After a baseline census, surveillance was conducted in study sites in Karachi, Pakistan; Calcutta, India; North Jakarta, Indonesia; and Hechi City, China, for 12 months to identify typhoid and paratyphoid cases from specific populations at high risk (Table). None of the sites had specific enteric fever control programs in the past. Hechi City, China, is located in Guangxi Zhuang Autonomous Region where Vi vaccines had been used in the past (8,9); however, no such intervention had taken place in Hechi City. The closest county with a vaccination program was ≈80 km away and vaccinated only students (29,000 doses in 2001).

During the surveillance period, persons with fever who lived in each study area were requested to visit participating healthcare providers. We collected 5–10 mL blood from adults with fever >3 days' duration into Bactec bottles (Becton Dickinson, Franklin Lakes, NJ, USA). We collected 3–8 mL from children with fever >3 days' duration into Pediatric Bactec bottles. The bottles were incubated at 37°C for 7–10 days and visually checked for growth every day. Bottles were subcultured on MacConkey agar on days 1, 2, 4, and 7 or when turbidity was detected. Suspected colonies were screened by using Kligler iron agar, sulfide-indole-motility medium, urea agar, and citrate. Colonies that showed biochemical reactions suggestive of salmonellae were confirmed serologically by Felix-Widal tube agglutination test with specific O and H antisera (Becton Dickinson). All Salmonella isolates were confirmed at a reference laboratory (University of Oxford, Wellcome Trust Clinical Research Unit, Ho Chi Minh City, Vietnam).

Incidence rates were calculated by using age-specific denominators of the population living in the catchment area based on the study census. We assumed that each person living in the study area contributed 12 months of person-time to the denominator. The number of disease episodes in eligible individuals was used as the numerator.

Each study received individual approval from the local ethical committees, the institutional review board of the International Vaccine Institute (Seoul, Korea), and the Secretariat Committee for Research Involving Human Subjects, World Health Organization (Geneva, Switzerland).

During the surveillance period, 285 S. Typhi episodes and 84 S. Paratyphi A episodes were detected at the 4 sites (Table). In Indonesia, 14% of enteric fever episodes were caused by S. Paratyphi A, in Pakistan 15%, in India 24%, and in China 64% (Figure). The highest S. Typhi incidence was observed in Pakistan (394/100,000/year), and the lowest S. Typhi incidence was found in China (15.2/100,000/year). The highest S. Paratyphi A incidence was also seen in Pakistan (72/100,000/year), and the lowest S. Paratyphi A incidence was seen in Indonesia (13.7/100,000/year).

The perception that a small proportion of enteric fever cases are caused by S. Paratyphi A is probably no longer true in many regions of Asia, especially in southeast China, where S. Paratyphi A is already more frequently isolated than is S. Typhi. This finding could be signaling the emergence of S. Paratyphi A as a pathogen in Asia. Comparison of S. Paratyphi A incidence during the last decade is needed to prove this hypothesis. However, none of the sites have comparable surveillance data on S. Paratyphi A over time. An alternative explanation is that the incidence of S. Typhi is decreasing. Previous reports from vaccine trials have shown successful control of S. Typhi but no changes in the incidence of S. Paratyphi A (8,10,11). Nonetheless, the reversal of the proportion of S. Typhi and S. Paratyphi A infections in Hechi City, China, is unlikely to be the result of typhoid fever control in other counties, considering the distance and oral-fecal transmission route of S. Typhi and S. Paratyphi A.

Economic growth in Asia has also resulted in improved water supply, sanitation, and hygiene; however, we have no reason to assume selective reduction only for transmission of S. Typhi. The incidence of typhoid fever per 100,000 in the different countries follows the same pattern as mean gross national income (2003) and the mortality ranking for children <5 years (12).

The sensitivity of blood culture for S. Typhi is well described, but little is known about S. Paratyphi A. Bacterial loads during infection are probably similar for both S. Typhi and S. Paratyphi A (J. Wain, unpub. data). Furthermore, S. Paratyphi A had been rarely isolated in these regions, which suggests that the increase in isolation rate of S. Paratyphi A is possibly caused by an increase in the number of cases of enteric fever caused by S. Paratyphi A rather than any bias toward blood culture–positive disease.

Besides the limitation that our studies only describe a 12-month period, the population varied between sites. In Pakistan, only children 2–16 years of age were included. As S. Paratyphi A infections are more frequently observed in adults, including older patients in surveillance may increase S. Paratyphi A incidence rates reported from Pakistan.

In China and India, countries with the largest populations in the world, S. Paratyphi A is the causal agent for a substantial proportion of enteric fever episodes that cannot be distinguished clinically from typhoid fever episodes. While similar treatment strategies may work for both organisms, future enteric fever prevention strategies in Asia must focus on S. Paratyphi A as well as on S. Typhi, especially when considering the emergence of drug-resistant strains (13–15). Future vaccination strategies should include bivalent vaccines that protect against S. Typhi as well as S. Paratyphi A. Otherwise, the protective effectiveness of typhoid fever vaccines (Vi, Ty21a) against enteric fever may diminish, which could result in a loss of public confidence and decrease public willingness to be vaccinated.