Emergence of nalidixic acid-resistant Shigella sonnei in acute-diarrhea patients on Andaman and Nicobar Islands, India.

Shigella infection has been occurring in both epidemic and sporadic forms on Andaman and Nicobar Islands in India (92 to 94°E longitude, 6 to 14°N latitude; Bay of Bengal). Although acute gastrointestinal infection clinically resembling shigellosis has been reported on the islands frequently, bacteriological surveillance was only started in 1994. The surveillance data showed that shigellosis is a common cause of bacterial diarrhea (1). Endemic infection is mainly due to Shigella flexneri 2a, whereas S. dysenteriae type 1 was responsible for the outbreaks that occurred among the general population and among the Andamanese tribe. The ampicillinand co-trimoxazole-resistant pandemic strain of S. dysenteriae type 1 was first observed during an epidemic that occurred in 1986 (4). By 1994, 100% of the Shigella isolates had become ampicillin resistant and 80.6% had become co-trimoxazole resistant. All co-trimoxazole-resistant strains were also resistant to ampicillin. Nalidixic acid resistance was first observed among S. dysenteriae type 1 strains in 1995 (15.1%), and gentamicin resistance was first observed in 1996 (37%) (2). The proportions of different strains among the isolates have been showing considerable variability from year to year. During 1994 and 1995 to 1999 and 2000, a total of 15 S. sonnei strains were isolated from acute-diarrhea patients. All of the isolates were sensitive to nalidixic acid (Table 1). Since September 1999 to August 2000, a rural-communitybased study was conducted to estimate the disease burden due to shigellosis. Three S. sonnei (phase I) isolates were recovered from patients who were identified in a follow-up study. The first isolate was sensitive to nalidixic acid, but two subsequent isolates were resistant. During 2001 and 2002, 106 samples from acute-diarrhea cases were processed. This yielded nine isolates (isolation rate, 8.49%). Seven were S. sonnei (phase I), one was S. flexneri 3a, and one was S. flexneri 4b. All seven of the S. sonnei isolates were resistant to nalidixic acid (30 g/disk; HiMedia), ampicillin, and co-trimoxazole (the dates of isolation of the first and last reported isolates were 27.06.01 and 15.03.02, respectively), whereas the S. flexneri isolates were sensitive to nalidixic acid. The emergence of nalidixic acid-resistant S. sonnei has been observed since the last week of September 1999 (29.09.99) and has continued during the reported study period. The MIC of nalidixic acid was estimated by using both the E-test (AB Biodisk) and agar dilution (3) methods. It ranged from 24 to more than 96 g/ml (mean, 40.88 26.06 g/ml) among the resistant strains. S. sonnei was also isolated in previous years (1, 2); however, those strains were sensitive to nalidixic acid. Besides nalidixic acid, ciprofloxacin and furazolidone are used in the treatment of bacillary dysentery in the hospital. Most of the S. sonnei isolates were isolated from children 5 years old admitted to G. B. Pant Hospital, the only referral hospital in Port Blair. The youngest child was 4 months old. They had fever, abdominal pain, and watery stools with mucus. During 2000 to 2001 and 2001 to 2002, S. flexneri 2a and S. dysenteriae type 1 were not isolated either from the community or from hospitalized patients. It appears that the multidrugresistant S. sonnei strain is replacing the endemic S. flexneri 2a and epidemic S. dysenteriae type 1 strains as the predominant strain of Shigella in Port Blair. The patients who were infected with this strain of S. sonnei were from different parts of South Andaman, indicating a wide distribution of the strain in Andamans. A shift in the predominant Shigella strain, usually from S. flexneri 2a to S. dysenteriae type 1, precedes shigellosis epidemics. The present shift from S. flexneri 2a to S. sonnei may also be an early signal of an emerging important problem.