Purpose
To describe the clinical outcomes of patients with cytomegalovirus (CMV) retinitis and AIDS treated with ganciclovir implant.
Design
Retrospective cohort study.
Methods
The charts of 115 patients (166 affected eyes) with CMV retinitis treated with ganciclovir implant in the Division of Ocular Immunology, Wilmer Eye Institute from April 1996 through November 2009 were reviewed. Ophthalmologic data collected included visual acuity, ocular complications, treatment, and presence of immune recovery. Kaplan-Meier analyses and Cox regression models were used to investigate relationships between potential risk factors and ocular outcomes.
Results
At implantation, 55% of patients were prescribed highly active antiretroviral therapy (HAART), 21% were formerly on HAART, and 24% were HAART-naïve. One hundred sixty-six eyes received 257 ganciclovir implants. Fifty-seven of the implanted eyes were diagnosed with a total of 126 ocular complications after implant surgery (rate = 0.19/eye-year [EY]), the 3 most common being cataract, vitreous hemorrhage, and retinal detachment. Despite these ocular complications, the development of severe vision loss (≥6 lines lost) was low (0.005/EY). Patients with immune recovery during follow-up were less likely to have ocular complications after implant surgery; however, only the risk reduction for retinal detachment achieved statistical significance (hazard ratio = 0.29, 95% CI: 0.08, 0.98).
Conclusions
Our data suggest that ocular complications after implant surgery, including cataract, vitreous hemorrhage, and retinal detachment, were relatively common after ganciclovir implantation but severe vision loss after surgery was low. Presence of immune recovery may lessen the risk of postoperative ocular complications.
Cytomegalovirus (CMV) retinitis is the most common ocular opportunistic infection among patients with AIDS in the United States despite a 75% to 90% reduction in incident cases since the widespread availability of highly active antiretroviral therapy (HAART). An intraocular implant for the sustained release of ganciclovir was approved by the Food and Drug Administration in 1996 as a treatment for CMV retinitis (Vitrasert; Bausch & Lomb, Madison, New Jersey, USA) and has been demonstrated to be effective in halting the progression of the disease. Because of the short survival time of the patients at the time the implant was approved, the aim of the treatment was to stop progression of disease and, ultimately, blindness without necessarily concern for long-term ocular complications after implant surgery. However, the widespread availability of HAART has resulted in a reduction in mortality among patients with AIDS, and thus has made understanding the potential for ocular complications of the implant over a longer period of observation important. Our study assessed the clinical outcomes of an urban HIV/AIDS population diagnosed with CMV retinitis and treated with the ganciclovir implant. We focused on the complications associated with the implant as well as how treatment with HAART and certain disease factors affected the incidence of certain ocular complications. Given that implants were placed during the HAART era, we also investigated the impact of immune recovery on complication rates.
Methods
Study Population
A retrospective chart review was conducted of patients that received treatment at the Division of Ocular Immunology, Wilmer Eye Institute for AIDS-related CMV retinitis. The charts of 115 patients who underwent ganciclovir implant surgery for treatment of AIDS-related CMV retinitis between April 1996 and November 2009 were reviewed. These patients were diagnosed with CMV retinitis while under standard follow-up care or upon first presentation to the ocular immunology clinic. All patients received at least 1 ganciclovir implant performed at the Wilmer Eye Institute as part of their treatment regimen for CMV retinitis.
Data Collection
Demographic, medical, and ophthalmologic data were collected from each patient chart. Demographic data included date of presentation to the ocular immunology clinic, age, sex, race, and, when applicable, date of death. Medical data included HIV risk factors, date of AIDS diagnosis, and history of extraocular CMV disease. A medication review captured the patient’s HAART and anti-CMV therapy at each clinic visit. Laboratory data included: nadir CD4+ T cell count; HIV viral load and CD4+ T cell count at time of CMV retinitis diagnosis; and CMV viral load, when available. For the purposes of this study, immune recovery was defined as an increase in the CD4+ T cell count to a level of ≥100 cells/μL in a patient receiving HAART. Ophthalmologic data collected specified the date of CMV retinitis diagnosis, presence and development of bilateral retinitis, location and percentage of retinal area involved, the date of ganciclovir implant(s), and visual acuity (VA) at the visit immediately preceding implant surgery and at each clinical visit after the implant surgery. Postoperative ocular complications were collected at each clinic visit after implant surgery. Each patient contributed at least 1 study eye; for those presenting with bilateral disease, both eyes were included in the study. The percentage of retinitis was calculated based on a standard classification system where the location of the lesion(s) is divided into 3 zones. Zone 1 describes the area where the retinitis is most sight-threatening and corresponds to an area 3000 μm from the fovea or 1500 μm from the margin of the optic nerve head. For zone 2, the lesion must extend from the zone 1 anterior border to the borders of the ampullae of the vortex veins. Zone 3 retinitis extends anteriorly from zone 2 to the ora serrata.
Main Outcome Measures
The main outcome measure was the development of an ocular complication occurring after placement of a ganciclovir implant. Postoperative complications were collected from the documentation in the patient’s medical record. For this study the complications were clinically established with standard parameters. New-onset cataract was defined as the presence of ≥1+ opacity in the lens in an eye that was free of cataract prior to implant surgery. Eyes with pre-existing cataract were censored from the analysis. Vitreous hemorrhage was established as a postoperative complication if vitreous bleeding was not documented in the operative report and was reported at a postoperative visit. Visual acuity loss was categorized as either loss of ≥3 lines of vision or a loss of ≥6 lines of vision. Visual acuity scores were determined using the logMAR formula and we used the method to quantitate VA loss of count fingers or worse as outlined by Kempen and associates. As transient visual loss is a common result of implant surgery, only losses that persisted through and outside of the 28-day postoperative period were included in the VA loss complication count. Lastly, clinical examination documented the following conditions: retinal detachment; uveitis; epiretinal membrane formation; hypotony; endophthalmitis; and presence of cystoid macular edema (CME), which was confirmed by fluorescein angiogram or optical coherence tomography when available. Immune recovery uveitis was defined as the presence of intraocular inflammation (either in the anterior chamber, the vitreous, or both) in a patient who has also undergone immune recovery.
Statistics
Complication rates and cumulative incidences at 1 and 2 years after ganciclovir implant surgery were determined through survival analysis with 95% confidence interval (CI) calculated using Poisson regression to account for the low number of events. For each study eye, time to complication was calculated from the date of implant to the clinic visit date when the complication was first discovered. Incidence rates were defined as the number of events divided by the number of eye-years (EY) at risk. Missing data were not imputed but were low (2% missing). Kaplan-Meier analyses and Cox proportional hazards regression models were used to investigate relationships between potential risk factors and ocular outcomes. Log rank tests were performed to assess the difference in Kaplan-Meier curves between categorical variables. Robust variance estimation was used in all Cox regression models in order to account for correlation between eyes in patients who presented with bilateral CMV retinitis. P values were nominal and 2-sided, with a P value of <.05 considered to be statistically significant. Analyses were performed with the Intercooled Stata 11.0 statistical package (Stata Corp, College Station, Texas, USA).
Results
Study Population Characteristics
Patient- and ocular-specific characteristics from within 1 month of CMV retinitis diagnosis are summarized as Tables 1, 2, and 3 . Sixty-four percent of 115 patients were men; 68% of patients were African-American. The most common HIV risk factor reported was men having sex with men (MSM) in 33%. Approximately 80% of the population had CD4+ T cell counts less than 50 cells/μL at the time of CMV retinitis diagnosis. The median duration of time between AIDS diagnosis and development of CMV retinitis was 1.1 years (range, 0 to 16.7 years) irrespective of the presence of HAART. Within our cohort, patients presented with multiple-zone retinitis involvement, with approximately 40% having zone 1 disease (ie, retinitis involving the macula and/or optic nerve) as the most posterior zone involved. Almost half of the cohort (n = 51) presented with bilateral CMV retinitis and 7 patients developed contralateral disease during the study period. The median survival time of the cohort from the time of implant was 1.8 years (range, 0 to 13 years), with a mortality rate of 0.23 per person-year (PY). During the postoperative follow-up period (range 1 month to 14.4 years), 61% of the cohort (n = 70) died. Of those who were alive at the end of the follow-up period, the median follow-up period was 2.5 years, while of those that died during follow-up, the median time was 0.9 years.
| Patient-Specific Characteristics (N = 115) | |
| Median age, years (range) | 38 (21–56) |
| Race, % | |
| White, non-Hispanic | 28.0 |
| Black, non-Hispanic | 68.4 |
| Hispanic | 2.6 |
| Sex, % | |
| Male | 64.3 |
| Female | 35.7 |
| HIV risk factors, % | |
| Men having sex with men | 33.3 |
| Injection drug use | 19.0 |
| Heterosexual contact | 20.9 |
| Median time between AIDS diagnosis and CMV retinitis diagnosis, years | 1.1 (0–16.7) |
| Median nadir CD4+ T-cell count (cells/μL) | 8 |
| CD4+ T-cell count (cells/μL) | |
| Median | 12 |
| Range | 0–295 |
| % patients with CD4+ T-cell count <50 | 78.0 |
| Immune recovery, % a | 33 |
| HIV viral load copies/mL | |
| Median | 203 000 |
| Range | 0 to 750 000 |
| Bilateral CMV retinitis, % | 44.3 |
| Eye-Specific Characteristics (N = 166) | |
| Visual acuity | |
| Median | 20/50 |
| Range | 20/10 to NLP |
| Categories, % | |
| Better than 20/50 | 55.5 |
| 20/50–20/200 | 24.3 |
| Worse than 20/200 | 20.2 |
| Extent of CMV retinitis lesion size, % | |
| < 25 | 69.2 |
| ≥ 25 | 30.8 |
| Location of CMV retinitis, % | |
| Zone 1 | 39.6 |
| Zone 2 | 31.7 |
| Zone 3 | 12.2 |
| Unknown | 16.5 |
| Immune recovery uveitis, % | 1.8 |
a Immune recovery defined as a CD4+ T cell count >100 cells/μL.
| Patient-Specific Characteristics (N = 115) | ||
|---|---|---|
| At Time of CMV Retinitis Diagnosis, n (%) | At Time of Ganciclovir Implant, n (%) | |
| HAART | ||
| Never | 42 (36%) | 28 (24%) |
| Former | 25 (22%) | 24 (21%) |
| Current | 48 (42%) | 63 (55%) |
| Anti-CMV treatment a | ||
| Ganciclovir | 81 (70%) | NA |
| Foscarnet | 10 (9%) | NA |
| Cidofovir | 6 (5%) | NA |
| None | 30 (26%) | NA |
| Unknown | 4 (4%) | NA |
| Known ganciclovir resistance (%) | 0.05 | NA |
a Percentage total surpasses 100% because of combinations of anti-CMV therapies.
| Patient-Specific Characteristics (N = 81) | |
|---|---|
| Route, n (%) | |
| Oral ganciclovir | 21 (26%) |
| Intravenous | 11 (18%) |
| Oral valganciclovir | 19 (23%) |
| Intravitreal | 8 (12%) |
| Unknown | 22 (35%) |
| Duration (months) of treatment with ganciclovir prior to implant surgery | |
| Mean ± SD | 2.69 ± 3.9 |
| Median (range) | 1 (0 to 21) |
| Known ganciclovir resistance, (%) | 0.05 |
a Percentage total surpasses 100% because of combinations of routes of ganciclovir treatment.
Treatment history for both HIV/AIDS and CMV infection are summarized in Tables 2 and 3 . At the time of CMV retinitis diagnosis over half of the study population (n = 68) was HAART-naïve. However, between CMV retinitis diagnosis and ganciclovir implant surgery, 14 of the HAART-naïve patients (33%) began antiretroviral therapy. More than 70% received some form of anti-CMV therapy at the time of CMV retinitis diagnosis. Ganciclovir was the most commonly prescribed anti-CMV therapy, with the median duration of ganciclovir treatment prior to implant surgery being 1 month (range, 0 to 21 months). Nearly 40% of the patient population (44 patients) required multiple implants and of those, only 2 patients had achieved immune recovery ( P = .05).
Incidence Rates of Ocular Complications
The study population experienced a total of 126 postoperative ocular complications in 57 eyes, with the rate of developing an ocular complication of 0.19/EY and cumulative incidences of 35% and 38% at 1 and 2 years after implant surgery respectively ( Table 4 ). Formation of cataract was the most common postoperative complication, with a median time to development of 10 months. Cataract was diagnosed in 25 study eyes for an incidence rate of 0.07/EY (95% confidence interval [CI]: 0.07/EY, 0.11/EY) and a cumulative incidence of 17% at 2 years after implant surgery. Vitreous hemorrhage, the second most common complication, was diagnosed in 24 eyes for an incidence rate of 0.06 EY (95% CI: 0.04/EY, 0.10/EY) and a cumulative incidence of 15% at 2 years after implant surgery. The majority of eyes with vitreous hemorrhage (18 of 24 eyes) were diagnosed with vitreous hemorrhage within 4 months of implant surgery. Retinal detachment after implant surgery occurred in 21 eyes for an incidence rate of 0.06/EY (95% CI: 0.04/EY, 0.09/EY) and a cumulative incidence of 15% at 2 years after implant surgery. Fortunately, the loss of VA was relatively low in our population, with 3 eyes having a loss of ≥3 lines of vision (rate = 0.007/EY; 95% CI: 0.002/EY, 0.02/EY; cumulative incidence = 2% at 2 years after implant surgery) and 2 eyes having a loss of ≥6 lines of vision (rate = 0.005/EY; 95% CI: 0.001/EY, 0.02/EY; cumulative incidence = 1% at 2 years after implant surgery). Minor postoperative ocular complications included 19 cases of subconjunctival hemorrhage and 6 cases of exposed surgical sutures.
