Immunologic and clinical responses to highly active antiretroviral therapy over 50 years of age. Results from the French Hospital Database on HIV

Objective: To study immunologic and clinical responses to HAART in patients over 50 years old.

Design and methods: A prospective cohort study which included 68 hospitals in France. A total of 3015 antiretroviral-naive patients, 401 of whom were aged 50 years or over, were enrolled following initiation of HAART. The influence of age on the mean CD4 cell count increase on HAART was studied by using a two-slope mixed model. Progression, defined by the occurrence of a new AIDS-defining event (ADE) or death, was studied by Cox multivariate analyses.

Results: Among patients with baseline HIV RNA above 5 log copies/ml, CD4 mean increase during the first 6 months on HAART was +42.9 × 10⁶ cells/l per month in patients under 50 years and +36.9 × 10⁶ cells/l per month in patients over 50 years (P < 0.0001); subsequently, the respective monthly changes were +17.9 and +15.6 × 10⁶ cells/l per month (P < 0.0001). Similar trends were observed in patients with baseline HIV RNA below 5 log copies/ml, and also after stratification for the baseline CD4 cell count. After a median follow-up of 31.5 months, 263 patients had a new ADE and 44 patients died. After adjustment for baseline characteristics, older patients had a significantly higher risk of clinical progression (hazard ratio (HR) = 1.52 [95% confidence interval (CI), 1.15-2.00]) and were more likely to achieve a viral load below 500 copies/ml [HR = 1.23, (95% CI, 1.11-1.38)].

Conclusion: Patients over 50 years of age have an immunologic response to HAART. However, their CD4 cell reconstitution is significantly slower than in younger patients, despite a better virologic response. This impaired immunologic response may explain their higher risk of clinical progression.

Introduction

The introduction of highly active antiretroviral therapy (HAART) has markedly improved the length of survival and the morbidity of HIV-infected patients through the improvement of the immunodeficiency caused by the HIV infection. Most published data on immunologic and clinical responses to HAART concern relatively young patients (median age, 35 to 40 years) [^1-7]. Indeed, the HIV pandemic has mainly concerned young adults, and older patients tend to be excluded from clinical trials because of the risk of age-related comorbidity.

Clinicians are now seeing an increasing number of older HIV-seropositive patients in their everyday practice. A recent report from the US Centers for Disease Control and Prevention indicates that the cumulative number of AIDS cases in adults over 50 years quintupled during the previous decade [⁸]. This trend is likely to become more pronounced as HIV infection becomes a quasi-chronic disease in wealthy countries, and as the HIV-infected population ages as a consequence of effective treatment.

Only few data specifically concerning older HIV-infected populations have been published since the advent of HAART. The most recent studies of responses to HAART among older patients mainly involved small populations and limited follow-up [^9-12]. Moreover, there is evidence that thymus involution in adults [^13,14] may negatively influence CD4 cell recovery, [^15-17] potentially leading to a lesser immunologic response to antiretroviral therapy and a higher risk of clinical progression than in younger subjects.

In the present study, we compared the immunologic and clinical responses to first-line HAART according to age (< and ≥ 50 years) in a large cohort of 3015 HIV-infected patients followed for a median period of 30 months.

Methods

Patients

Patients were selected from the French Hospital Database on HIV (FHDH), a nationwide hospital-based cohort described in detail elsewhere [¹⁸]. In brief, this epidemiological network was created in 1989. A total of 68 participating hospitals across France currently provide data on HIV-infected patients. FHDH inclusion criteria are HIV-1 or HIV-2 infection, and signature of a written informed consent. Trained research assistants prospectively collect clinical, laboratory and treatment data from medical records by using specialized software (DMI2) developed by the French Ministry of Health.

This analysis focuses on all antiretroviral-naive patients enrolled in FHDH between 1997 and 2001 and who initiated HAART. The HAART was defined as a combination of two nucleoside reverse transcriptase inhibitors (NRTI) and one protease inhibitor (PI), two NRTI and two PI, two NRTI and one non-nucleoside reverse transcriptase inhibitor (NNRTI), or three NRTI. Baseline was the date of the first HAART intake. Patients were eligible for this study if they attended at least one follow-up visit to an FHDH participating centre after HAART initiation, and if the baseline CD4 cell count and plasma HIV RNA value were available. Overall, 4695 individuals met these criteria. To study variations over time in the CD4 cell count after the beginning of HAART, we selected those patients who had at least three measurements of CD4 after baseline. Finally, 3015 patients were selected and followed until 31 December 2001.

Statistical analyses

The study population (3015 patients) was divided into two groups (< 50 years and ≥ 50 years) according to age at HAART initiation. This age cut-off is commonly used in other studies. Moreover, a large international study, the ART-cohort collaboration, recently showed, among a total of 12 594 patients, that older age was an independent predictor of clinical progression, with a threshold effect at 50 years [³].

Immune response to HAART

We studied the mean increase in the CD4 cell count over time from the beginning of HAART by using a two-slope mixed model (before and after month 6) with temporal random effects, including the age group and baseline HIV RNA value as covariates. This model allowed us to account for the two phases of CD4 cell reconstitution on HAART [^19,20]. Plasma HIV RNA at baseline was dichotomized: strictly as < 5 log copies/ml versus ≥ 5 log copies/ml.

Virologic response

Time to undetectability, defined as an HIV RNA measure < 500 copies/ml, was studied by Kaplan-Meier estimates. The curves of the two age-groups were compared using log-rank test and proportional Cox models used for multivariate analyses.

Clinical progression

Multivariate Cox proportional hazard models were used to analyse the time to the first new AIDS-defining event (ADE) or death after HAART initiation. This combined endpoint is referred in the study as clinical progression, only sensitivity analysis took into account only progression to AIDS. All deaths were included, regardless of the cause. We used the clinical definition of the 1993 Centers for Disease Control and Prevention revision of the AIDS case definition. Therefore, a CD4 cell count of less than 200 × 10⁶ cells/l was not considered as an AIDS-defining event. Disease recurrences were not considered for patients with a previous AIDS diagnosis. The relative risks of new ADE or death, and their corresponding 95% confidence intervals (CI), were estimated after controlling for the following baseline variables: age at HAART introduction (≥ 50 years versus < 50 years), gender (female versus male), HIV transmission group (homosexual versus others), AIDS status (AIDS versus no AIDS), CD4 cell count (log₂) and plasma HIV RNA (log₁₀). In order to homogenize the prescription of HAART over time, all the analyses were stratified on years of starting HAART. To account for the delay in reporting clinical events in the database, we used the following right-censoring strategy: patients free of events during the 6 months before the last database update (i.e. between June and December 2001) were censored on 31 December 2001. Patients who did not attend a follow-up visit during this 6-month period were considered lost to follow-up and were censored at the date of their last visit.

All tests were two-sided, and P values less than 0.05 were considered to denote statistical significance. All statistical analyses were performed using SAS software package version 8 (SAS Institute, Cary, North Carolina, USA). The PROC PHREG procedure was used for Cox models and the PROC MIXED procedure for linear mixed models.

Results

Study population

Of the 3015 patients included in the study, 2614 (86.7%) patients were less than 50 years old at HAART initiation and 401 (13.3%) were aged 50 years or older. Table 1 describes the patients' baseline characteristics according to the age group. Older patients were more often male, and the HIV transmission group was more often heterosexual intercourse or unknown than in the younger group. Half the older patients (49.5%) had a first positive HIV test in the month prior to FHDH inclusion, compared with 37.2% in the younger group (P < 0.0001). At HAART initiation, older patients were at a more advanced stage of HIV disease than younger patients: older patients were more likely to have had an AIDS-defining event, and tended to have lower CD4 cell counts (median 193 versus 252 × 10⁶ cells/l) and higher plasma viral load. These differences were already present at FHDH enrolment. One-third (32.4%) of the older patients had already developed AIDS at FHDH enrolment, compared with 20.2% of younger patients. The median year of HAART initiation was 1998 in both age groups. The precise HAART regimens did not differ significantly according to the age group, although older patients were less likely than younger patients to receive a combination of two NRTI and one PI (64.8 versus 70.6%). On average, HAART was initiated in older and younger patients respectively 4 and 5 months after enrolment in the database (P = 0.02).

Table 1

Immunologic response

When analysing the mean CD4 cell increase after initiation of HAART for each age group, we accounted for baseline plasma HIV RNA levels (< 5 log copies/ml versus ≥ 5 log copies/ml) (Fig. 1a) and used random-effect regression models to estimate the slopes.

Fig. 1

In each age group and baseline viral load category, the slopes of the mean CD4 cell count increment were significantly steeper during the first 6 months of HAART than subsequently.

During the first 6 months of HAART, among patients with baseline HIV RNA less than 5 log copies/ml, the CD4 count mean increase was +17.3 (95% CI, 14.9-19.7) × 10⁶ cells/l per month in the younger group, and +14.1 (95% CI, 7.6-20.6) × 10⁶ cells/l per month in the older group (P < 0.0001). Beyond 6 months, the respective mean increases were +11.1 (95% CI, 10.4-11.8) × 10⁶ cells/l per month in the younger group, and +9.8 (95% CI, 7.8-11.9) × 10⁶ cells/l per month in the older group (P < 0.0001).

Similar results were observed among patients with baseline HIV RNA above 5 log copies/ml. The slopes were +42.9 (95% CI, 39.4-46.3) × 10⁶ cells/l per month in the younger group and +36.9 (95% CI, 29.3-44.4) × 10⁶ cells/l per month in the older group during the first 6 months (P < 0.0001), and +17.9 (95% CI, 17.0-18.9) and +15.6 (95% CI, 13.3-17.8) × 10⁶ cells/l per month, respectively, thereafter (P < 0.0001). The mean CD4 cell count increase was also significantly higher in younger patients (P = 0.0001) when baseline CD4 cell counts were taken into account (Fig. 1b and c).

These models estimate that, among patients with baseline HIV RNA less than 5 log copies/ml, the mean time to increase the CD4 cell count by 200 × 10⁶ cells/l is 14.7 months (95% CI, 12.9-16.7) in younger patients and 17.7 months (95% CI, 12.4-25.8) in older patients. It is 4.7 (95% CI, 4.3-5.1) and 5.4 months (95% CI, 4.5-6.8), respectively, in patients with baseline HIV RNA above 5 log copies/ml.

Virologic response

The analyses of time to undetectability (defined as a viral load measure < 500 copies/ml) performed by Kaplan-Meier estimates showed that older patients progressed more rapidly to undetectability than younger patients (log-rank P = 0.0053). At 6 months, 70.6% (95% CI, 68.8-72.3) of the patients younger than 50 years of age reached a viral load less than 500 copies/ml; the corresponding value for patients ≥ 50 years old was 76.6% (95% CI, 72.4-80.7). After adjustments, older patients were significantly more likely to achieve a viral load less than 500 copies/ml [hazards ratio (HR) = 1.23 (95% CI, 1.11-1.38)].

Clinical progression

By 31 December 2001, after a median follow-up of 31.5 months (8241 person-years), 263 patients had had a new ADE, and 71 patients had died, 27 of whom had experienced an new ADE before dying. Older patients showed a significantly more rapid clinical progression than younger patients (Fig. 2). During the first year of HAART, 5.0% (95% CI, 4.2-5.9) of younger patients had a new ADE or died, compared to 10.2% (95% CI, 7.3-13.3) of older patients. Five years after HAART initiation, the respective proportions were 12.4% (95% CI, 10.7-14.1) and 21.9% (95% CI, 15.8-27.9). After adjustments for baseline characteristics (Table 2), the hazards ratio of clinical progression was significantly higher in the older group than in the younger group [HR = 1.52 (95% CI, 1.15-2.00)]. Similar results were observed when we took into account only the progression to a new ADE [adjusted HR = 1.50 (95% CI, 1.11-2.02), P = 0.0087].

Fig. 2

Table 2

Table 3 shows the incidence rates of the different opportunistic infections and tumours that occurred during follow-up in each age group. The incidence of cytomegalovirus (CMV) disease (HR = 5.1), HIV encephalopathy (HR = 2.8) and Kaposi's sarcoma (HR = 2.3) was significantly higher in the older group than in the younger group. Progressive multifocal leukoencephalopathy (PML) (HR = 3.0) and wasting syndrome (HR = 3.9) also tended to be more frequent in older patients. Fourteen of the 21 cases of CMV disease (six in the younger group, eight in the older group) concerned sites other than the retina. The 23 lymphomas that occurred during follow-up comprised 21 non-Hodgkin lymphomas and two primary central nervous system lymphomas. The incidence of non-Hodgkin lymphomas was slightly higher in the older group (3.5 per 1000 person-years versus 2.3) but that difference did not reach significance [HR = 1.5 (95% CI, 0.5-4.5)].

Table 3

Discussion

This study of a cohort of 3015 patients, of whom 401 were over 50 years of age, shows that older patients do exhibit a CD4 cell response to HAART although this response is significantly slower than for younger patients. This blunted immune response in older patients may explain why older patients have a more rapid clinical progression.

A large population of patients from a national study and the large number of patient-years of follow up, which allowed us to examine clinical progression up to almost 3 years after starting HAART, are important strengths of this study. The FHDH includes many clinics treating HIV-1-infected patients across France, with a wide variation in patient characteristics. Our results should therefore be applicable to many individuals with HIV-1 infection in France and elsewhere. Moreover, to our knowledge, this is the first study to have examined by longitudinal modelling, the CD4 response over time according to age at HAART initiation.

For our main analysis, we selected, from among the 4695 patients who started HAART for the first time, the 3015 patients who had at least three measures of CD4 cell counts since baseline. This selection may have biased the analysis of clinical progression. Indeed, these patients may have had more frequent CD4 cell counts because they were at a higher risk of clinical events. However, the same proportions of patients were excluded for the lack of the follow-up measures of CD4 cell count: 36.3% in the younger group and 32.5% in the older group (P = 0.07). Moreover, when sensitivity analyses were run on the 4695 patients (of whom 594 were over 50 years of age), the same results were obtained for clinical progression. The risk of clinical progression was 1.33-fold (95% CI, 1.06-1.69) higher among older patients than among younger patients.

One of the striking differences between the older group and the younger one is that the older patients were at a much more advanced stage of HIV disease than the younger patients, both at inclusion in this study and at FHDH enrolment. Moreover, the distribution of HIV transmission categories was quite particular in the older group. About half these patients (46.6%) were heterosexuals, very few were intravenous drug users (0.3 versus 13.7% of younger patients) and the HIV transmission group was frequently unknown (16.5 versus 7.9%). Similar findings were reported for American patients with AIDS in the pre-HAART era (1991-1996) [²¹]. Our data confirm that, in the HAART era, older HIV-infected patients still have delayed access to care and a later diagnosis of HIV infection [²²]. Several non-exclusive factors might explain this situation [^21,23,24]: older patients may be unaware they are at risk of HIV infection; physicians may be less likely to consider HIV infection among patients aged over 50 years; and routine HIV testing is uncommon in older patients. Furthermore, symptoms of some opportunistic infections such as encephalopathy may mimic other diseases associated with ageing and might therefore be responsible for a late diagnosis. Given this late detection of HIV infection, it is likely that older patients have failed to initiate HAART and appropriate prophylaxis in a timely manner. In contrast, we found no evidence of discrimination in terms of HAART prescription to older patients with known HIV infection. On the contrary, older patients tended to be prescribed HAART sooner after FHDH inclusion than younger patients. These findings call for HIV awareness programmes targeting both older patients and their physicians. Clinicians caring for older patients must become more familiar with AIDS in order to diagnose HIV infection sooner so that antiretroviral treatments could be initiated earlier.

We found that older age was associated with a better virological response to HAART. This confirms previous reports that younger age is associated with weaker virological responses and with viral rebound [^6,18,25-27]. Whether this effect is due to poorer adherence or to less favourable pharmacokinetics in younger patients remains to be determined. Adherence is not recorded in the FHDH database. In the Aproco Study (ANRS EP11), younger age was associated with poor adherence [²⁸]. However, in another analysis of the same study, in which both adherence and viral rebound were considered [²⁵], the detrimental effect of young age on the virological response persisted after adjusting for adherence to therapy at month 4, suggesting that it may not be related to poorer adherence. By contrast in another report [²⁹], older age was associated with better adherence but not with virological outcome.

Despite the better virologic response in older subjects, we, like others [^12,15,17], found that the immunologic response to HAART (CD4 cell count increment) was weaker than in younger subjects. Previous studies involving far fewer patients and non-longitudinal methods failed to show such a difference [^10,11,30]. The CD4 cell count increment according to the baseline viral load level was slower in older patients than in younger patients and, as expected, was more rapid during the first 6 months of HAART than subsequently, regardless of age. These findings were not influenced by the baseline CD4 cell count (< 200 or ≥ 200 × 10⁶ cells/l) (Fig. 1).

The slower CD4 cell reconstitution in older patients may be related to an impaired thymic function [¹³]. CD4 reconstitution has been shown to be age dependent [^16,17,31,32]. There are currently no specific clinical guidelines on HIV-infected patients over 50 years of age. Whether this slower immune response in older patients should have implications in the clinical management of elderly patients notably for the timing of HAART initiation, remain to be assessed after evaluating the relative contribution of age-related immune function on one hand, and late diagnosis and treatment on the other.

Older patients had a faster clinical progression to a new ADE or to death, as well as a faster progression to a new ADE regardless of death, than younger patients. The higher incidence of opportunistic infections in older patients was probably related to the poorer immunologic status and response to HAART of these subjects as it could not be related to an impaired virologic response. In our study, three-quarters of clinical events occurred in patients with baseline CD4 cell counts less than 200 × 10⁶ cells/l. When the data were stratified on baseline CD4 cell counts, older patients with baseline counts less than 200 × 10⁶ cells/l had a significantly higher risk of clinical progression compared with younger patients (P = 0.0013). As in the study by Perez et al. [⁹], we found no age-difference among patients with baseline CD4 cell counts of at least 200 × 10⁶ cells/l (P = 0.36), but the analyses were probably a little short of statistical power, owing to the small number of clinical events. Interestingly, the nature of AIDS-defining events during follow-up differed according to the age group, older patients having a significantly higher risk of CMV disease, HIV encephalopathy and Kaposi's sarcoma. In particular, we found that older patients had a five-fold higher risk of developing CMV disease. These CMV diseases occurred at a median CD4 cell count of 60 × 10⁶ cells/l (interquartile range, 20-85), with no difference between the two age groups. Despite initiation of HAART, older patients probably remained at a higher risk of developing these diseases, because of an inadequate CD4 cell response. Indeed, a CD4 count < 75 × 10⁶ cells/l after HAART initiation is associated with a higher risk of CMV infection [³³]. The three-fold higher risk of PML in the older group may also have been due to inadequate CD4 cell responses, as a recent study indicated that T-cell responses play a critical role in the control of JC virus infection, by preventing PML development [³⁴].

In conclusion, patients over 50 years of age do exhibit an immune response after HAART. However, relative to younger patients, their CD4 cells reconstitution is significantly slower, despite a better virological response. This may explain why older patients have a higher risk of clinical progression on HAART. Awareness campaigns targeting patients in this age group, and their practitioners, are needed to reduce the current delay in HIV diagnosis and treatment and thereby, preserving their chance of treatment success.

Acknowledgements

The authors are grateful to all FHDH participants and research assistants, without whom this work would not have been possible.

Sponsorship: FHDH is supported by Agence Nationale de Recherches sur le SIDA (ANRS), Fondation pour la Recherche Médicale, Institut National de la Santé et de la Recherche Médicale (INSERM) and the French Ministry of Health.