European hospitals as source of multidrug-resistant bacteria: analysis of travellers screened in Finland after hospitalization abroad

Abstract Background As hospitals have a high prevalence of multidrug-resistant organisms (MDRO), hospitalization abroad indicates for travellers an increased risk of acquiring MDRO—and carrying the strains home. Antimicrobial resistance (AMR) rates are highest in the (sub)tropics, whereas Europe is considered a lower risk region. Since AMR prevalences vary within Europe, we aimed to gather country-specific data on the risks for hospitalized travellers. Methods At hospitals of the Helsinki and Uusimaa district in Finland, patients hospitalized abroad over the past 12 months are systematically screened for methicillin-resistant Staphylococcus aureus (MRSA), extended-spectrum beta-lactamase (ESBL)-producing Enterobacterales (ESBL-PE), carbapenemase-producing bacteria and vancomycin-resistant Enterococcus spp. (VRE). Among patients screened 2010–19, we selected those hospitalized in Europe, recorded their MDRO findings, infections and mortality, and analysed MDRO-associated risk factors. Results Of the 1772 patients treated in 41 European countries, 16.6% (295) carried MDRO, 12.5% (221) ESBL-PE, 7.8% (138) solely ESBL-E. coli, 2.6% (46) MRSA, 2.2% (30) of those screened VRE and 2.2% (39) carbapenem-resistant Gram-negatives. Among those colonized, 9.8% (29) had symptomatic MDRO infections and 0.3% (one) died. Colonization was most frequently recorded for those treated in eastern and southern Europe, with Bulgaria, Cyprus and the Russian Federation scoring highest. MDRO colonization was associated with antibiotic treatment and showed a negative correlation with time from discharge to screening. Conclusions After hospitalization in European countries, ESBL-PE carriage was relatively common (12.5%), while other MDROs proved less frequent (<5%). Antibiotic treatment and short time since hospitalization abroad increased the risk of MDRO colonization. Clear differences between countries and regions were revealed, with highest rates in the east and the south.


Results
Of the 1772 patients treated in 41 European countries, 16 antibiotic treatment and showed a negative correlation with time from discharge to screening.

Conclusions
After hospitalization in European countries, ESBL-PE carriage was relatively common (12.5%), while other MDROs proved less frequent (<5%). Antibiotic treatment and short time since hospitalization abroad increased the risk of MDRO colonization. Clear differences between countries and regions were revealed, with highest rates in the east and the south. Travellers hospitalized abroad are at particular risk of acquiring resistant bacteria.
Indeed, several investigations in Europe report MDRO carriage rates as high as 44% after healthcare contact abroad [2][3][4][5][6][7][8][9] . According with other traveller studies, the risk varies by region: in our previous investigation in Finland screening for MDRO among patients hospitalized abroad, the highest risk was associated with the Indian subcontinent followed by Southeast Asia, Africa, and South America 10 . Those treated in Europe had considerably lower carriage rates than those treated in (sub)tropical regions 10

Study design and selection of participants
Drawing upon a regional infection control database run by the Helsinki and Uusimaa Hospital District with a population of 1.7 million, we identified patients screened for both methicillin-resistant Staphylococcus aureus (MRSA) and multidrug-resistant gram-negative bacteria (MDRGNB) within a 30-day time frame between January 2010 and August 2019.
Our basic inclusion criteria comprised adequate records of screens involving 1) three-site MRSA cultures (nose, throat, and groin or perineum); and 2) stool specimens or rectal swabs for MDRGNB cultures. Another criterion to be met in medical records was hospitalization for over 24 hours or a major invasive procedure while abroad in a European country within 12 months before screening. We defined as invasive any procedure which required anaesthesia or could not be carried out bedside, thus including for example any major surgery. A documented or deducible discharge date within a 30-day time frame was also required.
Documented travel outside Europe or hospitalization in multiple countries over the preceding 12 months led to exclusion.

Ethical statement
The research board of Helsinki University Hospital Department of Internal Medicine approved this study. In accordance with the Finnish Medical Research Act, an ethics committee review was not required, as there was no intervention.

Collection of data on MDRO colonization and clinical infections
We recorded all MDROs found in screening and clinical samples within one month (

Collection of data on risk factors
On the basis of medical records, the Charlson comorbidity index (CCI) 13 , verified or suspected alcohol abuse, surgery and intensive care unit (ICU) admission abroad were recorded. Antibiotic exposure abroad was listed for those with documented use, bacterial infection treatment or surgery routinely requiring prophylaxis; this included oral and parenteral antibiotics given before, during or after hospitalization prior to admission to our hospitals. These were all classified as negative, if not mentioned. Type of travel was recorded either as residence, known or suspected visit to friends and relatives (VFR), or work/holiday/other. The duration of hospitalization abroad was also listed, or in cases with multiple instances, the sum within one month.

Microbiological methods
The microbiological methods in routine clinical use at the Helsinki University Hospital during the study were described in our earlier paper 10

Statistics
The statistical analyses were performed using SPSS v. 25.0 (IBM Corp., Armonk, NY).
Univariate analysis of risk factors was carried out by χ 2 test, Fisher's exact test or binary logistic regression. Independent variables with p-values below 0.2 in univariate analysis were included in multivariable analysis if they did not correlate strongly. In multivariable logistic regression, backward selection based on Akaike information criteria was used.

Patient characteristics
The final study population comprised 1772 patients ( Figure 1) Fourteen patients showed carbapenemase gene-positive MDR Acinetobacter baumannii; two Acinetobacter strains had not been tested. Of the 12 individuals with MDRPA, three had carbapenemase gene-positive and five gene-negative strains; the isolates of four had not been tested.
The MDRO rates were higher among ICU-treated than non-ICU-treated patients: to screening showed a negative association. The same three associations were also found for ESBL-PE carriage. Antibiotic use and alcohol abuse were revealed as independent risk factors of acquiring MRSA.
To summarize the effects of the risk factors observed, we conducted a subgroup analysis of patients having been given antibiotics and screened within 30 days of discharge, versus no antibiotics and an interval of over 90 days, and found MDRO colonization rates of 26.1% (141/541) and 9.2% (20/217), respectively.

DISCUSSION
Although for travellers the risk of acquiring MDRO is highest in the (sub)tropics 1, 10 , colonization proved common (16.6%) also among our 1772 cases recently treated in European hospitals. Our data reveal substantial variation in the risk by country and European subregion.

MDRO colonization after hospitalization in various parts of Europe
The MDRO rates were highest in the east, followed by southern, western, and northern subregions, in this order. Although geographic grouping of countries is not ideal due to intercountry differences, this general finding accords with other studies, such as the north-to-south and west-to-east AMR gradients reported by EARS-Net 11 . Similar gradients have also been shown by others. Kaiser

Risk factor analyses
In addition to risks associated with specific countries and subregions, antibiotic use and a  The funding sources were not involved in study design, in the collection, analysis or interpretation of data, writing of the manuscript nor the decision to submit for publication.

Conflict of interest
MK has participated in a conference on the expense of Astellas Pharma. AK has received investigator-initiated grants (Valneva, Pfizer) and on an individual occasion consulted an advisory board (Valneva). None of the interests listed above are relevant to the current manuscript. TK reports no potential conflicts of interest.