- Email: [email protected]
Patterns of wound colonisation in patients with peripheral vascular disease
Journal of Infection (1987) 15, 21-26
P a t t e r n s o f w o u n d c o l o n i s a t i o n in patients w i t h p e r i p h e r a l v a s c u l a r disease J. S. Chesham and D. J. Platt University Department of Bacteriology, Royal Infirmary, Castle Street, Glasgow G4 oSF Accepted for publication 18 February 1987 Sutnnaary The bacterial colonisation of ulcers related to vascular disease, amputation sites and sites of surgical vascular repair was studied in 122 patients with vascular disorders. Wounds were graded clinically on the basis of inflammation, oedema and exudate. They were sampled at weekly intervals in order to determine the duratio_n of colonisation. The significance of potential pathogens was assessed. Vascular ulcers were of low mean wound grade (I'87), short duration (I'58 weeks) and potential pathogens were isolated from 82 % of 38 wounds. Amputation sites were of a higher mean wound grade (2"48) and longer duration (2"52 weeks) when compared to vascular ulcers. Potential pathogens were isolated from 69 % of 42 amputation sites. Sites of surgical vascular repair were similar to amputation sites with respect to mean wound grade (2"33) but of a mean duration (1"69 weeks) similar to that of vascular ulcers. Potential pathogens were isolated from 50 % of 42 surgical vascular repair sites but they were not evenly distributed among limb (33 %), groin (55 %) and abdominal wounds (71%). Coliforms were isolated from similar numbers of all wounds. Pseudomonas spp., Staphylococcus aureus and Streptococcus faecalis were evenly distributed with respect to the site of isolation. Pseudomonas spp. and S. aureus were associated more commonly with vascular ulcers and S. faecalis with sites of surgical vascular repair. The use of topical antiseptics and/or systemic antimicrobial agents had a minimal effect in reducing colonisation/infection in the wounds studied.
Introduction Assessment of the clinical significance of bacterial isolates f r o m w o u n d s is generally an area of uncertainty. It often relies on subjective distinction between ' c o l o n i s a t i o n ' and ' i n f e c t i o n ' unless the organism is a well recognised p a t h o g e n (e.g. Streptococcus pyogenes). Quantitative studies have suggested that if the n u m b e r s of bacteria isolated from a w o u n d are ~< IO~ colony f o r m i n g units per gram of tissue or per square centimetre of w o u n d surface t h e n healing will be unimpaired. T h i s approach has been applied to the rejection of skin-grafts 1, the healings of leg ulcers ~ and reclosure of surgical wounds, s' 4 Staphylococcus aureus has been reported as the m o s t c o m m o n l y isolated organism from both leg ulcers 5 and surgical wounds. 6 T h e range of other bacteria isolated was wide and m i x e d cultures were common. 5-8 Quantitative bacteriology in the investigation of w o u n d infection is rarely practicable in routine diagnostic laboratories. T h e distinction between inoi63-4453/87/o4oo2I + 06 $02.00/0
© I987 The British Society for the Study of Infection
22
J.s. CHESHAM
AND D. J. P L A T T
fection and colonisation is usually based on the subjective interpretation of combined clinical and microbiological data. T h e aim of this study was to compare ulcers related to vascular disease with surgical wounds, in respect of patterns of bacterial colonisation/infection, among patients in the same surgical unit. Correlation of these findings with differences in severity of the initial wounds, their rate of healing, as well as systemic and topical antimicrobial therapy was considered also. Patients
Materials and methods
W o u n d swabs were obtained from in-patients in the peripheral vascular disease unit of Glasgow Royal Infirmary between February I98I and June I983. Clinical data including the use of topical antiseptics and systemic antimicrobial therapy were recorded. W o u n d s were classified according to their appearance as follows : Grade I - moist, granulating, no evidence of infection, Grade II - inflammation present with or without oedema, no purulent exudate, Grade I I I - inflammation, oedema and purulent exudate present. Clean dry wounds were not sampled. W o u n d s were further divided clinically into vascular ulcers, amputation sites and sites of surgical vascular repair. Specimens were collected weekly and full details of the severity of each wound were recorded on each occasion. Most wounds were seen to heal with very few patients either dying or being transferred elsewhere without resolution of the wound. T h e n u m b e r of weeks that a given site was sampled provided an estimate of the relative rate of healing of the wounds studied. Specimens Material was collected by means of swabs and processed within 2 h of collection. Swabs were cultured on (I) cysteine-lactose-electrolyte-deficient ( C L E D ) agar (Mast D M I I O ) ( 2 ) Columbia agar base (Oxoid, C M 3 3 I ) containing 5 % defibrinated horse blood and (3) neomycin blood agar (Brucella agar base (Difco, o96ooi), which contained 7o m g / i neomycin (Sigma), 5 % defibrinated horse blood and o'1% (v/v) vitamin K haemin solution (Gibco)). C L E D and blood agar plates were incubated aerobically in an atmosphere of 5 % carbon dioxide at 37 °C overnight. Neomycin blood plates were incubated at 37 °C anaerobically for 48 h. Potential pathogens were purified from primary plates and identified by standard methods. 9 In addition, coliforms were further identified by the A P I 2oE system. Results Swabs were obtained from the wounds of 122 patients when the clinical appearance suggested delayed healing a n d / o r infection. T h e mean times for healing were I'58 weeks (range I - 4 weeks) for vascular ulcers, 2"52 weeks (range i-8 weeks) for amputation sites and I'69 weeks (range I-5 weeks) for sites of surgical vascular repair. These findings indicate that sites of surgical vascular repair take longer to heal than 'vascular ulcers' but that both of these types of w o u n d heal more rapidly than wounds at amputation sites.
Bacterial colonisation of wounds
23
Table I The classification of wounds in patients with peripheral vascular
disease ' Vascular ulcers' W o u n d grade
Amputation sites
Sites of surgical vascular repair
No.
(%)
No.
(%)
No.
(%)
9" 25 4
(24) (66) (IO)
3 i6 23
(7) (38) (55)
2 24 16
(5) (57) (38 )
I II III Mean
1-87
2"48
2"33
Table II The distribution of potential pathogens among the different types of
wounds studied 'Vascular ulcers' n = 38
Coliforms Pseudomonas spp. Staphylococcus aureus Streptococcus faecalis Beta-haemolytic streptococci Anaerobes Any potential pathogen isolated Potential pathogen not isolated
Amputation sites n = 42
Sites of surgical vascular repair n = 42
No.
(%)
No.
(%)
No.
(%)
I9 IO I5 3 I 2 3I
(5 ° ) (26) (39) (8)
(43) (2r) (2I) (I4)
I9 3 9 8
(45) (7) (2I) (I9)
(82)
r8 9 9 6 I 2 29
(69)
i 2I
(50)
7
(I8)
I3
(3I)
2I
(5o)
T h e classification of wounds into grades I, II and I I I according to severity and the percentage in each grade is shown in Table I. 'Vascular ulcers' were of the lowest grade with a mean value of I ' 8 7 . A m p u t a t i o n sites and sites of surgical vascular repair had mean grades of 2"48 and 2"33 respectively. Although ' vascular ulcers' were of short duration, having the lowest mean grade of wound, potential pathogens were isolated from 82% (Table II). W o u n d s at amputation sites were of longer duration and higher mean grade; 69 % yielded potential pathogens. W o u n d s at sites of surgical vascular repair were similar to 'vascular ulcers' in duration but their mean grade was close to that of wounds at amputation sites. Half the wounds at sites of surgical vascular repair did not yield any potential pathogen. Table II also shows the categories of organisms found in the wounds samples. Escherichia coli, Proteus spp., Enterobacter spp., Klebsiella spp., Citrobacter spp., Serratia spp., and Hafnia alvei were not separately distinguished and have been included under the general heading of coliforms. W h e n isolated
24
J.s. CHESHAM AND D. J. PLATT
Table I I I Potential pathogens isolated in the presence and absence of topical, systemic or a combination of both types of antimicrobial therapy Potential pathogens isolated
Topical and systemic antimicrobial therapy Topical therapy only Systemic therapy only No topical or systemic antimicrobial therapy
Potential pathogens not isolated
Total no. of wounds
No.
(%)
No.
(%)
60
48
(80)
I2
(20)
(55)
19 8
5
3
43 I4
24 6
2
(44)
from a given specimen, coliforms appeared to be evenly distributed among the three types of wound. Pseudomonas spp. were isolated more commonly from patients with 'vascular ulcers' and wounds at amputation sites than from patients with wounds at sites of surgical vascular repair. S.aureus was most often isolated from 'vascular ulcers' and S.faecalis from wounds at sites of surgical vascular repair. Beta-haemolytic streptococci and anaerobes were seldom isolated. T h e wounds of 65 (53 %) of the patients studied were being treated topically with either sodium hypochlorite (Milton), hydrogen peroxide, chlorhexidine gluconate (Hibitane), povidone-iodine (Betadine) or a combination of these antimicrobial agents (Table III). In addition, many different systemic antimicrobial agents were prescribed during the study. T h e s e included flucloxacillin, cephalosporins, metronidazole, penicillins V and G, ampiciUin, cotrimoxazole, tetracycline, gentamicin, augmentin, erythromycin and fucidin. At least one of these systemic antimicrobial agents was prescribed for lO3 patients, often for unrelated problems such as respiratory or urinary tract infections. T h e r e was no significant difference in the incidence of isolation of potential pathogens from wounds in patients treated with neither topical nor systemic antimicrobial agents, topical antimicrobial agents alone or systemic agents alone. T h e n u m b e r of patients treated with both topical a n d systemic antimicrobial agents was higher than those treated with none or either topical or systemic therapy alone. Potential pathogens were isolated from significantly more wounds when both types of therapy were administered. Only flucloxacillin, cephalosporins and metronidazole were prescribed for a sufficient n u m b e r of patients to determine any effect against organisms normally considered to be susceptible. Metronidazole was prescribed in 43 patients. Anaerobic organisms were isolated from the wounds of five patients, four of w h o m had been receiving metronidazole. Flucloxacillin was prescribed for 49 patients and S.aureus was isolated from the wounds of 12 (25 %) of these. Of 73 patients who did not receive flucloxacillin therapy 21 (29 %) had wounds which contained S.aureus. Similarly, of the wounds of 69 patients receiving cephalosporins 27 (39 %) contained coliforms whereas those
Bacterial colonisation of wounds
25
of 22 (42 %) of 53 patients not receiving cephalosporins yielded coliforms. Thus, organisms against which specific antimicrobial therapy was directed were not isolated more often from patients who did not receive those agents. This indicates that antimicrobial therapy did not exert a detectable effect on the bacterial flora of the wounds. Discussion
Factors which contribute to the colonisation or infection of wounds are numerous and include underlying pathology, type and site of the wound as well as the number and types of organisms isolated. Daltrey and colleagues 8 classified a wound as infected if there was a persistent presence of Io 4 microorganisms per square centimetre. Such criteria, however, pose a number of problems. For example, the authors made no attempt to define 'persistence'. Moreover, these results showed that there were notable variations in the ability and numbers of organisms required to impair wound healing. Robson and colleagues 4 have stated that, in view of the virulence of beta-haemolytic streptococci, no attempt was made to close a wound if they were present, regardless of their numbers. Although at least one species of potential pathogen was isolated from 82 % of 'vascular ulcers' the wounds were on average of short duration and low grade. This suggests bacterial colonisation rather than active infection. These results are consistent with the patients selected and the aims of surgery in that patients admitted for treatment would have been pre-selected such that expectation of success of treatment was high and that surgery was designed to improve the blood supply. These factors, together with that of good wound management, should lead to healing of the 'vascular ulcer'. Wounds at amputation sites were of longer duration and of a more severe grade compared to that of 'vascular ulcers'. Amputation sites were not a homogeneous group and although the amputation was performed through healthy tissue, circulation distal to the site was presumably poor and may explain the greater probability of colonisation/infection of these wounds by potential pathogens. T h e i r longer duration and higher grade of wound may be due in part to the severity of trauma associated with amputation. This together with the high incidence of a potential pathogens being isolated (69 %) suggests infection rather than colonisation. Sites of surgical vascular repair formed an intermediate group. Although they resembled 'vascular ulcers' in duration and amputation sites in severity, potential pathogens were isolated from a smaller proportion of sites (5o %) than from either of the former groups. Since little difference was seen in the duration and severity of these wounds, irrespective of whether potential pathogens were isolated, it seems likely that colonisation rather than infection was involved. Even so, the possibility that deep-seated infection (with failure to isolate the causative organism from surface samples) cannot be excluded. Morever, dividing sites of surgical vascular repair into limb, groin or abdominal wounds provides further support for the former argument. Of the ~5 limb wounds, 33 % yielded potential pathogens, whereas 55 ~ of the 20 groin wounds contained potential pathogens and 7 r % of the seven abdominal
26
J . s . CHESHAM AND D. J. PLATT
wounds contained potential pathogens. T h u s , the probability of colonisation appears to be related to the proximity of the w o u n d site to the alimentary tract. Faecal contamination, therefore, probably plays an important role in the colonisation/infection of these wounds. T h e approximately equal distribution of coliforms among all types of w o u n d suggests that contamination by the patients' gut flora may be important in colonisation/infection. T h e uneven distribution of Pseudomonas spp. however, and their association with 'vascular ulcers' and amputation sites suggests that environmental contamination is an important source of colonisation of peripheral sites. S. aureus was most often isolated from 'vascular ulcers', suggesting an endogenous source such as the nose or skin but S. faecalis was most commonly isolated from sites of surgical vascular repair and amputation sites, suggesting faecal contamination. It is difficult to assess the effect of both topical and systemic antimicrobial therapy objectively in terms of correlation with the isolation of potential pathogens. Few wounds that appeared 'clean' were treated whereas almost all wounds of grades II or I I I had been treated, at least with topical preparations, before sampling. Within these limitations, topical therapy did not reduce the numbers or types of organisms present in those wounds that were studied for several weeks. Similarly, systemic therapy appeared to exert a minimal, if any, effect on the colonisation and healing of wounds. (We are grateful to Mr. J. G. Pollock and Mr D. P. Leiberman for their co-operation during the collection of specimens and for ready access to their patients.) References I. Krizek TJ, Robson MC, Kho E. Bacterial growth and skin graft survival. Surg Forum I967; I8: 518-519. 2. Lookingbill DP, Miller SH, Knowles RC. Bacteriology of chronic leg ulcers. Arch Dermatol Res I978; I I 4 : I765-I768. 3- Robson MC, Lea CE, Dalton JB, Heggers JP. Quantitative bacteriology and delayed wound closure. Surg Forum 1968; I9: 5Ol-5O2. 4. Robson MC, Shaw RC, Heggers JP. The reclosure of postoperative incisional abscesses based on bacterial quantification of the wound. Ann Surg 197o; I7I : 279-282. 5. Dagher FJ, Alongi SV, Smith A. Bacterial studies of leg ulcers. Angiology I978; 29: 641-653. 6. Leigh DA. An eight year study of postoperative wound infection in two district general hospitals. J Hosp Infection I98I ; 2: 2o7-217. 7. Friedman SA, Gladstone ]'L. The bacterial flora of peripheral vascular ulcers. Arch Dermatol Res I969; xoo: 29-32. 8. DaltreyDC, Rhodes B, ChattwoodJG. Investigation into thernicrobial flora ofhealing and non-healing decubitus ulcers. J Clin Pathol 1981 ; 34: 7Ol-7O5. 9. Lennette EH, Balows A, Hausler WJ Jr, Traunt JP. Manual of clinical microbiology, Washington DC: American Society for Microbiology, I98O.
P a t t e r n s o f w o u n d c o l o n i s a t i o n in patients w i t h p e r i p h e r a l v a s c u l a r disease J. S. Chesham and D. J. Platt University Department of Bacteriology, Royal Infirmary, Castle Street, Glasgow G4 oSF Accepted for publication 18 February 1987 Sutnnaary The bacterial colonisation of ulcers related to vascular disease, amputation sites and sites of surgical vascular repair was studied in 122 patients with vascular disorders. Wounds were graded clinically on the basis of inflammation, oedema and exudate. They were sampled at weekly intervals in order to determine the duratio_n of colonisation. The significance of potential pathogens was assessed. Vascular ulcers were of low mean wound grade (I'87), short duration (I'58 weeks) and potential pathogens were isolated from 82 % of 38 wounds. Amputation sites were of a higher mean wound grade (2"48) and longer duration (2"52 weeks) when compared to vascular ulcers. Potential pathogens were isolated from 69 % of 42 amputation sites. Sites of surgical vascular repair were similar to amputation sites with respect to mean wound grade (2"33) but of a mean duration (1"69 weeks) similar to that of vascular ulcers. Potential pathogens were isolated from 50 % of 42 surgical vascular repair sites but they were not evenly distributed among limb (33 %), groin (55 %) and abdominal wounds (71%). Coliforms were isolated from similar numbers of all wounds. Pseudomonas spp., Staphylococcus aureus and Streptococcus faecalis were evenly distributed with respect to the site of isolation. Pseudomonas spp. and S. aureus were associated more commonly with vascular ulcers and S. faecalis with sites of surgical vascular repair. The use of topical antiseptics and/or systemic antimicrobial agents had a minimal effect in reducing colonisation/infection in the wounds studied.
Introduction Assessment of the clinical significance of bacterial isolates f r o m w o u n d s is generally an area of uncertainty. It often relies on subjective distinction between ' c o l o n i s a t i o n ' and ' i n f e c t i o n ' unless the organism is a well recognised p a t h o g e n (e.g. Streptococcus pyogenes). Quantitative studies have suggested that if the n u m b e r s of bacteria isolated from a w o u n d are ~< IO~ colony f o r m i n g units per gram of tissue or per square centimetre of w o u n d surface t h e n healing will be unimpaired. T h i s approach has been applied to the rejection of skin-grafts 1, the healings of leg ulcers ~ and reclosure of surgical wounds, s' 4 Staphylococcus aureus has been reported as the m o s t c o m m o n l y isolated organism from both leg ulcers 5 and surgical wounds. 6 T h e range of other bacteria isolated was wide and m i x e d cultures were common. 5-8 Quantitative bacteriology in the investigation of w o u n d infection is rarely practicable in routine diagnostic laboratories. T h e distinction between inoi63-4453/87/o4oo2I + 06 $02.00/0
© I987 The British Society for the Study of Infection
22
J.s. CHESHAM
AND D. J. P L A T T
fection and colonisation is usually based on the subjective interpretation of combined clinical and microbiological data. T h e aim of this study was to compare ulcers related to vascular disease with surgical wounds, in respect of patterns of bacterial colonisation/infection, among patients in the same surgical unit. Correlation of these findings with differences in severity of the initial wounds, their rate of healing, as well as systemic and topical antimicrobial therapy was considered also. Patients
Materials and methods
W o u n d swabs were obtained from in-patients in the peripheral vascular disease unit of Glasgow Royal Infirmary between February I98I and June I983. Clinical data including the use of topical antiseptics and systemic antimicrobial therapy were recorded. W o u n d s were classified according to their appearance as follows : Grade I - moist, granulating, no evidence of infection, Grade II - inflammation present with or without oedema, no purulent exudate, Grade I I I - inflammation, oedema and purulent exudate present. Clean dry wounds were not sampled. W o u n d s were further divided clinically into vascular ulcers, amputation sites and sites of surgical vascular repair. Specimens were collected weekly and full details of the severity of each wound were recorded on each occasion. Most wounds were seen to heal with very few patients either dying or being transferred elsewhere without resolution of the wound. T h e n u m b e r of weeks that a given site was sampled provided an estimate of the relative rate of healing of the wounds studied. Specimens Material was collected by means of swabs and processed within 2 h of collection. Swabs were cultured on (I) cysteine-lactose-electrolyte-deficient ( C L E D ) agar (Mast D M I I O ) ( 2 ) Columbia agar base (Oxoid, C M 3 3 I ) containing 5 % defibrinated horse blood and (3) neomycin blood agar (Brucella agar base (Difco, o96ooi), which contained 7o m g / i neomycin (Sigma), 5 % defibrinated horse blood and o'1% (v/v) vitamin K haemin solution (Gibco)). C L E D and blood agar plates were incubated aerobically in an atmosphere of 5 % carbon dioxide at 37 °C overnight. Neomycin blood plates were incubated at 37 °C anaerobically for 48 h. Potential pathogens were purified from primary plates and identified by standard methods. 9 In addition, coliforms were further identified by the A P I 2oE system. Results Swabs were obtained from the wounds of 122 patients when the clinical appearance suggested delayed healing a n d / o r infection. T h e mean times for healing were I'58 weeks (range I - 4 weeks) for vascular ulcers, 2"52 weeks (range i-8 weeks) for amputation sites and I'69 weeks (range I-5 weeks) for sites of surgical vascular repair. These findings indicate that sites of surgical vascular repair take longer to heal than 'vascular ulcers' but that both of these types of w o u n d heal more rapidly than wounds at amputation sites.
Bacterial colonisation of wounds
23
Table I The classification of wounds in patients with peripheral vascular
disease ' Vascular ulcers' W o u n d grade
Amputation sites
Sites of surgical vascular repair
No.
(%)
No.
(%)
No.
(%)
9" 25 4
(24) (66) (IO)
3 i6 23
(7) (38) (55)
2 24 16
(5) (57) (38 )
I II III Mean
1-87
2"48
2"33
Table II The distribution of potential pathogens among the different types of
wounds studied 'Vascular ulcers' n = 38
Coliforms Pseudomonas spp. Staphylococcus aureus Streptococcus faecalis Beta-haemolytic streptococci Anaerobes Any potential pathogen isolated Potential pathogen not isolated
Amputation sites n = 42
Sites of surgical vascular repair n = 42
No.
(%)
No.
(%)
No.
(%)
I9 IO I5 3 I 2 3I
(5 ° ) (26) (39) (8)
(43) (2r) (2I) (I4)
I9 3 9 8
(45) (7) (2I) (I9)
(82)
r8 9 9 6 I 2 29
(69)
i 2I
(50)
7
(I8)
I3
(3I)
2I
(5o)
T h e classification of wounds into grades I, II and I I I according to severity and the percentage in each grade is shown in Table I. 'Vascular ulcers' were of the lowest grade with a mean value of I ' 8 7 . A m p u t a t i o n sites and sites of surgical vascular repair had mean grades of 2"48 and 2"33 respectively. Although ' vascular ulcers' were of short duration, having the lowest mean grade of wound, potential pathogens were isolated from 82% (Table II). W o u n d s at amputation sites were of longer duration and higher mean grade; 69 % yielded potential pathogens. W o u n d s at sites of surgical vascular repair were similar to 'vascular ulcers' in duration but their mean grade was close to that of wounds at amputation sites. Half the wounds at sites of surgical vascular repair did not yield any potential pathogen. Table II also shows the categories of organisms found in the wounds samples. Escherichia coli, Proteus spp., Enterobacter spp., Klebsiella spp., Citrobacter spp., Serratia spp., and Hafnia alvei were not separately distinguished and have been included under the general heading of coliforms. W h e n isolated
24
J.s. CHESHAM AND D. J. PLATT
Table I I I Potential pathogens isolated in the presence and absence of topical, systemic or a combination of both types of antimicrobial therapy Potential pathogens isolated
Topical and systemic antimicrobial therapy Topical therapy only Systemic therapy only No topical or systemic antimicrobial therapy
Potential pathogens not isolated
Total no. of wounds
No.
(%)
No.
(%)
60
48
(80)
I2
(20)
(55)
19 8
5
3
43 I4
24 6
2
(44)
from a given specimen, coliforms appeared to be evenly distributed among the three types of wound. Pseudomonas spp. were isolated more commonly from patients with 'vascular ulcers' and wounds at amputation sites than from patients with wounds at sites of surgical vascular repair. S.aureus was most often isolated from 'vascular ulcers' and S.faecalis from wounds at sites of surgical vascular repair. Beta-haemolytic streptococci and anaerobes were seldom isolated. T h e wounds of 65 (53 %) of the patients studied were being treated topically with either sodium hypochlorite (Milton), hydrogen peroxide, chlorhexidine gluconate (Hibitane), povidone-iodine (Betadine) or a combination of these antimicrobial agents (Table III). In addition, many different systemic antimicrobial agents were prescribed during the study. T h e s e included flucloxacillin, cephalosporins, metronidazole, penicillins V and G, ampiciUin, cotrimoxazole, tetracycline, gentamicin, augmentin, erythromycin and fucidin. At least one of these systemic antimicrobial agents was prescribed for lO3 patients, often for unrelated problems such as respiratory or urinary tract infections. T h e r e was no significant difference in the incidence of isolation of potential pathogens from wounds in patients treated with neither topical nor systemic antimicrobial agents, topical antimicrobial agents alone or systemic agents alone. T h e n u m b e r of patients treated with both topical a n d systemic antimicrobial agents was higher than those treated with none or either topical or systemic therapy alone. Potential pathogens were isolated from significantly more wounds when both types of therapy were administered. Only flucloxacillin, cephalosporins and metronidazole were prescribed for a sufficient n u m b e r of patients to determine any effect against organisms normally considered to be susceptible. Metronidazole was prescribed in 43 patients. Anaerobic organisms were isolated from the wounds of five patients, four of w h o m had been receiving metronidazole. Flucloxacillin was prescribed for 49 patients and S.aureus was isolated from the wounds of 12 (25 %) of these. Of 73 patients who did not receive flucloxacillin therapy 21 (29 %) had wounds which contained S.aureus. Similarly, of the wounds of 69 patients receiving cephalosporins 27 (39 %) contained coliforms whereas those
Bacterial colonisation of wounds
25
of 22 (42 %) of 53 patients not receiving cephalosporins yielded coliforms. Thus, organisms against which specific antimicrobial therapy was directed were not isolated more often from patients who did not receive those agents. This indicates that antimicrobial therapy did not exert a detectable effect on the bacterial flora of the wounds. Discussion
Factors which contribute to the colonisation or infection of wounds are numerous and include underlying pathology, type and site of the wound as well as the number and types of organisms isolated. Daltrey and colleagues 8 classified a wound as infected if there was a persistent presence of Io 4 microorganisms per square centimetre. Such criteria, however, pose a number of problems. For example, the authors made no attempt to define 'persistence'. Moreover, these results showed that there were notable variations in the ability and numbers of organisms required to impair wound healing. Robson and colleagues 4 have stated that, in view of the virulence of beta-haemolytic streptococci, no attempt was made to close a wound if they were present, regardless of their numbers. Although at least one species of potential pathogen was isolated from 82 % of 'vascular ulcers' the wounds were on average of short duration and low grade. This suggests bacterial colonisation rather than active infection. These results are consistent with the patients selected and the aims of surgery in that patients admitted for treatment would have been pre-selected such that expectation of success of treatment was high and that surgery was designed to improve the blood supply. These factors, together with that of good wound management, should lead to healing of the 'vascular ulcer'. Wounds at amputation sites were of longer duration and of a more severe grade compared to that of 'vascular ulcers'. Amputation sites were not a homogeneous group and although the amputation was performed through healthy tissue, circulation distal to the site was presumably poor and may explain the greater probability of colonisation/infection of these wounds by potential pathogens. T h e i r longer duration and higher grade of wound may be due in part to the severity of trauma associated with amputation. This together with the high incidence of a potential pathogens being isolated (69 %) suggests infection rather than colonisation. Sites of surgical vascular repair formed an intermediate group. Although they resembled 'vascular ulcers' in duration and amputation sites in severity, potential pathogens were isolated from a smaller proportion of sites (5o %) than from either of the former groups. Since little difference was seen in the duration and severity of these wounds, irrespective of whether potential pathogens were isolated, it seems likely that colonisation rather than infection was involved. Even so, the possibility that deep-seated infection (with failure to isolate the causative organism from surface samples) cannot be excluded. Morever, dividing sites of surgical vascular repair into limb, groin or abdominal wounds provides further support for the former argument. Of the ~5 limb wounds, 33 % yielded potential pathogens, whereas 55 ~ of the 20 groin wounds contained potential pathogens and 7 r % of the seven abdominal
26
J . s . CHESHAM AND D. J. PLATT
wounds contained potential pathogens. T h u s , the probability of colonisation appears to be related to the proximity of the w o u n d site to the alimentary tract. Faecal contamination, therefore, probably plays an important role in the colonisation/infection of these wounds. T h e approximately equal distribution of coliforms among all types of w o u n d suggests that contamination by the patients' gut flora may be important in colonisation/infection. T h e uneven distribution of Pseudomonas spp. however, and their association with 'vascular ulcers' and amputation sites suggests that environmental contamination is an important source of colonisation of peripheral sites. S. aureus was most often isolated from 'vascular ulcers', suggesting an endogenous source such as the nose or skin but S. faecalis was most commonly isolated from sites of surgical vascular repair and amputation sites, suggesting faecal contamination. It is difficult to assess the effect of both topical and systemic antimicrobial therapy objectively in terms of correlation with the isolation of potential pathogens. Few wounds that appeared 'clean' were treated whereas almost all wounds of grades II or I I I had been treated, at least with topical preparations, before sampling. Within these limitations, topical therapy did not reduce the numbers or types of organisms present in those wounds that were studied for several weeks. Similarly, systemic therapy appeared to exert a minimal, if any, effect on the colonisation and healing of wounds. (We are grateful to Mr. J. G. Pollock and Mr D. P. Leiberman for their co-operation during the collection of specimens and for ready access to their patients.) References I. Krizek TJ, Robson MC, Kho E. Bacterial growth and skin graft survival. Surg Forum I967; I8: 518-519. 2. Lookingbill DP, Miller SH, Knowles RC. Bacteriology of chronic leg ulcers. Arch Dermatol Res I978; I I 4 : I765-I768. 3- Robson MC, Lea CE, Dalton JB, Heggers JP. Quantitative bacteriology and delayed wound closure. Surg Forum 1968; I9: 5Ol-5O2. 4. Robson MC, Shaw RC, Heggers JP. The reclosure of postoperative incisional abscesses based on bacterial quantification of the wound. Ann Surg 197o; I7I : 279-282. 5. Dagher FJ, Alongi SV, Smith A. Bacterial studies of leg ulcers. Angiology I978; 29: 641-653. 6. Leigh DA. An eight year study of postoperative wound infection in two district general hospitals. J Hosp Infection I98I ; 2: 2o7-217. 7. Friedman SA, Gladstone ]'L. The bacterial flora of peripheral vascular ulcers. Arch Dermatol Res I969; xoo: 29-32. 8. DaltreyDC, Rhodes B, ChattwoodJG. Investigation into thernicrobial flora ofhealing and non-healing decubitus ulcers. J Clin Pathol 1981 ; 34: 7Ol-7O5. 9. Lennette EH, Balows A, Hausler WJ Jr, Traunt JP. Manual of clinical microbiology, Washington DC: American Society for Microbiology, I98O.