- Email: [email protected]
Non-sporeforming anaerobic bacteria in clean surgical wounds—air and skin contamination
Journal
of Hospital
Infection
(1984)
Non-sporeforming surgical wounds
5, 38-49
anaerobic bacteria in clean - air and skin contamination
E. Benediktsdbttir Institute
of Clinical
and K. Kolstad”
Bacteriology and *Department Uniaersity of Uppsala, Uppsala,
of Orthopaedic Sweden
Surgery,
Summary: The contamination of clean surgical wounds with anaerobic and aerobic bacteria was studied in 52 hip operations. In addition to wound samples, samples from air and the patients’ skin were taken. The median of the total number of bacteria isolated from the wounds was 26 colony-forming units (cfu). ‘I’he median percentage of anaerobic bacteria in the wound counts was 30. Propionihacterium spp. were found in 71 per cent of the wounds and anaerobic or microaerophilic cocci, most often Peptococcus spp., were found in 23 per cent. In six of 43 patients the same bacterial flora was seen in the skin samples and wounds. The geometric mean of the total number of bacteria in the air was 70.3 cfu;m”. Of these the median percentage of anaerobic bacteria was 30.3. When operating clothing of a disposable fabric (Barrier 45O’Q, Johnson & Johnson) was used, the counts of airborne bacteria were a little less than half those found when conventional cotton clothing was worn. Probably because of the overall low air counts in the operating theatre and the great variability in the individual bacterial counts from the operation wounds, a significant decrease in wound contamination was not observed. A positive correlation was found between the duration of operation and wound contamination.
Introduction Anaerobic skin bacteria of relatively low pathogenicity have been reported to cause postoperative infections after total hip replacement (Schwan et al., 1977; Carlsson, Josefsson and Ilindberg, 1978; Launder and Hungcrford, 1981; Fitzgerald et al., 1982). ‘l‘hese infections often become manifest weeks or months after the operation has taken place, and probably the importance of the low-pathogenic skin bacteria as infecting agents has been underestimated (Kamme and Lindberg, 1981; Whyte et al., 1981). ‘I’hc transmission routes of aerobic bacteria in clean surgery have been investigated by many authors. In major implant surgery airborne infection has been considered important (Charnley, 1972) and a controlled multicentre study has shown that operations performed in ultra-clean air have a significantly lower infection rate than those performed in conventionally ventilated operating rooms (1,idwell et al., 1982). Results of studies on airborne anaerobic bacteria have indicated the possibility of contamination of the operation wound with anaerobic bacteria from exogenous sources 01954701
c> 1984 The IIospital
84~1110038 1 12 SO2.OO;O
38
Infection
Society
Contamination
of clean
surgical
wounds
39
(Hambraeus and Benediktsdottir, 1980). However, studies on the occurrence of anaerobic bacteria in clean operation wounds have been sparse, and the importance of different factors in the reduction of wound contamination from anaerobic bacteria is not clear. In this study the influence of different clothing on the anaerobic and aerobic bacterial flora in the air and in the wound during total hip replacement was examined. Attempts were made to relate the wound contamination to the airborne bacteria, to the bacteria on the skin of the patients, to the duration of the operation and to the operative diagnosis. We also report on the types of anaerobic bacteria that were found to contaminate the operation wounds. Patients,
materials
and
methods
Putients Sixty-four patients undergoing total hip replacement were included in the investigation. The operations were performed at the Department of Orthopaedic Surgery of the TJppsala University Hospital from December 1981 to December 1982. The diagnoses were osteoarthritis in 47 patients, rheumatoid arthritis in nine, and previously surgically treated (on average 31 months earlier; range 2-- 84 months) fracture of the femoral neck m five patients. Added to the ‘fracture group’ were three further patients. Two were patients with previous infections-one with a septic coxitis leading to an arthrodesis operation some 22 years ago, and one who had had an infected pelvic fracture 1.5 years before the hip replacement. The third patient had developed necrosis of the femoral head after kidney transplantation. Thus the ‘fracture group’ consisted in all of eight patients. The age and sex distribution of the patients is shown in Figure 1. All patients received antibiotic prophylaxis, including cloxacillin infusion 2 g three times during the first 24 h, starting in the morning of the.day of surgery, followed by oral cloxacillin 1 g three times per day for about 10 days postoperatively. The operations were carried out by 12 different surgeons. The length of time between the skin incision and the last suture was recorded. The operating suite The design of the operating unit has been described in detail elsewhere (Hambraeus, Rengtsson and Laurell, 1977, 1978). ‘I’he operations were performed in an operating theatre that has a zonal ventilation system, where the main part of the filtered air is introduced above the operating table through a clean air inlet installed in the ceiling. About 80 air exchanges per hour are achieved in the central part of the room, which is protected from contamination from the peripheral room air by ‘air curtains’ (Abel and Allander, 1966). The peripheral zone is supplied with air at 2@ -30 changes/h.
E. Benediktsdhttir
20-29
40-49 30-39
Figure 1. Age and sex distribution females.
and
K. Kolstad
60-69 SO-89 50-59 70-79 90-99 Age
of the patients
included
in the investigation.
q , males; H ,
Be-operatice routines The patients were washed over the whole body with a detergent solution containing 4 per cent chlorhexidine (‘Hibiscrub’) the day before surgery. On the day of surgery the operation field was washed with ‘Hibiscrub’ then disinfected by applying 0.5 per cent chlorhexidine in 70 per cent ethanol, and finally covered with adhesive film surrounded by cotton drapes. Operating clothing Members of the surgical team scrubbed up with ‘Hibiscrub’ and dressed in sterile surgical gowns. The scrubbed staff wore double gloves when working, and all staff and visitors in the operating room wore caps and masks. Every other week the non-scrubbed members of the staff wore trousers manufactured from disposable non-woven fabric (Barrier 450@, Johnson & Johnson) and the scrubbed staff wore surgical gowns of this same fabric. During the alternate weeks the staff wore conventional cotton operating clothing and a gown. Sampling routines Wound sampling. The wound sampling was performed with six sterile pads of 85 per cent viscose and 15 per cent polyester and polyamide (Benediktsdottir and Hambraeus, 1983). The pads were 5 x 6 cm in size. The surgeon imprinted the pads for a few seconds on the subcutaneous tissue after suture of the fascia. If the tissue was dry it was moistened with a few millilitres of saline before sampling. The pads were placed in Petri plates containing 1O-l 5 ml of modified Stuart agar. ‘I’he plates were transported to the laboratory in an anaerobic jar after the operation.
Contamination
of clean
surgical
wounds
41
Skin sampling. Two swab samples and one skin biopsy were taken from the skin of each patient. The first swab sample was taken from the proposed incisional area before the final disinfection procedure and the second immediately after the operation from an area adjacent to the incision. A sterile cotton applicator was moistened in 2 ml of a neutralizing fluid containing 3 per cent lecithin and 20 per cent Tween 80 in nutrient broth. An area of about 1 x 10 cm was rubbed with the applicator for 10 s and the applicator was then put back into the neutralizing fluid and was transported after the operation to the laboratory. The biopsy, a few square millimetres, was taken from the incisional edge at the beginning of the operation and transported to the laboratory in 1 ml of the neutralizing fluid. Air sampling. The air sampling was performed with a Sartorius filter sampler with a gelatin filter (Ilambraeus and Benediktsdottir, 1980) and a capacity of 40 l/min. The filter holder was placed in the immediate vicinity of the wound and eight to 10 lo-min samples were taken at each operation. After the operation the filters were transported to the laboratory in the filter holders. Bacteriological methods ‘I’he wound sampling pads were handled with sterile forceps on a laminar flow bench. The technicians disinfected their bare arms with 70 per cent ethanol and wore sterile gloves while working. ‘I’he pads were cultured after treatment in a Stomach& lab blender (Colworth, Stomacher 400) as described earlier (Benediktsdottir and Hambraeus, 1983). One sterile pad was always processed with the wound pads for control of laboratory contamination. The skin swabs were cultured as follows. After shaking the tube on a whirlmixer for 15 s, the swab was pressed against the inside wall of the tube to remove the excess fluid and discarded. Ten-step dilutions were made in Nbroth containing 0.3 per cent lecithin and 2 per cent Tween 80. Of the sample 0.1 ml was plated on supplemented BHI agar (Iloldeman, Cato and Moore, 1977) and incubated anaerobically for 7 days. The skin biopsy was homogenized in 2 ml of the neutralizing fluid with a manual glass homogenizer, 1 ml was plated on supplemented BHI agar for anaerobic incubation for 7 days. The gelatin filters were placed on supplemented BHI agar and incubated anaerobically for 7 days. ‘I’he anaerobic bacteria were identified according to the method of Holdeman et al. (1977) as described earlier (Benediktsdottir and Hambraeus, 1982). The aerobic bacteria were only classified by colonial and cellular morphology. From the pour plates of the wound cultures, up to 20 colonies were analysed from each pad. The distribution of aerobic and anaerobic bacteria in the whole sample was assumed to be the same as that in the colonies that were analysed.
42
E. Benediktsdbttir
and
K. Kolstad
Results
Wound sampling Fifty-two operation wounds were sampled. Cultures from all wounds except one were positive. Eighty-six per cent of the control pads were negative, four pads grew one colony-forming unit (cfu), two pads 3 cfu, and one 14 cfu. ‘The total bacterial count from the wounds ranged from 0 to 47,747; the median was 26. Figure 2 shows the cumulative distribution of the wound counts. Two operations had counts between 1000 and 9999 and two had counts greater than 10,000.
Cumulative
Figure 2. Cumulative the air (0).
distribution
of the number
percentage
of bacteria
isolated
from the wounds
CM)and
Anaerobic bacteria were isolated from 40 of the 52 operation wounds sampled. The proportion of anaerobic bacteria among the total number isolated ranged from 0 to 100 per cent, the median was 30 per cent. More than 100 cfu of anaerobic bacteria were isolated from nine wounds and of aerobic bacteria from eight wounds. In Table I the types of anaerobic and aerobic bacteria isolated from the wounds are shown. Propionibacterium acnes was isolated from 37 wounds, and P. granulosum was found in two wounds, in one together with I’. avidum. Anaerobic Gram-negative rods that were not typed further were isolated from two wounds, both of which had total bacterial counts exceeding 900. Anaerobic and microaerophilic cocci were isolated from 12 wounds. The type most often found was Peptococcus magnus, which was isolated from four wounds. Other types isolated were Pepto. saccharolyticus, Pepto, asaccharolyticus and Veillonella paroula. Gram-positive cocci that were not further typed were isolated from four wounds. In two wounds with total bacterial counts exceeding 600, the organisms isolated were dominated by anaerobic cocci, one by Pepto. saccharolyticus and the other by V. paroula. Members of the Micrococcaccae were the most common bacteria isolated.
Contamination Table
I. Types
of clean
surgical
of hactevia isolated from the operating
zwunds
So. of wounds
Organisms Anaerobic rods P. acnes P. granulosum P. az:idunz Gram-negative
43
wounds
37 2 rods*
Anaerobic cocci Pepto. magnus Pepto. saccharolyticus Pepto. asaccharolyticus Veillonella parwla Gram-positive cocci* Aerobic rods Coryneforms,
22
Bacilli
Aerobic cocci Micrococcaceae
45
*So further identification. ‘I’he total number of operations was 52.
They were found in 45 wounds. streptococci in nine wounds.
Aerobic
rods were found in 22 wounds
and
Air sampling Air samples were taken during 56 operations. The total bacterial,,counts in the air ranged from 15 to 290/m3; the geometric mean was 70.3/m3. The proportion of anaerobic bacteria in the total bacterial count ranged from 1.5 per cent to 67.7 per cent, with a median of 30.3 per cent. Figure 2 shows the cumulative distribution of the air counts. Propionibacterium acnes was the predominant anaerobic species isolated during 55 operations. Propionibacterium granulosum was isolated in the course of eight operations, P. azidum in 11 and Gram-negative anaerobic rods in one operation. Anaerobic and microaerophilic cocci were isolated during six operations. Skin sampling All three types of skin samples and wound samples were taken from 43 patients. Thirty-six of the swab samples taken before the operation and 39 taken after the operation had bacterial counts from 0 to 9;O.l ml, and 38 of the skin biopsies taken had counts under 10 per 1 ml. Bacterial counts from 10 to 2230 were obtained from one or more of the skin samples in 14 patients. In five patients these counts were obtained only from the first swab sample, in three only from the second swab sample and in four only from the biopsy sample. The predominant bacteria isolated from the skin samples were aerobic cocci and coryneforms, and anaerobic propionibacteria. In two patients
44
anaerobic exceeded
E. Benediktsdbttir
and
K. Kolstad
cocci were isolated. In eight of the 14 samples with counts 10 the samples were dominated by anaerobic bacteria.
that
Helutionship between skin and wound samples So quantitative or qualitative relationship was found between the skin samples with counts less than 10 and the wound samples. Of the 14 patients with bacterial counts of 10 or more in the skin samples, a corresponding bacterial flora was found in the wound in six, in numbers from 86 to 702 per wound. In one patient the bacteria were isolated from the swab taken after the operation. In another patient from both swab samples, and in the remaining four patients from the biopsy sample. The bacterial types involved were Pepto. saccharolyticus in one patient, P. acnes in two, Micrococcaceae in two and a mixed flora of P. avidum and Micrococcaceae in one patient. Relationship between levels of airborne bacteria and wound contaminution Roth air and wound sampling was performed during 44 operations. Regression analysis was carried out on data from 34 operations. It was assumed that the contamination of wounds which had counts greater than 1000 was not derived from the air. These operations and the six in which the same bacteria were found on the skin as in the wound were not included. The data were transformed by log conversion to improve the normality. ‘I’he regression coefficient for logi (count/wound) on log,, (air count/m3) was 0.72 t- 3 .O. Comparison of dijfevent clothing Effect on airborne bacteria. The geometric mean of the total bacterial count was 101 cfu/m3 (N= 28) when the staff wore conventional cotton operating dress and 49 cfu/m3 (N=28) when the disposable clothing was used (P
Contamination
of clean
surgical
wounds
45
Relationship between duration of operation and zound contamination The length of operation ranged from 90 to 275 min, with a mean of 160 min. Fifteen of the operations took 3 h or more to perform. The median count in the wound samples was 125 per wound in these operations, and four of the six operations in which the same bacteria were isolated from the skin of the patient and the wound were included in this group. The median count in the wound samples from the 37 operations lasting less than 3 h was 11 per wound. A regression analysis was performed on data from all 52 operations. The data were transformed by log conversion to improve the normality. The regression coefficient for log,, (count/wound) on log,, [operation time (min)] was 2.61 + 2.67. Relationship between operatiae diagnosis and wound contamination The operation wounds were sampled in 39 patients undergoing surgery for osteoarthritis, and the median of the bacterial count per wound was 43. Samples were also taken from the operation wounds in six patients who were operated on for rheumatoid arthritis and seven patients from the ‘fracture group’. The median bacterial counts in these two groups were 16 and 5 per wound respectively. All four patients with a bacterial count in the wound exceeding 1000 were osteoarthritis patients undergoing operation for the first time, and so far none of them have shown any sign of clinical infection. Patients ‘l’he observation time varied from 1 to 11 months. During the hospital stay postoperatively (2-3 weeks) there were no deep infections. One patient developed a deep infection with group B streptococci following extraction of an infected tooth 3 months after the operation. ‘I’his patient had a total bacterial wound count of four. Another patient had a clinical infection after several re-operations performed because of prosthetic dislocation, and had a total bacterial wound count of 83, with 82 anaerobic bacteria. Cultures of the later infection however showed growth of penicillinase-producing staphylococci, Enterobacter aerogenes and enterococci.
Discussion
Several investigations have been carried out on the contamination of operation wounds in hip surgery. Positive cultures have been found by different authors in 1.5 per cent to 36 per cent of the wounds (Fitzgerald et al., 1973; Murray, 1973; Fitzgerald and Washington, 1975; Lindgren, Elmros and Holm, 1976; Bechtol, 1979; Ilamilton et al., 1979; Blomgren, IIambraeus and Malmborg, 1982). In this study all but one of the wound cultures were positive. Although laboratory contamination cannot be totally ruled out, the most likely explanation for the large number of positive
46
E. Benediktsdbttir
and
K. Kolstad
cultures is that a highly sensitive sampling method was used (Benediktsdottir and Hambraeus, 1983). Hitherto few attempts have been made to investigate the contamination of clean operation wounds by anaerobic bacteria. Fitzgerald and Washington (1975) isolated P. acnes in 24 and peptococci in two of 157 samples taken from operation wounds during orthopaedic surgery. Lindgren et al. (1976) did not isolate any anaerobic skin bacteria from the wound drain in hip surgery in spite of an anaerobic culture procedure. In this study anaerobic skin bacteria were found in 77 per cent of the wounds, and in 17 per cent of them anaerobic bacteria were grown in counts that exceeded 100. The types of anaerobic bacteria found both in the air and wound samples were mostly members of the normal anaerobic skin flora, Propionibacterium spp. and Peptococcus spp. Xone of the patients participating in this study developed a primary postoperative infection (i.e. during the hospital stay). In view of the small number of patients, the antibiotic prophylaxis and the short observation period, the clinical significance of these observations is uncertain. Contamination of operation wounds in clean surgery can take place in several different ways: from the air, through direct settling or indirectly via instruments and gloves; by direct contact from the clothes or from punctured gloves of the surgical team; and from the skin of the patient. Transmission of bacteria from the air to the wound occurs continuously during all operations. It is possible to estimate the proportion of bacteria settling directly in the wound. By using the same approach as was used by Whyte, Hodgson and Tinkler (1982) it can be calculated that when the total count in the air is 70.3 cfu/m3 (the geometric mean found in this study), the amount of bacteria settling on to the 180 cm2 wound will be 30 cfu. This is very close to the median number found in the wound in this investigation. The lowest and the highest counts in the air were 15 and 290 cfu/m3, respectively, and the calculated bacterial deposit in the wound area sampled would therefore range from 6 to 125 cfu. The total count of bacteria isolated from the wounds varied, however, from 0 to 47,747 cfu. It is obvious that in some of the wounds a break-up of bacterial clusters had occurred or the bacteria had been brought to the wound by other routes than direct deposition from the air. In this study a 5 1 per cent reduction in air contamination was found when the staff wore trousers of a non-woven fabric and the scrubbed staff also wore gowns of the same fabric, compared with that when all the staff wore clothing of conventional cotton. This is consistent with results from other investigations (Dineen, 1973; Mitchell, Evans and Keen, 1978; Whyte et al., 1978). The levels of both anaerobic and aerobic bacteria were reduced, a finding in accordance with the results of dispersal experiments in a test chamber, which showed that the dispersal of anaerobic bacteria from the human skin occurs mainly from the same region as that of aerobic bacteria, the lower trunk (Benediktsdottir and Hambraeus, 1982). Whyte et al. (1982) reported that a 99 per cent reduction in air contamination was associated with
Contamination
of clean
surgical
wounds
47
a 97 per cent reduction in wound contamination. In this study the variation in the individual bacterial counts from the air and wounds was such that the regression of the latter on the former did not reach significance at the 5 per cent level. The same was true for the comparison between the wound values with use of cotton and disposable clothing, although the reduction of the median wound contamination, 45 per cent, was almost identical to that of the geometric mean levels of air contamination, 51 per cent. In an operating theatre with efficient ventilation as that used in this study, there will only be a low degree of airborne contamination of the wounds. Prolongation of the operations and a higher level of airborne bacteria will naturally change this situation. The disinfection methods used in the present study are among the most effective methods for skin disinfection known (Lilly, Lowbury and Wilkins, 1979). However, contamination of the wound with bacteria presumably derived from the patient’s skin was found in six cases, or 11.5 per cent of the wounds. Probably contamination from the skin occurred to an even greater extent, as negative skin samples do not preclude the skin as a contaminating factor. ‘l’he skin sampling techniques used, swab and biopsy, give a limited picture of the skin-the former reveals only the surface flora (Evans and Stevens, 1976) and the latter is limited to very small areas. In hip surgery, wound contamination from the patient’s skin has been reported by other authors. Whyte et al. (1982) found wound contamination associated with exceptionally high patient skin counts in two of 52 operations, and McLauchlan et al. (1976) observed major contamination of wounds (more than 100 cfu(operation) in 11 per cent of patients operated on in a surgical isolator that blocked both the air and the contact routes of crossinfection. A positive correlation was found between duration of operation and wound contamination. Increased contamination from exogenous sources is to be expected with prolongation of the operation. In this study evidence was also found for greater transmission of bacteria from the patient’s skin when the operation time exceeded 3 h. ‘I’his is to be expected, and is probably due to loosening of drapes and moistening of clothing with blood and fluid, leading to bacterial invasion from the patient’s skin into the wound. It was suspected that previous surgery and infection in the hip area might increase the bacterial counts in the operation wounds compared with patients being operated on for the first time for osteoarthritis or perhaps rheumatoid arthritis. In the present study no difference was found in the levels of bacterial contamination of the wounds between patients with different operative diagnoses. l’he results of this study indicate that in about 20 per cent of the operations transmission occurs from the skin of the patient and possibly by contact with the operating team. This contamination seems to be of a high level. Whatever the route of transmission, anaerobic bacteria constitute a large proportion of the contamination. The results of this study clearly de-
48
E. Benediktsdbttir
and K. Kolstad
monstrate the importance of anaerobic cultivation procedures in hygienic studies of major implant surgery. It is also evident that when estimating the necessary standard of air hygiene at clean surgery such as hip joint replacement, the presence of airborne anaerobic as well as aerobic bacteria must be taken into consideration. References Abel,
E. & Allander, C. (1966). Undersijkning av nytt inblasningssystem fiir rcna rum. C’vs, so. 8. Bechtol, C. 0. (1979). Environmental bacteriology in the unidirectional (vertical) operating 114, 784-788. room. Archiz~s of Szqery Benediktsdbttir, E. & Hambraeus, A. (1982). Dispersal of non-sporeforming anaerobic bacteria from the skin. Journal of IIygiene 88, 487-500. Benediktsdbttir, I<. & Hambracus, A. (1983). Isolation of anaerobic and aerobic bacteria from clean surgical wounds. An experimental and clinical study. Journal of Hospital Infection 4, 141-148. Blomgren, G., Hambraeus, A. & Malmborg, A.-S. (1982). Hygienic and clinical study of elective and acute hip operations. Journal of Hospital Infection 3, 111-121. Carlsson, A. S., Josefsson, G. & Lindberg, I,. (1978). Revision with gentamicin-impregnated cement for deep infections in total hip arthroplasties. Journal of Bone andJoint Surgery 60-A, 1059-1064. Charnley, J. (1972), Postoperative infections after total hip replacement with special reference to air contamination in the operating room. Clinical Orthopaedics and Related Research 87, 167-187. Dineen, I’. (1973). The role of impervious drapes and gowns in preventing surgical infection. Clinical Orthopaedics and Related Research 96, 210-212. quantitation of surface and subsurface Evans, C. A. & Stevens, R. J. (1976). D’ff1 crcntial bacteria of normal skin by the combined use of the cotton swab and the scrub methods. 3, 576 581. Journal of Clinical Microbiology Fitzgerald, R. H., Jr, Peterson, L. F. A., Washington, J. A., van Scoy, Ii. E. &Coventry, M. B. (1973). Bacterial colonization of wounds and sepsis in total hip arthroplasty. Journal of Bone and Joint Surgery S-A, 1242-I 250. Fitzgerald, R. I-I., Jr, Rosenblatt, J. E., Tenney, J. H. & Bourgault, A.-M. (1982). Anaerobic septic arthritis. Clinical Orthopaedics and Related Research 164, 141-148. Fitzgerald, Ii. II., Jr. & Washington, J. A. (1975). Contamination of the operative wound. Orthopedic Clinics of ATorth America 6, 1105 1114. IIambraeus, A., Bengtsson, S. & Laurell, G. (1977). Bacterial contamination in a modern operating suite. 1. Effect of ventilation on airborne bacteria and transfer of airborne particles. Journal of Hygiene 79, 121- 132. Hambraeus, A., Bcngtsson, S. & Laurell, G. (1978). Bacterial contamination in a modern operating suite. 2. Effect of a zoning system on contamination of floors and other surfaces. Journal of Hygiene 80, 57-67. Hambraeus, A. & Benediktsdbttir, E. (1980). Airborne non-sporeforming anaerobic bacteria. Journal of Hygiene 84, 181-189. Hamilton, H. W., Booth, A. ID., Lone, I;. J. & Clark, N. (1979). Penetration of gown material by organisms from the surgical team. Clinical Orthopaedics and Related Research 141, 237-246. Manual. Holdernan, L. V., Cato, E. I’. & Moore, W. E. C. (cds) (1977). A naerobe Laboratory Anaerobe T,aboratory, Virginia Polytechnic Institute and State University, Blacksburg. after Kamme, C. & T,indberg, I,. (1981). A erobic and anaerobic bacteria in deep infections total hip arthroplasty. Clinical Orthopaedics and Related Research 154, 201-207. Launder, W. J. & Hungerford, I>. S. (1981). Late infection of total hip arthroplasty with Propionibacterizlm acnes: A case and review of the literature. Clinical Orthopaedics and Related Research 157, 170-177. Lidwell, 0. M., Lowbury, E. J. L., Whyte, W., Blowers, R., Stanley, S. J. & I,owe, L). (1982). Effect of ultraclean air in operating rooms on deep sepsis in the joint after total hip or knee replacement: a randomized study. Ijritish Medical Journal 285, 10 -14.
Contamination Lilly,
of clean
surgical
wounds
49
H. A., I,owbury, E. J. I,. & Wilkins, M. D. (1979). Limits to progressive reduction of resident skin bacteria by disinfection. ~ournnl of Clinical Pathology 32, 382-383. Lindgren, IJ., I’lmros, ‘I’. & Holm, S. E. (1976). Bacteria in hip surgery. Acta Orthopaedica SEandinaCa 47, X20-323. McLauchlan, J., I,ogie, J. Ii. C., Smylie, II. G. & Smith, G. (1976). A study of the wound environment during total hip arthroplasty. Postgraduate Medical journal 52, 55&557. Mitchell, S. J., Evans, 1). S. & Kerr, A. (1978). Reduction of skin bacteria in theatre air with comfortable, non-woven disposable clothing for operating-theatre staff. IXtish Medicnl Journal 1, 696498. Murray, W. I<. (1973). Results in patients with total hip replacement arthroplasty. Clinical Ovthopaedics and Related Research 95, 80-90. Schwan, A., Bengtsson, S., IIambraeus, A. & Laurell, G. (1977). Airborne contamination and postoperative infection after total binI renlacement. Acta Ortho~aedica Scandintzeica I 48, 86 94. Whyte, W., Hodgson, Ii., Bailey, I’. V. & Graham, J. (1978). ‘The reduction of bacteria in the operation room through the use of non-woven clothing. British Journal of Surgery 65, 469474. Whyte, W., IIodgson, I<. & Tinkler, J. (1982). ‘I’he importance of airborne bacterial contamination of wounds. Journal of hospital Infection 3, 123-135. Whyte, W., IIodgson, IX., Tinkler, J. & Graham, J, (1981). Th e isolation of bacteria of low pathogenicity from faulty orthopaedic implants. Journal of Hospital Infection 2,219-230.
of Hospital
Infection
(1984)
Non-sporeforming surgical wounds
5, 38-49
anaerobic bacteria in clean - air and skin contamination
E. Benediktsdbttir Institute
of Clinical
and K. Kolstad”
Bacteriology and *Department Uniaersity of Uppsala, Uppsala,
of Orthopaedic Sweden
Surgery,
Summary: The contamination of clean surgical wounds with anaerobic and aerobic bacteria was studied in 52 hip operations. In addition to wound samples, samples from air and the patients’ skin were taken. The median of the total number of bacteria isolated from the wounds was 26 colony-forming units (cfu). ‘I’he median percentage of anaerobic bacteria in the wound counts was 30. Propionihacterium spp. were found in 71 per cent of the wounds and anaerobic or microaerophilic cocci, most often Peptococcus spp., were found in 23 per cent. In six of 43 patients the same bacterial flora was seen in the skin samples and wounds. The geometric mean of the total number of bacteria in the air was 70.3 cfu;m”. Of these the median percentage of anaerobic bacteria was 30.3. When operating clothing of a disposable fabric (Barrier 45O’Q, Johnson & Johnson) was used, the counts of airborne bacteria were a little less than half those found when conventional cotton clothing was worn. Probably because of the overall low air counts in the operating theatre and the great variability in the individual bacterial counts from the operation wounds, a significant decrease in wound contamination was not observed. A positive correlation was found between the duration of operation and wound contamination.
Introduction Anaerobic skin bacteria of relatively low pathogenicity have been reported to cause postoperative infections after total hip replacement (Schwan et al., 1977; Carlsson, Josefsson and Ilindberg, 1978; Launder and Hungcrford, 1981; Fitzgerald et al., 1982). ‘l‘hese infections often become manifest weeks or months after the operation has taken place, and probably the importance of the low-pathogenic skin bacteria as infecting agents has been underestimated (Kamme and Lindberg, 1981; Whyte et al., 1981). ‘I’hc transmission routes of aerobic bacteria in clean surgery have been investigated by many authors. In major implant surgery airborne infection has been considered important (Charnley, 1972) and a controlled multicentre study has shown that operations performed in ultra-clean air have a significantly lower infection rate than those performed in conventionally ventilated operating rooms (1,idwell et al., 1982). Results of studies on airborne anaerobic bacteria have indicated the possibility of contamination of the operation wound with anaerobic bacteria from exogenous sources 01954701
c> 1984 The IIospital
84~1110038 1 12 SO2.OO;O
38
Infection
Society
Contamination
of clean
surgical
wounds
39
(Hambraeus and Benediktsdottir, 1980). However, studies on the occurrence of anaerobic bacteria in clean operation wounds have been sparse, and the importance of different factors in the reduction of wound contamination from anaerobic bacteria is not clear. In this study the influence of different clothing on the anaerobic and aerobic bacterial flora in the air and in the wound during total hip replacement was examined. Attempts were made to relate the wound contamination to the airborne bacteria, to the bacteria on the skin of the patients, to the duration of the operation and to the operative diagnosis. We also report on the types of anaerobic bacteria that were found to contaminate the operation wounds. Patients,
materials
and
methods
Putients Sixty-four patients undergoing total hip replacement were included in the investigation. The operations were performed at the Department of Orthopaedic Surgery of the TJppsala University Hospital from December 1981 to December 1982. The diagnoses were osteoarthritis in 47 patients, rheumatoid arthritis in nine, and previously surgically treated (on average 31 months earlier; range 2-- 84 months) fracture of the femoral neck m five patients. Added to the ‘fracture group’ were three further patients. Two were patients with previous infections-one with a septic coxitis leading to an arthrodesis operation some 22 years ago, and one who had had an infected pelvic fracture 1.5 years before the hip replacement. The third patient had developed necrosis of the femoral head after kidney transplantation. Thus the ‘fracture group’ consisted in all of eight patients. The age and sex distribution of the patients is shown in Figure 1. All patients received antibiotic prophylaxis, including cloxacillin infusion 2 g three times during the first 24 h, starting in the morning of the.day of surgery, followed by oral cloxacillin 1 g three times per day for about 10 days postoperatively. The operations were carried out by 12 different surgeons. The length of time between the skin incision and the last suture was recorded. The operating suite The design of the operating unit has been described in detail elsewhere (Hambraeus, Rengtsson and Laurell, 1977, 1978). ‘I’he operations were performed in an operating theatre that has a zonal ventilation system, where the main part of the filtered air is introduced above the operating table through a clean air inlet installed in the ceiling. About 80 air exchanges per hour are achieved in the central part of the room, which is protected from contamination from the peripheral room air by ‘air curtains’ (Abel and Allander, 1966). The peripheral zone is supplied with air at 2@ -30 changes/h.
E. Benediktsdhttir
20-29
40-49 30-39
Figure 1. Age and sex distribution females.
and
K. Kolstad
60-69 SO-89 50-59 70-79 90-99 Age
of the patients
included
in the investigation.
q , males; H ,
Be-operatice routines The patients were washed over the whole body with a detergent solution containing 4 per cent chlorhexidine (‘Hibiscrub’) the day before surgery. On the day of surgery the operation field was washed with ‘Hibiscrub’ then disinfected by applying 0.5 per cent chlorhexidine in 70 per cent ethanol, and finally covered with adhesive film surrounded by cotton drapes. Operating clothing Members of the surgical team scrubbed up with ‘Hibiscrub’ and dressed in sterile surgical gowns. The scrubbed staff wore double gloves when working, and all staff and visitors in the operating room wore caps and masks. Every other week the non-scrubbed members of the staff wore trousers manufactured from disposable non-woven fabric (Barrier 450@, Johnson & Johnson) and the scrubbed staff wore surgical gowns of this same fabric. During the alternate weeks the staff wore conventional cotton operating clothing and a gown. Sampling routines Wound sampling. The wound sampling was performed with six sterile pads of 85 per cent viscose and 15 per cent polyester and polyamide (Benediktsdottir and Hambraeus, 1983). The pads were 5 x 6 cm in size. The surgeon imprinted the pads for a few seconds on the subcutaneous tissue after suture of the fascia. If the tissue was dry it was moistened with a few millilitres of saline before sampling. The pads were placed in Petri plates containing 1O-l 5 ml of modified Stuart agar. ‘I’he plates were transported to the laboratory in an anaerobic jar after the operation.
Contamination
of clean
surgical
wounds
41
Skin sampling. Two swab samples and one skin biopsy were taken from the skin of each patient. The first swab sample was taken from the proposed incisional area before the final disinfection procedure and the second immediately after the operation from an area adjacent to the incision. A sterile cotton applicator was moistened in 2 ml of a neutralizing fluid containing 3 per cent lecithin and 20 per cent Tween 80 in nutrient broth. An area of about 1 x 10 cm was rubbed with the applicator for 10 s and the applicator was then put back into the neutralizing fluid and was transported after the operation to the laboratory. The biopsy, a few square millimetres, was taken from the incisional edge at the beginning of the operation and transported to the laboratory in 1 ml of the neutralizing fluid. Air sampling. The air sampling was performed with a Sartorius filter sampler with a gelatin filter (Ilambraeus and Benediktsdottir, 1980) and a capacity of 40 l/min. The filter holder was placed in the immediate vicinity of the wound and eight to 10 lo-min samples were taken at each operation. After the operation the filters were transported to the laboratory in the filter holders. Bacteriological methods ‘I’he wound sampling pads were handled with sterile forceps on a laminar flow bench. The technicians disinfected their bare arms with 70 per cent ethanol and wore sterile gloves while working. ‘I’he pads were cultured after treatment in a Stomach& lab blender (Colworth, Stomacher 400) as described earlier (Benediktsdottir and Hambraeus, 1983). One sterile pad was always processed with the wound pads for control of laboratory contamination. The skin swabs were cultured as follows. After shaking the tube on a whirlmixer for 15 s, the swab was pressed against the inside wall of the tube to remove the excess fluid and discarded. Ten-step dilutions were made in Nbroth containing 0.3 per cent lecithin and 2 per cent Tween 80. Of the sample 0.1 ml was plated on supplemented BHI agar (Iloldeman, Cato and Moore, 1977) and incubated anaerobically for 7 days. The skin biopsy was homogenized in 2 ml of the neutralizing fluid with a manual glass homogenizer, 1 ml was plated on supplemented BHI agar for anaerobic incubation for 7 days. The gelatin filters were placed on supplemented BHI agar and incubated anaerobically for 7 days. ‘I’he anaerobic bacteria were identified according to the method of Holdeman et al. (1977) as described earlier (Benediktsdottir and Hambraeus, 1982). The aerobic bacteria were only classified by colonial and cellular morphology. From the pour plates of the wound cultures, up to 20 colonies were analysed from each pad. The distribution of aerobic and anaerobic bacteria in the whole sample was assumed to be the same as that in the colonies that were analysed.
42
E. Benediktsdbttir
and
K. Kolstad
Results
Wound sampling Fifty-two operation wounds were sampled. Cultures from all wounds except one were positive. Eighty-six per cent of the control pads were negative, four pads grew one colony-forming unit (cfu), two pads 3 cfu, and one 14 cfu. ‘The total bacterial count from the wounds ranged from 0 to 47,747; the median was 26. Figure 2 shows the cumulative distribution of the wound counts. Two operations had counts between 1000 and 9999 and two had counts greater than 10,000.
Cumulative
Figure 2. Cumulative the air (0).
distribution
of the number
percentage
of bacteria
isolated
from the wounds
CM)and
Anaerobic bacteria were isolated from 40 of the 52 operation wounds sampled. The proportion of anaerobic bacteria among the total number isolated ranged from 0 to 100 per cent, the median was 30 per cent. More than 100 cfu of anaerobic bacteria were isolated from nine wounds and of aerobic bacteria from eight wounds. In Table I the types of anaerobic and aerobic bacteria isolated from the wounds are shown. Propionibacterium acnes was isolated from 37 wounds, and P. granulosum was found in two wounds, in one together with I’. avidum. Anaerobic Gram-negative rods that were not typed further were isolated from two wounds, both of which had total bacterial counts exceeding 900. Anaerobic and microaerophilic cocci were isolated from 12 wounds. The type most often found was Peptococcus magnus, which was isolated from four wounds. Other types isolated were Pepto. saccharolyticus, Pepto, asaccharolyticus and Veillonella paroula. Gram-positive cocci that were not further typed were isolated from four wounds. In two wounds with total bacterial counts exceeding 600, the organisms isolated were dominated by anaerobic cocci, one by Pepto. saccharolyticus and the other by V. paroula. Members of the Micrococcaccae were the most common bacteria isolated.
Contamination Table
I. Types
of clean
surgical
of hactevia isolated from the operating
zwunds
So. of wounds
Organisms Anaerobic rods P. acnes P. granulosum P. az:idunz Gram-negative
43
wounds
37 2 rods*
Anaerobic cocci Pepto. magnus Pepto. saccharolyticus Pepto. asaccharolyticus Veillonella parwla Gram-positive cocci* Aerobic rods Coryneforms,
22
Bacilli
Aerobic cocci Micrococcaceae
45
*So further identification. ‘I’he total number of operations was 52.
They were found in 45 wounds. streptococci in nine wounds.
Aerobic
rods were found in 22 wounds
and
Air sampling Air samples were taken during 56 operations. The total bacterial,,counts in the air ranged from 15 to 290/m3; the geometric mean was 70.3/m3. The proportion of anaerobic bacteria in the total bacterial count ranged from 1.5 per cent to 67.7 per cent, with a median of 30.3 per cent. Figure 2 shows the cumulative distribution of the air counts. Propionibacterium acnes was the predominant anaerobic species isolated during 55 operations. Propionibacterium granulosum was isolated in the course of eight operations, P. azidum in 11 and Gram-negative anaerobic rods in one operation. Anaerobic and microaerophilic cocci were isolated during six operations. Skin sampling All three types of skin samples and wound samples were taken from 43 patients. Thirty-six of the swab samples taken before the operation and 39 taken after the operation had bacterial counts from 0 to 9;O.l ml, and 38 of the skin biopsies taken had counts under 10 per 1 ml. Bacterial counts from 10 to 2230 were obtained from one or more of the skin samples in 14 patients. In five patients these counts were obtained only from the first swab sample, in three only from the second swab sample and in four only from the biopsy sample. The predominant bacteria isolated from the skin samples were aerobic cocci and coryneforms, and anaerobic propionibacteria. In two patients
44
anaerobic exceeded
E. Benediktsdbttir
and
K. Kolstad
cocci were isolated. In eight of the 14 samples with counts 10 the samples were dominated by anaerobic bacteria.
that
Helutionship between skin and wound samples So quantitative or qualitative relationship was found between the skin samples with counts less than 10 and the wound samples. Of the 14 patients with bacterial counts of 10 or more in the skin samples, a corresponding bacterial flora was found in the wound in six, in numbers from 86 to 702 per wound. In one patient the bacteria were isolated from the swab taken after the operation. In another patient from both swab samples, and in the remaining four patients from the biopsy sample. The bacterial types involved were Pepto. saccharolyticus in one patient, P. acnes in two, Micrococcaceae in two and a mixed flora of P. avidum and Micrococcaceae in one patient. Relationship between levels of airborne bacteria and wound contaminution Roth air and wound sampling was performed during 44 operations. Regression analysis was carried out on data from 34 operations. It was assumed that the contamination of wounds which had counts greater than 1000 was not derived from the air. These operations and the six in which the same bacteria were found on the skin as in the wound were not included. The data were transformed by log conversion to improve the normality. ‘I’he regression coefficient for logi (count/wound) on log,, (air count/m3) was 0.72 t- 3 .O. Comparison of dijfevent clothing Effect on airborne bacteria. The geometric mean of the total bacterial count was 101 cfu/m3 (N= 28) when the staff wore conventional cotton operating dress and 49 cfu/m3 (N=28) when the disposable clothing was used (P
Contamination
of clean
surgical
wounds
45
Relationship between duration of operation and zound contamination The length of operation ranged from 90 to 275 min, with a mean of 160 min. Fifteen of the operations took 3 h or more to perform. The median count in the wound samples was 125 per wound in these operations, and four of the six operations in which the same bacteria were isolated from the skin of the patient and the wound were included in this group. The median count in the wound samples from the 37 operations lasting less than 3 h was 11 per wound. A regression analysis was performed on data from all 52 operations. The data were transformed by log conversion to improve the normality. The regression coefficient for log,, (count/wound) on log,, [operation time (min)] was 2.61 + 2.67. Relationship between operatiae diagnosis and wound contamination The operation wounds were sampled in 39 patients undergoing surgery for osteoarthritis, and the median of the bacterial count per wound was 43. Samples were also taken from the operation wounds in six patients who were operated on for rheumatoid arthritis and seven patients from the ‘fracture group’. The median bacterial counts in these two groups were 16 and 5 per wound respectively. All four patients with a bacterial count in the wound exceeding 1000 were osteoarthritis patients undergoing operation for the first time, and so far none of them have shown any sign of clinical infection. Patients ‘l’he observation time varied from 1 to 11 months. During the hospital stay postoperatively (2-3 weeks) there were no deep infections. One patient developed a deep infection with group B streptococci following extraction of an infected tooth 3 months after the operation. ‘I’his patient had a total bacterial wound count of four. Another patient had a clinical infection after several re-operations performed because of prosthetic dislocation, and had a total bacterial wound count of 83, with 82 anaerobic bacteria. Cultures of the later infection however showed growth of penicillinase-producing staphylococci, Enterobacter aerogenes and enterococci.
Discussion
Several investigations have been carried out on the contamination of operation wounds in hip surgery. Positive cultures have been found by different authors in 1.5 per cent to 36 per cent of the wounds (Fitzgerald et al., 1973; Murray, 1973; Fitzgerald and Washington, 1975; Lindgren, Elmros and Holm, 1976; Bechtol, 1979; Ilamilton et al., 1979; Blomgren, IIambraeus and Malmborg, 1982). In this study all but one of the wound cultures were positive. Although laboratory contamination cannot be totally ruled out, the most likely explanation for the large number of positive
46
E. Benediktsdbttir
and
K. Kolstad
cultures is that a highly sensitive sampling method was used (Benediktsdottir and Hambraeus, 1983). Hitherto few attempts have been made to investigate the contamination of clean operation wounds by anaerobic bacteria. Fitzgerald and Washington (1975) isolated P. acnes in 24 and peptococci in two of 157 samples taken from operation wounds during orthopaedic surgery. Lindgren et al. (1976) did not isolate any anaerobic skin bacteria from the wound drain in hip surgery in spite of an anaerobic culture procedure. In this study anaerobic skin bacteria were found in 77 per cent of the wounds, and in 17 per cent of them anaerobic bacteria were grown in counts that exceeded 100. The types of anaerobic bacteria found both in the air and wound samples were mostly members of the normal anaerobic skin flora, Propionibacterium spp. and Peptococcus spp. Xone of the patients participating in this study developed a primary postoperative infection (i.e. during the hospital stay). In view of the small number of patients, the antibiotic prophylaxis and the short observation period, the clinical significance of these observations is uncertain. Contamination of operation wounds in clean surgery can take place in several different ways: from the air, through direct settling or indirectly via instruments and gloves; by direct contact from the clothes or from punctured gloves of the surgical team; and from the skin of the patient. Transmission of bacteria from the air to the wound occurs continuously during all operations. It is possible to estimate the proportion of bacteria settling directly in the wound. By using the same approach as was used by Whyte, Hodgson and Tinkler (1982) it can be calculated that when the total count in the air is 70.3 cfu/m3 (the geometric mean found in this study), the amount of bacteria settling on to the 180 cm2 wound will be 30 cfu. This is very close to the median number found in the wound in this investigation. The lowest and the highest counts in the air were 15 and 290 cfu/m3, respectively, and the calculated bacterial deposit in the wound area sampled would therefore range from 6 to 125 cfu. The total count of bacteria isolated from the wounds varied, however, from 0 to 47,747 cfu. It is obvious that in some of the wounds a break-up of bacterial clusters had occurred or the bacteria had been brought to the wound by other routes than direct deposition from the air. In this study a 5 1 per cent reduction in air contamination was found when the staff wore trousers of a non-woven fabric and the scrubbed staff also wore gowns of the same fabric, compared with that when all the staff wore clothing of conventional cotton. This is consistent with results from other investigations (Dineen, 1973; Mitchell, Evans and Keen, 1978; Whyte et al., 1978). The levels of both anaerobic and aerobic bacteria were reduced, a finding in accordance with the results of dispersal experiments in a test chamber, which showed that the dispersal of anaerobic bacteria from the human skin occurs mainly from the same region as that of aerobic bacteria, the lower trunk (Benediktsdottir and Hambraeus, 1982). Whyte et al. (1982) reported that a 99 per cent reduction in air contamination was associated with
Contamination
of clean
surgical
wounds
47
a 97 per cent reduction in wound contamination. In this study the variation in the individual bacterial counts from the air and wounds was such that the regression of the latter on the former did not reach significance at the 5 per cent level. The same was true for the comparison between the wound values with use of cotton and disposable clothing, although the reduction of the median wound contamination, 45 per cent, was almost identical to that of the geometric mean levels of air contamination, 51 per cent. In an operating theatre with efficient ventilation as that used in this study, there will only be a low degree of airborne contamination of the wounds. Prolongation of the operations and a higher level of airborne bacteria will naturally change this situation. The disinfection methods used in the present study are among the most effective methods for skin disinfection known (Lilly, Lowbury and Wilkins, 1979). However, contamination of the wound with bacteria presumably derived from the patient’s skin was found in six cases, or 11.5 per cent of the wounds. Probably contamination from the skin occurred to an even greater extent, as negative skin samples do not preclude the skin as a contaminating factor. ‘l’he skin sampling techniques used, swab and biopsy, give a limited picture of the skin-the former reveals only the surface flora (Evans and Stevens, 1976) and the latter is limited to very small areas. In hip surgery, wound contamination from the patient’s skin has been reported by other authors. Whyte et al. (1982) found wound contamination associated with exceptionally high patient skin counts in two of 52 operations, and McLauchlan et al. (1976) observed major contamination of wounds (more than 100 cfu(operation) in 11 per cent of patients operated on in a surgical isolator that blocked both the air and the contact routes of crossinfection. A positive correlation was found between duration of operation and wound contamination. Increased contamination from exogenous sources is to be expected with prolongation of the operation. In this study evidence was also found for greater transmission of bacteria from the patient’s skin when the operation time exceeded 3 h. ‘I’his is to be expected, and is probably due to loosening of drapes and moistening of clothing with blood and fluid, leading to bacterial invasion from the patient’s skin into the wound. It was suspected that previous surgery and infection in the hip area might increase the bacterial counts in the operation wounds compared with patients being operated on for the first time for osteoarthritis or perhaps rheumatoid arthritis. In the present study no difference was found in the levels of bacterial contamination of the wounds between patients with different operative diagnoses. l’he results of this study indicate that in about 20 per cent of the operations transmission occurs from the skin of the patient and possibly by contact with the operating team. This contamination seems to be of a high level. Whatever the route of transmission, anaerobic bacteria constitute a large proportion of the contamination. The results of this study clearly de-
48
E. Benediktsdbttir
and K. Kolstad
monstrate the importance of anaerobic cultivation procedures in hygienic studies of major implant surgery. It is also evident that when estimating the necessary standard of air hygiene at clean surgery such as hip joint replacement, the presence of airborne anaerobic as well as aerobic bacteria must be taken into consideration. References Abel,
E. & Allander, C. (1966). Undersijkning av nytt inblasningssystem fiir rcna rum. C’vs, so. 8. Bechtol, C. 0. (1979). Environmental bacteriology in the unidirectional (vertical) operating 114, 784-788. room. Archiz~s of Szqery Benediktsdbttir, E. & Hambraeus, A. (1982). Dispersal of non-sporeforming anaerobic bacteria from the skin. Journal of IIygiene 88, 487-500. Benediktsdbttir, I<. & Hambracus, A. (1983). Isolation of anaerobic and aerobic bacteria from clean surgical wounds. An experimental and clinical study. Journal of Hospital Infection 4, 141-148. Blomgren, G., Hambraeus, A. & Malmborg, A.-S. (1982). Hygienic and clinical study of elective and acute hip operations. Journal of Hospital Infection 3, 111-121. Carlsson, A. S., Josefsson, G. & Lindberg, I,. (1978). Revision with gentamicin-impregnated cement for deep infections in total hip arthroplasties. Journal of Bone andJoint Surgery 60-A, 1059-1064. Charnley, J. (1972), Postoperative infections after total hip replacement with special reference to air contamination in the operating room. Clinical Orthopaedics and Related Research 87, 167-187. Dineen, I’. (1973). The role of impervious drapes and gowns in preventing surgical infection. Clinical Orthopaedics and Related Research 96, 210-212. quantitation of surface and subsurface Evans, C. A. & Stevens, R. J. (1976). D’ff1 crcntial bacteria of normal skin by the combined use of the cotton swab and the scrub methods. 3, 576 581. Journal of Clinical Microbiology Fitzgerald, R. H., Jr, Peterson, L. F. A., Washington, J. A., van Scoy, Ii. E. &Coventry, M. B. (1973). Bacterial colonization of wounds and sepsis in total hip arthroplasty. Journal of Bone and Joint Surgery S-A, 1242-I 250. Fitzgerald, R. I-I., Jr, Rosenblatt, J. E., Tenney, J. H. & Bourgault, A.-M. (1982). Anaerobic septic arthritis. Clinical Orthopaedics and Related Research 164, 141-148. Fitzgerald, Ii. II., Jr. & Washington, J. A. (1975). Contamination of the operative wound. Orthopedic Clinics of ATorth America 6, 1105 1114. IIambraeus, A., Bengtsson, S. & Laurell, G. (1977). Bacterial contamination in a modern operating suite. 1. Effect of ventilation on airborne bacteria and transfer of airborne particles. Journal of Hygiene 79, 121- 132. Hambraeus, A., Bcngtsson, S. & Laurell, G. (1978). Bacterial contamination in a modern operating suite. 2. Effect of a zoning system on contamination of floors and other surfaces. Journal of Hygiene 80, 57-67. Hambraeus, A. & Benediktsdbttir, E. (1980). Airborne non-sporeforming anaerobic bacteria. Journal of Hygiene 84, 181-189. Hamilton, H. W., Booth, A. ID., Lone, I;. J. & Clark, N. (1979). Penetration of gown material by organisms from the surgical team. Clinical Orthopaedics and Related Research 141, 237-246. Manual. Holdernan, L. V., Cato, E. I’. & Moore, W. E. C. (cds) (1977). A naerobe Laboratory Anaerobe T,aboratory, Virginia Polytechnic Institute and State University, Blacksburg. after Kamme, C. & T,indberg, I,. (1981). A erobic and anaerobic bacteria in deep infections total hip arthroplasty. Clinical Orthopaedics and Related Research 154, 201-207. Launder, W. J. & Hungerford, I>. S. (1981). Late infection of total hip arthroplasty with Propionibacterizlm acnes: A case and review of the literature. Clinical Orthopaedics and Related Research 157, 170-177. Lidwell, 0. M., Lowbury, E. J. L., Whyte, W., Blowers, R., Stanley, S. J. & I,owe, L). (1982). Effect of ultraclean air in operating rooms on deep sepsis in the joint after total hip or knee replacement: a randomized study. Ijritish Medical Journal 285, 10 -14.
Contamination Lilly,
of clean
surgical
wounds
49
H. A., I,owbury, E. J. I,. & Wilkins, M. D. (1979). Limits to progressive reduction of resident skin bacteria by disinfection. ~ournnl of Clinical Pathology 32, 382-383. Lindgren, IJ., I’lmros, ‘I’. & Holm, S. E. (1976). Bacteria in hip surgery. Acta Orthopaedica SEandinaCa 47, X20-323. McLauchlan, J., I,ogie, J. Ii. C., Smylie, II. G. & Smith, G. (1976). A study of the wound environment during total hip arthroplasty. Postgraduate Medical journal 52, 55&557. Mitchell, S. J., Evans, 1). S. & Kerr, A. (1978). Reduction of skin bacteria in theatre air with comfortable, non-woven disposable clothing for operating-theatre staff. IXtish Medicnl Journal 1, 696498. Murray, W. I<. (1973). Results in patients with total hip replacement arthroplasty. Clinical Ovthopaedics and Related Research 95, 80-90. Schwan, A., Bengtsson, S., IIambraeus, A. & Laurell, G. (1977). Airborne contamination and postoperative infection after total binI renlacement. Acta Ortho~aedica Scandintzeica I 48, 86 94. Whyte, W., Hodgson, Ii., Bailey, I’. V. & Graham, J. (1978). ‘The reduction of bacteria in the operation room through the use of non-woven clothing. British Journal of Surgery 65, 469474. Whyte, W., IIodgson, I<. & Tinkler, J. (1982). ‘I’he importance of airborne bacterial contamination of wounds. Journal of hospital Infection 3, 123-135. Whyte, W., IIodgson, IX., Tinkler, J. & Graham, J, (1981). Th e isolation of bacteria of low pathogenicity from faulty orthopaedic implants. Journal of Hospital Infection 2,219-230.