Cytomegalovirus retinitis after initiation of highly active antiretroviral therapy

Summary

Background

In previous natural history studies and clinical trials, AIDS-related cytomegalovirus (CMV) retinitis has occurred primarily in patients with absolute CD4 counts of 50 cells/μL or less (0·05×10⁹/L) at the time of diagnosis.

Methods

We report five patients identified from our clinical practices who were diagnosed with CMV retinitis while their CD4 counts were above 195 cells/μL. We also analysed, based on CD4 counts, 76 AIDS patients with newly diagnosed CMV retinitis whose CD4 lymphocyte enumerations were done in laboratories that maintained certification in a common external quality control programme.

Findings

5–24 weeks before retinitis was diagnosed, all five patients had had absolute CD4 lymphocyte counts of less than 85 cells/μL, and 4–7 weeks before diagnosis, all five patients had started taking highly active antiretroviral treatment (HAART) regimens. Only one (4%) of 27 patients enrolled in the trial between July, 1995, and February, 1996, had an absolute CD4 count of more than 50 cells/μL, and none of 27 had an absolute CD4 count of more than 100/μL on entry to the trial. However, from March, 1996 (when indinavir and ritonavir were approved by the FDA for marketing in the USA), to August, 1996, 14 (29%) of 49 patients had CD4 counts of more than 50/μL and seven (14%) of 49 had a CD4 count of more than 100 cells/μL on entry.

Interpretation

These findings suggest that the early immunological effects of HAART may not provide sufficient protection to prevent CMV retinitis in patients who have very low CD4 counts when therapy is started. Clinicians should note that CMV retinitis may now occur in patients who have CD4 counts of more than 100 cells/μL.

Introduction

AIDS-related cytomegalovirus (CMV) retinitis occurs primarily in patients with peripheral-blood absolute CD4 lymphocyte counts of 50 cells/μL or less (0·05×10⁹ L) and almost always in patients with CD4 counts of less than 200 cells/μL at the time of diagnosis.1, 2, 3 In a multicentre observational cohort study of 1002 patients with either AIDS or symptomatic HIV disease who had an absolute CD4 count of less than 250 cells/μL, the estimated risk of patients developing CMV end-organ disease within 3 months of having an absolute CD4 count of 100 cells/μL or more was less than 1%.¹ In a multicentre comparative trial of foscarnet versus ganciclovir as initial treatment for newly diagnosed CMV retinitis, 178 (86%) of 206 patients with CD4 entry data had absolute CD4 counts of less than 50 cells/μL. Only three (2%) patients had entry CD4 counts of 200 cells/μL or more (J Holbrook and colleagues, personal communication).² Also, in a multicentre comparative trial of oral versus intravenous ganciclovir for newly diagnosed CMV retinitis, none of the 91 eligible patients had an entry absolute CD4 count of 200 cells/μL or more (Charles Robinson, personal communication).³

Highly active antiretroviral treatment (HAART) regimens consisting of an HIV protease inhibitor combined with one or two dideoxynucleoside agents increase dramatically absolute CD4 counts, reduce HIV viral load, and improve survival, even in patients with very low absolute CD4 counts.4, 5 However, there is no evidence to show whether patients with very low absolute CD4 counts who start treatment with these drugs and who subsequently have a good immunological response (indicated by increased absolute CD4 counts to $200 cells/μL) will still be at risk for development of CMV end-organ disease.

We report five patients who had very low absolute CD4 counts and then developed CMV retinitis 4–7 weeks after initiation of this type of highly active treatment regimen, despite a concomitant large increase in absolute CD4 count. We also analysed CD4 counts in a group of 76 clinical-trial patients with newly diagnosed CMV retinitis.