After submission to NYC public health laboratories, representative colonies of Shigella isolates are identified with conventional biochemical tests and tested for susceptibility to ampicillin, cefixime, ciprofloxacin, azithromycin, and TMP/SMX using the Etest antibiotic gradient (bioMérieux, Durham, NC, USA). MICs are interpreted according to CLSI guidelines (5). After consultation with the Centers for Disease Control and Prevention (CDC), we defined DSA isolates as those with an MIC of azithromycin of >32 μg/mL (J. Whichard, CDC, pers. comm., 2013).

Using a standard questionnaire, we interviewed persons infected with DSA or ciprofloxacin-resistant Shigella isolates, diagnosed during March 22, 2013–May 31, 2015; we abstracted antibiotic use data from medical charts. Shigella-infected case-patients were matched to the NYC HIV Surveillance Registry (6). We determined neighborhood poverty level as described (7) and compared proportions of those infected by age group, sex, and HIV status using χ² tests. To identify factors associated with DSA or ciprofloxacin-resistant Shigella infection and with hospitalization, we used logistic regression analysis (SAS version 9.2; SAS Institute, Cary, NC, USA).

During 26 months, 978 Shigella isolates were submitted; 295 were associated with an outbreak (8) and analyzed separately, and 683 were defined as sporadic. Among patients with sporadic infections, 129 (19%) were infected with isolates displaying DSA, and 29 (4%) were infected with ciprofloxacin-resistant isolates; 5 isolates displayed both characteristics. The median age of case-patients was 27 years (range 0–93 years); 446 (65%) were male. Nearly all infections were caused by S. sonnei (65%) or S. flexneri (34%). Antibiotic resistance of isolates was as follows: 416 (61%) to ampicillin, 10 (1%) to cefixime, 29 (4%) to ciprofloxacin, and 481 (70%) to TMP/SMX (Table 1).

Persons infected with DSA or ciprofloxacin-resistant Shigella spp. were older and more likely to be male than those with DSA- or ciprofloxacin-susceptible isolates; no association with neighborhood poverty was found. Although most infections were caused by S. sonnei, most isolates displaying DSA were S. flexneri. Isolates displaying DSA or ciprofloxacin resistance were more likely to be ampicillin- and TMP/SMX-resistant than were azithromycin- and ciprofloxacin-susceptible isolates (Tables 1, 2).

Of the 683 shigellosis case-patients, 183 (27%) had diagnosed HIV infection. Among these, 76 (42%) were infected with DSA isolates, and 7 (4%) were infected with ciprofloxacin-resistant isolates; 108 (59%) were infected with S. flexneri, 73 (40%) with S. sonnei, and 1 each (0.5%) with S. boydii and S. dysenteriae. Of 47 (62%) HIV-diagnosed persons with DSA Shigella infection, 45 (95%) identified as MSM.

Of the 153 persons with DSA- and/or ciprofloxacin-resistant Shigella infection, chart reviews were completed for 111 (73%). Interviews were completed for 97 (64%), and isolates of 80 (82%) of those had DSA to Shigella, 15 (15%) had ciprofloxacin-resistant isolates, and 2 (2%) had isolates resistant to azithromycin and ciprofloxacin. Most case-patients were male (140 [91.5%]); of 120 who completed interviews or were listed in the HIV Surveillance Registry, 102 (85%) identified as MSM. Eleven (12%) of 93 interviewed case-patients who answered the question reported international travel. All interviewees reported symptoms; most common were diarrhea (98%) and abdominal cramps (82%). Median illness duration was 7 days (range 2–45 days). Of 31 (32%) reported hospitalizations, 28 (90%) were infected with DSA and 3 (10%) with ciprofloxacin-resistant isolates; median duration of stay was 3 days (range 1–10 days). Twenty-five (81%) hospitalized case-patients were infected with S. flexneri. In a model that considered age, sex, species (S. flexneri and S. sonnei), HIV status, and neighborhood poverty level, only infection with S. flexneri was associated with hospitalization (odds ratio 4.04, 95% CI 1.46–11.18).

Antibiotics, most commonly ciprofloxacin, were prescribed for 114 (89%) of 128 case-patients (for whom data were available); 16 (13%) received antibiotics to which their Shigella isolates was not susceptible. Fifteen (17%) of 90 patients had taken antibiotics in the 4 weeks before illness onset; all were men, 10 (67%) were HIV-positive, and 12 (92%) of 13 for whom data were available were MSM.

Median illness duration for the 15 (52%) interviewed case-patients infected with ciprofloxacin-resistant Shigella spp. was 7 days (range 2–17 days); 3 (20%) reported hospitalization, and 5 (33%) reported recent international travel.. Five (63%) of 8 case-patients for whom data were available identified as MSM.

In NYC, 19% of nonoutbreak shigellosis cases were caused by organisms with an azithromycin MIC >32 μg/mL, a much higher proportion than the national estimate of 3.8% (9).This finding is troubling because, when antibiotics are indicated, azithromycin is recommended (2,3).

DSA Shigella infection occurred almost exclusively among men, most infected with HIV. Of these, few reported travel, which suggests local acquisition. Transmission of Shigella spp. within networks of MSM has been described by using molecular characterization of isolates to link cases; some NYC cases are likely associated with these networks (10). Whether DSA Shigella infection results in more severe illness or in increased shedding time is unknown. Ten (59%) of 17 persons infected with ciprofloxacin-resistant Shigella spp. reported no exposure to disease-endemic areas (11). Half of non–travel-associated cases occurred among MSM, most were caused by S. sonnei, and few patients were hospitalized, suggesting milder illness. Strains exhibiting both ciprofloxacin resistance and DSA exist in NYC.

HIV-positive MSM may be at increased risk for acquiring infections caused by antibiotic-resistant Shigella spp. due to transmission-facilitating behavior (10) or because of increased exposure to macrolides and fluoroquinolones used to treat sexually transmitted infections, which could increase selective pressure on Shigella organisms (12). HIV infection may increase the risk of acquiring and transmitting Shigella infection due to increased carriage and shedding time or altered immune response (2).

To limit the emergence of resistance, the NYC Department of Health and Mental Hygiene and CDC recommend that antibiotics be avoided in treating Shigella infections except in cases of severe illness or among those at risk for systemic infection (13,14). When antibiotics are prescribed, therapy should be modified on the basis of sensitivity testing; however, without CLSI-defined clinical breakpoints for azithromycin, this process will be challenging (4,14,15).

We were unable to compare illness, exposures, and treatments between persons infected with susceptible versus resistant Shigella spp. We may have overestimated resistance if persons with resistant strains were more likely to have severe or persistent infections, seek care, and have cultures obtained. The clinical significance of resistance is not clear. Finally, molecular characterization of isolates to describe resistance mechanisms and transmission patterns was not done.

Although Shigella infections are generally self-limited, resistant organisms could lead to complications among those who develop systemic infection if they cannot be adequately treated. Studies of clinical breakpoints for azithromycin susceptibility and clinical outcomes are needed. In the meantime, providers should avoid treating otherwise healthy shigellosis patients with antibiotics. When antibiotics are indicated, providers should use available susceptibility results and monitor patient outcomes.