Is It Time To Change Fluoroquinolone Breakpoints for Salmonella spp.?

Fluoroquinolones are the drug of choice for treatment of invasive gastrointestinal infections in many parts of the world. Salmonella enterica is one of the most common causes of human gastroenteritis worldwide. Nontyphoidal Salmonella infections are primarily caused by consumption of contaminated foods primarily of animal origin, whereas S. enterica serovar Typhi and Paratyphi infections are spread directly or indirectly from human feces. Since the early 1990s, fluoroquinolones have in many countries been licensed for use in food production animals. The concomitant emergence and increased incidence of quinolone-resistant Salmonella strains have been an issue of intense worldwide debate.

In members of the family Enterobacteriaceae, resistance to quinolones is most commonly acquired by mutations in two steps. One mutation in the gyrA gene mediates full resistance to narrow-spectrum quinolones, such as nalidixic acid and decreased susceptibility to fluoroquinolones (3, 18). A second mutation in either gyrA or gyrB genes mediates full resistance to fluoroquinolones. The current National Committee for Clinical Laboratory standards (NCCLS) breakpoint for resistance to the fluoroquinolone ciprofloxacin is ≥4 μg/ml, while the breakpoint for resistance to its veterinary equivalent, enrofloxacin, is ≥2 μg/ml (14, 15). Even though the NCCLS does not provide specific breakpoints for bacteria associated with gastrointestinal infections, the current breakpoints are widely used by clinicians, veterinarians, microbiologists, and others involved with the issues involving selection for fluoroquinolone resistance.

Several studies have shown that resistance to nalidixic acid and decreased susceptibility to fluoroquinolones have increased among Salmonella spp. from food animals and infections in humans (8, 10, 13, 19-21). Some have used this as an argument against the continued usage of fluoroquinolones for food animals. Defenders of the continued usage of fluoroquinolones for food animals argue that on the basis of the NCCLS breakpoint, there is still very limited “clinical” fluoroquinolone resistance among Salmonella strains, despite several years of fluoroquinolone use in food animal production.

With this letter, we do not want to get into the debate regarding the continued usage of fluoroquinolones among food animals. However, we wish to stress the importance of establishing a common breakpoint as the basis for evaluating susceptibility data and interpreting these data in a clinical context. Isolates with a single mutation in gyrA are to some extent refractory to the bactericidal effect of fluoroquinolones (2; J. D. Knudsen, R. Skov, P. Gerner-Smidt, L. V. Pallesen, and N. Frimodt-Møller, Abstr. 41st Intersci. Conf. Antimicrob. Agents Chemother., abstr. A-2091, 2001). In our opinion, there is now compelling evidence that the treatment efficacy of fluoroquinolones is reduced in humans infected with strains of Salmonella enterica that demonstrate what is regarded as decreased susceptibility to fluoroquinolones. Thus, the use of the NCCLS breakpoint at ≥4 μg/ml for ciprofloxacin may have the effect of obscuring the true occurrence of resistance to quinolones among Salmonella strains. Indeed, in several cases, decreased susceptibility to fluoroquinolones has only been detected on the basis of resistance to nalidixic acid after the patient has failed to respond to treatment with ciprofloxacin. A number of reported cases of infections with S. enterica serovar Typhi and nontyphoid Salmonella in humans are shown in Table 1. One observation common to all reports is that fluoroquinolones were ineffective in treating patients who were infected with isolates for which the MICs of fluoroquinolones were between 0.06 and 2 mg/liter. In five case reports on nontyphoid Salmonella infections, resistance to nalidixic acid and decreased susceptibility to ciprofloxacin (MIC, 0.125 to 2 mg/liter) developed during therapy. Helms et al. (7) reported that patients infected with isolates for which MICs were between 0.06 and 0.38 mg/liter were five times more likely to die within 2 years after infection than patients infected with isolates for which MICs were lower. In the latter report, all isolates were isolated before initial therapy, and it is highly probable that the patients were originally infected with isolates with decreased susceptibility to fluoroquinolones.

TABLE 1.

Different studies reporting treatment failures in patients infected with Salmonella enterica serovar Typhi and non-Typhi Salmonella isolates with decreased susceptibility to fluoroquinolones

Serovar	No. of patients	MIC for isolate (μg/ml)	Case	Reference
Denmark
Enteritidis	1	Original strain, 0.032; after initial treatment, 1	82-yr-old male infected with S. enterica serovar Enteritidis; initial isolate fully susceptible to nalidixic acid and ciprofloxacin, and patient treated with ciprofloxacin; after 1 mo, infection reemerged, and strain showed full resistance to nalidixic acid and decreased susceptibility to ciprofloxacin; no effect of continued ciprofloxacin treatment	16
Typhimurium DT104	27	0.064-0.124	Outbreak including 25 culture-confirmed cases; 11 patients hospitalized, and 2 died; molecular epidemiology and data from patients confirm primary source Danish swineherd	12
Typhimurium	83	0.06-0.38	Cohort study showing 5 deaths out of 83 patients infected with Salmonella Typhimurium isolates with decreased susceptibility to ciprofloxacin, which is a 5-times-higher mortality than that of patients infected with fully susceptible strains	7
Typhi	1	0.19	8-yr-old girl of Pakistani origin with persistent bacteremia caused by a multidrug-resistant S. enterica serovar Typhi strain; patient treated for 10 days with ciprofloxacin (from day 3 also with ceftriaxone); infection relapsed 19 days after discharge	4
France
Typhi	1	0.12	45-yr-old man infected with S. enterica serovar Typhi; no response to ciprofloxacin treatment	11
India
Typhi	32	0.0625-0.5	32 bacteriologically positive children out of 140 not responding to ciprofloxacin treatment.	5
Spain
Enteritidis	2	Original strains, 0.06; after treatment, 0.5 and 1, respectively	(i) 76-yr-old diabetic male; original isolate fully susceptible to ciprofloxacin; treated with different antimicrobials, including ciprofloxacin, for 12 days; readmitted on day 30, now infected with an isolate for which MIC was 1 mg/liter	23
			(ii) 34-yr-old male with AIDS with S. enterica serovar Enteritidis bacteremia; original isolate fully susceptible to ciprofloxacin; received 10-day course of amoxicillin-clavulanate; 6 days later readmitted and received ciprofloxacin for 12 days; day 50, infection relapsed, and S. enterica serovar Enteritidis isolated again; MIC for evolved isolate, 0.5 mg/liter	23
United Kingdom
Typhi	1	0.5	65-yr-old woman infected with S. enterica serovar Typhi; no response to ciprofloxacin treatment; other treatment failures also reported	22
Typhimurium	1	Original strain, 0.03; after treatment, 2.0	61-yr-old male with chronic renal failure; S. enterica serovar Typhimurium isolated from feces; after 6 days of treatment with ciprofloxacin; resistant S. enterica serovar Typhimurium strain isolated from feces	9
Typhimurium	2	Original strains, 0.015 and 0.03; after treatment, 2 and 0.06-1	(i) 43-yr-old male with chronic upper urinary tract infection; initially treated with amoxicillin-clavulanic acid followed by a course of ciprofloxacin for 14 days; 1 mo later infected with isolate with decreased susceptibility to ciprofloxacin (2 mg/liter)	17
			(ii) 52-yr-old male undergoing aortic aneurysm surgery; susceptible isolates obtained from wound drainage and blood; after 10-day course of ciprofloxacin, lumbar osteomyelitis developed; oral ciprofloxacin treatment continued, but eventually changed to aztreonam; number of isolates with decreased susceptibility to ciprofloxacin recovered, for which MICs were between 0.06 and 1 mg/liter.	17
Bovismorbifica ns	1	Original isolate, 0.06; after treatment, 2, 16, and 4	40-yr-old man with acute pancreatitis; pure culture of S. enterica serovar Bovismorbificans isolated from a peripancreatic abscess on day 19; 5 days after surgery changed to intravenous ciprofloxacin, which was stopped after 17 days; S. enterica serovar Bovismorbificans isolated on days 15, 21, and 26 after initiation of ciprofloxacin treatment	6
Virchow	1	Original strain, 0.016; after initial, treatment, 0.75	24-yr-old male infected with S. enterica serovar Virchow; initial isolate fully susceptible to nalidixic acid and ciprofloxacin, and patient treated with ciprofloxacin; after 6 wk, infection reemerged, and strain showed full resistance to nalidixic acid and decreased susceptibility to ciprofloxacin; no effect of continued ciprofloxacin treatment	1
Vietnam
Typhi	150	0.125-1 (ofloxacin)	Comparison of clinical efficacy of ofloxacin treatment of 18 patients infected with nalidixic acid-resistant and 132 patients infected with susceptible S. enterica serovar Typhi isolates; clinical failure in 50% of patients infected with resistant isolates compared to 3% of patients infected with susceptible isolates, when treated with ofloxacin	24

In conclusion, the available literature has demonstrated that when patients are infected with Salmonella isolates with decreased susceptibility to fluoroquinolones, treatment failures may occur. Thus, to reduce the risk for humans, we recommend that for Salmonella, a breakpoint of ≥0.125 μg/ml for fluoroquinolones should be used when evaluating both laboratory sensitivity tests and for the surveillance of resistance to this important group of antimicrobials in both human and veterinary medicine. The most optimal solution may be to use testing for susceptibility to nalidixic acid as a screening tool and then determine the MIC of ciprofloxacin for all nalidixic acid-resistant isolates.

An artificially high breakpoint not only may lead to the wrong clinical decisions, but can also obscure surveillance data. And if the surveillance data are flawed, it is impossible to develop and improve policies for prudent drug use in both food production and human medicine.