Purpose
To investigate ocular involvement (prevalence, incidence, lesion characteristics) following postnatally acquired infection with an “atypical” genotype of Toxoplasma gondii during a well-characterized 2001 outbreak in Santa Isabel do Ivaí, Brazil, attributed to a contaminated municipal reservoir.
Design
Prospective longitudinal cohort study.
Methods
We performed ophthalmic examinations on 290 of 454 individuals with serologic evidence of T gondii infection during the epidemic (positive IgM antibody tests). Prevalence of ophthalmic findings (intraocular inflammatory reactions [including transient, isolated retinal whitening without clinically apparent retinal necrosis] and necrotizing retinochoroiditis) at initial examination (baseline) and incidence of new findings during 10.5 months of follow-up were calculated. Cumulative risks of ophthalmic events were determined (Kaplan-Meier technique).
Results
Ocular involvement was present in 33 of 288 IgM+ individuals (11.5%) at baseline, including 17 with focal retinal whitening only and 13 with necrotizing retinochoroiditis. Incidence of new ocular involvement was estimated to be 1.73 events per 100 person-months (PM); cumulative risk at 10.5 months was 30.1%. Incident necrotizing retinochoroiditis was more common among those with focal retinal whitening at baseline (6.7/100 PM) than among those with no ocular involvement at baseline (1.11/100 PM; hazard ratio 6.07 [1.94–19.01]; P < .0001).
Conclusions
Waterborne infection with an atypical genotype of T gondii is associated with substantial risk of ocular involvement. Lesions may continue to develop during the first year after infection. The increased risk of late necrotizing retinochoroiditis associated with isolated focal retinal whitening at presentation suggests the early presence of parasites in the retina, despite initial lack of observable retinal necrosis.
Occasional epidemics of toxoplasmosis provide an opportunity to study clinical features and course of disease in large numbers of patients with known times and routes of infection. During October-December 2001, a large cluster of Toxoplasma gondii infections was reported in Santa Isabel do Ivaí, a town in Parana, a state in southern Brazil. Epidemiologic investigations implicated a municipal surface water reservoir, believed to be contaminated with cat feces, as the source of infection. The parasite was subsequently found to have an “atypical” genotype, not belonging to 1 of the 3 major clonal lineages. Infection with a parasite having an atypical genotype is believed to be a risk factor for severe ocular disease.
Most reports about epidemics of toxoplasmosis have described ocular involvement at the time of infection, but little is known about the subsequent risk of eye disease. Serial ophthalmic examinations were performed on affected individuals following the Santa Isabel do Ivaí outbreak, as a part of the medical response to this public health emergency. Using these prospective data, we determined the incidence and characteristics of ocular involvement during the first year after infection and identified additional factors that may be related to the severity and course of retinal disease.
Methods
We performed this prospective longitudinal cohort study as part of the overall investigation of T gondii infections in Santa Isabel do Ivaí following identification of the outbreak. The study was determined to be exempt from institutional review board (IRB) approval by the Medical Ethics Committee of the Federal University of São Paulo because data were collected in conjunction with the investigation of a public health emergency. The investigation did adhere to the tenets of the Declaration of Helsinki. Analysis of anonymized data was also determined to be exempt by IRBs at UCLA and the United States Centers for Disease Control and Prevention.
In response to the epidemic, a local center was established for the study of ophthalmic disease. Using local media, public health authorities invited citizens to be examined at the center without charge. Everyone who presented for examination was evaluated, whether or not these presenting individuals had systemic or ophthalmic symptoms. Included in this study were those who had serologic evidence of recent infection (positive anti– T gondii IgM antibody tests).
Data Collection
Because of the large number of people who requested examinations, 2 sessions (January and February 2002) were organized for initial (baseline) ophthalmic evaluations; there were no criteria for assigning individuals to either session, and examinations were performed without knowledge of an individual’s IgM antibody status. Some people examined in January chose to return for repeat examination in February, providing 1-month follow-up data. To avoid bias, an attempt was made to reexamine all individuals in June or September 2002, whether or not they had findings on initial examination. To confirm findings, all examinations were performed by at least 2 of 3 ophthalmologists (C.S., C.M., R.B.). At all visits, fundi were examined by slit-lamp biomicroscopy and indirect ophthalmoscopy.
The following demographic and medical data were collected for all individuals: sex; age; anti– T gondii IgM antibody test results (obtained in November or December 2001); the presence of nonocular signs or symptoms of toxoplasmosis (eg, sore throat, malaise, lymphadenopathy); and antiparasitic treatment (either before baseline or during follow-up). For eyes with ocular involvement, the following information was collected at each examination: the presence of signs (redness) and symptoms (visual disturbance, pain, photophobia) of ocular involvement; presence or absence of intraocular inflammatory reactions (anterior chamber cells and flare; vitreous inflammatory reactions; retinal vascular sheathing; or focal retinal whitening without clinically apparent retinal necrosis, presumed to be retinal infiltrates [ Figure 1 ]); and presence or absence of fundus lesions (focal retinal whitening, as described above; foci of necrotizing retinochoroiditis; or retinochoroidal scars consistent with prior episodes of necrotizing retinochoroiditis). Focal retinal whitening was recorded separately from other signs of intraocular inflammation. Lack of retinal necrosis was assumed in areas of retinal whitening if there was no retinal granularity or if choroidal details could be visualized through the retinal haze on funduscopy. The size (≥1 disc area [DA] vs <1 DA) and location (macular vs extramacular) were determined for foci of retinal whitening and foci of necrotizing retinochoroiditis. Visual acuity and intraocular pressure were not considered, not being relevant to the specific purpose of this study.
Anti– T gondii IgM antibodies were identified by an ELISA technique (Abbott Laboratories, Abbott Park, Illinois, USA) at Lacen Laboratory (Curitiba, Brazil). Values ≥0.600 IU/mL were considered to be positive. Test specificity was 95% and sensitivity was 98%, as reported by the laboratory.
Conventions and Definitions
Retinal lesions were described using terms commonly employed by investigators who study ocular toxoplasmosis. Because individuals were examined infrequently, we established standard conventions for inferring disease course, based on findings. These definitions and conventions are listed in the Supplemental Materials (available at ajo.com ).
Data Analysis and Statistical Techniques
We analyzed findings for only 1 eye per person. For individuals with bilateral disease, the right eye was chosen as the study eye. Only individuals with baseline examinations were included in longitudinal analyses. We chose mid-November 2001 (the midpoint for the interval during which all new infections were reported) as time of infection for all individuals. Incidence was calculated for selected findings on the basis of events per 100 person-months (PM) of follow-up. For individuals with new retinochoroidal scars during follow-up, ocular disease was assumed to have occurred when the scar was identified, for purposes of calculating intervals. For longitudinal subgroup analyses based on the presence or absence of focal retinal whitening, baseline was used as time 0. For all other longitudinal analyses, mid-November 2001 (the assumed time of infection) was used as time 0.
Statistical analyses were performed using SAS software version 9.3 (SAS, Inc, Cary, North Carolina, USA). Cumulative risk of ocular involvement was estimated using the Kaplan-Meier method and compared using the log-rank test. Relative risks were expressed as hazard ratios (HR), estimated from Cox proportional hazards regression models. The Fisher exact test was used to compare proportions between subgroups at baseline.
Results
Santa Isabel do Ivaí had a population of 9147 at the time of the outbreak, approximately 6000 of whom lived in an urban setting ( http://tabnet.datasvs.gov.br/tabdata/cadernos/PR/PR_Santa Isabel_do_Ivai_Geral.xs, accessed March 2010). Serologic testing was performed on 3868 volunteers, 460 of whom were found to have anti– T gondii IgM antibodies. A total of 562 people presented for ophthalmic examination, 457 of whom had undergone IgM antibody testing. The 105 individuals without IgM antibody tests were not included in primary analyses, despite the fact that many had systemic symptoms, because infection during the outbreak had not been confirmed. Among the 457 examined individuals who had undergone IgM antibody testing, 454 had positive test results; these individuals constituted the study population.
Characteristics of the study population are shown in Table 1 . People of all ages were infected, with men and women involved equally. Symptoms suggestive of systemic toxoplasmosis were present in the majority of those examined. Baseline ophthalmic examinations were performed on 290 of the 454 IgM-positive individuals; they had characteristics similar to those of all IgM-positive individuals. Ophthalmic symptoms were present at baseline in 47% of those for whom data were available (133 of 282 individuals). Ophthalmic involvement at baseline (n = 33, Table 2 ) was similar between individuals with ophthalmic symptoms (n = 15 of 133, 11%) and those without symptoms (n = 18 of 149, 12%; Fisher exact test, P = .86). Ophthalmic symptoms were described by some individuals without findings on ophthalmic examination, as were systemic symptoms by some without serologic evidence of infection; these symptoms were attributed by investigators to unrelated disorders or to anxiety about the outbreak.
| Characteristic | All Individuals (N = 454) | Individuals With Baseline Examinations (N = 290) |
|---|---|---|
| Sex, n (percentage) | ||
| Male | 222 (49%) | 141 (49%) |
| Female | 232 (51%) | 149 (51%) |
| Age (y) | n = 442 a | n = 279 a |
| Mean ± SD | 26.1 ± 17.7 | 27.0 ± 16.8 |
| Median (range) | 22 (0–84) | 25 (0–84) |
| IgM antibody test result b | n = 269 a | n = 269 a |
| Median (range) | 3.5 (0.6–80.9) | 3.5 (0.6–80.9) |
| Clinically apparent nonocular toxoplasmosis, c n (percentage) | 395 (88%) (n = 450 a ) | 260 (91%) (n = 286 a ) |
| Ophthalmic symptoms, d n (percentage) | 227 (51%) (n = 445 a ) | 133 (47%) (n = 282 a ) |
| Antiparasitic treatment, n (percentage) | 14 (3%) | 10 (3%) |
| Received before initial eye examination | 5 (1%) | 3 (1%) |
| Received after initial eye examination | 9 (2%) | 7 (2%) |
a Number of individuals for whom values were known.
b Performed by ELISA technique. Values <0.6 are considered negative.
c Arthralgias, fatigue, fever, malaise, lymphadenopathy, sore throat, or a combination of these disorders.
d Blurring or other visual disturbance, redness, pain, or photophobia.
| Characteristic | January 2002 | February 2002 | June 2002 | September 2002 | ||||
|---|---|---|---|---|---|---|---|---|
| Baseline a | Baseline a | Follow-up b | Initial c | Follow-up d | Initial c | Follow-up | ||
| From Baseline | From June 2002 Only | |||||||
| Patients examined (n) | 74 e | 216 | 27 | 161 | 211 | 3 | 29 | 17 |
| Patients whose findings were available (n) | 72 e | 214 e | 25 e | 160 | 210 | 3 | 29 | 17 |
| Active ocular involvement (n) | 8 | 24 | 5 | 4 | 7 | 0 | 3 | 1 |
| Inflammatory reaction only f | 1 | 1 | 0 | 0 | 0 | – | 0 | 0 |
| Retinal lesions | 7 | 23 | 5 | 4 | 7 | – | 3 | 1 |
| Necrotizing retinochoroiditis g | 5 | 8 | 3 | 0 | 2 | – | 1 | 1 |
| New | 5 | 8 | 1 h | – | 1 | – | 0 | 0 |
| Recurrent | NA | NA | 2 | – | 1 | – | 1 | 1 |
| Focal retinal whitening only i | 2 | 15 | 2 | 4 | 5 j | – | 2 | 0 |
| Retinochoroidal scars only (n) | 1 | 0 | 2 | 13 | 13 j | – | 9 | 5 |
a Initial examination, if performed in January or February 2002, was considered baseline examination for longitudinal analyses.
b Individuals whose initial examination was in January 2002.
c Individuals who were not examined in January or February 2002. They were not included in longitudinal analyses.
d Individuals who had baseline examinations in either January or February 2002.
e The 2 individuals examined in January whose findings were not available were seen in follow-up in February; their findings in February were considered baseline findings for purposes of analysis; thus, the total number of individuals for whom baseline findings were available was 288.
f Anterior chamber cells, vitreous humor cells or haze, retinal vascular sheathing, or a combination of these findings in the absence of focal retinal whitening or necrotizing retinochoroiditis.
g Descrete foci of dense retinal opacification with thickening and overlying inflammatory material.
h The patient had progression to necrotizing retinochoroiditis from a focus of retinal whitening only (seen in January 2002).
i Descrete foci of retinal whitening without clinically apparent retinal necrosis.
j One person had both a retinochoroidal scar and an unrelated focus of retinal whitening; the case was categorized with those having retinal whitening.
Results of eye examinations are shown in Table 2 . Ophthalmic findings were known for 288 of 290 individuals with baseline examinations. Ocular involvement was present in 33 individuals (11.5%). We recalculated the prevalence of ophthalmic involvement at baseline in several models that also considered individuals excluded from our primary analyses, based on various assumptions (see Supplemental Materials , available at ajo.com ). Prevalence only varied from 9.6% to 12.0% in these models.
There was active disease in both eyes of 4 individuals at baseline; in each, lesion characteristics (type, size, location) were the same in both eyes. In 3, there were foci of retinal whitening only; in the other, there was necrotizing retinochoroiditis in both eyes.
An inactive retinochoroidal scar, without associated retinal or vitreous inflammation, suggestive of healed toxoplasmic retinochoroiditis, was present in an asymptomatic individual at baseline examination in January. We assumed that the scar was the result of prior endemic disease and excluded him from further analysis. He had a relatively low value for anti– T gondii IgM antibodies (2.6 units), and it is known that IgM+ antibodies can persist for 1 year or longer after infection. He had no active disease on each follow-up examination. Two individuals had only intraocular inflammatory reactions at baseline; neither had follow-up examinations. By June 2002, no one was found to have inflammatory reactions only.
Five patients were examined in both January and February. Two had necrotizing retinochoroiditis at both examinations; 1 had necrotizing retinochoroiditis in January that was resolving with evidence of new scar formation by February; 1 had focal retinal whitening in January that evolved to a focus of necrotizing retinochoroiditis by February; and 1 had only focal retinal whitening in both January and February.
The majority of patients had serial examinations; 236 of 290 patients (81%) with baseline examinations were seen at least twice, and 212 (73%) were seen at baseline and again at least 4 months after baseline. Among 255 individuals with no ocular involvement at baseline, cumulative follow-up was 811 PM. Any ocular involvement during follow-up was seen in 14 individuals (incidence, 1.73/100 PM). Foci of retinal whitening were seen during follow-up in 5 individuals (incidence, 0.62/100 PM). Evidence of incident necrotizing retinochoroiditis was seen in 9 individuals (1 with active necrotizing retinochoroiditis; 8 with new retinochoroidal scars; incidence, 1.11/100 PM).
When all individuals were considered, cumulative risk of any ocular involvement at 10.5 months after onset of the epidemic was 30.1% (95% confidence interval [CI], 14.0%–57.1%); among those who were untreated, it was 27.4% (95% CI, 11.5%–56.6%; Figure 2 ). Cumulative risk of necrotizing retinochoroiditis at 10.5 months after onset of the epidemic was 27.2% (95% CI, 12.2%–53.8%); among those who were untreated, it was 26.5% (95% CI, 10.6%–57.2%). Ophthalmic involvement at baseline or during the 10.5 months after onset of the epidemic was not related to age, sex, IgM antibody test values, or presence of symptomatic nonocular disease (data not shown; all P values ≥.11).
Outcomes differed on the basis of ophthalmic findings at baseline ( Figure 3 ). Among 17 individuals who had only focal retinal whitening at baseline, cumulative follow-up was 75 PM. Evidence of incident necrotizing retinochoroiditis was seen during follow-up in 5 individuals (1 with active necrotizing retinochoroiditis; 4 with retinochoroidal scars; incidence, 6.7/100 PM). The risk of developing necrotizing retinochoroiditis during follow-up was significantly higher for patients with focal retinal whitening at baseline than for individuals with no ocular involvement at baseline (HR 6.07, 95% CI, 1.94–19.01, P < .0001). At 7 months after baseline, the cumulative risk of incident necrotizing retinochoroiditis (active necrotizing retinochoroiditis or retinochoroidal scars) was 52.4% (95% CI, 20.4%–90.8%) for those with only focal retinal whitening at baseline and was 6.7% (95% CI, 2.6%–16.7%) for those with no ophthalmic involvement at baseline. No patient with focal retinal whitening at baseline had more than 7 months of follow-up. As shown in Figure 3 , cumulative risk among those with no ophthalmic involvement increased to 53.1% (95% CI, 11.6%–99.1%) at 8 months, based on the presence of a retinochoroidal scar in 1 of only 3 individuals with that length of follow-up; thus, we do not know whether the observed difference between subgroups was sustained.
Among 13 individuals who had necrotizing retinochoroiditis at baseline, cumulative follow-up was 69 PM. Recurrences were seen in 3 individuals (incidence, 4.3/100 PM). The incidence of recurrences was slightly lower among individuals with necrotizing retinochoroiditis at baseline than the incidence of new necrotizing retinochoroiditis lesions during follow-up among those with only focal retinal whitening at baseline, although the difference was not significant (HR, 0.81, 95% CI 0.19–3.40, P = .76).
Multiple examinations during periods of active disease were not available for most individuals, and as a result, evolution of focal retinal whitening could not be studied in most cases. As noted above, 1 focus of retinal whitening was known to evolve directly into a focus of necrotizing retinochoroiditis, with subsequent resolution to a scar. In contrast, we were able to confirm that retinal whitening resolved completely in another individual, leaving a clinically normal fundus, before a retinochoroidal scar was seen in the same area on later examination. Scars were eventually seen during follow-up, both in patients who had small foci of retinal whitening at baseline and in those with large foci, but the number of cases was too small to determine the relative risks associated with size or location of retinal whitening for eventual development of retinal necrosis. Three patients with necrotizing retinochoroiditis at baseline were noted to develop independent foci of retinal whitening at a later examination.
Lesion characteristics are shown in Table 3 . Lesions that developed during follow-up are similar to those seen at baseline, in terms of size and location, but numbers were too small for meaningful statistical comparisons. The majority of necrotizing retinochoroiditis lesions involved the macula, while the majority of focal retinal whitening lesions involved extramacular locations. Furthermore, macular lesions were more likely to be necrotizing retinochoroiditis; 8 of 14 macular lesions (57%) were necrotizing, while only 5 of 27 extramacular lesions (19%) were necrotizing retinochoroiditis ( P = .017).
