Multiple bacterial species reside in chronic wounds: a longitudinal study

Abstract

The aim of the study was to investigate the bacterial profile of chronic venous leg ulcers and the importance of the profile to ulcer development. Patients with persisting venous leg ulcers were included and followed for 8 weeks. Every second week, ulcer samples were collected and the bacterial species present were identified. More than one bacterial species were detected in all the ulcers. The most common bacteria found were Staphylococcus aureus (found in 93·5% of the ulcers), Enterococcus faecalis (71·7%), Pseudomonas aeruginosa (52·2%), coagulase‐negative staphylococci (45·7%), Proteus species (41·3%) and anaerobic bacteria (39·1%). Resident bacterial species were present in all the ulcers. In 76% of the ulcers, two or more (up to five) resident bacterial species were found. The most common resident bacterial species were S. aureus and P. aeruginosa. Furthermore, ulcers with P. aeruginosa were found to be significantly larger than ulcers without the presence of P. aeruginosa (P < 0·005). Our study demonstrated that the chronic wound is colonised by multiple bacterial species and that once they are established many of them persist in the wound. Our results suggest that the presence of P. aeruginosa in venous leg ulcers can induce ulcer enlargement and/or cause delayed healing.

Introduction

Chronic wounds represent a major but unfortunately neglected health care problem, resulting in distress and disability for the patients, potential loss of working capability, reduced quality of life and an increasing burden to health care providers 1, 2. It is estimated that around 50 000 persons in Denmark (approximately 1% of the population) suffer from wounds that need professional treatment (3).

The normal wound healing process is a complex physiological event which depends on interactions between epidermal keratinocytes, dermal fibroblasts, langerhans cells, endothelial cells and fibroblasts and which is coordinated via complex cell/cell and cell/matrix interactions. These responses are altered in chronic wounds with prolonged inflammation, a defective wound matrix, and failure of reepithelialisation (1). There is no single unifying theory as to the aetiology of these wounds and it is expected to be multifactorial. The chronic wounds can be conceived as normal wounds arrested in the middle of the healing process. It is likely that microorganisms in the wounds contribute to a non healing phenotype, some species of bacteria or combinations of such are likely to be more detrimental to wound healing than others. Furthermore, chronic wounds often occur in individuals and tissues that are at increased risk of bacterial invasion due to poor vascular supply and systemic factors (4). It is possible that the cure of a non healing, infected wound is to restore the beneficial equilibrium between bacteria and the surrounding tissue. This could be done by controlling the infection with antimicrobial agents or by addition of growth factors as suggested by Hayward and coworkers (5), who reported that basic fibroblast growth factor can overcome the defect in healing created by bacterial infection.

The aim of the present study was to investigate the bacterial profile of chronic venous leg ulcers and the importance of the bacterial profile — or changes of the bacterial profile — to ulcer development. Chronic venous leg ulcers were followed over time, and the bacterial species present were identified. Multiple bacterial species were found to colonise the chronic leg ulcers, and resident bacterial flora was found to be present in the ulcers.

Materials and methods

Patients

Fifty consecutive patients admitted to the Copenhagen Wound Healing Center, Bispebjerg Hospital, Denmark were included in the study. The project was approved by the Danish National Committee for Biomedical Research Ethics, Ministry of the Interior and Health, and written consents were obtained from patients upon information. The patients had persisting venous leg ulcers for more than 3 months prior to inclusion in the study; venous aetiology was proven on Duplex scanning. All patients included in the study had an ankle brachial pressure index above 0·6. Pregnant or breastfeeding women and patients with diabetes were not included in the study. This was to have a homogenous group of patients with no other diseases that could influence the chronic wound (except venous insufficiency). None of the patients had had any antibiotic treatment within 14 days prior to inclusion and did not receive any antibiotic treatment during the study. All patients in the study were treated with standard compression therapy. At the Copenhagen Wound Healing Center, antimicrobial treatment of chronic wounds is generally only used if the infection has spread to surrounding healthy tissue and/or if the patient’s general condition is affected (6).

Four of the 50 patients were excluded from the study, three of these patients’ ulcers were practically healed upon inclusion and the last patient did not wish to participate after all. Initial ulcer size and the length of the ulcer history were registered. Patients were followed for 8 weeks or until the ulcer healed. Every 2 weeks, the ulcer size was measured by tracing of the ulcer on a transparent foil sheet, and the condition of the ulcer was registered.

Sample collection from chronic venous leg ulcers

A sterile 10‐mm² filter paper pad was placed in the ulcer until it was saturated with exudates. The pad was then transferred to 2 ml of sterile isotonic sodium chloride solution. The ulcer surface was swabbed with a sterile charcoal swab [Statens Serum Institut (SSI) 40085, Denmark]. The swab was immediately placed in SSI Stuart transport medium (SSI 28733, Denmark). A 4‐mm punch biopsy was taken from the centre of the ulcer. The biopsy was weighed and immediately placed in SSI Stuart transport medium. The samples were transported to SSI and processed the same day as collected. Filter paper pad samples were taken from the ulcers at inclusion and every 2 weeks. A charcoal swab was taken at inclusion (week 0). Punch biopsy samples were taken at inclusion and after 4 weeks.

Culture analyses of the ulcer microflora

The processing of the samples was performed according to SSI standard procedures. Media and agar plates were from SSI, Denmark; detailed description of the medias can be obtained at www.ssi.dk.

The filter paper pad was vortexed for 15 seconds. One, 10 and 100 µl were plated on 5% blood agar plates (SSI 677, Denmark) in order to obtain single colonies, and 100 µl was plated on a blue agar plate selective for gram‐negative bacteria species [Modified Conradi‐Drigalski plates (7), SSI 694, Denmark]. The biopsy samples were bisected with a sterile scalpel; one part was used for cultural analysis and the second stored in 10% glycerol at−80°C. The tissue sample and the swabs were spread on a 5% blood agar plate (SSI 677, Denmark), a chocolate plate (SSI 700, Denmark), a blue agar plate (SSI 694, Denmark), a chocolate agar plate supplemented with cysteine and K‐vitamin for anaerobic cultures (with kanamycin and metronidasol neoSensitabs, Rosco, Taastrup, Denmark) and a susceptibility blood agar plate (SSI 723, Denmark) with the following neoSensitabs: methicillin, penicillin, gentamysin, mecillinam, erytromycin and polymyxin. The susceptibility plates were not used to obtain antibiotic susceptibility data but were used to facilitate isolation of representative colonies by suppressing growth of interfering bacterial species present. All plates, except for the anaerobe ones, were incubated under aerobic conditions at 37°C. Anaerobe plates were incubated under anaerobic conditions at 37°C. Primary isolation plates were examined after 24 and 48 h. Anaerobe plates were examined after 24 and 48 h and 4 days. All representative colonies on the primary isolation plates were isolated and identified by standard microbiological methods, and where appropriate commercial identification kits were used to assist identification (VITEK GNI+, ID 32 Kits, BioMerieux, Marcy L’Etoile, France). In this way, it was possible to identify the isolates to genus or species level.

Ribotyping

Cultures of Pseudomonas aeruginosa and Staphylococcus aureus isolates were processed by the RiboPrinter^® Microbial Characterization System (DuPont Qualicon, Wilmington, DE, USA) according to the manufacturer’s instructions. The restriction enzymes used were PvuII for the P. aeruginosa isolates and EcoRI for the S. aureus isolates. The results were analysed with the program included in the Qualicon RiboPrinter system (8).

Statistical analyses

We defined a bacterial species to be resident for a specific patient if the bacterial species was observed at all or all but one of the observation times for the patient in question. Differences in mean ulcer size depending on whether a given bacterial species was resident or not were evaluated. In order to make the normal approximation appropriate, the statistical analysis was based on the log‐transformed ulcer sizes. As multiple measurements were taken on the same patient during the study period, the observed ulcer sizes are serially correlated within each patient. To account for this, an autoregressive correlation structure of order 1 was applied when comparing log‐transformed ulcer sizes for various subgroups of the patients. Interactions between resident bacteria were not analysed due to the limited number of patients.

A linear regression of the log‐transformed ulcer sizes with the age of the wound as the explanatory variable was performed. A Wilcoxon–Mann–Whitney test was performed in order to evaluate the difference in the median age of the wound depending on the presence of any of the resident bacterial species in the wound. The analysis was performed on one bacterial species at a time.

The statistical analyses were performed using sas version 8·2 for Windows.

Results

Patients and ulcers information

The patients in the study consisted of 29 females and 17 males; their age ranging from 29 to 91 years (median 78 years). The patients had one ulcer each. On inclusion, the median ulcer size was 17·0 cm² (range 3·7–600 cm²) and a median duration of 21 months (range 3–300 months). During the study, three of the 46 ulcers healed and two ulcers healed more than 95%. Three of the chronic leg ulcers increased in size with more than 10% between sampling times, whereas 10 of the ulcers were reduced in size with more than 10% between sampling times. The remaining ulcers had a relatively stable ulcer size during the 8 weeks.

Culture analyses

A total of 342 samples were taken from the 46 chronic leg ulcers leading to 1236 bacteria isolates. A total of 37 bacterial species were identified from the ulcer samples. More than one bacterial species was detected in close to all the samples (94·4%). None of the patients’ chronic ulcers was colonised by a single bacterial species. Between four and six bacterial species were detected in 50% of the ulcers. Thirtynine per cent of the ulcers were colonised with more than six species (Figure 1). The mean number of bacterial species isolated per chronic wound was 6·3. Due to difficulties in isolating pure cultures of the anaerobic bacteria, not all of the anaerobic bacteria were identified at species or genus level but only as being anaerobic. As the results showed no correlation between anaerobic bacteria as a group and ulcer size (Table 2), no further attempt to identify the anaerobic bacteria was performed at this point.

Figure 1.

Number of bacterial species detected in the patients’ ulcers during the study period.

The most common bacterial species detected in the chronic ulcers was S. aureus, which was isolated at least once during the study period in 93·5% of the ulcers. Other common species found were Enterococcus faecalis (71·1%), P. aeruginosa (52·2%), anaerobic bacteria (39·1%) and coagulase‐negative staphylococci (e.g. Staphylococcus epidermidis) (45·7%). Proteus species were detected in 41·3% of the ulcers and β‐haemolytic streptococci (groups A and G) in 24·7% of the ulcers. Table 1 summarises the bacterial species which were found in two or more of the patients.

Table 1.

Bacterial species detected in two or more of the patients’ ulcers

Bacterial species	Number of ulcers	Frequency (%)
Staphylococcus aureus	43	93·5
Enterococcus faecalis	33	71·7
Pseudomonas aeruginosa	24	52·2
Coagulase‐negative staphylococci	21	45·7
Anaerobic bacteria	18	39·1
Enterobacter cloacae	17	37·0
Escherichia coli	15	32·6
Corynebacteria	14	30·4
Proteus mirabilis	12	26·1
Proteus vulgaris	12	26·1
Morganella morganii	10	21·7
Non haemolytic streptococci	8	17·4
Haemolytic streptococci group G	7	15·2
Acinetobacter	6	13·0
Klebsiella oxytoca	6	13·0
Pseudomonas maltophilia	6	13·0
Escherichia hermanii	3	6·5
Haemolytic streptococci group A	3	6·5
Klebsiella pneumoniae	3	6·5
Aeromonas hydrophila	2	4·3
Bacillus species	2	4·3
Citrobacter freundii	2	4·3
Leclericia adecarboxylata	2	4·3
Pseudomonas mendocina	2	4·3
Pseudomonas stutzeri	2	4·3
Other Pseudomonas	2	4·3

Resident bacteria in the ulcers

A comparison of the microbial findings from the same ulcer showed that one or more of the same species were re‐isolated from all the samples from 41 of 46 ulcers. In the remaining five ulcers, one or more of the same species was present in all but one of the samples. In this study, resident flora is characterised as a bacterial species present in all or all but one samples of a series of samples. According to this definition, resident bacterial species were found in all 46 ulcers.

In 35 of the 46 ulcers (76%), two or more (up to five) resident bacterial species were found.

The most common resident bacterial species were S. aureus (re‐isolated in 32 ulcers) and P. aeruginosa (re‐isolated in 17 ulcers). Other frequently occurring resident bacterial species were E. faecalis (re‐isolated in 12 ulcers), Proteus vulgaris (re‐isolated in eight ulcers), Proteus mirabilis (re‐isolated in seven ulcers) and Escherichia coli (re‐isolated in seven ulcers). RiboPrint analyses were performed on S. aureus and P. aeruginosa isolates; the results revealed that bacteria isolated from the same patient belonged to the same ribogroup (data not shown). Generally, none of the strains isolated from different patients belonged to the same ribogroup indicating that there was no common source of P. aeruginosa in the patient population. P. aeruginosa was found to be present as a resident bacterium in all the wounds that increased in size (three wounds) but was not resident in any of the wounds that were reduced in size (10 wounds).

Statistical analyses

The study was completed by 31 patients, while eight patients had one missing examination and seven patients attended only three of the five planned examinations. The 15 patients who did not complete the study were divided into the following groups: four patients cancelled their participation, two were hospitalised, one healed before the fourth examination, two healed before the fifth examination and six patients did not attend all the examinations for unreported reasons. The statistical analysis of the ulcer size is based on patients having at most one missing examination (39 patients). Furthermore, patients reported to have healed and one patient with an extreme ulcer size (600 cm²) were excluded. These considerations imply that the statistical analyses were based on observations from 36 patients.

Differences in the log‐transformed ulcer size depending on whether a bacterial species was resident or not were analysed for the six most frequent resident bacteria species (S. aureus, P. aeruginosa, E. faecalis, P. vulgaris, P. mirabilis and E. coli). The results summarised in Table 2 are given on the scale of measurement. For the two most commonly observed bacteria, we obtained significant differences in ulcer size depending on whether a specific bacterial species was resident or not. Ulcers with resident S. aureus were significantly smaller (P = 0·0034) than ulcers with non resident S. aureus. Ulcers with resident P. aeruginosa were significantly larger (P = 0·0014) than ulcers with non resident P. aeruginosa.

Table 2.

Difference in mean ulcer size according to whether a given bacterial species was resident or not

Bacteria	Resident *(no of ulcers)	Ulcer size Mean [95% CI] (cm2)†	P value‡
Staphylococcus aureus	Yes (25)	11·71 [7·74; 17·72]	<0·005
	No (11)	37·33 [20·02; 69·61]
Pseudomonas aeruginosa	Yes (13)	35·89 [20·49; 62·88]	<0·005
	No (23)	10·83 [7·09; 16·52]
Enterococcus faecalis	Yes (9)	15·02 [6·85; 32·93]	>0·05
	No (27)	17·28 [10·98; 27·18]
Proteus vulgaris	Yes (6)	20·56 [7·89; 53·59]	>0·05
	No (30)	16·00 [10·42; 24·57]
Proteus mirabilis	Yes (6)	11·91 [4·59; 30·86]	>0·05
	No (30)	17·85 [11·66; 27·35]
Escherichia coli	Yes (6)	15·49 [5·93; 40·51]	>0·05
	No (30)	16·93 [11·01; 26·04]
Anaerobic bacteria§	Yes (14)	26·88 [16·38; 44·14]	>0·05
	No (30)	16·86 [12·23; 23·24]

^* A bacterial species was resident for a specific patient if the bacterial species was observed at all or at all but one of the observation times for the patient in question.

^† Analyses were based on log‐transformed observations for 36 patients. The reported results are transformed back to the scale of measurement.

^‡ The P value for testing for difference in mean ulcer size between patients grouped according to whether the given bacteria species was resident or not.

^§ Analyses based on inclusion time (week 0), two patients were excluded due to extreme ulcer size. These two patients were excluded from all the statistical analyses.

Statistical analyses on the differences in log‐transferred ulcer size depending on whether anaerobic bacteria were present or not were based on data from inclusion (week 0). Excluded from the study were two patients due to extreme ulcer size; these two patients were also excluded from the above‐mentioned statistical analyses. No significant difference in ulcer size was observed whether anaerobic bacteria were present in the ulcers or not.

No difference was detected in the age of the wound depending on the presence of any of the resident bacterial species in the wound, and wound size was not found to depend on wound age.

Discussion

Bacteria were isolated from all the venous leg ulcers in this study. This was in agreement with previous studies 1, 9, 10, 11, 12. None of the chronic ulcers in this study was colonised with a single bacterial species, and the average number of bacterial species isolated from the ulcers was 6·3 species even though the anaerobic bacteria were regarded as one species. This was higher than the average found in other studies, where the number of bacterial species per ulcer was around three (when the anaerobic species were regarded as one species) 9, 11, 13. The reason for the higher average could be that the patient group of this study for some reason has more bacterial species colonising their ulcers; it could also be due to a more critical primary isolation by colony morphology of aerobic bacteria or to the fact that multiple samples and sampling techniques were used on the same ulcer at the same point of time.

The most common bacterial species detected in the chronic leg ulcers was S. aureus, which was isolated at least once during the study period in 93·5% of the patients. This frequency was considerably higher than that found in other studies, where S. aureus was isolated from 40 to 60% of the ulcers 11, 12, 14, but in agreement with results found by Madsen et al. (15).

P. aeruginosa was detected in 52·2% of the ulcers; most other studies found P. aeruginosa in 20–30% of the ulcers 1, 10, 14, 15, 16. Molecular analyses of the P. aeruginosa isolates indicated that there was no common source of P. aeruginosa in the patient population, as strains from different patients generally belonged to different ribogroups.

According to the definition of a resident bacterium as a bacterial species present in all or all but one sample of a series, resident bacteria were found in all the ulcers. In 76% of the ulcers, two or more (up to five) resident bacterial species were found. The second most common resident bacterial species was P. aeruginosa; the species was resident in 75% of the ulcers colonised with P. aeruginosa. Molecular analyses of the P. aeruginosa isolated from the same ulcer showed that it was the same strain that colonised the ulcer during the 8‐week period. These results indicate that once P. aeruginosa colonises a chronic ulcer, it continues to be present in the ulcer. Three of the chronic venous leg ulcers increased in size with more than 10% between sampling times, while 10 of the ulcers were reduced in size with more than 10% between sampling times. P. aeruginosa was present as a resident bacteria in all the ulcers that increased in size but non resident in all of the ulcers that were reduced in size. Due to the small number of wounds (3) that increased in size during the study and the diversity in the microbiological findings, it was not possible to perform any statistical analyses on the association between the presence of bacterial species and ulcer development. Therefore, statistical analyses were performed on the ulcer size depending on whether a bacterial species was resident or not. The statistical analyses revealed that ulcers with resident P. aeruginosa were significantly larger than ulcers without the presence of P. aeruginosa (P = 0·001). This result is in agreement with results described by Madsen et al. (15). The larger ulcer size observed in the presence of P. aeruginosa might be because P. aeruginosa induces ulcer enlargement and/or cause delayed healing. P. aeruginosa produces a large number of virulence factors including tissue‐destroying enzymes (17). Studies have shown that P. aeruginosa production of a metalloproteinase induces degradation of wound fluid and human skin proteins and inhibited fibroblast growth (18). P. aeruginosa can alter matrix metalloproteinase production and activation by epithelial cells (1); matrix metalloproteinases are directly implicated in cellular migration and wound‐healing responses. This indicates that P. aeruginosa is capable of altering the wound‐healing response and causes enlargement or delayed healing of chronic wounds.

Another explanation could be that P. aeruginosa primarily would colonise older wounds which are often of larger size. However, statistical analyses of data in this study showed no difference in the age of the wound depending on the presence of P. aeruginosa in the wound, and wound size was not found to depend on wound age. Hence, this hypothesis is not supported by results in this study.

Our study demonstrated that the chronic wound is colonised by multiple bacterial species (up to 14 species) and that once they are established, many of them persist in the wound. Our results suggest that the presence of P. aeruginosa in venous leg ulcers can induce ulcer enlargement and/or cause delayed healing.

Acknowledgements

The study was partly financed by the Danish Research Agency under the SUE‐programme (No. 9901187), a collaboration between Coloplast Research A/S, the University of Copenhagen, Bispebjerg University Hospital and Statens Serum Institut. We thank student Michael K. Sonnested for the ribotyping analyses and project nurses Hanne Vognsen and Lone Haase, Bispebjerg University Hospital for their valuable assistance.