Good syndrome and other causes of cytomegalovirus retinitis in HIV-negative patients—case report and comprehensive review of the literature

Introduction

Thymoma is an uncommon and slow-growing neoplasm that accounts for approximately 20 to 30 % of mediastinal masses in adults and 1 % in children [1]. Thymic tumors not only usually present with respiratory symptoms due to compression of the upper airways and/or superior vena cava syndrome but also can produce paraneoplastic or parathymic syndromes [1–4], and the most common of which are myasthenia gravis (MG), pure red cell aplasia (PRCA), and acquired partial immune deficiency or Good syndrome [5, 6].

Good syndrome was first described by the American hematologist-oncologist Dr. Robert Good in 1956 [7]. Good noted a direct relationship between the presence of thymoma and hypogammaglobulinemia causing immunosuppression in those patients. Good syndrome typically occurs in middle-aged adults and is associated most commonly with recurrent sinus and pulmonary infections, cytomegalovirus (CMV) disease (most often retinitis), fungal infections, pure red cell aplasia, and myasthenia gravis [8]. While hypogammaglobulinemia in the setting of thymoma defines Good syndrome, other, often partial, immune deficiencies have also been described, including decreased T cell function [9].

We describe a 65-year-old woman who developed CMV retinitis (CMVR) in the setting of Good syndrome. The patient subsequently developed vitritis with cystoid macular edema (CME) despite control of the retinitis with antiviral agents. Cases of CMVR in human immunodeficiency virus (HIV)-negative patients, including those with Good syndrome, identified in PubMed through December, 2014, were reviewed and are summarized. Search terms included “Cytomegalovirus AND eye” and “cytomegalovirus retinitis.” Additional publications were identified by reviewing collected references.

Case report

A 65-year-old Thai woman presented for evaluation of suddenly decreased vision with floaters in her right eye. Past ocular history was unremarkable. Past medical history was notable for PRCA diagnosed 2 years prior to presentation and for which she was treated for four months with erythropoietin and systemic corticosteroids. She also had two prior episodes of oropharyngeal candidiasis, which were treated successfully. There was no history of recent or current corticosteroid use.

The best-corrected vision was 20/100 on the right eye and 20/25 on the left eye. Intraocular pressure (IOP) was 15 mmHg bilaterally. No afferent pupillary defect was noted. Anterior segment examination of the right eye showed trace anterior chamber cell but was otherwise normal. Anterior segment examination on the left eye was unremarkable. Posterior segment examination on the right showed moderate vitreous inflammation and an advancing edge of necrotizing retinitis associated with scattered intraretinal hemorrhages and retinal vascular telangiectasis (Fig. 1a). Posterior segment examination of the left eye was unremarkable. The patient was diagnosed clinically with viral retinitis, an anterior chamber paracentesis was performed for viral DNA testing, a laser barrier was applied immediately posterior to the area of active retinitis, and the patient was given an intravitreal injection of 2 mg of ganciclovir followed by treatment with high-dose oral valaciclovir, 2 g three times daily. Analysis of the anterior chamber paracentesis was positive for CMV DNA, and the patient was switched from valaciclovir to valganciclovir, which resulted in resolution of the area of retinitis.

Within 1 month of the diagnosis of CMVR, the patient was hospitalized for acute pneumonia. During this hospitalization, a mediastinal mass was discovered on chest X-ray and evaluated further by chest CT (Fig. 2). Computer tomography guided thymus biopsy and subsequent thymectomy were performed, revealing histological changes consistent with thymoma that lead to the diagnosis of Good syndrome. Immunoglobulin A, G, and M levels remained low at the last testing 5 months following removal of the thymus. The patient then returned with worsening vision in the eye with CMVR while on maintenance valganciclovir therapy, 450 mg twice daily.

The best-corrected vision was 20/125 on the right eye and 20/25 on the left eye. Intraocular pressure was normal. No afferent pupillary defect was noted. Anterior segment examination on the right showed several stellate keratic precipitates on the corneal endothelium, one cell per high powered field in the anterior chamber, and occasional anterior vitreous cells. Anterior segment examination on the left was unremarkable. Posterior segment examination on the right showed mild to moderate vitreous inflammation, a posterior vitreous detachment, a large area of inactive retinal necrosis (Fig. 1b), continuous laser barrier scars immediately posterior to the area of retinitis, and loss of the foveal light reflex suggestive of CME. Posterior segment examination on the left was unremarkable. Fluorescein angiography confirmed the presence of severe CME on the right (Fig. 1c). Spectral domain optical coherence tomography (SD-OCT) imaging showed marked CME on the right (Fig. 1d) with a central macular thickness of 743 um. SD-OCT imaging of the left fovea revealed a normal contour with no evidence of subretinal or intraretinal fluid. The patient was treated with topical difluprednate four times daily for 1 month. The CME persisted, the difluprednate was stopped, and the patient was given two injections of 1.25 mg of intravitreal bevacizumab, 1 month apart. The CME failed to respond, and so the patient was given an intravitreal injection of triamcinolone acetonide, 2.0 mg, following which the CME resolved and vision improved to 20/80 in the affected eye. The CME subsequently recurred and the vision decreased to 20/100, but the patient refused further treatment. The retinitis remained inactive.

Comprehensive literature review

Although the retinitis in our patient responded promptly to intravitreal and systemic antiviral agents, the patient subsequently developed vitritis and CME of the type seen in patients with immune recovery uveitis (IRU) despite the fact that her total CD4+ T cell count was normal both before and after the occurrence of retinitis. In contrast, the total B cell count and immunoglobulin levels where low both before and after thymectomy. While hypogammaglobulinemia is required to diagnose Good syndrome and has been observed in all reported cases to date, including our patient, it is noteworthy that the total CD4+ T cell count was somewhat decreased in five of the eight previously reported cases with Good syndrome and CMVR (Table 3), indicating that partial CD4+ T cell depletion does occur in patients with Good syndrome and suggesting the possibility that selective loss of CMV-targeting CD4+ T cells may have occurred in our patient, facilitating the development of retinitis. To our knowledge, an IRU-like syndrome has not been reported previously following treatment of CMVR in a patient with Good syndrome.

Our review identified a total of 248 eyes of 178 patients previously reported with CMVR in the absence of either HIV infection, Good syndrome, or prior periocular or intraocular corticosteroid injection (Additional file 1: Table S1) [17–86]. Reported ages ranged from 1 week to 84 years, with a mean and median of 45.7 and 48.0 years, respectively. Men outnumbered women approximately two to one (M to F ratio = 1.88:1), and the vast majority (95.5 %) had an identifiable cause of systemic immunosuppression. The most common factors contributing to a decline in immune function included age over 60 years (33.1 %), an underlying malignancy (28.7 %), a systemic autoimmune disorder requiring treatment (19.1 %), organ (15.2 %) or bone marrow (16.3 %) transplantation requiring systemic immunosuppression, and diabetes mellitus (6.1 %). The most commonly reported cancers included leukemia (35 patients; 19.7 %) and lymphoma (14 patients; 7.9 %). Three patients had multiple myeloma (1.7 %). One patient each (0.6 %) had breast cancer and angiocentric immunoblastic lymphadenopathy with dysproteinemia. Several patients had a primary immune deficiency other than Good syndrome, including three patients (1.7 %) with severe combined immunodeficiency, two patients each (1.1 %) with unspecified primary immune deficiency and common variable immune deficiency, and one patient (0.6 %) with idiopathic CD4+ T cell lymphopenia. Among the 26 reported patients less 18 years of age, 9 (34.6 %) had acute lymphoblastic leukemia, 8 (30.8 %) had undergone bone marrow transplantation, 5 (19.2 %) had congenital CMV infection, 3 (11.5 %) had severe combined immunodeficiency and 1 each (3.8 %) had common variable immune deficiency and immunoglobulin 2 (Ig2) deficiency.

The use of systemic immunosuppressive medication was reported in 105 of 160 cases (65.6 %). Seventy-eight of these 105 patients (74.3 %) were on two or more immunosuppressive agents. Corticosteroids were the most common immunosuppressive agent used in 69 patients (65.7 %), followed by cyclophosphamide in 33 patients (31.4 %), azathioprine in 17 patients (16.2 %), vincristine in 16 patients (15.2 %), methotrexate in 15 patients (14.3 %), cyclosporine in 13 patients (12.4 %), tacrolimus in 11 patients (10.5 %), 6-mercaptopurine in 9 patients (8.6 %), mycophenolate mofetil in 8 patients (7.6 %), fludarabine in 7 patients (6.7 %), and adriamycin in 6 patients (5.7 %). Five patients each (4.8 %) used intravenous immunoglobulin G and rituximab. One patient each (1.0 %) was on mitoxantrone, ibritumomab tiuxetan, and hydroxychloroquine. Within the 105 cases reporting medication use, the use of either an antimetabolite or a leukocyte signaling inhibitor as a group (methotrexate, azathioprine, mycophenolate mofetil, tacrolimus, or cyclosporine) was reported in 46.7 % of cases, whereas chemotherapeutic agents as a whole (cyclophosphamide, vincristine, high-dose methotrexate, 6-mercaptopurine, adriamycin, rituximab, fludarabine, mitoxantrone, and ibritumomab tiuxetan) were reported in 48.6 % of cases. Clinically, the retinitis was unilateral in 108 out of 178 cases (60.7 %). Many case reports provided limited clinical information, but the location was either reported or illustrated in 97 out of the 248 eyes (39.1 %) and was found to involve zone 1 in 72 eyes (74.2 %). Additional clinical features were noted in 200 out of 248 eyes (80.6 %). Among these, 200 eyes, or just under one third (29.0 %), were noted to have anterior chamber inflammation, which was described as mild in 18 (9.0 %) and moderate in 13 (6.5 %), and severe in the remaining 27 (13.5 %). Just over one third of the reported eyes (37.5 %) were noted to have vitreous inflammation, among which inflammation was described as mild in 23 (11.5 %), moderate in 18 (9.0 %), and severe in 6 (3.0 %), with the remaining 28 (14.0 %) not quantifying severity. An occlusive vasculitis was noted in 47 eyes (23.5 %). Visual acuity at initial CMVR diagnosis was reported in 179 of 248 eyes (72.2 %). Among these 179 eyes, visual acuity at initial diagnosis was better than 20/40 in 34.1 %, between 20/40 and 20/200 in 39.1 %, and worse than 20/200 in 26.8 % of eyes. The method of diagnostic confirmation of CMVR was reported in 131 of the 178 cases (73.6 %), among which the diagnosis of CMVR was confirmed by PCR-based testing of intraocular fluids in 71.8 %. The retinitis responded to antiviral therapy in all cases. The treatment administered was reported in 126 of the 178 cases (70.8 %). Of these 126 cases, systemic treatment was administered in 45.2 % of patients, whereas both intravitreal and systemic therapy was given in 31.0 % of patients, and intravitreal treatment alone was employed in 23.8 % patients. Visual acuity at the last follow-up visit was reported in 171 eyes (mean 14.2 months; median 6.0 months; range 0 to 216 months) and was better than 20/40 in 30.4 %, between 20/40 and 20/200 in 37.4 %, and worse than 20/200 in 32.2 % of eyes.

The most complete and standardized description of CMVR in HIV-positive patients comes from the studies performed by the Studies of the Ocular Complications of AIDS (SOCA) research group [101–103]. Cytomegalovirus retinitis was the most frequently encountered complication of HIV infection in this cohort, occurring in 34.6 % of patients with CD4+ T cell counts <50 cells/μL and in 63.4 % of patients with CD4+ T cell counts <200 cells/μL [104]. Clinically, HIV-associated CMVR includes classic features of necrotizing retinitis with irregular sheathing of adjacent vessels and variable degrees of hemorrhage (sometimes referred to as “pizza pie retinopathy” or “cottage cheese with ketchup”), and which is sometimes coupled with a frost branch angiitis appearance without vascular occlusion and often associated with mild vitreous or anterior chamber inflammation [12, 20, 104]. In addition to these core clinical findings previously listed, the SOCA studies quantified the prevalence of a number of hallmark features of CMVR at presentation. Specifically, keratic precipitates were present in 36.8 % of the eyes with CMVR, anterior chamber inflammation in 46.2 %, and vitreous inflammation in 61.9 %. The prevalence of anterior chamber cells greater than 2+ was 1.9 % and vitreous haze greater than 2+ was 11.4 %. Macular edema, epiretinal membrane formation, and rhegmatogenous retinal detachment were each uncommon at presentation in the SOCA cohort, occurring in less than 10 % of the eyes, and posterior synechiae formation was not observed [104].

A number of studies have suggested that the clinical presentation of CMVR in HIV-negative patients can differ from that in HIV-positive patients [20–23, 27, 36, 94]. In a small series of eyes with CMVR in HIV-negative patients complied by Maguire and associates in 1995, three eyes were found to have spontaneously regressed or indolent appearing CMVR associated with vitritis and CME. The authors suggested that HIV-negative CMVR was more often associated with moderate to severe vitreous inflammation and CME [21]. The tendency for HIV-negative CMVR to have more severe intraocular inflammation as compared to CMVR in HIV-positive patients has since been noted by a number of authors, including Silverstein and colleagues [23], Voros and associates [27], Tajunisah and colleagues [25], Panthanapitoon and associates [22], and Schneider and colleagues [36] with specific mention of a similarity to acute retinal necrosis (ARN) in some eyes, including both the severity of the inflammation and the presence of occlusive vasculitis. In most instances, the more severe inflammation and vascular occlusion was ascribed to relative retention of anti-CMV immunoreactivity. However, other studies have failed to identify consistent difference in clinical presentation based on HIV status [18–20, 24, 26, 30–33, 37–40, 44, 94, 95]. In our review, of all the reported cases to date of HIV-negative CMVR not associated with Good syndrome or following intraocular or periocular administration of corticosteroids, clinical features of the inflammation were reported in 199 of the 248 eyes (80.2 %). Within these 199 eyes, 13 (6.5 %) were specifically described as having moderate to severe anterior chamber inflammation, 24 (12.1 %) as having moderate to severe vitreous inflammation, and 47 (23.6 %) were noted to have occlusive vasculitis. In HIV-negative CMVR associated with Good syndrome, clinical features of the inflammation were reported in all eyes. Within these 10 eyes, only 1 (10.0 %) was specifically described as having moderate to severe anterior chamber inflammation, whereas 5 out of 10 eyes (50.0 %) were described as having moderate to severe vitreous inflammation. None were noted to have occlusive vasculitis. In HIV-negative CMVR following intraocular or periocular administration of corticosteroids, clinical features of the inflammation were reported in 12 of the 21 eyes (57.1 %). Among these 12 eyes, 4 (33.3 %) each were described as moderate to severe anterior chamber inflammation or vitreous inflammation, and 6 (50.0 %) were noted to have occlusive retinal vasculitis. Hence, while an ARN-like picture including moderate to severe intraocular inflammation and the presence of occlusive vasculitis may be somewhat more common in HIV-negative as compared to HIV-positive CMVR, particularly in the setting of Good syndrome or following periocular or intraocular corticosteroids, the clinical presentation in these various cohorts appears, more often than not, to be fairly similar, with an overlapping clinical presentation (Table 5).

Immune recovery uveitis has been well characterized in HIV-positive patients and is currently one of the most common causes of vision loss in patients with CMVR receiving highly active antiretroviral therapy (HAART) [104, 105]. While the occurrence of IRU has varied widely in HIV-positive cohorts receiving HAART, with reported rates ranging from 3.0 to 63.3 % [106], the inflammation generally occurs several weeks to months after initiating HAART, as the number of circulating CD4+ T cells increases. The clinical spectrum of IRU includes vitritis, papillitis, CME, epiretinal membrane formation, vitreous hemorrhage, retinal neovascularization, vitreomacular traction syndrome, and proliferative vitreoretinopathy [104–107].

Of the 248 eyes of 178 patients with CMVR in the absence of either HIV infection or Good syndrome that have been described in the literature (Additional file 1: Table S1), 16 eyes (6.5 %) of 10 patients had one or more features consistent with IRU [17, 20, 45, 50, 70]. Reported ages of these 10 patients ranged from 15 to 68 years, with a mean and median of 47.8 and 54 years, respectively. Men outnumbered women approximately two to one (M to F ratio = 2.3:1), and the 100 % of the cases had an identifiable cause of systemic immunosuppression. The most common factors contributing to a relative decline in immune function included an underlying malignancy (n = 3; 30.0 %), age over 60 years (n = 2; 20.0 %), an autoimmune disorder (n = 2; 20.0 %), organ (n = 3; 30.0 %) or bone marrow (n = 1; 10.0 %) transplantation requiring systemic immunosuppression, and diabetes mellitus (n = 1; 10.0 %). The three reported cancers included two patients with acute lymphoblastic leukemia and one patient with chronic lymphocytic leukemia. The use of systemic immunosuppressive medication was reported in 8 out of the 10 cases (80.0 %). All of these patients were on two or more immunosuppressive agents. Corticosteroids were the most common immunosuppressive agent used in four patients (50.0 %), followed by cyclophosphamide in three patients (37.5 %), and two patients each (25.0 %) treated with vincristine and mycophenolate mofetil. One patient each (12.5 %) was treated with rituximab, fludarabine, methotrexate, and tacrolimus. Clinically, the retinitis was bilateral in 6 (60.0 %) out of 10 cases. Most case reports provided limited clinical information, and the location was either reported or illustrated in only 1 out of the 10 patients (10.0 %). In this patient, the retinitis was found to involve zone II. Additional clinical features were noted in 5 out of 16 eyes (31.3 %). Among these five eyes, just two eyes (40.0 %) were noted to have moderate anterior chamber inflammation. All five eyes (100.0 %) were noted to have vitreous inflammation, among which inflammation was described moderate to severe in all cases. An occlusive vasculitis was not present in any reported case of CMVR associated with IRU. Visual acuity at initial CMVR diagnosis was reported in 15 out of 16 eyes (93.8 %). Among these 15 eyes, visual acuity at initial diagnosis was better than 20/40 in 40.0 %, between 20/40 and 20/200 in 53.3 %, and worse than 20/200 in 6.7 % of eyes. The method of diagnostic confirmation of CMVR was reported in 1 of the 10 cases (10.0 %), among which the diagnosis of CMVR was confirmed serum PCR testing. The retinitis responded to antiviral therapy in all cases. The therapeutic treatment administered was reported in 6 of the 10 cases (60.0 %). Of these 6 cases, treatment was administered systemically only in 33.3 % of patients and intravitreally alone in 66.7 % patients. Visual acuity at last follow-up visit was reported in 8 patients and 13 eyes (mean 22.2 months; median 19 months; range 1 to 43 months). Visual acuity at last follow-up was better than 20/40 in 38.5 %, between 20/40 and 20/200 in 46.1 %, and worse than 20/200 in 15.4 % of eyes.

Since the December, 2014, cutoff for our literature review, there have been several publications of CMVR in HIV-negative patients [108–113]. While these publications are not incorporated into this review, the clinical context in which CMVR developed and the clinical characteristics of the retinitis and the treatment(s) given for the infection were not significantly different from those previously reported or from the conclusions drawn by our review [114].