HIV among people who inject drugs in the Middle East and North Africa: systematic review and data synthesis

Abstract

Background: It is perceived that little is known about the epidemiology of HIV infection among people who inject drugs (PWID) in the Middle East and North Africa (MENA). The primary objective of this study was to assess the status of the HIV epidemic among PWID in MENA by describing HIV prevalence and incidence. Secondary objectives were to describe the risk behavior environment and the HIV epidemic potential among PWID, and to estimate the prevalence of injecting drug use in MENA.

Methods and findings: This was a systematic review following the PRISMA guidelines and covering 23 MENA countries. PubMed, Embase, regional and international databases, as well as country-level reports were searched up to December 16, 2013. Primary studies reporting (1) the prevalence/incidence of HIV, other sexually transmitted infections, or hepatitis C virus (HCV) among PWIDs; or (2) the prevalence of injecting or sexual risk behaviors, or HIV knowledge among PWID; or (3) the number/proportion of PWID in MENA countries, were eligible for inclusion. The quality, quantity, and geographic coverage of the data were assessed at country level. Risk of bias in predefined quality domains was described to assess the quality of available HIV prevalence measures. After multiple level screening, 192 eligible reports were included in the review. There were 197 HIV prevalence measures on a total of 58,241 PWID extracted from reports, and an additional 226 HIV prevalence measures extracted from the databases. We estimated that there are 626,000 PWID in MENA (range: 335,000-1,635,000, prevalence of 0.24 per 100 adults). We found evidence of HIV epidemics among PWID in at least one-third of MENA countries, most of which are emerging concentrated epidemics and with HIV prevalence overall in the range of 10%-15%. Some of the epidemics have however already reached considerable levels including some of the highest HIV prevalence among PWID globally (87.1% in Tripoli, Libya). The relatively high prevalence of sharing needles/syringes (18%-28% in the last injection), the low levels of condom use (20%-54% ever condom use), the high levels of having sex with sex workers and of men having sex with men (15%-30% and 2%-10% in the last year, respectively), and of selling sex (5%-29% in the last year), indicate a high injecting and sexual risk environment. The prevalence of HCV (31%-64%) and of sexually transmitted infections suggest high levels of risk behavior indicative of the potential for more and larger HIV epidemics.

Conclusions: Our study identified a large volume of HIV-related biological and behavioral data among PWID in the MENA region. The coverage and quality of the data varied between countries. There is robust evidence for HIV epidemics among PWID in multiple countries, most of which have emerged within the last decade and continue to grow. The lack of sufficient evidence in some MENA countries does not preclude the possibility of hidden epidemics among PWID in these settings. With the HIV epidemic among PWID in overall a relatively early phase, there is a window of opportunity for prevention that should not be missed through the provision of comprehensive programs, including scale-up of harm reduction services and expansion of surveillance systems.

Figure 1. Map of the Middle East and North Africa region.

The defintion adopted in the review includes the following 23 countires: Afghanistan, Algeria, Bahrain, Djibouti, Egypt, Iran, Iraq, Jordan, Kuwait, Lebanon, Libya, Morocco, Oman, OPT, Pakistan, Qatar, Saudi Arabia, Somalia, Sudan (including the newly established Republic of South Sudan), Syria, Tunisia, United Arab Emirates (UAE), and Yemen.

Figure 2. PRISMA flow chart of study selection in the systematic search.

Figure 3. Trend of HIV prevalence among male people who inject drugs in (A) Iran and (B) Pakistan.

This graph displays all available HIV prevalence measures for these two countries as extracted from eligible reports (Table 3) and various databases (Table S4). Each dot represents one HIV prevalence measure for the specific year, and the bars around it define the 95% confidence interval. A pattern of established HIV epidemic is observed in Iran (A), while a trend of emerging HIV epidemic is observed in Pakistan (B).

Figure 4. Trend of HIV prevalence among people who inject drugs, and when available men who have sex with men, in repeated rounds of bio-behavioral surveillance surveys.

These graphs display the trend of HIV prevalence in repeated rounds of bio-behavioral surveillance surveys using state of the art sampling techniques for hard-to-reach populations including respondent driven sampling and time-location sampling. Country level and aggregate data of multiple cities/provinces are displayed. For consistency between countries and between different rounds within a given country, unadjusted sample estimates are displayed. Three main patterns of HIV epidemics among PWID are depicted. A pattern of emerging concentrated epidemics is observed in Pakistan (A) and Egypt (B); a pattern of established concentrated epidemic is observed in Iran (B); and a pattern of low-level HIV epidemic is observed in Tunisia (D). In Afghanistan (E), there is an emerging epidemic among PWID in apparently only part of the country; the effect of which was diluted in the second round with the inclusion of new cities with still very limited prevalence. The potential overlap of the HIV epidemics among PWID and MSM is depicted in Pakistan and Egypt. In Pakistan, an emerging HIV epidemic among transgender sex workers is observed, but lags the epidemic among PWID (A). In Egypt, the concentrated epidemic among MSM seems to have preceded the epidemic among PWID (B). In Tunisia, the potential link between the MSM and PWID epidemics is not clear because the studies were conducted after the epidemics had already risen.