Routine Eye Screening by an Ophthalmologist Is Clinically Useful for HIV-1-Infected Patients with CD4 Count Less than 200 /μL

Objective To investigate whether routine eye screening by an ophthalmologist in patients with HIV-1 infection is clinically useful. Methods A single-center, retrospective study in Tokyo, Japan. HIV-1-infected patients aged over 17 years who visited our clinic for the first time between January 2004 and December 2013 and underwent full ophthalmologic examination were enrolled. At our clinic, ophthalmologic examination, including dilated retinal examination by indirect ophthalmoscopy was routinely conducted by ophthalmologists on the first visit. The prevalence of ophthalmologic diseases and associated factors including the existence of ocular symptoms were analyzed. Results Of the 1,515 study patients, cytomegalovirus retinitis (CMV-R) was diagnosed in 24 (2%) patients, HIV retinopathy (HIV-R) in 127 (8%), cataract in 31 (2%), ocular syphilis in 4 (0.3%), and uveitis with unknown cause in 8 (0.5%). Other ocular diseases were diagnosed in 14 patients. The CD4 count was <200 /μL in all CMV-R cases and 87% of HIV-R. The prevalence of any ocular diseases, CMV-R, and HIV-R in patients with CD4 <200 /μL were 22%, 3%, and 15%, respectively, whereas for those with CD4 ≥200 /μL were 5%, 0%, and 2%, respectively. No ocular symptoms were reported by 71% of CMV-R cases and 82% of patients with any ocular diseases. Conclusions Routine ophthalmologic screening is recommended for HIV-1-infected patients with CD4 <200 /μL in resource-rich settings based on the high prevalence of ocular diseases within this CD4 count category and because most patients with ocular diseases, including those with CMV-R, were free of ocular symptoms.


Introduction
Antiretroviral therapy (ART) has dramatically improved the prognosis of patients with HIV-1 infection [1]. However, many patients are still diagnosed of advanced HIV-1 infection, with concurrent opportunistic infections [2,3]. HIV-1-infected patients are prone to develop ocular opportunistic infections, such as cytomegalovirus (CMV) retinitis, cryptococcosis, toxoplasmosis, and tuberculosis [4][5][6]. HIV retinopathy can also occur in patients with HIV-1 infection [7]. Of these ocular opportunistic infections, CMV retinitis occurs among patients with severely advanced HIV-1 infection, and can result in total blindness and increased mortality even in the ART era [8].
To date, there is no consensus on the importance of routine screening of HIV-1-infected patients for ophthalmologic diseases with dilated retinal examination using indirect ophthalmoscopy. The revised 2013 American CDC guidelines for opportunistic infections did not state recommendations for routine ophthalmologic screening for such patients and only described the expert opinion that "some specialists recommend yearly funduscopic examinations performed by an ophthalmologist for patients with CD4 counts <50 /μl" [9]. The aims of the present study were; 1) to determine the prevalence of ocular diseases according to CD4 count and 2) to determine factors associated with CMV retinitis and other ocular diseases, with a special focus on the presence of ocular symptoms, in order to assess the clinical utility of routine ophthalmologic screening by an ophthalmologist for patients with HIV-1 infection in a resource-rich setting.

Study design, setting, and participants
We conducted a single-center retrospective study to investigate the usefulness of routine ophthalmologic screening for patients with HIV-1 infection at AIDS Clinical Center, National Center for Global Health and Medicine (NCGM), Tokyo. AIDS Clinical Center is one of the largest referral centers for HIV-1 infection in Japan with more than 3,800 registered patients [10], and considering that the total reported number of patients with HIV-1 infection is 24,454 by the end of 2013, this clinic treats approximately 15% of the HIV-1 infected patients in Japan [11]. At our clinic, ophthalmologic examination, including dilated retinal examination using indirect ophthalmoscopy, is routinely conducted during the first visit to the clinic by an ophthalmologist, and if it is not possible during the first visit, patients are referred to an ophthalmologists during the second/third visit [12]. The following criteria were applied for enrollment of patients in the study. Inclusion criteria: 1) HIV-1-infected patients aged over 17 years who visited our clinic for the first time between January 2004 and December 2013 and underwent full ophthalmologic examination within one year from the first visit. Exclusion criteria: 1) patients who were already diagnosed with ocular diseases at the time of referral to our clinic, because the aim of this study was to evaluate the usefulness of routine ophthalmologic screening. Furthermore, it is often difficult to confirm retinal photography which are required for the diagnosis of CMV retinitis according to the standard ACTG criteria [12,13].
The study protocol was approved by the Human Research Ethics Committee of National Center for Global Health and Medicine (G-001623-01). Informed consent was waived because this study solely used the data gained from clinical practice. Patient information was anonymized and de-identified prior to analysis. The study was conducted according to the principles expressed in the Declaration of Helsinki.

Definitions of ophthalmologic diseases and measurements
The results of the first ophthalmologic examination conducted in each patient were extracted from the medical records, together with any or no ocular symptoms. CMV retinitis which fulfilled the standard ACTG criteria of "confirmed CMV retinitis" was included as CMV retinitis cases which required a diagnosis by an experienced ophthalmologist and documentation of CMV retinitis by retinal photography [13]. The diagnosis of HIV retinopathy was based on the findings of microaneurysms, telangiectasia, retinal hemorrhages, and cotton wool spots (CWS) [7]. The diagnosis of ocular syphilis was based on 1) ophthalmological examination documenting findings specific to ocular syphilis and 2) evidence of syphilis infection defined by a positive serum quantitative rapid plasma reagin (RPR) (titer 8) and positive Treponema pallidum hemagglutination assay [14,15]. When ruling out other ocular diseases was difficult, the response to syphilis treatment was also taken into account. At our hospital, if the diagnosis of ocular diseases could not be confirmed, ophthalmological examination was repeated within one to four weeks, and the diagnosis was confirmed by at least two ophthalmologists. Old lesions considered inactive were excluded. All ocular diseases were reviewed and confirmed by an experienced ophthalmologist (SY) and a HIV expert (TN). The basic characteristics [age, sex, ethnicity, the day (if not available, the month) of diagnosis of HIV-1 infection, history of AIDS, route of HIV-1 transmission, and treatment status of HIV-1 infection (either treatmentnaïve or experienced)], CD4 count, HIV-1 viral load, quantitative RPR were also collected. For the latter three variables, the data closest to and preceding the day of the first ophthalmologic examination were used. Quantitative RPR was routinely measured on the day of the first visit and on discretion of treating physician at our clinic. The variable CMV end-organ diseases other than retinitis was also collected; their diagnosis of each disease was based on the standardized ACTG criteria and confirmed within 4 weeks of ophthalmological examination [12,13]. The variables systemic steroid use, anti-CMV treatment, and chemotherapy were also recorded. They were defined as steroid, anti-CMV treatment, and chemotherapy, which were administered either orally or intravenously within one month preceding the ophthalmological examination [12].

Statistical analysis
All analyses were performed and all results were presented at the subject-level, rather than eye level. Baseline characteristics were described for the entire study patients, those with CMV retinitis, HIV retinopathy, cataract, and ocular syphilis. A univariate logistic regression model was constructed to estimate the association of each variable with CMV retinitis, HIV retinopathy, and cataract, respectively, and variables with p <0.05 in univariate analysis were incorporated into the multivariate model. For CMV retinitis, the variables anti-CMV treatment and history of AIDS were not added to the multivariate model because of multicollinearity with CMV diseases other than retinitis and CD4 count, respectively, and because of the small number of cases with CMV retinitis. For cataract, the variable systemic steroid use was added to the multivariate model because systemic steroid use is an established risk factor for cataract [16,17]. Sex was not added to any of the models because the study included only a small number of female patients. Statistical significance was defined as two-sided p values <0.05. We used odds ratios (ORs) with 95% confidence intervals (95% CIs). All statistical analyses were performed with The Statistical Package for Social Sciences ver. 21.0 (SPSS, Chicago, IL).

Discussion
In this single center cohort where an ophthalmologist routinely screens ocular diseases with dilated retinal examination using indirect ophthalmoscopy for HIV-1-infected patients who first visited the clinic [12], we aimed to determine whether routine ophthalmologic screening is clinically useful. The results of the present study showed the presence of ocular diseases and CMV retinitis in 26% and 5% of patients with CD4 <50 /μL, and in 22% and 3% of patients with CD4 <200 /μL, respectively, but only in 5.4% and 0% of patients with patients with CD4 count 200 /μL. Of the 24 patients with CMV retinitis, 14 patients had CD4 count <50 /μL, whereas 6 patients had CD4 50-99 /μL and 4 patients had CD4 100-199 /μL. Importantly, only 27 (13%) of 204 patients with any ocular diseases and 7 (29%) of 24 patients with CMV retinitis had ocular symptoms. The importance of early detection of ocular diseases, especially CMV retinitis is well established [18]. Based on the high prevalence of ocular diseases among patients with CD4 <200 /μL and high percentage of asymptomatic patients among those with ocular diseases, we recommend ophthalmological screening with dilated retinal examination using indirect ophthalmoscopy in HIV-1-infected patients with CD4 count of <200 /μL in resourcerich settings. To our knowledge, this is the first study that implemented systemic screening by ophthalmologists for all HIV-1-infected patients during the first visit to the HIV referral clinic regardless of CD4 count. The study design allowed comparison of the prevalence of ocular diseases according to CD4 count. The result showed that the prevalence of ocular diseases among HIV-1-infected patients with CD4 200 /μL was 5.4% (CMV retinitis 0%), far lower than 22% (CMV retinitis 3%) among patients with CD4 <200 /μL. Other studies that investigated the incidence and prevalence of ocular diseases among HIV-1-infected patients, such as the Longitudinal Study of the Ocular Complications of AIDS (LSOCA) [19], only enrolled AIDS patients with very low median nadir CD4 count of 30 /μL. Furthermore, in this study, most study patients underwent ophthalmological examination soon after the diagnosis of HIV-1 infection (median 1 month, IQR 0.3-2.6 months). Thus, the results of ophthalmological examination in this study would be similar to the results at the time of diagnosis of HIV-1 infection. Little evidence is available for the clinical usefulness of routine ophthalmological screening in HIV-1-infected patients. Although both the 2013 American CDC guidelines for the treatment of opportunistic infections and 2013 Primary Care Guidelines by Infectious Diseases Society of America recommend referral of patients with CD4 count <50 /μL to an ophthalmologist for examination, the recommendation is mostly based on expert opinion [9,20]. It is wellknown that CMV retinitis typically occurs in patients with CD4 count <50 /μL [21,22]. However, CMV retinitis can also occur in patients with CD4 >50 /μL [23], as in the present study the CD4 count was 50-199 /μL in 10 (42%) out of 24 CMV retinitis cases. Considering that 71% of CMV cases had no ocular symptoms, we recommend that routine screening by ophthalmologists of patients with CD4 <200 /μL is rational in resource-rich settings.
In the assessment of associated factors for each ocular disease, lower CD4 count and higher HIV-1 load were associated with HIV retinopathy, in agreement with the results of previous studies [24,25]. On the other hand, to our knowledge, this is the first study to report a significant association between old age and HIV retinopathy. HIV retinopathy is a well-known ocular disease characterized by micro aneurysms, telangiectasia, retinal hemorrhages, and CWS [7]. Pathogenesis and clinical significance of HIV retinopathy are not fully understood, and number of probable etiologies or associated factors, including viral immune complex, direct HIV infection of endothelial cells, and hyperviscosity due to red cell aggregation, have been suggested [7,26,27]. Clinically, most patients with HIV retinopathy do not have visual complaints [7] and CWS can disappear within a few weeks [28]. However, abrupt visual loss can occur in patients with HIV retinopathy [29], and HIV-induced CWS can cause permanent retinal destruction [26]. Our finding that old age is associated with HIV retinopathy might suggest the importance of active ophthalmologic screening for the elderly with HIV-1 infection.
The association between cataract and HIV-1 infection has been poorly documented. However, one Danish nationwide population-based cohort study reported almost twice higher risk of cataract surgery in patients with HIV-1 infection than in the general population [30]. One possible explanation for the susceptibility of HIV-1-infected patients to cataract is the occurrence of immune recovery uveitis/vitritis after the introduction of ART, with subsequent development of secondary ocular manifestations, including cataract [30][31][32]. The prevalence of cataract in this study was 2%, but it is difficult to directly compare this number with other studies, because older age is an established risk factor for cataract [33] and such comparison requires age matching. Although the Danish study identified that CD4 count 200 /μl and introduction of ART as risk factors for cataract [30], these findings were not reproduced in the present study.
Several limitations need to be acknowledged. First, due to the nature of a single-center cohort study, selection bias of study patients could not be ruled out. However, due to the very low prevalence of HIV-1 infection in Japan, our clinic treats approximately 15% of the HIV-1 infected patients in Japan [10,11]. Second, the majority of patients diagnosed with ocular diseases in the present study had HIV retinopathy, for which the clinical significance and sequelae are still not fully elucidated as described above. However, HIV-induced CWS can serve as a portal for CMV entry to the retina [34], and HIV retinopathy is an established risk factor for subsequent CMV retinitis [5,34,35]. Some experts even recommend ophthalmological examination every three months for patients with HIV retinopathy [7,27]. Third, 34% of the HIV-1-infected patients on the first visit to our clinic during the study period, including 12 patients with pre-diagnosed CMV retinitis, were excluded from the study (Fig 1), mostly because they did not receive ophthalmologic examination. If the characteristics of these patients differ substantially from the included cohort, this could influence the interpretation of the results.
In conclusion, the present study reported the results of systematic screening of HIV-1-infected patients by ophthalmologists with dilated retinal examination using indirect ophthalmoscopy among HIV-1-infected patients shortly after the diagnosis of HIV-1 infection. The prevalence of any ocular diseases and CMV retinitis were 22% and 3% in patients with CD4 <200 /μL, respectively, but were only 5.4% and 0% in patients with CD4 count 200 /μL. Furthermore, only 13% of patients with any ocular diseases and 29% of CMV cases complained of ocular symptoms. We recommend routine ophthalmologic screening by ophthalmologists for HIV-1-infected patients with CD4 count of <200 /μL in resource-rich settings.