Secondary Logo

Journal Logo

CLINICAL SCIENCE

Usefulness of total lymphocyte count in monitoring highly active antiretroviral therapy in resource-limited settings

Badri, Motasim; Wood, Robin

Author Information
  • Free

Abstract

Introduction

The survival benefits of highly active antiretroviral therapy (HAART) are well documented. However, owing to high cost, few in the developing countries currently have access to antiretroviral therapy (ART). Recent initiatives of the World Health Organization (WHO) for scaling up ART in resource-limited settings [1] will result in an increasing number of HIV-infected patients accessing ART. In well-resourced settings, commencement of ART is based predominantly on the presence of HIV-related symptoms and CD4 cell count. Laboratory capacity to measure CD4 cell count is limited in many areas with high HIV seroprevalence. The WHO has proposed a TLC of < 1200 × 106/l as a substitute indication for initiating ART in resource-limited settings [1]. TLC is an inexpensive and useful marker for disease staging and predicting progression to AIDS or death in HIV-infected patients [2]. A high correlation between TLC and CD4 cell count both in asymptomatic and symptomatic HIV-infected patients has been observed [3].

Restoration and/or preservation of immunological function together with suppression of plasma viraemia are important goals of ART. Capacity to monitor plasma viral load is very limited in resource-constrained settings. Use of TLC as a surrogate for other expensive markers, such as the CD4 cell count or viral load, would result in substantial reduction in cost associated with managing HIV-infected patients.

We conducted an observational study to determine whether changes over time in CD4 cell count and plasma HIV RNA could be monitored by changes in TLC in a cohort of indigent patients accessing HAART through participation in phase III randomized clinical trials in a public healthcare facility in Cape Town, South Africa between 1995 and 2001.

Methods

This prospective observational study was conducted in the New Somerset Hospital HIV clinic, University of Cape Town; a major public health care facility dedicated for HIV-infected patients in Cape Town. This clinic provides care for patients referred from a wide range of primary healthcare facilities in Cape Town.

Subjects were recruited into the study from the ongoing clinical trials in the hospital, which have been described in detail elsewhere [4]. In brief, patients > 16 years were recruited into the 12 trials conducted in the hospital between 1995 and 2001. Entry viral load and CD4 cell count requirements varied between studies. Patients were excluded if they presented with acute opportunistic infection, significant laboratory abnormalities, and active substance abuse or if they were treated with immune-modulating or systemic chemotherapeutic agents. Pregnant or lactating females were also excluded. All patients received at least three antiretroviral drugs: a non-nucleoside reverse transcriptase inhibitor with two nucleoside analogues; or three nucleoside analogues; or a protease inhibitor with two nucleoside analogues. Follow-up was repeated at 2–3-month intervals, or more frequently if clinically indicated. Total lymphocyte count (TLC), CD4 cell count and plasma HIV RNA were measured prospectively approximately every month. Plasma HIV RNA was determined by reverse transcriptase–polymerase chain reaction (Roche Amplicor HIV-1 RNA PCR assay version 1.5, Roche, Branchberg, New Jersey, USA), CD4 cell count by flow cytometry (Coulter, Hialeah, Florida, USA) and TLC by automated blood counter (CPICS, Hialeah, Florida, USA) on the same day the blood sample was obtained.

As all three parameters were non-normally distributed, when tested for normality using the Shapiro–Wilks’ W test, association between all pairs of TLC, CD4 cell count and plasma HIV RNA was determined using Spearman's partial rank order correlation. Median change difference (Δ) in baseline TLC, CD4 cell count and viral load at weeks 4, 8, 12 and 48 were assessed using the Friedman ANOVA non-parametric test. Changes in CD4 cell count and plasma HIV RNA relative to TLC was determined using the slope of the correlation's scatter graph. Sensitivity and specificity of the absolute change (i.e., the absolute increase/decrease) in CD4 cell count and TLC from baseline at the different time points during follow-up (i.e., every 4 weeks) as well as for TLC < 1250 × 106/l and CD4 cell count < 200 × 106/l were calculated. The choice of a TLC < 1250 × 106/l as surrogate for a CD4 cell count < 200 × 106/l was based on the significant correlation observed previously in patients presenting to our clinic [2]. All analyses were carried out using STATISTICA software (release 6.6,Tulsa, Oklahoma, USA).

Results

The study population consisted of 266 patients receiving HAART. At screening, 155 failed to meet entry requirements. Mean age of patients was 34.5 years (SD, 9 years). 115 (44%) patients were female and 122 (46%) presented with WHO clinical stage 3 or 4 at their initial clinic visit. At inclusion, median CD4 cell count was 254 × 106/l [inter-quartile range (IQR), 140–364 × 106/l], median TLC was 1480 × 106/l (IQR, 1100–1830 × 106/l) and mean (log10) plasma HIV RNA was 5.4 copies/ml.

At weeks 4, 8, 12 and 48, median increase in TLC (× 106/l) was 30, 52, 139 and 219, respectively. Median increase in CD4 cell count (× 106/l) was 8, 48, 88, and 145, respectively and median decrease in plasma HIV RNA (log10 copies/ml) was −1.6, −2.2, −2.5 and −2.7, respectively. Median increase from baseline in TLC (P < 0.0001) and CD4 cell count (P < 0.0001), as well as median decrease in plasma HIV RNA (P < 0.0001) was significant (Fig. 1).

Fig. 1.
Fig. 1.:
Median change in TLC (a), CD4 cell count (b) and HIV RNA (c) at weeks 4, 8, 12, and 48.

A significant correlation was observed between all pairs of ΔTLC and ΔCD4 cell count (r, 0.61; P < 0.01; Fig. 2a). The correlation between all pairs of Δ viral load (VL) and ΔCD4 cell count was significant (r, −0.26; P < 0.0001; Fig. 2c), but between all the pairs of ΔVL and ΔTLC it was not (r, −0.01; P = 0.73; Fig. 2b). However when the ΔTLC was log-transformed, the correlation reached significance (r, −0.011; P = 0.03; data not shown). The slope of ΔVL and ΔCD4 cell count relative to ΔTLC was [−2.02 − 0.00003(ΔTLC)] and [52.5 + 0.14(ΔTLC)] respectively, and the slope of ΔVL relative to ΔCD4 cell count was [−1.84 – 0.004(ΔCD4 cell count)].

Fig. 2.
Fig. 2.:
Correlation between (a) change in TLC and CD4 cell count, (b) change in TLC and HIV RNA, (c) change in CD4 cell count and HIV RNA, (d) median change in TLC and CD4 cell count, (e) median change in TLC and HIV RNA, and (f) median change in CD4 cell count and HIV RNA.

Reduction in the median plasma HIV RNA at weeks 4, 8, 12 and 48 correlated well with median increase in both ΔCD4 cell count (r, −0.96; P < 0.0001; Fig. 2f) and ΔTLC (r, −0.89; P < 0.0001; Fig. 2e). The correlation between median increase in ΔCD4 cell count and ΔTLC was also significant (r, 0.98; P < 0.0001; Fig. 2d). The correlation between all the pairs (P = 0.004) and median increase (P < 0.0001) in ΔVL and ΔCD4 cell count were significantly greater than that between ΔVL and ΔTLC. The slope of median ΔVL and ΔCD4 cell count relative to ΔTLC was [−1.071 − 0.005(Δ TLC)] and [−0.07 + 0.67(ΔTLC)] respectively, and the slope of median ΔVL relative to CD4 cell count was [−1.69 – 0.008(ΔCD4 cell count)] (Fig. 2).

Sensitivity and specificity of an increase or decrease in TLC for a similar trend in CD4 cell count were 83.4% [95% confidence interval (CI), 81.5–85.1] and 87.3% (95% CI, 83.6–90.4), respectively. To validate our results, we calculated the sensitivity and specificity for the categorical values of the change from baseline in TLC of < 1250 × 106/l and CD4 cell count of < 200 × 106/l which were 99.4% (95% CI, 98.8–99.7) and 92.8% (95% CI, 89.6–95.1), respectively.

Discussion

This study evaluated the usefulness of TLC in monitoring patients initiating HAART by quantifying the association between improvement in TLC, CD4 cell count and plasma HIV RNA following use of HAART. Our findings indicate that TLC can be used as a reasonable surrogate for CD4 cell count. Although the association between change in TLC and plasma HIV RNA is weak at the individual level, TLC can be used to monitor plasma HIV RNA at the group level. In addition to the proposed use of TLC for initiating ART in resource-limited settings in the recent WHO guidelines [1], TLC may have a role in inexpensive monitoring of the immunologic response to HAART. Given the significant correlation and the relatively high sensitivity and specificity rates, in clinical practice TLC can aid health care providers in monitoring the immunologic response to HAART using the slope of the correlation ΔCD4 cell count [52.493 + 0.14(Δ TLC)]. For example, an increase from baseline of 1 × 106/l in TLC is equivalent to an increase of 192.5 × 106/l CD4 cell count.

Recent ART strategies are based on the objective of reducing plasma HIV RNA to below the detection limits of assays in order to stop or reverse the pathogenic process. Although usefulness of TLC as a surrogate for CD4 cell count have been studied previously, our study is the first to assess the association between changes over time in these two parameters and the plasma HIV RNA in patients receiving HAART. In a study from the UK, Beck et al. observed a high correlation (r, 0.76) between TLC and CD4 cell count [3]. The correlation in their study was consistently significant in asymptomatic patients (r, 0.64), symptomatic non-AIDS HIV-infected patients (r, 0.72) and AIDS patients (r, 0.73). Studies by Lai [5], Fournier et al. [6] and Pulido et al. [7] have also demonstrated significant correlation between TLC and CD4 cell count, particularly in patients with CD4 cell count < 200 × 106/l. In addition, in a cohort of non-HAART users in Cape Town, Post et al. have demonstrated that TLC < 1250 × 106/l and CD4 count < 200 × 106/l predict similar progression to AIDS and mortality [2].

For HAART interventions to be cost-effective, targeting and timing are of utmost importance. More recently, the WHO proposed the use of a TLC of 1200 × 106/l as a substitute indication for ART treatment in resource-limited setting for CD4 cell count < 200 × 106/l [1]. Our sensitivity and specificity analyses suggest that a TLC of 1250 × 106/l is a reasonable surrogate for evaluating the outcome of HAART in patients failing to achieve immunologic reconstitution to a CD4 cell count of 200 × 106/l. Our sensitivity estimate concur with recent study by Flanigan et al. from India (80%) [8].

The exclusion criteria for patients entering HAART trials were protocol-determined but may have introduced bias. The major causes for exclusion were acute opportunistic infection and significant laboratory abnormalities as substance abuse among our patients is minimal. Initiation of HAART at the same time as treatment for acute opportunistic infections results in high pill-burdens and added potential for adverse drug interactions. In addition, acute opportunistic infection may cause transient lymphopaenia [9]. However, despite these exclusions, 46% of the patients included in this study presented with WHO stage 3 or 4, indicating that our cohort represents patients with wide spectrum of immune suppression.

In conclusion, our findings indicate that in addition to its use for initiating HAART, as proposed by the recent WHO guidelines, TLC can also be used as an inexpensive surrogate for monitoring the immunolgical response to HAART in resource-limited settings. This data is of particular relevance to sub-Saharan Africa where the laboratory infrastructure to perform CD4 cell and viral load measurement is frequently not available and current international initiatives for facilitating access to ART is increasing the number of patients requiring monitoring.

Acknowledgements

The authors thank the HIV Clinical Research Unit staff for their help in data collection.

Sponsorship: Supported in part by a grant from ‘Secure the Future', Bristol-Myers Squibb.

References

1.World Health Organization. Scaling up antiretroviral therapy in resource-limited settings: guidelines for a public health approach. Executive summary. Geneva: WHO; 2002.
2.Post FA, Wood R, Maartens G. CD4 and total lymphocyte counts as predictors of HIV disease progression.Q J Med 1996, 89: 50–508.
3.Beck EJ, Kupek EJ, Gompels MM, Pinching AJ. Correlation between total and CD4 lymphocyte counts in HIV infection: not making the good an enemy of the not so perfect.Int J STD AIDS 1996, 7:422–428.
4.Badri M, Wilson D, Wood R. Effect of highly active antiretroviral therapy on incidence of tuberculosis in South Africa: a cohort study.Lancet 2002, 359:2059–2064.
4.Lai KK. Correlation of absolute lymphocyte count with CD4 counts in HIV-infected patients.IX International Conference on AIDS. Berlin, June 1993 [abstract PO B03-0988].
6.Fournier AM, Sosenko JM. The relationship of total lymphocyte count to CD4 lymphocyte counts in patients infected with human immunodeficiency virus.Am J Med Sci 1992, 304: 79–82.
6.Pulido F, Rubio R, Castilla V, Sanchez-Manzano MD, Matteos R, Costa JR. Value of total lymphocyte count to start prophylaxis againstP.cariniipneumonia.IX International Conference on AIDS. Berlin, June 1993 [abstract PO B10-1462].
6.Flanigan T, Mahajan A, Kumarasamy N, Mayer K, Carpenter C, Solomon S. Total lymphocyte count as a surrogate for CD4 count to initiate and monitor HAART in resource-limited countries.Ninth Conference on Retroviruses and Opportunistic Infections. Seattle, February 2002 [abstract 456-W].
9.Badri M, Ehrlich R, Wood R, Pulerwitz T, Maartens G. Association between tuberculosis and HIV disease progression in a high TB prevalence. Int J Tuberc Lung Dis 2001, 5:225–232.
Keywords:

HIV; total lymphocyte count; CD4 T lymphocyte count; viral load; HAART

Copyright © 2003 Wolters Kluwer Health, Inc.