Introduction

There is an urgent need to address the public health and safety problems caused by war. High number of wounded people and severe character of injury, lack of medical care, and significant number of displaced persons led to high morbidity and mortality among people affected by war1. In addition, combatants usually receive soil-contaminated wounds, and they are typically treated in informal facilities with inadequate infection control, leading to high rates of wound infections. A survey conducted by Darvishi (2023) has shown that infection remains a prevalent and deadly consequence of combat-related injuries2.

Another important aspect of safety related to warfare is antimicrobial resistance (AMR). Some authors suggest that war-related wounds can accelerate the formation of multidrug resistant (MDR) organisms3 and human conflicts drive the growth of AMR4. Previous studies mainly indicated that armed conflicts occur primarily in low- and middle-income countries, which already face tremendous challenges in surveillance and control against antimicrobial resistance5. However, the extent to which ongoing armed conflicts contribute to the global AMR crisis remains poorly understood.

Several studies have attempted to examine the combat wound microbiome and the antibiotic resistance profile of isolated pathogens. Up to now, all reports on this topic could be divided into three groups. The first group includes investigations carried out in conflict and non-conflict-affected Middle East countries6. The second group reported data on civilian and non-civilian patients transported in the USA and Europe from the conflict setting7,8,9. Last group describes the spread of AMR among combatants in Ukraine since 2014 until 2022 years10,11,12. Overall, these studies illustrate the transformation of wound microbial flora and changes of antimicrobial profile over the past two decades. Similar trends were observed in conflict and non-conflict-affected countries. Two important themes emerge from the studies discussed so far: the prevalence of Gram-negative MDR bacteria (A. baumannii, K. pneumonia, and P. aeruginosa) among combat wound pathogens and the impact of the combat wound microflora on the general situation with AMR.

A better understanding of MDR carriage and infection in wounded military personnel will be useful in improving patient care and outcomes. In addition, it will decrease the risk of AMR spreading worldwide and limit the need to use empiric broad-spectrum perioperative and point-of-wounding antibiotics.

Therefore, permanent examination of combat-related wound microbiota with determination antibiotic resistance of the isolated pathogens are essential elements of improving trauma patient care as well as prevent global spread of antibiotic resistance.

We aimed to investigate the impact of war on wound microbiota and AMR distribution among military patients treated in Ukrainian civilian hospitals, to help revise empirical antibiotic prophylaxis and treatment protocols adopted in these settings.

Methods

Description of the patients

This was a retrospective cohort study conducted in Sumy Regional Center for Disease Control and Prevention of the Ministry of Health of Ukraine. We analyzed data obtained during January–April 2024 from the study of wound materials collected from patients hospitalized in three civilian hospitals of Sumy city (northeastern Ukraine, 30 km from the border with Russia).

The inclusion criterion for the study was the presence of a purulent wound in military patients receiving treatment in hospitals. Combatants were injured mainly in other regions of Ukraine, where fighting took place and were taken for treatment to Sumy hospitals. Specimens from the patients with signs and symptoms of infected wounds were sent to the laboratory center for testing. All details regarding gender, age, time of admission and collection of the sample, wound characteristics were extracted from the hospital records. All patients included in the investigation signed an informed consent at facilities where they were treated. All relevant ethical standards were followed.

Microbiological examination of wound samples

Samples were collected using a wound swab or tissue biopsy, then transported to a laboratory center and cultivated on a set of media (blood agar, Endo agar, malonate salt agar, enterococcal agar). Phenotypic identification of isolated strains was performed by determining their morphological, staining, biochemical, and antigenic features with the assistance of automated bacterial identification system (VITEK® 2 Compact 30 BioMérieux’s). Microorganisms were examined for susceptibility to cephalosporins, carbapenems, β-lactams, glycopeptides, aminoglycosides, fluoroquinolones, glycylcycline and macrolides according to EUCAST (European Committee on Antimicrobial Susceptibility Testing) guidelines with the disk diffusion method13. The strains resistant to at least one antibacterial drug in three or more antibacterial categories were determined as MDR according to the European Centre for Disease Prevention and Control (ECDC) and the US Centers for Disease Control and Prevention (CDC) definitions. MDR was determined for the following bacteria: Staphylococcus aureus, Enterococcus faecalis, Klebsiella pneumoniae, Escherichia coli, Pseudomonas aeruginosa and Acinetobacter baumannii. Statistical significance was determined using Student’s T-test and Kolmogorov–Smirnov test. The difference was significant at p < 0.05.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Results

Description of the patients and isolates

Seventy-three military patients were enrolled in the study within 4 months at the beginning of the year 2024. They accounted for 3.45% of all patients hospitalized in surgical departments that were included in the observation during the study period. All military patients were male with a median age of 40.76 ± 8.16 years. The average time from hospitalization to wound sampling was 2.63 ± 0.6 days. The median time from hospitalization to wound sampling was 1 day with an interquartile range 1 day. The distribution of patients depending on clinical diagnosis is shown in Supplementary Table 1.

The analysis of wound microbiota is shown in Fig. 1. As can be seen from the graph, 59 cultures were isolated from 41 (56.16 ± 0.67%) military patients with an average number of swab culture per wound equal 1.28 ± 0.44 (range 1–3). Thirty-two samples (43.84 ± 0.67%) showed no growth.

Fig. 1: Bacteria species isolated from the combatant wounds.
figure 1

Names of species and number of strains isolated from the wound of military patients during survey period are shown in bar graph.

Full size image

The monoculture was detected in 75.6 ± 1.01% positive samples. Two or three bacteria were isolated in ten samples (24.4 ± 0.67%). Gram-negative bacilli were the most prevalent bacteria. There were 32 isolated strains (54.2 ± 0.84%) Gram-negative rods, 19 strains (32.2 ± 0.79%) Gram-positive cocci, and 8 strains (13.6 ± 0.58%) Gram-positive rods. The dominant bacteria isolated from the military patients were A. baumannii (36 ± 0.81%), E. faecalis (12 ± 0.55%), and B. cereus (12 ± 0.55%).

Resistance profiles

The susceptibility of the main bacteria to antibiotics was examined for 43 strains isolated from military patients. We have tested bacteria on susceptibility to amikacin (Ami), tobramycin (Tob), gentamicin (Gen), ciprofloxacin (Cip), levofloxacin (Lev), moxifloxacin (Mox), ceftazidime (Cef), ceftazidime-avibactam (Cef-A), ceftriaxone (Cft), aztreonam (Azt), piperacillin/tazobactam (Pip-T), ampicillin sulbactam (Amp-S), trimethoprim (Tri), meropenem (Mer), doripenem (Dor), imipenem (Imi), norfloxacin (Nor), linezolid (Lnz), tigecycline (Tig), teicoplanin (Tec), vancomycin (Van), clindamycin (Cli), erythromycin (Er). The pattern of antibiotic susceptibility was species-specific. The susceptibility of the commonly isolated strains is shown in Fig. 2. Strains of A. baumannii were resistant to fluoroquinolones, carbapenems, and sulfanilamide. K. pneumoniae exhibited the highest proportion of resistant strains to cephalosporins and β-lactams. At the same time, 40% tested strains were resistant to carbapenems. All gram-negative bacteria isolated from combatants demonstrated a higher rate of resistant strains compared to gram-positive strains. All strains of E. faecalis (100 %) were resistant to imipenem and sensitive to norfloxacin, vancomycin, and sulbactam. Almost half of B. cereus strains were resistant to carbapenems. Assessment of antibiotic resistance found that 84.6 ± 0.83% strains isolated from combatants were MDR according to ECDC definition (Fig. 3).

Fig. 2: Susceptibility to antibiotics of the most frequently isolated bacteria from combatants.
figure 2

a Percentage of A.baumanii resistant strains to antibiotics: amikacin (Ami), tobramycin (Tob), gentamicin (Gen), ciprofloxacin (Cip), levofloxacin (Lev), trimethoprim (Tri), meropenem (Mer), doripenem (Dor), imipenem (Imi) is shown in bar graph. b Percentage of K.pneumoniae resistant strains to antibiotics: Ami, Tob, moxifloxacin (Mox), Cip, ceftazidime (Cef), ceftazidime-avibactam (Cef-A), Mer, aztreonam (Azt), piperacillin/tazobactam (Pip-T), ampicillin sulbactam (Amp-S) is shown in bar graph. c Percentage of E.faecalis resistant strains to antibiotics: norfloxacin (Nor), linezolid (Lnz), tigecycline (Tig), Imi, Amp-S, teicoplanin (Tec), vancomycin (Van) is shown in bar graph. d Percentages of B. cereus resistant strains to antibiotics: clindamycin (Cli), erythromycin (Er), Nor, Cip, Lev, Lnz, Mer, Imi are shown in bar graph. The total number of the examined strains for each species is written as number under the X-axes in (a-d).

Full size image
Fig. 3: The rate of MDR strains among the most frequently isolated pathogens from combatant wounds.
figure 3

The names of species and percentage of multidrug-resistant (MDR) strains among the most frequently isolated pathogens from the wounds of combatants during the survey period are shown in the bar graph.

Full size image

The full data set is reflected in the Supplementary data 1–3.

Discussion

Previous studies evaluating the bacteriological profile of wound infections indicated dependance of the microbiota and antibiotic resistance on the type of conflict, its location and condition of medical care10. The present study was designed to determine the war impact on the wound microbiota and spread of antimicrobial resistance in population as well as promote a revision of empirical antibiotic prophylaxis and treatment protocols.

Although other studies linked sex to the prevalence of wounds in the population14, we could not confirm this due to the artificial sex disproportion in the army. Previous reports noted the severe nature of the injuries, the slow speed of medical care15,16, and the primary contamination of the wound17 in the combatants. In our study, blast trauma was the leading reason for hospitalization. Obviously, these factors lead to poor regeneration and complicated healing, even in middle-aged combatant patients. The average time from hospitalization to wound sampling was short. Because more than half of military patients had clinical signs of purulent wounds at admission to the hospital.

The next question in this research was the examination of the wound microbiota. This study found that microorganisms were isolated from the wound samples in approximately half of the cases. Explanations for this might be severe character of injury that delays the healing process in this group even without microbial components or the presence of persisters. The results of this study also showed a high frequency of mixed cultures isolation (24.4% samples). This result is likely to be related to the fact that these patients usually are treated at several medical care facilities and have contact with microbiota of different hospitals.

Analysis of microbiota profile showed the dominance of gram-negative bacteria with prevalence A. baumannii. The data obtained in this research match those observed in study10 that was performed in Ukraine ten years ago at the beginning of the war. Our results also confirm data from a US military health system report on a significant increase in the incidence of Acinetobacter among military personnel who participated in military conflicts in Afghanistan and Iraq18. We suggest that the pathogens of combat wounds are represented by hospital microbiota. Military patients are typically treated in at least two facilities before their final hospital admission: a Stabilization Point for Medical Care and the nearest civilian or military hospital to the line of contact. After receiving primary medical care, they are transported to a hospital located in the interior of the country. The ability to fully comply with infection control rules and regulations at the stabilization point is minimal. These facilities, with a high circulation of MDR organisms, inadequate IPC measures, and weak antimicrobial stewardship programs, may contribute to wound contamination with nosocomial flora, leading to the further spread of these microorganisms in other medical institutions. This hypothesis is consistent with the previous report19 and results of a pilot study on nosocomial infections conducted by the Ukrainian Public Health Center in 2021. These highlights the primary role of hospital-acquired flora as a cause of wound infection in military patients and gram-negative bacteria as the dominant causative agents of this type of infection in Ukraine (https://phc.org.ua/news/ukraina-potrebue-vprovadzhennya-sistemi-ocinki-yakosti-nadannya-medichnoi-dopomogi-i-programi). Thus, an increase in the number of military personnel with infected wounds in civilian hospitals may contribute to the persistence of nosocomial microbiota. Strict adherence to IPC standards and treating military patients as primarily infected with nosocomial flora remain crucial to preventing its spread.

Examination of the wound pathogens susceptibility to antibiotics demonstrated the high rate of resistant strains among gram-negative bacteria isolated from the patients in Ukrainian civilian hospitals. Our findings are consistent with that of Krishtafor et al. 20 who reported growth of carbapenems resistance among wound pathogens in military patients. In contrast to this study, however, we demonstrated low susceptibility of wound bacteria to fluoroquinolones and aminoglycosides. This difference supports the idea that the level and pattern of antibiotic resistance are unique for each hospital and even department. According to these data, use of uniform national therapy protocols that is common for Ukrainian military medicine is a questionable idea. Wide implementation of local protocols with permanent monitoring of AMR could be a helpful solution to increase effectiveness of wound healing.

Another important finding was the high frequency of MDR among gram-negative rods isolated from combatants. Moreover, a significant difference was found between gram-negative and gram-positive bacteria groups regarding the number of MDR isolates (p-value < 0.05), with the predominance of the last one. Due to the limited number of samples, this study has been unable to demonstrate war contribution to the spread of antibiotic resistance in the population. However, there is a trend toward colonization of military patients with MDR strains from the WHO priority list during their treatment. Further research is needed to fully understand war impact on AMR spread.

According to the report for CDC made by Kuzin et al. 21 there is an inadequacy in implementation of infection prevention and control measures, such as recommended hand hygiene, and monitoring, evaluation, and feedback to the hospital staff members in Ukraine. Our data indirectly support this conclusion. It is obvious that to limit the spread of AMR in conflict areas, we need to implement compliance with IPC norms at each stage of providing medical care to military person and raise the level of awareness of health care workers on this issue.

Finally, several important limitations should be considered. First, the limited number of samples makes it difficult to generalize these findings. Second, the study of phenotypic resistance does not reflect the genetic potential of resistance. Finally, whole-genome studies of isolated strains could help to establish the origin of wound pathogens.

Conclusion

This paper argued about the impact of war on the wound microbiota and spread of antimicrobial resistance in population. This study found that wound infections in military patients were associated with age 40 years. Other finding was the dominance of gram-negative bacilli with prevalence A. baumannii. This investigation demonstrated a high level of resistance to carbapenems, fluoroquinolones, aminoglycosides and a significant prevalence of MDR in gram-negative bacteria compared to gram-positive microbes. Overall, this study strengthens the idea that the wide implementation of the IPC standards and local protocols with continuous monitoring of antimicrobial resistance across all types of healthcare in Ukraine can limit the spread of antimicrobial resistance and improve the effectiveness of treatment.