Three different patterns of CD4 recovery in a cohort of Chinese HIV patients following antiretroviral therapy

Abstract

To explore the heterogeneity of CD4 responses following highly active antiretroviral therapy, the patterns of CD4 recovery of HIV-1-infected Chinese patients who have been on their first antiretroviral regimen for ≥5 years were analysed. The CD4 trajectories were traced, smoothed and differentiated into three defined profiles. Half (56.3%) were ‘satisfactory responders’, with CD4 gain of >100 cells/μL and a peak of >350 cells/μL, plateauing before the end of Year 5. Thirty-three (24.4%) were ‘continuing responders’ whose CD4 rise persisted at Year 4–5. The remaining 26 (19.3%) were ‘poor responders’. Presentation with AIDS before therapy was common not just among ‘poor’ but also paradoxically the ‘continuing’ responders. While a majority had responded well to antiretroviral therapy, older patients and those with AIDS diagnosis before initiation of therapy may never achieve a satisfactory level even with effective treatment. Categorization of HIV patients by their CD4 trajectory may support the prediction of immunological outcome over time, and ultimately inform treatment choices.

Keywords

HIV HAART Asia AIDS treatment antiretroviral therapy immunological response CD4 count

Introduction

CD4 T-cell count is a useful surrogate of immune function and an important marker for monitoring the progression of HIV disease. Studies have shown that immunodeficiency is associated not only with AIDS but also non-AIDS defining diseases.¹ Once highly active antiretroviral therapy (HAART) is initiated, CD4 cell count increases rapidly in the first few weeks, followed by a slower recovery contributed to by newly-produced T cells.² Low baseline CD4 count before HAART has been shown to be a strong predictor of mortality and AIDS.³ Ineffective restoration of CD4 cells while on HAART is also associated with AIDS, non-AIDS disease⁴ and mortality.⁵ Factors influencing CD4 count after HAART include age,^6–10 gender,¹¹ baseline CD4 count,^5–10,12,13 HIV RNA level,^6,14 virological response to HAART,^6–8,12 duration of HIV infection^9,12 and clinical AIDS prior to HAART.¹³ While starting therapy at lower CD4 counts is undesirable, a large number of HIV-infected individuals in many countries only managed to commence HAART at a much lower baseline CD4 count,¹⁵ thus jeopardizing the clinical outcome.

Currently, there is significant heterogeneity in research findings on the trajectory of CD4 counts following HAART. Some studies suggested that the count plateaus after 2–4 years,^9,10,16 others showing a continued rise,^8,9,16 while the overall pattern appears to vary with the degree of viral suppression,^13,16,17 and either baseline or concurrent CD4 cell count.^9,10,16,17 Delineation of the CD4 count trajectory is further complicated by the variation in HAART regimens. The heterogeneity of CD4 responses could be particularly relevant in the Chinese populations who have been shown to have lower CD4 counts in healthy individuals.¹⁸ Through understanding the patterns of CD4 recovery, evaluation of specific effects of HAART regimens can become possible. In addition, patients can be classified into different response groups such that monitoring of each can be developed. Against this background, we set out to: firstly define and classify the CD4 trajectories since HAART initiation in a cohort of Chinese HIV patients and secondly explore factors which may be associated with the specific CD4 patterns.

Methods

Study population

Clinical data were collected from all HIV-1 infected, treatment-naive, Chinese patients receiving their first HAART regimen at the Integrated Treatment Centre (ITC) between 24 June 1997 and 31 December 2012. ITC serves as an outpatient clinic for about 60–70% of all HIV-positive patients under care in Hong Kong. The data retrieved included demographics, HIV status (HIV diagnosis date, transmission route and HIV-1 subtype), clinical status (diagnosis of AIDS defining illness, and date and cause of death, if appropriate) and therapy status (pre-HAART CD4 and HIV RNA measurements, first HAART regimen, its start and end date). CD4 count (cells/µL) and HIV RNA (copies/mL) were recorded regularly (3–4 times per year) for each patient by a central laboratory.

To evaluate the CD4 recovery patterns following effective treatment with HAART, cases were selected if they (a) were aged ≥16 at the time of diagnosis, (b) had more than three CD4 and plasma HIV RNA measurements, (c) had CD4 data at the beginning and month 60, (d) have been on first HAART regimen for ≥60 months, and (e) had achieved sustained HIV RNA suppression (≤500/mL) throughout the five years (without any AIDS-defining illness), though isolated blips (≤2000/mL) were allowed.

CD4 trajectories smoothing and classification

CD4 trajectories for enrolled patients were smoothed using a locally-weighted regression (Loess) by quadratic polynomial.¹⁹ Each smoothed value was determined by the neighbouring data points within the defined span, weighted by the distance from the smoothed value,²⁰ with missing values estimated in Loess. Effective CD4 recovery was defined by a rise of >100/μL and an ultimate level of ≥350/μL. The trajectories were classified into mutually-exclusive patterns, the criteria of which were derived from literature based on clinical rationale. Sensitivity analyses on parameters were performed to support the definition.

Statistical analysis

The Loess smoothing was conducted using the R statistical environment. Clinical characteristics and outcome were compared between the three groups by odds ratio (OR) with 95% confidence interval (CI) and Mann–Whitney U test in IBM SPSS Statistics® 21. Missing values were excluded from corresponding analysis.

Ethical consideration

This was a retrospective cohort study, involving anonymised data already collected for clinical management, without the need for contacting patients. Institutional approval for access to the data was obtained from Department of Health in compliance with the Personal Data (Privacy) Ordinance. Approval of the Joint Chinese University of Hong Kong – New Territories East Cluster Clinical Research Ethics Committee was obtained.

Results

Patient characteristics

Of the 1070 patients in the ITC database, 160 have been on their first regimen for at least five years. Twenty-five patients were excluded: 4 had fewer than three CD4 measurements, 13 did not have CD4 data up to month 60 and 8 had viral load rebound during treatment. Ultimately, 135 met the inclusion criteria. Among them, 90.4% were men with a median age of 38.1 years (IQR 33.1–47.3) when HAART was commenced (Table 1). Heterosexual transmission and infection in men-who-have-sex-with-men each accounted for about half of the study population. There were roughly equal numbers of CRF01_AE and B subtypes (35.6% and 37%, respectively). At baseline, 75.2% had a CD4 count of ≤200/μL while 58.6% had a viral load of >100,000/mL.

Table 1.

Characteristics of recruited Chinese HIV patients (n = 135).

	No.	%
Demographics
Male gender	122	90.4%
Status at diagnosis
HIV age group (years)
<35	56	41.5%
35–64	74	54.8%
>64	5	3.7%
Mode of transmission
Heterosexual	64	47.4%
Men-who-have-sex-with-men	70	51.9%
Injection drug users	1	0.7%
HIV subtype
CRF01_AE	48	35.6%
B	50	37.0%
CRF07_BC	3	2.2%
undetermined	34	25.2%
Late diagnosis^a	21	15.6%
Status pre-HAART
Baseline CD4 ≤200/μL (n = 129)	97	75.2%
Baseline VL >100,000/mL (n = 128)	75	58.6%
Diagnosis of AIDS before HAART	29	21.5%
>1 year from diagnosis to HAART	74	54.8%
HAART
Age at initiation of treatment
<35	40	29.6%
35–64	89	65.9%
>64	6	4.4%
Regimen category (with nucleoside reverse transcriptase inhibitors backbone)
Non-nucleoside reverse transcriptase inhibitor
Efavirenz	31	23.0%
Nevirapine	9	6.7%
Protease inhibitor
Atazanavir with or without booster	7	5.2%
Indinavir	20	14.8%
Boosted lopinavir	66	48.9%
Nelfinavir	1	0.7%
Saquinavir	1	0.7%
Outcome
≤3 months from HAART to SVL^b	80	59.3%
Development of AIDS during HAART (n = 29)	9	31.0%
Deceased	6	4.4%
Median maximum CD4 within 60 m (IQR)	477.1	(388.0–612.7)
Median maximum CD4 rise within 60 m (IQR) (n = 129)	328.6	(225.8–451.4)

Note: HAART: highly active antiretroviral therapy; VL: viral load.

Late diagnosis is defined by AIDS occurring within three months from HIV diagnosis.

SVL (suppressed viral load) is defined as viral load reading ≤500 copies/mL.

CD4 responses and trajectories

The total follow-up time, as defined by the interval between first and last CD4 measurements while on one’s first regimen, was 10,951 months (average = 81 months per person). Slightly more than half (53.3%) had achieved a CD4 count >500/μL by Year 5, while about 80% attained the minimum threshold of 350/μL. Overall, 26 (19.3%) were classified as ‘poor responders’, 76 (56.3%) as ‘satisfactory responders’, and 33 (24.4%) as ‘continuing responders’ (Supplementary Figure 1).

‘Poor responders’ had characteristically flat CD4 trajectories from month 0 to 60. Rise of CD4 above baseline, if present, was ≤100/μL, and that the ultimate CD4 count was ≤350/μL. A few presented with sharp rises and falls but without any ultimate net gain of CD4 cells. For ‘satisfactory responders’, the final CD4 count was invariably ≥350/μL. There was a rise above baseline by >100/μL in the early period (Year 1–2) followed by a plateau by Year 3–5, such that the net gain of CD4 at Year 3 and 5 was <100/μL. Some (n = 63) had continuous rise throughout the five years, but the amplitude of increase between Year 3 and 5 was below 100/μL. The remaining ‘continuing responders’ gave a rising trend of CD4 throughout the five years. Apart from the increased CD4 count above baseline by >100/μL and a final value >350/μL, there was a continuing rise of >100/μL between Year 3 and 5 (Figure 1).

Figure 1.

Selected examples of CD4 trajectory curves before and after smoothing.

Factors associated with CD4 patterns

CD4 patterns were compared between ‘satisfactory responders’ and (1) ‘poor responders’ and (2) ‘continuing responders’, respectively. Table 2 shows the factors associated with the differentiation of response patterns. Demographically, gender was not a significant factor for distinguishing between patterns. However, patients aged 35 or above at diagnosis were more likely to be ‘poor responders’ (80.8%) than ‘continuing responders’ (39.4%) (OR = 0.15, 95% CI = 0.05–0.51, p = 0.002). Similarly those aged above 40 at HAART commencement (OR = 0.31, p = 0.03) were more likely to be ‘poor responders’. Overall, the median pre-HAART CD4 for poor responders was 81/μL, whereas that for satisfactory responders was 157/μL and continuing responders was 156/μL. A baseline CD4 of ≤150/μL was a significant predictor for ‘poor’ compared to ‘satisfactory response’ (OR = 3.04, 95% CI = 1.13–8.17, p = 0.03). Other factors, such as mode of transmission, subtype, pre-treatment viral load levels and HAART regimen were not significant in differentiating between ‘satisfactory responders’ from the other two responder categories.

Table 2.

Comparisons between trajectories for different response profiles (n = 135) – ‘satisfactory responders’: CD4 rise ≥100/μl and reaching plateau ≥350/μl before month 60; ‘poor responders’: flat CD4 trajectory throughout; ‘continuing responders’: satisfactory response with rising CD4 count at month 36–60.

	Poor responders (n = 26)		Satisfactory responders (n = 76)		Continuing responders (n = 33)		Satisfactory vs. poor responders			Satisfactory vs. continuing responders			Poor vs. continuing responders
	No.	%	No.	%	No.	%	OR	95% CI	p	OR	95% CI	p	OR	95% CI	p
Demographics
Male gender	25	96.2%	69	90.8%	28	84.8%	2.54	0.3–21.66	0.40	0.57	0.17–1.94	0.37	0.22	0.02–2.05	0.19
Status at HIV diagnosis
Age ≥35	21	80.8%	45	59.2%	13	39.4%	2.89	0.99–8.5	0.05	0.45	0.19–1.03	0.06	0.15	0.05–0.51	0.002*
Men having sex with men	11	42.3%	43	56.6%	16	48.5%	0.56	0.23–1.39	0.21	0.72	0.32–1.64	0.44	1.28	0.46–3.61	0.64
Subtype	(n = 22)		(n = 57)		(n = 22)
CRF01_AE	13	59.1%	24	42.1%	11	50.0%	1.99	0.73–5.4	0.18	1.38	0.51–3.69	0.53	0.69	0.21–2.28	0.55
B	9	40.9%	31	54.4%	10	45.5%	0.58	0.21–1.57	0.29	0.7	0.26–1.88	0.48	1.20	0.36–3.97	0.76
Late diagnosis^a	5	19.2%	6	7.9%	10	30.3%	2.78	0.77–10.02	0.12	5.07	1.66–15.49	0.004*	1.83	0.54–6.22	0.34
Status pre-HAART
Baseline CD4 ≤150/μl	(n = 25)		(n = 72)		(n = 32)
Baseline CD4 ≤150/μl	18	72.0%	33	45.8%	15	46.9%	3.04	1.13–8.17	0.03*	1.04	0.45–2.4	0.92	0.34	0.11–1.05	0.06
Baseline VL >100,000/mL	(n = 25)		(n = 71)		(n = 32)
Baseline VL >100,000/mL	16	64.0%	38	53.5%	21	65.6%	1.54	0.6–3.95	0.37	1.66	0.7–3.94	0.25	1.07	0.36–3.21	0.90
AIDS before HAART	9	34.6%	9	11.8%	11	33.3%	3.94	1.36–11.45	0.01*	3.72	1.36–10.16	0.01*	0.94	0.32–2.79	0.92
>12 months from diagnosis to HAART	12	46.2%	46	60.5%	16	48.5%	0.56	0.23–1.37	0.20	0.61	0.27–1.4	0.25	1.10	0.39–3.08	0.86
HAART
HAART initiated at age ≥35	22	84.6%	52	68.4%	21	63.6%	2.54	0.79–8.18	0.12	0.81	0.34–1.91	0.63	0.32	0.09–1.14	0.08
Regimen
Efavirenz based	(n = 7)		(n = 26)		(n = 7)
Efavirenz based	6	85.7%	19	73.1%	6	85.7%	2.21	0.22–21.77	0.50	2.21	0.22–21.77	0.50	1.00	0.05–19.96	1.00
LPV vs. IDV^b	(n = 16)		(n = 45)		(n = 25)
LPV vs. IDV^b	10	62.5%	36	80.0%	20	80.0%	0.42	0.12–1.45	0.17	1	0.29–3.39	1.00	2.40	0.59–9.82	0.22
Protease inhibitor based	19	73.1%	50	65.8%	26	78.8%	1.41	0.53–3.79	0.49	1.93	0.74–5.04	0.18	1.37	0.41–4.56	0.61
Treatment outcome
≤3 months from HAART to SVL^c	17	65.4%	47	61.8%	16	48.5%	1.17	0.46–2.96	0.75	0.58	0.25–1.32	0.20	0.50	0.17–1.43	0.20
Development of AIDS during HAART	(n = 9)		(n = 9)		(n = 11)
Development of AIDS during HAART	1	11.1%	3	33.3%	5	45.5%	0.25	0.02–3.04	0.28	1.67	0.27–10.33	0.58	6.67	0.61–73.03	0.12
Deceased	1	3.8%	4	5.3%	1	3.0%	0.72	0.08–6.75	0.77	0.56	0.06–5.23	0.61	0.78	0.05–13.12	0.86

Note: HAART: highly active antiretroviral therapy; OR: odds ratio; VL: viral load.

Late diagnosis is defined by AIDS occurring within three months from HIV diagnosis.

LPV (lopinavir) coded as 1 and IDV (indinavir) coded as 0 in analysis.

SVL (suppressed viral load) is defined as viral load reading ≤500 copies/mL.

*p Value < 0.05.

From a different perspective, patients with late HIV diagnosis, as defined by AIDS occurring within three months from HIV diagnosis, were less likely to be ‘satisfactory responders’. Likewise, patients with AIDS diagnosis before HAART initiation were more likely to be either ‘poor responders’ (OR = 3.94) or ‘continuing responders’ (OR = 3.72). In this cohort, a majority of protease inhibitor (PI)-treated patients commenced indinavir or lopinavir. There was no difference between the two in their response category. As for treatment outcome, there was no significant difference in the time taken for achieving viral load suppression in the three CD4 response categories. In addition, there was no statistically significant difference in the change of CD4 in the first year between satisfactory and continuing responders (Mann–Whitney U = 1031, p = 0.14). It was not until year 3 that significant difference between the two response groups on CD4 change was observed (Mann–Whitney U = 910, p = 0.02). Multivariable logistic regression was performed by inputting variables that were significant in univariate analysis. AIDS before HAART remained the main predictor for distinguishing between ‘poor responder’ and ‘satisfactory responder’, while an age of ≥35 at HIV diagnosis was the distinguishing feature of both ‘poor responders’ and ‘continuing responders’ (results not shown).

Discussion

This is the first study conducted in an ethnic Chinese population for classifying HIV patients by their pattern of CD4 response following HAART. By including only patients who had been put on their first regimen, we removed the potential bias resulting from multiple changes in antiretroviral regimens. A CD4 cutoff of 350/μL has been chosen, as derived from the lower limit of the reference range in healthy Chinese populations (ranging between 292/μL and 415/μL),^18,21 a level about 100/μL lower than that in Caucasians.^22,23 Studies have shown that CD4 counts usually rise by >100/μL in the first year of treatment,^9,10,13,16 a value that we adopted as a surrogate of minimum immunological responses. While a vast majority in our cohort had started HAART with low CD4 counts <200/μL, most went on to reach 350/μL within five years of treatment. The proportion of those achieving a target of >500/μL was however lower than Caucasian studies (51.7% vs. 59%),^8,9 a discrepancy which may be related to a generally lower pre-treatment CD4 counts in the Asian population.^17,24

Although restoration of immunity does occur consistently following HAART, the question remains as to why interpatient variability is so high as regards the rates and duration of CD4 change. We took a novel approach by dividing all patients into three distinct groups based on their pattern of response. Our results showed that the main predictors of a satisfactory response are a good baseline CD4 count and an absence of AIDS at initiation of HAART, which are in line with results elsewhere.^5–10,12 Our own sensitivity analysis has shown that the numbers in each group was sufficient to generate reliable significance testing for comparing the CD4 patterns. Interestingly, both ‘continuing responders’ and ‘poor responders’ shared a similar association with a diagnosis of AIDS before treatment. Patients with continued increases in CD4 count beyond five years are likely those who started with a more impaired immune status but with good potential for immune recovery, while ‘poor responders’ had such severe immunological damage that functional recovery could not be achieved. This has been explained previously with reference to thymic involution occurring in the ageing population.²⁵ It is likely that the CD4 count of ‘continuing responders’ will eventually plateau once the T cells have reached their full potential. It is possible that the CD4 trajectory may vary with the HAART regimen, though this could not be confirmed in our study. A large meta-analysis reported that boosted PI therapy was significantly linked to greater rises in CD4 count than the other classes of drugs at 48 weeks; but no analysis was performed on individual regimen.²⁶ Another study reported no difference in on-treatment CD4 count between initial HAART regimens,¹⁷ although there were no restrictions on changing regimen once started.

To conclude, our study has enabled us to define the CD4 trajectories of an exclusively Chinese HIV population over a prolonged time frame of five years. Compared to Caucasian populations, a smaller proportion in our cohort has ultimately reached a level of >500/μl. In our study, patients with the lowest CD4 counts at HAART initiation and older patients had a worse outcome and they may never achieve a satisfactory CD4 count on treatment. We acknowledge that there are inherent limitations in the study. Foremost, almost all patients had acquired HIV through sexual transmission and very few were injection drug users, which may limit the generalizability of our results to HIV populations in China or to other ethnic populations. Patients having less than three CD4 measurements were excluded. This was however unlikely to have led to any significant deviation from the main observations. Other factors that may influence the CD4 trajectories but not measured included CD8 T-lymphocyte count^6,9 and hepatitis C coinfection status.¹⁰ The impacts of different HAART regimens also need to be further explored. Nevertheless, by aiming to initiate HAART earlier before disease progression, the potential for immune recovery is increased and therefore, risk of mortality and morbidity could be reduced in these patients, which may apply to all ethnic groups.

Footnotes

Acknowledgements

The authors thank all patients and staff at the Integrated Treatment Centre for their participation and assistance in the maintenance of the database.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Direct Grant of the Medical Faculty, The Chinese University of Hong Kong (project code: 4054074) and the Health and Medical Research Fund (project reference no. CU-14-03-01), of Food and Health Bureau of the Hong Kong Special Administrative Region Government. Li Ka Shing Institute of Health Sciences, The Chinese University of Hong Kong is acknowledged for supporting the conduct of the research.

Disclaimer

The opinions and assertions contained herein are private views of the authors and do not necessarily reflect those of the Centre for Health Protection, Hong Kong Department of Health. Initial exploration of related data has been presented at the Eleventh International Congress on Drug Therapy in HIV Infection 2012, Glasgow, UK.

References

Reekie

Kosa

Engsig

. Relationship between current level of immunodeficiency and non-acquired immunodeficiency syndrome-defining malignancies. Cancer 2010; 116: 5306–5315.

Pakker

Notermans

de Boer

. Biphasic kinetics of peripheral blood T cells after triple combination therapy in HIV-1 infection: a composite of redistribution and proliferation. Nat Med 1998; 4: 208–214.

Egger

May

Chêne

. Prognosis of HIV-1-infected patients starting highly active antiretroviral therapy: a collaborative analysis of prospective studies. Lancet 2002; 360: 119–129.

Baker

Peng

Rapkin

. CD4+ count and risk of non-AIDS diseases following initial treatment for HIV infection. AIDS 2008; 22: 841–848.

Van Griensven

Thai

. Predictors of immune recovery and the association with late mortality while on antiretroviral treatment in Cambodia. Trans R Soc Trop Med Hyg 2011; 105: 694–703.

Kaufmann

Perrin

Pantaleo

. CD4 T-lymphocyte recovery in individuals with advanced HIV-1 infection receiving potent antiretroviral therapy for 4 years: the Swiss HIV Cohort Study. Arch Intern Med 2003; 163: 2187–2195.

Moore

Keruly

. CD4+ cell count 6 years after commencement of highly active antiretroviral therapy in persons with sustained virologic suppression. Clin Infect Dis 2007; 44: 441–446.

Kelley

Kitchen

Hunt

. Incomplete peripheral CD4+ cell count restoration in HIV-infected patients receiving long-term antiretroviral treatment. Clin Infect Dis 2009; 48: 787–794.

Kaufmann

Furrer

Ledergerber

. Characteristics, determinants, and clinical relevance of CD4 T cell recovery to <500 cells/microL in HIV type 1-infected individuals receiving potent antiretroviral therapy. Clin Infect Dis 2005; 41: 361–372.

10.

Margolick

Jamieson

. CD4+ T-cell counts and plasma HIV-1 RNA levels beyond 5 years of highly active antiretroviral therapy. J Acquir Immune Defic Syndr 2011; 57: 421–428.

11.

Bastard

Soulinphumy

Phimmasone

. Women experience a better long-term immune recovery and a better survival on HAART in Lao People’s Democratic Republic. BMC Infect Dis 2013; 13: 27–27.

12.

Tarwater

Margolick

Jin

. Increase and plateau of CD4 T-cell counts in the 3(1/2) years after initiation of potent antiretroviral therapy. J Acquir Immune Defic Syndr 2001; 27: 168–175.

13.

Lifson

Krantz

Eberly

. Long-term CD4+ lymphocyte response following HAART initiation in a U.S. military prospective cohort. AIDS Res Ther 2011; 8: 2–6405-8-2.

14.

Santoro

Armenia

Alteri

. Impact of pre-therapy viral load on virological response to modern first-line HAART. Antivir Ther 2013; 18: 867–876.

15.

World Health Organization, UNICEF, UNAIDS. Global update on HIV treatment 2013: results, impact and opportunities – Summary, Switzerland: World Health Organization, 2013.

16.

Mocroft

Phillips

Gatell

. Normalisation of CD4 counts in patients with HIV-1 infection and maximum virological suppression who are taking combination antiretroviral therapy: an observational cohort study. Lancet 2007; 370: 407–413.

17.

Lok

Bosch

Benson

. Long-term increase in CD4+ T-cell counts during combination antiretroviral therapy for HIV-1 infection. AIDS 2010; 24: 1867–1876.

18.

Wong

Siu

. Reference ranges for lymphocyte subsets among healthy Hong Kong Chinese adults by single-platform flow cytometry. Clin Vaccine Immunol 2013; 20: 602–606.

19.

Cleveland

Devlin

. Locally weighted regression: an approach to regression analysis by local fitting. J Am Stat Assoc 1988; 83: 596–610.

20.

Jacoby

. Loess: a nonparametric, graphical tool for depicting relationships between variables. Elect Stud 2000; 19: 577–613.

21.

Kam

Leung

Kwok

. Lymphocyte subpopulation reference ranges for monitoring human immunodeficiency virus-infected Chinese adults. Clin Diagn Lab Immunol 1996; 3: 326–330.

22.

Jentsch-Ullrich

Koenigsmann

Mohren

. Lymphocyte subsets’ reference ranges in an age- and gender-balanced population of 100 healthy adults – a monocentric German study. Clin Immunol 2005; 116: 192–197.

23.

Lebranchu

Thibault

Degenne

. Abnormalities in CD4+ T lymphocyte subsets in patients with common variable immunodeficiency. Clin Immunol Immunopathol 1991; 61: 83–92.

24.

Wong

Chan

Lee

. Delayed progression to death and to AIDS in a Hong Kong cohort of patients with advanced HIV type 1 disease during the era of highly active antiretroviral therapy. Clin Infect Dis 2004; 39: 853–60.

25.

Taub

Longo

. Insights into thymic aging and regeneration. Immunol Rev 2005; 205: 72–93.

26.

Bartlett

Fath

Demasi

. An updated systematic overview of triple combination therapy in antiretroviral-naive HIV-infected adults. AIDS 2006; 20: 2051–2064.

Three different patterns of CD4 recovery in a cohort of Chinese HIV patients following antiretroviral therapy – a five-year observational study