We analyzed the electronic health records (EHRs) of persons >18 years of age who received care at the University of California San Diego Health System or the San Diego Veterans Affairs Medical Center from March 1, 2017, through May 31, 2020, for diarrhea and a diagnosis of shigellosis. We did not include persons <18 years of age because those data were not accessible in our EHR.

Stool samples from study participants were cultured or tested initially with a multiplex PCR panel ordered by providers at the University of California San Diego Health System or the San Diego Veterans Affairs Medical Center. Isolates that were positive for Shigella spp. on multiplex PCR were confirmed by stool culture for Shigella spp. on matrix-assisted laser desorption/ionization-time of flight mass spectrometry. Those isolates were then sent to the San Diego County Public Health laboratory to confirm speciation.

For persons for whom Shigella spp. isolation from stool culture was confirmed, we collected further information through EHR review. Information included patient ethnicity, race, age at diagnosis, history of travel out of San Diego County within 14 days before diagnosis, housing status, recent methamphetamine use, HIV status, use of highly active antiretroviral therapy (HAART), or use of preexposure prophylaxis (PrEP) for HIV prevention.

The most frequently reported antimicrobial susceptibility panel for S. flexneri and S. sonnei includes ampicillin, a fluoroquinolone, and trimethoprim-sulfamethoxazole (TMP/SMX); thus, we included these antimicrobials in the comparative analysis. The specific fluoroquinolone reported at each site differed. Most were ciprofloxacin; however, other fluoroquinolones, such as moxifloxacin or levofloxacin, were reported for a few persons. For the purpose of comparative analysis, we grouped fluoroquinolones together and reported them as either sensitive or resistant; we identified intermediate strains as resistant. We used broth microdilution for susceptibility testing. The standards for resistance were set by the Clinical and Laboratory Standards Institute (CSLI, https://clsi.org). The standards changed in 2022 (21), when CSLI began including azithromycin susceptibility, but we did not use these latest guidelines for our study because our data included only time points until May 2020.

We analyzed the data by using SPSS Statistics 27 (IBM, https://www.ibm.com) and descriptive statistics via cross-tabulations with the Fisher exact test and χ² by creating a 2-by-2 table comparing different variables: resistance of various antimicrobials cross-tabulated with HIV status, sexual orientation, use of HAART if living with HIV, use of PrEP, being unhoused, or recent use of methamphetamines. We included methamphetamine use because this information was readily available on chart review (based either on positive urine toxicology screen or documented recent use in physician notes). Data associated with other substance use on chart review were not available. We considered p<0.05 to indicate statistical significance.

Of the 140 EHRs for patients for whom Shigella spp. was isolated from >1 stool specimen, we included 128 in our analysis and excluded 12 (for 6, the Shigella isolate was not speciated; for 2, S. boydii was isolated from stool culture; for 2, there was not enough information to complete the analysis; and for 2, duplicate positive isolates were obtained within a 2-week period). Overall, 55% of Shigella spp. isolates were identified as S. flexneri and 45% as S. sonnei. Patient ages ranged from 15–79 years; mean age was 47 years (Table 1). Eighty-one percent were male. Most patients self-identified as non-Hispanic white. Fifty-six percent identified as gay or bisexual male (GBM), and 3 identified as transgender women who had sex with men. Fifty-one percent were living with HIV, among whom 85% were receiving HAART; for 11%, HIV status was unknown. Twenty-three percent were unhoused, and for 33%, recent methamphetamine use was documented.

S. flexneri infection was more common than S. sonnei infection among those who were living with HIV (68% vs. 43%; p = 0.008), were not unhoused (92% vs. 60%; p<0.001), did not use methamphetamines (77% vs. 54%; p = 0.006), and had not recently traveled (63% vs. 24%; p = 0.008) (Table 2). Sexual orientation and use of HAART and PrEP were not associated with S. flexneri or S. sonnei infection.

Of the 128 evaluable isolates, 23% were resistant to fluoroquinolones (11% of S. flexneri and 39% of S. sonnei isolates). Overall, 70% of Shigella spp. isolates were resistant to ampicillin (87% of S. flexneri and 47% of S. sonnei isolates), and 83% of all isolates were resistant to TMP/SMX (73% of S. flexneri and 95% of S. sonnei isolates). Only 4 (3%) isolates were susceptible to fluoroquinolones, TMP/SMX, and ampicillin; 38% were resistant to 1 antimicrobial (27% for S. flexneri vs. 53% for S. sonnei), 38% were resistant to 2 antimicrobials (62% for S. flexneri vs. 9% for S. sonnei), and 20% were resistant to 3 antimicrobials (7% for S. flexneri vs. 37% for S. sonnei) (Appendix).

Compared with persons without HIV infection, a higher percentage of Shigella isolates from persons with HIV infection and shigellosis were resistant to >2 antimicrobials (32% vs. 11%; p<0.001), and the rate of ampicillin resistance was higher for persons with than without HIV infection (48% vs. 30%; p<0.001). We found no difference between HIV status and fluoroquinolone or TMP/SMX resistance. Rates of Shigella resistance to 2 antimicrobials were higher for non-GBM/transgender persons than for GBM/transgender persons (41% vs. 3%; p<0.003).

Rates of antimicrobial resistance to 1 (25% vs. 13%; p = 0.016) or 2 antimicrobials (34% vs. 4%; p = 0.009) were higher for persons with stable housing than for those without stable housing; rates of resistance were higher for ampicillin (60% vs. 9%; p<0.001) and TMP/SMX (60% vs. 15%; p = 0.01). Compared with recent methamphetamines use, no methamphetamine use was associated with resistance to 2 antimicrobials (30% vs. 8%; p = 0.021) and higher rates of ampicillin resistance (51% vs. 19%; p = 0.042). Neither history of travel nor use of HAART was associated with increased antimicrobial resistance.

In the antimicrobial susceptibility panels at our sites, ceftriaxone and azithromycin were not commonly tested. Although isolates were not routinely tested against azithromycin, of the 36 isolates tested, 55% were resistant to azithromycin according to MIC breakpoints established by the 2022 CLSI m100 Performance Standards for Antimicrobial Susceptibility Testing for Shigella spp. (MIC >16 μg/mL for S. flexneri and >32 μg/mL for S. sonnei) (21). Azithromycin started to become more frequently tested at 1 of the sites by August 2019. Of the 128 samples analyzed, ceftriaxone susceptibility was reported for 37% of the cases without any reported resistance.

According to our retrospective review, S. flexneri and S. sonnei caused diarrheal disease in a large number of gender and sexual minorities who were living with HIV, persons of unhoused status, and persons who had recently used methamphetamines. Although the association between MSM and shigellosis has been reported, the rate of concomitant HIV infection and shigellosis in our study was higher than rates previously reported, including in a study of New York, New York, residents, in which 27% with shigellosis infection were living with HIV (12). In addition, although methamphetamine use by MSM can increase the risk for HIV and for hepatitis C virus infection (22,23), to our knowledge only a small number of reports have described an association between methamphetamine use and shigellosis, specifically among MSM (24,25). Unlike the studies that documented methamphetamine use in MSM living with HIV, we also found high rates among those who were unhoused.

Isolation of S. flexneri from 55% of stool specimens in our study is much higher than the national prevalence of 12.6% in 2016 reported by the Centers for Disease Control and Prevention (CDC) (26). Other studies have found high rates of S. flexneri causing shigellosis in certain populations, including in 34% of sporadic cases in a New York, New York, study (12) and 65% of cases from a cohort of MSM in the United States (22). Similarly, we found high rates of S. flexneri in GBM/transgender persons, persons living with HIV, and those receiving PrEP. Thus, although overall S. flexneri prevalence may be low in the United States, rates may be higher for certain populations, such as those reported here.

Increasing antimicrobial resistance of Shigella spp. has been documented globally, including in the United States. Clusters of ciprofloxacin-resistant S. sonnei have been reported, suspected to be associated with international travel, and 86% of isolates have been reported to be resistant (27). Conversely, in another study of shigellosis in MSM, all isolates were sensitive to ciprofloxacin (28). In our cohort, 23% of isolates were resistant to ciprofloxacin. Among MSM, decreasing susceptibility to azithromycin (MIC ≥32 μg/mL) in several S. sonnei outbreaks has been documented (29), including all isolates either resistant to (11) or with detected mphA or ermB macrolide resistance genes (8,11,29). We also detected a large number of isolates for which susceptibility to azithromycin was reduced.

Rates of antimicrobial resistance among the isolates tested were comparable to those reported in the most recent data from the CDC National Antimicrobial Resistance Monitoring System (NARMS) data (https://wwwn.cdc.gov/narmsnow), which included mostly S. flexneri and S. sonnei isolates. Among all Shigella spp. isolates, 55% of the isolates in our study were resistant to azithromycin compared with 54% from preliminary NARMS estimates from 2021, 23% were resistant to fluoroquinolones compared with 32% from NARMS (ciprofloxacin only), 70% were resistant to ampicillin compared with 79% from NARMS, and 83% were resistant to TMP/SMX compared with 75% from NARMS. We found no resistance to ceftriaxone in our study, which could be reflective of a smaller sample size, although of 28 isolates in 2021, only 1 (3.6%) was resistant to ceftriaxone according to NARMS data. The most recent NARMS data report a rate of 25% of resistance to ampicillin, ciprofloxacin, and TMP/SMX, which is comparable to the resistance rate of 20% to these 3 antimicrobial drugs in our study.

It is unclear why those who are unhoused in San Diego County may be at risk for shigellosis. Of note, a hepatitis A outbreak during 2016–2018 in San Diego County, which was the largest US hepatitis A outbreak in 2 decades, primarily involved persons experiencing unstable housing conditions, those who used drugs in unsanitary settings, or both (30). It is possible that unsanitary living conditions and illicit drug use also predispose unhoused persons to shigellosis, although further research would help clarify.

Our first study limitation is including data from only 2 medical centers in San Diego County, California, which means that our findings may not be representative of other locations within the United States or elsewhere globally. The high proportion of GBM/transgender, HIV-infected, unhoused, and methamphetamine-using persons may be unique to our setting and may not be reflective of other settings because it may not be reflective of the general population in San Diego County. Second, isolates were not routinely tested against azithromycin, which would have been helpful because this drug has traditionally been empirically used to treat shigellosis. Regardless, because most of the isolates tested were resistant to azithromycin according to the MICs, resistance to azithromycin in our setting is probably high, indicating that empiric use of azithromycin may not be beneficial. Third, given that the study was a retrospective review of EHRs, inconsistent availability of documentation of HIV status, travel, unhoused status, or methamphetamine use could have led to missing data.

In conclusion, in this retrospective study, we found shigellosis frequently diagnosed for GBM/transgender, HIV-seropositive, and unhoused persons who use methamphetamine; rates of S. flexneri prevalence were higher than those in national prevalence data. Overall, we found high rates of resistance to azithromycin, fluoroquinolones, ampicillin, and TMP/SMX, which were comparable to those most recently reported by CDC, except we found slightly lower resistance rates to ciprofloxacin and slightly higher resistance rates to TMP/SMX. Clinicians should remain aware of the risk for antimicrobial resistance among patients with shigellosis, particularly gender and sexual minorities, those living with HIV, persons of unhoused status, and persons who use methamphetamines.