Soc Sci Med. 2004 May;58(9):1751-6. HIV and Islam: is HIV prevalence lower among Muslims? Gray PB Department of Anthropology, Peabody Museum, Harvard University, 11 Divinity Avenue, Cambridge, MA 02138, USA. gray@fas.harvard.edu Abstract Religious constraints on sexuality may have consequences for the transmission of sexually transmitted diseases. Recognising that several Islamic tenets may have the effect, if followed, of reducing the sexual transmission of HIV, this paper tests the hypothesis that Muslims have lower HIV prevalence than non-Muslims. Among 38 sub-Saharan African countries, the percentage of Muslims within countries negatively predicted HIV prevalence. A survey of published journal articles containing data on HIV prevalence and religious affiliation showed that six of seven such studies indicated a negative relationship between HIV prevalence and being Muslim. Additional studies on the relationship of risk factors to HIV prevalence gave mixed evidence with respect to following Islamic sexual codes (e.g., vs. extramarital affairs) and other factors, but that benefits arising from circumcision may help account for lower HIV prevalence among Muslims. PMID: 14990375

1. Social Science & Medicine 58 (2004) 1751–1756
HIV and Islam: is HIV prevalence lower among Muslims?
Peter B. Gray*
Department of Anthropology, Peabody Museum, Harvard University, 11 Divinity Avenue, Cambridge, MA 02138, USA
Abstract
Religious constraints on sexuality may have consequences for the transmission of sexually transmitted diseases.
Recognising that several Islamic tenets may have the effect, if followed, of reducing the sexual transmission of HIV, this
paper tests the hypothesis that Muslims have lower HIV prevalence than non-Muslims. Among 38 sub-Saharan African
countries, the percentage of Muslims within countries negatively predicted HIV prevalence. A survey of published
journal articles containing data on HIV prevalence and religious affiliation showed that six of seven such studies
indicated a negative relationship between HIV prevalence and being Muslim. Additional studies on the relationship of
risk factors to HIV prevalence gave mixed evidence with respect to following Islamic sexual codes (e.g., vs. extramarital
affairs) and other factors, but that benefits arising from circumcision may help account for lower HIV prevalence
among Muslims.
r 2003 Elsevier Ltd. All rights reserved.
Keywords: HIV; AIDS; Islam; Religion
Introduction
Behaving in accordance with religious tenets may
have impacts on health and disease transmission (Ellison
& Levin, 1998; Reynolds & Tanner, 1995). In the
context of sexually transmitted diseases (STDs), religi-
osity and religious affiliation may be negatively related
to STDs because of the common constraints religions
place on sexuality (Seidman, Mosher, & Aral, 1992).
Religious practices such as circumcision can also affect
transmission rates of STDs. If religious factors asso-
ciated with HIV—which is largely transmitted sexu-
ally—can be identified, then this endeavour can be
important in helping to understand and predict the
course of the raging HIV epidemic (Gayle & Hill, 2001;
Piot, Bartos, Ghys, Walker, & Schwartlander, 2001).
Following this logic, this paper examines the relation-
ship between HIV and Islam. That is, it tests the
hypothesis that Islamic religious affiliation negatively
associates with HIV seropositivity. Though this hypoth-
esis has been proposed before (Lenton, 1997; Ridanovic,
1997), no one, to my knowledge, has tested it.
For several reasons, adherence to Islamic tenets may
confer protective benefits against the sexual transmission
of HIV. While Islamic marital codes permit men to
marry as many as four wives and divorce relatively
easily, potentially increasing the number of lifetime
sexual partners—a known risk factor for acquiring HIV
(Stanberrry & Bernstein, 2000; Wasserheit, Aral,
Holmes, & Hitchcock, 1991), prohibitions against sex
outside of marriage may outweigh these risks. If
followed, codes against sex outside marriage for both
males and females could reduce premarital and extra-
marital sex as well as reduce sexual activity with
commercial sex workers. Prohibitions against homo-
sexual sex could reduce the risks of, for example,
unprotected anal sex.
Islam also prohibits the consumption of alcohol. By
increasing risky sexual behaviour, including reduced use
of condoms (Bastani et al., 1996; Wilson, Lavelle,
Mwoboto, & Armstrong, 1992), alcohol consumption
may favour higher rates of sexually transmitted HIV.
Islamic attention to ritual washing could increase penile
hygiene, lessening the risk of STD transmission (see
ARTICLE IN PRESS
*Tel.: +1-617-496-4262; fax: +1-617-496-8041.
E-mail address: gray@fas.harvard.edu (P.B. Gray).
0277-9536/$ - see front matter r 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0277-9536(03)00367-8

2. Lerman & Liao, 2001). Lastly, circumcision has been
identified as a practice apparently decreasing HIV
transmission (Bailey, Plummer, & Moses, 2001; Weiss,
Quigley, & Hayes, 2000). Because all Muslims should be
circumcised, this practice may also reduce the acquisi-
tion of HIV. For these reasons, we may expect Islamic
religious affiliation to be negatively associated with HIV.
If adherence to tenets constraining sexuality distin-
guishes Muslims from members of other religious
groups, or Islamic practices systematically differ from
those of other religions (e.g., circumcision), then we may
also find that there is a stronger, inverse relationship
between HIV and Islam compared with HIV and other
religions.
Methods
Two methods of data collection were utilised. First, a
multivariate analysis was conducted which used infor-
mation obtained from several large on-line demographic
and AIDS databases. All 38 sub-Saharan African
countries with a minimum of 1 million inhabitants were
included in this analysis. Initially, countries from North
Africa, Asia, and the Pacific were also included in this
endeavour, but sub-Saharan Africa emerged as the only
region for which sufficient national variation in HIV
prevalence and percentage Muslims seemed to enable
meaningful multivariate comparisons. Attempts to
examine regional variation within sub-Saharan Africa
(e.g., west, central, east, and southern African countries)
were also made, but sample sizes were too small to allow
meaningful quantitative conclusions to be drawn.
Five independent (percentage Muslims, population
density, percentage urban population, annual per capita
purchasing power, and year of first recorded AIDS case)
and one dependent (HIV prevalence) variable were
obtained as follows for this multivariate analysis. The
percentage Muslims was obtained from the 2000 World
Almanac. Population density, percent urban population
and the annual per capita purchasing power were
taken from a 2002 on-line demographic database
(www.prb.org). The annual per capita purchasing power
standardises the average income relative to amounts in
$US of goods that could be purchased with it. The year
of first recorded AIDS case reports the earliest year in
which an AIDS case was noted in the UNAIDS
database at www.unaids.org.hivaidsinfo/statistics. To
obtain data on HIV prevalence, the on-line UN
database—(www.unaids.org/epidemic update/report/es-
timates.pdf)—was used. All comparisons involved the
percent of HIV positive adults aged 15–49 and were
estimates of HIV prevalence as of the end of 1999. All of
these variables utilised on-line data obtained on 19 May,
2002. The data used in multivariate analyses were
obtained from different sources and by different
methods (i.e., Schwartlander et al., 1999 for HIV
prevalence), and almost certainly vary with respect to
accuracy. Nonetheless, these data represent the best,
most recent national data available for examining the
variables considered in multivariate analyses.
The second method of data collection reviewed the
published literature for journal articles linking HIV, risk
factors for HIV, and Islam. Standard on-line searches
(e.g., MEDLINE under key words including ‘‘HIV and
Islam’’, ‘‘HIV and religion’’, and ‘‘HIV and risk
factors’’) helped identify studies that showed either (a)
relationships between Islam and HIV or (b) relation-
ships between Islam and risk factors (e.g., extramarital
affairs, alcohol consumption) for HIV. This second
method helps specify the immediate factors by which
HIV prevalence might be lower among Muslims. The
literature review was last updated on 21 May, 2002.
Results
The relationship between national HIV prevalence
and percentage Muslims within sub-Saharan African
countries is shown in Fig. 1. Across sub-Saharan Africa,
the multivariate regression model was significant
(F ¼ 10:60; df ¼ 24; po0:00005) and accounted for a
high proportion of the variance in national HIV
prevalence rates (R2
¼ 0:69). Of the five predictor
variables, only percentage Muslims (b ¼ À0:516;
p ¼ 0:0005) and annual per capita purchasing power
(b ¼ 0:636; p ¼ 0:0002) were significant (population
density: b ¼ À0:068; p ¼ 0:584; percentage urban popu-
lation: b ¼ À0:258; p ¼ 0:084; year of first recorded
AIDS case: b ¼ 0:244; p ¼ 0:059).
Table 1 summarises the results of studies that allow
links between Islamic religious affiliation and HIV to be
made. Six of seven studies observed a negative relation-
ship between HIV prevalence and Islam. The one study
which observed a positive relationship utilised a
different method than the other studies—it compared
HIV prevalence and percentage religious affiliation
among political units (wards) rather than individuals’
religious affiliation with HIV status. Of the two studies
on East African truck drivers (who constitute a high-risk
population), one observed that changes in HIV status
were non-significantly lower compared with Christians,
and significantly lower than members of other religious
groups. The second study of truck drivers found that
Muslims and Protestants both had lower HIV rates than
truck drivers of other religious groups. Of the four other
studies finding negative relationships between HIV and
Islam, some of these controlled for factors (e.g.,
circumcision (Gray et al., 2000) and lifetime number of
sex partners (e.g., Malamba et al., 1994) that themselves
can be construed as correlates of Islamic affiliation.
ARTICLE IN PRESS
P.B. Gray / Social Science & Medicine 58 (2004) 1751–17561752

3. Table 2 summarises the studies that contain links
between Islam and factors thought to increase the risk of
sexually acquiring HIV. These findings can be combined
with intervening factors (e.g., circumcision) reported in
Table 1 to provide a more extensive set of links between
risk factors and Islam. Two (out of two with data) of
these studies found lower rates of alcohol consumption
among Muslims, and one (out of one) study summarised
in Table 1 found a similar result. Thus, 3/3 studies for
which data exist observed negative relationships between
Muslims and alcohol consumption.
With respect to sexual relationships, 2/2 studies
observed higher degrees of polygyny among Muslims,
a possible contributor to more lifetime sexual partners
(a risk factor for HIV). One study found that Muslim
men, but not women, had less extramarital sex; another
study that Muslims did not report less casual sex the
previous year; and one study that Muslims were less
likely to have extramarital partners the previous year.
Of the studies in Table 1, one discovered that Muslims
were less likely to have had sex with a commercial sex
worker. Taken together, these data suggest a mixed
ARTICLE IN PRESS
Table 1
Studies linking Islam and HIV
+ Or À Link Study location and intervening factors Reference
À(But NS) Uganda. Among circumcised men, non-significantly lower HIV rate among
Muslims (but significant ages 20–29). Significantly lower HIV rate and alcohol
intake among circumcised (99.1% Muslims circumcised) than non-circumcised
men. Muslim post-intercourse cleaning may also reduce HIV.
Gray et al. (2000)
À Uganda. Muslims had significantly lower prevalence of HIV than non-Muslims, an
affect due to alcohol consumption (and fewer Muslims had a history of alcohol
consumption).
Mbulaiteye et al. (2000)
À Kenya (truck drivers). Islam associated with lower seropositive change than
Protestant (NS), Catholic (NS), or other religions (significant). Muslims were
significantly less likely to have engaged in sex with a Prostitute.
Rakwar et al. (1999)
À Kenya (truck drivers: also includes Ugandans and Rwandans). Muslims and
Protestants had lower HIV rates than Catholics, an effect apparently due to
circumcision status.
Bwayo et al. (1994)
+ Tanzania. Political units (wards) with a higher proportion of Muslims had a greater
prevalence of HIV.
Killewo et al. (1994)
À Uganda. Controlling for lifetime number of sexual partners and other factors,
Muslims had lower prevalence of HIV.
Malamba et al. (1994)
À Uganda. Muslims had significantly lower prevalence of HIV than members of other
religious groups.
Nunn, Kengeya-
Kayondo, Malamba,
Seeley, and Mulder
(1994)
Fig. 1. HIV prevalence vs. percentage Muslims among sub-Saharan African countries.
P.B. Gray / Social Science & Medicine 58 (2004) 1751–1756 1753

4. picture with respect to Muslims having fewer lifetime
sexual partners.
To the degree that circumcision confers protective
benefits, Muslims have apparently experienced these.
Four studies (out of four for which data exist) suggest
that Muslims were more likely to have been circumcised
than other members in the study populations. Two other
potential risk factors appeared once: Islam’s rules for
ritual washing may decrease HIV transmission, but
inferences that it does not allow condom use could
increase sexual HIV transmission.
Discussion
In a sample of sub-Saharan African countries, the
percentage of Muslims within countries negatively
predicted national HIV prevalence. These results sup-
port the hypothesis that HIV prevalence is lower among
Muslims. These results discount alternative explanations
based on the timing of HIV exposure (i.e., that HIV-1
appears to have evolved in west-central Africa and HIV-
2 in West Africa: Holmes, 2001) or HIV transmission
increasing more readily among urban populations since
the results remain after controlling for such variables.
That is, sub-Saharan African countries with earlier dates
of the first reported HIV case, or more urban countries,
did not have higher HIV prevalence. Moreover, addi-
tional analysis revealed that within sub-Saharan African
regions (west, central, east, and southern), percentage
Muslims and HIV prevalence were negatively related in
all four cases, but the small sample sizes (e.g., six central
African countries) argued against expanding on a
quantitative regional focus. The finding that per capita
purchasing power positively predicted HIV prevalence
in the multivariate analysis is interesting, but largely
outside the scope of this paper. Nonetheless, one
possible reason for it could be that higher per capita
purchasing powers are linked with relatively greater
economic costs to marriage (e.g., prohibitively expensive
brideprices), leading to an increase in sexual partners at
the same time that formalised marital relationships
decline (see Blanc & Gage, 2000).
The data on risk factors associated with HIV portray
a mixed picture with respect to any protective benefits
following from adherence to Islamic codes. The data
clearly show that these rules are not followed by
everyone, and in some cases do not appear to alter
patterns of behaviour by comparison with non-Muslim
members of the same population. This is especially the
case with sexual activity (see also Gibney, Choudhury,
Khawaja, Sarker, & Vermund, 1999; Kagimu et al.,
1998). In some cases, Muslims do not appear to engage
in less extramarital sex, for example, than non-Muslims
in the same population. On the other hand, there is no
evidence (discounting results in the Killewo, Dahlgren,
and Sandstrom (1994) study in Tanzania because these
data did not directly examine religious affiliation) that
Islamic religious affiliation increases sexual behaviours
that constitute risk factors for HIV. The summary of
this evidence on sexual behaviour may be that the link
between Islam and sexual risk factors is ambiguous and
varies among populations.
Through the practice of circumcision, which may
reduce the risk of HIV acquisition, and reduced
consumption of alcohol (which might reduce risky
sexual behaviour though, note the discussion about
sexual behaviour above), adherence to Islamic tenets
seems to reduce the risk of HIV transmission. Consistent
with this, Auvert et al. (2001) found that factors
unrelated to sexual behaviour accounted for differences
in HIV prevalence among four sub-Saharan African
urban populations. Ritual cleaning practices (e.g.,
after intercourse) may also decrease the chances of
sexually acquiring HIV. Higher degrees of polygyny,
observed in two studies, and a possible sentiment against
condom use (noted in one study) may increase the risks
of HIV.
ARTICLE IN PRESS
Table 2
Studies linking Islam and risk factors for HIV
+ Or À Link Study location and risk factor(s) Reference
À Tanzania. Muslims had higher circumcision rates than Christians and
members of other religions.
Nnko, Washija, Urassa,
and Boerma (2001)
Both + and À Senegal. Muslims report lower alcohol consumption, but equally likely to
report casual sex the previous 12 months and more likely to be polygynously
married than Christians.
Lagarde, et al. (2000)
Both + and À Uganda. Muslims less likely to consume alcohol, to have had extramarital
partners the previous year, more likely to have been circumcised but more
likely to be polygynously married and less likely to use condoms.
Bailey, Neema, and
Othieno (1999)
À and no relationship Nigeria. Muslim and Protestant men less likely than others to report an
extramarital affair the previous week. No relationship between religion and
affairs among women.
Isiugo-Abanihe (1994)
P.B. Gray / Social Science & Medicine 58 (2004) 1751–17561754

5. Conclusions
The hypothesis that Islamic religious affiliation is
negatively associated with HIV seropositivity is gener-
ally supported. The percentage of Muslims negatively
and significantly predicted the prevalence of HIV among
sub-Saharan African countries. Six of seven studies
enabling within-population comparisons revealed lower
HIV prevalence among Muslims. Examination of HIV
risk factors and HIV yielded more mixed results: Islamic
religious affiliation sometimes appeared to, but other
times not to, be associated with a reduction in sex
outside marriage (e.g., with commercial sex workers or
within extramarital affairs). However, in no study was
Islamic affiliation positively associated with such sexual
behaviours. Muslims seemed to have lower alcohol
consumption (which might partly underlie any differ-
ences in risky sexual behaviour) and higher rates of
circumcision compared with non-Muslims, but Islam’s
allowance for polygyny and discouragement of condom
use would work against reduced sexual transmission of
HIV. Future discussions of the predicted course of the
global HIV epidemic may consider Islamic religious
affiliation as a socio-demographic factor associated with
a reduced risk of HIV transmission.
Acknowledgements
I thank Robert C. Bailey, Frank Marlowe, Martin
Muller, Megan O’Connell Gray, John Polk, Richard
Sosis, and Richard Wrangham for helpful feedback. For
inspiring this research, I thank the men in Lamu, Kenya
who I interviewed during a project on marriage,
parenting, work, and hormones.
References
Auvert, B., Buve, A., Ferry, B., Carael, M., Morison, L., &
Lagarde, E., et al. (2001). Ecological and individual level
analysis of risk factors for HIV infection in four urban
populations in sub-Saharan Africa with different levels of
HIV infection. AIDS, 15(S4), S15–S30.
Bailey, R. C., Neema, S., & Othieno, R. (1999). Sexual
behaviors and other HIV risk factors in circumcised
and uncircumcised men in Uganda. Journal of AIDS,
22(3), 294–301.
Bailey, R. C., Plummer, F. A., & Moses, S. (2001). Male
circumcision and HIV prevention: Current knowledge and
future research directions. Lancet Infectious Diseases, 1(4),
223–231.
Bastani, R., Erickson, P. A., Marcus, A. C., Maxwell, A. E.,
Capell, F. J., & Freeman, H., et al. (1996). AIDS-related
attitudes and risk behaviours: A survey of a random
sample of California heterosexuals. Preventive Medicine,
25(2), 105–117.
Blanc, A. K., & Gage, A. J. (2000). Men, polygyny, and fertility
over the life-course in sub-Saharan Africa. In C. Bledsoe,
S. Lerner, & J. I. Guyer (Eds.), Fertility and the male life-
cycle in the era of fertility decline. Oxford: Oxford University
Press.
Bwayo, J., Plummer, F., Omari, M., Mutere, A., Moses, S., &
Ndinya-Achola, J., et al. (1994). Human immunodeficiency
virus infection in long-distance truck drivers in East Africa.
Archives of Internal Medicine, 154(12), 1391–1396.
Ellison, C. G., & Levin, J. S. (1998). The religion-health
connection: Evidence, theory, and future directions. Health
Education and Behavior, 25(6), 700–720.
Gayle, H. D., & Hill, G. L. (2001). Global impact of human
immunodeficiency Virus and AIDS. Clinical Microbiology
Reviews, 14(2), 327–335.
Gibney, L., Choudhury, P., Khawaja, Z., Sarker, M., &
Vermund, S. H. (1999). Behavioural risk factors for HIV/
AIDS in a low-HIV prevalence Muslim nation: Bangladesh.
International Journal of STD and AIDS, 10(3), 186–194.
Gray, R. H., Kiwanuka, N., Quinn, T., Sewankambo, N. K.,
Serwadda, D., & Wabwire Mangen, F., et al. (2000). Male
circumcision and HIV acquisition and transmission: Cohort
studies in Rakai, Uganda. AIDS, 14(15), 2371–2381.
Holmes, E. C. (2001). On the origin and evolution of the human
immunodeficiency virus (HIV). Biological Reviews, 76(2),
239–254.
Isiugo-Abanihe, U. C. (1994). Extramarital relations and
perceptions of HIV/AIDS in Nigeria. Health Transition
Review, 4(2), 111–125.
Kagimu, M., Marum, E., Wabwire-Mangen, F., Nakyanjo, N.,
Walakira, Y., & Hogle, J. (1998). Evaluation of the
effectiveness of AIDS health education intervention in the
Muslim community in Uganda. AIDS Education and
Prevention, 10(3), 215–228.
Killewo, J., Dahlgren, L., & Sandstrom, A. (1994). Socio-
geographical patterns of HIV-1 transmission in Kagera
region, Tanzania. Social Science & Medicine, 38(1),
129–134.
Lagarde, E., Enel, C., Seck, K., Gueye-Ndiaye, A., Piau, J-P., &
Pison, G., et al. (2000). Religion and protective behaviours
towards AIDS in rural Senegal. AIDS, 14(13), 2027–2033.
Lenton, C. (1997). Will Egypt escape the AIDS epidemic?
Lancet, 349(9057), 1005.
Lerman, S. E., & Liao, J. C. (2001). Neonatal circumcision.
Pediatric Clinics of North America, 48, 1539–1557.
Malamba, S. S., Wagner, H-U., Maude, G., Okongo, M.,
Nunn, A. J., & Kengeya-Kayondo, J. F., et al. (1994). Risk
factors for HIV-1 infection in adults in a rural Ugandan
community: A case-control study. AIDS, 8(2), 253–257.
Mbulaiteye, S. M., Ruberantwari, A., Nakiyingi, J. S.,
Carpenter, L. M., Kamali, A., & Whitworth, J. A. G.
(2000). Alcohol and HIV: A study among sexually active
adults in rural southwest Uganda. International Journal of
Epidemiology, 29(15), 911–915.
Nnko, S., Washija, R., Urassa, M., & Boerma, J. T. (2001).
Dynamics of male circumcision in northwest Tanzania.
Sexually Transmitted Diseases, 28(4), 214–218.
Nunn, A. J., Kengeya-Kayondo, J. F., Malamba, S. S.,
Seeley, J. A., & Mulder, D. W. (1994). Risk factors for
HIV-1-infection in adults in a rural Ugandan community:
A population study. AIDS, 8(1), 81–86.
ARTICLE IN PRESS
P.B. Gray / Social Science & Medicine 58 (2004) 1751–1756 1755

6. Piot, P., Bartos, M., Ghys, P. D., Walker, N., & Schwartlander,
B. (2001). The global impact of HIV/AIDS. Nature,
410(6831), 968–973.
Rakwar, J., Lavreys, L., Thompson, M., Jackson, D., Bwayo, J.,
& Hassanali, S., et al. (1999). Cofactors for the acquisition of
HIV-1 among heterosexual men: Prospective cohort study of
trucking company workers in Kenya. AIDS, 13(5), 607–614.
Reynolds, V., & Tanner, R. (1995). The social ecology of
religion. Oxford: Oxford University Press.
Ridanovic, Z. (1997). AIDS and Islam. Medicinski Archiv, 51,
45–46.
Schwartlander, B., Stanecki, K. A., Brown, T., Way, P. O.,
Monasch, R., & Chin, J., et al. (1999). Country-specific
estimates and models of HIV and AIDS: Methods and
limitations. AIDS, 13(17), 2445–2458.
Seidman, S. N., Mosher, M. D., & Aral, S. O. (1992). Women
with multiple sexual partners: United States, 1988. American
Journal of Public Health, 82(10), 1388–1394.
Stanberrry, L. R., & Bernstein, D. I. (2000). Sexually
transmitted diseases: vaccines, prevention and control. New
York: Academic Press.
Wasserheit, J. N., Aral, S. O., & Holmes, K. (Eds.), &
Hitchcock (associate Ed.) (1991). Research issues in
human behavior and sexually transmitted diseases in
the AIDS era. Washington, DC: American Society for
Microbiology.
Weiss, H. A., Quigley, M. A., & Hayes, R. J. (2000). Male
circumcision and risk of HIV infection in sub-Saharan
Africa: A systematic review and meta-analysis. AIDS,
14(15), 2361–2370.
Wilson, D., Lavelle, B., Mwoboto, B., & Armstrong, M.
(1992). Use of a retrospective timeline calendar to examine
alcohol use, sex behaviour and condom use among Zimbab-
wean men. International Conference on AIDS, 8, C332.
World Almanac and book of facts (2000). Mahwah, NJ: World
Almanac Books.
ARTICLE IN PRESS
P.B. Gray / Social Science & Medicine 58 (2004) 1751–17561756