Editor—As Parry et al point out,¹,² appropriate laboratory methods are crucial in detecting clinically important quinolone resistance.

We highlight the emergence of strains ofSalmonella enterica serovar Typhi (S Typhi) that show reduced susceptibility to the fluoroquinolones but are susceptible to nalidixic acid (minimum inhibitory concentration < 16 mg/l). In a review of all 692 isolates ofS Typhi sent to the Laboratory for Enteric Pathogens at the Health Protection Agency in London between 2000 and 2003 we detected 49 isolates that were susceptible to nalidixic acid but had reduced susceptibility to fluoroquinolone (minimum inhibitory concentration 0.125-1.0 mg/l;table).³ When the country of acquisition was known, 18 of these isolates were from patients who had visited India; eight, Pakistan; four, Bangladesh; and one, Kenya.

Overall, of 271 isolates with reduced susceptibility to fluoroquinolone, 18% were susceptible to nalidixic acid and therefore would not have been detected by routine screening with a nalidixic acid disc. No clinical outcome data are available and so the clinical relevance of these strains is uncertain, but an enhanced surveillance study is being undertaken by the Health Protection Agency.⁴

Although resistance to nalidixic acid (minimum inhibitory concentration > 256 mg/l) remains an important marker for failure of fluoroquinolone treatment in typhoid fever, several isolates show reduced susceptibility to fluoroquinolone while remaining susceptible to nalidixic acid. This is particularly true for isolates from the Indian subcontinent. Furthermore, in six of the 49 such isolates (table), none of the mutations in gyrA and parC commonly associated with reduced resistance to fluoroquinolone were present using the GAMA method.⁵ Therefore a new mechanism of fluoroquinolone resistance, which might not confer nalidixic acid resistance, may be emerging inS Typhi. Our findings underscore the importance of estimating the minimum inhibitory concentration of the agent selected for treatment.