Abstract
Keywords
Introduction
Ocular toxoplasmosis is a disease caused by the protozoan parasite Toxoplasma gondii. After gaining access to the eye through the retinal microvasculature, 1 T. gondii may cause a focal, necrotizing retinochoroiditis, classically associated with “fluffy” white or yellow fundus lesions. T. gondii is the most common cause of posterior uveitis worldwide, representing around 10% of patients presenting to uveitis services.2–4 Patients are treated with antibiotics (eg, trimethoprim-sulfamethoxazole or pyrimethamine-trisulfapyrimidine) with or without oral corticosteroids, though evidence regarding visual outcomes is relatively limited.5,6 In a survey of the American Uveitis Society regarding management of active ocular toxoplasmosis, 9 different drugs and 24 different regimens were described for treatment. 7 As a result, some clinicians may reserve treatment for symptomatic patients or those with at-risk posterior pole lesions. In immunocompetent patients, lesions generally resolve in 1 to 2 months, 6 whereas more severe retinochoroiditis can be found in immunosuppressed patients. 1 Some clinicians reserve treatment for symptomatic patients or those with at-risk posterior pole lesions.
A severe complication of ocular toxoplasmosis is retinal detachment (RD), which occurs in approximately 6% of patients 8 and is usually associated with severe vision loss, especially with macular toxoplasmosis involvement. RD can be rhegmatogenous (RRD) due to increased vitreoretinal interface anomalies or retinal necrosis from ocular toxoplasmosis, which could lead to retinal breaks,9,10 serous/exudative from the inflammatory component of the retinochoroiditis, which can resolve with treatment,11,12 or tractional (TRD) from proliferative membranes, causing traction on the surface of the retina. RRD and TRD can be amenable to pars plana vitrectomy (PPV) with or without scleral buckle, although fewer than 100 cases have been reported. 13 In a recent systematic review and meta-analysis by Amaral et al, 13 a lack of information was found regarding the extent of the reported RD cases, which may confound long-term visual prognosis, as well as the risk factors at presentation that influence surgical outcomes. This study aims to describe the initial presentation of ocular toxoplasmosis and the surgical outcomes associated with RRD and TRD.
Methods
Study Design and Population
This was a retrospective single-center case series of consecutive patients diagnosed with ocular toxoplasmosis at Wills Eye Hospital / Mid Atlantic Retina who subsequently developed RD between January 2013 and October 2022. Patients who were followed for at least 3 postoperative months at our institution were included in the study. Diagnosis of ocular toxoplasmosis was based on the clinical evidence of active primary or reactivated retinochoroiditis with or without the presence of an adjacent old chorioretinal scar, confirmed with serologic testing or polymerase chain reaction (PCR) on aqueous humor or vitreous fluid. Ocular PCR was also performed to exclude other possible viral etiologies. Patients who had developed serous or exudative RD were excluded from this analysis.
Data Collection
Complete baseline patient characteristics, characteristics of the ocular toxoplasmosis and the RD, and long-term outcomes were collected for each patient. For patients who underwent surgery for their RD, intraoperative and postoperative outcomes were also collected. Baseline characteristics included age at the time of RD development, sex, race, laterality, lens status, high myopia >6 diopters, smoking history, diabetes, immunosuppression, baseline best-available visual acuity (VA) with pinhole acuity or with habitual correction, intraocular pressure (IOP), previous history of glaucoma, previous retinochoroiditis, previous RD, previous eye trauma, and previous intraocular surgery.
Characteristics of ocular toxoplasmosis included presence of active retinochoroiditis at diagnosis, initial presentation (ie, retinochoroiditis or RD), presumed congenital ocular toxoplasmosis, pharmacologic treatment, location of ocular toxoplasmosis lesion, proximity of lesion relative to the retinal break, and recurrence of ocular toxoplasmosis during follow-up. The ocular toxoplasmosis lesion was located as defined by Holland et al 14 in which zone I represents the retina at 1500 μm from the optic nerve head margin, zone II represents the retina from zone I to the anterior border of the vortex veins, and zone III represents the retina from zone II to the ora serrata.
Characteristics of RD included the delay between diagnosis of ocular toxoplasmosis and RD, the type of RD (ie, RRD or TRD), macular involvement, presence of preoperative posterior vitreous detachment (PVD), location of the presumed causative break (eg, inferior RD with a break between 5 o’clock and 7 o’clock), the cause of the break (ie, toxoplasmosis scar or acute retinal necrosis), extent of the RRD in clock hours, presence of posterior pole break, presence of a macular scar or lesion, the number and type of breaks (ie, horseshoe tears, holes, dialysis), proliferative vitreoretinopathy (PVR) grading based on the updated Retina Society classification (1991) in cases of RRD, presence of vitreous hemorrhage, and presence of epiretinal membrane (ERM).
Intraoperative characteristics included type of surgery (ie, scleral buckle, PPV, or PPV-scleral buckle), gauge size for PPV cases, perfluorocarbon liquid use, drainage retinotomy, 360-degree peripheral laser, membrane peel, as well as the use and type of gas or silicone oil tamponade. A 360-degree encircling band was used for scleral buckle in all cases.
Postoperative outcomes included single-surgery anatomic success (defined as absence of return to the operating room during follow-up for recurrent RD), number and reason for surgeries during follow-up, final retina status, as well as postoperative best-corrected VA (BCVA) at 3 months, 6 months, 1 year, and final follow-up.
Statistical Analysis
Categorical variables were presented as frequencies (percentages), while continuous data were presented as median [first quartile, third quartile] in the case of nonnormally distributed variables and mean ± SD for normally distributed variables. Normal distribution was tested using the Shapiro-Wilk test. 15
Vision was converted from Snellen notation to logMAR for the purposes of analysis. 16 VA of counting fingers, hand motion, light perception, and no light perception were assigned Snellen values of 1/200, 0.5/200, 0.25/200, and 0.125/200 (logMAR equivalents, 2.3, 2.6, 2.9, and 3.2, respectively).17,18 The Wilcoxon signed-rank test was used to verify whether significant differences could be found between baseline and the different BCVA timepoints (ie, 3 months, 6 months, 1 year, and final follow-up) for each patient. 19
A multiple linear regression was built predicting final BCVA using ocular toxoplasmosis and RD characteristics, specifically considering presence of active retinochoroiditis, type of RD, macular involvement, extent of RD, and number of breaks. Given the sample size, a maximum of 2 subjects per predictor variable was considered for inclusion. 20 Variables were then selected using backward elimination with a threshold of P > .20. Unstandardized coefficients with 95% CI, as well as standardized coefficients, are presented. Statistical analyses were performed using IBM SPSS Statistics for Windows (version 29.0, IBM Corp). Analyses were conducted at the 0.05 significance level.
Results
Baseline Characteristics And Follow-Up
Between January 2013 and October 2022, a total of 422 patients were diagnosed with ocular toxoplasmosis, of which 36 patients (8.5%) developed RD. Of these 36, 17 patients did not meet the inclusion criteria due to RD having occurred before diagnosis of ocular toxoplasmosis (n = 3), RD did not develop in the eye with ocular toxoplasmosis (n = 3), RD secondary to another diagnosis (eg, acute retinal necrosis unrelated to toxoplasmosis [n = 2]), or lack of data on the RD within the database (n = 9). Twenty eyes of 19 patients met the criteria for study inclusion (Table 1). There were 10 (53%) male and 9 (47%) female patients. The mean age at diagnosis of RD was 46.8 ± 25.7 years (range, 10-90 years), including 4 patients (27%) who were minors. Additionally, there was a smoking history in 5 patients (26%) and a history of diabetes in 6 (32%). Patients were followed for an average of 3.6 ± 2.4 years.
Baseline Patient Demographics, RD Characteristics, and Treatment of 19 Patients (20 Eyes) Diagnosed With Ocular Toxoplasmosis–Associated RD.
Abbreviations: BCVA, best-corrected visual acuity; CF, counting fingers; NA, not available/applicable; PVR, proliferative vitreoretinopathy; RD, retinal detachment; RRD, rhegmatogenous retinal detachment; TMP-SMX, trimethoprim-sulfamethoxazole; TRD, tractional retinal detachment.
Of the eyes included in the study, none had high myopia, most were phakic (60% [n = 12]), and 1 patient had glaucoma (5%). There was a history of previous retinochoroiditis in 4 patients (20%) and previous RD in 1 patient (5%). No patient had a history of previous intraocular trauma.
Ocular Toxoplasmosis Presentation and Treatment
At presentation, the location of the ocular toxoplasmosis lesion and macular involvement were analyzed (Table 1). A total of 13 patients (68%) were treated with oral antibiotics (trimethoprim-sulfamethoxazole or clindamycin), of which 7 (37%) received additional oral steroids. No patient was treated with steroid monotherapy. Six patients (32%) had unknown treatments for ocular toxoplasmosis before presenting with RD. Three eyes (15%) did not show presence of a lesion on initial presentation, 3 eyes (15%) did not have adequate visualization of ocular toxoplasmosis lesions due to severity of RD on initial presentation, and 3 eyes (15%) did not have their ocular toxoplasmosis lesions identified at the time of evaluation. Nine eyes (47%) were found to have a macular scar or lesion secondary to ocular toxoplasmosis.
RD Presentation and Treatment
Most ocular toxoplasmosis–associated RDs were rhegmatogenous (75% [n = 15]), and half of the eyes had active retinochoroiditis at the time of presentation (50% [n = 10]). The mean time between initial diagnosis of ocular toxoplasmosis and RD was 1.6 years [84 days, 4.3 years]. Baseline BCVA was 2.30 [0.477, 2.60] (Snellen equivalent, counting fingers), with 10 eyes (50%) having macula-off RD. Baseline VA was comparable in patients with RRD (2.30 [0.48, 2.45]) compared with TRD (2.30 [0.44, 2.60]) (P = .85).
In 15 eyes (75%), the cause of the retinal break was deemed to be secondary to ocular toxoplasmosis. In 3 eyes of 3 patients, a causative break could not be identified. One patient developed widespread fulminant toxoplasmosis retinochoroiditis, which subsequently led to the retinal break. Three eyes (14%) had a retinal dialysis. Among the eyes with RRD, 2 developed PVR, with 1 case of grade C. Four eyes (20%) had retinal breaks associated with the ocular toxoplasmosis lesion, and 16 (80%) had peripheral retinal breaks. A preoperative PVD was present in 13 eyes (65%). A PVD was present in 2 of 5 eyes (40%) with TRD and 11 of 15 eyes (73%) with RRD (P = .18).
Among the 10 eyes (50%) with macula-on RD, baseline VA was 0.70 [0.40, 2.38] and improved to 0.57 [0.46, 2.3] by final follow-up. All eyes were operated on within 5 days, except for 1 with a TRD that was operated on 53 days after initial RD diagnosis. Three eyes had TRD (30%), and 7 had RRD (70%). Surgery was PPV only in most eyes (70% [n = 7]), followed by PPV-scleral buckle (20% [n = 2]) and scleral buckle (10% [n = 1]). Surgery included silicone oil tamponade in 4 eyes (40%) (Table 2).
Surgical Technique Used in 17 Patients Operated for Toxoplasmosis-Associated RD.
Abbreviations: NA, not available/applicable; PPV, pars plana vitrectomy; RD, retinal detachment.
Outcomes
In 3 eyes, a decision was made not to operate on the RD due to the advanced pathology at the time of diagnosis and limited visual potential, including in an 11-year-old patient with bilateral RD. In this patient, 1 eye with a total RD and poor vision was managed conservatively, while the other had a TRD that was considered for surgery after resolution of the active chorioretinitis. The patient was subsequently lost to follow-up for several years and returned with an attached retina that had scarred.
In the patients who were operated on, RRD were managed within 2 [0, 4] days (range, 0-29 days) of diagnosis compared with 27 [2, 52] days (range, 1-53 days) for TRD (P = .14). Cases of RRD were treated with PPV in 6 eyes, combined PPV-scleral buckle in 5 eyes, and primary scleral buckle in 2 eyes. All 4 eyes with TRD were repaired with PPV, and all but 1 received silicone oil tamponade. Of the eyes that received silicone oil tamponade at the initial surgery, 6 of 7 (86%) eventually had removal of their silicone oil, including a combined case with cataract surgery. The remaining patient was lost to follow-up after 3 months.
Overall, 14 of 14 eyes (100%) achieved single-surgery anatomic success at 3 months, including eyes with silicone oil tamponade, while the retina status of the remaining patients was unknown due to loss to follow-up at 3 months. At all further timepoints (ie, 6 months, 1 year, and final follow-up), single-surgery anatomic success was achieved in 15 of 17 eyes (88%), as 2 eyes required reoperation for recurrent RD at 4 months and 19 months. The former occurred due to reopening of the original break, while the latter occurred due to a new break. Both occurred in patients with originally macula-on RRD. Single-surgery anatomic success was not significantly different between patients with RRD or TRD (85% [n = 11/13] vs 100% [n = 4/4]; P = .40). Overall, the final retina status was attached in 17 of 17 eyes (100%) without silicone oil.
Compared with baseline, operated patients had VA improvement to 1.15 [0.40, 2.30] (Snellen equivalent, 20/283) at 3 months (P = .09), 1.00 [0.29, 1.30] (Snellen equivalent, 20/200) at 6 months (P = .06), 1.00 [0.20, 2.30] at 1 year (Snellen equivalent, 20/200) (P = .15), and 0.60 [0.40, 2.30] (Snellen equivalent, 20/80) at final follow-up (P = .28) (Figure 1). Patients with TRD had a worse final visual outcome (2.30 [0.87, 2.75]) compared with patients with RRD (0.54 [0.18, 2.30]), but again, this was not statistically significant (P = .28). Overall, a significant improvement was found in the early postoperative course that seemed to persist in the later follow-ups, although this varied widely among patients.
Boxplots illustrating the best-corrected visual acuity in logMAR at baseline, 3 months, 6 months, 1 year, and final follow-up in 17 eyes operated for toxoplasmosis-associated retinal detachment.
Among the 9 eyes (45%) with ERM at initial diagnosis, 5 underwent a membrane peel at the time of surgery, while 2 had a membrane peel at follow-up. No other patient warranted a membrane peel for a significant ERM during the postoperative follow-up. By final follow-up, 2 of 11 (18%) phakic eyes at baseline had undergone cataract surgery and were pseudophakic. No additional patient required re-intervention for glaucoma, but 1 did require intervention for postoperative hypotony. Finally, 2 patients developed cystoid macular edema during the postoperative follow-up that had resolved by the final visit.
In the multiple linear regression model using ocular toxoplasmosis and RD characteristics for prediction of final VA, active retinochoroiditis and macular involvement were the variables included in the final model. Active retinochoroiditis was most associated with decreased vision in logMAR (B = 1.699, 95% CI, 0.622-2.776; β = 0.801, P = .005) while accounting for macular involvement (B = −0.907, 95% CI, −1.984 to 0.170; β = −0.427, P = .09).
Conclusions
In this interventional case series, the anatomic and functional outcomes of 20 eyes with ocular toxoplasmosis–associated RD were analyzed. Although this is one of the largest cohorts of ocular toxoplasmosis–associated RD in the literature, the condition is rare. We documented an incidence of RRD and TRD after ocular toxoplasmosis of 8.5%, which is comparable to that found in the literature.8,21,22 Similar to other reported cases, we also confirm a generally poor visual prognosis for these patients, with a median final VA of approximately Snellen 20/80 with very variable results, despite good anatomic results.8,23,24 Single-surgery anatomic success was achieved in 88% of patients by final follow-up, with 100% of retinas reattached without silicone oil. This is similar to pooled retinal reattachment rates in the literature, ranging from 90% to 100% for ocular toxoplasmosis–associated RD. 13
Our study suggests that active retinochoroiditis at the time of RD presentation may be associated with worse final BCVA. A similar finding was previously identified in the literature whereby presence of active vitritis was a risk factor for RD development after ocular toxoplasmosis. 24 In their original description of 9 patients with RD and 7 patients with retinal breaks, Bosch-Driessen et al 8 described severe intraocular inflammation preceding the retinal complications. This association of active retinochoroiditis and worse final VA could be both a reflection of increased vitritis undermining visualization and greater severity of associated uveitic and retinal complications (eg, cystoid macular edema, retinal necrosis, hyaloidal contraction, ERM, and PVR).
Whether clearance of active retinochoroiditis with pharmacologic treatment before surgery could improve outcomes is still unknown and would be difficult to determine with the current analysis, given the retrospective nature of the study. A large sample size would be needed to achieve adequate statistical power. RRD also requires urgent intervention, and the eye typically cannot wait for complete resolution of inflammation. In the current series, RRD were generally treated promptly, within a median of 2 days, whereas the range was greater in TRD cases, up to 53 days. The latter allows greater flexibility in planning the surgery in relation to the presence of inflammation because it may not be immediately threatening to the macula, providing time for systemic treatment to reduce intraocular inflammation before surgery. Corticosteroids were used in a third of the patients; this practice is debated but generally used to reduce severe inflammation in immunocompetent patients. 7 It could theoretically also help reduce the retinal damage incurred due to the inflammation and possibly modulate the development of PVR, although the use of steroids in RRD for the latter purpose led to mixed results.25,26 The possibility of injecting intravitreal antibiotics like clindamycin 27 for better perioperative control of the infection could be explored in patients with persistently active retinochoroiditis at the time of surgery; however, there may be some concern about toxicity with injecting into gas- or oil-filled eyes. Additionally, in patients who cannot achieve preoperative control of retinochoroiditis, recent evidence suggests that PPV can still be considered during active inflammation to control ocular toxoplasmosis when conventional therapy fails or when complications, such as RD, develop. 28
It has been shown that there are greater vitreoretinal interface anomalies in active ocular toxoplasmosis that are detectable using optical coherence tomography (OCT) but missed on clinical examination.9,10 Cicatricial ocular toxoplasmosis scars and lesions may also produce focal vitreoretinal adhesion and resultant traction, which can cause breaks at various stages of follow-up. 29 These RD can be complex to manage and may require silicone oil tamponade to maintain retinal reattachment,30,31 especially in cases of TRD, as was seen in our study, with 3 of 4 eyes with TRD being managed with silicone oil. However, long-term silicone oil tamponade is avoided when possible in uveitic eyes, given the increased risk of complications, including silicone oil emulsification. This did not occur in the current series, possibly because ocular toxoplasmosis is characterized by bouts of acute inflammation, but is a concern in patients with chronic intraocular inflammation and associated strong emulsifiers such as fibrinogen, fibrin, and gamma globulins.32,33
It is crucial to remain vigilant regarding the etiology of RD in patients with ocular toxoplasmosis, given its influence on the approach to RD repair, the high proportion of patients developing serous RD from active ocular toxoplasmosis, and the consequences of inappropriate treatment of the ocular toxoplasmosis. 34 It is thought that choroidal ischemia may contribute to the development of exudative subretinal fluid, which can range from about 25% of patients 35 to more than half of patients when screening with OCT. 11 These patients do not require surgical intervention and may have complete resolution of their serous RD after appropriate pharmacotherapy for ocular toxoplasmosis. 12 On the other hand, TRD are typically repaired with vitrectomy, and the goal is to release the traction without creating iatrogenic retinal breaks. Once the traction is released, the retinal pigment epithelium pumps the fluid out to attach the retina. Long-term tamponade is not required in such cases. However, as we saw in our study, most cases of RD in ocular toxoplasmosis have rhegmatogenous components. A primary scleral buckle is preferred for young patients or eyes with attached posterior hyaloid and a peripheral break. However, most eyes in our series had a PVD, so a PPV-based procedure was performed. A concurrent scleral buckle is recommended in eyes with tractional components or at high risk for PVR.
Given the severe visual prognosis of ocular toxoplasmosis–associated RD despite anatomic success and the worse visual prognosis associated with active toxoplasmosis retinochoroiditis, antimicrobial prophylaxis for patients with ocular toxoplasmosis is another consideration. Although antimicrobial prophylaxis has not yet been shown to improve final VA or reduce lesion size, it was found to reduce retinochoroiditis recurrence. 36 Recurrent bouts of inflammation could subsequently increase the risk of RD, which could be important in monocular patients and patients with fovea-threatening lesions who are at risk of severe vision loss in the case of reactivation near a previous scar. In selected patients, antibiotic prophylaxis may therefore be considered.
This study has limitations. The retrospective nature of this study limited our ability to include patients who had minimal information in the electronic medical record, which was the case of nearly half the patients who had developed ocular toxoplasmosis–associated RD at our center. This includes certain variables that could have been of interest, such as inflammation of the preoperative anterior chamber during the active episode. At follow-up, given that certain patients were followed elsewhere, we may not have the full records to be able to provide detailed incidences of certain postoperative complications, such as glaucoma. Additionally, the variable presentation of disease severity complicates the task of summarizing the clinical picture of ocular toxoplasmosis–associated RD. Time from the initial diagnosis of ocular toxoplasmosis to the development of RD could not be established in a third of cases, given that many patients presented from outside institutions with no record of the initial diagnosis date. Similarly, the small sample size makes it difficult to establish the time between initial surgery and silicone oil removal, given the complex course of many of these patients, and with some needing multiple surgeries.
RRD or TRD occurred in 8.5% of patients after ocular toxoplasmosis. Anatomic results were favorable with surgical intervention, but functional outcomes remained guarded, especially in eyes with active toxoplasmosis retinochoroiditis at the time of RD presentation and those with macular involvement. Further work could help best select patients in whom ocular toxoplasmosis treatment should be optimized before surgery and in whom antibiotic prophylaxis may be considered to preserve vision.
Footnotes
Author Note
Previous versions of this work were presented at the 2023 Association for Research in Vision and Ophthalmology meeting, April 2023, New Orleans, LA, USA.
Ethical Considerations
This study adheres to the tenets of the Declaration of Helsinki, was compliant with the Health Insurance Portability and Accountability Act of 1996, and received institutional review board approval.
Consent to Participate
Informed consent to participate was waived.
Consent for Publication
Not applicable.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: J. Arch McNamara, MD Fund for Research and Education (Wills Eye Hospital, Philadelphia, PA).
Declaration of Conflicting Interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Dr. Hébert has received research grants from Bayer. Dr. Deaner is a consultant for AbbVie, Alimera, Bausch + Lomb, EyePoint, Genentech, and Regeneron. Dr. Klufas is on the speakers bureau of Genentech and Regeneron; is a consultant for Allergan, Alimera, DORC, Genentech, Regeneron, and Zeiss; and received an educational grant from Bausch + Lomb. Dr. Ho is a consultant for and received research grants from Alcon Laboratories and AcuSurgical. Dr. Hsu is a consultant for Astellas and Gyroscope Therapeutics; and received grant support from Astellas, Genentech/Roche, and Stealth Biotherapeutics. Dr. Garg is a consultant for Allergan, American Academy of Ophthalmology, Apellis Pharmaceuticals, Bausch + Lomb, Boehringer Ingelheim, Johnson & Johnson, Merck Manuals, Regeneron Pharmaceuticals, Inc., West Pharmaceutical Services, and Zeiss; is on the speakers bureau of Apellis; and has research contracts with Alcon, Apellis, Boehringer Ingelheim, Genentech, Kodiak Biosciences, NGM Bio, Regeneron, and RegenexBio. Dr. Xu is a consultant for AbbVie, Apellis, Bausch + Lomb, DORC, and Gyroscope Therapeutics. Dr. Yonekawa is a consultant for Alcon, Bausch + Lomb, Genentech, Long Bridge Medical, Pykus, and Versant Health; and received research grants from EyeBio, Genentech, Kyowa Kirin, Ocugen, and Regeneron. None of the other authors declared potential conflicts of interest with respect to the research, authorship, and/or publication of the article.