In vitro activity of azithromycin, newer quinolones and cephalosporins in ciprofloxacin-resistant Salmonella causing enteric fever

In India, ciprofloxacin has been the treatment of choice for enteric fever for over a decade. Unfortunately, after its introduction there was a decrease in the susceptibility to this drug, with consequent therapeutic failure (Capoor et al., 2006; Cooke et al., 2006; Kownhar et al., 2007; Renuka et al., 2005; Saha et al., 2006). This was detected in the laboratory as nalidixic acid resistance. It was followed by the widespread occurrence of isolates with decreased susceptibility to ciprofloxacin (indicated by nalidixic acid resistance, with ciprofloxacin MICs of 0.125–1.0 µg ml^–1), associated with an impaired response to ciprofloxacin treatment, and the subsequent emergence of highly resistant isolates (ciprofloxacin MICs >1.0 µg ml^–1). High-level ciprofloxacin-resistant enteric fever has evolved in Asian countries, including India (Capoor et al., 2006; Cooke et al., 2006; Gaind et al., 2006; Kownhar et al., 2007; Pokharel et al., 2006; Renuka et al., 2005; Saha et al., 2006). Resistance is also emerging to extended-spectrum cephalosporins (Capoor et al. 2006; Pokharel et al., 2006; Saha et al., 1999). These alternative regimens have several disadvantages such as expense, the intravenous route of administration required and prolonged defervescence time (Saha et al., 2006).

Azithromycin, a broad-spectrum azilide, has prolonged intracellular concentrations and a prolonged half-life. A large number of experimental and clinical trials carried out in the 1990s support its efficacy in traditional multidrug-resistant (ampicillin, chloramphenicol, co-trimoxazole and tetracycline) enteric fever (Butler et al., 1999; Girgis et al., 1999; Gordillo et al., 1993). However, studies reporting the MICs of azithromycin and newer quinolones in the current scenario of ciprofloxacin-resistant enteric fever are scarce (Frenck et al., 2004; Parry, 2004; Parry et al., 2007). The aim of this study was to determine the in vitro MIC patterns of various therapeutic alternatives available for the treatment of enteric fever in an endemic region reporting a recent increase in ciprofloxacin resistance.

Quinolone-resistant Salmonella isolates were recovered from patients with enteric fever admitted to a 1570-bed tertiary care centre at Vardhman Mahavir Medical College and Safdarjung Hospital and Majeedia Hospital, in New Delhi, India, from December 2004 to December 2006. Of 384 isolates of Salmonella enterica serovar Typhi (S. Typhi) and S. enterica serovar Paratyphi A (S. Paratyphi A), 31 (8.1 %) demonstrated ciprofloxacin (5 µg) resistance on screening using the Kirby–Bauer disc diffusion method. These were confirmed by biochemical reactions and serotyping with specific antisera using polyclonal, monovalent O, H and A antisera (Central Research Institute, Kasauli, India). They were also tested by the agar dilution method for their nalidixic acid MICs. The isolates were subjected to Etest strip (AB Biodisk) and agar dilution MIC testing to ciprofloxacin, cefotaxime and cefepime (Sigma) on cation-adjusted Mueller–Hinton agar (Difco). In addition, the MICs for ofloxacin, gatifloxacin, levofloxacin, cefotaxime, cefixime, cefepime and azithromycin were determined using the Etest strip method. The results were interpreted following Clinical and Laboratory Standards Institute guidelines (CLSI, 2006). The breakpoint MIC levels for azithromycin have not been determined for isolates of S. Typhi and S. Paratyphi A (CLSI, 2006). In a previous in vitro study, azithromycin had an MIC range of 4–16 µg ml^–1 against S. Typhi (Girgis et al., 1999). Of the 31 resistant isolates, 26 were S. Typhi and 5 were S. Paratyphi A. Tables 1 and 2 show the MICs of ciprofloxacin-resistant S. Typhi and S. Paratyphi A, respectively, to ciprofloxacin, cefotaxime and cefepime as determined by agar dilution. Tables 3 and 4 show the MICs of the newer antimicrobials tested against the 31 ciprofloxacin-resistant S. Typhi and S. Paratyphi A isolates, respectively, as determined using Etest strips. All of the isolates had ciprofloxacin MICs ≥32 µg ml^–1 by the Etest strip test and had nalidixic acid MICs ≥256 µg ml^–1 by the agar dilution method (results not shown). By agar dilution, five S. Typhi isolates and one S. Paratyphi A isolate showed MICs ≥512 µg ml^–1 (Tables 1 and 2). Gatifloxacin resistance was seen in 80.8 and 80 % of S. Typhi and S. Paratyphi A isolates, respectively. S. Typhi showed MIC₉₀ values of 0.50, 0.25 and 0.38 µg ml^–1 for cefixime, cefotaxime and cefepime, respectively, by Etest strip test. The MIC₅₀ values of these agents were 0.19, 0.125 and 0.25 µg ml^–1. For the cephalosporins tested, the difference in MIC₉₀ and MIC₅₀ for S. Typhi and S. Paratyphi A was minimal. A single isolate of S. Typhi showed a high azithromycin MIC (64 µg ml^–1), and the MIC₉₀ value for azithromycin was 24 µg ml^–1 for both S. Typhi and S. Paratyphi A.

Table 1. Agar dilution MICs of ciprofloxacin-resistant S. Typhi for ciprofloxacin, cefotaxime and cefepime The CLSI (2006) interpretive criteria for sensitive, intermediate and resistant strains, respectively, are: ciprofloxacin (CIP), ≤1, 2 and ≥4 µg ml–1; cefotaxime (CTX), ≤8, 16–32 and ≥64 µg ml–1; cefepime (CEP), ≤8, 16 and ≥32 µg ml–1.

Table 2. Agar dilution MICs of ciprofloxacin-resistant S. Paratyphi A for ciprofloxacin, cefotaxime and cefepime For CLSI interpretive criteria see Table 1.

Table 3. MICs of ciprofloxacin-resistant S. Typhi by the Etest strip method The CLSI (2006) interpretive criteria for sensitive, intermediate and resistant strains, respectively, are: ofloxacin (OFX) and levofloxacin (LVX), ≤1, 2 and ≥4 µg ml–1; gatifloxacin (GA), ≤2, 4 and ≥8 µg ml–1; cefixime (CFX), ≤1, 2 and ≥4 µg ml–1; cefotaxime (CTX), ≤8, 16–32 and ≥64 µg ml–1; cefepime (CPM), ≤8, 16 and ≥32 µg ml–1. Azithromycin (AZ) MIC breakpoints have not been defined.

Table 4. MICs of ciprofloxacin-resistant S. Paratyphi by the Etest strip method For CLSI interpretive criteria see Table 3.

There was a discrepancy in the MICs observed in the Etest strip and agar dilution tests, which has been reported previously (Capoor et al. 2006). The first- and second-generation quinolones had varying results. Gatifloxacin (80.8 % resistance) demonstrated better in vitro activity compared with other quinolones (96.2 % resistance for ofloxacin and 92.3 % for levofloxacin) in S. Typhi. This finding was consistent with the results of a study from Nepal (Pokharel et al., 2006). These observations indicate that fluoroquinolones should be tested individually and that ciprofloxacin does not represent this group adequately. The target-specific action of quinolones was originally studied in Streptococcus pneumoniae where quinolones have different binding-target affinities (Richardson et al., 2001). Disparity in the MIC levels of quinolones has been attributed to differences in the additional fluoro group and other substitutions in their chemical structure. There are no similar studies in Salmonella spp.

Amongst the cephalosporins tested against S. Typhi and S. Paratyphi A, cefotaxime had the lowest MIC₅₀ and MIC₉₀ levels. In a previous study, we showed that cefepime also had a high activity (Capoor et al., 2006). Cefixime is widely used in India due to its oral route of administration. This could be the reason for the rising MIC levels of third- and fourth-generation cephalosporins (Capoor et al., 2006; Saha et al., 1999), and their overuse can induce strains with extended-spectrum ß-lactamases (Pokharel et al., 2006).

Only one isolate of S. Typhi showed a high azithromycin MIC (64 µg ml^–1) and the MIC₉₀ was 24 µg azithromycin ml^–1. This was slightly higher than previous reports, which have found the MIC range to be 4–16 µg ml^–1 (Girgis et al., 1999; Butler et al., 1999). There are as yet no data on the break points of azithromycin for enteric fever (CLSI, 2006). Thus, molecular analysis of such strains with higher MICs is warranted. The MIC₉₀ value observed by Butler et al. (1999) was higher for S. Paratyphi A compared with S. Typhi. A single isolate of S. Typhi with an MIC of ≥32 µg ml^–1 was detected as early as 1999. In our study, the MIC₉₀ values for S. Typhi and S. Paratyphi A were 24 µg ml^–1. In enteric fever, the role of azithromycin needs to be appreciated, as it is highly effective in removing intracellular salmonellae, defervescence is rapid, gastrointestinal carriage is eradicated and it represents a potential alternative in paediatric populations where quinolones are contraindicated (Girgis et al., 1999). The higher clinical and bacteriological cure rate is attributable to the >100-fold intracellular concentrations of azithromycin in macrophages compared with serum (Butler et al., 1999; Frenck et al., 2004; Girgis et al., 1999; Gordillo et al., 1993; Parry, 2004; Parry et al., 2007). Thus, there is speculation that intracellular MICs may not be represented fully by the currently available in vitro MIC testing methods and therefore such testing should be coupled with therapeutic trials. As this was a retrospective study, the therapeutic efficacy of this drug was not determined. Due to its negligible relapse rate and faecal carriage, and its favourable outpatient compliance, azithromycin could become the preferred drug of choice over ceftriaxone, ofloxacin and chloramphenicol.

The indiscriminate use in patients of the existing therapeutic options for enteric fever, and a concomitant rise in MIC levels demonstrated for these antimicrobials for S. Typhi and S. Paratyphi A in the current study and in prior studies (Frenck et al., 2004; Gordillo et al., 1993; Kownhar et al., 2007; Parry, 2004) is a serious concern. The effects of some of the newer drugs, such as tigecycline and carbapenems, against salmonellae have yet to be elucidated by in vitro and in vivo trials. The presence of fluoroquinolone resistance warrants a review of the current therapy and initiation of the search for a new effective and affordable drug. Meanwhile, azithromycin and other available antimicrobials will require large-scale, randomized clinical trials to establish their population efficacy.