Absence of a recognizable seroconversion illness in Africans infected with HIV-1

It is difficult to assess the proportion of individuals who experience an acute seroconversion illness after infection with HIV-1. We found that five out of 27 recent HIV-1 seroconverters (18.5%) in a population-based cohort in rural Uganda and four out of 22 HIV-negative controls (18.2%) reported a flu-like illness. More symptoms were reported by seroconverters, but the duration of illness was similar in both groups. We found no association between symptoms and infection with HIV-1 subtype A or D. Many of the studies giving details of seroconversion illnesses are from early in the epidemic in industrialized countries, and few of these give an unbiased report of the proportion of individuals with HIV-1 infection who have a flu- or glandular fever-like illness during seroconversion. Reported rates vary from 53 to 93% of those infected [1–5]. More studies [6–8] report the features of a seroconversion illness and the association with a faster progression, and are almost exclusively reported from industrialized countries where clade B is the main circulating sub-type. In one of the few studies from Africa, Lavreys et al. [9] reported 11 signs and symptoms that were significantly associated with acute HIV-1 infection in Kenyan commercial sex workers. Of the women who had seroconverted, 81% had at least one of these symptoms or signs compared with 52% of women at non-seroconverting visits, and 52 and 18% respectively, had three or more symptoms. A study in Barbados [10] reported that black individuals had a low incidence of severe seroconversion illnesses. A clinical cohort was established in 1990 in rural Uganda, and recruits incident cases of HIV infection identified during annual serosurveys of a large general population cohort [11]. Randomly selected HIV-negative controls are also enrolled. The incident cases of infection have estimated dates of seroconversion, which are mid-way between their last negative and first positive HIV test in the population cohort. Since 1997, all participants are asked if they have had a ‘flu-like illness in the past 12 months’ at their enrolment visit. All staff at the clinic are blind to the HIV status of participants. All participants gave informed consent. Twenty-seven participants with HIV-1 infection were enrolled within one year of their estimated date of seroconversion (mean time 9.2, SD 1.7 months). The mean time between their last HIV-1 negative and first HIV-1 positive test was 12.1 months (SD 2.2). Twenty-two HIV-1-negative controls were enrolled over the same period. Details of these participants are shown in Table 1;. Five (18.5%) HIV-1 positive and four (18.2%) HIV-1-negative participants reported a flu-like illness in the preceding 12 months. Fever was the commonest symptom reported by all five HIV-1-positive and three negative controls. The symptoms reported by participants are shown in Table 1, and although sore throat, malaise and joint pains were reported more commonly in the seroconvertors, the numbers are too small to show statistically significant differences. Of the five seroconvertors reporting an illness, one reported six symptoms, one five symptoms and three reported four symptoms (mean 4.6 symptoms). One HIV-1-negative control reported four symptoms, another three symptoms and two reported two symptoms (mean 2.8 symptoms). The illness was present for a mean of 20 days (range 2–60) in the HIV-positive group and 18.5 days (range 7–30) in the HIV-negative group.Table 1: Details of the HIV-1-positive and HIV-1-negative participants and the numbers reporting a flu-like illness and the actual symptoms reported. The HIV-1 subtype, using sequence data from the env, p17 and p24 gag regions, was available for all five seroconvertors reporting an illness. Two were subtype A and two were subtype D in all three regions, and one was env A and recombinant in gag regions. Thirteen seroconvertors not reporting an illness had subtype data from all three regions. Five were subtype A and seven were subtype D in all three regions, and one was an env A–gag D recombinant. This suggests that infection with subtypes A or D was not associated with a greater risk of a seroconversion illness. In summary, we found that five out of 27 participants (18.5%) in a population-based cohort in rural Uganda who had become infected with HIV-1 in the previous year and four of 22 HIV-negative controls (18.2%) reported a flu-like illness in the previous 12 months. More symptoms were reported by the seroconvertors, but the duration of illness was similar in both groups. We found no association between reporting a possible seroconversion illness and infection with HIV-1 subtypes A or D. Although we had relatively small numbers, we thus found that a seroconversion illness is not commonly reported in this population. There may be immunological differences in the responses to HIV-1 infection in African populations, our numbers may be too small to detect a difference, or the background level of malaria and other illnesses may mean that seroconversion illnesses are not readily distinguishable from background morbidity. Further work is needed to confirm these findings, but unfortunately there are few studies in Africa with the quality of data and HIV-1-negative control groups to do this. Dilys Morgan Cedric Mahe Jimmy Whitworth