Extensively Drug-Resistant Carbapenemase-Producing Pseudomonas aeruginosa and Medical Tourism from the United States to Mexico, 2018–2019

Ian Kracalik; D. Cal Ham; Gillian McAllister; Amanda R. Smith; Maureen Vowles; Kelly Kauber; Melba Zambrano; Gretchen Rodriguez; Kelley Garner; Kaitlyn Chorbi; P. Maureen Cassidy; Shannon McBee; Rhett J. Stoney; Kathleen Moser; Margarita E. Villarino; Oscar E. Zazueta; Amelia Bhatnagar; Erisa Sula; Richard A. Stanton; Allison C. Brown; Alison L. Halpin; Lauren Epstein; Maroya Spalding Walters

Disclosures

Emerging Infectious Diseases. 2022;28(1):51-61. 

In This Article

Results

Epidemiologic Investigation

During August 1, 2018–December 31, 2019, we identified 44 cases from 19 states; 25 were confirmed cases, 13 were probable cases, and 6 were suspected cases (Figure 1). Among the 38 patients who had confirmed or probable cases, 34 (89.5%) were female, and the median age at time of specimen collection was 39 (interquartile range 31–48) years (Table 1). Sleeve gastrectomy was the most common surgical procedure, reported by 34 (89.5%) of 38 case-patients. Median time from surgery to specimen collection was 12 (range 3–40) days. After surgery in Tijuana, 16 (42.1%) of 38 case-patients were hospitalized in the United States. Among the 14 hospitalized patients for which the duration of hospitalization was known, the median stay was 7 (range 1–19) days.

Figure 1.

Confirmed and probable cases of infection with Verona integron–encoded, metallo-β-lactamase–producing, carbapenem-resistant Pseudomonas aeruginosa, by state in which bacterium was identified, among US medical tourists undergoing elective invasive procedures in Tijuana, Mexico, January 2018–December 2019. Six suspected cases, from Arizona (n = 1), Georgia (n = 3), Michigan (n = 1), and Washington (n = 1) are not shown.

Four hospitalized case-patients were admitted to the intensive care unit; for 8 case-patients, this admission status was unknown. Six hospitalized case-patients underwent surgery because of their infection; for 5, surgery for postsurgical infection management status was unknown. One of the 16 hospitalized case-patients died in the in the hospital 9 days after sleeve gastrectomy surgery. For this patient, who underwent a procedure at facility E by surgeon 1, VIM-CRPA was identified from a screening rectal swab specimen at admission. Whether the patient had VIM-CRPA infection or the surgery at facility E otherwise contributed to death is unclear from available medical records. All other case-patients had VIM-CRPA infections on the basis of results from clinical cultures. From our investigation, no evidence of onward transmission in the US healthcare facilities in which case-patients were hospitalized was identified.

For the confirmed and probable cases, 37 (97%) case-patients named 10 Tijuana facilities in which they underwent invasive procedures. Most reported surgery at facility A (27/38; 71.1%) (Table 1). Among the 35 case-patients who reported the name of their surgeon, 31 (88.6%) named surgeon 1, including the 27 case-patients who underwent surgery at facility A and 4 case-patients who underwent surgery at other or unknown facilities.

Surgery dates ranged from August 2018 to August 2019 (Figure 2). Confirmed and probable cases in patients who underwent surgery at facility A peaked during January 2019 (epidemiologic weeks 2–5); 13 (48.2%) of 27 case-patients who reported surgery at facility A had a procedure during this 4-week period when the surgical suite was reported closed. Facility A closed permanently in early March 2019; 4 case-patients associated with surgeon 1 had surgery after this date at facility I (n = 2 case-patients), facility E, and an unknown facility in the Tijuana area. Ongoing monitoring through December 2019 did not identify additional cases linked to facility A after January 2019 (epidemiologic week 5) or to surgeon 1 after July 2019 (epidemiologic week 29).

Figure 2.

Confirmed and probable cases of infection with Verona integron–encoded, metallo-β-lactamase–producing, carbapenem-resistant Pseudomonas aeruginosa, by week of surgery, among US medical tourists undergoing elective invasive procedures in Tijuana, Mexico, January 2018–December 2019. Dark blue bars show cases associated with surgery performed at facility A by surgeon 1; light green bars show cases associated with surgery at facilities D–H and J by surgeons other than surgeon 1; and light blue bars show cases associated with facilities B, C, and I by surgeon 1; and dark green bar shows a case associated with surgeon 1 and an unknown facility.1. A confirmed case was isolation of Verona integron–encoded, metallo-β-lactamase–producing, carbapenem-resistant P. aeruginosa from a patient who had an elective invasive medical procedure in Mexico during January 2018–December 2019 and within 45 days before specimen collection. A probable case was isolation of carbapenem-resistant P. aeruginosa, with an isolate unavailable for carbapenemase testing, from a patient who had an elective invasive medical procedure in Mexico during January 2018–December 2019 and within 45-days before specimen collection. No cases were identified from patients who underwent surgery before August 2018 (week 34). The peak of the outbreak encompassed epidemiologic weeks 2–5 (January 2019).

Patient Notification and Active Case Finding

During August 1, 2018–March 1, 2019, travel agency A referred 793 persons from 6 countries for surgery at facility A; of these persons, 743 (94%) were US residents. Health authorities in the other countries were contacted to inform them of the outbreak, and we were not notified of any cases. We interviewed 160 (21%) US residents who underwent surgery, including 92 (46%) of 200 persons in the higher risk group targeted for active outreach and 68 (13%) of 543 persons in the lower risk group and for whom some health jurisdictions performed active outreach. Fifteen cases were identified through interviews. Overall, passive and active case finding identified 7 confirmed case-patients and 6 probable case-patients who underwent surgery at facility A and who were among the 200 persons in the higher risk group; therefore, the attack rate for VIM-CRPA for persons who had surgery at facility A during January–March 2019 was 13/200 (6.5%, 95% CI 3.6%–10.8%).

Interviewed persons who were not confirmed or probable case-patients (n = 148) were demographically similar to confirmed and probable case-patients (n = 38). The most common reason for undergoing surgery abroad was lower cost, reported by 132 (82.5%) of the 160 interviewed persons. Among the 41 persons who reported being aware of the CDC travel advisories or negative media stories before their surgery, cost was the most common reason for proceeding with surgery (n = 22, 53.7%) (Table 2, https://wwwnc.cdc.gov/EID/article/28/1/21-1880-T2.htm)..

Microbiologic Investigation

Isolates from 22 of 25 confirmed cases underwent WGS. All isolates harbored bla VIM-2; 21 were sequence type (ST) 111 and 1 was a novel ST. Overall, isolates varied by 0 to 4,375 single nucleotide variants (SNVs) over a 90.08% core genome (Figures 3, 4). Seventeen isolates formed a distinct cluster varying by 0 to 4 SNVs over a 93.29% core genome and were associated with surgeon 1 (n = 16) or an unknown surgeon (n = 1) and with ≥3 different facilities. One isolate associated with surgeon 1 was the novel ST and differed by 4,375 SNVs, indicating that it was not closely related to other isolates associated with facility A (Figure 3). The remaining 4 isolates differed by 5–18 SNVs over an 96.34% core genome; among these, 2 were associated with facility E but were not more closely related to each other than to isolates associated with facilities B and G. Antimicrobial susceptibility testing was performed for 10 isolates at the request of health departments to guide treatment. All isolates were XDR and resistant to ceftazidime/avibactam and ceftolozane/tazobactam (Table 3).

Figure 3.

Whole-genome sequencing analysis and selected epidemiologic data for 22 Verona-integron-encoded metallo-β-lactamase-producing carbapenem-resistant Pseudomonas aeruginosa clinical isolates from US medical tourists who underwent surgery in Tijuana, Mexico, August 2018–December 2019. Phylogenetic tree includes an outlier isolate from Arkansas. On the right, the first group of 8 columns indicates facilities (A, B,C, E, F, G, I, and unknown), and the second group of 4 columns indicates surgeons (1, 2, 3, and unknown). Scale bar indicates nucleotide substitutions per site.

Figure 4.

Whole-genome sequencing analysis and selected epidemiologic data for 21 Verona-integron-encoded metallo-β-lactamase-producing carbapenem-resistant Pseudomonas aeruginosa clinical isolates from US medical tourists who underwent surgery in Tijuana, Mexico, August 2018–December 2019. Phylogenetic tree excludes an outlier isolate from Arkansas. On the right, the first group of 8 columns indicates facilities (A, B,C, E, F, G, I, and unknown), and the second group of 4 columns indicates surgeons (1, 2, 3, and unknown). Scale bar indicates nucleotide substitutions per site.

processing....