AIDS RESEARCH AND HUMAN RETROVIRUSES Volume 18, Number 16, 2002, pp. 1167–1173 © Mary Ann Liebert, Inc.

Low Level Viremia and High CD4% Predict Normal Survival in a Cohort of HIV Type-2-Infected Villagers NEIL BERRY,1,2 SHABBAR JAFFAR,3 MAARTEN SCHIM VAN DER LOEFF,4 KOYA ARIYOSHI,4,5 ELIZABETH HARDING,4 PA TAMBA N’GOM,4 FRANCISCO DIAS,6 ANDREW WILKINS,4 DOMINIC RICARD,1 PETER AABY,7 RICHARD TEDDER,1 and HILTON WHITTLE4

ABSTRACT A community-based study of human immunodeficiency virus type 2 (HIV-2) infection was conducted in a rural village in northern Guinea Bissau, West Africa to assess the relationship between plasma HIV-2 RNA levels, CD4 lymphocyte percentage, and survival over an 8-year period. The cohort of 133 HIV-2-infected individuals and 160 HIV-uninfected controls enrolled in 1991 were followed up at home until 1998. Thirty-one (23%) HIV-2-infected and 24 (16%) HIV-uninfected individuals died over the follow-up period (mortality hazard ratio 1.7, 95% CI 1.0, 2.9; p 5 0.06). In HIV-2-infected individuals, the median HIV-2 RNA level was 347 copies/ml and the mean CD4% was 28.6. Both plasma viremia and CD4% were independent predictors of survival, with hazard ratios increasing by 1.6 (95% CI, 1.1, 2.3) for each log10 increase of plasma viremia and 1.7 (1.1, 2.6) for each 10% decrease of CD4%. Infected subjects with a plasma viral load $ the median (347 copies/ml) and a CD4% # the mean (28.6%) had a mortality hazard ratio of 3.1 (95% CI 1.7, 5.8) compared to uninfected controls, whereas the remaining infected subjects had a mortality rate similar to uninfected controls, the mortality hazard ratio being 1.0 (95% CI, 0.5, 2.1.) In those who survived between 1991 and 1996, HIV-2 RNA levels were unchanged overall and CD4 lymphocyte counts remained high. In conclusion, baseline HIV-2 RNA levels predicted a normal survival for the majority, with low and stable levels of plasma viremia characterizing HIV-2 infections in this rural West African community.

INTRODUCTION

H

UMAN LENTIVIRAL INFECTION ,

which today affects over 34 million persons, is caused by two related but distinct viruses, human immunodeficiency viruses types 1 and 2 (HIV1 and HIV-2), which differ in their natural history, level of viremia, transmission rates, and disease associations.1–3 In Guinea-Bissau the prevalence of HIV-2 among adults is estimated to be about 8%.3 A cohort study conducted in the capital, Bissau Town, showed that the mortality rate among HIV2-infected subjects was 2.6 per 100 person-years, which was just over twice that among HIV-uninfected individuals.4 In con-

trast, a number of community-based studies from East Africa suggest that the mortality rate among HIV-1-infected individuals is over 10-fold greater than that of HIV-uninfected individuals.5,6 The factors that influence survival in HIV-infected African people and the reasons why survival appears to be much better in HIV-2 than in HIV-1 infections are not fully understood. In studies of HIV-1 conducted in developed countries, plasma viremia and CD4 cell count are important predictors of disease progression and survival.7–9 They are now central to the effective management and the counseling of HIV-1-infected persons and are often used as primary end points in trials eval-

1 Department

of Virology, Royal Free and University College Medical School, London, UK. of Retrovirology, National Institute for Biological Standards and Control, South Mimms, Herts, UK. 3 Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, UK. 4 Medical Research Council Laboratories, Banjul, The Gambia, West Africa. 5 AIDS Research Centre, National Institute of Infectious Diseases, Tokyo, Japan. 6 Laboratório Nacional de Saude Publica, MINSAP, Bissau, Guinea-Bissau, West Africa. 7 Department of Epidemiology Research, Danish Epidemiology Centre, Statens Serum Institut, Copenhagen, Denmark. 2 Division

1167

1168

BERRY ET AL.

uating new treatment regimens, rather than clinical end points. Whether these variables will have similar prognostic value in African communities where the spectrum of HIV-associated disease, the exposure to pathogens, access to health care and a variety of other factors all differ from developed countries is unknown. To date there have been no community-based studies, either HIV-1 or HIV-2, that have related plasma viral load to survival in Africa. Thus we report our findings of a cohort study of HIV-2 infection in a rural region of Guinea-Bissau where access to health care is limited and examined the association between HIV-2 plasma viremia, CD4%, and mortality.

MATERIALS AND METHODS Study population The study was conducted in Caio, a remote village in northwest Guinea-Bissau.10 The village has a population of approximately 10,000 dispersed among several main settlements on tracks that run through several miles of woodland. Rice growing, cashew nuts, and palm oil and palm wine production are the main economic activities of the area. The adult sex ratio is unbalanced, since many men leave to work elsewhere in Guinea-Bissau, neighboring countries in West Africa, or Europe. Women may also leave in search of paid domestic work, or to trade, or to work as commercial sex-workers in the region’s urban centers. Annual censuses have shown that the proportion of adults who were claimed to be residents but were absent at the time varies between 30 and 40% according to the season of the year. Following informed consent, HIV testing of subjects aged 15 years and over who were considered residents was started in May 1989 and continued until June 1991 to allow for inclusion of subjects who were away at the start of the survey. Overall 220 (7.9%) of the 2770 subjects tested HIV-2 seropositive. Seroprevalence was 9.3% in women, peaking at 17.2% in the 35–44 year age group, and 6.6% in men, peaking at 19.1% in the 45–54 age group.10 Age and sex-matched HIV seronegative controls were selected from the same area of the community as each of the infected subjects. Between March and June 1991, 133 of the HIV-2 subjects identified from the original 220 cases and 160 controls were recruited into a cohort study.11 The remaining 87 cases and 60 controls were away from the village at the time, or had died, or were unwilling to participate. The mean age (range) of the seven cases who had died was 43 (30–68) years; such data were not documented for the controls. In 1991 one of the HIV-2 cases was found to be infected with both HIV-1 and HIV-2 and one of the seronegative controls became HIV-2 positive; both were excluded from the study. Blood was taken for CD41 and viral load measurements. Subjects who were largely asymptomatic were examined clinically and generalized lymphadenopathy was the only clinical finding associated with HIV-2 infection; three cases and one control had signs and symptoms of AIDS.11 Both cases and controls had access to the project doctors and were visited at home by a field worker every 3 to 6 months. Free care was offered in accordance with local and national guidelines when they presented to the clinic, including treatment for general in-

fections. None of the study subjects received any form of antiretroviral treatment. A complete census of all villagers was conducted yearly and in addition between April and October 1996 study subjects were traced and a second venous blood specimen taken for the determination of levels of HIV-2 plasma viremia and CD4%. From February 1 to September 7 1998, a further field census involving home visits was conducted to establish the survival status of the study subjects. Throughout the studies, field workers and doctors involved with the field studies did not know the participant’s HIV status. In most cases the exact cause of death could not be ascertained. The study was approved by the National AIDS Committee, Guinea Bissau and The Gambia Government/MRC Joint Ethics Committee. All subjects were counseled about HIV before informed consent was obtained. Test results were available on request from a counselor who was independent of the clinician and the field workers.

Laboratory procedures Plasma samples were collected and within 3 hr stored in liquid nitrogen. HIV-2 plasma viremia was measured using a quantitative RNA PCR modified from a previously published technique. 12 In brief, RNA was extracted according to the method of Boom et al.13 and reverse transcriptase polymerase chain reaction (RT-PCR) was performed, in duplicate, using HIV-2specific long-terminal repeat (LTR) oligonucleotide sequences in a single-tube RT-PCR assay (Titan reagents, Roche Diagnostics). Products of amplification were quantified by chemiluminescence. The reliable level of sensitivity for quantification in the assay was 100 copies/ml, though additional sensitivity, estimated to be 50 copies/ml by Poisson end point distribution of RNA molecules, was obtained by qualitative nested PCR visualized on agarose gels. Samples that failed to signal in the PCR assays were assigned an arbitrary value of 20 copies per ml. Plasma samples from 1991 and 1996 samples were tested together for each individual in the same assay run. FACS analysis and serological investigations on cases and controls were as previously described.11 HIV-1- and HIV-2specific PCR assays were used to confirm or exclude serological dual reactions as previously described.14 We chose to analyze the CD41 cell measurements by percentage as lymphocyte counts, which had to be done manually, were subject to considerable observer variation; in addition a number of slides were damaged due to an accident in transit, thus allowing no possibility of an estimate of lymphocyte count.

Statistical analyses Proportions were compared with a x2 test or Fisher’s exact test if numbers were small. Continuous data were analyzed using a t test if normally distributed, with correlation assessed using Pearson’s correlation coefficient, and by a Wilcoxon test if not normally distributed, with correlation assessed using Spearman’s correlation coefficient. Rates of survival were analyzed by Kaplan–Meier methods, and Cox regression, with timesince-entry used as the underlying time. Subjects lost to followup were censored at the time that they were last ascertained to be alive. The analysis was conducted using the SAS System for Windows and Stata.

LOW LEVEL HIV-2 VIREMIA AND SURVIVAL

1169

RESULTS Baseline characteristics in 1991 of HIV-2-infected individuals and controls Data were available from 133 HIV-2-infected individuals and 160 control subjects recruited in 1991 (Table 1). The median age was 43 (range 16, 88) years and mean CD4% 28.6 [standard deviation (SD) 10.0]. The median plasma viremia was 347 (range 20–286,000) copies/ml. Plasma viremia was undetectable (below 50 copies/ml) in 19/130 (14.6%) of HIV-2 infections. Median age and mean CD4% were similar for men and women. Median plasma viremia was significantly higher in men, 755 (range 20–286,000) copies/ml compared to 137 (range 20–127,000) copies/ml, p 5 0.04) in women. This was reflected in the plasma viremia being 100 copies/ml or below in 10/42 (24%) men compared with 38/88 (43.2%) women (p 5 0.032). Seven subjects became infected with HIV-1 over the course of the study. The distribution of age, sex, CD4%, and plasma viral load at baseline of these individuals are similar to that of those who remained uninfected by HIV-1. Forty-eight subjects were not retested in 1996: 26 had died, 7 were lost to follow-up, and 15 were alive but away from village at the time of the resurvey. Of the 160 control subjects recruited in 1991, 3 were found to be infected with HIV-1 in 1996, 3 had acquired HIV-2 infection, 3 had acquired dual infection, 42 were not tested (17 had died, 7 had been lost to follow-up, and 18 were alive but away at the time of survey), and 109 were known to have remained HIV uninfected throughout. All seroconverters

were retained in the analyses of survival and classified according to their baseline HIV status. The median age of control subjects in 1991 was 43 (range 16, 80) years; 117 (73%) were women and their mean CD4% was 34.5 (SD 8.6). In HIV-2 subjects where both CD4% and plasma viremia were measured, neither was significantly associated with age (r 5 20.17; p 5 0.06 and r 5 0.06; p 5 0.5, respectively). However, a weak inverse association existed between CD4% and plasma viremia (r 5 20.29, p 5 0.001), the median number of virus copies/ml being 15,110 (range 300, 56,000), 1056 (20, 127,000), and 134 (20, 286,000) in the CD4% categories ,14, 14–28, and .28, respectively (p 5 0.01, Kruskal–Wallis test). The proportion of HIV-2-infected individuals with low or undetectable plasma viral load (#100 copies/ml) was 0, 29, and 43% for the above categories of CD4%, respectively (p 5 0.01, x2 test for trend).

Survival of HIV-2-infected subjects and controls Survival data for HIV-2-infected individuals (n 5 133) and uninfected controls (n 5 160) are shown in Table 1. The overall median follow-up time of HIV-2-infected individuals was 7.0 years (range 0.40, 7.4). Thirty-one (23%) died during the study period, 95 (71%) survived, and 7 (5%) were lost to follow-up. Control subjects were followed for a median 7.2 years (range 1, 7.7), during which time 24 (16%) died, 129 (80%) survived, and 7 (4%) were lost to follow-up. The crude mortality rates of the HIV-2-infected and control subjects were 3.9 (95% CI 2.5, 5.3) and 2.3 (95% CI 1.4, 3.3) per 100 person

TABLE 1. BASELINE CHARACTERISTICS OF 133 HIV-2-INFECTED SUBJECTS 160 CONTROLS STUDIED IN 1991 AND THEIR RELATION TO SURVIVAL

AND

HIV-2 infected

Number died (%) Age ,30 years 30–39 40–49 50–59 60–69 $70 Sex, (%) Women Men CD4% ,14 14–28 .28 Plasma viremia, copies/ml (%) #100 101 to 1,000 1001 to 10,000 10,000 to 100,000 100,001 to 1,000,000 $100

1/18 4/34 8/33 6/21 5/15 7/12

(6) (12) (33) (29) (33) (58)

Controls

Mortality rate per 100 person years 95% CI 0.9 1.8 4.1 4.5 5.9 12.7

(0, 2.6) (0, 3.7) (1.2, 6.9) (0.8, 8.2) (0.6, 11.2) (3.1, 22.3)

21/91 (24) 10/42 (23)

3.8 (2.1, 5.4) 4.1 (1.5, 6.7)

4/6 (67) 16/56 (29) 9/63 (14)

16.4 (0, 32.7) 4.8 (2.4, 6.9) 2.3 (0.8, 3.8)

7/48 7/34 6/26 9/20 1/2 23/82

(15) (21) (23) (45) (50) (28)

2.2 3.2 3.4 3.8 9.1 4.8

(0.5, (0.6, (0.8, (0.7, (3.0, (2.8,

3.9) 5.8) 6.0) 7.0) 15.1) 6.9)

Number died (%) 1/27 3/39 5/38 4/27 7/18 5/11

(4) (8) (13) (15) (39) (45)

11/43 (26) 13/117 (11)

Mortality rate per 100 person years 95% CI 0.5 1.2 1.6 2.3 6.1 8.3

(0, (0, (0, (0, (0, (0,

1.6) 2.6) 3.2) 4.6) 10.6) 15.7)

4.3 (1.7, 7.0) 1.7 (0.7, 2.6)

1170

BERRY ET AL.

years, respectively. Age at recruitment was associated strongly with survival in both groups; for each 10 year increment in age, the rate ratio increased similarly and linearly by a multiplicative factor of 1.5 (95% CI 1.2, 1.9; p 5 0.0002) among cases and 1.7 (95% CI 1.3, 2.3; p , 0.0001) among the controls. However, after adjusting for age and gender, the mortality rate among the HIV-2-infected subjects was higher but not significantly different from that among control subjects (hazard ratio 1.7, 95% CI 1.0, 2.9; p 5 0.06).

Association between age, sex, CD4%, and plasma viremia at recruitment and survival Age, CD4%, and plasma viremia were associated significantly with survival (Table 1 and Fig. 1); gender was not. The poorest outcome was in subjects with plasma viremia of 10,000 copies/ml or greater. Overall, the rate ratio increased linearly by a multiplicative factor of 2.1 (95% CI 1.4, 3.2; p 5 0.002) for each 10% decrease in CD4%, and by 1.7 (95% CI, 1.2, 2.3; p 5 0.002) for log10 increase in copies per ml of plasma viremia. There was no evidence of a nonlinear increase in rate ratios in either case. Subjects whose plasma viremia was #100 copies/ml were more likely to survive than those in whom this was in excess of 100 copies/ml [rate ratio 2.2, (95% CI 0.96, 5.2); p 5 0.06]. Those who died had a lower CD4% and a higher plasma viremia than those who survived. In those who died their mean CD4% was 22.9 (SD 10.3) compared with 30.3 (SD 9.3, p 5 0.0004) and their median plasma viremia was 1640 (range 20, 127,000) copies/ml compared with 207 (range 120, 28,600) copies/ml, respectively (p 5 0.007).

II, III, and IV was 2.2, which was similar to the mortality rate for HIV-2-uninfected controls, which was 2.3 per 100 personyears. In a Cox regression model, after adjusting for age and sex, the mortality hazard ratios in quadrant I and in quadrants II, III, and IV combined compared to HIV-uninfected controls were 3.1 (95% CI 1.7, 5.9) and 1.0 (95% CI 0.5, 2.1), respectively. In the same model, the hazard ratio of quadrants I compared with the other three quadrants was 3.1 (95% CI 1.5, 6.5).

Interaction between CD4%, plasma viremia and survival

Multivariate analysis of CD4%, plasma viremia and survival

The combined interaction between CD4%, plasma viremia, and survival was examined after stratification according to their mean and median values, respectively (Fig. 2). The mortality rate of subjects with a CD4% below the mean and a plasma viremia above the median (quadrant I) was 8.0 per 100 personyears, which was higher than the mortality rates in quadrants II, III, and IV, which were 2.6, 2.7, and 1.7 per 100 personyears, respectively. The combined mortality rate for quadrants

FIG. 2. HIV-2 plasma RNA levels and CD4% in 1991 available for 122 HIV-2-infected individuals, stratified according to their median and mean values of 347 copies/ml and 28.6%, respectively. Deaths (n 5 28) are denoted by solid symbols.

There were 122 subjects from whom there were complete data in 1991 for CD4% and plasma viremia, of whom 28 died. The hazard ratio for each 10 yearly increase in age (after adjustment for plasma viremia and CD4%) was 1.5 (95% CI 1.2, 1.9; p 5 0.0004), which was similar to that of controls, their hazard ratio being 1.7 (95% CI 1.3, 2.3) per 100 person years. In a Cox regression model controlling for age, both CD4% and plasma viremia were independently and significantly associated with survival. The hazard ratio increased by 1.6 (95% CI 1.1, 2.3) and 1.7 (95% CI 1.1, 2.6) for each log increase in plasma viral load and each 10% decrease in CD4%, respectively. Gender was not associated with survival after adjusting for these factors (p 5 0.7).

Changes in CD4% and plasma viremia between 1991 and 1996

FIG. 1. Kaplan–Meier survival estimates according to baseline levels of HIV-2 plasma viremia (copies/ml) available for 130 HIV-2-infected individuals at recruitment.

Among the 69 survivors who were tested for CD4 count and were HIV-2 positive on both occasions, there was a trend for higher CD4 percentages in 1996 compared to 1991, the mean CD4% being 32.1 (SD 8.9) and 29.9% (SD 9.1), respectively. These changes were not significant in any of the three categories of CD4% or between categories and did not differ from those of the controls (data not shown). In 76 subjects in whom plasma viremia was tested and who were confirmed as HIV-2 infected on both occasions, median HIV-2 RNA levels were the same in 1996 as in 1991, the values being 170 (interquartile range 50, 1313) and 150 (interquartile range 20, 1325)

LOW LEVEL HIV-2 VIREMIA AND SURVIVAL copies/ml, respectively. Changes did not vary significantly within the three strata #100, 101–1000, and $1000 copies/ml or between strata. Five of the 31 subjects who died had plasma viremia levels measured twice. These were below the limits of detection at both time points in two and increased by 31,600, 158,000, and 800 copies/ml in the other three.

DISCUSSION This study is the first report to relate survival of HIV-2-infected individuals living in a community to the level of plasma HIV-2 RNA and CD4 percentage. Our previous study focused on a selected group of patients attending clinics in a hospital, many of whom were sick with symptoms of AIDS.15 Hence, we extended our studies of a cohort of HIV-2-infected villagers in a rural part of Guinea Bissau,11 where the overall prevalence is high but individuals are largely asymptomatic. The median baseline level of plasma viremia of these subjects was low (347 copies/ml) with 37% of individuals having levels of HIV-2 RNA of 100 copies/ml or below, which were associated with high CD4 percentages. The mortality rate of HIV-2-infected subjects was 3.7 per 100 person-years, which was just 1.7-fold greater than that of HIV-uninfected controls and was broadly similar to the findings from an urban study conducted in Bissau Town. 4 In both communities large numbers of HIV-2-infected persons with a low plasma viral load may have a near normal life-span, contrasting sharply with HIV-1 infections in East Africa, where the mortality rate of infected individuals is at least 10-fold greater than that of HIV-uninfected individuals.5,16 Our findings also show that the mortality rate of HIV-2-infected subjects in the community is substantially lower than those reported from hospital-based studies, which has indicated that the mortality in HIV-2-infected subjects to be only slightly less than that of HIV-1-infected subjects.1 This may reflect the selection bias of studying subjects presenting to hospital who are more likely already to be ill due to HIV disease, whereas our findings reported here are more representative of HIV-2 infections in the broader community. In a cohort of sex workers in Senegal with known seroconversion dates, the incidence of AIDS was six times lower in HIV-2- than HIV-1-infected women. 2 Although we knew neither the time of infection nor the cause of death among most of the study subjects, their high average age renders it likely that some of the deaths were not related to HIV, a factor that will have diluted the association between prognostic markers and survival. However, there was no direct evidence of a differential effect of age on survival between HIV2-infected subjects and controls, which is interesting as age is a crucially important determinant of prognosis in HIV-1-infected subjects.17 Hence, the notion that our cohort represents a highly selected group of elderly long-term nonprogressors in which those with progressive disease would have died at a younger age seems unlikely. Moreover, it has been shown that HIV-2 infection occurs at an older age in women,18 which suggests that the difference in virulence between the two infections may be even more marked, given that elderly HIV-1-infected subjects have a very high mortality rate. Despite these factors, we were able to demonstrate for the

1171 first time in a community-based study of HIV-2 infection that plasma viremia and CD4% were associated significantly with survival after adjusting for age. Both markers were found to be independent predictors of disease progression, but the accuracy of prognosis was strengthened if both were used, a finding that has implications for trial design and monitoring of therapies in HIV-2-infected subjects. Gender was not associated with survival, both sexes having similar survival rates despite a significantly lower plasma viremia among women then men. Similar findings have been reported for HIV-1 infections in developed countries.19 The low level of expression of virus, reflected by low plasma viremia, is likely to be a critical feature of HIV-2 infections, which relates to reduced virulence and prolonged survival. However, levels of proviral DNA are similar to those found at comparable stages of HIV-1 infection,12,20–23 yet the plasma viremia is typically 30- to 40-fold less in HIV-2 infections during the asymptomatic period.24–26 The reasons need defining: perhaps HIV-2 provirus is under tighter transcriptional control than HIV-1 during the early stage of infection leading to a lower level of infectious virus,12,27 or perhaps HIV-2 less readily evades the cellular immune response and is more likely to be contained within cells.28 Whatever the reason, a low level of plasma viremia is the hallmark of asymptomatic HIV-2 infection. Overall, the level of HIV-2 viremia, which remained largely unchanged over the 5 years in those who survived, was much lower than that found in comparable studies of HIV-1 infection. This observation is compatible with the findings of Andersson et al.26 who have described a lower plasma virus set point in HIV-2 infection and a stable level of viremia in the 2 years following seroconversion. We observed relatively stable levels of plasma viremia, which remained below 1000 copies/ml in a very high proportion of those who survived, and which is likely to be extended over prolonged periods. However, these data do not preclude a rise in viral load as a near terminal event in those who died as we have already observed high HIV-2 RNA levels in patients with severe HIV-2-related disease.12 CD4 levels also remained constant during the study period in survivors. This finding, also noted in another HIV-2 cohort,29 contrasts markedly with descriptions of HIV-1 infection where the CD41 count falls inexorably over time in the majority of patients.30,31 Thus, the long period of low level and relatively stable viremia that was found in a considerable proportion of subjects seems likely to explain the improved survival and reduced pathogenicity of the majority of HIV-2 infections and the low sexual and mother-to-child transmission rates that are typical of HIV-2 infection.17 Differences in viral genotype, but not subtype, have been linked to mortality in this cohort in Caio,32 though why the balance between virus and immune response is maintained in the majority but is disturbed in a minority who fared badly is unclear. Certainly increased replicative capacity is likely to result in more pathogenic variants33 and host genetic differences such as HLA or cell receptor polymorphisms may have also played a role in regulating the relationship between virus and host. Hence, infection with, or emergence of, a more pathogenic virus in a subject with a genetically determined less efficient immune response may result in higher levels of viral replication, increased chances of transmission, and more rapid progression of disease and death. Identification of such subjects with a demon-

1172

BERRY ET AL.

strable poor prognosis is important for targeting with antiviral therapy as the majority of HIV-2-infected individuals with a low plasma viral load and a high CD4 count probably do not need treatment. It seems reasonable to speculate that low but sustained levels of virus may be responsible for stimulating effective immunological responses, which may confer resistance to disease. Thus intervention strategies such as those based on immunization with DNA vaccines that in macaques have suppressed viral replication and stabilized CD41 T cell counts34 may provide sustained clinical benefit. In conclusion, in this community-based study, we have demonstrated that HIV-2 plasma viral load and CD4% are independently associated with prognosis, findings that have important implications for trial design and for monitoring therapies in HIV-2-infected patients. Low levels of plasma viremia coupled with a well-preserved CD4% count are an underlying feature of HIV-2 infections, as seen in the community, which reflect the ability of the majority to survive for prolonged periods with this human lentivirus infection.

ACKNOWLEDGMENTS We are grateful to the subjects participating in the studies over the years. We also thank Alhajie Bayang, Mamady Njie, and Bakary Sanneh for laboratory work; Bryan Savage, Glyn Taylor, Stephanie Constantine, Chadwick Anderson, Tim Vincent, and Margaret Buckner for field work; Carlos Da Costa, Anders Naucler, and Olav Larsen for clinical care; Aveika Akum for data management; Mamae Barbosa (Ministerio de Saude Publica, Bissau) for her support; and Peter Langfield for transport and Mbemba Ceesay for his driving and negotiating skills. This work was supported by the UK Medical Research Council (in part Grant G9632130) and a gift from Glaxo-Wellcome.

REFERENCES 1. Whittle HC, Morris J, Todd J, et al.: HIV-2-infected patients survive longer than HIV-1-infected patients. AIDS 1994;8:1617 – 1620. 2. Marlink R, Kanki P, Thior I, et al.: Reduced rate of developmen t after HIV-2 infection as compared to HIV-1. Science 1994; 265:1587–1590. 3. Schim van der Loeff M and Aaby P: Towards a better understanding of the epidemiology of HIV-2. AIDS 1999;13(Supp l A):S69–84. 4. Poulsen AG, Aaby P, Larsen O, et al.: 9-year HIV-2 associated mortality in an urban community in Bissau, West Africa. Lancet 1997;349:911 –914. 5. Morgan D, Maude GH, Malamba SS, et al.: HIV-2 disease progression and AIDS defining disorders in rural Uganda. Lancet 1997;350:245 –250. 6. Boerma JT, Nunn AJ, and Whitworth JA: Mortality impact of the AIDS epidemic: Evidence from community studies in less developed countries. AIDS 1998;12(Suppl 1):S3–14. 7. Mellors JW, Rinaldo CR, Gupta P, White R, Todd JA, and Kingsley LA: Prognosis of HIV-1 infection predicted by the quantity of virus in plasma. Science 1996;272:1167 –1170. 8. Loveday C and Hill A: Prediction of progression to AIDS with serum HIV-1 RNA and CD4 count. Lancet 1995;345:790 –791.

9. O’Brian WA, Hartigan PM, Martin D, et al.: Changes in plasma HIV-1 RNA and CD41 lymphocyte counts and the risk of progression to AIDS. N Engl J Med 1996;334:426 –431. 10. Wilkins A, Ricard D, Todd J, Whittle H, Dias F, and Paulo Da Silva A: The epidemiology of HIV infection in a rural area of Guinea Bissau. AIDS 1993;7:1119 –1122. 11. Ricard D, Wilkins A, N’Gum PT, et al.: The effects of HIV-2 infection in a rural area in Guinea Bissau. AIDS 1994;8:977 –982. 12. Berry NJ, Ariyoshi K, Jaffar S, Sabally S, Corrah T, Tedder RS, and Whittle HC: Low peripheral blood viral HIV-2 RNA in individuals with high CD4 percentage differentiates HIV-2 from HIV1 infection. J Hum Virol 1998;1:457 –468. 13. Boom R, Sol CJ, Salimans M, Jansen C, Wertheim-van Dillen P, and van der Noorda J: Rapid and simple methods for purification of nucleic acids. J Clin Microbiol 1990;28:495 –503. 14. Ariyoshi K, Cham F, Harding E, et al.: Diagnosis of HIV-1/2 dual infection using dilutional analysis of type-specific antibody. AIDS 1998;12:2504 –2505. 15. Ariyoshi K, Jaffar S, Alabi A, Berry N, Schim van der Loeff M, Sabally S, Tamba N’Gom P, Corrah T, Tedder R, and Whittle H: Plasma RNA viral load predicts the rate of CD4 T cell decline and death in HIV-2-infected patients in West Africa. AIDS 2000; 14:339–344. 16. Todd J, Balira R, Grosskurth H, et al.: HIV-associated adult mortality in a rural Tanzanian population. AIDS 1997;11:801 –807. 17. Collaborative Group on AIDS Incubation and HIV Survival: Time from HIV-1 seroconversion to AIDS and death before widespread use of highly-active antiretroviral therapy: A collaborative re-analysis. Lancet 2000;355:1131 –1137. 18. Aaby P, Ariyoshi K, Buckner M, et al.: Age of wife as a major determinant of male to female transmission of HIV-2 infection: A community study from rural West Africa. AIDS 1996;10: 1585–1590. 19. Farzadegan H, Hoover DR, Astemborski J, et al.: Sex differences in HIV-1 viral load and progression to AIDS. Lancet 1998; 7:352:1510–1514. 20. Ariyoshi K, Berry N, Wilkins A, et al.: A community-based study of HIV-2 proviral load in a rural village in West Africa. J Infect Dis 1996;173:245 –248. 21. Berry N, Ariyoshi K, Jobe O, et al.: HIV type 2 proviral load measured by quantitative polymerase chain reaction correlates with CD41 lymphopaenia in HIV type-2-infected individuals. AIDS Res Hum Retroviruses 1994;10:1031 –1037. 22. Sarr A, Popper S, Thior I, et al.: Relationship between HIV-2 proviral load and CD41 lymphocyte count differs in HIV monotypic and dual infections. J Hum Virol 1999;2:45–51. 23. Norrgren H, Marquina S, Leitner T, et al.: HIV-2 genetic variation and DNA load in asymptomatic carriers and AIDS cases in Guinea Bissau. J AIDS Hum Retrovirol 1997;16:31 –38. 24. Popper SJ, Dieng Sarr A, Travers K, et al.: Lower human immunodeficiency (HIV) type 2 viral load reflects the difference in pathogenicity of HIV-1 and HIV-2. J Infect Dis 1999;180:1116 –1121. 25. O’Donovan D, Ariyoshi K, Milligan P, et al.: Maternal plasma viral RNA levels determine marked differences in mother-to-child transmission rates of HIV-1 and HIV-2 in The Gambia. AIDS 2000;4:441–448. 26. Andersson S, Norrgren H, da Silva Z, Biague A, Bamba S, Kwok S, Christopherson C, Biberfeld G, and Albert J: Plasma viral load in HIV-1 and HIV-2 singly and dually infected individuals in Guinea Bissau, West Africa: Significantly lower plasma virus setpoint in HIV-2 infections than in HIV-1 infection. Arch Intern Med 2000;160:3286 –3293. 27. Popper SJ, Dieng-Sarr A, Gueye-Ndiaye A, Mboup S, Essex M, and Kanki P: Low plasma human immunodeficiency virus type 2 viral load is independent of proviral load: Low virus production in vivo. J Virol 2000;74:1554 –1557.

LOW LEVEL HIV-2 VIREMIA AND SURVIVAL 28. Desrosiers RC: Strategies used by human immunodeficiency virus that allow persistent viral replication. Nature Med 1999;5:723 –725. 29. Lisse I, Poulsen AG, Aaby P, Knudsen K, and Dias F: Serial CD4 and CD8 T-lymphocyte counts and associated mortality in an HIV2-infected population in Guinea-Bissau. J AIDS Hum Retroviral 1996;13:355 –362. 30. de Wolf F, Spijkerman I, Schellekens PT, et al.: AIDS prognosis based on HIV-1 RNA, CD41 T-cell count and function: Markers with reciprocal predictive value over time after seroconversion . AIDS 1997;11:1799 –1806. 31. Kim S, Hughes M, Hammer S, Jackson B, DeGruttola V, and Katzenstein DA: Both serum HIV type 1 RNA levels and CD41 lymphocyte counts predict clinical outcome in HIV type-1-infected subjects with 200 to 500 CD41 cells per cubic millimeter. AIDS Res Hum Retroviruses 2000;16:645 –653. 32. Grassly N, Xiang Z, Ariyoshi K, et al.: Mortality among human immunodeficiency virus type 2-positive villagers in rural Guinea-

1173 Bissau is correlated with viral genotype. J Virol 1998;72: 7895–7899. 33. Kimata JT, Kuller L, Anderson D, Dailey P, and Overbaugh J: Emerging cytopathic and antigenic simian immunodeficiency virus variants influence AIDS progression. Nature Med 1999;5:535 –540. 34. Barouch DH, Santra S, Schmitz JE, et al.: Control of varemia and prevention of clinical AIDS in Rhesus monkeys by cytokine-aug mented DNA vaccination. Science 2000;290:486 –492.

Address reprint requests to: Hilton Whittle Medical Research Council Laboratories P.O. Box 273 Banjul The Gambia, West Africa E-mail: [email protected]