During the follow-up exercise between July and November 2001, of 812 women, 655 (81%) were found to be alive, 64 (8%) had died, 8 (1%) had refused further follow-up before 2001, and 85 (10%) were lost during follow-up (Table 2). The rate of loss during follow-up was 17 cases per 1000 pyo (95% CI, 14-21). The median follow-up was 6.9 years (range, 3 months to 8.4 years). No significant differences were found in HIV status, age, CD4%, PVL, or parity between women who refused or who were lost during follow-up and women with known survival status; however, those lost during follow-up had less education, were more often of the Jola ethnic group, were more often from the coastal area, and tended to have a poorer water supply (data not shown but available upon request).
For quality assurance purposes, different field-workers visited 43 women (5.3%) on 2 occasions in 2001. No discrepancies in recorded survival status occurred. There were 4 discrepancies in the date that a woman was last known to be alive (median difference, 10 months); this was caused by interviewing different informants. Despite intensive counseling by the field-workers, only 12 women opted to have a new HIV test.
Thirty-two of 101 HIV-1-infected, 23 of 243 HIV-2-infected, and 9 of 468 seronegative women died. The mortality rates were 56 deaths per 1000 pyo for HIV-1-infected (95% CI, 39-79), 16 deaths per 1000 pyo for HIV-2-infected (95% CI, 10-23), and 3.1 deaths per 1000 pyo for seronegative (95% CI, 1.6-6.0) women. Table 2 and Figure 2 compare the mortality rate and survival probability for the 3 groups (log-rank test, P < 0.0001 for all 3 pairwise comparisons). The mortality rate among women with HIV-1 infection was 18-fold higher (95% CI, 8.5-37; χ2 test, P < 0.001) than that among HIV-uninfected women. The mortality rate among women with HIV-2 infection was 5-fold higher (95% CI, 2.3-11; χ2 test, P < 0.001) than that among HIV-uninfected women and 3.6-fold lower (95% CI, 2.1-6.1; χ2 test, P < 0.001) than that among HIV-1-infected women. There were no significant differences in mortality rate by age group for HIV-1-infected (χ2 test, P = 0.6), HIV-2-infected (P = 0.11), or seronegative (P = 0.5; Table 2) women.
Among HIV-1-infected women, the mortality rate increased over time from 25 deaths per 1000 pyo (95% CI, 12-53) in the first period to 63 deaths per 1000 pyo (95% CI, 26-152) in the third period (P = 0.02), and among HIV-2-infected women, it increased from 7.5 deaths per 1000 pyo (95% CI, 3.1-18) to 34 deaths per 1000 pyo (95% CI, 17-68; P = 0.005). Among HIV-uninfected women, the mortality rate increased marginally from 2.3 deaths per 1000 pyo (95% CI, 0.74-7.1) to 4.3 deaths per 1000 pyo (95% CI, 1.1-17; P = 0.5).
Mortality by PVL and CD4%
Among HIV-1-infected study subjects, the mortality rate increased with the log of PVL, although not significantly (P = 0.09) (Table 3); among HIV-2-infected women, the mortality rate increased significantly with PVL (P = 0.002) (Table 3). When comparing the mortality rates between HIV-1- and HIV-2-infected women within the same PVL category, the rates of HIV-2 infection were lower in 2 categories and higher in 1. The Mantel-Haenszel mortality rate ratio (comparing HIV-2- with HIV-1-infected women) controlling for PVL category was 0.78 (95% CI, 0.44-1.4; P = 0.4; Table 3).
Among HIV-1-infected subjects, the mortality rate increased with decreasing CD4% (P = 0.008) (Table 3). Among HIV-2-infected women, the mortality rates also increased with decreasing CD4% (P < 0.0005). Among women with CD4% of >28%, the mortality rate was significantly lower in HIV-2-infected subjects than in HIV-1-infected subjects (χ2 test, P < 0.001), but in the 2 lower CD4% categories, there were no significant differences in mortality rates between the 2 infections (P = 0.4 each). The Mantel-Haenszel mortality rate ratio controlling for CD4% was 0.38 (95% CI, 0.22-0.67; P < 0.0005; Table 3).
The following variables were included in an initial Poisson model as potential confounders: parity, ethnic group, educational status, area of residence, serological syphilis, number of woman's live-born children who had died, circumcision status, number of people sharing the same bedroom, smoking, and whether living with the father of the last child. In the multivariate analysis, only PVL and CD4% were independent predictors of mortality. Age was a priori in the model, and time period was a confounder and kept in the model as well. After adjusting for the other variables, each 10% decrease in CD4% was associated with a 1.6-fold increase in mortality rate (95% CI, 1.1-2.3; P = 0.02) among HIV-1-infected women and a 1.5-fold increase (95% CI, 1.0-2.3; P = 0.046) among HIV-2-infected women (Table 4). After adjusting for the other variables, a 1-log increase in PVL was associated with a 1.8-fold increase in mortality rate (95% CI, 0.9-3.4; P = 0.08) for HIV-1-infected women. Because 41% of HIV-2-infected women had a PVL below the detection limit, we could not calculate the effect of a 1-log increase in PVL in HIV-2 infection. In a multivariate analysis using categories, the mortality rate was 1.3-fold (95% CI, 0.36-4.5) higher among women with PVLs between 1000 and 9999 copies/mL and 8.7-fold (95% CI, 2.8-28) higher among women with PVLs of ≥10,000 copies/mL compared with women with PVLs of <1000 copies/mL. The mortality rate among HIV-2-infected women with normal CD4% and undetectable PVL was not significantly different from that among seronegative women (P = 0.6).
In an analysis combining HIV-1 and HIV-2, each 10% decrease in CD4% was associated with a 1.7-fold increased risk of mortality (95% CI, 1.3-2.2; P < 0.0005). PVL was also associated with mortality: rate ratios were 1.8 (95% CI, 0.66-4.9) for those with PVLs between 1000 and 9999 copies/mL and 4.4 (95% CI, 1.6-12) for those with PVLs of ≥10,000 copies/mL compared with the baseline group (<1000 copies/mL). When adjusted for CD4%, PVL, age group, and time period, the mortality hazard ratio for HIV-2 compared with HIV-1 was 0.74 (95% CI, 0.36-1.5), which was not significant (P = 0.4).
To our knowledge, this is the first community-based study to compare the mortality of both HIV-1 infection and HIV-2 infection in persons from the same population, confirming that the excess mortality of HIV-2 infection is much less than that of HIV-1 infection.7,10,12
The study confirms data from other community-based studies showing that mortality rates among African HIV-1-positive patients are comparable with those among HIV-1-infected patients in industrialized countries before the introduction of highly active antiretroviral treatment.7,8 Eight years after recruitment into this seroprevalent cohort, >50% of HIV-1-infected women were still alive.
Finally, this study shows that PVL and CD4% are independent and significant predictors for mortality of HIV-2 infection in West Africa, confirming the results of an earlier community-based study in Guinea-Bissau19 and clinic-based studies in The Gambia.25,26 HIV-2-infected women with normal CD4% had a significantly lower mortality than did HIV-1-infected women with normal CD4%, but among women with low CD4%, there were no significant differences in mortality between the 2 HIV types. These findings are in agreement with findings from a clinical cohort in The Gambia.5
A possible source of selection bias is that women were pregnant when they were recruited, and fertility is reduced among HIV-infected women.27-29 Because fertility declines with duration of HIV-1 infection,28 it is likely that on average HIV-1-infected women in this study were infected relatively recently.29 Thus, they are less representative of a seroprevalent adult population cohort, and the HIV-1-associated mortality rates among populations are generally higher than those estimated in this study.30
Up to 18 months after delivery, the women were visited quarterly, and the loss to follow-up was limited. The next visit was 4-7 years later, and this long time gap and the associated loss to follow-up could be a source of bias. However, there were no significant differences in the key baseline variables (age, CD4%, PVL, and parity) between those lost to follow-up and those who were not.
We had anticipated that many women would wish to obtain their HIV status after being counseled a second time, but this did not transpire. Potential reasons for the poor uptake could be the stigma attached to the disease, lack of effective antiretroviral treatment at the time of test offering, and counseling by male field-workers. Voluntary counseling and testing are considered valuable tools in the fight against HIV and AIDS,31 but this low acceptance ratio of HIV testing suggests that the uptake may be very low in societies where HIV disease is stigmatized and effective antiretroviral therapy is unavailable.
The association between PVL and mortality in HIV-1 infection has been confirmed in various studies14 and also for women,15 but it was of borderline significance (P = 0.08) in our study. This may have been due to low power, because the number of HIV-1-infected women with available PVLs was small (94 women).
This study shows that, like in HIV-1 infection,14 PVL is a key predictor of mortality in HIV-2 infection. Although in univariate analysis the mortality of HIV-2 infection was much lower than that of HIV-1 infection, HIV type was not an independent predictor of mortality after adjusting for PVL, CD4%, age group, and time period. This suggests that the lower mortality of HIV-2 infection can be explained by the generally lower PVL and higher CD4% in HIV-2 infection,24,26 indicating a lower virulence of HIV-2.
An intriguing question remains: why do some people infected with HIV-2 develop a high PVL and rapid decay of the immune system and others do not? This is not due to subtype (all HIV-2-infected patients in The Gambia have subtype A)32 but could be due to intrasubtype variations.33 The transmission route, infecting dose, genetic factors such as HLA type, and the HLA type of the infecting partner may be important as well. Further studies to elucidate these agent, host, and environment factors are needed.
This study indicates that PVL and CD4% are of great value in predicting the outcome of infection: the mortality rate among women with a normal baseline CD4% and a baseline PVL of <500 copies/mL was not significantly different from that among seronegative women (P = 0.6). Because clinical markers alone do not reliably predict mortality,34,35 baseline CD4% or PVL measurement may be important to guide decisions to start antiretroviral treatment. Currently, there are no guidelines for treatment of HIV-2 infection. Our data suggest that the same clinical, viral load, and CD4 cell count criteria should be applied for HIV-2 infection as for HIV-1 infection when considering the use of antiretroviral treatment. Antiretroviral drugs are effective against HIV-2,36,37 with the exception of nonnucleoside reverse transcriptase inhibitors.38,39 Most HIV-2-infected people live in Africa, but unfortunately, few HIV-2-infected patients in Africa are receiving antiretroviral treatment. Cohort studies comparing the response of HIV-1- and HIV-2-infected patients to treatment are needed.
The authors thank all the women and their families who participated in the study; Alieu Jatta, Baba Danso, and Saikou Badjie (counselors/field-workers); Aji Fatou Jallow and Majulla Saidy for data entry; the divisional health teams in The Gambia; and Dr. Martin Weber (WHO, Geneva), Dr. Tumani Corrah and Professor Keith McAdam (MRC Laboratories, The Gambia), Saihou Ceesay (National AIDS Secretariat, The Gambia), Dr. Gisela Schneider (WEC International), Dr. Shabbar Jaffar and Professor Richard Hayes (LSHTM), and Professor Roel Coutinho (GG&GD, Amsterdam) for their advice and support.
2. Marlink R, Kanki P, Thior I, et al. Reduced rate of disease development after HIV-2
infection as compared to HIV-1
3. Kanki PJ, Travers KU, MBoup S, et al. Slower heterosexual spread of HIV-2
4. 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
in The Gambia. AIDS
5. Schim van der Loeff MF, Jaffar S, Aveika AA, et al. Mortality
dually infected patients in a clinic-based cohort in The Gambia, West Africa
6. 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
7. Morgan D, Muwonge R, Mahe C, et al. HIV-1
infection in rural Africa
: is there a difference in median time to AIDS and survival compared with that in industrialised countries? AIDS
8. Crampin AC, Floyd S, Glynn JR, et al. Long-term follow-up of HIV-positive and HIV-negative individuals in rural Malawi. AIDS
9. Ricard D, Wilkins A, NGum PT, et al. The effects of HIV-2
infection in a rural area of Guinea-Bissau. AIDS
10. Poulsen AG, Aaby P, Larsen O, et al. 9-year HIV-2
in an urban community in Bissau, West Africa
11. Norrgren H, da Silva ZJ, Andersson S, et al. Clinical features, immunological changes and mortality
in a cohort of HIV-2
infected individuals in Bissau, Guinea-Bissau. Scand J Infect Dis
12. Nunn AJ, Mulder DW, Kamali A, et al. Mortality
associated with HIV-1
infection over five years in a rural Ugandan population: cohort study. BMJ
13. Leroy V, Msellati P, Lepage P, et al. Four years of natural history
infection in African women
: a prospective cohort study in Kigali (Rwanda), 1988-1993. J Acquir Immune Defic Syndr
14. Mellors JW, Rinaldo CR Jr, Gupta P, et al. Prognosis in HIV-1
infection predicted by the quantity of virus in plasma. Science
15. Anastos K, Kalish LA, Hessol N, et al. The relative value of CD4
cell count and quantitative HIV-1
RNA in predicting survival in HIV-1
: results of the Women
's Interagency HIV Study. AIDS
16. Katzenstein DA, Hammer SM, Hughes MD, et al. The relation of virological and immunological markers to clinical outcomes after nucleoside therapy in HIV infected adults with 200 to 500 CD4
cells per cubic millimeter. N Engl J Med
17. O'Brien WA, Hartigan PM, Martin D, et al. Changes in plasma HIV-1
RNA and CD4
+ lymphocyte counts and the risk of progression to AIDS. N Engl J Med
18. O'Brien WA, Hartigan PM, Daar ES, et al. Changes in plasma HIV RNA levels and CD4
+ lymphocyte counts predict both response to antiretroviral therapy and therapeutic failure. Ann Intern Med
19. Ho DD. Viral counts in HIV infection. Science
20. Berry N, Jaffar S, Schim van der Loeff MF, et al. Low level viremia and high CD4
% predict normal survival in a cohort of HIV-2
infected villagers. AIDS Res Hum Retroviruses
21. Ota M, O'Donovan D, Alabi AS, et al. Maternal HIV-1
infection and child survival in The Gambia. AIDS
22. Ministry of Health, National AIDS Control Programme, The Gambia. Policies and Guidelines on HIV and AIDS
. Banjul: Ministry of Health; 1995.
23. Boom R, Sol CJ, Salimans M, et al. Rapid and simple methods for purification of nucleic acids. J Clin Microbiol
24. Berry N, Ariyoshi A, Shabbar J, et al. Low peripheral blood viral HIV-2
RNA in individuals with high CD4
percentage differentiates HIV-2
infection. J Hum Virol
25. Ariyoshi K, Jaffar S, Alabi A, et al. Plasma RNA viral load
predicts the rate of CD4
T cell decline and death in HIV-2
-infected patients in West Africa
26. Alabi AS, Jaffar S, Ariyoshi K, et al. Plasma viral load
% and survival of HIV-1
, and dually infected Gambian patients. AIDS
27. Ryder RW, Batter VL, Nsuami M, et al. Fertility rates in 238 HIV-1
in Zaire followed for 3 years post-partum. AIDS
28. Gray RH, Wawer MJ, Serwadda D, et al. Population-based study of fertility in women
infection in Uganda. Lancet
29. Glynn JR, Buvé A, Caraël M, et al. Decreased fertility among HIV-1
attending antenatal clinics in three African cities. JAIDS
30. Boerma JT, Nunn AJ, Whitworth JA. Mortality
impact of the AIDS epidemic: evidence from community studies in less developed countries. AIDS
. 1998;12(Suppl 1):S3-S14.
31. UNAIDS. Report on the global HIV/AIDS epidemic
. Geneva: UNAIDS; 2002.
32. Berry N, Ariyoshi K, Balfe P, et al. Sequence specificity of the human immunodeficiency virus type 2 (HIV-2
) long terminal repeat u3 region in vivo allows subtyping of the principal HIV-2
viral subtypes A and B. AIDS Res Hum Retroviruses
33. Grassly NC, Xiang Z, Ariyoshi K, et al. Mortality
among human immunodeficiency virus type 2-positive villagers in rural Guinea-Bissau is correlated with viral genotype. J Virol
34. Kilmarx PH. Commentary: virus, host, or environment? BMJ
35. Morgan D, Mahe C, Mayanja B, et al. Progression to symptomatic disease in people infected with HIV-1
in rural Uganda: prospective cohort study. BMJ
36. van der Ende ME, Schutten M, Ly TD, et al. HIV-2
infection in 12 European residents: virus characteristics and disease progression. AIDS
37. Smith NA, Shaw T, Berry N, et al. Antiretroviral therapy for HIV-2
infected patients. J Infect
38. Cox S, Aperia K, Albert J, et al. Comparison of the sensitivities of primary isolates of HIV type 2 and HIV type 1 to antiretroviral drugs and drug combinations. AIDS Res Hum Retroviruses
39. Witvrouw M, Pannecouque C, Laethem KV, et al. Activity of non-nucleoside reverse transcriptase inhibitors against HIV-2
and SIV. AIDS
Keywords:© 2005 Lippincott Williams & Wilkins, Inc.
HIV-2; mortality; viral load; CD4; Africa; women; natural history; HIV-1