0% (99%

CI: 1.6–2.4) (Figure 2). Estimates of cumulative incidence among the nine studies and data sources ranged from 0.96% to 3.59%. The overall incidence density was 2.9 conversions per 1000 person-months (99% CI: 2.5–3.4). The cumulative incidence scatter plot shows that the risk of conversion was relatively constant over the average duration of travel seen in the studies (Figure Romidepsin 3). In contrast, the incidence density scatter plot appears to demonstrate a decrease in conversion rates as average travel duration increased (Figure 4). Calculation of an incidence density rate assumes that the rate of infection is constant over the interval studied, but the data in Figure 4 violate that assumption. Therefore, the remaining analyses use only the cumulative incidence measures. There was marked heterogeneity among studies estimating cumulative incidence (χ2 heterogeneity statistic, p < 0.0001). Attempts to explain this heterogeneity for most variables

was limited due to the small number of studies in each subgroup and limited data on other risk factors for TB infection, but stratification was used to explore this heterogeneity to the extent possible. Examination of meta-influence see more (not shown) suggested that no single study substantially affected the overall estimate. Exclusion of the large US Army data set from MEDPROS and the Navy study by Bowman increased the cumulative incidence estimate to 2.3% (99% CI: 2.0–2.7).30 Mannose-binding protein-associated serine protease When stratifying by military or civilian studies, the cumulative incidence risk estimate was 2.0% (99% CI: 1.6–2.4) for military studies and 2.3% (99% CI: 2.1–2.5) for civilian studies. Stratifying the analysis by published and unpublished studies resulted in a cumulative incidence of 2.0% (99% CI: 1.6–2.4) for published studies and 2.0% (99% CI: 1.0–3.1) for unpublished studies. Stratifying by travel to recent conflicts in SWA only versus travel elsewhere resulted in an estimated cumulative incidence of 1.7% (99% CI: 0.6–2.9) for data from SWA and 2.3% (99%

CI: 1.6–3.0) from all other locations. Stratifying by deployments from North America (United States and Canada) versus deployments from other countries resulted in a cumulative incidence of 1.9% (99% CI: 1.5–2.4) for North America and 2.5% (99% CI: 1.2–3.8) for others. Finally, temporal trends were considered by stratifying the analysis by data sources which only contained military data after 2001, which marked the beginning of Operation Enduring Freedom (OEF) combat operations in Afghanistan, compared to those civilian and military sources obtained prior to 2001. This resulted in estimates of 2.0% (99% CI: 1.0–3.1) for data after 2001 and 2.1% (99% CI: 1.4–2.9) for data sources including travel from before 2001.