Highly active antiretroviral therapy (HAART) has dramatically changed the prognosis of HIV infection, with a decrease in the number of HIV-related events, hospitalizations and deaths in various trials and clinical settings[1-5]. It has been shown that immune restoration occurs after the first months of HAART, and clinical and immunological observations have led to the suggestion that primary prophylaxes for opportunistic infections be discontinued in subjects showing sustained increases in CD4 cell counts during HAART. Nevertheless, the frequency of therapeutic failure is greater than expected for a number of reasons: drug toxicity and consequent therapy discontinuation, poor compliance, the incomplete abatement of virus replication, and the selection of resistant mutants[8-10]. Moreover, it has been reported that up to 21% of the patients receiving HAART have discrepant immunological and virological responses. It is clearly important to evaluate the long-term clinical outcome of HAART on the basis of the occurrence of new or recurrent AIDS-related events and death, and to identify the factors predictive of failure (in particular, the relative weights of immunological and virological responses), as this may lead to differentiating the therapeutic regimen and/or increasing the frequency of control examinations.
We therefore decided to evaluate the frequency and type of new and recurrent AIDS-related events, death and any independent predictive factors of their occurrence in a large cohort of severely immunodepressed patients receiving HAART.
Patients and methods
This observational study involved all of the patients who initiated a first HAART regimen including two reverse transcriptase inhibitors (NRTIs) and one protease inhibitor (PI) at the University of Milan‚s Institute of Infectious and Tropical Diseases between April 1996 and April 1998. The antiretroviral drugs were given at standard doses: zidovudine, 300 mg twice daily; didanosine, 200 or 100 mg twice daily (depending on whether body weight was more or less than 60 kg); zalcitabine, 0.75 mg three times daily; stavudine, 40 or 30 mg twice daily (depending on whether body weight was more or less than 60 kg); lamivudine, 150 mg twice daily, indinavir (IDV), 800 mg twice daily; ritonavir (RTV), 600 mg twice daily; hard-gel saquinavir (SQV), 600 mg three times daily.
Primary prophylaxis for cytomegalovirus (CMV) disease and non-tuberculous mycobacterial infection were not applied as standard practice at our Institution, but primary prophylaxis for Pneumocystis carinii pneumonia (PCP) and Toxoplasma gondii infection was given to patients with CD4 cell counts below 200×106cells/l according to International Guidelines; in the case of an increase in CD4 cell count to above this limit during HAART, the decision as to whether or not to discontinue prophylaxis was taken by the physician on an individual basis. In some cases, primary prophylaxis was discontinued in the absence of this immunological condition when requested by the patient. All the secondary prophylaxes were given according to international guidelines. All secondary prophylaxes for opportunistic infections were continued during HAART, independently from CD4 cell count.
The following information was collected at baseline: age, sex, risk factors for HIV infection, Centers of Disease Control stage according to the 1993 classification, the date of starting first antiretroviral regimen, the date of HAART initiation and the drug regimen.
Lymphocyte subset (Elite flow cytometer; Coulter Corporation, Miami, Florida, USA) and HIV RNA measurements (branched DNA, Chiron Inc., Emeryville, California, USA; detection limit: 500 copies/ml) were made at baseline and then every 3 months.
New or recurrent AIDS-defining events and death, as well as the discontinuation of HAART and the reason for it were also recorded. The causes of death were carefully determined by means of clinical charts and autopsy reports in order to ascertain the role of AIDS-related phenomena and any toxic events. The last date of follow-up was 30 November 1998.
Clinical failure was defined as the occurrence of a new or recurrent AIDS-defining event or death, whichever occurred first.
The data were analysed by means of intention-to-treat analysis. Categorical variables between groups were compared using Pearson‚s χ2 and Fisher‚s exact test, and the distributions of continuous variables by means of the Mann-Whitney U test. In order to identify any factor predictive of new or recurrent AIDS-defining events or death, univariate and multivariate analyses were performed using Cox‚s regression model. End-point was clinical failure; persons who did not reach the end-point were censored at their last clinical visit. The factors considered in the analyses were those thought to affect outcome: sex; age; risk factor for HIV infection; AIDS, CD4 cell counts and HIV-RNA copies at baseline. Cox‚s hazard proportional model was also used to evaluate the influence of 3-month CD4 cell counts on clinical outcome. Only patients reaching the third month of follow-up without clinical failure and with 3-month CD4 cell count determination were included in this last analysis. We considered the immunological response at the third month in two different ways: not reaching CD4 > 200×106cells/l in those patients starting HAART with counts lower than 200×106cells/l, and not reaching an increase of at least 20% from baseline count in those patients starting HAART with CD4 cell counts higher than 200×106cells/l. The SPSS software package (SPSS version 6.1.2; SPSSInc., Chicago, IL, USA) was used to perform all of the analyses.
Between 1 April 1996 and 30 April 1998, 585 patients (423 males: 72.3%) with a median age of 34 years (range: 20-66) started a first HAART regimen containing IDV (n=444), RTV (n=78) or SQV (n=63). Thirty-seven (6.3%) were active parenteral drug addicts (PDA); 230 (39.3%) former PDA; 98 (16.8%) homosexual men; and 220 (37.6%) were infected by a heterosexual or unknown route. The median baseline CD4 cell count was 141×106cells/l (range: 1-695): 165 patients (28.2%) had baseline counts of <50×106cells/l and 236 (40.3%) of >200×106cells/l. The median number of HIV-RNA copies at baseline was 27774/ml (range: <500-1164834): 44 patients (7.5%) had undetectable plasma viraemia and 126 (21.5%) HIV-RNA >100000 copies/ml. HIV-RNA was not determined at baseline in 119 cases (19.8%). One hundred and sixteen (19.8%) of the subjects were on their first antiretroviral treatment; the remaining 469 had received a median of 937 days (range: 7-3501) of previous non-HAART regimens, of whom 32% added the protease inhibitor to their current regimen without changing or adding any new NRTIs.
Among the 121 patients with AIDS at baseline (20.7%), the following AIDS-defining diseases had been diagnosed with a frequency of more than 5%: PCP (n=41; 33.9%); oesophageal candidiasis (n=28; 23.1%); CMV disease (n=21; 17.4%); Mycobacterium avium complex (MAC) infection (n=14; 11.6%); brain toxoplasmosis (n=13; 10.7%); extra-pulmonary tuberculosis (n=11; 9.1%) and Kaposi‚s sarcoma (n=10; 8.3%).
Primary prophylaxis for PCP was being received at HAART initiation by 330 of 349 patients with CD4 cell counts below 200×106cells/l: cotrimoxazole in 315, and aerosolized pentamidine in 15.
During a median follow-up period of 483 days (range: 33-1334), 17 patients (2.9%) died, 38 (6.5%) experienced a total of 45 new AIDS-defining events, and nine of the 121 AIDS patients (7.4%) experienced one recurrent AIDS-defining event each.
One hundred and sixty-six patients (50.3%) discontinued primary PCP prophylaxis after a median of 268 days (range: 1-937): 102 because of persistent CD4 cell counts above 200×106cells/l, and 64 upon their own decision despite CD4 cell counts of <200×106cells/l.
The deaths occurred after a median of 229 days (range: 37-559): ten were AIDS-related (due to new AIDS-defining events in seven and pre-existing diseases in three). Seven other patients died as a result of hepatic failure (n=3), suicide (n=2), renal failure (n=1) and cerebral haemorrhage (n=1).
The 45 new AIDS-defining events were diagnosed a median of 98 days (range: 1-777) after the initiation of HAART, with a median CD4 count of 98×106cells/l (range: 1-675) and a median HIV-RNA copy number of 4466copies/ml (range: 499-490000). Table 1 shows the AIDS-defining events in relation to their time of occurrence: 24 of 45 (53.4%) occurred within the first 3 months of HAART treatment, eleven (24.4%) between the fourth and sixth month, five (11.1%) were diagnosed between the seventh and twelfth month and five (11.1%) occurred after the twelfth month. The most commonly diagnosed diseases were: CMV disease and oesophageal candidiasis (six cases each), MAC infection and recurrent bacterial pneumonia (five cases each), progressive multifocal leukoencephalopathy (PML), tuberculosis and Kaposi‚s sarcoma (four cases each). All but seven patients experiencing a new AIDS-defining event had CD4 counts below 200×106cells/l at HAART initiation; 24 (66.7%) had CD4 cell counts of less than 50×106cells/l. The median CD4 cell count at diagnosis increased with the length of time from the start of HAART but, as this was an intent-to-treat analysis, not all of the patients were still on HAART treatment.
As shown in Table 2, 11 of 45 AIDS-defining events (24.4%) occurred in patients with CD4 cell counts >200×106cells/l: six of 24 (three Kaposi‚s sarcoma, one PML, one MAC infection and one oesophageal candidiasis) within 3 months of the start of HAART, two of 11 (both non-Hodgkin lymphoma) between the fourth and sixth month, two of five (one MAC infection and one primary brain lymphoma) between the seventh and twelfth month, and one oesophageal candidiasis (not shown in the table) after the twelfth month. Furthermore, two of the 21 tested patients had undetectable viraemia at the time of the diagnosis of the new event. The table also shows the relative risks of the occurrence of events at the different times in relation to CD4 cell counts below 200×106cells/l and >500 HIV-RNA copies/ml at the previous determination: only a 3-month CD4 cell count below 200×106cells/l statistically correlated with a higher risk of disease occurrence in the subsequent 3 months [relative risk 5.26: 95% confidence intervals (CI) 1.1-25].
Furthermore, the three cases of PCP and the two cases of brain toxoplasmosis were diagnosed during the first 3 months of HAART in patients still receiving primary prophylaxis whose CD4 cell counts were <200×106cells/l.
The nine relapses of pre-existing AIDS-defining diseases were diagnosed after a median of 288 days (range: 44-749) from HAART initiation, with a median CD4 cell count of 134×106cells/l (range: 23-187) and a median HIV-RNA copy number of 5534 copies/ml (range: 499-184850). All of the relapses occurred in patients with CD4 cell counts of less than 200×106cells/l: five cases of CMV retinitis (all occurring during anti-CMV maintenance treatment), and one each of brain toxoplasmosis, MAC infection (both during maintenance therapy), oesophageal candidiasis and bacterial pneumonia. The frequency of relapses among the 121 AIDS patients in relation to their pre-existing condition was five of 21 (23.8%) for CMV disease, one of 14 (7.1%) for MAC infection, one of 13 (7.7%) for brain toxoplasmosis and one of 28 (3.6%) for oesophageal candidiasis.
Clinical failure occurred in 55 patients (9.4%) after a median of 130 days (range: 1-777) as a result of new AIDS-defining events in 38 patients (6.5%), recurrent AIDS-defining events in six (1.0%), and death (including non-AIDS related events) in 11 (1.9%). Fifty-one of 55 patients with clinical failure were receiving the initial HAART regimen. In particular, clinical failure occurred in 35 out of 400 patients who never changed their HAART regimen, and in 20 out of 185 patients who shifted the initial regimen; of these, clinical failure occurred in 16 before and in four after shifting regimen.
Both mean CD4 cell counts and mean HIV-RNA copy numbers between HAART initiation and clinical failure or the last follow-up examination were determined in 536 cases (26 with and 510 without clinical failure): two of 26 patients (7.7%) with and 126 of 510 (24.7%) without clinical failure had mean CD4 cell counts >200×106cells/l and mean HIV-RNA levels of <500 copies/ml; three of 26 (11.5%) with and 226 of 510 (44.3%) without clinical failure had mean CD4 counts of >200×106cells/l and mean HIV-RNA levels >500 copies/ml; nine of 26 (34.6%) with and 55 (10.8%) without clinical failure had mean CD4 cell counts of <200×106cells/l and mean HIV-RNA levels of <500 copies/ml; and 12 (46.2%) with and 103 (20.2%) without clinical failure had mean CD4 cell counts of <200×106cells/l and mean HIV-RNA levels >500 copies/ml (χ2 test: 26.75; P<0.00001). Figure 1 shows the mean CD4 cell count and mean log10 HIV-RNA copies/ml in relation to clinical failure. The mean (median; range) CD4 cell counts were 144×106cells/l (128; 4-529) in patients with and 322×106cells/l (288; 14-1162) in patients without clinical failure (P<0.0001); mean log10 plasma viraemia was 3.52 (3.1; 2.7-5.6) log10 copies/ml in patients with and 3.3 (2.99; 1.9-5.4) in patients without clinical failure (not significant).
Table 3 shows the results of the univariate and multivariate analyses of the factors predictive of clinical failure. At univariate analysis, a low CD4 cell count or high level of plasma viraemia at baseline and a previous diagnosis of AIDS were significantly associated with clinical failure. In the multivariable Cox model, the diagnosis of AIDS before HAART initiation was associated with a worse prognosis [hazard risk (HR) 1.80; 95% CI: 0.98-1.66], and baseline CD4 cell counts >200×106cells/l with a good prognosis (HR: 0.41; 95% CI: 0.19-0.91 versus CD4 cell counts of <50×106cells/l), whereas the predictivity of high plasma viraemia disappeared. The patients who acquired HIV infection as a result of heterosexual or unknown exposure had a better prognosis than parenteral drug addicts at both univariate and multivariate analysis.
Univariate and multivariate analyses were also performed on a selected population of 294 patients who reached the third month of HAART without clinical failure and had a 3-month CD4 cell count determination (Table 4). The findings observed in the whole cohort were confirmed in the multivariate analysis on these selected patients: baseline CD4 cell count >200×106cells/l were predictive of good prognosis (HR: 0.17; 95% CI: 0.03-0.86 versus CD4 cell counts of <50×106cells/l). After adjusting for the 3-month CD4 cell count determination, i.e. persistence of CD4 cell count <200×106cells/l in patients with this baseline condition, or lack of at least 20% increase from baseline count in patients with baseline CD4 cell count >200×106cells/l, this marker was the only one strongly associated with clinical failure (HR 4.79; 95% CI: 1.40-16.47).
This study related both immunological and virological markers to the clinical outcome of HAART, and allows a number of conclusions to be drawn. First, half of the new AIDS-defining events occurred within the 3 months following HAART initiation and may have been due to delayed immune restoration because having CD4 cell counts of more than 200×106cells/l did not protect against the occurrence of AIDS in the first months, but was related to a reduction in risk thereafter. Second, all of the recurrences of opportunistic infections occurred in severely immunodepressed patients, with CMV disease being the most frequent. Third, no cases of PCP and brain toxoplasmosis occurred in the patients discontinuing prophylaxis or in those with CD4 cell counts above 200×106cells/l; furthermore, as both diseases were only diagnosed during the first 3 months of HAART and even if no definitive conclusion can be drawn, these findings support the safety of discontinuing primary prophylaxis for these infections. Fourth, we observed a close relationship between mean CD4 counts during HAART and clinical failure, whereas no association was found with mean plasma HIV-RNA levels. This last measure might be not as sensitive as mean CD4 cell counts in detecting clinical failure. Fifth, in the multivariable model, the independent factors predictive of clinical failure were a diagnosis of AIDS and baseline CD4 cell counts of less than 200×106cells/l. Sixth, in the patients failing more than 3 months after HAART initiation, a diagnosis of AIDS and baseline HIV-RNA levels of more than 100000 copies/ml, and CD4 cell counts of less than 200×106cells/l after 3 months of HAART, were all independently predictive of failure regardless of the baseline CD4 cell counts.
Looking at the details of our study, clinical failure was observed in 55 out of 585 patients (9.4%) during the median follow-up of 16 months, as shown by the presence of new AIDS-defining events in 38 (6.5%), recurrent AIDS-defining events in six (1.0%) and death in 11(1.9%). Most of the patients were severely immunodepressed at baseline, and more than 80% had been heavily pre-treated with suboptimal antiretroviral regimens. This frequency of clinical failure partially fits with that reported in other recent studies. The ritonavir placebo-controlled trial  showed a higher frequency of disease progression (21.9% during a median follow-up of 230 days), but included more severely immunodepressed patients who just added ritonavir to their current failing regimen. The indinavir trial  found a lower frequency of clinical failure (6% in the three-arm group) during a shorter period of follow-up (median: 38 weeks) than in our cohort. The higher frequency of failure in the Aquitaine group, in which disease progression was observed in 79 out of 556 patients (14.2%) during a median follow-up of 230 days, might be partially explained by the large proportion of patients with AIDS (44%). The frequency of disease progression in Michelet‚s cohort  was similar to ours even if the follow-up was shorter: 46 patients (9.5%) with mean baseline CD4 cell counts of 22×106cells/l experienced a clinical event during a 6-month period.
More than half of the new AIDS-defining events observed in this study occurred within the first 2 months of HAART, in agreement with other reports. There was no predictive advantage in reaching CD4 cell counts of 200×106cells/l or more in these cases, which is consistent with the delayed restoration of immune function during HAART[6,16].
CMV disease (mainly retinitis) was the most common new or recurrent disease. As reported by other authors, the earliest episodes occurred in patients with low CD4 cell counts (median 45×106cells/l; range: 3-98) and mainly during the first months of HAART. Relapsing CMV disease occurred at higher than expected CD4 counts (median: 145×106cells/l; range: 57-187), which might raise doubts to the safety of discontinuing anti-CMV maintenance during HAART. However, it has to be emphasised that the CMV relapses occurred in a smaller percentage (23%) of patients than those experiencing relapses in the pre-HAART era[19,20]. MAC infection was a quite common opportunistic disease too (five out of 45 AIDS-defining events). It has to be emphasised that MAC prophylaxis was not given to our patients, due to the low incidence of MAC infections previously reported.
Death was the first event in 11 cases. Two cases of suicide were recorded, one of them in a subject with a good clinical, immunological and virological response to therapy.
One important finding of this study is that, as documented by mean CD4 and HIV-RNA levels, only 23.9% of the subjects had a concordantly good immuno-virological response to HAART, and a concordant immuno-virological failure was observed in another 21.5%. This means that more than half of the patients had discordant immunological and virological responses: in 11.9%, mean plasma viraemia was undetectable despite mean CD4 cell counts of less than 200×106cells/l and, in 42.7%, was detectable although their mean CD4 cell counts were above 200×106cells/l. Figure 1 strikingly shows that mean CD4 cell counts were predictive of clinical failure, whereas mean viraemia was not.
A similar analysis has been made by Piketty et al.. Using a weaker definition of response (i.e. CD4 cell counts reaching at least 50×106cells/l and HIV-RNA decreasing by at least one log10 copies), they found discordant immuno-virological responses in a smaller number of patients (21% of 162 patients); the incidence of AIDS-defining events was significantly different between responders and non-responders; and immunological response was found to be a predictor of clinical success even in the presence of virological failure.
We have recently reported our observations of T lymphocyte subset variations during HAART in subjects with previous severe CD4 cell depletion. The patients with good virological responses but different levels of increase in CD4 cells after six or more months of therapy had qualitatively different CD4 cell subsets: in particular, the percentages of CD4 naive cells were significantly lower in the patients with a poor immunological response. The difficulty in achieving immunorestoration in the absence of a functional thymus gland in adult patients, and the possibility that HAART alone is not sufficient to guarantee it, has been wisely debated. The present study documents the fact that a poor immunological response is a predictor of disease progression even in subjects with low or undetectable plasma viremia.
As expected, and in agreement with other reports, univariate analysis showed that the markers of disease severity at baseline (CD4 cell counts less than 50×106cells/l, HIV-RNA>100000 copies/ml and AIDS) were all predictive of clinical failure. In the multivariable model, HIV-RNA levels lost their predictive role, whereas baseline CD4 cell counts >200×106cells/l had a protective effect on clinical failure; AIDS remained a predictor of worse prognosis. After the first months of therapy, the 3-month immunological response represents the best marker of clinical outcome. This might suggest that, independently from baseline conditions, once an immunological response has been reached, at least in terms of number of circulating cells, this may positively affect the clinical outcome.
Finally, the analyses showed that heterosexually infected subjects had a better prognosis than parenteral drug addicts and homosexuals. Behavioural categories may be related to treatment compliance  and thus have an indirect effect on outcome; moreover, heterosexually infected patients more frequently have a short history of antiretroviral treatments and, as previously shown by our group, are more likely to be naive to anti-HIV drugs, both of which may explain the better outcome in this population.
In conclusion, the results of this study suggest that a more aggressive antiretroviral treatment including four drugs or more potent combinations may be indicated in subjects with very high plasma viraemia or AIDS; in addition, antiretroviral resistance testing may help with treatment choices in persons with previous mono or dual therapy; that particular attention needs to be paid during the first months of HAART, especially in severely immunodepressed subjects; and, finally, that the determination of 3-month CD4 cell counts is crucial for predicting the clinical outcome. Further studies are needed in order to explore the possible role of CD4 cell subsets as additional prognostic indicators.
1. Cameron DW, Hearth-Chiozzi M, Danner S et al. Randomized placebo-controlled trial of ritonavir in advanced HIV-1 disease. Lancet 1998, 351:543-549.
2. Hammer SM, Kathleen ES, 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. N Engl J Med 1997, 337:725-733.
3. Clumeck N on behalf of the Saquinavir International Phase III Trial (SV 14604) Group. Clinical benefit of Saquinavir (SQV) plus Zalcitabine (ddC) plus Zidovudine (ZDV) in untreated/minimally treated HIV-infected patients. Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, 1997 [abstract LB-4].
4. Palella FJ, Delaney KM, Moorman AC, et al. Declining morbidity and mortality among patients with advanced Human Immunodeficiency Virus Syndrome. N Engl J Med 1998, 338:853-860.
5. Mouton Y, Alfandari S, Valette M, et al. Impact of protease inhibitors on AIDS-defining events and hospitalizations in 10 French AIDS reference centres. Federation National des Centres de Lutte contre le SIDA. AIDS 1997, 11:F101-F105.
6. 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. Science 1997, 277:112-116.
7. Schneider MME, Borleffs JCC, Stolk RP, Japers CAJJ. Hoepelman IM. Discontinuation of Pneumocystits carinii pneumonia prophylaxis in HIV-1 infected patients treated with highly active antiretroviral therapy. Lancet 1999, 353:201-203.
8. Fatkenheuer G, Theisen A, Roskstroh J, et al. Virological treatment failure of protease inhibitor therapy in an unselected cohort of HIV-infected patients. AIDS 1997, 11:F113-F116
9. d‚Arminio Monforte A, Testa L, Adorni F et al. Clinical outcome and predictive factors of failure of highly active antiretroviral therapy in antiretroviral-experienced patients in advanced stages of HIV-1 infection. AIDS 1998, 12:1631-1637.
10. Vanhove GF, Schapiro JM, Winters MA, Merigan TC, Blaschke TF. Patient compliance and drug failure in protease inhibitor monotherapy. JAMA 1996, 276:1955-1956.
11. Piketty C, Castiel P, Belec L, et al. Discrepant responses to triple combination antiretroviral therapy in advanced HIV disease. AIDS 1998, 12:745-750.
12. US Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention (CDC). 1997 USPHS/IDSA guidelines for the prevention of opportunistic infections in persons infected with Human Immunodeficiency Virus. MMWR 1997, 46:RR-12.
13. 1993 Revised Classification System for HIV infection and Expanded Surveillance Case Definition for AIDS among adolescents and adults. MMWR 1992, 41:19-23
14. Chene G, Binquet C, Moreau J-F, et al. Changes in CD4+ cell count and the risk of opportunistic infection or death after highly active antiretroviral treatment. AIDS 1998, 12:2313-2320.
15. Michelet C, Arvieux C, François C, et al. Opportunistic infections occurring during highly active antiretroviral treatment. AIDS 1998, 12:1815-1822.
16. Powderly WG, Landay A, Lederman MM. Recovery of the immune system with antiretroviral therapy. The end of opportunism? JAMA 1998, 280:72-77.
17. Jacobson MA, Zegans M, Pavan PR, et al. Cytomegalovirus retinitis after initiation of highly active antiretroviral therapy. Lancet 1997, 349:1443-1445.
18. Tural C, Romeu J, Andreu D, et al. Long-lasting remission of cytomegalovirus retinitis without maintenance therapy in human immunodeficiency virus-infected patients. J Infect Dis 1998, 177:1080-1083.
19. Gallant JE, Moore RD, Richman DD, Keruly J, Chaisson RR and the 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.
20. d‚Arminio Monforte A, Mainini F, Testa L, et al. Predictors of cytomegalovirus disease, natural history and autopsy findings in a cohort of patients with AIDS. AIDS 1997, 11:517-524.
21. d‚Arminio Monforte A, Vago L, Gori A, et al. Clinical diagnosis of mycobacterial diseases versus autopsy findings in 350 patients with AIDS. Eur J Clin Microbiol Infect Dis 1996, 15:453-458
22. Meroni L, Varchetta S, Manganaro D, Moscatelli GC, Moroni M, Galli M. CD4+CD7- lymphocyte subset is expanded in HIV infected patients with low CD4+ cell count rescue during HAART. AIDS 1999, in press.
23. Levy JA. Caution: should we be treating HIV infection early? Lancet 1998, 352:982-983.
24. Wenger N, Gifford A, Liu H, et al. Patient characteristics and attitudes associated with antiretroviral (AR) adherence. 6th Conference on Retroviruses and Opportunistic Infections. Chicago, January 31-February 4, 1999 [abstract n 98].
25. Ridolfo AL, Mainini F, Santambrogio S, et al. AIDS features in patients unaware of their HIV-seropositivity. 4th Conference on Retroviruses and Opportunistic Infections. Washington, 22-26 January, 1997, [abstract 503].
© 1999 Lippincott Williams & Wilkins, Inc.