Reunion, a sentinel island for multidrug-resistant bacteria surveillance in South-western Indian ocean: a retrospective survey using hospitalised patients screening, 2015-2017

Methods Based on exhaustive hospital MRB screening laboratory data, we provided the first estimation of MRB colonization rates in hospitalized patients residing in SWIO (2015-2017). Using Reunion (European Union) MRB colonization rate as a baseline we identified at risk patients according to their territory of residence. The survey out significantly higher overall rates in patients from Mayotte, and when compared to Reunion with extended-spectrum betalactamase producing Enterobacteriaceae as a generalized public health issue for these territories. Specific epidemiological dynamics were pointed out in Mayotte and significantly higher and with higher carbapenemase producing Enterobacteriaceae (CPE) colonization rate. These results


Methods
Based on exhaustive hospital MRB screening laboratory data, we provided the first estimation of MRB colonization rates in hospitalized patients residing in SWIO (2015SWIO ( -2017. Using Reunion (European Union) MRB colonization rate as a baseline we identified at risk patients according to their territory of residence.

Results
The survey pointed out significantly higher overall MRB colonization rates in patients from Comoros, Madagascar, Mayotte, and Seychelles when compared to Reunion (France) with extended-spectrum betalactamase producing Enterobacteriaceae as a generalized public health issue for these territories. Specific epidemiological dynamics were pointed out in Mayotte and Seychelles patients with significantly higher methicillin-resistant Staphylococcus aureus (MRSA) colonization rates; and in Mauritius patients with higher carbapenemase producing Enterobacteriaceae (CPE) colonization rate.

Conclusion
These results could be the first step to set up a regional MRB surveillance in SWIO hospitals.

Background 3
Since the 90's, an increase of the prevalence of multidrug-resistant bacteria (MRB) was observed in Reunion (France). 1

Data collection, inclusion criteria, and statistical analyses
We performed a retrospective survey on all patients admitted to the Felix-Guyon University hospital in Reunion from 2015 to 2017 (main hospital). Only SWIO residing patients were included. All these patients were screened for MRB detection (i.e. anal and nasal swabbing) and patients admitted to intensive care unit were used as baseline for odds ratio estimation as all patients from Reunion intensive care were screened.
The MRB included were MRSA, ESBL-E, carbapenemase-producing Enterobacteriaceae A patient was considered MRB positive if carrier of one, or more than one MRB. MRB colonization rates in patients were compared according to their country of residence using χ² test with Fisher correction or Fisher exact test if theoretical frequency were less than 5.
Odds ratio were calculated with Reunion patients as reference and threshold was set at pvalues <0.05. Confidence intervals were estimated using the Woolf method. The statistical analyses were performed using R software version 3.4.2.
The study was approved by the French national commission on data protection and liberties (reference 2210228 v0, 10 th of January 2019).

Bacterial isolates and antibiotic susceptibility testing
Bacterial isolates were routinely obtained at hospital laboratory and species identified for all isolates using MALDI-TOF mass spectrometry (Bruker Daltonics, Breme, Germany).
ESBL-E, and CPE were screened using two selective chromogenic agar ChromID-ESBL and ChromID CARBA SMART (bioMérieux, Marcy l'Étoile, France). ESBL-E phenotypes were confirmed using combined disk synergy testing. MRB colonized patients were significantly higher in SWIO (excluding Mauritius patients) when compared to patients from Reunion as a baseline, ESBL-E colonization rates followed the same trend (Table 1). MRSA colonization rates were significantly higher in Seychelles and Mayotte hospitalized patients than in Reunion patients. CPE colonization rates were significantly higher in patients residing in Mauritius when compared to Reunion patients.

Consent for publication
Not applicable.

Availability of data and materials
The data were obtained from Felix-Guyon hospital laboratory, Saint Denis, La Réunion.
Data were provided for the survey according to the French law (reference 2210228 v0, 10 th of January 2019) but are not publicly available.

Competing interest
The authors declare that they have no competing interests.

Funding
This project was ordered and funded by the Indian Ocean Health Agency (Noellie Gay PhD fellowship).

Contributions
NG conceived the study. NG, OB and EC defined the study methods. OB, GM, NL provided the database. NG performed the analyses and drafted the manuscript. All authors provided comments and approved the final version of the manuscript.