Secondary Logo

Journal Logo


An HIV-1 natural history cohort and survival times in rural Uganda

Morgan, Dilys1,3; Malamba, Samuel S.1; Maude, Gillian H.2; Okongo, Martin J.1; Wagner, Hans-Ulrich1; Mulder, Daan W.1; Whitworth, James A.1

Author Information
  • Free



The World Health Organization estimates that, of the 16 million people infected with HIV-1 world-wide, over 10 million live in Sub-Saharan Africa [1]. Despite the enormous impact it is having, little is known about the natural history of HIV-1 infection amongst the general population in Africa [2]. Studies from Africa on the natural history of HIV-1 infection have included recipients of HIV-infected blood [3], hospital personnel cohorts [4], commercial sex workers [5], and cohorts of women recruited at perinatal or paediatric clinics [6] or maternity wards [7]. In this paper, we describe a continuing population-based natural history cohort (NHC) of HIV-prevalent and HIV-incident cases, negative controls, and negative partners of HIV-positive cases in rural Uganda. We also report survival times to death, which we believe to be the first such reports from a population-based cohort in Africa.

Uganda has been experiencing a severe AIDS epidemic for more than a decade. By the end of 1995, over 48 000 cases of AIDS had been reported (92.2% of cases were in adults) [8], and around 1.3 million adults are estimated to be infected with HIV-1, with an overall percent adult HIV prevalence of 14.5% [9]. In rural Masaka district, where the current study is based, the prevalence of HIV-1 infection in adults (over 12 years) is 8%. Almost half of all deaths in adults are associated with HIV-1. In young adults this figure reaches 80%. In the age group 13–44 years, those who are HIV-positive are 60-fold times more likely to die over a 2-year period than those who are HIV-negative [10]. The major route of transmission of HIV-1 in this population is heterosexual contact [11,12].

Materials and methods

In 1989, the Medical Research Council Programme on AIDS in Uganda established a general population cohort which includes approximately 5000 adults in a cluster of 15 villages in rural south-west Uganda. This studies the population dynamics of HIV-1 infection by annual census and serosurveys [13,14]. Most of the population are subsistence farmers growing bananas and beans. Coffee is grown as a cash crop. Around 65% of the population are Roman Catholic, with approximately 25% Muslim and 10% Protestants. Although there are no health facilities in the study area, nearby are two dispensaries and a health centre. The nearest hospital is 20 km away.

In 1990, a random selection of one-third of the HIV-1 seropositive adults identified from the initial survey round of the general population cohort were invited to enrol into an NHC as prevalent cases. Attempts were made to recruit all seroconvertors identified during subsequent annual surveys as incident cases. Seronegative controls for the prevalent and incident cases, matched for age (in 5-year bands) and village of residence, were randomly selected from the general population cohort and invited to enrol in the NHC. There was no replacement selection if the case or the control failed to enrol. In addition, negative spouses of all the identified HIV-positive cases in the general population cohort were also requested to enrol (negative discordants).

Following informed consent, the four categories of participants described above were invited to attend a study clinic every 3 months. During these regular appointments they were seen by clinicians who administered a detailed medical and sexual history questionnaire and undertook a physical examination. Participants were also encouraged to attend the clinic with medical problems occurring between routine appointments. At any visit, we provided diagnostic, therapeutic and clinical services based on simple and affordable drugs. If participants were too ill to attend the clinic, they were visited at home by a clinician and referred to hospital as required. Transport and treatment fees were paid for by the project.

Four local home visitors acted as mobilizers and visited participants to inform them of their clinic appointments and arrange transport where necessary. They also ascertained deaths and unreported illnesses during their frequent visits to the villages. The home visitors traced participants who had missed appointments. Information on deaths of study participants, including those occurring outside the study area, became available within 3 months.

All clinic staff are blind to the participants' HIV-1 serostatus. Participants for enrolment were selected by the project statistician (S.M.) at Entebbe, using a computer system, who then informed the clinic. Similar proportions of HIV-positive and HIV-negative individuals were recruited at each round. All data handling and analysis, and linkage with laboratory results were performed at the main office in Entebbe. The above system ensured that the staff did not know the HIV status of participants. This allowed unbiased reporting of early signs and symptoms, reduced follow-up and treatment biases, and estimates the background prevalence of infections and other conditions in this population against which the presenting features of the HIV-infected participants can be compared. Also, in a rural community such as this, it is essential that confidentiality of the participants' HIV serostatus is maintained. Stigmatization of participants in the NHC was avoided by continually emphasizing that there were more HIV-uninfected than HIV-infected people in the NHC, and that no-one in the clinic knew, or could find out the HIV status of any participant. General medical services for residents of the study area were also provided at the clinic and so attendance at the clinic did not necessarily indicate that they were in the NHC. The community assuming all persons selected for the NHC were HIV-positive, however, may have deterred potential participants from enrolling in the early stages of the study. This also may have caused subjects to delay in enrolling, either until they were in need of medical attention or until they were reassured of the benefits of participation.

Comprehensive counselling services were provided in study villages away from the clinic for safe-sex education and HIV testing [15]. Participants were actively encouraged to use these services, but were told that they should not inform or discuss their HIV status with the clinic staff. Condoms were freely available in the study clinic.

Serum for testing HIV-1 status was taken at each visit to detect participants who seroconverted within the NHC [16].

Statistical methods

Person-years (PY) of observation were calculated from date of enrolment for the prevalent and negative (control and discordant) cases, and from the estimated date of seroconversion of the incident cases, to death or censoring at 31 December, 1995. All enrolled participants, including those who subsequently refused or moved, were accounted for at censoring and were therefore included in the survival analyses. It was assumed that their HIV status remained unchanged from their last visit. Seroconvertors within the NHC initially contributed observation time to the HIV-negative and then to the incident cases from their estimated date of seroconversion. The estimated date of seroconversion for all incident cases was taken as midway between last negative and first positive HIV-1 test. STATA computer software (Stata Corporation, College Station, Texas, USA) was used to estimate median survival times to death, produce Kaplan-Meier survival curves and calculate cumulative survival probabilities.

Age-specific mortality rates were obtained by calculating follow-up times for participants in 10-year age groups, over which period the mortality rates were assumed to be constant. Thus, persons who changed age group during follow up contributed PY to more than one age group [17]. Mortality rates were age-standardized by the direct method of standardization, using the PY of observation distribution of the negative controls, and the age-adjusted mortality rate ratios were calculated using the Mantel-Haenszel method. Mantel-Haenszel methods were also used to adjust for confounding effects when rates were compared.


By the end of 1995, 390 of the 491 people asked to join the NHC had done so. The enrolment status for each HIV category is shown in Table 1. The reasons for not enrolling were: having moved out of the study area (34); refused (40); died (26); and severe mental retardation (one). Of those who died before enrolling: 15 were prevalent cases (median age, 34 years; range, 20–70); five were incident cases (median age, 64 years; range, 50–78); five were HIV-negative (median age, 39 years; range, 23–76); and one was a negative discordant aged 84 years. Six prevalent cases died after selection but before recruitment started. Although there was gender imbalance in some HIV categories, equal numbers of males and females were recruited into the cohort overall. Of those selected for enrolment, similar proportions of males and females in each HIV category enrolled into the cohort. The median age of enrolment of the female incident cases was 24 years, 7 years younger than that of the male cases, reflecting the different patterns of infection in men and women reported previously [13,18]. The median ages of the male and female prevalent cases were similar.

Table 1:
. Age on enrolment and gender of study population.

Forty-three (65%) of the participants enrolled as incident cases had a negative result at one round of the general population cohort and a positive result at the next round 1 year later. The remainder had an interval of 2 years or more. The mean time between last negative and first positive test was 1.6 years (range 0.42–5.0 years). The median interval between the estimated date of seroconversion and enrolment of the incident cases was 15 months (range, 1.4–41.5 months). The seroconversion dates for the prevalent cases were unknown, but the mean interval between being invited to join the NHC and being enrolled was 5.4 months (range, 1 day-60.7 months).

Compliance rates and follow up

By the end of the study period, 1353 PY of observation of the NHC had been achieved, 599 PY of observation in HIV-positive and 754 in HIV-negative cases (Table 1). For the incident cases, 279 PY of observation were achieved from seroconversion, including 189 PY of observation from enrolment in the NHC. High compliance rates have been maintained throughout the period of follow up, with an overall attendance rate of 74.2% (i.e., actual attendances/scheduled attendances). There was no difference in the attendance rates between the HIV categories. Of those participants living in the study area during 1995, over 96% were seen at least once in the clinic during the year. Migration was high among the participants of the NHC. For example, 52 (13.3%) of enrolled participants had moved out of the study area and were not available for follow up during 1995, of whom 19 were HIV-1-positive and 33 were HIV-1-negative. The proportion of surviving participants who moved were similar among HIV-positive and HIV-negative cases (15.2% and 16.4% respectively). Reasons for moving included: people returning to Rwanda; migration to urban areas for employment, especially the young; and women returning to their family following divorce, separation or death of their husbands. Although the number of participants who moved was high, many visited and were seen at the study clinic on such occasions and some returned to live in the study area and thus could be included in follow up.

Seroconversions during follow up

Twenty participants seroconverted within the NHC during follow up. Nine were enrolled as negative controls: four males whose median age at serocoversion was 33 years (range, 20–59); and five females whose median age was 26 years (range, 21–48). Eleven were negative discordants: four males with a median age of 32 years (range, 25–52); and seven females with a median age of 23 years (range, 17–31). The median duration from enrolment to the estimated date of seroconversion was 25 months for the negative controls and 6 months for negative discordants (range, 5–37 months and 2–26 months, respectively). The age-standardized incident rates were 14.1/1000 PY of observation (95% CI, 7.4–27.2) for the negative controls and 86.4/1000 PY of observation (95% CI, 47.2–158.2) for the negative discordants. These individuals were then reclassified as incident cases of infection, making 86 cases in total.

Progression to death

Sixty-four participants died over the 5-year period. Fifty-four deaths occurred in the HIV-positive (47 prevalent, and seven incident cases) and 10 in the HIV-negative cases (nine controls and one negative discordant). The negative controls and negative discordants were combined for further analysis. Three of the seven deaths in the incident category were in participants aged 26, 59 and 52 years who seroconverted in the NHC. They had intervals between last negative and first positive HIV result of 10, 16 and 22 weeks and they died 54, 86 and 60 weeks respectively after their estimated date of seroconversion. All died of AIDS.

The mortality rates by age group, and for males and females by HIV category are shown in Table 2. The age-standardized mortality rates for the prevalent, incident and negatives per 1000 PY of observation were 156.5 (95% CI, 115.8–211.4), 35.0 (95% CI, 16.4–75.0), 13.5 (95% CI, 7.3–25.1) respectively. The age-group and gender adjusted mortality rate ratios for prevalent and incident categories per 1000 PY of observation in comparison with the HIV-negative category were 14.3 (95% CI, 6.5–31.8) and 2.6 (95% CI, 1.0–6.8), respectively.

Table 2:
. Mortality rates per 1000 pyo by HIV category for gender and for age group.

There were no significant differences in the gender-specific mortality rates in any HIV category [χ2adjusting for HIV category and age group, 0.117; degrees of freedom (df) = 1; P = 0.73]. Overall, the death rates per 1000 PY of observation in males and females were 48.9 (95% CI, 34.8–68.8) and 45.7 (95% CI, 32.1–65.0), respectively.

There was a significant trend of increasing mortality rates with increasing age group in the incident and negative cases, but not in the prevalent cases (Table 2). The median ages at death of the prevalent, incident and negative participants were 33, 53 and 53 years, respectively.

The Kaplan-Meier curves for the prevalent, incident and negatives (controls and discordants together) are shown in Fig. 1. The median survival time from enrolment to death for the prevalent cases was 4.5 years (95% CI, 3.5->5.2); >5.4 years from seroconversion for the incident cases and >5.2 years from enrolment for the HIV-negative cases. The cumulative probabilities of survival at 1 and 3 years from enrolment for prevalent cases and negatives and from seroconversion for the incident cases are shown in Table 3. At 5 years, the cumulative probabilities of survival for the prevalent, incident and negative cases were 46% (95% CI, 36–57), 83% (95% CI, 65–93) and 94% (95%CI, 89–97), respectively. The survival probabilities for participants aged over 55 years were less than for the younger age groups, but since there were few participants in these older groups the confidence intervals are wide (Table 3).

Fig. 1:
. Kaplan-Meier survival curves for participants of the natural history cohort. 1=HIV-prevalent cases, n=93; 2=HIV-incident cases, n=86; 3=HIV-negative participants, n=231.
Table 3:
. Cumulative survival at 1, 3 and 5 years by HIV category and age group.


Disease progression rates of HIV infection in Africa remain uncertain [2]. We are not aware of previous studies which have looked at the natural history of HIV-1 infection in a population-based rural cohort in Africa. We have described an NHC in rural south-west Uganda which consists of prevalent and incident cases of HIV-1 infection, HIV-negative controls and discordant spouses. We do not know when the prevalent cases acquired their infection, but they must have been infected prior to 1989/1990 when the general population cohort (from which the participants of the NHC are recruited) was established. We have estimated dates of seroconversion for the 86 incident cases, (including fairly precise details on 20 of these who seroconverted within the NHC) and therefore are able to calculate survival times from the estimated date of seroconversion. We also have enrolled negative controls, and HIV-negative discordant spouses. The NHC was established over 5 years ago, and is already providing valuable information, although further follow up is required to obtain more precise estimates in the incident category.

We have achieved good compliance rates over the duration of the study, with an overall attendance rate of 74.2%. Sixteen per cent of the enrolled surviving participants had moved by 1995. The proportion of participants who moved away from the study area, however, did not vary in their HIV-1 infection rates from those who stayed. Previous studies in this area reported that 13% of individuals recruited into the large general population cohort had left the area by 2 years of follow-up [13] and this had increased to 28% after 4 years [19]. Compliance rates were almost identical in all HIV categories showing that ascertainment bias between the groups was minimal.

Twenty participants seroconverted within the NHC. Nine of these were HIV-negative controls who had been randomly selected from the general population cohort. The age-standardized incident rate in the controls was 14.1 (95% CI, 7.4–27.2) per 1000 PY of observation, which is not significantly different from the reported 9.2 per 1000 PY of observation reported in the general population cohort [13]. This further suggests that the NHC participants are representative of the general population in this area. Eleven HIV-1-negative spouses of positive partners seroconverted over the 5-year period, with an age-standardized incidence rate of 86.4 per 1000 PY of observation. This agrees with the findings of a study in the nearby Rakai District where the reported incidence amongst discordant partners was 90 per 1000 PY of observation [20].

The age-standardized mortality rates for the HIV-prevalent, -incident and -negative cases were 156.5, 35.0 and 13.5 per 1000 PY, respectively. Thus, the rate was over 10-fold higher in the prevalent cases, and about threefold higher in incident cases compared with the HIV-negative subjects. These rates are similar to results reported from the population cohort which found that the age-standardized death rates in HIV-positive and HIV-negative cases were 114 (95% CI, 93.2–134.8) and 10.4 (95% CI, 9.0–11.8) per 1000 PY of observation, respectively (Medical Research Council Programme on AIDS in Uganda, unpublished data). Again this suggests that the participants of the NHC are representative of the general population.

Median survival times to death were 4.5 years for prevalent and >5.4 years for incident cases of infection. The PY of follow up for the incident cases were fewer than for the prevalent cases and relatively few incident cases died. Therefore the median survival times for the incident cases cannot be estimated. For this reason, we also presented the cumulative probability of survival up to 5 years. Our survival times and probabilities for the HIV-positive cases are shorter than most reported studies, including those from Africa. The 4-year cumulative probability of survival for a cohort of HIV-seroprevalent women recruited at a maternity ward in Kigali was 86% [7]. In our NHC, the cumulative probability of survival for prevalent cases was only 53% at 4 years. Another HIV-infected women's cohort study in Kigali reported a 2-year mortality rate of 7% [21] compared with our 2-year mortality rate of 22%. For HIV-infected attenders at a sexually transmitted diseases clinic in Zambia, the reported mortality rates at 18 months were 4% for males and 6% for females [22]. Differences in the enrolment source, age composition, duration of infection, and HIV-1-related symptoms on enrolment make direct comparison with other studies difficult, however.

Our cumulative survival probability for the incident cases at 5 years was 83%, and this is lower than that reported from similar cohorts in Europe and North America [25–25]. We are not aware of other HIV-incident cohorts in Africa reporting survival times, although rapid progression to AIDS (4.4 years) has been reported for a cohort of female sex workers in Kenya [5]. We routinely see participants every 3 months, and provide open access to the clinic. Any medical complaints or treatable findings receive medication at these visits. Also, for ethical reasons, we provide access to secondary medical care, which would otherwise be unavailable to most individuals. Thus, although our reported survival times are shorter than other reported studies, they may in fact be longer than those experienced by persons outside the NHC who rely on routine government health services. Five incident cases died before enrolment into the NHC and were not included in the analysis. These deaths suggest that the mortality in recent seroconvertors may be high and our survival times are likely to be overestimated.

We found no gender differences in mortality rates in the HIV-positive and HIV-negative cases. Although some studies have suggested that survival maybe poorer in HIV-positive females compared with HIV-positive males [26], most studies have found no difference [27,28].

As might be expected, the death rates in the negatives increased significantly with increasing age. Although death rate in the incident cases was increased, the incident cases had not been infected long enough to disrupt this normal mortality trend, and the median age at death was 55 years, the same as in the negatives. In the prevalent cases (all of whom were infected with HIV before 1990), however, the increased mortality rates in young persons resulted in a loss of the relationship between increasing age and increasing death rates and so reduced the median age at death in this group to 33 years. Many studies have reported an increased HIV-1 disease progression and mortality rate in older age groups [22,29–32], and the older prevalent cases may have had a more rapid progression to death and died before the NHC was established or before they could enrol. Although the numbers are small and so the wide CI make interpretation of the data difficult, there is a suggestion that incident cases aged over 55 years have shorter survival times. Longer follow-up times should confirm whether the survival rates in older incident cases are significantly reduced.

In conclusion, we have described an NHC of HIV-positive and HIV-negative participants which is representative of the general population in south-west Uganda. Evidence for this includes comparable recruitment and attendance rates in all HIV categories, as well as the overall death, seroconversion and migration rates being similar to that in our general population cohort. The NHC was established over 5 years ago; it is continuing and good compliance rates are being maintained. This paper reports survival probabilities which are lower than in most reported studies, and we plan to describe clinical manifestations and other parameters of disease progression in the cohort in future papers.


We thank L. Carpenter for reviewing this manuscript and thank all the clinic staff in Kyamulibwa: B. Mayanja, J. Babuwe, R. Lubega, H. Eotu, N. Emojong, J. Musisi, E. Namara, M. Wanyana, J. Nalwadda, H. Bayiye Musoke, R. Senono (deceased), F. Nakisozi, H. Kawooya, Lupepe and the participants themselves.


1. World Health Organization: The HIV/AIDS Pandemic: 1994 Overview. Geneva: WHO/Global Programme on Aids/TCO/SEP; 1994.
2. Colebunders RL, Latif AS: Natural history and clinical presentation of HIV-1 infection in adults: AIDS 1991, 5 (suppl 1):S103–S112.
3. Colebunders R, Ryder R, Francis H, et al.: Seroconversion rate, mortality, and clinical manifestations associated with the receipt of a human immunodeficiency virus-infected blood transfusion in Kinshasa, Zaire. J Infect Dis 1991, 164:450–456.
4. Mann JM, Colebunders RL, Khonde N, et al.: Natural history of human immunodeficiency virus in Zaire. Lancet 1986, ii:707–709.
5. Anzala OA, Nagelkerke NJD, Bwayo JJ, et al.: Rapid progression to disease in African sex workers with human immunodeficiency virus type 1 infection. J Infect Dis 1995, 171:686–689.
6. Allen S, Lindan C, Serufilira A, et al.: Human immunodeficiency virus infection in urban Rwanda. Demographic and behavioural correlates in a representative sample of childbearing age. JAMA 1991, 266:1657–63.
7. Leroy V, Msellati P, Lepage P, et al.: Four years of natural history of HIV-1 infection in African women: a prospective cohort study in Kigali (Rwanda), 1988–93. J Aquir Immune Defic Syndr 1995, 9:415–421.
8. STD/AIDS Control Programme: HIV/AIDS surveillance report March 1996. Entebbe: Ministry of Health; 1996.
9. World Health Organization Global AIDS Statistics. Provisional working estimates of adult HIV prevalence as of the end of 1994 by country. AIDS Care 1996, 8:249–254.
10. Mulder DW, Nunn AJ, Kamali A, Nakiyingi J, Wagner HU, Kengeya-Koyondo JF: Two-year HIV-1-associated mortality in a Ugandan rural population. Lancet 1994, 343:1021–1023.
11. Sewankambo NK, Carswell JW, Mugerwa RD, et al.: HIV infection through normal heterosexual contact in Uganda. AIDS 1987, 1:113–116.
12. Malamba SS, Wagner H-U, Maude G, et al.: Risk factors for HIV-1 infection in adults in a rural Ugandan community: a case-control study. AIDS 1994, 8:253–257.
13. Mulder DW, Nunn AJ, Wagner U-H, Kamali A, Kengeya-Kayondo JF: HIV-1 incidence and HIV-1-associated mortality in a rural Ugandan population cohort. AIDS 1994, 8:87–92.
14. Wagner H-U, Kamali A, Nunn AJ, Kengeya-Kayondo JF, Mulder DW: General and HIV-1-associated morbidity in a rural Ugandan community. AIDS 1993, 7:1461–1467.
15. Seeley JA, Wagner HU, Mulemwa J, Kengeya-Kayondo JF, Mulder DW: The development of a community-based HIV/AIDS counselling service in a rural area in Uganda. AIDS Care 1991, 3:207–217.
16. Nunn AJ, Biryahwaho B, Downing RG, Van der Groen G, Ojwiya A, Mulder DW. Algorithms for detetecting antibodies to HIV-1: results from a rural Ugandan cohort. AIDS 1993, 7:1057–1061.
17. Clayton D, Hills M: Time. In Statistical Models in Epidemiology. Oxford: Oxford University Press; 1993:53–62.
18. Nunn AJ, Kengeya-Kayondo JF, Malamba SS, Seeley JA, Mulder DW: Risk factors for HIV-1 infection in a rural Ugandan community: a population study. AIDS 1994, 8:81–86.
19. Mulder D, Nunn A, Kamali A, Kengeya-Kayondo J: Decreasing HIV-1 seroprevalence in young adults in a rural Ugandan cohort. Br Med J 1995, 311:833–836.
20. Serwadda D, Gray RH, Wawer MJ, et al.: The social dynamics of HIV transmission as reflected through discordant couples in rural Uganda. AIDS 1995, 9:745–750.
21. Lindan CP, Allen S, Serufilira A, et al.: Predictors of mortality among HIV-infected women in Kigali, Rwanda. Ann Intern Med 1992, 116:320–328.
22. Hira SK, Ngandu N, Wadhawan D, et al.: Clinical and epidemiological features of HIV infection at a referral clinic in Zambia. J Acquir Immune Defic Syndr 1990, 3:87–91.
23. Veugelers PJ, Page KA, Tindall B, et al.: Determinants of HIV disease progression among homosexual men registered in the Tricontinental seroconverter study. Am J Epidemiol 1994, 140:747–758.
24. Lee CA, Phillips AN, Elford J, Janossy G, Griffiths P, Kernoff P: Progression of HIV disease in a haemophiliac cohort followed for 11 years and the effect of treatment: Br Med J 1991, 303:1093–1096.
    25. Brettle RP, McNeil AJ, Burns S, et al.: Progression of HIV: follow-up of Edinburgh injecting drug users with narrow seroconversion intervals in 1983–1985. AIDS 1996, 10:419–430.
    26. Melnick SL, Shere R, Louis TA, et al.: Survival and disease progression according to gender of patients with HIV infection. JAMA 1994, 272:1915–1921.
    27. Phillips AN, Antunes F, Stergious G, et al.: A sex comparison of rates of new AIDS-defining disease and death in 2554 AIDS cases. AIDS 1994, 8:831–835.
    28. Lepri AC, Pezzotti P, Dorrucci M, Phillips AN, Rezza G, the Italian Seroconversion Study: HIV disease progression in 854 women and men infected through injecting drug use and heterosexual sex and followed for up to nine years from seroconversion. Br Med J 1994, 309:1537–1542.
    29. Carré N, Deveau C, Belanger F, et al.: Effect of age and exposure group on the onset of AIDS in heterosexual and homosexual HIV-infected patients. AIDS 1994, 8:797–802.
    30. Brettle RP, McNeil AJ, Gore SM, Bird AG, Leen CSL, Richardson A: The Edinburgh City Hospital cohort: analysis of enrolment, progression and mortality by baseline covariates. Quart J Med 1995, 88:479–491.
    31. Rosenberg PS, Goedert JJ, Biggar RJ, the Multicenter Hemophilia Cohort Study and the International Registry of Seroconvertors: Effect of age at seroconversion on the natural AIDS incubation distribution. AIDS 1994, 8:803–810.
    32. Darby SC, Ewart DW, Giangrande PLF, Spooner RJD, Rizza CR, the UK Haemophilia Centre Directors' Organisation: Importance of age at infection with HIV-1 for survival and development of AIDS in UK haemophilia population. Lancet 1996, 347:1573–1579.

    Africa; progression; epidemiology; natural history

    © Lippincott-Raven Publishers.