Sex Differences in Tuberculosis Burden and Notifications in Low- and Middle-Income Countries: A Systematic Review and Meta-analysis

Abstract

Background: Tuberculosis (TB) case notification rates are usually higher in men than in women, but notification data are insufficient to measure sex differences in disease burden. This review set out to systematically investigate whether sex ratios in case notifications reflect differences in disease prevalence and to identify gaps in access to and/or utilisation of diagnostic services.

Methods and findings: In accordance with the published protocol (CRD42015022163), TB prevalence surveys in nationally representative and sub-national adult populations (age ≥ 15 y) in low- and middle-income countries published between 1 January 1993 and 15 March 2016 were identified through searches of PubMed, Embase, Global Health, and the Cochrane Database of Systematic Reviews; review of abstracts; and correspondence with the World Health Organization. Random-effects meta-analyses examined male-to-female (M:F) ratios in TB prevalence and prevalence-to-notification (P:N) ratios for smear-positive TB. Meta-regression was done to identify factors associated with higher M:F ratios in prevalence and higher P:N ratios. Eighty-three publications describing 88 surveys with over 3.1 million participants in 28 countries were identified (36 surveys in Africa, three in the Americas, four in the Eastern Mediterranean, 28 in South-East Asia and 17 in the Western Pacific). Fifty-six surveys reported in 53 publications were included in quantitative analyses. Overall random-effects weighted M:F prevalence ratios were 2.21 (95% CI 1.92-2.54; 56 surveys) for bacteriologically positive TB and 2.51 (95% CI 2.07-3.04; 40 surveys) for smear-positive TB. M:F prevalence ratios were highest in South-East Asia and in surveys that did not require self-report of signs/symptoms in initial screening procedures. The summary random-effects weighted M:F ratio for P:N ratios was 1.55 (95% CI 1.25-1.91; 34 surveys). We intended to stratify the analyses by age, HIV status, and rural or urban setting; however, few studies reported such data.

Conclusions: TB prevalence is significantly higher among men than women in low- and middle-income countries, with strong evidence that men are disadvantaged in seeking and/or accessing TB care in many settings. Global strategies and national TB programmes should recognise men as an underserved high-risk group and improve men's access to diagnostic and screening services to reduce the overall burden of TB more effectively and ensure gender equity in TB care.

Fig 1. PRISMA flow diagram.

Mtb, Mycobacterium tuberculosis.

Fig 2. Global map showing countries in which prevalence surveys have been conducted.

Yellow indicates low- and middle-income countries for which sex-disaggregated data are available from at least one prevalence survey (n = 24). Red indicates low- and middle-income countries in which at least one prevalence survey has been conducted but sex-disaggregated data are not available (n = 4). Dark gray indicates low- and middle-income countries where no prevalence survey has been identified (n = 107). Labels show the total number of surveys identified within each country for which at least one prevalence survey was identified (n = 88).

Fig 3. Male-to-female ratios of participation among eligible or invited individuals (n = 29).

Analysis includes surveys that report the number of individuals who were eligible for screening and the number of individuals screened by sex. See S2 Table for survey details and references. Lao PDR, Lao People’s Democratic Republic.

Fig 4. Male-to-female ratios in bacteriologically positive (n = 56) and smear-positive (n = 40) TB prevalence by WHO region.

Regional and overall summaries from random-effects models for bacteriologically positive and smear-positive TB. See S2 Table for survey details and references. Lao PDR, Lao People’s Democratic Republic.

Fig 5. Random-effects weighted male-to-female prevalence ratios for bacteriologically positive TB by age group (n = 19).

Analysis includes surveys that report the number of individuals screened and the number of bacteriologically positive TB cases by sex and age. Horizontal axis shows age groups in years. Vertical axis shows random-effects weighted M:F ratios in prevalence of bacteriologically positive TB per 100,000 individuals with 95% confidence intervals.

Fig 6. Male-to-female ratios in prevalence-to-notification ratios (n = 34).

Analysis includes surveys that report the prevalence of smear-positive TB by sex and for which corresponding national notification and population data are available. See S2 Table for survey details and references.