newly identified more virulent type 12 strains.^29,143 In South America, type 1 and type 3 strains were the predominant strains, type 2 strains were rarely found, and atypical, more virulent strains were also detected in South America^28,30,31 and appeared to be responsible for the higher incidence of severe ocular disease in children with CT.³⁸ Atypical strains have also been implicated in severe CT cases in South America and in Europe.^46,144 Genotyping data from Central America were more limited, with type 1 and type 1–related strains being detected in cases of CT¹⁴⁵; type 1 and type 3 strains found in chickens,¹⁴⁶ and atypical strains found in wild animals.¹⁴⁵ Third, although there has been evidence of genetic predisposition for symptomatic CT,^147,148 it is not known whether allelic differences between European and certain US populations could also account for host-pathogen interaction differences that could lead to detected differences in disease severity. Fourth, possible differences in gestational age at the time of acute primary maternal infections between the United States and Europe could contribute to differences in disease severity. Although the time of acute primary maternal infections in the United States is not known for the majority of CT cases,¹²⁴ recent data from France showed that there were differences in the trimester during which acute primary infections were diagnosed (42% [684 of 1624] diagnosed during the first trimester, 31% [497 of 1624] diagnosed during the second trimester, and 27% [443 of 1624] diagnosed during the third trimester).³ 

A recent study^38,149 of 281 CT cases from the EMSCOT and 30 cases from Brazil showed that T gondii causes more severe ocular disease in Brazil. Children in Brazil (1) developed chorioretinitis more frequently than children in Europe by their first year of life (50% [15 of 30 CT cases] versus 10% [29 of 281 CT cases]), (2) had a fivefold higher risk of chorioretinitis than in Europe by 4 years of age, (3) more frequently had multiple eye lesions, and (4) more frequently had larger lesions and lesions that were more likely to be located in the posterior pole and cause visual impairment and threaten vision than in Europe (visual impairment occurred in 87% of cases in Brazil versus 29% in Europe).³⁸ 

The observed differences in the frequency of chorioretinitis and in the number and size of the ocular lesions could be attributed to the lack of antepartum treatment and the higher frequency of more virulent strains infecting children in South America when compared with Europe.³⁸ It is important for US physicians to appreciate that there are geographic differences in the spectrum and severity of clinical manifestations between geographic areas, especially when evaluating children from such international settings.

In a systematic review by Brown et al,¹⁵⁰ the prevalence of sensorineural hearing loss (SNHL) associated with CT ranged from 0% to 26%. In subgroup analysis of children who received no postnatal treatment or only limited treatment (less than 1 year), the prevalence of SNHL was 28% (8 of 29 CT cases), whereas in children who were prescribed 12 months of postnatal therapy but in whom treatment was not confirmed to have started before 2.5 months of age, the prevalence of SNHL was 12% (2 of 17 CT cases). In children treated for 12 months postnatally and with therapy initiated before 2.5 months of age, the prevalence of SNHL was 0% (0 of 69 CT cases).¹⁵⁰ 

The largest study in US children who received treatment of 12 months suggested that SNHL was a rare occurrence in treated children.¹¹² Nevertheless, it has been suggested that children with CT who received anti-Toxoplasma therapy initiated before 2.5 months of age for a total duration of 12 months may need to have follow-up audiometric evaluation at 24 to 30 months of age. In contrast, children who had received no treatment, received partial treatment, or had delayed onset of treatment (eg, after 2.5 months of age) may need to undergo annual audiologic monitoring until they are able to reliably self-report hearing loss.¹⁵⁰ A recent otopathologic evaluation of temporal bones from 3 infants who died of CT identified T gondii in the dormant cystic form as well in the tachyzoite form with a surrounding inflammatory response, suggesting that the mechanism of hearing loss seen in CT can be sensory, neural, or sensorineural¹⁵¹ and is directly mediated by the parasite and host immune response.

Laboratory tools for the diagnosis of T gondii infection include serologic tests, such as the Toxoplasma IgG, IgM, IgA, IgE, IgG-avidity, and differential agglutination (AC/HS) tests; PCR assays; histologic and cytologic examination of tissue and body fluids; and attempts to isolate the parasite with mice subinoculation (this test is only performed in reference laboratories). Commercial laboratories in the United States usually offer Toxoplasma IgG and IgM testing; some of them also offer Toxoplasma IgA and IgG-avidity tests (FDA approved since 2011) and PCR assays. In reference laboratories for toxoplasmosis (eg, http://www.pamf.org/serology/), panels of serologic tests are offered that include different combinations of all of the aforementioned serologic tests (according to the clinical indication), and in addition, they include the IgG dye test (the gold standard for the detection of IgG, a sensitive and specific neutralization test in which live tachyzoites are lysed in the presence of the patient’s IgG T gondii–specific antibodies and complement), the AC/HS differential agglutination test (which uses 2 antigen preparations that express antigenic determinants found early after acute infection [AC] or in the later stages of infection [HS]; AC/HS ratios are interpreted as acute, equivocal, nonacute patterns of reactivity or nonreactivity), and well-standardized T gondii PCR assays. In consultation with a medical expert, the toxoplasmosis reference laboratory in the United States offers appropriate interpretations of serologic and PCR test results for each case.

Classification systems for T gondii infection in pregnant women used in Europe¹⁵² differ from those used in the United States. In Europe, classification is based on routine antepartum screening (every month [eg, France and Italy] or every 2–3 months [eg, Austria, Lithuania, Slovenia, and Germany¹⁰]) throughout pregnancy. Frequent screening allows for the identification of previously seronegative pregnant women who seroconvert during pregnancy.¹⁵³ In contrast, in the United States, the interpretation of maternal infection status has typically been based on testing at a single time point (performed usually because of fetal ultrasonographic abnormalities or maternal signs or symptoms suggestive of toxoplasmosis). The classification system for T gondii infections that has been used by the PAMF-TSL for maternal, fetal, and infant infectious status is depicted in Table 4.

When there is clinical suspicion of acute toxoplasmosis during pregnancy (eg, in the setting of certain ultrasonographic abnormalities or a high clinical suspicion of acute infection on the basis of symptoms), samples should be sent to a reference laboratory for toxoplasmosis to avoid any unnecessary delays in the establishment of diagnosis and initiation of prenatal treatment. Prompt initiation of prenatal treatment as soon as possible after acute maternal infection has been recently shown in observational studies to decrease MTCT and ameliorate the severity of clinical manifestations.

There is variation in the diagnostic performance of serologic screening tests for toxoplasmosis in commercial versus reference laboratories. Calderaro et al¹⁶² evaluated the diagnostic performance of 4 commercially available tests for toxoplasmosis. For all 4 commercial IgG tests, the sensitivity (point estimates) was 100% and the specificity (point estimates) of these different IgG tests ranged between 97% and 100%. The sensitivity (point estimates) of the different commercial IgM tests ranged between 82.3% and 100%, and the specificity (point estimates) of these different IgM tests ranged between 99.7% and 100%. There were commercially available tests that had 100% sensitivity and 100% specificity for both of their Toxoplasma IgG and IgM tests (eg, the Vidas Biomerieux). In this validation study, the “status” of samples as positive or negative was defined on the basis of concordant results obtained by at least 3 of the 4 assays compared in this study. Moreover, according to Wilson et al,¹⁶³ who performed a study sponsored by the FDA for the validation of 6 commercial IgM tests, the sensitivity (point estimates) of these 6 different IgM tests ranged between 93.3% and 100% and the specificity (point estimates) ranged between 77.5% and 98.6%. In this validation study, the reference gold standard for the Toxoplasma IgM was the double-sandwich enzyme-linked immunosorbent assay (ELISA) of the National Reference Laboratory for Toxoplasmosis (the PAMF-TSL). In an area of low seroprevalence, tests with a high specificity (but <100%) could lead to a large number of false-positive results. In reference laboratories, the specificity of the confirmatory testing (for the diagnosis of a recently acquired acute infection) is 100%.¹⁶³ 

These results generally exclude maternal T gondii infection. It is useful to order Toxoplasma IgM tests concomitant with IgG tests. It is possible to have a negative Toxoplasma IgG test result and a positive Toxoplasma IgM test result when testing is performed shortly after acute primary infection (eg, fewer than 2 weeks from time of infection).

The pregnancy-specific serologic panels of tests for a gestational age less than 16 weeks versus more than 16 weeks used by PAMF-TSL¹⁶⁴ help estimate the most likely time of maternal infection and thus differentiate chronic infections (acquired in the distant past, before conception) from acute infections. The majority of the tests in those panels are not available in nonreference laboratories, including the Toxoplasma IgG dye test, the Toxoplasma IgM immunosorbent agglutination assay (ISAGA), the Toxoplasma IgE ELISA, and the AC/HS differential agglutination test. Recently, the Toxoplasma IgG avidity test was approved by the FDA and currently also is performed by some nonreference laboratories.¹⁶⁵ However, the Toxoplasma IgG avidity test alone cannot differentiate between an acute and chronic infection, because there are cases in which low Toxoplasma IgG avidity can persist for years after primary infection. In such cases, a comprehensive panel of tests generally is necessary.

For pregnant women at ≤16 weeks of gestation, the PAMF-TSL panel contains the Toxoplasma IgG dye test, the Toxoplasma IgM ELISA, and the Toxoplasma IgG-avidity test. (Referring physicians can obtain the specific laboratory request forms online from the PAMF-TSL Web site at: http://www.pamf.org/serology/SpecimenRequirements.pdf). They can also contact the laboratory directly (telephone: 650-853-4828; fax: 650-614-3292; e-mail: toxolab@pamf.org) regarding the specimens to be tested (volume, shipping temperature, turnaround times, etc) (Tables 5, 6, 7).

For women at a gestational age of >16 weeks, the PAMF-TSL panel includes the Toxoplasma IgG dye test, the Toxoplasma IgM ELISA, and the AC/HS differential agglutination test (an acute AC/HS indicates an infection acquired <12 months from the time of testing, and a nonacute AC/HS is suggestive of a chronic infection acquired >12 months from the time of testing).

Toxoplasma IgG avidity is included in the routine panel only for women at ≤16 weeks of gestation, because a high avidity can exclude an acute infection acquired during pregnancy or close to conception. However, a high avidity in a woman at >16 weeks of gestation may not exclude an infection acquired early in pregnancy or very close to conception (IgG avidity becomes high 16 weeks after a primary infection).

In certain situations, additional reflex testing sometimes is needed. Reflex testing includes the Toxoplasma IgA ELISA, the Toxoplasma IgE ELISA, and the Toxoplasma IgG-avidity test, even for women at >16 weeks of gestation. A guide for the interpretation of the individual serologic test results performed at PAMF-TSL can be found on the PAMF-TSL Web site (http://www.pamf.org/serology/clinicianguide.html).^164,166 

In general, such results are confirmed in a reference laboratory for toxoplasmosis.¹⁶⁶ A positive Toxoplasma IgM could indicate either a very recently acquired infection or a false-positive IgM result.¹⁶⁷ Additional tests performed in a reference laboratory¹⁶⁶ and follow-up testing within 2 to 3 weeks will help differentiate between these 2 possibilities. If the Toxoplasma IgG test result becomes positive during follow-up, then this finding would be consistent with seroconversion and of a very recently acquired infection during pregnancy. If the Toxoplasma IgG test result remains negative, then the Toxoplasma IgM result probably was a false-positive result.

These results obtained during the first half of the pregnancy (<20 weeks’ gestation) could exclude, in most patients, an infection acquired during pregnancy, particularly if the Toxoplasma IgG titers are “low” (eg, IgG dye test titers ≤512 [value refers to a dilution of 1:512]). However, for results at ≥20 weeks’ gestation or when there is high clinical suspicion of CT (ie, abnormal fetal ultrasonographic findings), these results should be confirmed in a reference laboratory for toxoplasmosis¹⁶⁶ with an additional panel of tests according to gestational age.

In addition, more recently, there have been some reports of atypical seroconversion with negative or transient IgM responses, but the clinical significance and how common this phenomenon is remain unclear.¹⁶⁸ (Among 4500 documented maternal seroconversions across 12 centers in France, negative IgM or only transient IgM responses were reported to occur in 0.58% of cases [n = 26 of 4500] and this was considered to represent an atypical seroconversion serologic profile.¹⁶⁸) Even though these are rare cases, they could be associated with significant clinical implications if not appropriately managed.

In general, such results are confirmed in a reference laboratory before being interpreted as evidence of acute primary infection. Approximately 60% of cases with positive Toxoplasma IgM results from nonreference laboratories were not associated with recently acquired infections when tested at the PAMF-TSL with the use of the comprehensive pregnancy panel of tests.^103,154 These positive Toxoplasma IgM assays represented either false-positive IgM assays (20%) or positive IgM assays in the context of chronic infections (40%).^167,169 The persistence of IgM antibodies in low titers beyond 1 year is not uncommon.¹⁵⁴ In those cases, reflex testing with additional tests at a toxoplasmosis reference laboratory will help confirm or exclude that possibility.

CT can occur in 3 ways: (1) transmission of T gondii infection to the fetus from a previously seronegative, immunocompetent mother who acquired acute primary infection during pregnancy or shortly before conception (even within 3 months before conception)¹⁵⁶; (2) reactivation of a Toxoplasma infection in previously T gondii–immune women who have been immunocompromised during gestation (eg, women with altered immune function attributable to HIV infection or immunosuppressive medications); and (3) reinfection of a previously immune mother with a new, more virulent strain¹⁴⁴ (eg, after international travel or after eating undercooked meat from areas where more virulent atypical strains predominate^{29,33,141,142,144}).

AF PCR assay has many advantages as a method of diagnosis of fetal infection. The evidence for the diagnostic performance of AF PCR assay is summarized in Table 8, according to the results of the most recent studies published over the past 10 years.^114,170,171 The clinical value of AF PCR assay results in estimating a woman’s risk of delivering an infant with CT varied according to gestational age at maternal seroconversion. Figure 8 shows how the probability of fetal infection (posttest probability) changed (as compared with the pretest probability) after a positive or a negative AF PCR assay result according to the gestational age at which the maternal primary infection most likely had occurred.

A negative AF PCR assay result for maternal primary infections thought to be acquired in the first or second trimester may be reassuring with regard to excluding fetal infection. The negative predictive value (NPV) of AF PCR assay for such early maternal primary infections was very high (Table 8). On the contrary, for infections thought to be acquired in the third trimester, the NPV was lower, and a negative AF PCR assay result in such cases should be interpreted cautiously and would not exclude fetal infection (Fig 8).^171,172 (The exact reason for this situation is not known. Some possible explanations could be a dilution effect attributable to a larger amount of AF in the third trimester and/or maternal treatment before AF PCR testing [eg, in cases in which P/S was initiated immediately after the diagnosis and before the AF was tested].)^114,170,171 Moreover, it probably takes several weeks for the infection to cross the placenta from the mother to the fetus, in large enough quantities to spill into the AF. T gondii likely first infects the placenta, replicates, then crosses to the fetus.

The positive predictive value (PPV) of an AF PCR assay result for maternal infections acquired at any time during pregnancy was very high (Table 8), and a positive AF PCR assay result would indicate CT and could be used to recommend treatment for fetal infection (assuming there is no laboratory contamination). (Estimates of the diagnostic performance of AF PCR assay always refer to the gestational age at which maternal infection was acquired and not to the time the amniocentesis was performed.)

In general, AF PCR assay should be performed at least 4 weeks after acute primary maternal infection and at ≥18 weeks of gestation. The majority of the available data on the diagnostic performance of an AF PCR assay came from women who had amniocentesis performed at ≥18 weeks of gestation, and there were no data regarding the interpretation of negative results of AF PCR assay performed at <18 weeks of gestation (false-negative results cannot be excluded, for example, because of the small amounts of AF that can be obtained at this gestational age).

Data from France revealed that only 28% of women infected at >24 weeks of gestation underwent amniocentesis,¹¹⁴ either because it was not offered to them (because of concern for preterm birth) or because of patient refusal. In such cases, it is possible that physicians elected to treat those women empirically with P/S because of the high risk of MTCT with T gondii infections during the third trimester of pregnancy.

All infants born to mothers with suspected/confirmed T gondii infection during pregnancy, even if they had negative AF PCR assay results, will need a complete clinical, radiologic, and laboratory diagnostic evaluation to rule out CT as well as serial serologic monitoring every 4 to 6 weeks during their first year of life to document disappearance of Toxoplasma IgG antibodies.

Quantitative AF PCR assay still is considered experimental and is not used routinely in commercial or even reference laboratories for the diagnosis of CT. In a retrospective study conducted in 88 consecutive AF samples that tested positive for T gondii by real-time quantitative PCR assay,¹⁷³ after adjusting for gestational age at maternal seroconversion, higher AF parasite concentrations were associated with more severely symptomatic CT cases (OR: 15.38 per parasite log [parasites per mL AF]; 95% CI: 2.45–97.7).¹⁷³ Pregnant women infected before 20 weeks of gestation with a parasite load ≥100/mL determined by AF PCR assay had the highest risk of giving birth to infants with severe CT.¹⁷³ 

In addition to maternal serologic testing and AF PCR assay, fetal ultrasonography also may help determine whether the fetus has been affected, although a normal fetal ultrasonogram cannot exclude a fetal infection. Fetal ultrasonographic findings reported to be associated with CT include ventriculomegaly, intracranial calcifications, increased placental thickness, intrahepatic calcifications, ascites, fetal growth restriction, and echogenic bowel. Of note, microcephaly has not been reported in fetal ultrasonographic findings of CT, possibly because of the associated ventriculomegaly.¹⁷⁴ Crino¹⁷⁴ reported a pooled sensitivity of 49% and a pooled specificity of 99% of any fetal ultrasonographic findings for CT on the basis of results from 4 studies.^{123,175,–177} 

In general, any maternal Toxoplasma IgG antibodies transferred transplacentally are expected to decrease by 50% per month after birth and usually disappear by 6 to 12 months of age.⁹⁸ Maternal treatment during pregnancy and/or postnatal treatment could affect the production and kinetics of Toxoplasma IgG antibodies in the infant. In infants with CT, postnatal therapy might decrease the Toxoplasma IgG titers to below detectable levels; however, rebound will occur after discontinuation of therapy.⁹⁸ Disappearance of Toxoplasma IgG antibodies in the presence of postnatal treatment cannot be used as an indication that the child does not have CT. (Berrebi et al¹³¹ used as a criterion the persistence of Toxoplasma IgG antibodies after 18 months of age.)

The disappearance of Toxoplasma IgG before 12 months of age in an infant capable of producing IgG antibodies and who has not received anti-Toxoplasma treatment essentially excludes the diagnosis of CT.

A positive Toxoplasma IgM ISAGA result (at or after 5 days of age) and/or a positive Toxoplasma IgA test result (at or after 10 days of age) is considered diagnostic of CT in infants with a positive Toxoplasma IgG. Because IgM and IgA antibodies do not cross the placenta, they reflect the infant’s response to T gondii infection. Of note, positive Toxoplasma IgM antibodies before 5 days of age or positive Toxoplasma IgA antibodies before 10 days of age also could represent false-positive results resulting from contamination of the infant’s blood with maternal blood because of a materno-fetal blood leak, and the tests should be repeated after 5 and 10 days, respectively. Moreover, false-positive Toxoplasma IgG and IgM test results can occur after recent transfusion of blood products or receipt of immune globulin intravenously, and they should be repeated at least 1 to 2 weeks after the last transfusion. However, consideration should be given to testing infants with high suspicion for CT as soon as possible after birth.

These neonatal Toxoplasma IgM and/or IgA tests could fail to identify approximately 20% to 50% of CT cases, with the highest false-negative rates being reported in Western European countries,¹⁷⁸ where infants have been exposed to antepartum treatment, and the lowest false-negative rates being reported in the United States, where the majority of infants with CT had mothers who did not receive antepartum treatment.¹²⁴ False-negative Toxoplasma IgM and Toxoplasma IgA test results in infants with CT could also occur in cases of fetal infection early in gestation. In those cases, the early transient positive Toxoplasma IgM and/or IgA antibody responses might have disappeared by the time of delivery.¹⁷⁸ Moreover, if maternal infection was acquired very late in pregnancy, initially negative Toxoplasma IgM ISAGA and IgA ELISA results at birth could be attributable to delayed production of those antibodies and would not exclude the diagnosis of CT. Repeat testing 2 to 4 weeks after birth and every 4 weeks thereafter until 3 months of age could be considered in such cases.

In the study by Olariu et al,¹²⁴ the reported rates of false-negative results in the United States for neonatal Toxoplasma IgM ISAGA, IgA ELISA, and IgM or IgA tests were much lower (false-negative IgM ISAGA results occurred in 13% of CT cases [22 of 164]; 95% CI: 9%–20%; false-negative IgA ELISA results occurred in 23% of CT cases [37 of 164]; 95% CI: 16%–30%; false-negative IgM ISAGA and IgA ELISA results occurred in 7% of CT cases [11 of 164]; 95% CI: 3%–12%) when infants were tested within the first 6 months of age.

According to the EUROTOXO 2006 systematic review of diagnostic tests, in 8 of 19 diagnostic studies that had the best study design and used appropriate reference standards for the ascertainment of the infant’s infection status,¹¹ the reported false-negative rates for neonatal Toxoplasma IgM or IgA tests in infants with CT (the majority of whose mothers received antepartum treatment) were 19% to 45% (when both tests were performed and if tests were performed within the first 2 weeks after birth) and 21% to 37% if performed within 2 weeks to 3 months after birth. The reported false-negative rates for the neonatal Toxoplasma IgA test were 29% to 48% if the test was performed within 2 weeks after birth and 40% to 43% if performed within 2 weeks to 3 months of age. The reported false-negative rates for neonatal Toxoplasma IgM ISAGA were 33% to 56% if performed within 2 weeks after birth and 29% to 39% if performed 2 weeks to 3 months after birth.^11,179 

The observed differences in the sensitivity of these tests between the European and US cohorts could be attributed to the presence of antepartum treatment in most CT cases in the European Union. Such treatment could have affected the production of Toxoplasma IgM and IgA in those infants. In the US cohort,¹²⁴ all infants had mothers who did not receive antepartum treatment. There was no statistically significant difference in the rates of false-negative results whether these were performed at or before 1 month of age versus at 1 to 6 months of age.¹²⁴ However, the number of cases was small.

The role of Toxoplasma IgE antibodies for routine screening of infants has been considered to be limited,^124,180,181 and currently, IgE antibodies are not included in the recommended infant diagnostic panel for CT at the PAMF-TSL.¹⁶⁴ 

In the largest cohort of CT cases without antepartum treatment in the United States¹²⁴ over the past 15 years (N = 164), results of PCR assays of peripheral blood, CSF, and urine were reported in 7, 58, and 10 infants, respectively. The sensitivity was 29% (95% CI: 4%–71%) for blood PCR assay, 46% (95% CI: 33%–60%) for CSF PCR assay, and 50% (95% CI: 19%–81%) for urine PCR assay.¹²⁴ The uncertainty around these estimates was large for blood and urine and is attributable to the small sample sizes.¹²⁴ The CSF PCR assay result was positive in 71% (22 of 31) of infants with CT and hydrocephalus, in 53% (23 of 43) of infants with intracranial calcifications, and in 51% (26 of 51) of infants with eye disease.¹⁸² CSF PCR assay had the potential to increase the frequency of cases in which the diagnosis of CT was confirmed (95% of CT cases [55 of 58] had a positive IgM ISAGA, a positive IgA ELISA, or a positive CSF PCR assay result), and CSF PCR assay was successful in detecting CT in 3 infants who had negative Toxoplasma IgM and IgA test results at birth (3 of 6 infants with CT and negative IgM ISAGA and IgA ELISA results had positive CSF PCR assay results).¹⁸² 

The 2 most recent studies on the diagnostic performance of placental PCR from European cohorts of women who had received antepartum treatment showed a sensitivity ranging between 71%¹⁸³ and 79%¹⁸⁴ and a specificity ranging between 92%¹⁸⁴ and 97%¹⁸³ (Robert-Gangneux et al¹⁸³ tested 102 placentas [28 CT cases]; Sterkers et al¹⁸⁴ tested 238 placentas [39 CT cases]). The reported PPV of placental PCR assay was 67% (95% CI: 54%–81%) and the NPV was 96% (95% CI: 93%–99%).¹⁸⁴ A positive placental PCR assay result may provide some evidence of CT but is not diagnostic of CT per se. Whole placentas of at least 200 g were used, and several samples were obtained for testing from different placental sites.¹⁸³ Robert-Gangneux et al also reported the diagnostic performance of placental testing by mice subinnoculation and reported a sensitivity of 67% and a specificity of 100%.¹⁸³ 

The diagnostic performance of placental PCR assay in the United States, where the majority of pregnant women with T gondii infection are not treated, is unknown. In general, for PCR assay, the specimen could be sent frozen, whereas for mice subinnoculation (ie, for isolation and serotyping of the T gondii strain), the specimen should not be frozen and should be sent at 4°C. Cord blood PCR assay does not appear to be clinically useful, because its sensitivity was shown to be very low (16% [4 positive cord blood PCR assay results/25 CT cases])¹⁷²; however, its specificity was 100%.¹⁷² 

This situation is rare, but these newborn infants would be considered to be infected and be treated until proven otherwise, assuming that other etiologies for the clinical manifestations have been reasonably excluded. In the Olariu et al¹²⁴ study, 7% (11 of 164) of infants with CT had negative Toxoplasma IgM and IgA test results. As mentioned previously, initially negative neonatal Toxoplasma IgM ISAGA and IgA ELISA results could be attributable to delayed production of those antibodies and would not exclude the diagnosis of CT. Repeat testing at 2 to 4 weeks after birth and every 4 weeks thereafter until 3 months of age could be considered in such cases.

A detailed depiction of the diagnostic workup of infants with suspected CT, according to the PAMF-TSL, is shown in Table 9. In addition, technical details for the preparation and shipping of clinical specimens to toxoplasmosis reference laboratories are provided in Tables 5–7.

The antepartum treatment algorithm (according to the PAMF-TSL) based on gestational age at which primary infection most likely occurred is shown in Fig 9 and Table 10 .

Consideration should be given to initiating treatment of the pregnant woman as soon as possible, either with spiramycin, if maternal primary infection was acquired close to conception and up to 18 weeks of gestation (the request for spiramycin should be made to the FDA; details regarding how to obtain spiramycin are provided in the footnote of Table 10), or with P/S and folinic acid, if maternal infection was acquired at >18 weeks of gestation. (Spiramycin does not readily cross the placenta and thus is not reliable for treatment of infection in the fetus; P/S crosses the placenta and, if fetal infection has occurred, can provide treatment of the fetus.¹⁵⁴)

In addition, an AF PCR assay should be performed, if feasible and safe, as soon as possible after 18 weeks of gestation and sent to a reference laboratory. The risk of complications for amniocentesis performed at >24 weeks of gestation should be taken into consideration. In all cases, treatment should be continued until delivery, but the AF PCR assay result will determine whether optimal treatment for the duration of pregnancy is spiramycin or P/S plus folinic acid. Monthly fetal ultrasonographic monitoring throughout pregnancy is indicated.

Spiramycin should be continued throughout pregnancy to prevent MTCT, even if the AF PCR assay result is negative, because the placenta still might be infected and fetal infection through an infected placenta can occur at any time during pregnancy.

Treatment with P/S and folonic acid should be initiated in the pregnant woman. If the pregnant woman already was receiving spiramycin, spyramycin should be stopped and treatment should be switched to P/S. A French randomized clinical trial (TOXOGEST) was launched to compare the efficacy of P/S versus spiramycin in preventing MTCT. Enrolled pregnant women are those at ≥14 weeks of gestation with confirmed acute primary infection during pregnancy. The anticipated date of completion was 2015, and the target sample size is 330 pregnant women. Secondary study endpoints include the safety of therapy and the severity of CT.²⁰³ 

The incidence of CT in the children of such women has been shown to be extremely rare unless the woman is severely immunocompromised (eg, advanced HIV disease, receiving corticosteroids or immunosuppressive drugs). Anti-Toxoplasma treatment is not indicated in this case unless the patient is immunocompromised. However, there are a few exceptions to this dictum, illustrated by case reports of presumed T gondii reinfections during pregnancy with a different T gondii strain in previously immune women with serologic evidence of chronic T gondii infection.^144,204 

There are a few published cases reports for HIV-infected, previously immune pregnant women with serologic evidence of chronic T gondii infection who transmitted the infection to their fetuses.^205,–207 This situation has been reported to occur in <4% of cases (in 0.09% [1 CT case in 1058 HIV-infected women] of HIV-infected women followed as part of the European Collaborative Study and Research Network on CT²⁰⁶; and in 0.5% [95% CI: 0.24%–0.91%; 10 CT cases in 2007 HIV-exposed infants] of HIV-exposed infants from Brazil between 1998 and 2011²⁰⁸; and in 3.7% [3 CT cases in 82 HIV-infected pregnant women] of HIV-infected pregnant women with chronic T gondii infection in Brazil²⁰⁷).

In this setting, the risk of MTCT appeared to be greater when the pregnant woman’s latent infection was reactivated with or without clinical manifestations (eg, Toxoplasma encephalitis, pneumonia, fever). Reactivation was more likely to occur if the maternal CD4+ T-lymphocyte count decreased below 200 cells/mm³ in the absence of effective anti-Toxoplasma prophylaxis. However, in HIV-exposed infants, CT has been documented to occur even when mothers were not severely immunosuppressed (eg, 6 of 10 CT cases in HIV-exposed infants occurred with maternal CD4+ T-lymphocyte counts >200/mm³, with 1 CT case occurring with a maternal CD4+ T-lymphocyte count >500/mm³).²⁰⁸ Cautious monitoring of HIV-infected pregnant women is indicated.

In the absence of clinical trial information, the following recommendations have been made by PAMF-TSL: (1) if the mother’s CD4+ T-lymphocyte count decreases below 200 cells per mm³ in the absence of clinical manifestations of toxoplasmosis, prophylaxis with agents known to be effective against Toxoplasma (eg, trimethoprim/sulfamethoxazole) should be instituted; (2) for mothers with clinical manifestations of toxoplasmosis, anti-Toxoplasma treatment with effective regimens (eg, P/S and folinic acid) should be initiated as soon as possible; and (3) if the mother’s CD4+ T-lymphocyte count remains at ≥200 cells per mm³, as a cautionary measure, spiramycin should be continued throughout pregnancy.

There are very few case reports of previously immune pregnant women with serologic evidence of chronic T gondii infection who delivered an infant with CT^151,155; this situation was presumably attributable to reinfection during pregnancy by a different T gondii strain (eg, in the case report from France by Elbez-Rubinstein et al,¹⁴⁴ the previously immune pregnant woman had ingested imported raw horse meat). This unusual situation might be considered when there is a strong clinical suspicion with recent international travel, but currently there are no routinely used laboratory tests that could help distinguish reinfections from a new T gondii strain from preexisting infection. If this clinical scenario is suspected, consultation with reference toxoplasmosis centers in the United States regarding treatment is encouraged.

When acute T gondii infection is suspected in a pregnant woman after international travel, it is important to remember that more virulent T gondii strains circulating outside the United States and Europe have been reported to cause more severe clinical manifestations in infants with CT.^38,155 

There is a large variation between the different postnatal treatment protocols followed in different centers in the United States and Europe, as shown in Table 11. In the United States, the postnatal treatment protocol that is used is the one proposed by the NCCCT study (Tables 10 and 11).¹¹² Currently, there is no evidence on the comparative efficacy or effectiveness of the postnatal treatment strategies, and the selection of different strategies relies on the experience of the different centers.

Treatment of infants with P/S was frequently associated with serious adverse events, reported to occur in 20% to 50% of cases.¹⁴ However, data from Denmark have shown that the medication was not well tolerated in only 14% (4 of 29) of treated infants.²⁰⁹ The main adverse effect was neutropenia, reported to occur more often with higher doses of P/S and especially when folinic acid was not administered.¹⁴ Moreover, seizures have been reported with cases of pyrimethamine overdose resulting from prescription dosing errors.²¹⁰ 

Adverse effects with pyrimethamine/sulfadoxine (Fansidar, F. Hoffman-LaRoche Ltd, Basel, Switzerland) were much less frequent, mainly skin rashes and urticarial reactions, and reported to occur in approximately 1% to 2% of cases.¹⁴ The disadvantage of the long half-life of sulfadoxine may need to be taken into account. Treatment discontinuations of pyrimethamine/sulfadoxine attributable to adverse events in adults with Toxoplasma chorioretinitis have been reported to occur in up to 26% of patients.²¹¹ The frequency of serious adverse events with pyrimethamine/sulfadoxine has been reported to occur in 1 in 2100 prescriptions, skin reactions have been reported to occur in 1 in 4900 prescriptions, and death has been reported to occur in 1 in 11 100 prescriptions.²¹² 

In French centers before 2001, infants were treated with P/S for the first 3 weeks of life, changed to spiramycin until 2 months of age, and then changed to pyrimethamine/sulfadoxine to complete 12 months of total therapy. However, more recently published protocols from several European centers did not include spiramycin in their postnatal treatment regimens, because spiramycin does not have CNS penetration and is a parasitostatic medication. In contrast, P/S penetrates the CNS and is parasitocidal.

Some centers have used a daily dosing scheme for P/S, whereas other centers have used pyrimethamine/sulfadoxine dosed every 10 to 15 days (because of the long half-life of sulfadoxine). In the latter regimen, the pyrimethamine dose was higher. In the United States, the initial daily dosing scheme was kept during the first 2 months (or 6 months [considered for symptomatic CT]) and was then transitioned to a 3-times/week dosing scheme in the following months.

The majority of centers treated infants with CT for 12 months, but some centers used longer treatment courses of up 24 months.¹³¹ Some centers in Germany individualized the duration of the postnatal treatment according to the severity of CT (eg, 3 months of total treatment of asymptomatic infants, 6 months for mildly symptomatic infants [ie, minor ventricular dilation or intracranial calcifications but with a normal neurologic examination and the presence of retinal scars but without active inflammation], and 12 months for severely symptomatic infants [seizures, abnormal neurologic examination, or chorioretinitis]).

The different postnatal monitoring protocols, according to the PAMF-TSL, the Toxoplasmosis Center at the University of Chicago, and European toxoplasmosis centers, are outlined in Table 12. There is significant variation across centers of the frequency of recommended clinical and serologic testing follow-up of infants with CT and the duration of follow-up.

The management of infants with confirmed or suspected CT, according to the Toxoplasmosis Center at the University of Chicago and the PAMF-TSL, is outlined in Tables 10–12.

There is no clear consensus on the management of asymptomatic infants in whom CT cannot be confirmed or ruled out. Some experts in France would elect not to treat infants who have no proof of infection at birth: cases with documented maternal seroconversion during pregnancy but with a negative evaluation (negative AF PCR assay result, normal fetal ultrasonogram, normal clinical and ophthalmologic evaluation at birth; normal imaging evaluation at birth; and no detection of Toxoplasma IgM and IgA at birth). Instead, they elect to simply follow those infants serologically every 2 months until complete disappearance of the Toxoplasma IgG antibodies (that were presumably transplacentally transferred maternal Toxoplasma IgG antibodies). However, if follow-up serologic testing indicates congenital infection, these infants generally require treatment. A rationale for deferring treatment in those infants is that those infants had already received the benefit of antepartum treatment.

Cautious use of this “defer treatment” strategy for newborn infants in the United States is advised, because the majority of these infants’ mothers did not receive antepartum treatment. According to the PAMF-TSL, all of the following criteria should be fulfilled before deciding to defer postnatal treatment of an infant in the United States: (1) acute primary maternal infection that was most likely acquired in the first or second trimester of pregnancy, (2) negative AF PCR assay result at ≥18 weeks, (3) no laboratory evidence of CT in the newborn (positive neonatal Toxoplasma IgG test result, at lower titers than maternal IgG titers, but negative Toxoplasma IgM and IgA test results), and (4) no clinical signs or symptoms of CT at birth after a complete clinical and radiologic evaluation (head ultrasonography may be used for CNS imaging in such infants).

The NPV of a negative AF PCR assay result for infections acquired in the first and second trimester is nearly 100% and, when all the aforementioned criteria are fulfilled, the risk of fetal T gondii infection is nearly zero. However, when maternal primary infection has occurred in the third trimester, even a negative AF PCR assay result cannot completely exclude the possibility of fetal infection. The risk of MTCT and fetal infection increases steeply with advancing gestational age; therefore, the decision to defer postnatal treatment in those settings should be carefully discussed with the parents.

A decrease in neonatal Toxoplasma IgG titers below the detection level (during serial follow-up after birth every 4 to 6 weeks) can exclude CT in infants who have not received postnatal treatment. However, a decrease in titers below detection during postnatal therapy cannot be used as an indication that the infant was not congenitally infected, because antibody titers may become negative during therapy but can rebound after therapy discontinuation in infants with CT.⁹⁸ 

In a study by Boyer et al,²⁵ only 48% of mothers of infants with CT had clinical symptoms suggestive of acute toxoplasmosis during pregnancy or had reported risk factors for T gondii exposure (eg, exposure to undercooked meat or to cat feces). Only routine serologic screening during pregnancy would have identified the rest of the women, who were nevertheless at risk of delivering infected infants.

In addition, only 49% of mothers of infants with CT who had serologically documented evidence of acute T gondii infection acquired from oocysts had also reported that they had significant risk factors for such exposures.²⁴ (Of note, laboratory tests to detect antibodies to sporozoite-specific antigens are performed only in research laboratories.⁵⁵)

Among Western European countries, only France³ and Austria²¹³ have implemented long-standing free national routine antepartum screening programs (monthly screening until the end of pregnancy in France since 1978¹¹³ and once per trimester in Austria since 1975,^214,215 although shorter intervals [maximum every 8 weeks] were later advocated by experts). Screening once per trimester leaves a “blind period” during maternal infections late in pregnancy.²¹⁶ 

Their approach was followed by other countries, including Italy and Slovenia. In 1998, Italy passed legislation that requires screening for toxoplasmosis of all pregnant women. This program was implemented mostly in the northern parts of Italy, and according to the latest Italian recommendations, screening should begin by 13 weeks of gestation and continue monthly until the end of pregnancy.²¹⁷ In Switzerland, screening was applied only regionally in the Basel region and only for surveillance purposes.^214,218 In many other European countries, antepartum screening was widely adopted by experts and physicians despite the absence of national mandatory antepartum screening programs; this phenomenon has been termed “wild screening.”²¹⁴ 

Neither France nor Austria complemented their antepartum screening programs with additional postnatal surveillance programs for CT for approximately 3 decades. A recent European survey of 28 European countries in 2007 identified only 4 countries that had surveillance screening programs specifically for CT (France, Germany, Italy, and Denmark).¹⁰ Twelve other countries had surveillance systems only for symptomatic toxoplasmosis (CT or noncongenital toxoplasmosis), and another 12 did not have any surveillance systems for toxoplasmosis. The neonatal screening program for CT in France started only in 2007. In Italy, neonatal screening for CT is only regionally applied to the Campania region (since 1997). Denmark¹¹⁶ had a national neonatal screening program since 1997, but this program was discontinued in August of 2007 because it was thought that there was no benefit from postnatal treatment, because some infants with CT developed new eye lesions despite postnatal treatment (3 new eye lesions developed during a 3-year follow-up period among 32 infants evaluated²¹⁹). However, children with CT in this program were treated for a very short period of time (3 months) as opposed to the longer 12-month postnatal treatment courses proposed in other European countries (Table 11). In 2007, a dedicated surveillance network for CT also was instituted in Greece.²²⁰ 

The preventive effect of toxoplasmosis screening of pregnant women depends on both the magnitude of disease caused by CT (the incidence of maternal infection during pregnancy multiplied by the risk of MTCT multiplied by the proportion of symptomatic infected children) and the preventable proportion of disease (the sensitivity of the screening strategy multiplied by the efficacy of the preventive treatment multiplied by the adherence to therapy).

Early studies evaluating the cost-effectiveness of antepartum toxoplasmosis screening programs yielded conflicting results.^8,221 The 1999 study by Mittendorf et al²²¹ concluded that antepartum screening was not cost-effective, whereas the most recent 2011 decision analysis by Stillwaggon et al⁸ concluded that implementation in the United States of a universal monthly screening of pregnant women, following the French protocol, is cost-effective and leads to a savings of $620 per screened infant if the incidence of CT in the United States is greater than 1 per 10 000 live births. This decision analysis made a number of assumptions, including a cost of $12 per test, an estimated cost of fetal death of over $6 million, and an incidence of acute primary maternal infection during pregnancy of 1 in 1000 (including also sensitivity analysis for an incidence of acute maternal infections as low as 0.2 cases/1000 pregnant women). It also assumed that treatment was highly efficacious and inexpensive. Although the study concluded that screening in the United States would be cost-effective, it remains unclear whether these conclusions would be reached if data were used assuming higher costs of screening, lower costs of loss, and less efficacy of treatment. Previous decision analyses have not arrived at the same conclusions.⁹ However, in this previous decision analysis, among the strategies analyzed was pregnancy termination for documented fetal infections. The early Cochrane systematic review on treatments for toxoplasmosis during pregnancy published in 2000 suggested that, in countries where screening or treatment is not routine, these technologies should not be introduced outside the context of a carefully controlled trial.⁶ 

Several assumptions made in the earlier decision analysis by Mittendorf et al²²¹ differ from current evidence. Mittendorf et al²²¹ assumed that antepartum screening tests had a very low specificity, which translated to a very low PPV for a relatively rare disease like CT. This assumption was incorrect for samples sent for further confirmatory testing at the PAMF-TSL,¹⁶⁴ because confirmatory testing at the PAMF-TSL has been shown to have a specificity of 100% for the differentiation of acute from chronic T gondii infections.^162,222,223 The diagnostic tests used at the PAMF-TSL have been extensively cross-validated against national reference laboratories in France by using sera from pregnant women with complete ascertainment of their T gondii infection status. Furthermore, AF PCR assay has a specificity and a PPV of nearly 100%,^114,172 which means that a positive AF PCR assay result is diagnostic of fetal infection. In addition, in the Mittendorf et al study, pregnancy termination rates were assumed to be very high (they assumed that 12.1 fetuses without CT would be aborted for every fetus truly diagnosed with CT).²²¹ According to recent evidence, pregnancy termination would not be routinely performed once the diagnosis of CT is made.¹³³ Confirmatory serologic testing in a reference laboratory and communication and interpretation of the results by an expert decreased the rate of unnecessary abortions by approximately 50%.¹⁶⁹ AF PCR assay and fetal ultrasonography to confirm or exclude fetal infection and symptomatic CT may further decrease the number of elective abortions. Further support against pregnancy terminations for CT is provided by the analysis of Berrebi et al¹³³; even among fetuses infected in the first trimester and with normal fetal ultrasonograms, 97% (35 of 36) would be either completely asymptomatic (78% [28 of 36]) or only slightly affected at birth (19% [7 of 36] had chorioretinitis and/or moderate ventricular dilatation). Of note, all of these fetuses had mothers who received antepartum treatment. The updated review of the French experience over the past 2 decades (1987–2008) described a fetal loss rate of 0.9% (N = 18) and a pregnancy termination rate of 1% (N = 21).³ Moreover, in the SYROCOT international consortium of CT, the reported fetal death rate was similar (∼2%; 22 terminations of pregnancy [1%] and 13 fetal deaths [0.7%] among the 1745 pregnant women with primary T gondii infections from the 26 international cohorts).¹ 

In the more recent 2011 Stillwaggon et al⁸ study, universal antepartum screening in the United States following the French protocol of monthly serologic screening during pregnancy (started at 11 weeks of gestation), antepartum treatment, fetal ultrasonography/AF PCR assay, and postnatal serologic and clinical follow-up of infants and treatment of infants as indicated was cost-saving. The findings from this analysis were robust to changes in the incidence of acute maternal primary infections, value of life, and test costs. Specifically, the screening approach included initial screening with commercial tests for IgG and IgM (at a cost of $12 per test) and subsequent confirmatory testing in reference laboratories of any positive results by using a panel of tests with high diagnostic accuracy and a specificity of 100%¹⁶³ (at a cost of $385 per confirmatory panel of tests). The cost of $12 per test included the costs for phlebotomy and reagents in a community hospital. Universal screening strategy remained a cost-saving approach, even for an incidence of maternal primary infections as low as 0.2 acute infections per 1000 women. This incidence would translate to approximately 0.5 CT cases per 10 000 children if a 25% MTCT risk is assumed. Moreover, the estimated cost of $12 per test is also in the cost range of several recently developed (already commercially available) tests for toxoplasmosis with novel technologies.

According to Calderaro et al,¹⁶² who validated the diagnostic performance of 4 commercially available IgG and IgM tests for toxoplasmosis, there are commercially available IgG and IgM tests that have 100% analytical sensitivity and specificity. Moreover, according to the Wilson et al¹⁶³ validation published in 1997, of 6 commercial IgM tests, there are commercially available IgM tests with a specificity up to 98.6% (eg, the Vidas-Biomerieux IgM). In more detail, Calderaro et al showed that the sensitivity (point estimate) of all 4 commercial IgG assays tested was 100% and the specificity (point estimates) of the different IgG tests ranged between 98.5% and 100%. In the same study, the sensitivity (point estimates) of 4 commercial IgM tests ranged between 82.4% and 100% and the specificity ranged between 99.7% and 100%. Moreover, according to Wilson et al,¹⁶³ in an FDA-sponsored validation study of 6 commercial IgM tests, the sensitivity (point estimates) of these 6 IgM tests ranged from 93.3% to 100% and the specificity (point estimates) ranged from 77.5% to 98.6%. In reference laboratories, the specificity of confirmatory testing (for the diagnosis of a recently acquired acute infection) is 100%.¹⁶³ 

Assumptions used in the decision analysis by Stillwaggon et al⁸ were as follows:

1. a rate of acute maternal primary infections during pregnancy of 1.1 in 1000 (this is likely an overestimate, because it was based on US data from 1959–1966 and might not represent the incidence of acute T gondii infection during pregnancy in the current era with the decreasing overall seroprevalence rates; however, sensitivity analyses in this article revealed that the screening strategy would remain cost saving even at a rate of acute maternal infections as low as 0.2 per 1000 women);

2. a probability of fetal infection after acute maternal primary infection of 50%^{98,224,–228};

3. a probability of fetal death attributable to CT of 5%;

4. an estimated cost for fetal death attributable to CT of $6 million;

5. a probability of visual impairment from undiagnosed CT and no antepartum treatment of 48%;

6. a probability of visual and cognitive impairment of 45%; and

7. varying probabilities of visual impairment and visual/cognitive impairment for fetal infections by trimester.

A study limitation that was acknowledged by Stillwaggon et al⁸ was that this decision analysis model assumed that all mothers in the United States would receive care by 12 weeks of gestation and would adhere to monthly follow-ups. However, this assumption might not represent a reality in the United States. Experience from France and Austria, where antepartum screening has been conducted for many years, indicated that late initial testing was common (25% of French pregnant women had their first test performed late) and there was also poor adherence to screening (only 30%–40% of pregnant women in France and Austria had all their mandatory antepartum screening tests completed).^97,214,216 It remains unclear how much a low adherence rate in pregnant women in the United States will affect the conclusions of this decision analysis. A follow-up decision analysis study should try to explore how changes in adherence rates with regard to the time of initiation of screening and the frequency of screening will affect the conclusions of this analysis.

Low-cost tests in state laboratories, combining screening for toxoplasmosis with screening for other congenital infections, may further reduce the screening cost (test cost and shipping cost) while keeping a centralized quality control for those screening tests (in the state laboratories) and maintaining high-quality confirmatory testing (at reference laboratories). Moreover, novel technologies are already widely used by several companies around the globe that provide low-cost serologic tests (eg, using microfluidics technologies in lateral flow devices, the recently introduced plasmonic gold technology).²²⁹ Some of these technologies are already commercially available in the United States. Of course, previous rigorous validation of these novel tests is required to better understand their role in universal screening programs.

Moreover, the following changes in the future may make the universal screening strategy even more cost-effective: (1) “point-of-care test” technologies for serologic screening, (2) screening with the use of saliva specimens, and (3) multiplex screening²²⁹ (multiplex IgG/IgM, which already are under evaluation and will soon provide further cost-saving alternatives for screening). In addition, screening for >1 congenital disease at the same time will provide further cost savings. Assessments of the feasibility of any universal antepartum screening strategy should take into account several strict criteria that should be fulfilled before any such strategy is considered feasible but should also take into account the rapidly changing picture in the field of novel diagnostic technologies.

In summary, the decision analysis article by Stillwaggon et al⁸ indicated that prenatal screening and treatment of toxoplasmosis in the United States could be a cost-saving approach for congenital toxoplasmosis but only under the assumptions included in the model. Although several sensitivity analyses in their model (including variations for the main assumptions such as incidence of acute maternal infections, cost of tests and life-value equivalent) showed that the results were robust, the implementation of such a screening program would, nevertheless, require a significant reduction in the current cost of screening laboratory tests and significant changes in the structure of the US prenatal care system.

Aside from universal antepartum screening for toxoplasmosis, additional approaches to consider include the following: (1) antepartum education of women to avoid T gondii–related exposures during pregnancy and (2) ascertainment of adherence to antepartum screening, at least in high-risk women (eg, those with reported exposures through the oral route [oocyst/sporozoite-related exposures to cat feces and/or tissue cysts/bradyzoite-related exposures via ingestion of undercooked meat]; occupational exposures; immigrants from high-incidence areas; women traveling internationally during pregnancy; women with immunosuppressive diseases). However, physicians should be aware that selective screening of pregnant women on the basis of self-reported exposure risk factors has the potential to miss more than approximately 50% of women who give birth to infants with CT and thus probably a much larger number of acutely infected pregnant women.

If adherence to antepartum screening is poor, neonatal screening programs can at least capture infants with CT who were not detected in utero who would benefit from postnatal treatment. Nevertheless, the risk of adverse long-term sequelae in children in the United States who were only postnatally treated (and whose mothers did not receive antepartum treatment) is significant (85% with vision impairment, 36% with recurrences in eye disease, 27% with abnormal cognition at or after 3.5 years of age, and 16% with a decrease in IQ of >15 points).¹¹² 

Primary prevention strategies in seronegative individuals are shown in Table 13. A survey of pregnant women in the United States indicated that the majority were not aware that acquisition of T gondii infection was associated with the consumption of undercooked meat.²³⁰ 

The efficacy of preventive educational interventions targeting pregnant women has been studied by 2 cluster-randomized controlled trials with a total of 5455 women,²³¹ but both of them were of low methodologic quality and did not target any objective clinical outcomes. Studies in Belgium from 1979–1982 and 1983–1990 in which antepartum education started around the tenth week of pregnancy, and therefore could not significantly affect transmission rates in the first trimester, have shown that educational tools targeting pregnant women reduced seroconversion rates by 63% to 92%.²³² However, other epidemiologic studies, summarized in the systematic review by Gollub et al,¹³ failed to replicate those findings.

A survey in 2006 of a random sample of 1200 US obstetrician/gynecologists revealed that although all respondents routinely counseled their pregnant women about risks associated with handling cat litter, fewer had counseled them about risks associated with eating undercooked meat (78%), handling raw meat (67%), gardening (65%), or the need to wash fruit and vegetables (34%). Of note, 73% of the respondents were not aware that Toxoplasma IgM tests from nonreference laboratories could have a high false-positive rate, and most (91%) were not aware of the Toxoplasma IgG-avidity test that could be used to determine the most likely time at which T gondii infection was acquired during pregnancy. The identified educational knowledge gaps were subsequently addressed in a practice bulletin from the American College of Obstetricians and Gynecologists. In 2012, an updated survey was circulated to members as well as nonmembers of the Collaborative Ambulatory Research Network to follow up on any accomplished changes.²³³ Eighty percent of providers in that survey had diagnosed a nonacute primary T gondii infection in their patient population over the previous 5 years, 43% had performed serologic screening for at least some asymptomatic pregnant women (62% of those had used appropriate serologic tests), and 13% had correctly identified the role of Toxoplasma IgG avidity. Providers from the northeastern United States were 2 times more likely to routinely screen for T gondii than those in the West, and female providers were 1.5 times more likely to screen than male providers.²³³ 

It is important that physicians, general practitioners, and specialists caring for pregnant women and/or pediatric patients become more familiar with Toxoplasma serologic tests. For samples sent to the toxoplasmosis reference laboratories, clinical interpretation of the serologic test results and estimation of the most likely time at which acute primary maternal infection with T gondii might have been acquired are provided to the referring physicians. (Additional practical information for the interpretation of the test results can be found on the PAMF-TSL Web site [http://www.pamf.org/serology/testinfo.html], and in the “Treatment” section of this report.)

1. Further studies understanding the benefits and limits of routine antepartum screening of all pregnant women in the United States for toxoplasmosis should be conducted. The cost estimates for universal CT prenatal screening per the French model appear to be unrealistic in the current American health care system. The implementation of such a screening program would require a significant change to the structure of American prenatal care.

2. Performing serologic screening only in pregnant women with known risk factors for T gondii infection or only in women with symptoms will fail to identify up to 50% of primary T gondii–infected women who are at risk of transmitting T gondii infection to their fetuses.

3. Large-scale observational studies support the benefit of early diagnosis and treatment of primary T gondii infection during pregnancy.

4. Observational studies support the benefit of postnatal treatment of infants with CT.

5. The time of initiation of antepartum treatment after acute maternal primary infection is critical and supports iterative screening.

6. According to the literature, the Centers for Disease Control and Prevention, FDA, and PAMF-TSL maternal serologic test results suggestive of acute T gondii infection in a pregnant woman (eg, positive IgM test result) should be confirmed in reference laboratories for toxoplasmosis, because false-positive test results in nonreference laboratories are not uncommon. Serologic testing of women with clinical suspicion of acute toxoplasmosis during pregnancy should be performed at a reference laboratory for toxoplasmosis to avoid unnecessary delays in the establishment of diagnosis and initiation of treatment that can affect the risk of MTCT and the risk of severe clinical manifestations of CT.

   1. Optimally, newborn infants with suspected CT should be evaluated by specialists, including experienced neonatologists, retinal specialists, neurologists, and pediatric infectious disease specialists.

   2. According to the PAMF-TSL, serologic testing of infants with suspected congenital toxoplasmosis should be performed at a reference laboratory for toxoplasmosis because of the availability of the highly sensitive IgM-ISAGA test as well as PCR testing for body fluids such as blood, urine, and CSF.

The authors acknowledge the significant contributions of Dr Despina Contopoulos-Ioannidis, MD, FAAP, Department of Pediatrics, Division of Infectious Diseases, Stanford University School of Medicine, and Dr Jose G. Montoya, MD, FACP, FIDSA, Department of Medicine, Division of Infectious Diseases and Geographic Medicine, Stanford University School of Medicine, and Palo Alto Medical Foundation Toxoplasma Serology Laboratory.

Abbreviations

AC/HS

differential agglutination (test)

AF

amniotic fluid

CI

confidence interval

CNS

central nervous system

CSF

cerebrospinal fluid

CT

congenital toxoplasmosis

ELISA

enzyme-linked immunosorbent assay

EMSCOT

European Multicenter Study of Congenital Toxoplasmosis

EUROTOXO

European Toxoprevention Study

FDA

US Food and Drug Administration

GRADE

Grading of Recommendations Assessment, Development, and Evaluation

HR

hazard ratio

Ig

immunoglobulin

ISAGA

immunosorbent agglutination assay

MTCT

mother-to-child transmission

NCCCT

National Collaborative Chicago-Based Congenital Toxoplasmosis Study

NPV

negative predictive value

OR

odds ratio

PAMF-TSL

Palo Alto Medical Foundation Toxoplasma Serology Laboratory

PCR

polymerase chain reaction

PPV

positive predictive value

P/S

pyrimethamine/sulfadiazine

RCT

randomized controlled trial

SNHL

sensorineural hearing loss

SNSD

serious neurologic sequelae or death

SYROCOT

Systematic Review on Congenital Toxoplasmosis

Yvonne A. Maldonado, MD, FAAP Jennifer S. Read, MD, MS, MPH, DTM&H, FAAP

Athena P. Kourtis, MD, PhD, MPH, FAAP

Carrie L. Byington, MD, FAAP, Chairperson Yvonne A. Maldonado, MD, FAAP, Vice Chairperson Elizabeth D. Barnett MD, FAAP H. Dele Davies, MD, MS, MHCM, FAAP Kathryn M. Edwards, MD, FAAP Ruth Lynfield, MD, FAAP Flor M. Munoz, MD, FAAP Dawn Nolt, MD, MPH, FAAP Ann-Christine Nyquist, MD, MSPH, FAAP Mobeen H. Rathore, MD, FAAP Mark H. Sawyer, MD, FAAP William J. Steinbach, MD, FAAP Tina Q. Tan, MD, FAAP Theoklis E. Zaoutis, MD, MSCE, FAAP

David W. Kimberlin, MD, FAAP – Red Book Editor Michael T. Brady, MD, FAAP – Red Book Associate Editor Mary Anne Jackson, MD, FAAP – Red Book Associate Editor Sarah S. Long, MD, FAAP – Red Book Associate Editor Henry H. Bernstein, DO, MHCM, FAAP – Red Book Online Associate Editor H. Cody Meissner, MD, FAAP – Visual Red Book Associate Editor

Dennis L. Murray, MD, FAAP Gordon E. Schutze, MD, FAAP Rodney E. Willoughby Jr, MD, FAAP

Despina Contopoulos-Ioannidis, MD, FAAP – Department of Pediatrics, Division of Infectious Diseases, Stanford University School of Medicine Jose G. Montoya, MD, FACP, FIDSA – Department of Medicine, Division of Infectious Diseases and Geographic Medicine, Stanford University School of Medicine, and Palo Alto Medical Foundation Toxoplasma Serology Laboratory and Department of Medicine, Division of Infectious Diseases and Geographic Medicine, Stanford University School of Medicine

Douglas Campos-Outcalt, MD, MPA – American Academy of Family Physicians Karen M. Farizo, MD – US Food and Drug Administration Marc Fischer, MD, FAAP – Centers for Disease Control and Prevention Bruce G. Gellin, MD, MPH – National Vaccine Program Office Richard L. Gorman, MD, FAAP – National Institutes of Health Natasha Halasa, MD, MPH, FAAP – Pediatric Infectious Diseases Society Joan L. Robinson, MD – Canadian Paediatric Society Marco Aurelio Palazzi Safadi, MD – Sociedad Latinoamericana de Infectologia Pediatrica (SLIPE) Jane F. Seward, MBBS, MPH, FAAP – Centers for Disease Control and Prevention Geoffrey R. Simon, MD, FAAP – Committee on Practice Ambulatory Medicine Jeffrey R. Starke, MD, FAAP – American Thoracic Society

Jennifer M. Frantz, MPH