Burn wound infections are a global health problem in both developing and underdeveloped countries [1]. In India, more than 6 to 7 million people are exposed to moderate or severe burns annually [2]. Seventy-five percent of all deaths in these patients are due to infectious complications such as sepsis and bacteremia [1]. Burn wound surfaces are sterile immediately following thermal injury, but within 48 hours, they become colonised with the patient’s own skin commensals. After approximately seven days, the wound becomes colonised with endogenous organisms from the patient’s gastrointestinal or respiratory tract, or exogenously from the hospital environment [1]. Exogenous organisms from the hospital environment are generally more resistant to antibiotics than endogenous organisms. It is therefore desirable to conduct periodic reviews of the bacterial flora of burn wounds in all centres so that preventive strategies can be designed accordingly.
Temporal changes in microbial communities reflect constant fluctuations and recurrent variations over time in the community structure, composition, or function of microbes and are governed by both intrinsic and extrinsic factors [1,3]. Burn wounds serve as a susceptible site for opportunistic colonisation by endogenous and exogenous organisms due to local and systemic immunosuppression [4]. This leads to changes in the bacterial infection profile over time depending on which organisms the wounds are exposed to at any given time in hospitalised patients [4,5]. Colonisation during prolonged hospitalisation increases the risk of invasion, systemic complications and death. The characteristics of different organisms also play major roles in their invasiveness. Invasive infection is indicated by inflammation of surrounding uninjured tissue, positive blood cultures with a pathogen isolated and systemic signs of sepsis with no other identifiable source of infection [1]. Staphylococcus aureus is the most common cause of early burn wound infection [1,6]. Other organisms that may cause infection during the course of admission to the ward include Pseudomonas aeruginosa, Klebsiella species, Proteus species, Acinetobacter species and Enterococcus species [1,5,7,8]. The aim of the present study was to determine the temporal changes in the bacterial profile of burn wound infections. The primary objective was to identify the frequency of association of various species of bacteria with burn wounds, to determine their susceptibility patterns, and to identify any temporal associations. The secondary objective of the study was to identify the risk factors for invasion.
Materials and Methods
This prospective cohort study was conducted over a period of 15 months, from June 2018 to September 2019. Since this was a time-bound prospective cohort study, samples from all patients admitted for burn wound infections during the aforementioned period were considered as study participants (n=131). Permission was obtained from the Institutional Ethics Committee (IEC) before commencing this study (EC/113/2017 dated 28th November 2017). The procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional) and the Helsinki Declaration of 1975, as revised in 2000.
Inclusion criteria: All patients admitted to the burns unit in this hospital during this period, including those referred from outside, were enrolled after receiving written informed consent from adult participants and assent from children aged over seven years, along with parental consent were included in the study.
Exclusion criteria: Patients who did not consent to the study were excluded from the study.
A detailed history and associated risk factors, as per the available literature such as age, days of hospitalisation, percentage of Total Burned Surface Area (TBSA), depth of burns, type of referral, diabetes, hypertension and pregnancy, were recorded [7-9].
Specimen collection: Two swab specimens were aseptically collected from the burn wound using sterile cotton swabs moistened with sterile normal saline. Swabs were collected periodically on days 0, 4, and 8 and thereafter every week until the completion of the hospital stay or until the death of the patient. Specimens were processed according to the standard protocol and isolates were identified up to the species level by routine phenotypic tests. Antimicrobial susceptibility testing was carried out using the Kirby-Bauer disc diffusion method [9-13]. Growth of three or more colony morphologies with no polymorphonuclear cells on Gram’s stain was excluded. A repeat specimen was used in these cases.
For reporting Coagulase-Negative Staphylococci (CoNS) from the burn wound, a detailed history, including clinical signs of infection and associated risk factors, was taken, along with repeat testing with a freshly collected sample. If a patient developed an invasive infection, systemic specimens, such as blood, sputum, tracheal aspirate, urine, or tissue, as appropriate, were processed for culture according to the standard protocol [8,10,11]. Specimens that showed no growth on wound swabs were considered non infected, whereas those that showed growth only in wound swabs were considered localised infections/colonisation. Those that showed growth in both wound swabs and blood cultures were considered invasive infections.
Depending on the depth of injury, burns were classified into superficial (involving only the epidermal layer of skin), partial-thickness (involving the epidermis and portions of the dermis) and full-thickness (extending through and destroying all layers of the dermis and often injuring the underlying subcutaneous tissue) [14].
For estimating the percentage of TBSA burned for risk factor analysis, the ‘Rule of Nines’ was followed: head - 9%, each arm - 9%, anterior chest and abdomen - 18%, posterior chest and back - 18%, each leg - 18%, and perineum - 1% [15].
Statistical Analysis
Data were analysed using Statistical Package for the Social Sciences (SPSS) software version 20.0. The organisms isolated and their susceptibility were expressed as percentages, which were analysed by linear regression analysis, and the p-value was calculated. For risk factor analysis, participants were categorised into those with invasive infections and those with non invasive infections, and the p-value and Odds Ratio (OR) were calculated. The risk factors, i.e., age, days of hospitalisation, percentage of TBSA, depth of burns, type of referral, diabetes, hypertension, and pregnancy, were analysed using the Chi-square test. Statistical significance was set at p-value <0.05. The strength of the association between the risk factors was studied by calculating the OR. A value of >1 indicates a positive association.
Results
A total of 131 patients were enrolled in this study. As the burns ward in this hospital is only for adult female patients and children (≤18 years) of both sexes, 95 (72.5%) female patients were included, of which 74 (78%) were adults and 21 (22%) were children. Of the 36 (27.5%) male patients enrolled, 33 (91.7%) were from the pediatric age group, and only three were adult males who were admitted to the surgical ward.
A total of 131 patients with burns were included in the present study. From all patients, swabs were collected on days 0, 4, 8, and weekly until discharge or death. In total, 503 swabs were collected from 131 patients at periodic intervals, and 373 culture positives were obtained over the entire study period. Five cultures showed polymicrobial growth and were excluded from analysis, so 368 isolates were considered as infections. The results of the swabs collected are mentioned in [Table/Fig-1].
Of the 131 swabs collected on day 0 from an equal number of patients, 32 showed growth, of which four were contaminants.
The most common organism isolated was Klebsiella pneumoniae (9, 32.14%), followed by Pseudomonas aeruginosa (8, 28.57%) and Acinetobacter baumannii (5, 17.86%). Gram-negative bacteria were the predominant isolates until day eight [Table/Fig-2a]. Samples collected from the end of the second week showed predominant growth of Staphylococcus aureus [Table/Fig-2b].
Microbiological profile in burn wounds.
Organism | Distribution of isolates (n=368) n (%) |
---|
P. aeruginosa | 100 (27.17) |
S. aureus | 84 (22.83) |
A. baumannii | 74 (20.11) |
K. pneumoniae | 65 (17.66) |
Enterobacter species | 15 (4.08) |
Proteus species | 15 (4.08) |
Enterococcus species | 7 (1.90) |
Coagulase negative Staphylococcus species (CoNS) | 5 (1.36) |
E. coli | 3 (0.82) |
*Five samples were contaminants and were removed from analysis
Temporal changes of bacterial profile in burn wounds (Gram negative bacteria).
Wound swabs | 0 day | 4th day | 8th day | 2nd week | 3rd week | 4th week | >5th weeks | Total |
---|
Total no. of samples | 131 | 101 | 99 | 61 | 47 | 29 | 35 | 503 |
No growth | 99 (75.57%) | 18 (17.82%) | 33.03% | 3 (4.92%) | 4 (8.51%) | 1 (3.45%) | 2 (5.71%) | 130 (25.85%) |
Polymicrobial growth | 4 (3.05%) | 0 | 0 | 1 (1.64%) | 0 | 0 | 0 | 5 (0.99%) |
Total isolates | 28 (21.4%) | 83 (82.2%) | 96 (96.9%) | 57 (93.4%) | 43 (91.5%) | 28 (96.6%) | 33 (94.3%) | 368 (73.2%) |
Total gram negative isolates | 24 | 71 | 77 | 39 | 22 | 15 | 24 | 272 |
Non-fermenting Gram negative bacilli |
• A. baumannii | 5 (17.86%) | 23 (27.71%) | 23 (23.96%) | 14 (24.56%) | 6 (13.95%) | 1 (3.57%) | 2 (6.06%) | 74 (20.11%) |
• P. aeruginosa | 8 (28.57%) | 21 (25.30%) | 27 (28.13%) | 16 (28.07%) | 11 (25.58%) | 6 (21.43%) | 11 (33.33%) | 100 (27.17%) |
Enterobacterales |
• K. pneumoniae | 9 (32.14%) | 20 (24.10%) | 20 (20.83%) | 4 (7.02%) | 2 (4.65%) | 2 (7.14%) | 8 (24.24%) | 65 (17.66%) |
• Enterobacter species | 2 (7.14%) | 3 (3.61%) | 4 (4.16%) | 4 (7.02%) | 0 | 2 (7.14%) | 0 | 15 (4.08%) |
• E. coli | 0 | 2 (2.41%) | 1 (1.04%) | 0 | 0 | 0 | 0 | 3 (0.82%) |
• Proteus group | 0 | 2 (2.41%) | 2 (2.08%) | 1 (1.75%) | 3 (6.98%) | 4 (14.29%) | 3 (9.09%) | 15 (4.08%) |
Temporal changes of bacterial profile in burn wounds (Gram positive bacteria).
Wound swabs | 0 day | 4th day | 8th day | 2nd week | 3rd week | 4th week | >5th weeks | Total |
---|
Total no. of samples | 131 | 101 | 99 | 61 | 47 | 29 | 35 | 503 |
Total gram positive isolates | 4 | 12 | 19 | 18 | 21 | 13 | 9 | 96 |
S. aureus (Total) | 1 (3.57%) | 9 (10.84%) | 16 (16.67%) | 17 (29.82%) | 20 (46.51%) | 13 (46.43%) | 8 (24.24%) | 84 (22.83%) |
1. MRSA* | 0 | 5 (55.56%) | 11 (68.75%) | 14 (82.35%) | 17 (85%) | 9 (69.23%) | 4 (50%) | 60 (71.43%) |
2. MSSA† | 1 (100%) | 4 (44.44%) | 5 (31.25%) | 3 (17.65%) | 3 (15%) | 4 (30.77%) | 4 (50%) | 24 (28.57%) |
Coagulase negative Staphylococcus species (CONS) | 2 (7.14%) | 2 (2.41%) | 1 (1.04%) | 0 | 0 | 0 | 0 | 5 (1.36%) |
Enterococcus species | 1 (3.57%) | 1 (1.20%) | 2 (2.08%) | 1 (1.75%) | 1 (2.33%) | 0 | 1 (3.03%) | 7 (1.90%) |
*MRSA: Methicillin resistant Staphylococcus aureus; †MSSA: Methicillin sensitive Staphylococcus aureus
A total of 74 blood culture specimens were collected from 47 patients suspected of having invasive infection based on signs and symptoms of systemic infection clinically and no other identifiable source of infection. Of these, 37 specimens (78.7%) were culture positive. The most common isolate was Klebsiella pneumoniae (16, 43.2%), followed by Pseudomonas aeruginosa and Acinetobacter baumannii (5, 13.5% each). Other bacteria isolated included Enterococcus species, Proteus species, and Enterobacter species (3, 8.1% each), as well as Staphylococcus aureus (2, 5.4%).
Flames were an important cause of burns in 47 patients with invasive infections (OR=4.2, p-value=0.0007) [Table/Fig-3]. A longer hospital stay of >2 weeks was associated with higher invasion rates (OR=3.07, p-value=0.006). Invasive infections were also more associated with a higher degree and depth of burns (burns with >50% TBSA: OR=2.23, p-value=0.074) (deep burns OR=8.25, p-value=<0.00001) [Table/Fig-3].
Association of risk factors with invasive infections and non invasive infections (Based on signs and symptoms of invasive infection clinically).
Patient characteristics | (Invasive infections)n=47 | (Non-invasive infections)n=84 | Chi-square | p-value | Odds Ratio (OR) | 95% CI |
---|
Cause of burns• Flame/blast/Self-infliction (n=76)• Scalding/electric/chemical/miscellaneous (n=55) | 37 (48.7%)10 (18.2%) | 39 (51.3%)45 (81.8%) | 11.613 | 0.0007 | 4.20.234 | 1.88-9.690.10-0.53 |
Days of hospitalisation• <2 weeks (n=88)• >2 weeks (n=43) | 24 (27.3%)23 (53.5%) | 64 (72.7%)20 (46.5%) | 7.527 | 0.006 | 0.323.07 | 0.15-0.691.43-6.56 |
Total body surface area (TBSA) %• <50% (n=97)• >50% (n=34) | 30 (30.9%)17 (50%) | 67 (69.1%)17 (50%) | 3.195 | 0.074 | 0.452.23 | 0.20-0.991.00-4.96 |
Depth of burnsPartial thickness burn (n=68)• Deep partial/Full thickness burn (n=63) | 10 (14.7%)37 (58.7%) | 58 (85.3%)26 (41.3%) | 25.670 | <0.00001 | 0.128.25 | 0.05-0.283.57-19.08 |
Type of referral• Direct walk-in (n=70)• Referred (n=61) | 21 (30%)26 (42.6%) | 49 (70%)35 (57.4%) | 1.742 | 0.187 | 0.581.7 | 0.28-1.180.84-3.56 |
Diabetes (n=8)Non diabetic (n=123) | 5 (62.5%)42 (34.15%) | 3 (37.5%)81 (65.85%) | 1.537 | 0.215 | 3.210.31 | 0.73-14.110.07-1.36 |
Hypertension (n=12)Non hypertensive (n=119) | 6 (50%)41 (34.45%) | 6 (50%)78 (65.54%) | 0.569 | 0.4506 | 1.900.5256 | 0.57-6.270.15-1.73 |
Pregnancy (n=3) | 2 (66.7%) | 1 (33.3%) | 0.266 | 0.6059 | - | - |
Twenty-four patients (18.32%) showed no growth on wound swabs and were considered non infected, whereas 76 patients (58.02%) showed growth only in the wound swab and were considered localised infections/colonisation. Thirty-one patients (23.66%) showed growth in both wound swabs and blood cultures and were classified as having invasive infections.
Mortality was higher among those with invasive infections (n=28; 90.32%) compared to those with localised infections (n=15; 19.74%) (OR=37.9, p-value <0.001) [Table/Fig-4]. The mortality rate was higher in culture-positive sepsis (28, 90.32%) compared to those who were suspected of having an invasive infection but were culture negative (2, 12.5%). This difference was statistically significant (p-value <0.0001) [Table/Fig-4].
Outcome of patients with localised infection/colonisation and invasive infection and its association with mortality.
Type of infection | Death | Discharge | Chi-square test | p-value | Odd’s ratio | 95% CI |
---|
Non infected (n=24) | 3 (12.5%) | 21 (87.5%) | - | - | - | - |
Localised infections (n=76) | 15 (19.74%) | 61 (80.26%) | 42.751 | <0.001 | 0.0263 | 0.01-0.09 |
Invasive infection infections (n=31) | 28 (90.32%) | 3 (9.68%) | | | 37.9 | 10.16-141.78 |
Total (n=131) | 46 (35.11%) | 85 (64.89%) | | | | |
Antibiotic susceptibility testing was performed for routinely used antibiotics according to CLSI standards [10,12,13]. The resistance pattern is mentioned in [Table/Fig-5a,b].
Resistance percentage (Gram negative bacteria to different antibiotics).
Antibiotic agent | Klebsiella pneumoniae (n=65) | Pseudomonas aeruginosa (n=100) | Acinetobacter baumannii (n=74) |
---|
Amikacin | 22 (33.85%) | 37 (37%) | 63 (85.14%) |
Gentamicin | 29 (44.6%) | 43 (43%) | 67 (90.54%) |
Netilmycin | - | - | 61 (82.44%) |
Cefepime | 37 (56.93%) | 37 (37%) | 67 (90.54%) |
Ceftriaxone | 45 (69.23%) | - | 72 (97.30%) |
Ceftazidime | - | 37 (37%) | 67 (90.54%) |
Ciprofloxacin | 39 (60%) | 43 (43%) | 67 (90.54%) |
Piperacillin | - | 10 (10%) | - |
Piperacillin-tazobactam | 20 (30.77%) | 9 (9%) | 66 (89.19%) |
Ampicillin-sulbactam | - | - | 0 |
Trimethoprim-sulfamethoxazole | - | - | 0 |
Aztreonam | - | 19 (19%) | - |
Imipenem | 27 (41.54%) | 38 (38%) | 66 (89.19%) |
Meropenem | 22 (33.85%) | 33 (33%) | 66 (89.19%) |
Colistin | 0 | 0 | 0 |
Not tested
Resistance percentage Gram positive bacteria to different antibiotics.
Antibiotic agent | Resistance in MSSA (n=24) | Resistance in MRSA (n=60) | Resistance in Staphylococcus aureus (MSSA+MRSA) (n=84) |
---|
Gentamicin | 10 (41.7%) | 31 (51.67%) | 41 (48.81%) |
Ciprofloxacin | 22 (91.67%) | 39 (65%) | 61 (72.62%) |
Trimethoprim-sulfamethoxazole | 19 (79.17%) | 51 (85%) | 70 (83.33%) |
Clindamycin | 2 (8.33%) | 23 (38.33%) | 25 (29.76%) |
Erythromycin | 4 (16.67%) | 32 (53.33%) | 36 (42.86%) |
Tetracycline | 2 (8.33%) | 15 (25%) | 17 (20.24%) |
Penicillin | 24 (100%) | 60 (100%) | 84 (100%) |
Linezolid | 0 | 0 | 0 |
Teicoplanin | 0 | 0 | 0 |
Vancomycin | 0 | 0 | 0 |
In the present cohort, Acinetobacter species demonstrated higher resistance to any antibiotic tested as compared to Klebsiella pneumoniae and Pseudomonas aeruginosa. Pseudomonas aeruginosa exhibited the least resistance to piperacillin and aztreonam. Meropenem resistance was lower than that of imipenem in both Klebsiella pneumoniae and Pseudomonas aeruginosa. For the other antimicrobial testing agents, the resistance was greater than 30% and in Acinetobacter species, it was more than 80%.
Discussion
Burn injuries are a common global health problem that is also seen in India [2]. Burn wounds not only lower immunity but also provide a good nidus for bacteria to grow as these are exposed surfaces [10,16]. It is important to distinguish burn wound colonisation from infection [1,17]. Unnecessary usage of antibiotics in burn wounds that are colonised and not infected may cause more harm by affecting patient’s protective normal skin flora thus exposing the wound to opportunistic as well as multidrug resistant pathogens [1,17].
In the present study on the temporal association of bacteria in burn wound infections, it was found that gram-negative bacteria predominated in the first week, whereas the association with Staphylococcus aureus increased from the third week onwards. Additionally, the bacterial species present in the same patient would differ over time. This observation contrasts with a few other studies where S. aureus was identified as the most common pathogen in early burn wound infections [1,5,18,19]. Typically, within the first 48 hours after a burn, gram-positive bacteria, such as Staphylococcus species present in the surrounding skin, heavily colonise the wound surface [20,21]. After 5-7 days, these wounds become colonised by other microbes, including gram-negative bacteria and yeasts, which are primarily derived from the patient’s own gastrointestinal tract and upper respiratory tract or from the hospital environment, or acquired from healthcare personnel [1,21]. Among the gram-negative bacteria, Pseudomonas aeruginosa was found to be the most common (100, 27.17%), followed by Staphylococcus aureus (84, 22.83%), Acinetobacter baumannii (74, 20.11%), and Klebsiella pneumoniae (65, 17.66%). Most cultures were negative on day zero; however, as hospital stays increased, the wounds became colonised with various opportunistic organisms present endogenously or in the environment [1]. In the present study, colonisation of burn wounds was detected as early as within 48 hours. This study also aimed to determine the bacterial spectrum of invasive burn wound infections. Of the 47 patients suspected of having invasive infections based on clinical signs and symptoms of systemic infection, 37 (78.7%) were culture positive. The most common isolate was Klebsiella pneumoniae (16, 43.2%).
Among the factors known to be associated with a higher risk of invasive infection, flame burns (OR=4.2), a length of hospitalisation greater than 2 weeks (OR=3.0), involvement of a larger body surface area (OR=2.23), and deep burns (OR=8.25) were associated with a higher risk. Similar findings have been reported in other studies [9,22,23]. Invasive infections were significantly associated with deep burns compared to superficial burns (OR=0.12). Deep burns expose blood vessels in wounds directly to the environment, facilitating the entry of organisms into the bloodstream. These findings are consistent with those from previous studies [16,22,24,25].
In the present study, no significant association was found between patients referred from outside and direct patients. In contrast, Dhopte A et al., reported higher mortality rates in outpatients compared to those who came in as direct walk-ins. Early referral to a tertiary care centre helps prevent microbial colonisation by maintaining wound sterility, ensuring early wound coverage, and providing adequate nutritional support [23].
Co-morbid conditions known to increase the risk of infection in burn wound patients include diabetes, hypertension and pregnancy [7,8]. In this study, diabetes and pregnancy were identified as associated factors for invasive infection. Hypertension was also evaluated as a possible risk factor, but the results were not found to be significant. Diabetes is a neurovascular disorder that increases the likelihood of developing any injury, including burns, due to decreased sensitivity [26].
In patients with diabetes, delayed wound healing may contribute to an increased risk of infection. However, in this study, only 8 (6.11%) burn patients had diabetes mellitus, and no significant association with invasive infections was found. Three (2.26%) patients were pregnant, but no significant association was found between invasive infections and pregnancy.
The antibiotic susceptibility pattern demonstrated high rates of resistance to most of the antibiotics tested for gram-negative bacteria. These results align with the trends of resistance observed in most pathogens from India [2,4,24,27]. Sixty (71.4%) isolates of S. aureus were methicillin-resistant. An interesting observation in this study was that drug resistance was more common in strains isolated after the first week of admission, which may be attributed to selection pressure resulting from prolonged broad-spectrum antibiotic use. For Staphylococcus aureus, ciprofloxacin, co-trimoxazole, erythromycin and gentamycin were found to be effective. Sharma S et al., reported similar findings [4].
Limitation(s)
Although tissue biopsy along with blood culture is a better predictor of invasive burn wound infections, it could not be performed in these patients to differentiate true pathogens from mere colonisers. Since very few cases demonstrated involvement of the urinary and respiratory systems, the data were insignificant for analysis.
Conclusion(s)
As seen in this study, the most common organisms leading to wound infections in burn patients were Pseudomonas aeruginosa, Staphylococcus aureus, Acinetobacter baumannii and Klebsiella pneumoniae, in that order, with gram-negative organisms predominating in the first week and Staphylococcus aureus in the second week. There was a changing trend in the bacterial profile of burn wound infections over time, with gram-negative bacteria predominating in the first week and Staphylococcus aureus in the second week. This study highlights the need for periodic monitoring of the bacterial profile in burn wounds, along with their antimicrobial susceptibility patterns over time, thus ensuring early and appropriate diagnosis and treatment in patients, thereby improving outcomes.
*Five samples were contaminants and were removed from analysis*MRSA: Methicillin resistant Staphylococcus aureus; †MSSA: Methicillin sensitive Staphylococcus aureusNot tested