点击显示 收起
Ethio-Netherlands AIDS Research Project, Ethiopian Health and Nutrition Research Institute, Addis Ababa, Ethiopia
University of Nijmegen, Nijmegen
Cluster, Infectious Diseases, Department of Research, Municipal Health Service Amsterdam, Department of Clinical Viro-Immunology, Sanquin Research at CLB and Landsteiner Laboratory, and Departments of Human Retrovirology and Clinical Epidemiology and Biostatistics, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
Background.
Human immunodeficiency virus type 1 (HIV-1)uninfected Ethiopians have lower CD4 T cell counts than do other populations in Africa and industrialized countries. We studied whether this unique immunological profile results in shorter survival times in HIV-1infected Ethiopians.
Methods.
Data from an open cohort study of 149 HIV-1infected factory workers in Ethiopia for 19972002 were used. To estimate survival times, a continuous-time Markov model was designed on the basis of CD4 T cell counts and World Health Organization clinical staging. By use of a random-effects model, decline in CD4 T cell counts was compared between HIV-1infected Ethiopian and Dutch individuals.
Results.
Median survival times were in the range of 9.113.7 years, depending on the approach used. This range is similar to that for populations in industrialized countries before the advent of antiretroviral therapy. Ethiopians had a lower annual decline in CD4 T cell counts than did Dutch individuals, which remained when groups with similar CD4 T cell count categories were compared. Moreover, the slower decline in CD4 T cell counts was not due merely to lower HIV-1 RNA loads or an absence of syncytium-inducing/X4 HIV-1 subtype C strains in Ethiopians.
Conclusions.
Low baseline CD4 T cell counts do not imply shorter survival times in Ethiopians than in other populations, presumably because of a slower decline in CD4 T cell counts
Clear understanding of the natural history of HIV-1 disease progression is critical for planning and developing appropriate therapeutic strategies for the HIV-1infected population and also for allocating health-care resources [1, 2]. It is widely held that HIV-1 infection might progress faster in individuals in African countries than in individuals in Western countries. It has been argued that this would be related to a higher burden of infectious diseases in Africa, which is known to induce chronic immune activation and thereby provide more viral target cells and accelerated loss of CD4 T cells once the patient becomes infected with HIV-1 [3]. This hyperimmune activation hypothesis is in line with the more-recent insights in AIDS pathogenesis, which show that the level of bystander immune activation induced in the host determines the rate of loss of CD4 T cells and progression to AIDS [4, 5].
Although the majority of HIV-1 infections occur in sub-Saharan Africa, only very limited information is available about that region. Early studies on disease progression and survival times indeed showed a faster disease progression in individuals in some African countries than in individuals in industrialized countries [69]. Recent reports based on long-term follow-up prospective cohort studies in Uganda, Malawi, and Tanzania indicate that, before the widespread use of antiretroviral therapy, HIV-1 infection had a similar natural disease progression in individuals in Africa and in individuals in industrialized nations [1013].
CD4 T cell count is a leading indicator of HIV-1 disease progression in an individual [14, 15] and, together with HIV-1 RNA load, is used as the basis for the initiation and monitoring of therapy [16]. The preseroconversion CD4 T cell count [4] and the rate of decline in CD4 T cell counts after infection are strong predictors of the time to AIDS and death [17, 18]. Information on CD4 T cell counts for individuals at different stages of disease progression, as available for cohorts with prevalent cases of HIV-1 infection, is used to estimate the HIV-1 incubation period and survival time [19].
Several studies have shown that HIV-1uninfected Ethiopians have lower CD4 T cell counts than does any other population in Africa or in industrialized countries. The average CD4 T cell count in HIV-1uninfected Ethiopians reportedly ranged from 591 × 106 to 775 × 106 cells/L [2025]. CD4 T cell counts ranged from 838 × 106 to 1256 × 106 cells/L elsewhere in Africa and were, for example, 1067 × 106 cells/L in Dutch homosexual men [2630]. In addition, it has been shown that the level of immune activation in Ethiopians is higher than that in Dutch individuals, even before HIV infection [24].
Since (1) the extent of immune activation during HIV infection and CD4 T cell counts are the best predictors of disease progression; (2) the level of immune activation in Ethiopians is higher than that in Dutch individuals, even before HIV-1 infection; and (3) Ethiopians start off with significantly lower CD4 T cell counts, one would expect progression of HIV-1 infection in this African population to be faster than that in the Dutch population. Here, we analyze whether the unique immunological profile of healthy Ethiopians results in a shorter survival time after HIV-1 infection in Ethiopian than in Dutch individuals.
PATIENTS AND METHODS
Study population.
The data used originated from a cohort study of HIV-1 infection and disease progression in workers from 2 factories in Ethiopia. Cohort study procedures have been described in detail elsewhere [31]. The Ethiopian Health and Nutrition Research Institute Ethics Committee and the National Ethical Clearance Committee approved the study protocol. The cohort enrolled both HIV-1uninfected and HIV-1infected factory workers between April 1997 and December 2001, for whom follow-up visits were scheduled every 6 months. Data on follow-up until June 2002 were used for the present study. During this period, antiretroviral therapy was not available to the Ethiopian cohort participants. We compared the rate of decline in CD4 T cell counts for 149 HIV-1infected Ethiopians with that for 306 HIV-1infected Dutch homosexual men and 181 HIV-1infected Dutch drug users between 1990 and 1995 (i.e., before highly active antiretroviral therapy was available in The Netherlands). The Dutch individuals were participants in the ongoing Amsterdam cohort studies [19, 32]. Follow-up visits were scheduled every 3 and 4 months for Dutch homosexual men and drug users, respectively.
Clinical examination.
Clinical information on participants in the Ethiopian cohort was obtained through a medical examination that focused on the conditions included in the World Health Organization (WHO) staging system [33]. The staging system includes a clinical axis made up of 32 conditions divided into 4 stages (stage 4 being equivalent to AIDS).
Most diagnoses in our cohort were presumptive, because of our limited access to laboratory investigations. As a result, very few HIV-1infected individuals in the cohort presented with an AIDS-defining illness during follow-up. Within a few days or weeks, individuals died from conditions presumably related to HIV-1 infection. Of the 29 deaths included in this analysis, 14 were attributable to tuberculosis, 8 were attributable to wasting syndrome, 3 were attributable to pneumonia, and 4 were attributable to unknown causes.
Laboratory analyses.
For Ethiopians, HIV-1 screening was performed by use of HIVSPOT (Genelabs Diagnostics) and ELISA (Vironostika HIV-1 Uni-Form II; Organon Teknika) kits. All positive or discrepant results were confirmed by use of a Western blot assay (HIV-1 BLOT 2.2; Genelabs Diagnostics). The absolute number of leukocytes per microliter of whole blood was obtained by use of a Coulter counter (T540). Lymphocyte subsets were determined by flow cytometry (FACScan; Becton Dickinson). For the Amsterdam cohort study, HIV-1 screening was performed by use of 2 commercially available ELISAs (Axzym [Abbot Laboratories] and Vironostika [Organon Teknika]) and confirmed by use of a Western blot assay. Lymphocyte subsets were determined by flow cytometry with a FACScan. Assessment of HIV-1 RNA load in plasma was performed on 100-L samples by use of nucleic acidbased amplification assays: for Ethiopian samples, the Nuclisens assay was used, and, for Dutch samples, the NASBA assay was used. The detection threshold for the Nuclisens assay is 80 copies/mL, and that for the NASBA assay is 1000 copies/mL. In analyses (see next subsection), all negative samples were given a value of 100 copies/mL. Analyses were repeated by use of a higher number of copies per milliliter for negative samples (1000 copies/mL), but results were similar. Values above the detection limits of the 2 assays were comparable [25]. Results are presented as log10 RNA copies per milliliter of plasma.
RESULTS
Characteristics of the study population at enrollment.
The present study included 149 HIV-1infected Ethiopians, of whom 132 were prevalent cases and 17 were incident cases (table 1). They completed a total of 791 visits, with a median of 5 visits/individual (interquartile range , 28 visits/individual) and a median of 3 years (IQR, 2.04.6 years) of follow-up. The median time between 2 successive visits was 6.3 months (IQR, 6.16.9 months).
A total of 35 deaths until June 2002 were documented, of which 6 occurred >1 year after an individual's last visit; these individuals were censored at the last visit. The incidence of death was 6.7 deaths/100 person-years. The median time between the last visit and death was 5.2 months (IQR, 3.98.2 months). Most of the deaths (n = 24 [82.8%]) occurred after the individuals were seen at CD4 stage 4 (figure 1A), and 4 individuals (13.8%) died after being seen at CD4 stage 3. Only 1 individual died after being seen at CD4 stage 2, and nobody died after being seen at CD4 stage 1. The median time from the last follow-up visit to death was 4.4 months (IQR, 2.07.6 months) for Dutch drug users and 2.6 months (IQR, 1.35.1 months) for Dutch homosexual men. The median CD4 T cell count at the last visit before death was 121 × 106 cells/L; that for Dutch drug users was 180 × 106 cells/L (IQR, 20 × 106380 × 106 cells/L), that for Dutch homosexual men was 20 × 106 cells/L (IQR, 10 × 10660 × 106 cells/L), and that for Ethiopians was 119 × 106 cells/L (IQR, 70 × 106187 × 106 cells/L). At the last visit before death, 12 individuals (40%) were at WHO stage 3 (figure 1B), whereas only 3 individuals (10%) were at WHO stage 4.
For the individuals who remained alive, the median time from the last follow-up visit until the end of follow-up was 6.3 months (IQR, 4.610.0 months) for Ethiopians, 3.0 months (IQR, 1.44.6 months) for Dutch drug users, and 1.8 months (IQR, 0.93.0 months) for Dutch homosexual men. The median duration of follow-up was 4.4 (IQR, 2.85.0), 5.9 (IQR, 2.86.0), and 6.0 (IQR, 5.16.0) years for these 3 populations, respectively. The median duration of follow-up until death was 2.1 years (IQR, 1.23.4 years) for Ethiopians, 2.9 years (IQR, 1.44.5 years) for Dutch drug users, and 3.3 years (IQR, 2.14.6 years) for Dutch homosexual men.
We compared the survival times and the decline in CD4 T cell counts for HIV-1infected Ethiopians with those for 306 HIV-1infected Dutch homosexual men and 181 HIV-1infected Dutch drug users, of whom 33 and 37 were incident cases, respectively, and who participated in the Amsterdam cohort studies. At enrollment in our study population, the 3 cohorts had similar age distributions (mean ages, 34.0, 35.1, and 32.3 years for Ethiopians, Dutch homosexual men, and Dutch drug users, respectively), but they showed differences in CD4 T cell counts, CD4 T cell percentages, CD4 : CD8 T cell ratios, and HIV-1 RNA loads. The median CD4 T cell count for Ethiopians (333 × 106 cells/L) was lower than that for Dutch drug users (480 × 106 cells/L) but was similar to that for Dutch homosexual men (370 × 106 cells/L). Also, the CD4 T cell percentage was lower for Ethiopians (20%) than for Dutch drug users (32%) and Dutch homosexual men (28%), as was the CD4 : CD8 T cell ratio, which was 30% for Ethiopians, 63% for Dutch drug users, and 50% for Dutch homosexual men. The median log10 HIV-1 RNA load was lower for Ethiopians (4.0 log copies/mL) than for Dutch drug users (4.5 log copies/mL) and Dutch homosexual men (4.4 log copies/mL).
Survival time.
Using the maximum-likelihood approach, we obtained estimates of the transition rates of moving between adjacent HIV-1 disease progression stages (table 2). We then estimated the survival probability distribution functions for the 4 different models (figure 2). The shortest estimated median time from seroconversion to death was 9.1 years, for the model based on WHO clinical stages. The model based on raw CD4 T cell counts gave a survival time of 11.0 years. Correcting for misspecification of CD4 T cell counts yielded a median survival time of 13.0 years (95% confidence interval [CI], 10.815.0 years) for model 2 and 13.7 years (95% CI, 10.718.0 years) for model 4. After applying the hidden Markov model to the Dutch study population, we obtained a median survival time of 10.5 years (95% CI, 8.810.3 years) for homosexual men and 8.4 years (95% CI, 5.810.6 years) for drug users.
Decline in CD4 T cell counts in Ethiopian versus Dutch HIV-1infected individuals.
In both Ethiopian and Dutch HIV-1infected individuals, the rate of loss of CD4 T cells was strongly dependent on the CD4 T cell count. The higher the CD4 T cell count, the more rapid the loss of CD4 T cells. Since Ethiopians had lower baseline CD4 T cell counts, they thus lost their cells more slowly than did Dutch HIV-1infected individuals. However, even when Dutch and Ethiopian individuals with similar CD4 T cell counts were compared, the annual loss of CD4 T cells was larger in the Dutch individuals (figure 3). For example, at a CD4 T cell count of 333 cells/L (the median baseline value at which the Ethiopians had their first seropositive visit), the annual loss for Ethiopians was 32 × 106 cells/L, whereas it was 68 × 106 cells/L and 79 × 106 cells/L for Dutch homosexual men and drug users, respectively. When the decline in CD4 T cell counts was compared by CD4 T cell count category (figure 4), the overall rate of decline in CD4 T cell counts for Ethiopians appeared to be significantly lower than that for Dutch homosexual men and drug users, in all CD4 T cell categories (all P < .01). Conversely, Dutch homosexual men and drug users had similar rates of decline in CD4 T cell counts, except for the last category (0200 × 106 cells/L), in which Dutch homosexual men had a somewhat faster decline, although the difference was not statistically significant (P = .088).
Since none of the HIV-1infected Ethiopians developed infection with SI/X4 HIV-1 strains, we tested whether the more rapid decline in CD4 T cell counts in the Dutch individuals could be attributed to the fraction of Dutch individuals who developed infection with SI/X4 HIV-1 (i.e., 38% of the Dutch homosexual men and 14% of the Dutch drug users) by omitting them in the analyses after SI/X4 HIV-1 strains developed. Although those Dutch homosexual men and drug users who had no evidence of SI/X4 strains had a slower decline in CD4 T cell counts than did those who developed infection with SI/X4 HIV-1 strains (figure 4), they still had a significantly higher rate of decline in CD4 T cell counts than did Ethiopians (P < .05). We also examined whether the differences in HIV-1 RNA load between Ethiopians and Dutch individuals could explain the differences in decline in CD4 T cell counts, by incorporating HIV-1 RNA load in the model as a cofactor. Although the differences in the slopes of decline in CD4 T cell counts decreased, Ethiopians still had a significantly slower loss of CD4 T cells than did Dutch homosexual men and drug users, independent of the HIV-1 RNA load and even if only Dutch individuals who had no evidence of infection with SI/X4 HIV-1 strains were considered.
DISCUSSION
For Ethiopian factory workers, the median survival time from HIV-1 seroconversion (range, 9.113.7 years) was not shorter than that for individuals in industrialized countries before the introduction of HAART (range, 813 years) [2, 19, 3840]. Considering the unique immunological profile (i.e., lower CD4 T cell counts) of Ethiopians, this finding is remarkable. We used both clinical markers and CD4 T cell counts to define stages of disease progression in Ethiopians, since these have previously been shown to be independent predictors of death in the same study population [41]. This also allowed us to make comparisons with previous studies in other populations that used either clinical data [42, 43] or CD4 T cell counts [15, 19, 32] to estimate HIV-1 incubation period and survival time. In the present study, the median survival time varied depending on the data used, but, interestingly, even the shortest survival time estimate in Ethiopians, by use of clinical staging, was within the range of previous estimates from other countries.
We analyzed why survival times were not shorter in Ethiopians despite their low baseline CD4 T cell counts. Ethiopians turned out to lose their CD4 T cells at least twice as slowly as the Dutch individuals, even when Ethiopian and Dutch individuals with similar CD4 T cell counts were compared. The annual loss in CD4 T cells in individuals in industrialized countries before the widespread use of antiretroviral therapy has been estimated to be 40 × 10680 × 106 cells/L [44, 45]. A possible cause for the slower decline in CD4 T cell counts could be the lack of SI/CXCR4-using HIV-1 strains in HIV-1 subtype C (HIV-C)infected Ethiopians [46]. Although this indeed partially accounted for the observed differences, a slower decline in CD4 T cell counts in Ethiopians was still apparent when they were compared with only those participants in the Amsterdam cohorts who were infected with only non-SI/R5 strains. The HIV-1 RNA load is known to correlate with the rate of loss of CD4 T cells. Moreover, the slower decline in CD4 T cell counts in HIV-1infected Ethiopians could not be explained by their lower HIV-1 RNA loads, compared with those of Dutch homosexual men and drug users. Notably, the slower loss of CD4 T cells in HIV-Cinfected Ethiopians has remarkable similarity with the slower loss of CD4 T cells in HIV-2 infection, which may provide some leads to investigate possible reasons for a slower loss.
Interestingly, similar to the present study, 3 other recent African studies with at least 10 years of follow-up data from Uganda, Malawi, and Tanzania did not demonstrate a shorter survival time [10, 11, 13]. These findings seem difficult to reconcile with the hyperimmune activation hypothesis. It may be that, compared with people in the Western world, Ethiopians and people living in the developing world in general have been strongly selected for the ability to survive despite chronic immune activation because of the high burden of infectious diseases. Studies have shown that, with aging, there is gradual erosion of the naive T cell pool, which is likely the effect of exposure to foreign antigens and decreasing renewal [47]. To survive longer with frequent infectious diseases, one has to have an efficient immune response but should at the same time minimize the accelerated aging of the immune system by curtailing bystander responses. In developing countries, there may have been a strong selection for that low immune activationrelated phenotype over many centuries. Notably, it is exactly this phenotype that has been shown to be associated with slow progression of HIV-1 infection in cohorts in the Western world [4, 48].
Through their participation in the cohort study, the participants had access to free treatment for their general health problems and for some of the opportunistic infections they might have had during the course of HIV-1 infection. Such a standard of care was not available for the general HIV-1infected population in the country. Therefore, one may expect better survival with HIV-1 infection in these factory workers than in the general HIV-1infected Ethiopian population. In Uganda, no difference in survival time was found between participants of a population cohort study who received intensive treatment with standard drugs from the WHO essential drug list and those with no such access [11]. If conditions are comparable with this population, our finding may well be representative of survival in the general HIV-1infected population of Ethiopia. Participants were factory workers, implying that they were all healthy enough to perform their duties in the factories when enrolled into the study, and most were still actively working. Indeed, the majority of prevalent cases of HIV-1 infection were asymptomatic. Thus, there might have been a preselection of healthy individuals for the present study. Note that in developed countries, no difference has been found in HIV-1 incubation time and decline in CD4 T cell counts between homosexual men and injection drug users, even though the latter group has a worse general health status. Hence, there is no reason to assume that preselection, if at all present, could have significantly influenced our results. Although individuals who were too sick to work in the factory (and who potentially had low CD4 T cell counts) could still continue participating in the cohort, we checked whether selective study dropout dependent on CD4 T cell counts occurred and influenced our results on differences in decline in CD4 T cell counts. In a Cox regression analysis with dropout as the end point, the relative risk for dropout at CD4 T cell counts between 200 and 350 cells/L was 1.4 (P = .49), that for dropout at CD4 T cell counts between 350 and 500 cells/L was 1.3 (P = .56), and that for dropout at CD4 T cell counts between 500 and 2000 cells/L was 0.2 (P = .15), compared with dropout at CD4 T cell counts <200 cells/L. Overall, no statistically significant effect of CD4 T cell counts was noted (P = .11), indicating that the slower decline in CD4 T cell counts in Ethiopians was not significantly biased by study dropout. Moreover, since we do know whether the individuals were still alive after loss to follow-up, a model for development of CD4 T cell counts that corrects for informative dropout due to death was fitted, and results did not change.
In conclusion, survival time in HIV-1infected Ethiopians is similar to survival time in individuals from industrialized countries before the widespread use of antiretroviral therapy. The generally low baseline CD4 T cell counts and high levels of immune activation in healthy Ethiopians do not lead to shorter HIV-1 survival times, because individuals with HIV from Ethiopia lose their CD4 T cells at a slower pace than do individuals with HIV from Western countries.
Acknowledgments
The Ethio-Netherlands AIDS Research Project is a collaborative effort of the Ethiopian Health and Nutrition Research Institute in Addis Ababa, the Municipal Health Service in Amsterdam, the Department of Human Retrovirology of the Academic Medical Center (University of Amsterdam), and the Central Laboratory of the Netherlands Red Cross Blood Transfusion Service (Sanquin). We thank Aster Tsegaye, Maria Prins, Anneke Krol, and Georgios Pollakis, for their helpful and enthusiastic discussions.
References
1. French N, Mujugira A, Nakiyingi J, Mulder D, Janoff EN, Gilks CF. Immunological and clinical stages in HIV-1infected Ugandan adults are comparable and provide no evidence of rapid progression but poor survival with advanced diseases. J Acquir Immune Defic Syndr 1999; 22:50916. First citation in article
2. Collaborative Group on AIDS Incubation and HIV-1 survival including the CASCADE EU Concerted Action. Time from HIV-1 seroconversion to AIDS and death before widespread use of a highly active antiretroviral therapy: a collaborative re-analysis. Lancet 2000; 355:11317. First citation in article
3. Bentwich Z, Kalinkovich A, Weisman Z. Immune activation is a dominant factor in the pathogenesis of African AIDS. Immunol Today 1995; 16:18791. First citation in article
4. Hazenberg MD, Otto SA, van Benthem BH, et al. Persistent immune activation in HIV-1 infection is associated with progression to AIDS. AIDS 2003; 17:18. First citation in article
5. Hazenberg MD, Hamann D, Schuitemaker H, Miedema F. T cell depletion in HIV-1 infection: how CD4+ T cells go out of stock. Nat Immunol 2000; 1:2859. First citation in article
6. Anzala OA, Nagelkerke NJ, Bwayo JJ, et al. Rapid progression of disease in Africa sex workers with human immunodeficiency virus type 1 infection. J Infect Dis 1995; 171:6869. First citation in article
7. N'Galy B, Ryder RW, Bila K, et al. Human immunodeficiency virus infection among employees in an African hospital. N Engl J Med 1988; 319:11237. First citation in article
8. Whittle H, Egboga A, Todd J, et al. Clinical and laboratory predictors of survival in Gambian patients with symptomatic HIV-1 or HIV-2 infection. AIDS 1992; 6:6859. First citation in article
9. Mann JM, Bila K, Colebunders RL, et al. Natural history of human immunodeficiency virus infection in Zaire. Lancet 1986; 2:7079. First citation in article
10. Crampin AC, Floyd S, Glynn JR, et al. Long-term follow-up of HIV-1infected and HIV-1uninfected individuals in rural Malawi. AIDS 2002; 16:154550. First citation in article
11. Morgan D, Mahe C, Mayanja B, Okongo JM, Lubega R, Whitworth JA. HIV-1 infection in rural Africa: is there a difference in median time to AIDS and survival compared with that in industrialized countries AIDS 2002; 16:597603. First citation in article
12. Morgan D, Whitworth J. The natural history of HIV-1 infection in Africa. Nat Med 2001; 7:1435. First citation in article
13. Urassa W, Bakari M, Sandstrom E, et al. Rate of decline of absolute number and percentages of CD4 lymphocytes among HIV-1infected adults in Dar es Salaam, Tanzania. AIDS 2004; 18:4338. First citation in article
14. Phillips AN, Pezzotti P, Lepri AC, Rezza G, the Italian Seroconversion Study. CD4 lymphocyte count as a determinant of the time for HIV-1 seroconversion to AIDS and death from AIDS: evidence from the Italian Seroconversion Study. AIDS 1994; 8:1299305. First citation in article
15. Longini IM Jr, Clark WS, Gardner LI, Brundage JF. The dynamics of CD4+ T-lymphocyte decline in HIV-1infected individuals: a Markov modeling approach. J Acquir Immune Defic Syndr 1991; 4:11417. First citation in article
16. 1997 revised guidelines for performing CD4+ T cell determinations in persons infected with human immunodeficiency virus (HIV). Centers for Disease Control and Prevention. MMWR Recomm Rep 1997; 46:129. First citation in article
17. Lang W, Perkins H, Anderson RE, Royce R, Jewell N, Winkelstein W Jr. Patterns of T-lymphocyte changes with human immunodeficiency virus infection: from seroconversion to the development of AIDS. J Acquir Immune Defic Syndr 1989; 2:639. First citation in article
18. Phillips AN, Lee CA, Elford J, et al. Serial CD4 lymphocyte counts and the development of AIDS. Lancet 1991; 337:38992. First citation in article
19. Hendriks JC, Satten GA, van Ameijden EJ, et al. The incubation period to AIDS in injecting drug users estimated from prevalent cohort data, accounting for death prior to an AIDS diagnosis. AIDS 1998; 12:153744. First citation in article
20. Wolday D, Tsegaye A, Messele T. Low absolute CD4 counts in Ethiopians. Ethiop Med J 2002; 40(Suppl 1):116. First citation in article
21. Pollack S, Fuad B, Etzioni A. CD4 T-lymhocytopenia without opportunistic infection in HIV-1uninfected Ethiopian immigrants to Israel. Lancet 1993; 342:50. First citation in article
22. Kalinkovich A, Weisman Z, Burstein R, Bentwich Z. Standard values of T-lymphocyte subsets in Africa. J Acquir Immune Defic Syndr Hum Retrovirol 1998; 17:1835. First citation in article
23. Kassa E, Rinke de Wit TF, Hailu E, et al. Evaluation of the World Health Organization staging system for HIV-1 infection and disease in Ethiopia: association between clinical stages and laboratory markers. AIDS 1999; 13:3819. First citation in article
24. Messele T, Abdulkadir M, Fontanet AL, et al. Reduced naive and increased activated CD4 and CD8 cells in healthy adult Ethiopians compared with their Dutch counterparts. Clin Exp Immunol 1999; 115:44350. First citation in article
25. Rinke de Wit TF, Tsegaye A, Wolday D, et al. Primary HIV-1 subtype C infection in Ethiopia. J Acquir Immune Defic Syndr 2002; 30:46370. First citation in article
26. Hazenberg MD, Otto SA, Cohen Stuart JW, et al. Increased cell division but not thymic dysfunction rapidly affects the T cell receptor excision circle content of the naive T cell population in HIV-1 infection. Nat Med 2000; 6: 103642. First citation in article
27. Tugume SB, Piwowar EM, Lutalo T, et al. Hematological reference ranges among healthy Ugandans. Clin Diagn Lab Immunol 1995; 2:2335. First citation in article
28. Levin A, Brubaker G, Shao JS, et al. Determination of T-lymphocyte subsets on site in rural Tanzania: results in HIV-1infected and non-infected individuals. Int J STD AIDS 1996; 7:28891. First citation in article
29. Zekeng L, Sadjo A, Meli J, et al. T-lymphocyte subsets values among healthy Cameroonians. J Acquir Immune Defic Syndr Hum Retrovirol 1997; 14:823. First citation in article
30. Embree J, Bwayo J, Nagelkerke N, et al. Lymphocyte subsets in human immunodeficiency virus type 1infected and uninfected children in Nairobi. Pediatr Infect Dis J 2001; 20:397403. First citation in article
31. Mekonnen Y, Sanders E, Aklilu M, et al. Evidence of changes in sexual behaviors among male factory workers in Ethiopia. AIDS 2003; 17:22331. First citation in article
32. Hendriks JC, Satten GA, Longini IM, et al. Use of immunological markers and continuous-time Markov models to estimate progression of HIV-1 infection in homosexual men. AIDS 1996; 10:64956. First citation in article
33. The WHO International Collaborating Group for the Study of the WHO Staging System. Proposed "World Health Organization staging system for HIV-1 infection and disease": preliminary testing by an international collaborative cross-sectional study. AIDS 1993; 7:7118. First citation in article
34. Sypsa V, Touloumi G, Kenward M, Karafoulidou A, Hatzakis A. Comparison of smoothing techniques for CD4 data in a Markov model with states defined by CD4: an example on the estimation of the HIV-1 incubation time distribution. Stat Med 2001; 20:366776. First citation in article
35. Satten GA, Longini, IM. Markov chains with measurement error: estimating the "true" course of a marker of the progression of human immunodeficiency virus disease (with discussion). Appl Stat 1996; 45:275309. First citation in article
36. Ihaka R, Gentleman R. R: a language for data analysis and graphics. J Comp Graph Stat 1996; 5:299314. Available at: http://www.r-project.org. Accessed 18 July 2005. First citation in article
37. Taylor JMG, Cumberland WG, Sy JP. A stochastic model for analysis of longitudinal AIDS data. J Am Stat Assoc 1994; 89:72736. First citation in article
38. Multicohort Analysis Project Workshop. Immunological markers of AIDS progression: consistency across five HIV-1infected cohorts, part 1. AIDS 1994; 7:91121. First citation in article
39. Pehrson P, Lindback S, Lidman C, Gaines H, Giesecke J. Longer survival after HIV-1 infection for injecting drug users than for homosexual men: implications for immunology. AIDS 1997; 11:100712. First citation in article
40. Koblin BA, van Benthem BH, Buchbinder SP, et al. Long-term after infection with human immunodeficiency virus type 1 (HIV-1) among homosexual men in hepatitis B vaccine trial cohorts in Amsterdam, New York City, and San Francisco, 19871995. Am J Epidemiol 1999; 150:102630. First citation in article
41. Mekonnen Y, Dukers NH, Sanders E, et al. Simple markers for initiating antiretroviral therapy among HIV-1infected Ethiopians. AIDS 2003; 17:8159. First citation in article
42. Schechter MT, Le N, Craib KJ, Le TN, O'Shaughnessy MV, Montaner JS. Use of the Markov model to estimate the waiting times in a modified WHO staging system for HIV-1 infection. J Acquir Immune Defic Syndr Hum Retrovirol 1995; 8:4749. First citation in article
43. Longini IM, Clark WS. Statistical analysis of the stages of HIV-1 infection using Markov model. Stat Med 1989; 8:83143. First citation in article
44. Munoz A, Carey V, Saah AJ, et al. Predictors of decline in CD4 lymphocytes in a cohort of homosexual men infected with human immunodeficiency virus. J Acquir Immune Defic Syndr 1988; 1:396404. First citation in article
45. Easterbrook PJ, Emami J, Gazzard B. Rate of CD4 cell decline and prediction of survival in zidovudine treated patients. AIDS 1993; 7:95967. First citation in article
46. Abebe A, Demissie D, Goudsmit J, et al. HIV-1 subtype C syncytium- and non-syncytium-Ethiopian patients with AIDS. AIDS 1999; 13:130511. First citation in article
47. Fagnoni FF, Vescovini R, Passeri G, et al. Shortage of circulating naive CD8+ T cells provides new insights on immunodeficiency in aging. Blood 2000; 95:28608. First citation in article
48. Silvestri G, Sodora DL, Koup RA, et al. Nonpathogenic SIV infection of sooty mangabeys is characterized by limited bystander immunopathology despite chronic high-level viraemia. Immunity 2003; 18:44152. First citation in article