Treatment of Shigellosis: IV. Cefixime Is Ineffective in Shigellosis in Adults

Methods

The study was similar in design to previous antimicrobial trials in shigellosis that we conducted at the International Centre for Diarrhoeal Disease Research in Bangladesh [4-6].

Patient Selection

Patients eligible for the study were men 18 to 60 years of age who had dysentery lasting 72 hours or less and who sought care at the International Centre for Diarrhoeal Disease Research in Dhaka, Bangladesh. Patients were enrolled from July to November 1993. We excluded patients who had received antimicrobial therapy possibly effective for shigellosis, who had hematophagous trophozoites of Entamoeba histolytica in their stool, or who had other serious illnesses. All patients gave informed consent, and the Ethical and Research Review Committees of the International Centre for Diarrhoeal Disease Research approved the study protocol.

Study Protocol

When patients were admitted, histories were taken, physical examinations were done, and blood and stool samples were obtained. Patients received either 400 mg of cefixime once daily for 5 days or 400 mg of pivamdinocillin every 6 hours for 5 days. Treatment was randomized and double-blind. To blind the study, we gave each patient an active drug and a placebo formulation of the other study drug.

Patients remained in the hospital for 6 days. Symptoms and physical findings were recorded daily, and body temperatures were measured every 6 hours. Individual stools were examined for consistency and for the presence of blood and mucus. A rectal swab or stool sample for culture was obtained daily, and a microscopic examination of stool was repeated on study days 3 and 5. Blood samples to measure serum concentrations of cefixime were obtained 90 minutes, 4 hours, and 24 hours after patients received the first cefixime dose. Cefixime concentrations in stool were measured on samples collected on study days 2 and 4. Serum and stool concentrations were measured using a microbiological assay. Pivamdinocillin concentrations could not be measured because of the instability of this drug in serum and stool.

Seven days after discharge, patients returned for follow-up examinations that included a history, a physical examination, and a rectal swab or stool sample collection for culture.

Evaluation of Outcome

Treatment was considered to have failed clinically if patients continued to have frank dysentery on study day 3, had six or more stools on study day 5, had any grossly bloody-mucoid or watery stools, or had a fever (oral temperature greater than 37.8 °C). Bacteriologic failure of therapy occurred if Shigella organisms could be isolated from a stool or rectal swab sample after study day 2. Patients in whom initial therapy failed clinically were treated with 500 mg of ciprofloxacin every 12 hours for 5 days, with no identification of the initial treatment.

Sample Size Determination

To be clinically useful, the rate of clinical cure with cefixime had to be at least 65%. We expected the rate of clinical cure in patients treated with pivamdinocillin to be 95% [7]. With a type I error of 5% and a type II error of 20%, we needed a sample size of 33 in each treatment group to reject the null hypothesis. Because the clinical efficacy of cefixime was uncertain when we began the study, we planned an interim analysis after we enrolled 30 evaluable patients.

Statistical Analysis

We did statistical analyses using the Statistical Package for the Social Sciences, version 6.0 for Windows (SPS Inc., Chicago, Illinois), and Epi Info, version 5.0 (USD Inc., Stone Mountain, Georgia). We calculated 95% CIs using Confidence Interval Analysis, version 1.0 (British Medical Journal, London, United Kingdom). We used the Student t-test, the Mann-Whitney U test, the 2 × 2 chi-square test without continuity correction, and the Fisher exact test when appropriate to determine the significance of differences between the groups. Because we planned two comparisons between the groups (an interval analysis and a final analysis), we considered P < 0.029 to be significant [8]. All tests were two-tailed. We did the initial analysis blinded.

Results

We stopped enrolling patients after the interim analysis of the first 30 evaluable patients showed a significant difference in rates of clinical cure between the two treatment groups. Thirty-four patients, 17 in each treatment group, had been enrolled by that time. We excluded 4 patients (11.8%) from analysis because Shigella organisms were not isolated from stool cultures taken on admission. Of the 30 remaining patients, 15 had received cefixime and 15 had received pivamdinocillin.

Admission Characteristics

The admission characteristics of the 30 evaluable patients were similar and are shown in Table 1.

Clinical and Bacteriologic Outcomes

Table 2 lists clinical and bacteriologic outcomes for the two study groups. Therapy was clinically successful in all 15 patients treated with pivamdinocillin but in only 8 of the 15 patients (53%) treated with cefixime (P = 0.006). Cefixime therapy failed in 7 patients—after day 3 in 3 patients, after day 4 in 1 patient, and after day 5 in 3 patients. Rates of bacteriologic success were 87% (n = 13) for pivamdinocillin-treated patients and 40% (n = 6) for cefixime-treated patients (P = 0.009). The median number of stools during the study, the median duration of bloody-mucoid stools, and the median duration of fever were all significantly (P < 0.029) greater in the cefixime group. Patients in that group also had significantly more stool leukocytes on study day 3 (P = 0.014) and more stool erythrocytes on study day 5 (P = 0.042). The median duration of excretion of shigellae was 1 day in the pivamdinocillin group and 3 days in the cefixime group (P = 0.064). However, 6 patients in the cefixime group received alternate therapy with ciprofloxacin; Shigella organisms were isolated from stool from 4 of them until ciprofloxacin was given, but not thereafter.

Outcome by Infecting Serotype or Species of Shigella

Our subgroup analysis is limited in power because the number of patients was small, but the results indicate that cefixime was less effective than pivamdinocillin in patients infected with Shigella dysenteriae type 1 and those infected with other Shigella serotypes. Table 2 shows outcome by infecting serotype or species of Shigella.

Antimicrobial Susceptibility and Concentrations of Cefixime in Serum and Stool

All 30 isolates of Shigella were susceptible to the two study drugs and to ciprofloxacin when evaluated by the disc-diffusion method. Fifty-seven percent of the 30 isolates were resistant to ampicillin, 70% to trimethoprim-sulfamethoxazole, and 50% to nalidixic acid. The median minimum inhibitory concentrations to cefixime and pivamdinocillin were 0.125 µg/mL (range, 0.063 to 2.0 µg/mL) and 0.031 µg/mL (range, 0.008 to 2.0 µg/mL), respectively. Median serum concentrations of cefixime at 90 minutes and 4 hours were 0.98 µg/mL and 1.98 µg/mL, respectively, which were substantially greater than the minimum inhibitory concentrations of the infecting strains of Shigella. Median concentration at 24 hours was 0.10 µg/mL, which was equivalent to the minimum inhibitory concentration. In most patients, cefixime could not be detected in the stool on either study day 2 or 4.

Follow-up

Ten (67%) of the patients in the pivamdinocillin group and 9 (60%) of the patients in the cefixime group returned for follow-up visits. None of the patients treated with pivamdinocillin had continued symptoms of dysentery compared with 2 (22%) of the patients treated with cefixime (P = 0.211), and shigellae were isolated from a culture of stool from one of the cefixime-treated patients.

Discussion

In our study, cefixime was markedly less efficacious than pivamdinocillin for the treatment of shigellosis in adults. The therapeutic failure rate was 47% (7 of 15 patients) in the cefixime group and was similar to the 56% (19 of 34 patients) failure rate found in a previous study when patients received de facto placebo therapy (ampicillin treatment of ampicillin-resistant strains of Shigella) [5].

Our results differ from those of the only other study of cefixime in the treatment of shigellosis [9]. That study, done mostly in outpatients in Israel, found rates of clinical cure or improvement of 97% (37 of 38) in children treated with cefixime and 74% (29 of 39) in children treated with trimethoprim-sulfamethoxazole. Eighty-one percent of patients (62 of 77) in the Israeli study, however, were infected with S. sonnei, which usually produces a self-limited illness much less severe than that caused by the S. flexneri and S. dysenteriae type 1 infections encountered in our study. In addition, 82% (32 of 39) of the patients treated with trimethoprim-sulfamethoxazole were infected with Shigella strains resistant to this drug, which limited the value of the comparison group. By eradicating susceptible enteric flora but leaving the multiply resistant Shigella organisms unaffected, trimethoprim-sulfamethoxazole may have actually aggravated the illness.

Differences in dosing regimen (twice daily in the Israeli study; once daily in our study) are unlikely to explain the disparity in findings. The total daily dosage in the two studies was equivalent, and the long half-life of cefixime has allowed it to be used successfully in a once-daily dosing format for various infections [10]. Although the Israeli study enrolled children and we enrolled adults, the pharmacokinetics of cefixime are similar in adults and children [11], and drugs previously identified as useful in the treatment of shigellosis have been equally effective in adults and children.

Two older, orally administered cephalosporins, cephalexin and cephaloglycin, and the parenterally administered cephalosporin cefamandole are also ineffective in treating shigellosis [12-14]. Ceftibuten is the only extended-spectrum oral cephalosporin other than cefixime that has been evaluated to treat shigellosis [15]. Although the authors of the one study evaluating ceftibuten concluded that it was equivalent in efficacy to trimethoprim-sulfamethoxazole, only 12 patients with Shigella infection were treated with ceftibuten and only 8 with trimethoprim-sulfamethoxazole, severely limiting the power of the study to detect differences between groups. In addition, 3 of the 8 patients treated with trimethoprim-sulfamethoxazole were infected with strains resistant to this drug [15].

Why oral cephalosporins are less effective than other β-lactam agents in treating shigellosis, despite the in vitro susceptibility of the infecting strains of Shigella and serum concentrations in excess of the minimum inhibitory concentrations for those strains, is uncertain. Cefixime could not be detected in the stools of most patients in this study; this is consistent with the finding that nonabsorbed drug is degraded in the intestine and that active drug is thus usually not present in the stool [16]. The apparent efficacy of the parenterally administered cephalosporin ceftriaxone in shigellosis [17] might be explained by its excretion in an active form in the bile [18]. To treat shigellosis, however, intraluminal concentrations of antibiotic are probably a less critical determinant of success than are intracellular concentrations [19], because Shigella multiply intracellularly, and cefixime, like most cephalosporins, does not readily penetrate intracellularly [20].

The results of this study show the importance of selecting an appropriate comparison treatment when evaluating the efficacy of new antimicrobial agents, the effect that the severity of illness can have on the apparent utility of a drug, and the importance of repeatedly evaluating, under different circumstances, the efficacy of experimental therapies. Based on these findings, we conclude that cefixime, when used in the standard recommended dosage, is not effective to treat adults who are moderately to severely ill with shigellosis.

Dr. Bennish: New England Medical Center, 750 Washington Street, Boston, MA 02111.

@copy; 1995 American College of Physicians