Ocular Cytomegalovirus Infection: A Challenging Diagnosis

The two case reports on cytomegalovirus (CMV) optic neuritis (papillitis) in this month's Journal of Neuro-Ophthalmology juxtapose two methods of establishing the diagnosis of CMV ocular disease-one by relying on clinical history and examination (1) and the other by relying on laboratory data (2). The first case involved a patient with AIDS who presented with visual loss, optic disc swelling, a focal area of retinitis, and spot hemorrhages on the disc margin (1). Cerebrospinal fluid (CSF) CMV IgG was positive, but CMV polymerase chain reaction (PCR) from CSF or peripheral blood was not performed. No other etiology was discovered, and the patient responded well to ganciclovir treatment for presumed CMV retinitis and optic neuritis. In this case, the lack of compelling laboratory evidence for CMV infection was balanced by the patient's reduced immune status (CD4 count of <200 L), the characteristic retinal lesions, and a negative workup for other opportunistic infections. Treatment with ganciclovir was reasonable, given the risk of CMV retinitis in this patient population (3). A successful response to treatment further supported the diagnosis. The second case involved a young immunocompetent patient who presented with fever, headache, visual changes, and ophthalmoscopic evidence of bilateral optic disc swelling (2). Although ocular involvement with CMV is exceedingly rare in immunocompetent patients, the high serum CMV IgM, positive serum CMV PCR results, pancytopenia, splenomegaly, and absence of other infection made a compelling case for CMV papillitis and justified treatment. This patient also responded to intravenous ganciclovir with complete resolution of symptoms and examination findings. The definitive diagnostic test for CMV end-organ disease is a tissue sample demonstrating CMV inclusion bodies within parenchymal cells (4). Serologic testing can demonstrate previous (IgG) or recent (IgM) exposure to CMV but is not proof of current end-organ disease (5). Culture techniques are laborious; even early antigen detection methods (shell vial cultures) require 2-3 days to complete (6). Furthermore, culturing CMV from a body fluid does not necessarily indicate end-organ infection as this organism can sometimes persist after infection in immunocompromised patients (7). PCR has recently been used to detect CMV DNA in bodily fluids including blood, CSF, vitreous, bronchial washings, and urine (8). Serum PCR and quantitation of CMV have been helpful in diagnosing CMV disease and predicting risk of end-organ disease in transplant and HIV-infected patients (9,10). However, as with culture techniques, the presence of CMV does not always indicate disease, especially in immunocompromised patients who can shed virus without end-organ disease (11). Ocular CMV is particularly challenging to diagnose because pathologic samples are difficult to obtain. In these two patients with optic neuritis, biopsy was not an option (1,2). Vitreous sampling for culture or PCR is sometimes useful, but this procedure is not without risk (12). In HIV-positive patients, CMV retinitis is generally a diagnosis made on the basis of characteristic retinal changes in a patient with a CD4 count of <200 cells/L. This was true in the first case with the presence of characteristic hemorrhages and a low CD4 count in a patient who was not taking antiretroviral medications (1). In the pre-HAART era, 21%-44% of patients with AIDS developed CMV retinitis (3). Therefore, even with little laboratory evidence, treatment for CMV seemed necessary in this patient. The second case involved an immunocompetent patient with bilateral papillitis, a very unusual complication of CMV (2). In this patient, the laboratory evidence (CMV serology and PCR results) suggested the diagnosis of CMV papillitis and was adequate justification for treatment with ganciclovir. On the other hand, the nonspecific clinical presentation of a mononucleosis-like syndrome would not, by itself, have been compelling enough evidence to warrant therapy.