Incidence of typhoid and paratyphoid fever in Bangladesh, Nepal, and Pakistan: results of the Surveillance for Enteric Fever in Asia Project

Abstract

Background: Precise enteric fever disease burden data are needed to inform prevention and control measures, including the use of newly available typhoid vaccines. We established the Surveillance for Enteric Fever in Asia Project (SEAP) to inform these strategies.

Methods: From September, 2016, to September, 2019, we conducted prospective clinical surveillance for Salmonella enterica serotype Typhi (S Typhi) and Paratyphi (S Paratyphi) A, B, and C at health facilities in predetermined catchment areas in Dhaka, Bangladesh; Kathmandu and Kavrepalanchok, Nepal; and Karachi, Pakistan. Patients eligible for inclusion were outpatients with 3 or more consecutive days of fever in the last 7 days; inpatients with suspected or confirmed enteric fever; patients with blood culture-confirmed enteric fever from the hospital laboratories not captured by inpatient or outpatient enrolment and cases from the laboratory network; and patients with non-traumatic ileal perforation under surgical care. We used a hybrid surveillance model, pairing facility-based blood culture surveillance with community surveys of health-care use. Blood cultures were performed for enrolled patients. We calculated overall and age-specific typhoid and paratyphoid incidence estimates for each study site. Adjusted estimates accounted for the sensitivity of blood culture, the proportion of eligible individuals who consented and provided blood, the probability of care-seeking at a study facility, and the influence of wealth and education on care-seeking. We additionally calculated incidence of hospitalisation due to typhoid and paratyphoid.

Findings: A total of 34 747 patients were enrolled across 23 facilitates (six tertiary hospitals, surgical wards of two additional hospitals, and 15 laboratory network sites) during the study period. Of the 34 303 blood cultures performed on enrolled patients, 8705 (26%) were positive for typhoidal Salmonella. Adjusted incidence rates of enteric fever considered patients in the six tertiary hospitals. Adjusted incidence of S Typhi, expressed per 100 000 person-years, was 913 (95% CI 765-1095) in Dhaka. In Nepal, the adjusted typhoid incidence rates were 330 (230-480) in Kathmandu and 268 (202-362) in Kavrepalanchok. In Pakistan, the adjusted incidence rates per hospital site were 176 (144-216) and 103 (85-126). The adjusted incidence rates of paratyphoid (of which all included cases were due to S Paratyphi A) were 128 (107-154) in Bangladesh, 46 (34-62) and 81 (56-118) in the Nepal sites, and 23 (19-29) and 1 (1-1) in the Pakistan sites. Adjusted incidence of hospitalisation was high across sites, and overall, 2804 (32%) of 8705 patients with blood culture-confirmed enteric fever were hospitalised.

Interpretation: Across diverse communities in three south Asian countries, adjusted incidence exceeded the threshold for "high burden" of enteric fever (100 per 100 000 person-years). Incidence was highest among children, although age patterns differed across sites. The substantial disease burden identified highlights the need for control measures, including improvements to water and sanitation infrastructure and the implementation of typhoid vaccines.

Funding: Bill & Melinda Gates Foundation.

Figure 1

Recruitment, eligibility, study consent, and laboratory culture positivity *Two detected cases of S Paratyphi B were excluded. †Population for the crude incidence rates includes age-eligible patients from within the catchment area recruited from outpatient departments, inpatient departments, surgical wards, hospital laboratories, and laboratory network sites. ‡Population for the adjusted incidence rates includes age-eligible patients from within the catchment area recruited from outpatient departments, inpatient departments, and surgical wards.

Figure 2

Unadjusted and adjusted incidence rates of laboratory-confirmed typhoid and paratyphoid cases per 100 000 person-years, by adjustment factor and by study site Adjustments are as follows: adjustment 1=culture sensitivity; adjustment 2=study consent; adjustment 3=care-seeking; adjustment 4=differential care-seeking. The points indicate the median estimate and shaded areas indicate the density. The numbers to the right of each graph indicate the incidence estimate and the 95% confidence interval for each adjustment.

Figure 3

Adjusted incidence estimates of typhoid and paratyphoid cases per 100 000 person-years, by age group and by study site The points indicate the median estimate and shaded areas indicate the density. The numbers to the right of each graph indicate the incidence estimate and the 95% confidence interval for each age group. There are no data plotted for Salmonella Typhi and Paratyphi A incidence among people aged 16–25 years and >25 years for the Bangladesh sites because these were paediatric facilities. There are no data plotted for Salmonella Paratyphi A incidence among people aged <2 years and 2–4 years in Dulikhel Hospital (Nepal), aged <2 years in Kathmandu Medical College (Nepal), and aged 5–15 years, 16–25 years, and >25 years in Kharadar General Hospital (Pakistan), because there were no observed cases for these age groups and therefore insufficient data to estimate incidence.