Severe Acquired Toxoplasmosis in Immunocompetent Adult Patients in French Guiana

Abstract

The most common presentation of symptomatic postnatally acquired toxoplasmosis in immunocompetent patients is painless cervical adenopathy. Acute visceral manifestations are associated in rare cases. We report 16 cases of severe primary toxoplasmosis diagnosed in French Guiana during a 6.5-year period. All of the subjects were immunocompetent adults hospitalized with clinical presentations consisting of a marked, nonspecific infectious syndrome accompanied by an altered general status with at least one visceral localization, mainly pulmonary involvement (14 cases). Acute toxoplasmosis was diagnosed according to the results of serological tests suggestive of recent primary infection and the absence of an alternative etiology. Recovery was rapid following specific antitoxoplasmosis treatment. Thirteen of the 16 patients had consumed game in the 2 weeks before the onset of the symptoms, and in eight cases the game was considered to have been undercooked. Toxoplasma strains, which were virulent in mice, were isolated from three patients. Microsatellite analysis showed that all of these isolates exhibited an atypical multilocus genotype, with one allele found only for isolates of this region.

Toxoplasmosis is a major problem in French Guiana, due to the severity of the disease in AIDS patients and the fact that French Guiana is the French region with the highest prevalence of human immunodeficiency virus (HIV) (40). It is also a problem in immunocompetent subjects, due to the relatively high risk of primary infection during pregnancy and thus congenital transmission, as shown by the steadily increasing prevalence of seropositivity in the 15- to 40-year-old age group. The prevalence of toxoplasmosis seropositivity in 269 patients hospitalized in Cayenne, French Guiana, for reasons other than acute toxoplasmosis in 1999 to 2001 is shown in Table 1.

TABLE 1.

Toxoplasmosis seropositivity in 269 patients hospitalized in Cayenne, French Guiana for reasons other than acute toxoplasmosis in 1999 to 2001 as determined by ELISA (Axsym, Abbott)

Age (yr)	No. of patients	Seroprevalence
		%	95% CI
1-4	40	12.5	4.7-25.5
5-9	36	19.4	8.9-34.7
10-14	14	28.6	19.8-55.5
15-19	16	37.5	16.8-62.4
20-29	30	60	41.9-76.2
30-39	58	70.7	58.1-81.3
40-49	26	81	62.4-92.6
50-59	33	84.8	69.6-94.2
≥60	16	100	82.9-100

The recent identification of severe primary toxoplasmosis in immunocompetent patients is a third factor indicating the serious problems posed by this well-known parasitosis. Reports of two cases were published in 1998 (5, 16), and 11 other cases have been reported in the last decade. Only 2 of these 11 cases have been reported in detail (6, 18). Four of these cases were in the armed forces and have only been mentioned (17). Five were pediatric cases and were considered to be acquired toxoplasmosis, i.e., noncongenital (7). The last case of acute disseminated toxoplasmosis acquired by an immunocompetent patient was reported recently (6).

We report 16 new cases observed at the Cayenne and Saint Laurent du Maroni Hospitals over a 6.5-year period (October 1995 to March 2002). Thus, in the last 10 years at least 30 clinical cases of severe primary toxoplasmosis have been observed in immunocompetent individuals in French Guiana, which has an estimated population of 160,000 inhabitants. This clinical form has rarely been reported in the literature.

MATERIALS AND METHODS

Environment: French Guiana.

French Guiana is an overseas French region with an area of 86,000 km² It is bordered to the north by the Atlantic Ocean and separated from Brazil to the east by the Oyapock river, and from Suriname, formerly known as Dutch Guiana, to the west by the Maroni river. The Amazon rain forest covers 92% of the territory. The flora and fauna are both abundant, and most of the South American species are represented. Mammals are typical of the neotropical fauna and are a major reservoir for the parasites responsible for cutaneous leishmaniasis, Chagas' disease, and toxoplasmosis (9). The human population is strikingly diverse. Despite this diversity and the rapid population growth over the last 20 years, French Guiana had only 160,000 inhabitants in 1998 (versus 73,000 in 1982). The overall population density is especially low, and most of the population is clustered in coastal towns and villages. Fewer than 20% of the population live in inland areas in the unmodified Amazonian environment. Town dwellers and the inhabitants of the coastal region may occasionally come into contact with the forest for their work and leisure activities. There are no restrictions on hunting.

Case selection and description.

All of the 16 cases reported occurred between October 1995 and March 2002. All of the subjects were hospitalized in the two main hospitals in French Guiana: Cayenne Hospital, which is located in the largest town and administrative capital of this French department, and Saint Laurent du Maroni Hospital, located in the second-largest town subprefecture. The first four patients (October 1995 and October 1997) were not initially diagnosed with acute toxoplasmosis. A negative etiological assessment and serology in favor of or compatible with primary toxoplasmosis infection led to a presumptive diagnosis and treatment for toxoplasmosis. Subsequently, some information was obtained from interviews and complementary tests. In November 1997, a concerted action between the two hospitals (Cayenne and Saint Laurent du Maroni) led to systematic tests for acute toxoplasmosis and specific serological tests for all patients hospitalized with a marked, nonspecific infectious syndrome. The cases of severe primary toxoplasmosis included in this study all met the following criteria: (i) serological evidence of recent primary toxoplasmosis; (ii) adult subject admitted to the Cayenne or Saint Laurent Hospital with symptoms of acute infection with at least one visceral localization; (iii) negative serological tests for HIV and human T-lymphotropic virus (HTLV); (iv) no signs of immunodeficiency and no previous medical history of immunodeficiency or medication that could lead to immunodeficiency; (v) inefficacy of empirical antibiotic and antimalarial treatments when given; and (vi) absence of any other etiology that could account for the signs and symptoms, in particular, malaria, Epstein-Barr virus, cytomegalovirus, Legionella, Chlamydia, tuberculosis, Q fever, or bartonellosis.

The epidemiological, clinical, biological, and therapeutic data were obtained from the patients' medical charts. Details concerning the circumstances in which the infection occurred and dietary habits were obtained prospectively by interviewing 13 of the 16 patients individually.

All sera were examined for toxoplasma-specific immunoglobulin (IgG) and IgM antibodies by use of the corresponding enzyme immunoassay (EIA) (Axsym enzyme-linked immunosorbent assay [ELISA] system; Abbott). All paired samples were tested together in the same test to minimize test variation. Specific IgA was detected (Axsym ELISA system) and/or Toxoplasma gondii-specific IgG avidity levels (IgG avidity test; Vidas bioMérieux) were determined for the four cases for which we only had one serum sample.

For HIV serological tests, two screening techniques were used: (i) the HIV [1/2] gO Axsym microparticle EIA system (Abbott) and (ii) the Genescreen Plus HIV Ag-Ab EIA (Bio-Rad). The results of these tests were confirmed with the Strip Immunoblot assay (Chiron Riba HIV-1/HIV-2 SIA). We used Murex HTLV-I + II EIA (Abbott) to screen for HTLV and the INNO-LIA HTLVI/II (Innogenetics) to confirm the results. We checked the HIV and HTLV statuses of seven patients between 3 months and 3 years later. They were negative in all cases. None of the nine other patients were registered as being HIV positive on 30 March 2002.

Detection of T. gondii DNA in blood samples and strain isolation.

An attempt to detect T. gondii DNA and to isolate strains was performed in the Limoges, France, laboratory from blood collected on EDTA from patients 6, 7, 10, 14, 15, and 16 and from lymphadenopathy biopsy sample from patient 8. Samples from French Guiana were received in the Limoges laboratory between 3 and 12 days after collection. DNA was extracted from the blood samples by use of a QIAamp DNA minikit (Qiagen, Courtaboeuf, France) and from the lymphadenopathy sample by use of proteinase K and phenol-chloroform. The DNA was then subjected to a PCR-based assay for the detection of the T. gondii B1 gene (26). Two or three Swiss mice were inoculated intraperitoneally with the blood samples. Mice were not inoculated with the lymphadenopathy biopsy sample because it had been treated with a 1% cetylpiridium solution for the detection of Mycobacterium tuberculosis.

Genotype analysis of Toxoplasma isolates.

Microsatellite (MS) marker selection, DNA extraction, PCR amplification, and electrophoresis were carried out as described previously (1). Briefly, the QIAamp DNA minikit (Qiagen) was used to isolate DNA from tachyzoites isolated from mouse peritoneal exudate and from bradyzoites from the brains of infected mice. PCR was used to amplify eight MS sequences present in the introns of the genes coding for myosin A (TgM-A) and for beta-tubulin (TUB2) and in six expressed sequence tags (GenBank accession numbers AA519150, W35487, N61191, N82375, N83021, and N60608). The amplification reaction mixture was as described by Ajzenberg et al. (1). The forward primers were 5′ end labeled with fluorescein (6-carboxyfluorescein or 4,7,2′,4′,5′,7′-hexachloro-6-carboxyfluorescein) to allow sizing of PCR products with an automatic sequencer (Abiprism 310 collection 1.0; Applied Biosystems, Courtabœuf, France). The data were stored and analyzed with GeneScan analysis software (version 2.1; Applied Biosystems).

As previously described (1), evolutionary genetic analyses of diversity, based on genotyping with eight MS markers, showed that the T. gondii population consists of two clonal genetic groups (MS group 1 and MS group 2). However, when the allelic associations were carried out using the three least polymorphic MS markers (TUB2, expressed sequence tag W35487, and TgM-A), these two groups could be further subdivided. Of the 147 isolates that have been typed with MS markers (reference 1 and unpublished data), 131 (89.1%) have only three allelic associations with these three MS markers, corresponding to classical type I, II, or III or zymodemes 1, 2/4, and 3. The other isolates with different combinations of TUB2, W35487, and TgM-A or with new alleles were considered to have atypical genotypes. In this study, the isolates were assigned to MS groups according to the MS data obtained with the eight MS markers by correspondence analysis using the Genetix software.

Statistical analysis.

Numeric variables were compared as means and/or medians (Student's t test, Mann-Whitney test, or Kruskal-Wallis test). Ninety-five percent confidence intervals (CI) are indicated only for means because the sample size was too small to indicate 95% CI for the medians. Proportions were compared by use of the chi-square test.

RESULTS

Serological, clinical, and epidemiological data.

Tables 2, 3, and 4 summarize the serological, clinical, and epidemiological data for the 16 cases.

TABLE 2.

Serological results for toxoplasmosis, determined by IgG, IgM, and IgA ELISA (Axsym, Abbott) and IgG avidity test (Vidás, Biomérieux)

Case no.	Delaya (days)	IgG (IU/ml)b	IgM (optical density)c	IgA (optical density)c	IgG avidity
1	12	30	4.4	6.6	NDd
	79	2,775	2.7	ND	ND
2	90	4,730	4.7	1.2	Weak
3	23	50	4.7	4.2	Weak
4	40	1,250	10.8	2.2	Weak
5	12c	57	17.8	ND	ND
	21	785	18.1	ND	ND
6	6e	4.2	4.1	2.0	ND
	22	2,375	8.7	ND	ND
7	15	59	10.8	2.6	ND
	32	2,540	11.6	2.7	ND
8	−320e	Negative	Negative	ND	ND
	17	290	3.1	ND	ND
	26	2,850	3.8	ND	ND
9	8e	32	11	ND	ND
	20	238	9.3	ND	ND
10	7	Negative	Negative	ND	ND
	8	Negative	1.9	ND	ND
	9	4.7	3.7	ND	ND
	10	13	7.3	ND	ND
	11	27	9.5	ND	ND
11	7e	47	11.1	ND	ND
	14	165	5.1	ND	ND
12	16	41	9.3	ND	Weak
13	17	174	8.4	ND	ND
	20	457	8.8	ND	ND
	34	2,390	9.4	ND	ND
14	21	71	10.3	ND	ND
	27	244	8.1	ND	ND
	34	1,119	7.7	ND	ND
15	14	354	9	ND	Weak
	17	1,747	8.6	ND	ND
	26	2,280	9.1	ND	ND
16	11	283	12.6	ND	ND
	17	431	11.8	ND	ND
	26	2,545	8.6	ND	ND

^aDelay between the onset of symptoms and collection of blood sample used for serological testing. A negative number indicates that testing was carried out before the onset of symptoms.

^bPositivity threshold, IgG: 4 IU/ml.

^cPositivity threshold, > 0.60.

^dND, not determined.

^eSerological tests were carried out on frozen sera.

TABLE 3.

Characteristics of the 16 patients with severe acquired toxoplasmosis

Characteristic	Result for patienta:																Totalb
	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
Yr-mo of illness	95-10	96-1	96-4	97-10	97-12	98-3	98-3	99-1	00-3	00-4	01-2	01-3	01-4	01-11	02-1	02-3
Residence	Rural	SLM	Rural	Cayenne	Rural	Rural	Rural	Rural	SLM	Rural	Rural	Rural	St. Georges	Rural	Cayenne	Rural
Hospital	SLM	SLM	SLM	Cayenne	Cayenne	Cayenne	Cayenne	Cayenne	SLM	Cayenne	SLM	SLM	Cayenne	Cayenne	Cayenne	Cayenne
Gender	Male	Male	Male	Male	Female	Female	Male	Male	Male	Female	Male	Male	Male	Male	Male	Male	—c
Age (yr)	17	47	27	27	33	65	24	24	29	21	31	28	53	17	22	40	—d
Origin	Creole	Metro	Bush- inenge	Creole	Metro	Amerindian	Amerindian	Metro	Metro	Amerindian	Bushinenge	Metro	Metro	Amerindian	Creole	Amerindian
Delay before admission (days)	8	90	21	31	11	7	15	17	21	7	7	14	15	18	12	10	—e
Antibiotics before diagnosis	Yes	Yes	Yes	?	Yes	Yes	Yes	No	No	Yes	Yes	Yes	Yes	No	Yes	Yes	13
Antimalarial(s)	H	No	No	No	Q-D	Yes (?)	Yes (?)	H	No	H	H	No	H	No	No	H	10
Game consumption
Recent consump- tion	Yes	?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	No	Yes	Yes	No	13
Animal(s) eaten	Peccary		Peccary	?	Tapir	?	?	Armadillo	Tapir	Deer	Deer	Peccary		Tapir	Various
Delay before symptoms (days)	10		8		12			10	14	6	7	14		12	ND		—f
Undercooked game	Yes		Yes		Yes			Yes	No	Yes	Yes	Yes		Yes	No		8

^aSLM, Saint Laurent du Maroni; Metro, French mainland; H, halofantrine; Q-D, quinine plus doxycycline; ND, not determined.

^bUnless otherwise noted, values are total numbers of patients with characteristic.

^cTotal number of patients by gender: 13 males and 3 females.

^dMedian age, 28 years; mean age, 32 years (95% CI, 25.2 to 45.3 years).

^eMedian delay, 15 days; mean delay, 20 days (95% CI, 8.3 to 32.3 days).

^fMedian delay, 10 days; mean delay, 10.3 days (95% CI, 8.4 to 12.2 days).

TABLE 4.

Clinical features and treatment for the 16 patients with severe acquired toxoplasmosis

Symptom or treatment	Result for patient:
	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	Total
General symptoms
Temp of >39°C	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	16
Wt loss (kg)	3	?	3	10	8	Yes (?)	Yes (?)	10	6	6	3	10	7	5	?	10	14
Elevated liver enzymes	Yes	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	15
Pulmonary involvement	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	No	No	Yes	Yes	Yes	Yes	Yes	14
Cough without pneumonia	Yes	No	Yes	Yes	No	No	No	No	Yes	No	No	Yes	No	No	No	No	5
Pneumonia	No	Yes	No	No	Yes	Yes	Yes	Yes	No	No	No	No	Yes	Yes	Yes	Yes	9
Chest X ray		Bilateral interstitial infiltrates			Bilateral interstitial infiltrates	Bilateral interstitial infiltrates	Bilateral interstitial infiltrates	Bilateral interstitial infiltrates					Bilateral alveolar- interstitial infiltrates	Bilateral alveolar infiltrates	Bilateral interstitial infiltrates	Bilateral interstitial infiltrates
Respiratory distress		No			No	Yes	Yes	No				No	No	Yes	Yes	Yes	4
Severe Headache	No	Yes	Yes	Yes	Yes	No	No	Yes	Yes	No	Yes	No	Yes	Yes	No	No	9
Splenomegaly	Yes	Yes	No	No	Yes	No	No	Yes	No	Yes	Yes	Yes	No	No	Yes	No	8
Peripheral adenopathy	Axillary inguinal	No	Cervical	Axillary	No	No	No	No	Cervical	Axillary	No	Cervical	No	Inguinal	Axillary inguinal	Cervical, inguinal	9
Hepatomegaly	Yes	Yes	No	No	No	Yes	Yes	No	No	No	No	No	Yes	No	No	No	5
Diarrhea	No	Yes	Yes	No	No	No	No	No	No	Yes	Yes	Yes	No	Yes	No	No	6
Chorioretinitis	No	Yes	No	Yes	No	No	Yes	No	No	No	No	No	No	No	No	No	4
Decreased visual     acuity		Yes (unilateral)		Yes (unilateral)			Yes (unilateral)
No. of lesions		2		2			1
Topography of     lesions		Peripheral		Juxtamacular			Peripheral
Hyalitis		Yes		Yes			Yes
Treatment
Antitoxoplasmosis     drug(s)a	Py + Sul	Py + Sul	Py + Sul	Sp	Py + Sul	Py + Sul	Py + Sul	Py + Sul	Sp	Py + Sul	Py + Sul	Sp	Py + Sul	Py + Sul	Py + Sul	Py + Sul	—b
Delay for apyrexia     (days)	7	5	4	15	7	10	?	6	13	6	3	10	5	5	30	12	—c

^aPy, pyrimethamine; Sul, sulfadiazine; Sp, spiramycin.

^bTotal number of patients treated with spiramycin, 3; total number of patients treated with pyramethamine plus sulfadiazine, 13.

^cFor treatment with spiramycin, the median delay was 13 days (mean, 12.7 days), and for treatment with pyramethamine plus sulfadiazine, the median delay was 6 days (mean, 6.4 days). The delay for patient 15 was not taken into account when calculating the mean and median, because this patient, who had sickle cell anemia, had persistent fever due to nosocomial complications.

A significant and rapid increase in the number of IgG antibodies in two separate samples, tested at the same time, associated with the presence of specific IgM confirmed that 12 patients had recently undergone seroconversion. For the other four cases, only one result was available, but the presence of IgM antibodies associated with specific IgA and/or a low avidity of IgG was highly suggestive of recent infection.

In all cases, the clinical presentation consisted of a marked nonspecific infectious syndrome. This was accompanied by an altered general status with consequent weight loss (>10%) in more than half of the cases. Fever was the main symptom; it was typically high (39 to 40°C), prolonged (more than 10 days), and resistant to the antibiotics and antimalarial drugs that were usually given on a presumptive clinical basis before hospitalization.

Elevated liver enzyme levels and lung, lymph node, and spleen involvement were noted in 15, 14, 9, and 8 cases, respectively. Four patients presented acute retinitis, whereas no cardiac or cerebral involvement was recorded, although a severe and persistent headache was noted in nine patients. Three patients presented up to two active chorioretinis foci with unilateral perimacular lesions. For those (two patients) whose short-term ophthalmic outcome was known, improvement of eyesight occurred between 3 and 4 weeks with typical developing scars. In four patients, severe sepsis-like infection and acute respiratory distress necessitated a period of treatment in intensive care. Two of these patients may have been infected by the same strain of T. gondii, as they lived together and fell ill at the same time. The mean delay between the appearance of the first symptoms and hospitalization was 20 days (95% CI, 8.3 to 32.3 days). The initiation of specific therapy was always followed by recovery. For patient 15, who was homozygous for sickle cell anemia, long-lasting apyrexia was delayed due to nosocomial complications caused by infection with Pseudomonas aeruginosa and an episode of phlebitis.

Twelve of the patients lived in tropical rain forest areas. Thirteen had consumed game in the 2 weeks before the onset of symptoms. Two other patients had also eaten game, but it is unknown whether they had done so recently. In eight cases, the meat was clearly undercooked. In nine cases, the eaten animal was identified: there were three peccaries, including two collared peccaries (Tayassu tajacu) and 1 white-lipped peccary (Tayassu pecari); three tapirs (Tapirus terrestris); two deer (Mazama sp.); and one nine-banded armadillo (Dasypus novemcinctus). The mean delay between the meal that was believed to have caused the infection and the start of the symptoms was 10.3 days (95% CI, 8.4 to 12.2 days). Only one patient (patient 16) had not eaten game or any other meat, because he is vegetarian. Young adults were most often affected Similarly, men were predominantly affected (13 men and 3 women). The Creole population seemed to be underrepresented, with only 3 cases out of 16 (19%), even though Creoles account for >50% of the population living in French Guiana.

Detection of T. gondii in samples and genotype analysis.

The results for detection of T. gondii are reported in Table 5. The blood sample from patient 14 was used to inoculate three mice; one mouse died at 18 days postinoculation (p.i.), another died at 36 days p.i (convulsions occurred at 1 month p.i), and the third one was not infected. A T. gondii isolate (GUY-DOS) was isolated from the brains of both infected mice. One blood sample was taken daily for 3 days from patient 15. All seven of the mice inoculated with these samples died between 13 and 20 days p.i. A T. gondii isolate (GUY-MAT) was isolated from mouse peritoneal exudate at 8 days p.i. All six mice inoculated with the three daily blood samples taken from patient 16 died between 15 and 16 days p.i. A T. gondii isolate (GUY-KOE) was isolated from mouse peritoneal exudate at 7 days p.i.

TABLE 5.

Detection of T. gondii in patient samples

Case	Delay (days) between:		Sample	B1 PCR result	Mouse inoculation result (no. positive/total)
	Beginning of symptoms and sampling	Sampling and inoculation of mice
6	16	11	Blood	Negative	0/3
7	9	11	Blood	Negative	0/3
8	19	8	Lymphadenopathy	Negative	NDa
10	9	12	Blood	Negative	0/3
14	21	3	Blood	Positive	2/3
15	14	4	Blood	Positive	3/3
		5	Blood	Inhibitors	2/2
		4	Blood	Inhibitors	2/2
16	11	7	Blood	Positive	2/2
		6	Blood	Positive	2/2
		5	Blood	Positive	2/2

^aND, not done.

The length of PCR product was assigned to an allele for each MS marker as recently described (1). The results of multilocus genotyping with the eight MS markers are reported in Table 6 for the GUY-DOS, GUY-MAT, and GUY-KOE isolates and for two isolates (VAND and RUB) previously isolated from French Guiana (5, 16). These eight MS markers revealed that these five isolates from French Guiana have different multilocus genotypes. Correspondence analysis showed that all five isolates belonged to MS group 2. For the GUY-DOS, GUY-MAT, and GUY-KOE isolates, allelic associations of MS markers TUB2, W35487, and TgM-A show atypical genotypes, with combinations between MS group 1 alleles (126 in TUB2 for GUY-MAT or 101 in W35487 for GUY-DOS and GUY-KOE isolates) and MS group 2 alleles (124 in TUB2 for GUY-DOS and GUY-KOE isolates or 95 in W35487 for GUY-MAT). All three isolates had the very rare allele 115 in TgM-A. This allele is shared only by VAND, which was also isolated from French Guiana (5).

TABLE 6.

Multilocus genotypes with eight MS markers for five T. gondii isolates from humans infected in French Guiana

Isolate	Allele in marker:
	TUB2	W35487	TgM-A	N60608	N82375	N83021	N61191	AA51915
GUY-DOS	124	101	115	137	111	133	120	140
GUY-KOE	124	101	115	133	109	133	146	128
GUY-MAT	126	95	115	133	105	135	132	156
VANDa	126	95	115	137	109	129	126	154
RUBa	124	95	117	137	107	133	120	136

^aIsolate previously isolated from French Guiana (5, 14).

DISCUSSION

The incidence of this severe form of primary T. gondii infection is high in French Guiana, given the small size of the population, the number of recent cases (29 reported observations in the last decade), and the small number of cases published in the overall literature. Consultation of medical databases (PubMed for National Library of Medicine, Current Contents/Clinical Medicine, Pascal [INIST/CNRS], EMBASE [Excepta Medica], and Research Alert) showed that there have been fewer than 50 documented cases of severe primary toxoplasmosis in immunocompetent subjects worldwide in the last 30 years. However, the real number of cases is likely to be much higher, as a much larger number of severe cases are not reported.

The infectious syndrome associated with a marked change in general condition appeared to be unusually severe. The frequency of pulmonary complications was unusually high, as they have seldom been reported in the literature (2, 38). Four of the five soldiers who became infected in 1994 during the same journey in the deep forest of French Guiana developed severe pulmonary toxoplasmosis; one of these cases has been reported in detail (16), and the other three cases were unproven (17) but were almost certainly genuine cases (M. Pinon, personal communication). A pulmonary infection location, associated with cardiac and neurological involvement, has also been observed previously in French Guiana (5, 6, 16). Other locations, which were not seen in our cases, include classic myocarditis (11, 33) and/or pericarditis (37), polyradiculoneuritis (30, 39), and acute neurological syndrome (20). The largest series of cases with neurological involvement in immunocompetent subjects was reported by a French reference unit in Paris over a 13-year period (1982 to 1995) (13). However, only six of these patients had acquired toxoplasmosis with central nervous system or meningeal involvement. Three of these cases were in adults and three were in children, including the previously described pediatric case from French Guiana (18). In comparison, 45 cases of ocular involvement were reported by the same unit during the same period, but only one-quarter of these cases corresponded to a primary infection. Frequent ocular involvement has also been reported in southern Brazil (23), but this is not specifically associated with recent primary infection or with a wild toxoplasmosis cycle. It is noteworthy that other complications have occasionally been reported: hepatic (22), renal (34), subcutaneous (41). and muscular (42).

A significant serological response in terms of IgG antibody levels occurred at an early stage in most cases. For the observations that give a precise result, the positivity threshold was reached within a week after the onset of the symptoms. High concentrations were reached within 1 month in all cases. Eleven of our 16 patients lived in rural communities, as do one-third of the general population of French Guiana. In all cases except one (patient 16), the source of infection was thought to be the consumption of game. Consumption of game is not uncommon in French Guiana, but game is traditionally well cooked, at least by the Creoles, Amerindians, and Noir Marrons. In 8 out of 16 cases, undercooked game was eaten between 1 and 2 weeks before the onset of the clinical signs and between 2 and 3 weeks before the first positive serological test. This is consistent with a contamination pattern involving “wild” strains of T. gondii. The suspected animals are species that are often infected with Toxoplasma in the region. The prevalence of antibodies for toxoplasmosis differs greatly between different neotropical species of free-ranging mammals in French Guiana (10). The prevalence was zero or very low for tree-dwelling animals such as sloths (Choleopus didactylus), porcupines (Coendou prehensilis), and howler monkeys (Alouatta seniculus). Conversely, the prevalence of infection was 40 to 60% among burrowing animals, granivores, and insectivores, such as deer (Mazama sp.), armadillos (Dasipus novemcinctus), pacas (Agouti paca), and peccaries (Tayassu tajacu). Given the absence of domestic cats from the regions of the primary tropical rain forest inhabited by humans, the noncarnivorous species (i.e., folivores, frugivores, or granivores) are probably infected via the ingestion of oocysts (10), which are eliminated into the soil or water in the stools of wild cats before being dispersed. These results are consistent with a study carried out on the neotropical fauna of Panama at the beginning of the 1980s (21). Wild cats, including jaguars, ocelots, pumas, jaguarundis, and margays, were quite numerous and are still present in the forest of French Guiana. At least the first four of these felines are known to eliminate Toxoplasma oocysts in their stools (29, 36). However, the number and frequency of eliminated oocysts and the length of time during which these felines eliminate oocysts during natural infections are not known. A study to answer these questions is under way in French Guiana, but all attempts to isolate oocysts from stools collected in the natural environment, which are difficult to find, have failed so far. The only time that Toxoplasma oocysts were found was with a young captive ocelot that had been fed artificially for a month (unpublished data). Human infection may also occur from the consumption of water infected with excreted oocysts. The two familial cases (patients 3 and 4) in our study came from a village where the population occasionally consumes stream water and where witnesses consistently report the presence of wild cats in the area. Patient 16, who carried one of the three strains of T. gondii isolated, is vegetarian. He regularly stays in the forest, where he drinks untreated river water. The authors of the 1992 study in French Guiana (16) also suggested that the consumption of infected water was the most likely route of infection because the patients studied had not eaten wild animals. This was also thought to be the most likely infection route for the American soldiers on operation in a region of Panama in which wild cats were prevalent (4). This transmission route was also suggested for an epidemic of toxoplasmosis in British Columbia (Canada) in which the local drinking water system, based on the use of unfiltered surface water, was implicated (8), in a zone in which two pumas (Felis concolor vancouverensis) were found to be carrying T. gondii oocysts (3). The strain involved was later characterized and was found to have an uncommon genotype (31).

It is impossible to obtain a real picture of the genetic diversity of a parasite with a worldwide distribution, such as T. gondii, because of the limited geographical origins of the isolates. Most of the isolates that have been typed were from North America or Europe (12, 14, 15, 27, 28, 32, 35). The few studies that genetically characterized isolates from other geographical origins or from unusual host species showed unexpected findings (5, 16, 19). Thus, these exotic isolates are of special interest for genotype analyses. All of the isolates from our French Guianan patients clustered in MS group 2, but they had atypical genotypes with unusual allelic combinations of the three least polymorphic MS markers and a new allele for the TGM-A locus. In a collection of 147 isolates (136 from North America and Europe, 9 from South America, 1 from Australia, and 1 from Japan) typed with MS markers (reference 1 and unpublished data), allele 115 for TgM-A was found in only four isolates from French Guiana. This is the first report of an association between a specific allele and a geographical area. It does not mean that all isolates from French Guiana have this allele (see isolate RUB in Table 5, which has the classical allele 117 for TgM-A). The existence of new alleles does not modify the separation of T. gondii isolates into two groups, as all atypical isolates can be clustered either in MS group 2 or in MS group 1, with genotypes mixing alleles of both groups. All three isolates had the very rare allele 115 for TgM-A. This allele is shared only by VAND, which was also isolated from French Guiana (5). In spite of the existence of a truly novel allele in very atypical isolates (including VAND), exotic strains seemed to be derived from the mixing of the proposed two ancestral lineages (24). New alleles, such as allele 115 for the TgM-A gene, are probably markers of genotypes that are well adapted to wild hosts from uninhabited regions of South America, whereas classical alleles (e.g., allele 119 for the TgM-A gene) are markers of genotypes that are particularly well adapted to domestic hosts (e.g., pigs, sheep, and cats) from North America or Europe. Furthermore, as reported for other atypical isolates, these isolates are virulent in mice, unlike most human isolates (14).

In humans, asymptomatic or benign toxoplasmosis is the rule in immunocompetent patients. This study confirms that there is an association between atypical genotypes and severe toxoplasmosis acquired by immunocompetent adults in French Guiana (5, 16). In the same way, it has recently been reported that an unusual abundance of atypical strains is associated with human ocular toxoplasmosis (25) and with severe forms of congenital toxoplasmosis (1a). Thus, these atypical isolates are of special interest for pathogenicity studies with humans. In particular, the isolation of these atypical isolates should lead to greatly reinforced treatment and clinical surveillance.

In conclusion, we have described the incidence of severe primary toxoplasmosis in immunocompetent adults in French Guiana. It is important that physicians consider the diagnosis of acute toxoplasmosis from the onset in patients who live in or have recently visited the Amazon region and who present a severe infectious syndrome with visceral involvement, especially lung involvement. Serological tests should be promptly considered in such cases, and if recent T. gondii infection is diagnosed, antitoxoplasmosis treatment (sulfadiazine plus pyrimethamine) should be given. This unusual form of toxoplasmosis is associated with an unusual genotype of T. gondii and may be linked to a neotropical forest-based cycle involving wild cats and their prey. The “wild-type” strain of T. gondii should be isolated from neotropical mammals so that we can study its virulence and genetic characteristics and compare them to those of strains isolated from human also infected in French Guiana.