Incidence and risk factors for developing cytomegalovirus retinitis in HIV-infected patients receiving protease inhibitor therapy

Casado, Jose L.; Arrizabalaga, Julioa; Montes, Milagrosa; Martí-Belda, Paloma; Tural, Cristinab; Pinilla, Javierc; Gutierrez, Carolina; Portu, Josebad; Schuurman, Robf; Koldo Aguirrebengoae for the Spanish CMV-AIDS Study Group

Clinical: Original Papers

Objective: To assess the incidence and risk factors for cytomegalovirus (CMV) retinitis in HIV-infected patients who initiated protease inhibitor-containing antiretroviral therapy.

Design and setting: Prospective, multicentre study.

Patients: A cohort of 172 HIV-infected patients with a CD4 cell count below 100×106cells/l at the time of protease inhibitor introduction.

Main outcome measures: Confirmed CMV retinitis and mortality, according to CD4 cell count, HIV load, and CMV viraemia.

Results: The cumulative incidence of CMV retinitis was 5% at 1 year and 6% at 2years. Only a positive CMV polymerase chain reaction (PCR) test at therapy initiation was significantly associated with the development of disease (relative hazard, 4.41; 95% confidence interval, 2.12--8.93; P<0.00001). The 12-month Kaplan-Meier CMV retinitis event rate was 38% in patients who were CMV PCR-positive compared with 2% in those who were CMV PCR-negative (P<0.001). Mean CMV load was significantly higher in those individuals who went on to develop CMV retinitis (3700 versus 384copies/ml, P=0.002). Only 2% of patients remained CMV PCR-positive after 3 months of protease inhibitor therapy, and CMV viraemia was not associated with a worse therapy response or shorter survival. Transient CMV positivity without a higher risk of disease was observed in 7% of patients at the first month on therapy.

Conclusions: Protease inhibitor-containing antiretroviral therapy significantly reduces the incidence of CMV viraemia and disease. Although a positive CMV PCR test identifies those patients on therapy at highest risk of CMV retinitis, it is not associated with an increased risk of death or a worse response to protease inhibitor therapy.

Author Information

From the Infectious Diseases Units, Hospital Ramon y Cajal, Madrid, aHospital Ntra. Sra de Aránzazu, San Sebastian, bHospital Germán Trias i Pujol, Badalona, cHospital San Millán, Logroño, dHospital Txarrigortu, Vitoria, eHospital Cruces, Bilbao, España and fDepartment of Virology, Utrecht University, The Netherlands. *See Appendix.

Sponsorship: This study was supported, in part, by Roche Laboratories providing us qualitative and quantitative CMV PCR determinations.

Note: Presented in part at the 6th Conference on Retroviruses and Opportunistic Infections, Chicago, 1999. [Abstract 251].

Correspondence to Dr Jose L. Casado, Infectious Diseases Unit, Ramon y Cajal Hospital, Cra. Colmenar, km 9.1. 28034, Madrid, Spain. Date of receipt: 7 January 1999; revised: 17 May 1999; accepted: 27 May 1999.

Article Outline
Back to Top | Article Outline


Cytomegalovirus (CMV) disease, especially retinitis, is a common opportunistic complication of advanced HIV infection. The incidence increases markedly when the CD4 cell count falls below 100×106cells/l [1-3], and before the use of protease inhibitors, at least 20% of patients with counts below that level will develop CMV-related disease over a 2year period [4]. Data from several studies show that methods for the detection of CMV viraemia, such as polymerase chain reaction (PCR) or the detection of pp65 antigenaemia, could identify those patients at highest risk of developing CMV disease and also those patients at risk of decreased survival [5-8].

However, the immunological and clinical benefit observed with the widespread use of protease inhibitor-containing therapy has changed the natural history of CMV disease [9]. Preliminary data suggest an important reduction in the incidence of opportunistic infections, including CMV retinitis, for patients treated with highly active antiretroviral therapy (HAART) [10]. Moreover, observational studies show cases of CMV disease in patients with more than 100×106cells/l, a rare presentation before protease inhibitors were introduced into the clinical management of HIV patients [11]. However, little information exists on the incidence, risk factors, and clinical presentation of CMV disease for patients receiving HAART. In addition, no data are available on the evolution of CMV markers during protease inhibitor therapy, or on the association of CMV viraemia and outcome in a large cohort of HIV-infected patients in the era of HAART.

Therefore, the aim of this study was to estimate the incidence of CMV retinitis and to determine risk factors for the development of the disease in a cohort of HIV-infected patients with a CD4 cell count below 100×106cells/l, and who initiated protease inhibitor-containing therapy.

Back to Top | Article Outline

Patients and methods

Patients were enrolled in the cohort at 10 AIDS care units participating in a multicentre group for the study of CMV disease in AIDS. Patients with HIV infection were enrolled if they had a CD4 cell count <100×106cells/l, positive CMV immunoglobulin G serology, and were to initiate antiretroviral therapy containing a protease inhibitor. Patients were evaluated in this study for whom baseline and follow-up plasma samples were available. The first patient was included in February 1996 and the last patient in April 1997. Those patients with past or present CMV disease, or using anti-CMV therapy, were excluded.

A complete medical history and physical examination, CD4 cell count, and HIV RNA estimate were performed at baseline, 1 month, 3 months, and every 3 months thereafter. Plasma samples for CMV detection was collected at the same time points during the first 6 months of therapy at each of the participating study centres. CD4 cell count was assessed by flow cytometry, and HIV RNA was determined using PCR (Amplicor HIV Monitor, Roche Molecular Systems, Branchburg, New Jersey, USA). CMV viraemia was detected by a commercially available qualitative PCR technique (Amplicor CMV, Roche Diagnóstica, Madrid, Spain). CMV DNA quantification was also performed on the positive samples using a quantitative CMV microwell plate assay, developed by Roche Molecular Systems [12].

In each follow-up visit, patients were evaluated for signs or symptoms of CMV retinitis. Those with visual signs or symptoms were referred to an ophthalmologist experienced in the diagnosis of CMV retinitis. Retinal disease due to CMV was diagnosed in the presence of white and granular lesions associated with haemorrhage that usually follows the vascular distribution producing large areas of necrosis. Funduscopic retinal photographs were not required.

Screening ophthalmological examinations were not required at baseline or during follow-up for asymptomatic patients.

Back to Top | Article Outline
Statistical analysis

Differences in demographic details according to initial PCR status were tested using the Mann-Whitney U test for quantitative variables and the χ2 test for qualitative variables. Time to development of disease was estimated using the Kaplan-Meier method, and the relationships between PCR status and time to CMV retinitis were compared by log-rank test, with patient follow-up censored at the time of death or on the 31 May 1998 if the patient had not developed CMV disease. A Cox proportional hazards model was developed to determine the influence of the baseline covariates on the development of CMV retinitis.

Back to Top | Article Outline


During the study period, 418 HIV-infected patients with CD4 cell counts below 100×106cells/l initiated protease inhibitor therapy. Baseline plasmas from 172 patients with a positive IgG serology against CMV were available, and were evaluated for CMV DNA load by qualitative PCR. Overall, 11% (18 out of 172) of the patients were CMV PCR-positive at study entry. Baseline characteristics of the patients according to initial PCR status were similar. Most patients had been heavily pre-treated with nucleoside analogues (86%), and had experienced a prior AIDS-defining condition (76%). Median CD4 cell count at entry was 28×106cells/l, and median HIV RNA was 5.12log10copies/ml.

On initiation of protease inhibitor therapy, a rapid improvement of CD4 cell count and HIV load was observed. By the third month, median CD4 cell count was 97×106cells/l (range, 6-640) and HIV RNA was 3.62log10copies/ml (range, 2.3-6.4). At the same time, and excluding those patients who developed CMV disease, the number of patients who were initially CMV PCR-positive decreased from 11% at study entry to 7% (11 out of 151) at the end of the first month of protease inhibitor therapy, to 2% (two out of 99) at the end of the third month, and to 0% (of 69 patients) at the end of the sixth month of therapy. Seven per cent of patients (nine out of 137) who were CMV PCR-negative at baseline had PCR-positive results at the end of the first month of protease inhibitor therapy. However, by the third month of treatment their CMV PCR results were once again negative.

In a median follow-up of 24 months (range, 13-26), the cumulative incidence of CMV retinitis was 5% at 1 year and 6% at 2 years. Most cases of disease were observed for patients who had a baseline CD4 cell count below 50×106cells/l, and occurred during the first 90 days after the initiation of protease inhibitor-containing therapy (six out of nine patients, 67%). A detailed analysis of the nine cases of CMV retinitis showed that the disease developed in three patients with CD4 cell counts above 50×106cells/l, and in one patient with a count above 200×106cells/l (Table 1).

A positive CMV PCR result at baseline was significantly associated with the development of CMV retinitis (P<0.001, log-rank test; Fig. 1). Thus, the 12 months Kaplan-Meier CMV retinitis event rate was 38% in patients who were CMV PCR-positive, compared with 2% for those who were CMV PCR-negative. In a Cox proportional hazards model, a positive CMV PCR result at study entry was also significantly associated with the development of CMV retinitis (relative hazard, 4.41; 95% confidence interval 2.12-8.93; P<0.001). There was no significant association with CMV disease risk for CD4 lymphocyte count, HIV load, protease inhibitor used, previous AIDS diagnosis, or prior antiretroviral treatment (Table 2). The association between a positive baseline CMV PCR result and CMV retinitis was not observed for those patients with transient positive CMV PCR results at the first month of protease inhibitor therapy (P=0.23; log-rank test), and no patient developed CMV retinitis. Quantitative analysis of patients who were initially CMV PCR-positive showed that mean CMV load at entry was significantly higher in patients who went on to develop CMV retinitis than those who did not (3700 versus 384copies/ml; P=0.002). Moreover, 75% of patients with a CMV load higher than 1000copies/ml developed the disease. When combined with a CD4 cell count below 50×106cells/l, this cut-off value identified 100% of the patients who went on to develop CMV disease. Characteristically, mean CMV load in patients with transient CMV PCR positivity was low (559 copies/ml, range 275-1256).

Despite the strong association between CMV PCR status and development of CMV retinitis, clinical, virological and immunological outcome was similar for patients with or without initial CMV viraemia. After 15 months on protease inhibitor therapy, median CD4 cell count (240 versus 245×106cells/l; P=0.91) and HIV RNA (3.01 versus 3.32log10copies/ml; P=0.46) were similar and independent of initial CMV PCR status. During follow-up, 13 patients died, three of them shortly after the diagnosis of CMV retinitis. Causes of death were non-Hodgkin lymphoma in three cases, MAC disease in two, chronic hepatitis in four, and gram-negative sepsis, interstitial pneumonia, disseminated tuberculosis and central nervous system lymphoma in one each. No death was attributed to CMV disease. There was no statistically significant difference in mortality rate according to initial CMV viraemia (P=0.87; log-rank test).

Back to Top | Article Outline


This study shows an important reduction in the incidence of CMV retinitis for patients who initiated protease inhibitor therapy. For patients with CD4 cell counts below 100×106cells/l at study entry, the cumulative incidence after 2 years was 6%. This is significantly lower than the 20-28% at 1 year reported in recent studies for a similar population of patients not receiving protease inhibitor therapy [13,14]. Furthermore, the incidence in this study is lower than in a study of CMV prophylaxis in patients receiving oral ganciclovir that reported a CMV event rate of 14% at 1 year [15]. This impressive benefit appears to be attributable to the improvement of the patient‚s immunological status while on protease inhibitor therapy.

Most cases of CMV retinitis in our study were observed during the first three months of therapy. During this period, the increase in CD4 cell count is considered to be insufficient to provide adequate protection [16,17]. Indeed, recent data demonstrate that disruptions in the T-cell receptor repertoire observed in advanced HIV disease may not be corrected in the first weeks following initiation of protease inhibitor therapy [16]. Evidence from previous studies also suggests that patients in our study may have already seeded their retinas with CMV, or may have had an asymptomatic peripheral retinitis, before initiating protease inhibitor therapy. However, this fact has been found in a minority of patients [18]. Furthermore, when patients with more than three months of protease inhibitor therapy are considered separately, the cumulative incidence of CMV retinitis in our study was 2% at 2 years, a figure similar to that found in CMV PCR-negative patients receiving prophylactic oral ganciclovir [15].

The results of this study are strikingly similar to recent findings for the incidence of other opportunistic infections such as Mycobacterium avium complex infection or cryptosporidiosis, in severely immunosuppressed patients taking protease inhibitors [19]. However, we found a lower incidence of patients with CMV viraemia at baseline in our study compared with similar cohorts of patients in other studies (11 versus 28-45%) [7,13]. This lower number of CMV PCR-positive results can be explained by differences in the sensitivity of PCR methods. In a comparison of several CMV markers, Boivin et al. described a sensitivity of 65% for plasma-based PCR (Amplicor CMV; Roche Molecular Systems, Branchburg, New Jersey, USA) and 100% for leukocyte-based PCR assays [20]. However, the positive predictive value of the plasma-based PCR was higher than the leukocyte-based assay (64 versus 40%). Therefore, despite having fewer CMV PCR-positive patients in our study, we believe that the low incidence of CMV retinitis could be attributed to the initiation of protease inhibitor therapy.

As with previous studies, our data shows a 4.4-fold increased risk of developing CMV disease for patients who had CMV viraemia [5-8,14]. A positive CMV PCR result was also a better predictor for the risk of developing CMV disease than CD4 lymphocyte counts. In addition, quantitative PCR identified those patients at highest risk for CMV retinitis, a group in which it is worth performing controlled trials of pre-emptive therapy. This study provides important data that may alter the current prevention and management of CMV retinitis. Firstly, our study confirms the progressive decrease of CMV load after the initiation of antiretroviral therapy containing protease inhibitors, reflected by the increase in negative qualitative CMV PCR results in patients with low CMV load [21]. The results also suggest that an increase in CD4 cell count may be sufficient to avoid the development of CMV disease, allowing prophylaxis or pre-emptive therapy to be reserved for those high-risk patients in the short time period after the initiation of protease inhibitor therapy. Secondly, our study suggests that a transient positivity of CMV PCR, with low CMV load, in the first few weeks after the initiation of therapy is not necessarily indicative of increased risk of disease. The cause of this transient positivity is not known, and one hypothesis could be that it reflects the ongoing immune restoration. Finally, at variance with previous reports, patients in our study who were initially CMV PCR-positive and who did not develop CMV disease, had a similar clinical, immunological, and virological response while on therapy to those who were CMV PCR-negative. These data indicate that the risk of death directly related to CMV viraemia in persons with advanced AIDS declines after the initiation of protease inhibitor therapy, and change the previously-reported role of CMV viraemia as an independent predictor of poor prognosis [5,7]. It is not known if the survival benefit seen in our study is directly related to immune restoration or if it is secondary to the reduction of CMV viraemia.

In summary, our study indicates that protease inhibitor therapy is effective in reducing the incidence of CMV viraemia and disease in HIV-infected patients with a CD4 cell count below 100×106cells/l at the time of therapy initiation. The results also offer a new approach to the management of CMV disease in HIV-positive patients, where prevention strategies could be implemented during a short period after the initiation of protease inhibitor therapy. Furthermore, the significant improvement in survival for patients who were initially viraemic supports the fundamental role of protease inhibitors in this immunocompromised population. Finally, controlled trials of CMV prophylaxis or pre-emptive therapy should be conducted in patients who have a high CMV load and who are therefore at high risk of CMV retinitis. These studies should also take into account the contribution of protease inhibitor therapy to the reduction of CMV viraemia.

Back to Top | Article Outline


We are indebted to M. Garrido and I. Diaz, from Roche Pharmaceuticals, for their assistance and support. Also, thanks to R. Manjiry and C. Boucher for the determination of CMV load.

Back to Top | Article Outline


1. Jacobson MA, Mills J. Serious cytomegalovirus disease in the acquired immunodeficiency syndrome (AIDS). Clinical findings, diagnosis, and treatment. Ann Intern Med 1988, 108: 585-594.
2. Gallant JE, Moore RD, Richman DD, Keruly J, Chaisson RE, Zidovudine Epidemiology Study Group. Incidence and natural history of cytomegalovirus disease in patients with advanced human immunodeficiency virus disease treated with zidovudine. J Infect Dis 1992, 166: 1223-1227.
3. Chan IS, Neaton JD, Saravolatz LD, Crane LR, Osterberger J. Frequencies of opportunistic diseases prior to death. AIDS 1995, 9: 1145-1151.
4. Pertel P, Hirschtick R, Phari J, Chmiel J, Poggense L, Murphy R. Risk of developing cytomegalovirus retinitis in persons infected with the human immunodeficiency virus. J Acquir Immune Defic Syndr 1992, 5: 1069-1074.
5. Bowen EF, Wilson P, Cope A, et al. Cytomegalovirus retinitis in AIDS patients. Influence of cytomegaloviral load on response to ganciclovir, time to recurrence and survival. AIDS 1996, 10: 1515-1520.
6. Dodt KK, Jacobsen PH, Hofmann B, et al. Development of cytomegalovirus (CMV) disease may be predicted in HIV-infected patients by CMV polymerase chain reaction and the antigenemia test. AIDS 1997, 11: F21-F28.
7. Spector SA, Wong R, Hsia K, Pilcher M, Stempien MJ. Plasma cytomegalovirus (CMV) DNA load predicts CMV disease and survival in AIDS patients. J Clin Invest 1998, 101: 497-502.
8. Brosgart CL, Louis TA, Hillman DW, et al. A randomized, placebo-controlled trial of the safety and efficacy of oral ganciclovir for prophylaxis of cytomegalovirus disease in HIV-infected individuals. AIDS 1998, 12: 269-277.
9. Casado JL, Perez-Elías MJ, Martí-Belda P, et al. Improved outcome of cytomegalovirus retinitis in AIDS patients after introduction of protease inhibitors. J Acquir Immune Defic Syndr Hum Retrovirol 1998, 19: 130-134.
10. Palella FJ Jr, Delaney KM, Moorman AC, et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators. N Engl J Med 1998, 338: 906-908.
11. Mallolas J, Arrizabalaga J, Loncá M, et al. Cytomegalovirus disease in HIV-1 infected patients treated with protease inhibitors [letter]. AIDS 1997, 11: 1785-1787.
12. Schuurman R, Blank B, Pouw W, et al. CMV viral load determination by quantitative DNA PCR in relation to other virological parameters for the diagnosis of CMV infection in HIV-infected individuals. Fifth Conference on Retroviruses and Opportunistic Infections. Chicago, February 1998 [Abstract 764].
13. Bowen EF, Sabin CA, Wilson P, et al. Cytomegalovirus (CMV) viraemia detected by polymerase chain reaction identifies a group of HIV-positive patients at high risk of CMV disease. AIDS 1997, 11: 889-893.
14. Shinkai M, Bozzette SA, Powderly W, Frame P, Spector SA. Utility of urine and leukocyte culture and plasma DNA polymerase chain reaction for identification of AIDS patients at risk for human cytomegalovirus disease. J Infect Dis 1997, 175: 302-308.
15. Spector SA, McKinley GF, Lalezari JP, et al. Oral ganciclovir for the prevention of cytomegalovirus disease in persons with AIDS. N Engl J Med 1996, 334: 1491-1497.
16. Connors M, Kovacs JA, Krevat S, et al. HIV infection induces changes in CD4 T-cell phenotype and depletion within the CD4 T-cell repertoire that are not immediately restored by antiviral or immune-based therapies. Nature Med 1997, 3: 533-540.
17. Kelleher AD, Carr A, Zaunders J, Cooper DA. Alterations in the immune response of human immunodeficiency virus (HIV)-infected subjects treated with an HIV-specific protease inhibitor, ritonavir. J Infect Dis 1996, 173. 321-329.
18. Brosgart CL, Pulling C, Chaloner K, et al. Prevalence of asymptomatic CMV retinitis in AIDS patients. Fifth Conference on Retroviruses and Opportunistic Infections. Chicago, February 1998 [Abstract 756].
19. Michaels S, Clark R, Kissinger P: Differences in the incidence rates of opportunistic processes before and after the availability of protease inhibitors. Fifth Conference on Retroviruses and Opportunistic Infections. Chicago, July 1998 [Abstract 180].
20. Boivin G, Handfield J, Toma E, Lalonde R, Murray G, Bergeron MG: Evaluation of the Amplicor CMV test for the diagnosis of CMV disease in HIV patients. 37th Interscience Conference on Antimicrobial Agents and Chemotherapy. Toronto, 27 September-1 October 1997 [Abstract H104].
21. Gerna G, Baldanti F, D‚Arminio Monforte A, et al. Sustained disappearance of human CMV in blood of AIDS patients following HAART. 12th World AIDS Conference, Geneva, July 1998 [Abstract 22237].
Back to Top | Article Outline
Grupo de Estudio de CMV y SIDA

In addition to the authors, the following investigators are members of the CMV-AIDS study group. Hospital Aránzazu (San Sebastian): J.A. Iribarren, F. Rodriguez-Arrondo, M.A. von Wichmann, M. Dolores de Juan. Hospital de Cruces (Bilbao): M. Montejo. Hospital Germán Trías i Pujol (Badalona): B. Clotet. Hospital Miguel Servet (Zaragoza): P. Arazo. Hospital Provincial La Rioja (Logroño): J.A. Oteo. Hospital Ramón y Cajal (Madrid): M.J. Perez-Elías, A. Antela, F. Dronda. Hospital Reina Sofía (Tudela, Navarra): M.T. Rubio. Hospital San Millan (Logroño): P. Labarga. Hospital de Teruel (Teruel): M. Yuyol. Hospital Txagorritxu (Vitoria): J.M. Agud.


Cytomegalovirus; retinitis; opportunistic infections; protease inhibitor; HIV

© 1999 Lippincott Williams & Wilkins, Inc.