Background Widespread use of fluoroquinolones has resulted in emergence of Salmonella typhi strains with decreased susceptibility to fluoroquinolones. These strains are identifiable by their nalidixic acid-resistance. We studied the impact of infection with nalidixic acid-resistant S. typhi (NARST) on clinical outcomes in patients with bacteriologically-confirmed typhoid fever. Methods Clinical and laboratory features, fever clearance time and complications were prospectively studied in patients with blood culture-proven typhoid fever, treated at a tertiary care hospital in north India, during the period from November 2001 to October 2003. Susceptibility to amoxycillin, co-trimoxazole, chloramphenicol, ciprofloxacin and ceftriaxone were tested by disc diffusion method. Minimum inhibitory concentrations (MIC) of ciprofloxacin and ceftriaxone were determined by E-test method. Results During a two-year period, 60 patients (age [mean ± SD]: 15 ± 9 years; males: 40 [67%]) were studied. All isolates were sensitive to ciprofloxacin and ceftriaxone by disc diffusion and MIC breakpoints. However, 11 patients had clinical failure of fluoroquinolone therapy. Infections with NARST isolates (47 [78%]) were significantly associated with longer duration of fever at presentation (median [IQR] 10 [7-15] vs. 4 [3-6] days; P = 0.000), higher frequency of hepatomegaly (57% vs. 15%; P = 0.021), higher levels of aspartate aminotransferase (121 [66–235] vs. 73 [44–119] IU/L; P = 0.033), and increased MIC of ciprofloxacin (0.37 ± 0.21 vs. 0.17 ± 0.14 μg/mL; P = 0.005), as compared to infections with nalidixic acid-susceptible isolates. All 11 patients with complications were infected with NARST isolates. Total duration of illness was significantly longer in patients who developed complications than in patients who did not (22 [14.8–32] vs. 12 [9.3–20.3] days; P = 0.011). Duration of prior antibiotic intake had a strong positive correlation with the duration of fever at presentation (r = 0.61; P = 0.000) as well as the total duration of illness (r = 0.53; P = 0.000). Conclusion Typhoid fever caused by NARST infection is associated with poor clinical outcomes, probably due to delay in initiating appropriate antibiotic therapy. Fluoroquinolone breakpoints for S. typhi need to be redefined and fluoroquinolones should no longer be used as first-line therapy, if the prevalence of NARST is high.