Evidence of Failure of Oral Third-Generation Cephalosporin Treatment for Shigella sonnei Infection

Abstract

In 2017, state health departments notified the Centers for Disease Control and Prevention about 4 patients with shigellosis who experienced persistent illness after treatment with oral third-generation cephalosporins. Given increasing antibiotic resistance among Shigella, these cases highlight the need to evaluate the efficacy of oral cephalosporins for shigellosis.

Shigella sonnei causes shigellosis, a self-limited infection characterized by diarrhea, nausea, and abdominal cramps, with or without fever. Typical risk groups include young children, travelers to developing countries, men who have sex with men, and immunocompromised persons. Antibiotic therapy is indicated for patients with severe illness or immunocompromising conditions and to limit the risk of transmission. Treatment decisions have been complicated by use of culture-independent diagnostic tests, which allow for a diagnosis but do not provide an isolate on which antibiotic susceptibility testing (AST) can be performed. Current clinical guidelines recommend azithromycin, ciprofloxacin, or ceftriaxone for empiric treatment of suspected shigellosis. Because of widespread resistance to ampicillin and trimethoprim-sulfamethoxazole, treatment with these drugs is not advised unless isolates are susceptible [1]. National surveillance data show a recent rise in Shigella with decreased susceptibility to ciprofloxacin and azithromycin [2]. However, clinical breakpoints have not been established for azithromycin, and most clinical laboratories only routinely report AST results for ampicillin, trimethoprim-sulfamethoxazole, and fluoroquinolones [3].

Parenteral ceftriaxone is recommended when patients with shigellosis cannot receive oral antibiotics [1]. Although resistance to ceftriaxone is relatively uncommon in the United States, the route of administration can be inconvenient. Oral third-generation cephalosporins may be a tempting alternative; however, the following cases highlight treatment failures due to these agents.

CASE SERIES

Case 1

A 51-year-old man with poorly controlled HIV infection was hospitalized with 25-pound weight loss, abdominal pain, and bloody diarrhea in July 2017. He reported sexual contact with men. His CD4 count was 41 cells/mm³, and an abdominal computed tomography (CT) scan demonstrated proctocolitis. A stool culture yielded S. sonnei; the isolate was susceptible to ceftriaxone and resistant to ampicillin, trimethoprim/sulfamethoxazole, and ciprofloxacin (Figure 1, Isolate 1-A). He received 4 days of intravenous (IV) ceftriaxone with resolution of symptoms and was transitioned to oral cefpodoxime (400 mg twice daily). One week later, he reported recurrence of diarrhea. A stool culture yielded S. sonnei (Figure 1, Isolate 1-B). Diarrhea resolved after 4 weeks of IV ceftriaxone.

Case 2

A 32-year-old man who was HIV-negative developed fever and nonbloody diarrhea in July 2017. He reported sexual contact with men. Oral ciprofloxacin was empirically prescribed. When a stool culture yielded S. sonnei resistant to ampicillin, trimethoprim/sulfamethoxazole, and ciprofloxacin (Figure 1, Isolate 2-A), therapy was changed to oral azithromycin. Despite a 3-day course, diarrhea persisted, and a stool culture again yielded S. sonnei (Figure 1, Isolate 2-B). After a 5-day course of oral cefixime (200 mg twice daily), he continued to have diarrhea. A stool culture obtained 6 days after completion of treatment yielded S. sonnei. He then received a 5-day course of IV ceftriaxone. Diarrhea continued, but a stool culture obtained the day after completion of treatment did not grow Shigella. Another stool specimen collected 4 days later yielded S. sonnei (Figure 1, Isolate 2-C) and was positive for S. sonnei by multiplex polymerase chain reaction (PCR). He reported resolution of symptoms after several weeks.

Case 3

A 41-year-old man with HIV infection developed vomiting, diarrhea, and abdominal cramps in July 2017. He reported sexual contact with men. One week after symptoms began, he submitted a stool specimen for culture, and he took ciprofloxacin for 7 days. The stool culture yielded S. sonnei resistant to ampicillin, trimethoprim/sulfamethoxazole, and levofloxacin (Figure 1, Isolate 3-A). His symptoms initially improved. However, 3 weeks after completion of treatment, he continued to experience diarrhea. A stool specimen at that time did not yield Shigella, but a multiplex stool PCR was positive for S. sonnei. He then completed a 10-day course of oral cefdinir (300 mg twice daily). Diarrhea persisted, and a stool culture yielded S. sonnei 18 days after completion of treatment (Figure 1, Isolate 3-B). Five weeks later, because of continued diarrhea, the patient received a 7-day course of outpatient ceftriaxone; a stool specimen collected 5 weeks after ceftriaxone treatment ended was negative for Shigella by PCR.

Case 4

While visiting Ireland in July 2017, a 22-year-old previously healthy woman developed fever, vomiting, and nonbloody diarrhea. After returning to the United States, she submitted a stool specimen and began taking oral ciprofloxacin. Symptoms initially improved. The stool culture grew S. sonnei susceptible to ampicillin and resistant to levofloxacin and trimethoprim-sulfamethoxazole (Figure 1, Isolate 4-A). Five days after completing treatment, fever and diarrhea recurred, and she was given 5 days of oral cefixime (200 mg twice daily). Her symptoms improved. A stool culture obtained 3 days after completion of cefixime did not yield Shigella, but a specimen obtained 2 days later yielded S. sonnei (Figure 1, Isolate 4-B). After treatment with a second 5-day course of cefixime, symptoms resolved, and no further stool testing was done.

ANTIBIOTIC RESISTANCE AND ISOLATE RELATEDNESS

The Centers for Disease Control and Prevention’s (CDC’s) National Antimicrobial Resistance Monitoring System (NARMS) laboratory performed antimicrobial susceptibility testing (AST) using a standard broth microdilution protocol [4]. To determine isolate relatedness, whole-genome sequencing (WGS) with high-quality single nucleotide polymorphism (hqSNP) analysis was completed according to the PulseNet protocol (https://www.cdc.gov/pulsenet/pathogens/wgs.html). Isolates were screened for resistance genes and mutations using ResFinder 3.0. The CDC’s Human Research Protection Office determined that this work was exempt from human subjects research regulations.

The CDC laboratory confirmed that each isolate was resistant to trimethoprim-sulfamethoxazole and ciprofloxacin based on Clinical and Laboratory Standards Institute breakpoints (Figure 1). Isolates from cases 1–3 also demonstrated resistance to ampicillin and decreased susceptibility to azithromycin [3]. All isolates were susceptible to amoxicillin-clavulanate, cefoxitin, and ceftriaxone. No cephalosporin resistance mechanisms were detected by WGS analysis. For each case, isolates collected before and after cephalosporin therapy were closely related (≤2 SNPs different) (Figure 1). Isolates from cases 1–3 were genetically linked to each other (≤12 SNPs different) and to other cases in a multistate shigellosis outbreak. We classified antibiotic treatment failure as clinical failure, which we defined as patient-reported continued diarrhea following treatment, or microbiological failure, which we defined as a stool culture that yielded S. sonnei after treatment.

Figure 1.

High-quality single nucleotide polymorphism (hqSNP) analysis of Shigella sonnei isolates. This hqSNP analysis contains 9 isolates from 4 patients. SNP differences for each patient’s isolates before cephalosporin treatment (indicated by *) and after are shown. The isolates in Clade A cluster in 0–12 SNPs. The isolates in Clade B are related by 0–2 SNPs but are unrelated to the other isolates (67–81 SNPs). Clinical laboratory antimicrobial susceptibility testing (AST) was performed at the time of patient illness. Surveillance AST was performed later by the Centers for Disease Control and Prevention; results were not available during patients’ illnesses. Interpretive categories for AST results are based on Clinical and Laboratory Standards Institute breakpoints, according to the following minimum inhibitory concentration (MIC) values (S, susceptible; R, resistant): ampicillin (S: ≤8 µg/mL; R: ≥32 µg/mL); amoxicillin-clavulanate (amox-clav; S: ≤8/4 µg/mL; R: ≥32/16 µg/mL); trimethoprim-sulfamethoxazole (TMP-SMX; S: ≤2/38 µg/mL; R: ≥4/76 µg/mL); ciprofloxacin (S: ≤0.25 µg/mL; R: ≥1 µg/mL); levofloxacin (S: ≤0.5 µg/mL; R: ≥2 µg/mL); ceftriaxone: (S: ≤1 µg/mL; R: ≥4 µg/mL); cefoxitin (S: ≤8 µg/mL; R: ≥32 µg/mL). AST results for azithromycin were interpreted using epidemiological cutoff values; isolates were reported as susceptible (MIC ≤ 16 ug/mL) or with decreased susceptibility (MIC ≥ 32 µg/mL). Abbreviations: CDC, Centers for Disease Control and Prevention; EDLB, Enteric Diseases Laboratory Branch; NCBI, National Center for Biotechnology Information.

DISCUSSION

The cases described in this report demonstrate antibiotic treatment failure after oral cephalosporin treatment for shigellosis despite in vitro susceptibility. After initial oral third-generation cephalosporin therapy, patients 1–3 experienced both clinical and microbiological failure, and patient 4 experienced microbiological failure. The matching antibiotic resistance patterns and close genetic relatedness of each patient’s isolates before and after oral cephalosporin therapy indicate persistent infection despite antibiotic therapy rather than reinfection with a new strain. The limited sensitivity of stool culture for Shigella may account for patients 2–4 having initial negative cultures following antibiotic therapy. Reinfection with the same strain, such as from the same sexual partner, is another possibility.

The Infectious Diseases Society of America does not recommend oral cephalosporins as first-line or second-line antibiotics for shigellosis [1]. Cefixime, an oral third-generation cephalosporin, is sometimes recommended as an alternative therapy [5, 6]. However, studies evaluating the effectiveness of oral third-generation cephalosporins for shigellosis have yielded variable results. Randomized controlled trials in the 1990s found that children with shigellosis who received cefixime had better outcomes than control groups who received trimethoprim-sulfamethoxazole or ampicillin-sulbactam [7, 8]. The high proportion of control group isolates that were resistant to trimethoprim-sulfamethoxazole (82%) and ampicillin-sulbactam (41%) was the likely reason. A randomized controlled trial of cefixime vs oral pivamdinocillin therapy for shigellosis conducted in 1993 demonstrated clinical failure in 47% and microbiological failure in 60% of the 15 adults receiving cefixime, much higher than for pivamdinocillin [9]. A US study conducted in 2000 found clinical failure in 20% and microbiological failure in 11% of children who received a 5-day course of cefixime for shigellosis; a higher percentage of children (55%) had microbiological failure with a course of only 2 days [10]. In 2003, a trial that randomized 75 children with shigellosis to azithromycin or cefixime found higher bacteriologic eradication with azithromycin, but similar clinical effectiveness for both drugs [11]. This study may have been underpowered to detect meaningful clinical differences between groups given its small sample size. The above studies have limited generalizability; most were conducted among children in low-resource settings, and only 1 compared cefixime with any current first-line antibiotic (azithromycin) available in the United States. The studies are at least 16 years old; changes in antibiotic resistance patterns may limit their usefulness.

Shigella can become resistant to third-generation cephalosporins through the acquisition of extended-spectrum beta-lactamases (ESBLs). However, in our series, based on both phenotypic AST and WGS analysis, antibiotic resistance was not a contributing factor to treatment failure. Short duration of therapy, inadequate dosing, or low oral bioavailability may have contributed to these oral cephalosporin treatment failures [12].

Two patients who received IV ceftriaxone had protracted infections. Advanced HIV infection may have been a contributing factor in Patient 1. Patient 2 had severe infection, for which 5 days of IV therapy may have been insufficient. The optimal duration of ceftriaxone therapy for shigellosis has not been adequately studied in adults or people with immunocompromising conditions.

Contemporary data on the effectiveness of oral cephalosporins for shigellosis are lacking. Prospective studies comparing oral third-generation cephalosporins with recommended agents, particularly among adult and immunocompromised populations, are needed. AST remains essential in guiding antibiotic choice. Given case reports such as these and the limitations of available clinical trial data, providers should be aware of the potential for clinical and microbiological treatment failure with oral third-generation cephalosporin treatment of shigellosis.