Among patients with sustained viral suppression in a resource-limited setting, CD4 gains are continuous although gender-based differences occur.

INTRODUCTION: There is conflicting data on long-term CD4 immune recovery after combination antiretroviral therapy (ART) in resource-limited settings. Virologic suppression is rarely documented in cohorts from sub-Saharan Africa so objective evidence of adherence is biologically unsubstantiated. We sought to investigate long-term patterns of immune recovery in Ugandan patients on ART with sustained viral suppression.
METHODS: A prospective cohort of patients starting ART between April, 2004 and April, 2005 at the Infectious Diseases Institute with sustained viral suppression (viral load ≤ 400 copies/ml at month 6 and 12) while on first-line ART. Propensity scores were used to adjust for treatment allocation (nevirapine or efavirenz) at ART initiation. Data were analyzed using Kaplan Meier methods and cross-sectional time series regression.
RESULTS: Three hundred and fifty-six patients were included in the analysis.71.6% were female, 87% in WHO stage 3 or 4, median age was 37 years, (IQR:32-43), and median CD4 count was 108 cells/µL, (IQR:35-174) at ART start. At multivariable analysis, lower immune recovery (measured by change in CD4 from ART start at each time interval) was associated with male-gender (-59, 95% CI: 90, -28, P<0.001), baseline CD4 count of 101-200 cells/µL (-35, 95% CI: 62, -9, P=0.009) and >200 (-64, 95% CI: 101, -26, P=0.001), and use of AZT at baseline (-47, 95% CI: -74, -20, P=0.001). Median time to reach >400 cells/µL was longer in males (197.4 weeks, IQR:119.9-312.0), compared to females (144.7 weeks, IQR:96.6-219.7, P<0.001). The cumulative probability of attaining CD4 >400 cells/µL over 7 years was higher in females compared to males (P<0.001).
CONCLUSIONS: There was long-term, continuous, immunologic recovery up to 7 years after ART initiation in an urban Ugandan cohort. Virologically suppressed women had better sustained immune recovery than men. Men take longer to immune reconstitute and have a lower probability of reaching a CD4 cell count >400 cells/µL. The biologic mechanisms of these gender differences need further exploration.

Introduction

Potent combination antiretroviral therapy (ART) leads to suppression of the viral load, and immune reconstitution, usually defined by an increase in CD4+ T cell (CD4) count, and consequently, to a reduction in occurrence of AIDS and mortality [1]. However, increases in CD4 counts vary by region [2,3], gender with a higher increase in CD4 count in women [3,4], the duration of antiretroviral therapy [3,5,6] and on the co-existence of other infections, particularly tuberculosis [7]. The greatest increases in CD4 count occur during the first year of ART initiation [4,5,8,9,10], especially in patients with low baseline CD4 count [11]. However patients with low CD4 count are more likely to take longer to return to normal CD4 count levels [5], and may remain at increased risk of opportunistic infection hence morbidity [12]. Two studies in resource-rich countries with a minimum follow-up time of 6 years have shown a significant increase in CD4 count among patients who initiated ART at lower CD4 count level compared to those who started ART at high CD4 count [4,9]. However some patients tend to have a suboptimal and lower rate of immune recovery even if they achieve viral suppression [13] and they remain at increased risk for opportunistic infections and AIDS-related mortality [2].

A few studies have been done in the area of immune recovery in the Sub-Saharan Africa (SSA) but there are no specific data for patients with sustained viral suppression [3,6] since viral load testing is not routinely performed due to the high cost [14]. We sought to investigate long-term patterns of immune reconstitution after ART for up to 7 years, and factors associated with higher CD4 count increase in patients with documented sustained viral suppression.

Methods

Study setting and population

The Infectious Diseases Institute (IDI) of Makerere University is a center of excellence which delivers HIV treatment and care to over 10,000active HIV patients with over 8,000 patients on antiretroviral therapy. Nested within the clinic at IDI is a prospective cohort study of 559 patients who started ART between April 2004 and April 2005 and have been followed for over 8 years. Patients were started on stavudine, lamivudine and nevirapine (provided by Global Fund) or zidovudine plus lamivudine plus efavirenz (provided by the US President’s Emergency Plan for AIDS Relief) according to the WHO [15], and national guidelines [16]. Details of this cohort have been previously described in detail [17,18]. Briefly, the study participants have clinic visits every 3 months. Laboratory tests are performed every 6 months, which include CD4 count by FACSCalibur (Becton Dickson, Mountain View, California, USA) and HIV-1 viral loads (Amplicor HIV-1 Monitor PCR Test, version 1.5, Roche Diagnostics, GmbH Molecular Systems, Pleasanton, California, USA, with a lower limit of detection of 400 copies/ml).

This analysis included all patients who had viral load measurements <400 copies/ml at month 6 and 12 and who were taking first-line ART for up to 7years of follow-up; patients were included in the analysis for only the time that they had sustained viral suppression defined as viral load ≤400 copies/μL while still on first-line treatment. Patients were categorized in 3 groups depending on the CD4 count at ART start: 1) ≤100 cells/µL, 2) 101-200 cells/µL, and 3) >200 cells/µL.

Ethics statement

This study was approved by the institutional review board of Makerere University College of Health Sciences and the Uganda National Council for Science and Technology (MV853). All patients were counseled and provided a written informed consent.

Statistical analysis

We calculated the increase in CD4 count from baseline to all subsequent 6-monthly CD4 count readings. To obtain unadjusted estimates for the association between change in CD4 count from baseline and each predictor, we used cross-sectional time series regression while controlling for time interval when the laboratory tests were done. Factors considered in the analysis were gender, age, baseline viral load (<5 log10copies/ml and ≥5 log10copies/ml), and WHO staging (stage 1/2 and stage 3/4). Fortime-updated covariates, we considered time intervals of the CD4 count measurement, weighted estimates of baseline ART regimen, and hemoglobin in the analysis. Predictors with a P-value ≤0.2 in the unadjusted analysis and those of clinical significance were included in the multivariable analysis. A multivariable time series regression model was used to determine factors associated with CD4 count increases in these patients. We calculated the time to reach and the probability of achieving a CD4 count >400 cells/µL, which is the lower normal limit of CD4 count in Ugandan adults [19] using the Kaplan Meier method. Curves were compared using the log rank test. We also compared the proportions of mortality and opportunistic infections that occurred during the study period in the 3groups using the chi-square test. Since baseline ART treatment was not randomly assigned, we accounted for the selection bias using the predicted probability of treatment allocation (stavudine plus lamivudine plus nevirapine, and zidovudine plus lamivudine plus efavirenz), and generated propensity scores.

Another sub-analysis was done to study the differences in immunological recovery between male and female gender. We compared the mean age between males and females using a Student’s T-test. We used a chi-square test for trend to analyze for differences in baseline BMI (underweight, normal, and overweight), and baseline CD4 count groups categorized by gender. Chi-square test was used to determine the association between gender and other baseline variables (i.e. regimen, WHO stage, hemoglobin, and HIV RNA viral load). We calculated the time to reach and the probability of achieving a CD4 count >400 cells/μL stratified by gender using the Kaplan Meier method. Wilcoxon rank sum test was used to compare median time, and curves were compared using the log rank test. A P<0.05 was considered statistically significant.

Data management and statistical analyses were performed using Stata (StataCorp. STATA 11.1, College Station, Texas 77845 USA)

Results

Three hundred fifty-six patients out of the 559 enrolled, were eligible for our study. We excluded202 patients by month 12 (98 had viral load >400copies/ml month 6 and/or 12,79 died, 20 got lost to follow up, 4 were transferred out, and one withdrew consent before reaching 12 months), and one was excluded because of lack of a viral load result at month 6 [20]. The treatment outcomes of our cohort at 6 months of follow up have been described elsewhere [18] (Figure 1).

Figure 1

Study flow diagram.

Of the remaining 356 patients 71.6% were female, 87% had WHO stage 3 or 4, the median age at ART start was 37 years, (IQR: 32-43), and the medianCD4 count at ART start was 108 cells/µL, (IQR 35-174); 72% were started on stavudine plus lamivudine plus nevirapine, and 28% on zidovudine plus lamivudine plus efavirenz. The baseline characteristics of the patients included in the analysis were similar to those of the patients excluded (data not shown). Of the patients included 164 (46%) started ART with a CD4 count ≤100 cells/µL, 133 (37%) with a CD4 count 101-200 cells/µL, and 59 (17%) with a CD4 count >200 cells/µL. Gender distribution, hemoglobin, and ART regimen at start were similar across the three groups except for viral load which was <5 log10copies/ml in a higher proportion of patients with a CD4 count >200 cells/µL at ART start, and median age which was higher in the same group. Although the difference was only borderline significant, there were more patients in WHO stage 3 and 4 in the group of patients who started ART with a CD4 count ≤100 cells/µL (Table 1).

Table 1

Characteristics of the study participants categorized by CD4 counts at ART initiation (N=356).

Variable	<=100	101-200	>200	P-value
Gender: Male, n(%)	52 (31.7)	35 (26.3)	14 (23.7)	0.406
Age in years: median(IQR)	34 (29-40)	36 (32-41)	37 (31-45)	0.047
ART regimen:n(%)
Stavudine/lamivudine/nevirapine	118 (72.0)	93 (69.9)	46 (78.0)
Zidovudine/lamivudine/efavirenz	46 (28.0)	40 (30.1)	13 (22.0)	0.516
WHO stage:n(%)
Stage I-II	13 (7.9)	21 (15.8)	10 (17.0)
Stage III-IV	151 (92.1)	112 (84.2)	49 (83.1)	0.062
Hemoglobin:n(%)
≤8g/dL	7 (4.3)	4 (3.0)	2 (3.4)
>8 g/dL	157 (95.7)	129 (97.0)	57 (96.6)	0.841
HIV RNA Viral load:n(%)
<5 log10copies/ml	27 (16.5)	32 (24.1)	27 (45.8)
≥5 log10copies/ml	137 (83.5)	101 (75.9)	32 (54.2)	<0.001

Note: ART: Antiretroviral therapy; WHO: World Health Organization; n = number, IQR = interquartile range

Clinical outcomes

Twenty-four patients died during follow-up with a mortality rate that was similar for both males 5 (5%) and females 19 (7.5%) (P=0.396). Mortality was higher in patients who had a baseline CD4 count ≤100 cells/µL, 13 (7.9%) compared to patients who had a baseline CD4 count 101-200 cells/µL 7 (5.3%) (P=0.0001), and >200 cells/µL 4 (6.8%) (P=0.0001). Sixteen patients were lost to follow-up (n=7) or transferred (n=9) with no observed differences by gender. Only one woman withdrew her consent. The overall median time of follow up was 312 weeks (192–336), and was similar across the CD4 categories. One hundred and thirty-six patients(38.2%) were censored during follow up because they had a viral load measurement >400 copies/ml. The mean (SD) follow up time among those who were censored was 185 weeks(7 weeks) and this was similar across gender(P=0.285).

Immunological outcomes

Overall, the median CD4 count increased from 108 cells/µL (IQR: 35-174) at baseline to 534 cells/µL (IQR: 389, 717) at 7 years (P<0.001). The median increase (IQR) during follow-up among those who had ≤100 cells/µL at ART start was 478 cells/µL (IQR: 310, 606) and was significantly higher compared to those with baseline CD4 count of 101-200 cells/µL, (390 cells/µL (IQR:304, 591) (P=0.044)), and to those with >200 cells/µL (343 cells/µL (IQR: 118, 496) (P=0.008). During the first year of ART, the increase in CD4 count was higher in patients with baseline CD4 count ≤100 cells/µL (136 cells/µL, IQR: 80, 211) compared to those with baseline CD4 count of 101-200 cells/µL, (119 cells/µL, IQR:37, 205, P=0.036), and to those who started ART with >200 cells/µL (47 cells/µL, IQR: 0, 122, P<0.001). Annual median CD4 count trajectories are shown in detail in Figure 2. Overall, the trajectories of the median CD4 count continued to rise significantly in all the groups up to 7years after ART initiation (Figure 2).

Figure 2

Median CD4 count over time stratified by baseline CD4 count.

The full line with diamonds represents people with baseline CD4 count ≤100 cells/µL.

The long dash line with ‘x’ s represents people with baseline CD4 count 101-200 cells/µL.

The long dash-dotted line with triangles represents people with baseline CD4 count >200 cells/µL.

As expected the probability of attaining a CD4 >400 cells/µL across the 3 groups was lower in patients who started ART at CD4 count ≤100 cells/µL (87.4%,95% CI78.4,93.9) compared to those started at 101-200 cells/µL (94.4%, 95% CI85.3,98.7, P=0.018), and >200 cells/µL (88.9, 95% CI75.0,97.0, P<0.001) (Figure 3). The median time to reach >400 cells/µL was longer in patients with baseline CD4 count of ≤100 cells/µL (1,183 days, IQR 839, 1,531), compared with patients with CD4 count of 101-200 cells/µL (939 days, IQR 673, 1354), and >200 cells/µL (672 days, IQR: 506, 710, P<0.001).

Figure 3

Probability of achieving CD4 count >400 cells/µL and number at risk by baseline CD4 count.

The full line represents people with baseline CD4 count ≤100 cells/µL.

The long dash line represents people with baseline CD4 count 101-200 cells/µL.

The long dash-dotted line represents people with baseline CD4 count >200 cells/µL.

Risk factors for lower immune recovery

The bivariate analysis (Table 2) showed that, age, male gender, and baseline CD4 count >200 and 101-200 cells/µL compared to ≤100 cells/µL were associated with an annual CD4 count reduction of-12, -48, -41, and -82cells/μL respectively. Patients with a baseline viral load of ≥5 log10copies/ml had higher rates of immune recovery compared to patients with < 5 log10copies/ml at ART start.

Table 2

Cross-sectional time series regression model of factors associated with lower immune recovery (N=356).

Variable	Unadjusted model				Adjusted model
	Change in CD4 count	95% C.I.	P-value		Change in CD4 count	95% C.I.	P-value
Baseline :
Age (5-year increase)	-12	-18, -4	0.003		-5	-12,3	0.206
Gender, Male	-48	-77, -19	0.001		-59	-90, -28	<0.001
CD4 count: cells/μL
≤ 100	Reference				Reference
101-200	-41	-69, -13	0.004		-36	-63, -9	0.009
>200	-82	-119, -45	<0.001		-64	-102, -26	0.001
HIV RNA Viral load
<5 log10copies/ml	Reference				Reference
≥5 log10copies/ml	52	21,83	0.001		46	17, 76	0.002
Allocation of NNTRI based regimen[**]	65	-179, 309	0.600		162	-103, 427	0.232
AZT based ART regimen					-47	-74, -20	0.001
Time updated :
Follow-up time (years)	52	50, 55	<0.001		53	50, 55	<0.001
Hemoglobin: >8g/dl	44	-10,98	0.113		41	-13, 95	0.134

NNTRI based regimen (efavirenz or nevirapine) was analyzed using propensity scores

Note: WHO stage, hemoglobin at ART initiation, Body Mass Index (BMI) and opportunistic infections were not associated with immune recovery at bivariate and multivariable analysis.

CI = confidence interval, n = number, ART = antiretroviral therapy, NNRTI = non-nucleoside reverse transcriptase inhibitor, AZT = zidovudine

In the multivariable analysis, factors associated with lower yearly immune recovery were: CD4 count of 101-200 cells/µL at ART start (-36 cells/µL,95%CI: 63, -9, P=0.009) and CD4 count >200 cells/µL (-64cells//μL,95%CI:-102, -26, P=0.001) compared to those with ≤100 cells/µL at ART start), and male gender (-59 cells/µL, 95%CI: 90, -28, P<0.001). Baseline viral load ≥5 log10copies/ml was also associated with a higher immune recovery (46 cells/µL,95%CI:17,76, P=0.002), per year.

Immunologic recovery by gender

At baseline, males and females had similar characteristics except that the males were older than women (38 years, IQR:35, 42 versus 34 IQR:29, 40; P<0.001) (Table 3). Other characteristics were similar including the median CD4 count at ART initiation within the CD4 count strata: median time to reach >400 cells/µL was longer in males (197weeks, IQR:120-312) compared to females (145, IQR:97, 220) (P<0.001). The cumulative probability of attaining CD4 >400 cells/µL during 7year follow-up was also higher among females (92.1%, 95% CI: 90.7, 93.3) compared to males (63.9% (95%CI: 60.5-67.3) (P<0.001) (Figure 4). The women had also a higher immune recovery compared to men across all the CD4 count strata (Figure 5).

Table 3

Characteristics of study participants stratified by gender at ART initiation.

Variable	Female n=255	Male n=101	Total n=356	P-value
Age in years; median(IQR)	34 (29-40)	38 (35-42)	35 (30-42)	<0.001*
Body mass Index; Kg/m2				0.113***
<18.5	69 (27.0)	35 (34.7)	104 (29.2)
18.5-24.9	160 (62.8)	59 (58.4)	219 (61.5)
≥25	26 (10.2)	7 (6.9)	33 (9.3)
CD4 cell count (cells/μL);n(%)				0.181***
≤100	112 (43.9)	52 (51.4)	164 (46.0)
101-200	98 (38.4)	35 (34.7)	133 (37.4)
>200	45 (17.7)	14 (13.9)	59 (16.6)
Regimen:n(%)				0.305**
Nevirapine-based	188 (73.7)	69 (68.3)	257 (72.2)
Efavirenz-based	67 (26.3)	32 (31.7)	99 (27.8)
WHO stage:n(%)				0.05**
1/2	37 (14.5)	7 (6.9)	44 (12.4)
3/4	218 (85.5)	94 (93.1)	312 (87.6)
Hemoglobin (g/dl);n(%)				0.411**
≤8	8 (3.1)	5 (5.0)	13 (3.7)
>8	247 (96.9)	96 (95.0)	343 (96.3)
HIV RNA log10copies/ml;n(%)				0.227**
<5	66 (25.9)	20 (19.8)	86 (24.2)
≥5	189 (74.1)	81 (80.2)	270 (75.8)

* p-value from t-test; ** p-value for chi-square test; ***p-value for chi-square test for trend

IQR = interquartile range, n = number

Figure 4

Probability of achieving CD4 count normalization (i.e. CD4 >400 cells/µL) and number at risk by gender.

The full line represents women.

The long dash line represents men.

Figure 5

Median CD4 count trajectory stratified by gender within baseline CD4 count groups.

The full line represents women.

The long dash line represents men.

Discussion

The CD4 count of patients with sustained viral suppression enrolled in a prospective cohort in Uganda continues to rise over 7years of follow-up, similar to other analyses from cohorts in Europe and North America with a follow up of ≥ 6 years [11,21]. In a UK cohort with a relatively high proportion(2056/7069, 29.1%) of Black Africans, CD4 counts continued to rise in the 8years of follow-up among patients with viral load consistently <1000copies/ml without attaining a plateau regardless of CD4 count at initiation [22]. Our study findings were also similar to other sub-Saharan African cohorts in Malawi, Uganda, and Kenya where patients had continued CD4 gains up to 6 years on ART, although the authors did not limit their analysis to patients with sustained viral suppression and patients had generally higher baseline CD4 counts compared to patients in our cohort [23]. In contrast, some studies have shown an initial increase in CD4 count and a leveling off after 4years of follow-up [9,24]. The level of CD4 counts in patients on ART reached a plateau, especially in those with higher CD4 cell counts at baseline [4,25,26].

In our cohort, the magnitude of the increase in CD4 count after every measurement generally becomes smaller (i.e. CD4 count increases at a decreasing rate) and is most consistent with an asymptote rather than a plateau. This has also been described in other studies [21,27] were patients with higher change in CD4 counts (>200 cells/µL) from ART to 2 years’ experience lower increases after 2years on ART. The increased immune recovery seen in the early stages of ART initiation is mainly due to the decreasing viral load levels of the patients [28,29]. Some authors have suggested that the initial trajectory of CD4 count after ART initiation is due to redistribution of memory CD4 cells [30] from lymphoid tissues towards the blood compartment. This trajectory evolves into an asymptote after long-term ART because of an impaired activity in the bone marrow [31] and/or reduced thymic activity.

Consistent with other studies, patients who started ART at higher CD4 counts >200 cells/µL were more likely to achieve a CD4 count >400 cells/µL [4,9,10,23,25]. This suggests that despite the higher immune recovery seen immediately after initiation of ART among patients with lower CD4 count thresholds, thereis an increased risk of the hazards associated with patients with lower baseline CD4 counts since they remain at a CD4 count <400 cells/μL for a longer time period [32,33,34]. These results have been reported by other studies [12,35]. A possible explanation could be that while CD4 counts remain low, the bone marrow generates IL-7 which is responsible for the increased stimulation of the production of T-cells as well as their proliferation and survival [12], which is increased in lymphopenic conditions [36].

Contrary to other studies [4,9,23,37,38,39], in our analysis older age at initiation was not associated with lower immune recovery which has been attributed to decreasing thymic reserves [40,41]. Notwithstanding the differences in the definition of viral suppression, our findings were consistent with another study in Senegal with median age 37 years that did not demonstrate an effect of age on immune reconstitution [27].

Male gender was associated with lower immune recovery compared to the female gender. Previous studies have also attributed the difference to higher physiological CD4 count levels in females compared to males [23,42,43] and lower CD4 counts at ART start in males [44]. In a multi-center analysis of patients on ART from 27 sites largely in resource limited settings, slower immune recovery was found in men compared to women [6]. It has been suggested that gender differences could be as a result of other unmeasured confounders including hormones. Studies in resource rich settings have shown that female hormones may stimulate increased production of CD4 T cells compared to male hormones [45,46]. Another explanation may be that the thymic output is higher in women than in men; however, studies to definitely show this are lacking [35].

In conclusion, patients who start ART at lower CD4 count have the greatest CD4 increase in the first years compared to those with higher CD4 counts, however, these patients take longer and are less likely to reach CD4 count normalization. Females have a better immune response and are more likely to achieve earlier CD4 count normalization compared to the males. This study has shown that patients with good adherence to ART as demonstrated by viral suppression have the potential to immune reconstitute beyond normalization, thereby reducing the chances of opportunistic infections and death.