Spatial and temporal clustering of typhoid fever in an urban slum of Dhaka City: Implications for targeted typhoid vaccination

by Faisal Ahmmed, Farhana Khanam, Md Taufiqul Islam, Deok Ryun Kim, Sophie Kang, Md Golam Firoj, Asma Binte Aziz, Masuma Hoque, Xinxue Liu, Hyon Jin Jeon, Suman Kanungo, Fahima Chowdhury, Ashraful Islam Khan, Khalequ Zaman, Florian Marks, Jerome H. Kim, Firdausi Qadri, John D. Clemens, Birkneh Tilahun Tadesse, Justin Im

Salmonella enterica serotype Typhi (Salmonella Typhi) causes severe and occasionally life-threatening disease, transmitted through contaminated food and water. Humans are the only reservoir, inadequate water, sanitation, and hygiene infrastructure increases risk of typhoid. High-quality data to assess spatial and temporal relationships in disease dynamics are scarce.

We analyzed data from a prospective cohort conducted in an urban slum area of Dhaka City, Bangladesh. Passive surveillance at study centers identified typhoid cases by microbiological culture. Each incident case (index case) was matched to two randomly selected index controls, and we measured typhoid incidence in the population residing in a geographically defined region surrounding each case and control. Spatial clustering was evaluated by comparing the typhoid incidence in residents of geometric rings of increasing radii surrounding the index cases and controls over 28 days. Temporal clustering was evaluated by separately measuring incidence in the first and second 14-day periods following selection. Incidence rate ratios (IRRs) were calculated using Poisson regression models.

We evaluated 141 typhoid index cases. The overall typhoid incidence was 0.44 per 100,000 person-days (PDs) (95% CI: 0.40, 0.49). In the 28 days following selection, the highest typhoid incidence (1.2 per 100,000 PDs [95% CI: 0.8, 1.6]) was in the innermost cluster surrounding index cases. The IRR in this innermost cluster was 4.9 (95% CI: 2.4, 10.3) relative to the innermost control clusters. Neither typhoid incidence rates nor relative IRR between index case and control populations showed substantive differences in the first and second 14-day periods after selection.

In the absence of routine immunization programs, geographic clustering of typhoid cases suggests a higher intensity of typhoid risk in the population immediately surrounding identified cases. Further studies are needed to understand spatial and temporal trends and to evaluate the effectiveness of targeted vaccination in disrupting typhoid transmission.