More than two thirds of the 40 million HIV-1-infected patients in the world live in sub-Saharan Africa.1 In Africa, AIDS affects women more than it does elsewhere in the world. In some areas, more than 60% of women aged between 15 and 24 years and 40% of pregnant women are infected.2 A few studies have been conducted in eastern Africa, but unlike industrialized countries, no overall approach to HIV/AIDS prevention and treatment has been developed.3 This situation is now a major public health concern, and the World Health Organization (WHO) has called for unprecedented action to ensure access to antiretroviral treatment in the developing world. The available few data about virus subtypes and their potential drug resistance mutations threaten such initiatives with failure. Three genetic groups of HIV-1 circulate in Africa, designated M, O, and N. Group M includes a wide diversity of non-B subtypes and circulating recombinant forms (CRFs).4,5 This genetic diversity has substantial consequences for the emergence of antiretroviral drug (ARV) resistances.6-8 Furthermore, there is a risk of oversimplification if only data from high-prevalence countries or industrialized countries, where subtype B is predominant, are considered.
To assess the HIV-1-infected population, the seroprevalence, and the genetic heterogeneity of circulating subtypes, a national study was conducted for monitoring the general population of Djibouti.
The prevalence of HIV infection in the population aged 15 to 54 years living in the Republic of Djibouti was estimated in March 2002, using an anonymous noncorrelated cluster sampling method.9 In brief, 30 clusters in the city of Djibouti and 30 other clusters in the rural districts were investigated. The clusters were randomly selected from the lists of quarters used by the National Directorate of Statistics in the City of Djibouti and from the lists of towns used by the Expanded Program on Immunizations for the districts. A starting household was randomly selected, and the nearest households were investigated until a total of 44 (City of Djibouti) or 37 (other districts) permanent resident adults were recorded for each cluster.
Blood samples were collected anonymously in accordance with the recommendations of the Djiboutian Ministry of Health, which also gave ethical approval for the study. Whole blood was collected into EDTA tubes. Plasma was separated from cells by centrifugation, and stored at −80°C in aliquots. The samples were screened locally with a third-generation anti-HIV assay (Genscreen HIV ½ version 2; Biorad), and samples scoring positive for HIV-1 antibodies were shipped to the Military Hospital Val de Grâce (Paris, France) for quantification of the viral load (Amplicor HIV-1 Monitor 1.5 Assay; Roche Diagnostic Systems). The HIV-RNA isolated from plasma during the specimen preparation process was used for reverse transcriptase (RT) polymerase chain reaction (PCR) of target regions of the genome for sequencing. The pol and env genes were amplified with the primers SK 38/MJ4 and JA 9/JA 12 of the consensus, using the technique of the Agence Nationale de Recherche sur le SIDA-ANRS (available at: http://www.anrs.fr). Sequencing was carried out with the CEQ Dye Terminator Cycle sequencing “Quick-start” kit and a Genetic Analysis System (Beckman Coulter) CEQ 8000. Phylogenetic analysis was performed by estimating the relatedness of the pol sequences and of env sequences (fewer of which were successfully amplified). Reference sequences for HIV genetic subtypes and CRFs were obtained from the GenBank database. Nucleotide sequences were aligned by Clustal W 1.7 software.10 The phylograms were constructed with the MEGA program using the Jukes-Cantor algorithm and the neighbor-joining method with 500 bootstrapped data sets.11 Bootstrap values greater than 75% were regarded as providing evidence for a phylogenetic grouping.
To compare categoric variables of the initial population, χ2 or Fisher exact tests were used as appropriate. The Mann-Whitney U test was used to compare continuous variables. Data were computed using Epi-Info 6.04d.
A panel of 2423 subjects representing the general population of Djibouti was included (Table 1). HIV-1 antibodies were detected in 53 (2.2%) of 2423 samples tested. The serologic status was previously unknown. Neither obvious risk factors nor previous ARV treatment was found for the 53 patients. The ratio of infected patients to the initial population indicated an HIV-1 infection prevalence of 2.2% in the general population of Djibouti. There was no significant difference in the sex ratio (female [F]/male [M]) between the noninfected population (F/M = 1.4) and the 53 infected patients (F/M = 2.1). The mean age was 31.8 years in the noninfected population and 29.6 years in the 53 infected patients (P = 0.05). There was no difference between the mean age in men and women in the infected group (F = 30.2 years/M = 29.2 years; P = 0.95). Study subjects were predominantly Djiboutians, and 54.5% have lived in the city.
All 53 patients were viremic, confirming the validity of the serologic screening. The results according to the sex are presented in Table 2. The viral load was greater than 30,000 copies/mL for 34 (62%) of 53 plasma samples and greater than 300,000 copies/mL for 15 (28%) samples. A phylogenetic tree (Fig. 1) was constructed on the basis of their HIV-1 nucleotide sequences, referenced as Dj1 to Dj34. All the patients harbored HIV-1 non-B subtypes. Twenty-five sequences clustered together and belonged to subtype C (73%). Six isolates were close to subtype CRF 02_AG (18%), 2 isolates belonged to subtype D (6%), and 1 belonged to subtype A (3%). The sequences of 2 samples (Dj29, Dj30) were close to each other, but the patients were not epidemiologically linked and lived in 2 different areas. The env sequence (330 base pairs [bp]) was successfully amplified from 14 samples. The same phylogenetic analysis performed on the 14 env sequences was similar to the pol sequences for 10 samples, as indicated by an asterisk in Figure 1. Four samples gave discordant results: Dj11 (C pol/CRF02_AG env), Dj17 (C pol/A env), Dj 27 (D pol/CRF02_AG env), and Dj28 (D pol/C env).
Despite the dramatic impact of HIV-1 in sub-Saharan Africa, HIV-1 subtypes circulating in populations with potential access to ARV have not been fully characterized. Djibouti, surrounded by Erythrea, Ethiopia, and Somalia, is an important crossroad for trade and exchange. The subjects enrolled in this study are representative of the general population in Djibouti, and the samples were collected using a valid sampling scheme and were designed to represent urban and rural populations. Indeed, appropriate population sampling is essential for a study of this kind. Because women are the main targets for controlling the HIV-1 pandemic in Africa (eg, education, mother-to-child transmission, prostitution), it is useful to consider the data obtained according to sex. We found no significant difference between the sex ratio of the noninfected Djiboutians and infected patients, but there are twice as many women among infected patients. The mean viral load was higher in infected women than in infected men, but the difference was not significant. In contrast to most African populations, the mean age of the infected women was not younger than that of the infected men.
The first HIV-1 prevalence studies in Djibouti took place in 1990. The rate of AIDS/HIV infections among street prostitutes has been rising continually: 9% in 1989, 41% in 1990, 43% in 1991, and 70% in 2002.12 Among bar prostitutes, however, there is a downward trend: 13.1% in 1991 and 7% in 2002. These data suggest better education and improved condom use in this group.13 In 1993, HIV-1 prevalence among men with sexually transmitted diseases was 10.4%.14 In patients suffering from tuberculosis, the HIV-1 rate rose from 9% in 1990 to 19% in 2001.15 Some studies attempted to describe the HIV-1 prevalence in the population outside such risk groups. In 1990, a study in secondary school students revealed a rate of 0%, underlining the effectiveness of information campaigns at school.16 In 1997, the prevalence among blood donors was estimated to 3.4%,17 but a study conducted in 2003 at the only blood transfusion center in Djibouti revealed a lower prevalence of 1.9% (X. Dray, MD, unpublished data, 2004), confirming previous suggestions.18 Our study clearly documents the low prevalence of HIV infection (2.2%) in the general population of this area for the first time. It is lower than the prevalence in neighboring countries,19 particularly Kenya (14%), Uganda (10%), and Ethiopia (9%). It is close to the rate in Somalia, where a prevalence of 1% was reported by the United Nations Program on HIV/AIDS (UNAIDS) in 1990 and 2000.20 No obvious explanation for this low prevalence has been reported. The common characteristics of Somalia and Djibouti include the way of life (pastoral in rural districts and small shopkeepers in the city) and the majority ethnic groups (Issas and Somalis). Host factors that confer relative protection against HIV-1 in the same way as for some exposed uninfected patients in Vietnam could also be suspected (genetic factors or immune activation).21 Our study is not informative about these possibilities.
The small number of isolates tested does not provide enough information for detailed conclusions about the HIV-1 subtypes in the general population of Djibouti, but it is possible to delineate some features. Phylogenetic analysis revealed that all the isolates clustered with African isolates, with a large majority of subtype C. In the Horn of Africa, subtype C prevalence has been described mostly among commercial sex workers, where it can be as high as 45%.22 The other isolates were grouped as subtypes A and D and CRF02_AG recombinants. Subtype A viruses are mostly found in East Africa, particularly in Kenya,23 and 1 subtype A virus was characterized in a Djiboutian. A recent study revealed that 50% of the blood donors in Khartoum (Sudan) were infected by subtype D, which is also the majority subtype in Uganda.24 Two isolates were characterized in this study clustering with a subtype D. CRF02_AG recombinant isolates were previously described in West and Central Africa.25,26 The detection of CRF02_AG strains in Djibouti indicates that they are still circulating in Djibouti, the only country in East Africa in which this recombinant virus was found some years ago.27,28 It would be valuable to have more isolates so that the CRF02_AG subtype prevalence could be determined. The discordances observed between pol and env sequences in 4 isolates are probably attributable to the high variability in the env region of the genome. No subtype B viruses in the general population were found. According to a former survey among prostitutes in Djibouti, subtype B seems to be limited to this risk group.29
These different subtypes were previously geographically separated. The civil wars and ethnic conflicts in the Horn of Africa have caused large population displacements and led to a cocirculation of HIV-1 viruses in Djibouti. As suggested by similar studies in other countries, this heterogeneity may generate novel strains and scattering. This potential diversity could have an impact on susceptibility to ARV drugs.
The authors thank Diane Descamps and Florence Damond of the Laboratory of Virology, Hôpital Bichat, Paris, France, for their assistance in reviewing earlier versions of the manuscript; Michel Etchepare and the Centre de Recherche d'Etude et de Documentation en Economie de la Santé (CREDES), Paris, France, for their valuable contribution and support to the study; and Christian Tosi for his technical assistance in the field.
1. UNAIDS. UNAIDS report for 2003. Most deaths and new infections ever; some good news. AIDS Treat News
2. De Cock KM, Mbori-Ngacha D, Marum E. AIDS in Africa: shadow on the continent: public health and HIV/AIDS in Africa in the 21st century. Lancet
3. Hanson S. HIV/AIDS control in sub-Saharan Africa. Science
4. Janssens W, Buvé A, Nkengasong JN. The puzzle of HIV-1 subtypes
in Africa. AIDS
5. Kantor R, Katzenstein D. Polymorphism in HIV-1 non subtype B protease and reverse transcriptase and its potential impact on drug susceptibility and drug resistance evolution. AIDS Rev
6. Handema R, Terunuma H, Kasolo F, et al. Prevalence of drug-resistance-associated mutations in antiretroviral drug-naive Zambians infected with subtype C HIV-1. AIDS Res Hum Retroviruses
7. Weidle PJ, Kityo C, Mugyenyi P, et al. Resistance to antiretroviral therapy among patients in Uganda. J Acquir Immune Defic Syndr
8. Hermans P, Schmit JC, Kabeya K, et al. Virological response to salvage therapy at 6 months in patients with B or non-B subtypes. Presented at: Ninth Conference on Retrovirus and Opportunistic Infections; 2002; Seattle.
9. Henderson RH, Sundaresan T. Cluster sampling to assess immunization coverage: a review of experience with a simplified sampling method. Bull World Health Organ
10. Thompson JD, Higgins DG, Gibson TJ. CLUSTAL W; improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res
11. Kumar S, Tamura K, Nei M. Molecular Evolutionary Genetics Analysis
[computer program]. Version 1.01. University Park, PA: Pennsylvania State University; 1993.
12. Couzineau B, Bouloumie J, Hovette P, et al. Prevalence de l'infection par le virus de l'immunodéficience humaine (VIH) dans une population cible en République de Djibouti
. Med Trop (Mars)
13. Marcelin AG, Grandadam M, Flandre P, et al. Kaposi's sarcoma herpes virus and HIV-1 seroprevalence in prostitutes in Djibouti
. J Med Virol
14. Rodier G, Couzineau B, Gray CG, et al. Trends of human immunodeficiency virus type-1 infection in female prostitutes and males diagnosed with a sexually transmitted disease in Djibouti
, East Africa. Am J Trop Med Hyg
15. Renoux E, Barreh Matan A, Sevre JP, et al. Tuberculose et infection VIH. L'expérience du programme national de lutte antituberculeuse de Djibouti
: 1990-1996. Med Trop (Mars)
16. Rodier G, Morand JJ, Olson JS, et al. HIV infection among secondary school students in Djibouti
, horn of Africa: knowledge, exposure and prevalence. East Afr Med J
17. Massenet D, Bouh A. Les aspects de la transfusion sanguine à Djibouti
. Med Trop (Mars)
18. Fox E, Abatte EA, Wassef HH, et al. Low prevalence of HIV infection in Djibouti
-has the AIDS epidemic come to a stop at the Horn of Africa? Trans R Soc Trop Med Hyg
19. UNAIDS. AIDS epidemic update, 2004. Available at: http://www.unaids.org
. Accessed March 2004.
20. UNAIDS. Somalia: epidemiological fact sheets on HIV/AIDS and sexually transmitted infections, 2002 update. Available at: http://www.unaids.org
. Accessed February 2004.
21. Truong LX, Luong TT, Scott-Algara D, et al. CD4 cell and CD8 cell-mediated resistance to HIV-1 infection in exposed uninfected intravascular drug users in Vietnam. AIDS
22. Hussein M, Abebe A, Pollakis G, et al. HIV-1 subtype C in commercial sex workers in Addis Ababa, Ethiopia. J Acquir Immune Defic Syndr
23. Robb ML, McCutchan FE, Carr JK, et al. Support of HIV-1 vaccine development in East Africa: sequencing of 98 full-genome sequences from Uganda, Kenya, and Tanzania. Presented at: AIDS Vaccine Conference; 2001; Philadelphia.
24. Hierholzer M, Graham RR, El Khidir I, et al. HIV type 1 strains from East and West Africa are intermixed in Sudan. AIDS Res Hum Retroviruses
25. McCutchan F. The global spread of subtypes and inter-subtypes recombinant HIV-1. Presented at: Conference on the Laboratory Science of HIV; 1998; Rockville, MD.
26. Vidal N, Koyalta D, Richard V, et al. High genetic diversity of HIV-1 strains in Chad, West Central Africa. J Acquir Immune Defic Syndr
27. Carr J, Salminen M, Albert J, et al. Full genome sequences of human immunodeficiency virus type 1 subtype G and A/G intersubtype recombinant. Virology
28. Cornelissen M, Van Den Burg R, Zorgdrader F, et al. Spread of distinct human immunodeficiency virus type 1 AG recombinant lineages in Africa. J Gen Virol
29. Lasky M, Perret JL, Peeters M, et al. Presence of multiple non-B subtypes and divergent subtype B strains of HIV-1 in individuals infected after overseas deployment. AIDS
Nucleotide Sequences GenBank Accession Numbers [(pol = Djn), (pol/env = Djn)]
(AY920378 = Dj1), (AY920374/AY923056 = Dj2), (AY920364/AY920916 = Dj3), (AY920383/AY920387 = Dj4), (AY920384/AY920388 = Dj5), (AY906969 = Dj6), (AY920367 = Dj7), (AY920377 = Dj8), (AY923060 = Dj9), (AY906963 = Dj10), (AY920370/AY923054 = Dj11), (AY906966 = Dj12), (AY920365 = Dj13), (AY920372/AY923055 = Dj14), (AY906964 = Dj15), (AY906968 = Dj16), (AY920363/AY920915 = Dj17), (AY906970 = Dj18), (AY906965 = Dj19), (AY920366/AY920917 = Dj20), (AY920382/AY920386 = Dj21), (AY906967 = Dj22), (AY906971 = Dj23), (AY920373 = Dj24), (AY920374 = Dj25), (AY920369 = Dj26), (AY920368/AY923053 = Dj27), (AY920381/AY923059 = Dj28), (AY920380 = Dj29), (AY920371 = Dj30), (AY906962 = Dj31), (AY920376/AY923057 = Dj32), (AY920379/AY923058 = Dj33), (AY920385/AY920389 = Dj34)