Gastroenteritis due to typhoidal Salmonella: a decade of observation at an urban and a rural diarrheal disease hospital in Bangladesh.

BACKGROUND: The study aimed to compare the socio-demographic, host and clinical characteristics, seasonality and antimicrobial susceptibility of Typhoidal Salmonella (Salmonella enterica serovar Typhi and Paratyphi) (TS) with diarrhea between urban and rural Bangladesh.

METHODS: Relevant information of 77/25,767 (0.30%) and 290/17,622 (1.65%) patients positive with TS (in stool) were extracted from the data archive of Diarrheal Disease Surveillance System of icddr,b (urban Dhaka and rural Matlab Hospitals respectively) during 2000-2012. Comparison group (diarrhea patients negative for TS) was randomly selected from the database (1:3 ratio). Two poisson regression models were investigated for modelling seasonal effects on the number of cases.

RESULTS: Salmonella Typhi was more frequently isolated in Dhaka than Matlab (57% vs. 5%, p < 0.001); while Salmonella Paratyphi was more frequent in Matlab than Dhaka (96% vs. 43%; p < 0.001). Fever [adj. OR-5.86 (95% CI: 2.16, 15.94)], antimicrobial use at home [5.08 (2.60, 9.90)], and fecal red blood cells [2.53 (1.38, 4.64)] were significantly associated with detection of TS in stool of patient from Dhaka. For Matlab, the correlates were, vomiting [1.88 (1.35, 2.64)], fecal macrophage [1.89 (1.29, 2.74)] in addition to fever and duration of diarrhea and antimicrobial use. At Dhaka, all Salmonella Typhi isolates were susceptible to ceftriaxone; while in Dhaka and Matlab however, for ciprofloxacin it was 45% and 91%, respectively. Susceptibility to chloramphenicol, ampicillin, trimethoprim-sulphamethoxazole and nalidixic acid ranged from 12%-58%. Salmonella Paratyphi were susceptible to ceftriaxone (99%). A significant seasonal trend and year difference (before and after 2007) for Matlab was observed (p < 0.001 for all effects). Dhaka does not show significant year or seasonal effects (p = 0.07 for years and p = 0.81 and p = 0.18 for the cos and sin components, respectively). While not significant, two seasonal peaks were observed in Dhaka (January-February and September-November); while a single peak (August-November) was observed in Matlab.

CONCLUSIONS: Proportion of serovar distribution of TS and their clinical characteristics, antimicrobial susceptibility and seasonal pattern were different among diarrhea patients in urban Dhaka and rural Matlab of Bangladesh.