CMV retinitis recurs after stopping treatment in virological and immunological failures of potent antiretroviral therapy : AIDS

Secondary Logo

Journal Logo


CMV retinitis recurs after stopping treatment in virological and immunological failures of potent antiretroviral therapy

Torriani, Francesca J.a; Freeman, William R.b; Macdonald, J. Christopherb; Karavellas, Marietta P.b; Durand, Deborah M.c; Jeffrey, Douglas D.c; Meylan, Pascal R. A.d; Schrier, Rachel D.c

Author Information
  • Free



The administration of highly active antiretroviral therapy (HAART) is associated with dramatic decreases in plasma HIV RNA levels and rises in CD4 cell counts [1,2]. Although CD4 cell count increases resulting from HAART are less in patients with advanced HIV disease, earlier predictions that reconstitution of antigen-specific immune responses would not occur in this population have proved untrue [3]. In 1997, Autran et al. [4] demonstrated that HAART was associated with polyclonal expansion of the memory T cell repertoire and improved in vitro T cell responses to opportunistic pathogens even in patients with advanced HIV disease. Epidemiological and clinical reports confirm that functional reconstitution of immunity occurs. Patients experiencing sustained CD4 cell count increases above 100 cells/mm3 have a greatly reduced risk of developing cytomegalovirus (CMV) or other opportunistic infections [5–8]. Finally, Furrer et al. [9] have demonstrated that primary prophylaxis for Pneumocystis carinii can be safely stopped in patients whose CD4 cell counts have risen above 200 cells/mm3 in response to HAART.

Before the availability of HAART, patients with CMV end-organ disease were committed to life-long therapy [10,11]. In the era of HAART, CMV maintenance therapy has been successfully discontinued in several pilot studies for up to 15 months [12–16]. However, the definition of response to antiretroviral agents differed significantly among those studies, and ranged from the complete suppression of viral replication and CD4 cell count increases above 150 cells/mm3 to an isolated increase in CD4 cell counts above 60 cells/mm3.

Previously, we described a cohort of subjects treated with HAART who discontinued CMV maintenance and did not exhibit recurrent disease up to 1 year [14]. Among that cohort who we have continued to monitor, five participants have experienced reactivation of CMV retinitis. To evaluate the virological and immunological parameters of HIV disease and measures of T cell immune responses to CMV in predicting CMV relapse, we compared the clinical and laboratory characteristics of patients who reactivated CMV retinitis to those who did not.

Materials and methods

The institutional review board approved the study. Written informed consent was obtained from all patients before enrolment.

Study population

Between October 1995 and September 1998, 20 AIDS patients with two consecutive CD4 cell counts above 70 cells/mm3 in response to HAART, and healed and treated CMV retinitis, discontinued their anti-CMV therapy at the AIDS Ocular Research Unit. HAART was defined as the use of an HIV protease inhibitor in combination with at least one nucleoside or non-nucleoside reverse transcriptase inhibitor.

Clinical evaluations

Participants were followed for recurrence or progression of CMV retinitis or contralateral disease with monthly dilated bilateral funduscopic examinations and wide-angle fundus photographs as described previously [17]. If CD4 cell counts dropped below 50 cells/mm3, the frequency of ophthalmological examinations was increased to twice a month and the re-initiation of systemic anti-CMV therapy was recommended. CMV recurrence was defined as an increase of activity of the border of two or more steps on the scale of border activity of the fundus photographic reading center.

HIV RNA quantitation by the Amplicor HIV-1 Monitor Test (Roche Diagnostic Systems, Branchburg, NJ, USA) and CD4 cell counts were performed as part of clinical care at 2–3 month intervals. Plasma HIV RNA values below the level of detection were assigned a value of 399 or 2.6 log10 copies/ml. Other clinical and laboratory data were extracted from the patients' clinic charts and entered in a database until 31 March 1999.

The leukocyte CMV DNA content was measured in 10 patients using a recently described quantitative competitive polymerase chain reaction (PCR) on samples drawn for the first and second proliferative assay [18].

In vitro lymphoproliferative assays to CMV, phytohemeagglutinin (PHA) and other antigens (Mycobacterium aviumcomplex, Candida, Toxoplasma, herpes simplex virus, mumps) were performed at 3 month intervals or on the day of scheduled ophthalmological visits using standard methods and following the guidelines established by the AIDS Clinical Trials Group. Positive responses were defined as a stimulation index above or equal to 3 [19].

Statistical analysis

Mean values of continuous variables, including lymphoproliferative assays were compared with the use of Students' two samples or paired t-tests as appropriate. A P value of 0.05 or less in a two-sided test was considered to indicate statistical significance. CD4 cell counts and plasma HIV RNA provided by the primary physicians were included if taken within 2 months of the ophthalmology clinic visit. Plasma HIV RNA levels were log10 transformed before analysis.

Because multiple measurements of the lymphoproliferative assays were performed on the 17 participants, the arithmetic mean from each patient was used for continuous analyses. The Yates-corrected chi-square test for 2 × 2 contingency tables was used for categorical analysis of lymphoproliferative assays using stimulation index cutpoints of 3 or 10. All available lymphoproliferative assays were used for this analysis without accounting for repeated measures.

Because it is more robust in the setting of small datasets, discriminant analysis was performed to determine which variables could identify patients at high risk of reactivation of CMV retinitis [20]. The variables utilized were CD4 cell counts and HIV RNA levels at maintenance discontinuation, maximal CD4 cell count and nadir HIV RNA obtained after HAART (continuous variables) and median or mean lymphoproliferative response for each patient above a stimulation index (SI) of 3 (categorical variables) (SPSS for Windows 6.1.33, SPSS Inc., Chicago, IL, USA).


Study participants

Of 20 individuals who stopped CMV maintenance between 11 October 1995 and 25 September 1998, 17 consented to participate in the prospective study of withdrawal of CMV maintenance and are therefore included in this analysis. Fifteen participants (88%) were men, 11 (65%) were Caucasian and six (35%) were Hispanic. Upon entering the study, median age was 40 (range: 27–63) years and median time with AIDS by the 1987 definition was 2.0 (range: 0.2–7.9) years.

Clinical course

CMV retinitis was first diagnosed a mean of 14.0 (range: 0.2–27.6; median: 12.0) months before the discontinuation of CMV maintenance. HAART was started a mean of 5.2 (range: −3.4–16.5; median: 3.3) months after the first diagnosis of CMV retinitis. CMV maintenance was discontinued a mean of 8.8 (range: −6.0–23.3; median: 9.5) months after the initiation of HAART. The first in vitro lymphoproliferative assay was performed a mean of 17.9 (range: 8.5–34.2; median: 17.1) months after starting HAART and 9.3 (range: −0.9–23.2; median: 8.2) months after stopping CMV treatment. Assays were repeated at 3 month intervals in all participants, achieving a median of four assays per patient (range: 2–6). Patients were followed off CMV maintenance for a mean of 25.7 (range: 10.0–36.1; median 26.6) months.

Five out of 17 (29%) participants experienced a recurrence of CMV retinitis over a median of 26 months off maintenance therapy. Intervals from initial CMV diagnosis to the initiation of HAART, time on CMV therapy, and HAART before stopping CMV treatment did not differ significantly between patients who reactivated CMV disease and those who did not. These intervals were, respectively [patients who reactivated versus those who did not, mean ± SE (median)] 7.5 ± 2.2 (7.5) versus 4.3 ± 1.8 (2.6) months from CMV diagnosis to HAART (P  = 0.29), and 15.4 ± 3.6 (19.6) versus 13.4 ± 2.1 (11.6) months on anti-CMV treatment (P  = 0.65) and 7.9 ± 3.9 (9.2) versus 9.1 ± 1.1 (9.7) months from HAART to the discontinuation of CMV maintenance (P  = 0.77).

Clinical characteristics

The clinical characteristics of the five patients who reactivated CMV retinitis are shown in Table 1. CMV retinitis reactivated a median of 14.5 months after discontinuing CMV maintenance. Similar to patients diagnosed with an initial episode of CMV retinitis, retinitis recurred from 8 days to 10 months after CD4 cell counts dropped below 50 cells/mm3. Median (range) CD4 cell counts at the time of reactivation were 37 (23–154) cells/mm3. Median (range) plasma HIV-1 RNA were 5.3 (4.4–5.9) log10 copies/ml. Patient no. 4 had CD4 cell counts below 50 cells/mm3 for 10 months. One month before recurring, this individual was started on HAART, which resulted in a transient rise in CD4 cell counts to 150 cells/mm3 at the time of reactivation. Unlike the other four participants who recurred in the eyes with previous disease, patient no. 4 presented with incident CMV retinitis and immune recovery vitritis in the previously unaffected contralateral eye. Both CMV retinitis and vitritis responded to ganciclovir administered through an intraocular device. Of note is the fact that in this patient CMV DNA was detected in the aqueous humor retrieved upon implantation of the ganciclovir device [21]. Patient no. 2 was diagnosed with a recurrence of CMV retinitis and pancreatitis that was attributed to CMV, and both of which resolved with anti-CMV treatment.

Table 1:
Clinical characteristics of five patients with reactivation of cytomegalovirus retinitis.

Response to anti-cytomegalovirus treatment

Three patients healed with a single intraocular cidofovir injection, one was treated with one dose of intravenous cidofovir, just before receiving an intraocular ganciclovir device; the last participant was treated with a ganciclovir implant. All five individuals received systemic maintenance with oral ganciclovir given in a dose of 1 g three times a day.

Virological and immunological studies

The virological and immunological characteristics of HIV infection since the diagnosis of CMV retinitis were compared between patients who reactivated CMV and those who did not, and are shown in Fig. 1 and Fig. 2.

Fig. 1.:
Evolution of CD4 cell counts (cells/mm3) at clinically relevant events in five patients with (REACT) and 12 patients without (NO REACT) reactivation of cytomegalovirus (CMV) retinitis. CMV, at first CMV retinitis diagnosis; LAST, at last follow-up; MAX, maximal increase achieved on HAART; PostHAART, 8 months after starting a HAART regimen; REACT, at reactivation of CMV retinitis; ST MT, at CMV maintenance discontinuation. Values are represented as box plot charts. The top, bottom and middle lines represent the 75th, 25th percentile and median. The whiskers on the top and bottom represent the 10th and 90th percentile. Squares represent arithmetic means.
Fig. 2.:
Evolution of plasma HIV RNA (log10 copies/ml) at specific clinical events in five patients with (REACT) and 12 patients without (NO REACT) reactivation of cytomegalovirus (CMV) retinitis. LAST, at last follow-up; NADIR, lowest value achieved on HAART; REACT, at reactivation of CMV retinitis; ST MT, at CMV maintenance discontinuation. Values are represented as box plot charts. The top, bottom and middle lines represent the 75th, 25th percentile and median. The whiskers on the top and bottom represent the 10th and 90th percentile. Squares represent arithmetic means.

The level of immunosuppression (CD4 cell counts) at the time of first diagnosis of CMV retinitis [patients who reactivated versus those who did not, mean ± SE (median): 22 ± 8 (22) versus 35 ± 6 (28) cells/mm3;P  = 0.24] and 8 months after starting HAART tended to be higher in the patients who reactivated but the difference did not make statistical significance [138 ± 33 (156) versus 227 ± 38 (191) cells/mm3;P  = 0.10]. However, in the patients with CMV reactivation, maximal CD4 cell counts after starting HAART were less dramatically increased [224 ± 41 (176) versus 425 ± 69 (382) cells/mm3;P  = 0.02] and were not sustained [CD4 cell counts at last follow-up: 48 ± 16 (50) versus 335 ± 44 (320) cells/mm3;P  < 0.0001]. Because CD4 cell counts continued to increase in the patients who did not recur, the maximal increase in CD4 cell counts was achieved later [mean ± SE (median): 12.4 ± 2.2 (13) versus 24.9 ± 1.8 (28) months;P  = 0.001]. CMV maintenance was discontinued at lower CD4 cell counts in patients who later reactivated CMV disease, but this difference did not reach statistical significance [124 ± 27 (161) versus 200 ± 30 (185) cells/mm3;P  = 0.08].

Nadir plasma HIV RNA obtained with HAART did not differ between groups [3.64 ± 0.4 (3.51) versus 2.85 ± 0.2 (2.60) copies/ml;P  = 0.16]. However, measures of sustained viral suppression were greater in patients without reactivation compared with those who reactivated. Mean (median) HIV RNA at the time CMV maintenance was stopped was 5.3 ± 0.1 (5.32) and 3.5 ± 0.3 (2.88) log10 copies/ml, respectively (P  = 0.0002). At last follow-up the HIV RNA level was also 5.24 ± 0.4 (5.82) in patients who reactivated versus 3.4 ± 0.3 (2.67) in those who did not (P  = 0.003). Plasma HIV RNA was below the limits of detection at maintenance discontinuation only in one out of five (20%) participants with recurrence compared with five out of 12 (42%) without, and at last follow-up in none compared with six out of 12 (50%) without reactivation of CMV retinitis.

Lymphoproliferative responses

In vitro lymphoproliferative responses to CMV antigens are shown in Fig. 3. Higher and nearly non-overlapping responses [mean ± SE (median)] were observed in the participants without recurrence of CMV: 22.1 ± 6.3 (15.9) SI versus 2.4 ± 0.7 (2.2) SI;P  = 0.01. A single patient in the group without reactivation had consistently low responses to CMV (mean SI: 2.6) in the face of sustained elevation in CD4 cell counts (mean CD4 cell counts 198 cells/mm3) and low level HIV RNA (mean HIV RNA 2.62 log10 copies/ml).

Fig. 3.:
In vitro lymphoproliferative responses to cytomegalovirus (CMV) (stimulation index; SI) in five patients with (REACT) and 12 patients without (NO REACT) reactivation of CMV retinitis. The mean lymphoproliferative response for each patient was used for this graph. Values are represented as box plot charts. The top, bottom and middle lines represent the 75th, 25th percentile and median. The whiskers on the top and bottom represent the 10th and 90th percentile. Squares represent arithmetic means. Mean values for each patient are plotted as full circles on the right side of the boxplots.

Of a total of 69 lymphoproliferative assays performed during the mean 26 months of observation off CMV maintenance (mean of 17 months since the first assay), 50 were run in the 12 patients who did not reactivate CMV retinitis and 19 in the five who experienced reactivation. Forty-two out of 50 (84%) assays in the non-reactivating group were above a SI of 3 compared with six out of 19 (32%) in the reactivating group (P  < 0.001). In the group without reactivation, 29 out of 50 (58%) lymphoproliferative assays were above 10 SI compared with none out of 19 (0%) in the group who reactivated (P  < 0.001).

The frequency of positive T cell proliferative responses to other antigens (M. avium complex, Candida, Toxoplasma), recall antigens (herpes simplex virus, mumps) was inhomogeneously distributed within and among patients (data not shown). Responses to PHA were positive in 82% of patients who did not reactivate versus 68% of those who did.

Predictors of recurrence in a model of discriminant analysis

Results from four models are shown in Table 2. Neither CD4 cell counts at discontinuation of CMV maintenance (percentage of grouped cases correctly classified as non-reactivators or reactivators 58.8%;P  = 0.15) nor the maximal CD4 cell count (70.6%;P  = 0.09) obtained after the initiation of HAART or both variables (64.7%;P  = 0.25) predicted reactivation. HIV RNA load at maintenance discontinuation (model 1) classified 76.5% of group cases correctly (P  < 0.005) and resulted in four false positives. In contrast, nadir HIV RNA (model 2) missclassified two patients who reactivated (false negatives) and two who did not reactivate (false positives) (76.5%;P  = 0.06). Model 1 would thus have prompted continued CMV maintenance therapy in four patients who did not reactivate, whereas model 2 would have left two patients unprotected.

Table 2:
Predictors of recurrence in a model of discriminant analysis.

When four variables (HIV RNA and CD4 cell counts at maintenance withdrawal, nadir HIV RNA and maximal CD4 cell count) were entered in model 3, the proportion of correctly classified cases remained unchanged (76.5%;P  = 0.02; four false positives). Therefore, the addition of nadir HIV RNA and CD4 cell counts failed to improve the power to discriminate.

Finally, when lymphoproliferative responses (mean/median lymphoproliferative responses > 3) were added to model 3, the percentage of correctly classified cases rose to 88.2% (P  = 0.04, one false positive, one false negative), compared with 72% when lymphoproliferative responses alone were used. Interestingly, the false-negative patient reactivated CMV retinitis at a CD4 cell count of 154 cells/mm3 (mean CD4 cell counts: 85 cells/mm3; HIV RNA 4.66 log10 copies/ml in the year after the first lymphoproliferative response) and had a relatively high mean lymphoproliferative response of 5.05 SI throughout the study. The false-positive patient had controlled HIV infection (mean CD4 cell counts: 198 cells/mm3; HIV RNA 2.62 log10 copies/ml), but exhibited consistently mean low lymphoproliferative responses of 2.26 SI to CMV. Both patients were thus missclassified primarily because of an unexpected lymphoproliferative response to CMV.


In this prospective observational study, five out of 17 (29%) CMV retinitis patients who discontinued CMV maintenance after responding to HAART experienced a recurrence of CMV retinitis. Although we have previously shown that sustained suppression of HIV replication is not a prerequisite for preventing reactivation of retinitis, these patients were immunological and virological failures of HAART (median CD4 cell counts of 37 cells/mm3 and plasma HIV RNA of 5.5 log10 copies/ml) at the time of reactivation of CMV disease. These data suggest that the CD4 cell threshold for the risk of recurrent CMV retinitis in patients without anti-CMV treatment and failing HAART is similar to that of initial episodes of CMV retinitis in HAART-na]ve patients (CD4 cell count < 50 cells/mm3) [22].

Because CMV retinitis recurred anywhere between 8 days and 10 months after CD4 cell counts dropped below 50 cells/mm3, CD4 cell counts below this threshold are insufficient to predict at what time patients will reactivate disease. In the pre-HAART era, CMV viremia was highly predictive of CMV retinitis occurring within 3 months in patients with CD4 cell counts below 50 cells/mm3[23]. Similarly, in HAART failures one could expect CMV viremia to precede and predict imminent recurrence. However, neither of the two participants tested had detectable CMV in peripheral blood mononuclear cells (PBMC) 1 and 7 months before reactivation when CD4 cell counts were 39 and 38 cells/mm3, respectively. CMV DNA was not detected in the PBMC of eight additional patients without reactivation of CMV retinitis. Because of the limited number of measures available in this study, no conclusions on the predictive value of quantitative CMV PCR in blood are warranted. Several ongoing studies will provide an answer to this important question.

None of the participants followed the recommendation to restart CMV maintenance upon confirmation of a CD4 cell count drop below 50 cells/mm3. Factors contributing to this decision included drug toxicity and the need for intravenous access to administer available anti-CMV agents. At present, intravenous or intraocular ganciclovir, cidofovir or foscarnet are the mainstay of anti-CMV therapy. Without symptoms or signs of the recurrence of disease, patients were thus not willing to embark on a course of one of these therapies.

Our study suggests that the decision not to resume CMV therapy can have serious consequences. Contralateral eye involvement and disseminated disease occurred in two patients, suggesting that recrudescent CMV viremia may disseminate to many organs in the absence of maintenance therapy once the CD4 cell counts drop below 50 cells/mm3. Moreover, because of the frequent central nervous system involvement in patients with CMV retinitis [24], the threat of neurological disease increases with untreated infection. Although retinal disease can be easily detected by systematic retinal examinations and treated before damage vital to the vision ensues, neurological disease can be difficult to diagnose and treat.

Our finding of robust lymphoproliferative responses to CMV in 11 out of 17 patients on HAART suggests that the restoration of specific immune responses is possible. Although these individuals were not tested before treatment with potent antiretroviral agents, our earlier studies indicate that responses to CMV are absent in patients with active CMV disease [25]. These data are consistent with those recently reported by Komanduri et al. [26], who, using a flow-cytometry based assay measuring single cytokine expression, demonstrated that effective restoration of CMV-specific CD4 lymphocyte responses was possible in HAART responders with previously active CMV disease. The persistence of strong CMV-specific responses more than one year after the initiation of HAART in most of our patients suggests that CMV-specific memory T cells are reconstituted during HAART and are maintained if the CD4 cell response to HAART is sustained.

In our study, five out of the six patients with low lymphoproliferative responses experienced reactivation of CMV retinitis a median of 14.5 months after the discontinuation of CMV maintenance. These patients did not have significant lymphoproliferative responses to CMV even when CD4 cell counts were between 50 and 100 cells/mm3. Our data thus suggest that functional lymphoproliferative responses to CMV are restored with effective HAART, but are lost with a failing regimen.

We found that a model containing HIV RNA and CD4 cell counts at maintenance withdrawal, maximal CD4 cell count and nadir HIV RNA ever on HAART, and average lymphoproliferative response above 3 SI could correctly group 88.2% of patients and thus predict who would experience reactivation of CMV retinitis. Of note was the fact that lymphoproliferative response alone contributed to 72% of the model.

An important limitation of this study is that the first lymphoproliferative assay was performed on average 9.3 months after discontinuing maintenance and more than one year after starting HAART. A larger study with systematic measurements is thus needed to determine the independent contribution of a single lymphoproliferative assay before stopping CMV therapy and to establish its value in determining whether it is safe to stop maintenance and when to resume treatment. Meanwhile, immediate re-institution of CMV therapy appears to be indicated in patients who stopped CMV maintenance and who are failing HAART immunologically (CD4 cell count < 50 cells/mm3). For patients who elect not to restart anti-CMV treatment, frequent ophthalmological examinations may detect early reactivation and allow the resumption of CMV therapy before clinically significant progression or extraocular disease occur.


The authors thank Richard H. Haubrich, MD, and Jimmy Hwang, PhD, from the UCSD Treatment Center Data and Biostatistical Unit of the University of California, San Diego, for conducting the complex statistical analyses. The authors also thank Diane Havlir, MD, and J. Allen McCutchan, MD, for critical review of the manuscript.


1. Hammer SM, Squires KE, Hughes MD. et al. A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less. :AIDS Clinical Trials Group 320 Study Team [see Comments]. N Engl J Med 1997, 337: 725 –733.
2. Hirsch M, Meibohm A, Rawlins S, Leavitt R. Indinavir (IDV) in combination with zidovudine (ZDV) and lamivudine (3TC) in ZDV-experienced patients with CD4 cell counts < 50 cells/mm3.4th Conference on Retroviruses and Opportunistic Infections, Washington, DC; 1997:207 [abstract LB7].
3. Connors M, Kovacs JA, Krevat S. et al. HIV infection induces changes in CD4+ T-cell phenotype and depletions within the CD4+ T-cell repertoire that are not immediately restored by antiviral or immune-based therapies. Nat Med 1997, 3: 533 –540.
4. Autran B, Carcelain G, Li TS. et al. Positive effects of combined antiretroviral therapy on CD4+ T cell homeostasis and function in advanced HIV disease [see Comments]. Science 1997, 277: 112 –116.
5. Gulick RM, Mellors JW, Havlir D. et al. Treatment with indinavir, zidovudine, and lamivudine in adults with human immunodeficiency virus infection and prior antiretroviral therapy [see Comments]. N Engl J Med 1997, 337: 734 –739.
6. Baril L, Jouan M, Caumes E, Mengual X, Bricaire F, Katlama C. The impact of highly active antiretroviral therapy on the incidence of CMV disease in AIDS patients.37th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, Canada, 1997:248 [abstract I-31].
7. Palella FJ, Delaney KM, Moorman AC. et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. :HIV Outpatient Study Investigators [see comments]. N Engl J Med 1998, 338: 853 –860.
8. Jacobson MA, French M. Altered natural history of AIDS-related opportunistic infections in the era of potent combination antiretroviral therapy. AIDS 1998, 12 (Suppl. A) : S157 –163.
9. Furrer H, Egger M, Opravil M. et al. Discontinuation of primary prophylaxis againstPneumocystis cariniipneumonia in HIV-1-infected adults treated with combination antiretroviral therapy. :Swiss HIV Cohort Study [see Comments]. N Engl J Med 1999, 340: 1301 –1306.
10. Jabs DA, Enger C, Bartlett JG. Cytomegalovirus retinitis and acquired immunodeficiency syndrome. Arch Ophthalmol 1989, 107: 75 –80.
11. Gallant JE, Moore RD, Richman DD, Keruly J, Chaisson RE. Incidence and natural history of cytomegalovirus disease in patients with advanced human immunodeficiency virus disease treated with zidovudine. :The Zidovudine Epidemiology Study Group [see Comments]. J Infect Dis 1992, 166: 1223 –1227.
12. Tural C, Romeu J, Sirera G. et al. Long-lasting remission of cytomegalovirus retinitis without maintenance therapy in human immunodeficiency virus-infected patients. J Infect Dis 1998, 177: 1080 –1083.
13. Macdonald JC, Torriani FJ, Morse LS, Karavellas MP, Reed JB, Freeman WR. Lack of reactivation of cytomegalovirus (CMV) retinitis after stopping CMV maintenance therapy in AIDS patients with sustained elevations in CD4 T cells in response to highly active antiretroviral therapy. J Infect Dis 1998, 177: 1182 –1187.
14. Vrabec TR, Baldassano VF, Whitcup SM. Discontinuation of maintenance therapy in patients with quiescent cytomegalovirus retinitis and elevated CD4+ counts. Ophthalmology 1998, 105: 1259 –1264.
15. Jabs DA, Bolton SG, Dunn JP, Palestine AG. Discontinuing anticytomegalovirus therapy in patients with immune reconstitution after combination antiretroviral therapy. Am J Ophthalmol 1998, 126: 817 –822.
16. Jouan M, Saves M, Tubiana R, et al. RESTIMOP (ANRS 078): a prospective multicentre study to evaluate the discontinuation of maintenance therapy for CMV retinitis in HIV-patients receiving HAART.6th Conference on Retroviruses and Opportunistic Infections, Chicago, IL; 1999:153 [abstract A56].
17. SOCA. Assessment of cytomegalovirus retinitis. :Clinical evaluation vs. centralized grading of fundus photographs. Studies of Ocular Complications of AIDS Research Group, AIDS Clinical Trials Group. Arch Ophthalmol 1996, 114: 791 –805.
18. Chatellard P, Sahli R, Iten A, von Overbeck J, Meylan PR. Single tube competitive PCR for quantitation of CMV DNA in the blood of HIV+ and solid organ transplant patients. J Virol Methods 1998, 71: 137 –146.
19. Schrier RD, Wiley CA, Spina C, McCutchan JA, Grant I. Pathogenic and protective correlates of T cell proliferation in AIDS. J Clin Invest 1996, 98: 731 –740.
20. Afifi AA, Clark V. Discriminant analysis. In:Computer-aided multivariate analysis. Edited by Chapman and Hall. New York: Reinhold; 1990. pp. 271–314.
21. Smith IL, Macdonald JC, Freeman WR, Shapiro AM, Spector SA. Cytomegalovirus (CMV) retinitis activity is accurately reflected by the presence and level of CMV DNA in aqueous humor and vitreous. J Infect Dis 1999, 179: 1249 –1253.
22. Kuppermann BD, Petty JG, Richman DD. et al. Correlation between CD4+ counts and prevalence of cytomegalovirus retinitis and human immunodeficiency virus-related noninfectious retinal vasculopathy in patients with acquired immunodeficiency syndrome. Am J Ophthalmol 1993, 115: 575 –582.
23. 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.
24. McCutchan JA. Cytomegalovirus infections of the nervous system in patients with AIDS. Clin Infect Dis 1995, 20: 747 –754.
25. Schrier RD, Freeman WR, Wiley CA, McCutchan JA. Immune predispositions for cytomegalovirus retinitis in AIDS. J Clin Invest 1995, 95: 1741 –1746.
26. Komanduri KV, Viswanathan MN, Wieder ED. et al. Restoration of cytomegalovirus-specific CD4+ T-lymphocyte responses after ganciclovir and highly active antiretroviral therapy in individuals infected with HIV-1. Nat Med 1998, 4: 953 –956.

cytomegalovirus maintenance discontinuation; cytomegalovirus-specific responses; highly active antiretroviral therapy responders; prognostic factors of reactivation; reactivation of cytomegalovirus disease

© 2000 Lippincott Williams & Wilkins, Inc.