Characteristics and determinants of T-cell phenotype normalization in HIV-1-infected individuals receiving long-term antiretroviral therapy


  • The data were originally presented in part at the 6th International AIDS Society Conference on HIV Pathogenesis, Treatment and Prevention, July 2011, Rome, Italy (Abstract MOPE203).



Although combination antiretroviral therapy (cART) can restore CD4 T-cell numbers in HIV infection, alterations in T-cell regulation and homeostasis persist. We assessed the incidence and predictors of reversing these alterations with cART.


ART-naïve adults (n = 4459) followed within the Canadian Observational Cohort and exhibiting an abnormal T-cell phenotype (TCP) prior to cART initiation were studied. Abnormal TCP was defined as having (1) a low CD4 T-cell count (< 532 cells/μL), (2) lost T-cell homeostasis (CD3 < 65% or > 85%) or (3) CD4:CD8 ratio dysregulation (ratio < 1.2). To thoroughly evaluate the TCP, CD4 and CD8 T-cell percentages and absolute counts were also analysed for a median duration of 3.14 years [interquartile range (IQR) 1.48–5.47 years]. Predictors of TCP normalization were assessed using adjusted Cox proportional hazards models.


At baseline, 96% of pateints had CD4 depletion, 32% had lost homeostasis and 99% exhibited ratio dysregulation. With treatment, a third of patients had normalized CD4 T-cell counts, but only 85 individuals (2%) had normalized their TCP. In a multivariable model adjusted for age, measurement frequency and baseline regimen, higher baseline CD4 T-cell counts and time-dependent viral suppression independently predicted TCP normalization [hazard ratio (HR) for baseline CD4 T-cell count = 1.42 (1.31–1.54) per 100 cells/μL increase; P ≤ 0.0001; HR for time-dependent suppressed viral load = 3.69 (1.58–8.61); P-value ≤ 0.01].


Despite effective cART, complete TCP recovery occurred in very few individuals and was associated with baseline CD4 T-cell count and viral load suppression. HIV-induced alterations of the TCP are incompletely reversed by long-term ART.


Homeostatic regulation occurs in all living organisms and is critical in keeping many biological parameters within a physiological range [1, 2]. The immune system of healthy individuals is characterized by the maintenance of T-cell homeostasis and a balanced T-cell phenotype (TCP). This is achieved through complex and tightly regulated processes such as thymic output, access to cytokines, naïve T-cell differentiation into memory cells, and antigen-independent peripheral T-cell proliferation [3-5]. Although progressive CD4 T-cell depletion is the hallmark of HIV disease, impaired T-cell homeostasis and profound immune dysregulation are often observed during the course of the infection.

In the early 1980s, prior to the identification of HIV, the earliest diagnostic marker of AIDS was a reversion of the CD4:CD8 ratio [6, 7]. The imbalance of these T-cell subsets defines HIV-mediated immune dysregulation. It begins in early disease preceding the progressive loss of CD4 T-cells and deteriorates during untreated infection. Furthermore, altered T-cell homeostasis occurs mostly in the late stage of HIV disease and is manifested by a failure to maintain physiologically normal levels of T cells [8]. All T cells have a CD3 phenotype and characteristically express either CD4 or CD8 cell surface molecules. T-cell homeostasis was first described in 1993, as the normal physiological state in which the human body maintains a constant number of circulating (CD3) T-cells irrespective of fluctuations within the CD4 and CD8 T-cell compartments [9]. Maintaining the size and diversity of the peripheral T-cell pool is of crucial importance for a healthy and balanced immune system [10, 11]. Low CD4 T-cell counts, a low CD4:CD8 ratio and loss of T-cell homeostasis are part of a continuum of immune abnormalities that occur during progressive HIV infection [12].

The use of potent combination antiretroviral therapy (cART) has resulted in sustainable reductions of HIV viral load to undetectable levels and in improved CD4 T-cell counts [13]. However, cART-mediated CD4 T-cell restoration may not accurately reflect complete recovery of the immune phenotype as T-cell dysregulation and altered T-cell homeostasis may persist.

In non-HIV immune-based disease states such as rheumatoid arthritis, Crohn's disease, systemic lupus erythematosus (SLE) and Sjogren's syndrome, impaired T-cell homeostasis has been linked to several deleterious clinical outcomes [11, 14]. Burn injury is associated with altered T-cell homeostasis and subsequent decreased resistance to infections [15]. Patients with active pulmonary tuberculosis also have loss of T-cell homeostasis and a decreased CD4:CD8 T-cell ratio [16]. Individuals with common variable immune deficiency (CVID) have persistently low CD4:CD8 T-cell ratios that are associated with poor clinical outcomes [17].

Most notably, immune dysregulation has been associated with accelerated immune senescence (the gradual decline in immune function occurring with age) and with increased risks of non-AIDS-related comorbidities such as coronary artery diseases, diabetes and liver impairment [18]. A low CD4:CD8 T-cell ratio is an important component of the immune risk phenotype (IRP); it has been found to be associated with increased morbidity and mortality in seronegative individuals over the age of 60 years [19, 20]. In HIV-infected individuals, the potential risks for comorbidities from failure to restore T-cell homeostasis and to normalize the CD4:CD8 T-cell ratio are currently unknown.

Despite previous findings that altered T-cell homeostasis has predictive value in determining impending AIDS and that CD4:CD8 ratio dysregulation correlates with higher risks of developing coronary disease, few, if any, studies have assessed the effect of long-term successful cART on altered T-cell homeostasis and T-cell ratio dysregulation [8, 21, 22]. Thus, the degree to which these HIV-mediated immune alterations can be reversed by effective cART remains to be elucidated. We therefore assessed the incidence and predictors of complete TCP normalization in antiretroviral-naïve HIV-positive patients initiating cART.


Cohort description

The Canadian Observational Cohort (CANOC) collaboration is an observational cohort study of antiretroviral-naïve HIV-positive patients initiating cART on or after 1 January 2000 [23]. CANOC participants represent nearly a quarter of Canadians currently on cART after this date. This collaboration is open to all Canadian HIV treatment cohorts with more than 100 eligible patients and currently includes eight participating cohorts across Canada. Eligibility criteria for inclusion in CANOC include: documented HIV infection, residence in Canada, age 18 years and older, initiation of a first antiretroviral regimen comprised of at least three individual agents, and at least one measurement of HIV-1 RNA viral load and CD4 T-cell count within 6 months of initiating cART. Patient selection and data extraction are performed locally at the data centres of the participating cohort studies. Nonnominal data from each cohort on a predefined set of demographic, laboratory and clinical variables are then pooled at the Project Data Centre in Vancouver, British Columbia. The last date of follow-up in the cohort for the current analysis was 22 August 2010. All participating cohorts have received approval from their institutional ethics boards to contribute nonnominal patient-specific data.

Ethical consideration

The human subjects activities of CANOC were approved by the Simon Fraser University Research Ethics Board, the University of British Columbia Research Ethics Board and the following local institutional review boards of the participating cohorts: Providence Health Care Research Institute Office of Research Services, The Ottawa Hospital Research Ethics Board, University Health Network (UHN) Research Ethics Board, Véritas Institutional Review Board (IRB), Biomedical C (BMC) Research Ethics Board of the McGill University Heath Centre (MUHC), University of Toronto HIV Research Ethics Board (HIV REB), and Women's College Hospital Research Ethics Board. Local cohort studies have obtained written consent except for the following: HAART Observational Medical Evaluation and Research (IRB approves the retrospective use of anonymous administrative data without requiring consent; an information sheet for participants is provided in lieu of a consent form); Ottawa Hospital Cohort (IRB approves the anonymous use of data retrospectively abstracted from clinical care databases without requiring consent); UHN (REB approves the anonymous use of data retrospectively abstracted from clinical care databases without requiring consent); MUHC (IRB approves the anonymous use of data retrospectively abstracted from clinical care databases without requiring consent; patients sign a general waiver on opening a medical chart at the hospital but no specific study related consent); Maple Leaf Medical Clinic (REB has approved the anonymous use of data retrospectively abstracted from clinical care databases without requiring consent); and Effective Anti-Retroviral Therapy cohort (REB approves the anonymous use of data retrospectively abstracted from clinical care databases without requiring consent; patients sign a general waiver on opening a medical chart at the hospital but no specific study related consent).


Participants included in the current analysis had an altered TCP at baseline. Participants were eligible for analysis if (a) they came from sites able to provide electronic data for all of the following immunological markers: CD4 and CD8 T-cell counts and percentages, CD3 T-cell percentages and CD4:CD8 T-cell ratios, (b) they had at least one record with all six markers within 2 years prior to starting cART, and (c) they had at least two follow-up records with all six markers > 30 days apart following the initiation of treatment. Note that inclusion criteria for enrolment in CANOC require that each individual have CD4 T-cell counts within 6 months of starting cART. Therefore, all 4459 individuals included in the study had CD4 T-cell measurements within 6 months of treatment initiation; of those, 4359 (97.8%) had all three baseline T-cell measurements (CD4, CD8 and CD3) within 6 months of starting therapy. Only 100 individuals (2.2%) had complete TCP values between 6 months and 2 years. Thus, our 2-year pre-cART window does not introduce any significant bias in terms of baseline evaluation. The data for all patients meeting eligibility criteria at participating sites were included in CANOC.

Primary outcome

HIV-induced altered TCP involves one or more of: (1) low CD4 T-cell count (< 532 cells/μL), (2) lost T-cell homeostasis [low (< 65%) or high (> 85%) CD3 T-cell percentage] and (3) ratio dysregulation (CD4:CD8 ratio < 1.2). For patients without CD3 T-cell percentage measurements, CD3 T-cell percentage was calculated as the sum of CD4 and CD8 T-cell percentages. Considering that CD4 and CD8 T-cell percentages and absolute counts are the main components of total T-cell levels and CD4:CD8 T-cell ratio, we also assessed each of these markers in order to ensure a thorough evaluation of the TCP.

The primary outcome of interest was the achievement of a healthy TCP on at least two sequential visits at least 30 days apart. TCP recovery was defined as meeting all six of the following criteria: CD3 T-cell percentage 65−85%, CD4:CD8 ratio 1.2−3.3, CD4 T-cell count 532−1170 cells/μL, CD4 T-cell percentage 39−55%, CD8 T-cell count 236−651 cells/μL and CD8 T-cell percentage 18−31%. These values were derived from 124 healthy controls (62 male and 62 female) recruited at the Montreal General Hospital. These controls had a median age of 39 years [interquartile range (IQR) 32–47 years], were all HIV-negative and received thorough physical examinations. Furthermore, these individuals were screened for the presence of primary immune deficiencies and autoimmune diseases.

Patients who failed to maintain at least one of these parameters within its physiological range were considered to have an altered TCP.

Statistical methods

Demographic and clinical characteristics at baseline are summarized using medians and IQRs for continuous variables and frequencies and proportions for categorical variables. Baseline values were defined as the closest values within 2 years of initiating cART. Duration of the follow-up period was measured from the time of cART initiation. Time to normalization of TCP was assessed using the Kaplan−Meier survival method. Predictors of TCP normalization were assessed using univariate and multivariable Cox proportional hazards models. The assumption of proportional hazards was checked and met for each covariate. Based on previous studies analysing the effect of HIV infection on the immune system, sociodemographic and clinical covariates potentially associated with immune dysregulation were considered for inclusion in the analyses [24, 25]. As there were few events, we were conservative with regard to the number of variables included in the model. Because of the great deal of missing data for variables such as ethnicity, HIV risk factors and hepatitis C virus (HCV) coinfection, and because these factors were neither associated with the primary outcome in an adjusted model nor changed the inference of the other covariates under consideration, we did not include these variables in the final multivariable model.

All analyses were performed using sas software version 9.3 (SAS Institute, Cary, NC).


Among the 6673 initially ART-naïve HIV-positive individuals followed within CANOC, 2214 were excluded from the analysis. Of these, 1071 came from two sites that did not have electronic records of T-cell percentages available. Information on excluded participants and reason for exclusion is given in Figure 1. A total of 4459 patients met the inclusion criteria for this study. These individuals were studied for a median duration of 3.14 years (IQR 1.48–5.47 years), with the median year of cART initiation being 2005 (IQR 2002–2007). Most individuals were on a nonnucleoside reverse transcriptase inhibitor (NNRTI)-based or protease inhibitor (PI)-based regimen at the time of treatment initiation.

Figure 1.

Flow chart of inclusion criteria, giving information on excluded participants and reason for exclusion. Among the 6673 initially antiretroviral-naïve HIV-positive individuals followed within The Canadian Observational Cohort (CANOC), 4459 patients met the inclusion criteria for this study. ARV, antiretroviral; TCP, T-cell phenotype.

The demographic and clinical baseline characteristics of the study population are summarized in Table 1. At baseline, 96% of the patients had CD4 T-cell counts < 532 cells/μL, 32% had CD3 T-cell percentages outside the homeostatic range (65–85% of circulating lymphocytes) and 99% had a CD4:CD8 ratio < 1.2. Among individuals with low CD4 T-cell counts, 53% had CD4 T-cell counts < 200 cells/μL, which connotes severe HIV disease [26]. Six hundred and five individuals exhibited AIDS-defining illnesses at baseline.

Table 1. Demographic and clinical baseline characteristics for 4459 HIV-positive patients on combination antiretroviral therapy
VariableIncluded (n = 4459)
  1. Values are number and percentage, or median and interquartile range.
  2. ARV, antiretroviral; NNRTI, nonnucleoside reverse transcriptase inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; PI, protease inhibitor.
British Columbia2066 (46%)
Ontario1727 (39%)
Quebec666 (15%)
Age at first ARV treatment (years)40.0 (34.0–46.7)
Male3668 (82%)
Female790 (18%)
Caucasian1368 (31%)
Black231 (5%)
Aboriginal208 (5%)
Mixed136 (3%)
Other224 (5%)
Unknown/missing2292 (51%)
HIV risk factor 
Men who have sex with men1501 (34%)
Injecting drug users935 (21%)
From endemic country365 (8%)
Unknown/missing1471 (33%)
Year of first ARV treatment2005 (2002–2007)
Baseline regimen 
NNRTI-based2022 (45%)
Boosted PI-based2006 (45%)
Single PI-based352 (8%)
NRTI only79 (2%)
Baseline CD4 T-cell count (cells/μL)190 (105–280)
< 200 cells/μL2275 (51%)
200–349 cells/μL1505 (34%)
350–531 cells/μL492 (11%)
532–1170 cells/μL (within normal range)179 (4%)
> 1170 cells/μL8 (0%)
Baseline CD8 T-cell count (cells/μL)764 (500–1130)
< 236 cells/μL255 (6%)
236–651 cells/μL (within normal range)1504 (34%)
652–1200 cells/μL1746 (39%)
> 1200 cells/μL954 (21%)
Baseline CD3 T-cell percentage76 (69–83)
< 50%167 (4%)
50–64%549 (12%)
65–85% (within normal range)3040 (68%)
> 85%703 (16%)
Baseline CD4:CD8 ratio0.22 (0.14-0.36)
< 0.53874 (87%)
0.5-1.1552 (12%)
1.2–3.3 (within normal range)32 (1%)
> 3.31 (0%)
Baseline viral load (log10 copies/mL)4.9 (4.4–5.0)
Presence of AIDS-defining illness at baseline605 (14%)
Hepatitis B virus coinfection 
Yes250 (6%)
No1233 (28%)
Unknown/missing2976 (66%)
Hepatitis C virus coinfection 
Yes965 (22%)
No2413 (54%)
Unknown/missing1081 (24%)

Following treatment initiation, 68% of patients achieved normal T-cell homeostasis, 6.6% a balanced CD4:CD8 ratio and 30% a normal CD4 T-cell count throughout the course of their follow-up. Of the individuals who did not normalize their CD4:CD8 ratio (93%) during the study period, 96% had elevated CD8 T-cell percentages and 68% had elevated CD8 T-cell counts.

Only 85 individuals (2%) reached the primary endpoint of normalizing all the components of the TCP during the follow-up period. The probability of normalizing the complete TCP after 5 years of treatment was 0.08 (0.03−0.14) for people who initiated therapy with CD4 T-cell counts within the normal range and 0.008 (0.003−0.012) percent for those who initiated therapy at CD4 T-cell counts < 200 cells/μL.

Figure 2 shows Kaplan−Meier (KM) curves comparing time to TCP normalization according to the degree of immune alteration at baseline. Panels (a), (b) and (c) display the KM curves for time to TCP normalization by baseline CD4 T-cell count, baseline CD4:CD8 ratio and baseline CD3 T-cell percentage, respectively. In all three panels, time to TCP normalization was shorter among individuals who maintained their immune parameters within physiological ranges at baseline. Note that, for the purpose of the survival analysis, the CD4 T-cell categories 532–1170 and > 1170 cells/μL were collapsed because of the very small number of patients with baseline CD4 T-cell counts > 1170 cells/μL (n = 8).

Figure 2.

Time to T-cell phenotype normalization, showing Kaplan−Meier survival curves for time to T-cell phenotype normalization by degree of immune alteration. CD4 T-cell counts, CD4:CD8 ratio and CD3 T-cell percentages were stratified according to levels of dysregulation vs. the normal physiological range. (a) Kaplan−Meier plot of time to T-cell immunophenotype profile normalization by baseline CD4 T-cell count (the CD4 T-cell count group 532–1170 cells/μL includes eight individuals with a CD4 T-cell count > 1170 cells/μL). (b) Kaplan−Meier plot of time to T-cell immunophenotype profile normalization by CD4:CD8 ratio (the CD4:CD8 ratio group 1.2–3.3 includes one individual with a CD4:CD8 ratio > 3.3). (c) Kaplan−Meier plot of time to T-cell immunophenotype profile normalization by CD3 T-cell percentage.

Table 2 shows the hazard ratios (HRs) of TCP normalization associated with covariates of interest from univariate proportional hazards models. Participants with a high CD8 T-cell percentage at baseline and those that were injecting drug users (IDUs) were less likely to normalize TCP. Conversely, participants with a high CD4:CD8 ratio at baseline and those with sustained viral suppression were more likely to normalize their TCP. In a multivariable proportional hazards model adjusted for age, rate of measurement of immune markers and baseline regimen, higher baseline CD4 T-cell counts and HIV viral load suppression were associated with increased likelihood of TCP normalization [HR = 1.42 per 100 cells/μL increase in baseline CD4 T-cell count; 95% confidence interval (CI) 1.31, 1.54; P ≤ 0.0001; and HR = 3.69; 95% CI 1.58, 8.61; P ≤ 0.01, respectively] (Table 3).

Table 2. Univariate analysis of time to immunophenotype normalization
VariablenHazard ratio95% confidence intervalP-value
  1. ARV, antiretroviral; NNRTI, nonnucleoside reverse transcriptase inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; PI, protease inhibitor; VL, viral load.
Demographic parameters    
Age (continuous per 10 years)44591.26(1.02, 1.56)0.03
Female 1
Male 1.59(0.82, 3.07)0.17
British Columbia 1
Ontario 1.37(0.85, 2.20)0.19
Quebec 1.40(0.76, 2.60)0.28
Caucasian 1
Black 0.19(0.03, 1.37)0.10
Aboriginal 0.67(0.21, 2.19)0.51
Mixed 0.56(0.13, 2.31)0.42
Other 0.63(0.19, 2.05)0.44
Risk factors2998   
Men who have sex with men 1.41(0.84, 2.36)0.19
Injecting drug users 0.53(0.28, 1.02)0.06
From endemic country 0.80(0.34, 1.86)0.60
Baseline regimen4459   
NNRTI-based 1
NRTI only 0.48(0.07, 3.54)0.48
Single PI-based 1.08(0.52, 2.25)0.83
Boosted PI-based 1.33(0.84, 2.10)0.23
Rate of measurement4459   
< 3 measurements per year 1.26(0.60, 2.64)0.55
3–4 measurements per year 1
4–6 measurements per year 2.26(1.21, 4.22)0.01
> 6 measurements per year 2.50(1.21, 5.18)0.01
Year of first ARV (continuous per year)44591.01(0.91, 1.12)0.80
Hepatitis C virus coinfected33780.78(0.44, 1.38)0.39
Hepatitis B virus coinfected14830.38(0.12, 1.23)0.11
AIDS at baseline42580.76(0.39, 1.47)0.42
Baseline immunological parameters    
Baseline CD4 T-cell count4459   
< 200 cells/μL 0.11(0.05, 0.25)<0.0001
200–349 cells/μL 0.40(0.19, 0.81)0.01
350–531 cells/μL 1.14(0.56, 2.33)0.72
532–1170 cells/μL (within normal range) 1
Continuous (per 100 cells/μL increase)44591.36(1.27, 1.47)<0.0001
Baseline CD4 T-cell percentage4459   
< 14% 0.02(0.01, 0.04)<0.0001
14–24% 0.11(0.05, 0.22)<0.0001
25–38% 0.38(0.18, 0.78)<0.01
39–55% (within normal range) 1
Continuous (per 0.1 unit increase)44593.24(2.69, 3.89)<0.0001
Baseline CD8 T-cell count4459   
< 236 cells/μL 1.14(0.51, 2.55)0.76
236–651 cells/μL (within normal range) 1
652–1200 cells/μL 0.82(0.51, 1.31)0.40
> 1200 cells/μL 0.40(0.19, 0.84)0.02
Continuous (per 100 cells/μL increase)44590.95(0.90, 0.99)0.02
Baseline CD8 T-cell percentage4459   
18–31% (within normal range) 1
32–60% 0.40(0.18, 0.87)0.02
> 60% 0.07(0.03, 0.17)<0.0001
Continuous (per 0.1 unit increase)44590.56(0.49, 0.65)<0.0001
Baseline CD3 T-cell percentage4459   
< 50% 0.26(0.04, 1.85)0.18
50–64% 0.55(0.25, 1.21)0.14
65–85% (within normal range) 1
> 85% 0.74(0.39, 1.41)0.36
Continuous (per 0.1 unit increase)44591.11(0.92, 1.35)0.28
Baseline ratio4459   
< 0.5 0.03(0.01, 0.05)<0.0001
0.5–1.1 0.14(0.07, 0.29)<0.0001
1.2–3.3 (within normal range) 1
Continuous (per 0.1 unit increase)44591.20(1.17, 1.24)<0.0001
Baseline lymphocyte count (continuous per 100 cells/μL increase)44591.00(0.97, 1.03)0.99
Suppressed VL at baseline44591.14(0.36, 3.61)0.82
Viral suppression (time-dependent)44592.90(1.44, 5.84)<0.01
Viral load continuous (time-dependent)44590.65(0.47, 0.92)0.01
Table 3. Multivariable proportional hazards models of time to normalization of immunophenotype profile
ParameterHazard ratio95% confidence intervalP-value
  1. n = 4459. NRTI, nucleoside reverse transcriptase inhibitor; PI, protease inhibitor; NNRTI, non-nucleoside reverse transcriptase inhibitor.
Age (continuous per 10 years)1.14(0.92, 1.40)0.23
Rate of measure   
< 4 measures per year1 
4–6 measures per year 2.23(1.36, 3.68)0.002
> 6 measures per year2.64(1.38, 5.05)0.003
Baseline regimen   
NNRTI-based 1
NRTI only 0.41(0.06, 3.01)0.38
Single PI-Based 0.89(0.42, 1.88)0.75
Boosted PI-based 1.28(0.80, 2.04)0.30
Baseline CD4 count (continuous per 100 cells)1.44(1.33, 1.55)<0.0001
Suppressed viral load (< 50 copies/mL) (time-dependent)2.84(1.43, 5.63)0.003


We studied the predictors of complete TCP normalization in a well-described, large cohort of treatment-naïve HIV-seropositive patients with an altered immune phenotype who were initiating cART. At baseline, although almost all the patients exhibited CD4 T-cell depletion and CD4:CD8 ratio dysregulation, only a third of the population had lost their CD3 T-cell homeostasis. Despite a trend of increasing CD4 T-cell counts, only one-third of our sample normalized their CD4 T-cell counts during the study period and very few (2%) normalized their complete TCP. The overall probability of normalizing the complete TCP 5 years after treatment initiation was greater in those with baseline CD4 T-cell counts within the normal range compared with those with baseline CD4 T-cell counts < 200 cells/μL (0.08 vs. 0.007, respectively). We found that a high CD4 T-cell count at baseline was associated with an increased hazard of normalizing the complete TCP, and that individuals who were unable to suppress their viral load over time were less likely to have normal TCP.

As a change in the marker will be documented sooner on average for a patient monitored closely than for a patient followed infrequently, the rate of measurement of a marker is associated with the chance of observing a change in that marker. Our previous paper on rates of viral load measurement documented associations of rates of measurement with characteristics such as geographical region, HIV risk factor and age [27]. Failure to adjust for rates of measurement of CD4 T-cell counts may result in spurious associations of normalization of phenotype with covariates associated with more frequent measurement. Our multivariable analysis shows that, after adjusting for the frequency of laboratory measurements, CD4 T-cell counts and HIV viral load suppression still have a statistically significant effect on normalization. We detected no statistically significant associations of age and type of baseline regimen with the likelihood of normalizing the complete TCP.

Despite the fact that a third of the study population restored their CD4 T-cell counts and 68% achieved T-cell homeostasis, only 7% of the patients were able to normalize their CD4:CD8 ratio. This was attributable to a disproportionate contribution of the CD8 T-cell compartment to the overall ratio. Indeed, the majority of patients with a dysregulated ratio had abnormally high levels of circulating CD8 T-cells. This phenomenon was reflected in the univariate analysis, where high percentages of CD8 T-cells were associated with a decreased likelihood of normalization.

Although immune recovery has been extensively studied in numerous cohorts, research has focused almost exclusively on CD4 T-cell recovery [28, 29]. There is a paucity of data available on long-term CD4:CD8 ratio recovery and, to our knowledge, no studies on T-cell homeostasis restoration in treated HIV-positive patients. The findings from our study are in line with those of previous studies showing that the mechanism of T-cell homeostasis is a ‘blind’ process that occurs irrespective of changes in the CD4 and CD8 compartment [30, 31]. In HIV disease, the failure of T-cell homeostasis has been shown to coincide with the onset of clinically defined AIDS.

While previous studies have measured T-cell homeostasis in terms of CD3 T-cell counts, in this study we chose to look at T-cell homeostasis in terms of CD3 T-cell percentages as we consider percentage to be a more reliable marker with less between-measurement variation [32, 33]. Furthermore, while CD3 T-cell count reflects the total lymphocyte count in the blood, CD3 T-cell percentage avoids the impact of lymphocyte fluctuations and is a better measure of proportional stability of circulating T cells [32, 34, 35].

In a study by Margolick et al., a group of 372 seroconverters enrolled in the Multicenter AIDS Cohort Study (MACS) were followed for 8 years post-seroconversion. The study data showed that individuals who did not develop clinical AIDS were able to maintain homeostatic levels of T cells for several years, despite significant declines in their CD4 T-cell counts. However, among those who developed AIDS, a loss of T-cell homeostasis (characterized by a substantial decline in the total level of circulating CD3 T-cell) was observed approximately 18 months prior to the onset of AIDS [36]. Using artificial intelligence tools, we previously published on the utility of CD3 T-cell percentages in predicting mortality and morbidity in HIV-positive treated individuals [37]. These findings highlight the potential role of blind T-cell homeostasis as an independent marker of HIV disease progression and AIDS onset. Studies of HIV-negative individuals have emphasized the importance of T-cell homeostasis in maintaining the integrity of the immune system. Disruption of T-cell homeostasis has been associated with autoimmune diseases such as SLE, rheumatoid arthritis and multiple sclerosis [14, 38, 39]. There is a paucity of studies that have investigated the link between long-term disrupted T-cell homeostasis and morbidity in the context of HIV disease. We have recently reported data from CANOC, showing for the first time that loss of T-cell homeostasis is associated with poor clinical outcomes in those treated for HIV infection [40].

It is unclear to what extent HIV accounts for persistent immune dysregulation following successful therapy. As HIV primarily leads to CD4 T-cell depletion, it was not surprising that successful ART significantly increased CD4 T-cell counts in our cohort of patients. Although the CD8 T-cell compartment rapidly expands early in HIV infection, other comorbid viral infections may contribute to and sustain this expansion. These viral coinfections include infections with cytomegalovirus (CMV) and Epstein−Barr virus (EBV), which are known to sustain the expansion of CD8 T-cells during their chronic infection phase [41, 42]. A recent study by Naeger et al. showed that the CMV-specific CD8 T-cell response is high in successfully treated HIV-infected individuals [43]. Thus, the persistent CD8 T-cell lymphocytosis and the low CD4:CD8 ratio observed in long-term treated and virologically suppressed HIV-infected patients might be partially explained by a subclinical CMV infection. In HIV-seronegative individuals, CMV infection is also associated with immune senescence [44, 45]. An important characteristic of immune senescence is the large expansion of CD8+CD28 T cells. The frequency of these cells increases with progressive HIV disease and can account for up to 50% of the CD8 T-cell compartment, thus contributing to the dysregulation of the CD4:CD8 ratio [46]. Geriatric longitudinal studies found a CD4:CD8 T-cell ratio of < 1 and CMV seropositivity to be part of an IRP which is associated with higher mortality and morbidity rates [44]. Elderly individuals with an IRP have increased susceptibility to infections, reactivation of latent pathogens, and decreased responses to vaccination [47], reflecting an age-related loss of T-cell responsiveness.

Because of the strict definition of the outcome variables, a significant number of patients were excluded from the analysis because of missing data on one or more of the six required immunological markers at baseline and/or at follow-up. Finally, because there is a fair amount of heterogeneity in T-cell reference values across various studies and populations, it is possible that our definition of TCP may not apply to patients outside of our study demographics.

In conclusion, disruptions in T-cell homeostasis resulting from infections or medical interventions are generally expected to be transient events [48]. However, our data show a striking lack of recovery of the TCP despite successful treatment. It remains to be determined if HIV infection results in irreversible phenotypic as well as functional changes within the T-cell compartment or whether comorbid chronic viral infections such as CMV infection prevent the normalization of T-cell homeostatic and regulatory processes. For those with HIV infection, the clinical importance of maintaining a normal phenotype as it relates to long-term morbidities remains to be elucidated. The data presented in this study suggest that the concept of immune reconstitution should not be restricted to CD4 T-cell counts, as HIV infection is associated with other changes within the T-cell compartment that are not immediately restored by antiretroviral therapy. These residual immune abnormalities in many patients resemble the IRP of elderly people and might thus be associated with deleterious clinical outcomes in the long term. The data presented in this study reflect the need for more in-depth research into this complex area of long-term management of HIV infection.


We would like to thank all of the participants for allowing their information to be a part of the CANOC collaboration. CANOC is funded by an Emerging Team Grant from the Canadian Institutes of Health Research (CIHR) and is supported by the CIHR Canadian HIV Trials Network (CTN242). PN is supported through a CANOC Scholarship Award, a collaborative programme of CANOC, CTN and REACH. ANB is supported by a CIHR New Investigator Award. CC and JR are supported by Career Scientist Awards from the OHTN. MBK is supported by a Chercheur-Boursier Clinicien Senior Career Award from the Fonds de recherche en santé du Québec (FRSQ). MRL receives salary support from CIHR. JSGM is supported by an Avant-Garde Award from the National Institute on Drug Abuse, National Institutes of Health.

Conflicts of interest: There were no conflicts of interest for this study.

Appendix: Appendix

The CANOC collaboration includes the following contributors.

Investigators. Gloria Aykroyd [Ontario HIV Treatment Network (OHTN)], Louise Balfour [University of Ottawa and OHTN Cohort Study (OCS) Co-Investigator], Ahmed Bayoumi (University of Toronto and OCS Co-Investigator), Ann Burchell (OHTN), John Cairney (University of Toronto and OCS Co-Investigator), Liviana Calzavara (University of Toronto and OCS Co-Investigator), Angela Cescon (British Columbia Centre for Excellence in HIV/AIDS), Curtis Cooper (University of Ottawa and OCS Co-Investigator), Kevin Gough (University of Toronto and OCS Co-Investigator), Silvia Guillemi (British Columbia Centre for Excellence in HIV/AIDS and University of British Columbia), P. Richard Harrigan (British Columbia Centre for Excellence in HIV/AIDS and University of British Columbia), Marianne Harris (British Columbia Centre for Excellence in HIV/AIDS), George Hatzakis (McGill University), Robert Hogg (British Columbia Centre for Excellence in HIV/AIDS and Simon Fraser University), Sean Hosein (Canadian AIDS Treatment Information Exchange), Don Kilby (University of Ottawa and OHTN), Marina Klein (Montreal Chest Institute Immunodeficiency Service Cohort and McGill University), Richard Lalonde (The Montreal Chest Institute Immunodeficiency Service Cohort and McGill University), Viviane Lima (British Columbia Centre for Excellence in HIV/AIDS and University of British Columbia), Mona Loutfy (University of Toronto, Maple Leaf Medical Clinic and OCS Co-Investigator), Nima Machouf (Clinique Medicale l'Actuel and Université de Montréal), Ed Mills (British Columbia Centre for Excellence in HIV/AIDS and University of Ottawa), Peggy Millson (University of Toronto and OCS Co-Investigator), Julio Montaner (British Columbia Centre for Excellence in HIV/AIDS and University of British Columbia), David Moore (British Columbia Centre for Excellence in HIV/AIDS and University of British Columbia), Alexis Palmer (British Columbia Centre for Excellence in HIV/AIDS), Janet Raboud (University of Toronto, University Health Network and OCS Co-investigator), Anita Rachlis (University of Toronto and OCS Co-Investigator), Stanley Read (University of Toronto and OCS Co-Investigator), Sean Rourke (OHTN and University of Toronto), Marek Smieja (McMaster University and OCS Co-Investigator), Irving Salit (University of Toronto and OCS Co-Investigator), Darien Taylor (Canadian AIDS Treatment Information Exchange and OCS Co-Investigator), Benoit Trottier (Clinique Medicale l'Actuel and Université de Montréal), Chris Tsoukas (McGill University), Sharon Walmsley (University of Toronto and OCS Co-Investigator), and Wendy Wobeser (Queens University and OCS Co-Investigator).

Analysts and staff. Mark Fisher (OHTN), Sandra Gardner (University of Toronto), Nada Gataric (British Columbia Centre for Excellence in HIV/AIDS), Guillaume Colley (British Columbia Centre for Excellence in HIV/AIDS), Sergio Rueda (OHTN), and Benita Yip (British Columbia Centre for Excellence in HIV/AIDS).