- Email: [email protected]
Causes and prevention—(ii) bacteriology
Causes and Prevention
(ii) Bacteriology
G. C. B a n n i s t e r
that subsequently involve the joint. The coagulase negative staphylococcus is a much more frequent skin commensal. It has the capacity to acquire multiple resistance and to colonise prosthetic heart valves when the bacterial ecology has been disturbed. Resistance to antibiotics is by no means the sole province of cardiothoracic surgery and paediatric intensive care units. In 1981 Whyte reported the bacteriology of 51 cases of infected joint replacements in Glasgow of which only 17% yielded organisms normally regarded as pathogenic and 82% those considered to be commensal. 2 Of the total, 45% were staphylococcus epidermidis of which 78% were resistant to penicillin.
The Organisms Failure of a joint replacement as a result of infection has declined from 9% in the early '70s to approximately 1% at ~ present. The majority of cases present within two years of surgery appearing sporadically thereafter. The infection involves the bone-cement interface. A broad spectrum of bacteria are isolated from deep joint infections (Table 1). They comprise predominantly staphylococci, the remainder being skin and bowel organisms. Whilst the proportion of most skin and faecal flora has remained constant there has been a gradual decline of staphylococcus aureus and an emergence of coagulase negative organisms in cultures over the last 15 years. The staphylococcus aureus has a predilection for rheumatoids and for causing major wound infections
Source of Organisms Organisms infecting joint replacements come from the operating theatre, the wound and distant foci via the blood system in that order. The most accurate method of localising the source of bacteria is the correlation of cultures taken from the area under suspicion with the isolates from subsequent deep joint infection. In the Medical Research Council (MRC) trial on ultraclean air, it was possible to identify similar organisms to those cultured from the theatre where the operation was performed in 86% of staphylococcal infections by phage typing. 3 In that study some 24% of deep joint infections were associated with prior wound infection. 4 Thirty percent of major wound infections culturing staphylococci subsequently manifested as deep infection. Of staphylococcal joint infections 79% demonstrated the same strain as found in the wound and 86% of these organisms could be traced to a source in theatre. Whilst there was a strong association between theatre staphylococcal wound infection and joint sepsis, overall less than half of all organisms isolated from wounds could be correlated with a theatre source
1--% Organisms Isolated from Medical Research Council Trial on Ultraclean Air Table
94 cultures from 86 joints
% 30
Staphylococcus Aureus Other Skin Organisms
Faecal Organisms
Coagulase -ve staphylococci Pripionibacter spp Corynebacteria Gram positive cocci Anaerobic cocci Group B streptococci Mixed Strep. E. Colifaecalis Pseudomonas spp Alkaligenes faecalis Proteus spp Mixed Flora
111 29 iJ
i1
13 8
Gordon C. Bannister, MCh.Orth, FRCS, FRCSEd.(Orth) Lecturer in Orthopaedics, Department of Orthopaedics, Bristol Royal Infirmary, Bristol. Current Orthopaedics (1986)1,35-39 © 1986LongmanGroupUKLtd
35
36
CAUSESAND PREVENTION--(ii) BACTERIOLOGY
and thus were acquired between leaving the operating room and wound healing. Haematogeneous spread from other septic foci is well described. A wide range of organisms have been isolated but the single most potent source appears to be from the _urinary tract in post-operative retention. Wroblewski reported a 6.2% infection rate from such patients at Wrightington when the concurrent incidence of deep sepsis was 0.5%. Transient asymptomatic bacteraemia has been reported in 6.8% of blood donors but this is a theoretical mechanism.
Culture Techniques Once infected, the septic joint replacement can elude accurate culture if sampling is limited to a superficial swab whilst the patient is currently taking antibiotics. The best returns from the laboratory come when antimicrobial chemotherapy has been stopped for 2 days, several biopsies of l g of tissue are taken from the area of active infection, and these are rapidly transported to the laboratory in culture medium. The samples should then be homogenised and set up for aerobic and anaerobic culture. Gram staining gives returns of only 20% in the most optimistic series and anaerobic cultures continue to grow and should be left for 5 days.
Virulence of Organisms The relative virulence of infecting organisms can be inferred from Fitzgerald's studies at the Mayo Clinic 6'7 (Table 2). He was able to obtain positive cultures from the operation sites of 30% of patients undergoing joint replacement in conventional theatres with 12-30 air exchanges per hour. The staphylococcus aureus was not isolated from theatre air and in only 1% of wound cultures, yet was responsible for 24% of infections in a subsequent series of hip replacements from the same unit. Gram negative organisms expressed at a slightly higher rate of deep infection than their appearance in theatre whilst skin flora were less aggressive. The large numbers of coagulase negative staphylococci in the theatre appear to have governed their prevalence in joint infections.
Inoculum and Host Resistance The capacity to infect a joint replacement is a function of the inoculum of bacteria and the ability of the host to resist them. From bacterial counts in operating Table 2--Relationship of Theatre and Wound Organisms to Deep Joint Infection (%)
Table ~Inoculum of Bacteria Required to Infect a Cemented Arthroplasty
Species
Type of Inoculation
InoculumSize
Man Rabbit Rabbit Rabbit
Intraoperative Intraoperative Immediately Postoperative After wound healing
10 50 100000 380 000
theatre air Charnley and Lidwell separately concluded that fewer than 10 organisms are required to infect a human joint replacement 3'8 (Table 3). In rabbit experimental arthroplasty, Southwood found that less than 50 bacteria would infect a joint if given locally intraoperatively. 9 An intravenous bolus of 100000 organisms was required before sepsis ensued in the immediate post-operative period, whilst after wound healing had taken place 380 000, a near lethal dose, was needed. The cemented arthroplasty is the one constant to which the majority of patients are subjected. Charnley noted a wide variation in infection rates between hip arthrodeses and arthroplasties. 8 Although he did not incriminate acrylic cement, Elson demonstrated its potent adjuvant effect in experimental burr holes in rat tibiae. ~° Salvage of infected arthroplasties has been more effective in both knee and hip arthroplasties when they have been uncemented. Other variables that appear to depress the capacity of the host to resist infection are rheumatoid arthritis and obesity severe enough to be recorded in hospital records (Table 4). Wound infection, particularly with Table 4~-Factors Increasingthe Risk of Deep Joint Infection
Rheumatoid Arthritis Obesity Non-staphylococcalmajor wound sepsis Wound haematoma Staphylococcalminor wound sepsis Diabetes mellitus Revision surgery Constrained knee arthroplasty Concurrent urinary retention Staphylococcalmajor wound sepsis
Factor 2 2 2 2.5 3 3.5 4 5 6 10
staphylococci and haematomata, increase sepsis by a factor of at least two. Hip replacement in diabetics is more hazardous and previous surgery carries a fourfold greater risk of infection in a knee arthroplasty. A series from a secondary referral clinic in Boston records a twenty-fold increased risk of infection after constrained knee arthroplasty but the literature would suggest considerably less. Methods of Prevention
Wound Joint Air Culture Culture Staphylococcusaureus
0
1
24
Other Coagulase -re staphylococci 42 Skin Corynebacteria 30 Organisms Streptococcus viridans 12
36 7 4
20 0 12
Faecal Gram -re organisms Organisms Fungi
4 18
18 0
12 3
Infection of joint replacements can be prevented by reducing the inoculum of bacteria, creating an inhospitable environment with antibiotics, or both.
Theatre Air Whilst spores measure 1/, the vast majority of bacteria are carried by dust or scales of skin measuring 10t*. The
CURRENT ORTHOPAEDICS scales are spread by theatre personnel and five members of staff in a conventional operating theatre increase colony counts by a factor of 34. The most potent spreaders of scales and bacteria are excessive skin 'shedders'. Thirteen percent of these are male, 5% post-menopausal and 1% pre-menopausal women. Skin disorders such as eczema result in high shedding rates but screening requires time-consuming enclosure techniques. Shedders can reduce their output by application of alcohol-based lotions. The organisms dispersed are representative of body flora. These are located in the outer 0.3mm of skin, and can be reduced by regular bathing with antiseptic soap. The staphylococcus aureus forms a small percentage of skin organisms and can normally only be isolated from the perineum. However 33% of theatre personnel are nasal carriers of the organism. They are no more frequent amongst surgeons than patients or other theatre staff and surgeons are no more likely to infect patients than anyone else in the operating theatre. Dispersal of organisms can be reduced by wearing theatre gowns and nasopharyngeal bacteria redirected away from the wound by wearing a mask. Charnley reduced the infection rates in hip arthroplasty at Wrightington from 7-0.5% by a variety of measures to limit bacteria in theatre air. 8 The large numbers required to compare his techniques were amassed in the M R C trial on ultraclean air. ~ In this study conventional opeating theatres had a mean 164 colony forming particles per m 3 of air, body exhaust suits in a conventional theatre 50 and vertical laminar flow and plastic isolates less than 10. The study incorporated a number of variables but the randomised comparison was of ultraclean air and conventional operating theatres. O f over 8000 joint replacements 3.4% performed in conventional theatres became septic compared with 1.2% of those implanted in ultraclean air. The additions of other measures to reduce colonisation were additive and a body exhaust suit further reduced infection rates to less than 0.5% (Table 5). Table
~Measures Reducing the Risk of Deep Joint Infection
Gentamicin cement Systemic antibiotics Ultraclean air Plastic isolators Body exhaust systems
Factor 11 4.8 2.6 2.2 2.2
A cautionary note concerning the use of horizontal laminar flow is made by Salvati wh~o\recorded double the number of infected knee i m p l a n t s ~ m p a r e d with those performed in conventional theatre~suggesting that the interposition of operating personnel between the air flow and the wound may have caused an increase in bacterial contamination. 11
Antibiotic Prophylaxis Surgical incisions undergo a period during which they are particularly vulnerable to bacterial colonisation.
37
Fitzgerald's studies demonstrate that this occurs whilst the wound is open 7 and Elson, in the rat tibia, l° and Southwood, in experimental rabbit arthroplasty, confirm that the period extends for some time after wound closure. The principle that antibiotics given before this period could reduce infection was established by Burke in experimental skin lesions. 12 He found that preoperative antibiotics reduced the area of inflammation whereas when agents were given 3 h after the lesion was fashioned no benefit accrued. W h e n this principle was applied to orthopaedic surgery Burke reduced the infection rate in hip fractures by a factor of 6 and similar results were achieved in total hip replacement. 13 It appears that the vulnerable period in human wounds lasts for about 24 h and that antibiotic cover should continue for this period. The evidence for this derives from a number of randomised prospective controlled trials with variable lengths of follow-up. Using cefamandole the Barcelona group found one dose to carry 10 times the risk of wound infection of five doses in internally fixed fractures. Pollard demonstrated that 12 h of cephaloridine was as effective as 14 days of flucloxacillin, 14 whilst Nelson confirmed that 1 day of nafcitlin and cefazolin was as effective as 7 days. is The Paris group compared 2 days of cefamandole with 5 of cefazolin with no infections recorded in 966 arthroplasties after 2 years. The two studies that demonstrated the benefit of prophylactic antibiotics in total hip replacement were multicentre based on Lund 16 and the H6pital Cochin, Paris 17 (Table 6). The Swedish study used cloxacillin in the early days of high infection rates. They demonstrated the increasing benefit of perioperative antibiotics with the passage of time reducing infection by a factor of 8. The Paris study used the broader spectrum cefazolin with a m u c h greater number of patients. Their treated group retained a constant infection rate of 0.9%, whilst the control group manifested increasing numbers of infections over the 6-year follow-up period. The second Swedish study compared gentamicin cement with cloxacillin and cephalaxin, demonstrating a fourfold reduction in deep infection rates although there was a slight but significantly higher incidence of superficial wound infection in the gentamicin group, is They achieved infection rates of 0.4% in conventional operating theatres after a 2-year followup. The difference was highly significant. This finding is not isolated. The infection rate of patients receiving gentamicin cement was 0.4% in the M R C trials whilst Murray achieved similar results using erythromycin and colistin mixed with methylmethacrylate.19 There are three essential differences between these methods of antibiotic delivery. The antibiotic impregnated cements are broad spectrum, they achieve high concentrations at the bone-cement interface where infection occurs and give prolonged action at this site. The addition of 1 g of antibiotic powder to a 40 g packet of cement has a negligible effect on the strength of the methylmethacrylate cement.
38
CAUSES AND PREVENTION
(ii) BACTERIOLOGY
Table ~-Controlled Trials on Prophylactic Antibiotics in Hip Replacement
Centre
Number of cases
Lurid
118
Paris
2137
Length of Administration
Drug Cloxacillin vs Control
14 days
Cefazolin
5 days
% Infected Mean Follow-up 2 yrs 6 yrs 0
3
12
23
0.9
0.9
Control
3.4
4.6
Gentamicin Cement
0.4
Ratio 8
5
VS
Lund
1685
4
VS
Cloxacillin or Cephalexin
Choice of Antibiotic T h e present practice is to use a systemic a gen t that has g o o d activity against the s t a p h y l o c o c c u s aureus. This m a y be a methicillin related penicillin or a 1st or 2nd g e n e r a t i o n c e p h a l o s p o r i n . T h i r d g e n e r a t i o n c e p h a l o s p o r i n s have a c q u i r e d activity against g r a m negative o r g a n i s m s at the expense o f s t a p h y lo c o cci and are n o t generally used. T h e evidence indicates that the s p e c t r u m o f infecting o r g a n i s m s is b r o a d a n d that the c e m e n t v e c t o r is superior to systemic a d m i n i s t r a t i o n . Systemic and c e m e n t - i n c o r p o r a t e d antibiotics are m i n i m a l l y additive. A n y a n t i b i o t i c a d d e d to c e m e n t m u s t be heat stable. It w o u l d a p p e a r logical to use antibiotic l o a d e d c e m e n t with an a n t i s t a p h y l o c o c c a l systemic agent to reduce w o u n d infection. T h e period o f a d m i n i s t r a t i o n o f the latter need n o t exceed 24 h.
Cost of Prevention Lidwell gives an a c c o u n t o f the cost o f p r e v e n t i t i v e measures using Swedish records o f 90 extra hospital days an d an a v e r a g e o f 1.6 extra o p e r a t i o n s per infected case. z° These are c r u d e ly calculated to cost £14 000 per case a n d L i d w e l l' s figures are modified to include g e n t a m i c i n c e m e n t (Table 7). It is a s s u m e d that antibiotics cost an extra £30 per patient, that ultraclean air can be p r o v i d e d by a plastic isolator at a p p r o x i m a t e l y £50 per case and that disposable suits costing £20 per case can be used in theatres p e r f o r m i n g fewer than 50 a r t h r o p l a s t i e s per a n n u m . In o p e r a t i n g r o o m s with a greater l o a d the installation o f ultraclean air w o u l d cost a b o u t £2500 a n d b o d y exhaust suits a b o u t £1000 per a n n u m , b o t h including m a i n t e n a n c e and w e a r an d tear. T h e costs o f clean air w o u l d fall in units with a higher v o l u m e o f cases.
1.6
It can be seen that the single m o s t cost-effective m e a s u r e is antibiotic prophylaxis. T h e isolated cost o f clean air w o u l d be repaid in high v o l u m e units b u t even in small centres the overall saving o f a p p l y i n g all these m e a s u r e s w o u l d a m o u n t to over £3000 per case. W i t h the additive effects o f antibiotics and clean air infection rates well u n d e r 0.2% should be achieved.
References i. Lidwell O M, Lowbury E J L, Whyte W, Blowers R, Stanley S J, Lowe D 1982 Effect of ultraclean air in operating rooms on deep sepsis of the joint after total hip or knee replacement: A randomised study. British Medical Journal. 285:1(~14 2. Whyte W, Hodgson R, Tinkler J, Graham J 1981 The isolation of bacteria of low pathogenicity from faulty orthopaedic implants. Journal of Hospital Infection. 2:219-230 3. Lidwell O M, Lowbury E J L, Whyte W, Blowers R, Stanley S J, Lowe D 1983 Bacteria isolated from deep joint sepsis after operation for total hip or knee replacement and the sources of the infections with staphylococcus aureus. Journal of Hospital Infection. 4:19-29 4. Lidwell O M, Lowbury E J L, Whyte W, Blowers R, Stanley S J, Lowe D 1984 Infection and sepsis after operations for total hip or knee-joint replacement: influence of ultraclean air, prophylactic antibiotics and other factors. Journal of Hygiene.
93:505-529 5. Wroblewski B M, De1 Sel H J 1980 Urethral instrumentation and deep sepsis in total hip replacement. Clinical Orthopaedics. 146:209-212 6. Fitzgerald R H, Nolan D R, Ilstrup D M, Van Scoy R E, Washington J AII, Coventry M B 1977 Deep wound sepsis following total hip arthroplasty. Journal of Bone and Joint Surgery. 59A: 847-855 7. Fitzgerald R H, Peterson L F A, Washington J A, Van Scoy R E, Coventry M B 1973 Bacterial colonisation of wounds and sepsis in total hip arthroplasty. Journal of Bone and Joint Surgery. 55A: 1242-1250 8. Charnley J, Eftekhar N 1969 Post-operative infection in total prosthetic replacement arthroplasty of the hip joint. British Journal of Surgery. 56:641-649
Table 7 Costings of Infected Joint Replacement per 100 Cases in £
Type of Theatre Conventional Gentamicin cement and antibiotics + Ultraclean air + Body exhaust suits
Infection Rate %
Cost of Treatment
Cost of Prevention
Financial Savings of Prevention
3.4 0.4
44 200 5 200
Nil 3 000
Nil 36000
0.15
1950
5 000
34 250
0.06
780
2 000
33420
CURRENT ORTHOPAEDICS 9. Southwood R T, Rice J L, McDonald P J, Hakendorf P H, Rozenbilds M A 1985 Infection in experimental hip arthroplasties. Journal o f Bone and Joint Surgery. 67B: 229-231 10. Elson R A, Jephcott A E, McGechie D B, Veretta D 1977 Bacterial infection and acrylic cement in the rat. Journal of Bone and Joint Surgery. 59B: 452-457 11. Salvati E A,. Robinson R P, Zeno S M, Koslin B L, Brause B D, Wilson P D 1982 Infection rates after 3175 total hip and total knee replacements performed with and without a horizontal unidirectional air-flow system. Journal of Bone and Joint Surgery. 64A: 525-535 12. Burke J F 1961 The effective period ofpreventitive antibiotic action in experimental incisions and dermal lesions. Surgery. 50:161-168
13. Boyd R J, Burke J F, Colton T 1973 A double-blind clinical trial of prophylactic antibiotics in hip fractures. Journal of Bone and Joint Surgery. 55A: 1251-1258 14. Pollard J P, Hughes S P F, Scott J E, Evans M J, Benson M K D 1979 Antibiotic prophylaxis in total hip replacement. British Medical Journal. 707-709 15. Nelson C L, Green T G, Porter R A, Warren R D 1983 One day
16.
17.
18.
19. 20.
39
versus 7 days of preventitive antibiotic therapy in orthopaedic surgery. Clinical Orthopaedics and related research. 176: 258-263 Ericson C, Lidgren L, Lindberg L 1973 Cloxacillin in prophylaxis of post-operative infection of the hip. J ournal of Bone and Joint Surgery. 55A: 808-813 Hill C, Mazas F, Flamant R, Evrard J 1981 Prophylactic Cefazolin versus placebo in total hip replacement. Lancet. 795-797 Josefsson G, Lindberg L, Wiklander B 1981 Systemic antibiotics and gentamicin-containing bone cement in the prophylaxis of post-operative infections in total hip arthroplasty. Clinical Orthopaedics and related research. 159:[94-200 Murray W R 1984 Use of antibiotic-containing bone cement. Clinical Orthopaedics and related research. 190:89-95 Lidwell O M 1984 The cost implications of clean air systems and antibiotic prophylaxis in operations for total joint replacement. Infection Control. 5:36-37
Address for reprints: Gordon C. Bannister, Lecturer in Orthopaedics, Bristol Royal Infirmary, Bristol BS2 8HW.
(ii) Bacteriology
G. C. B a n n i s t e r
that subsequently involve the joint. The coagulase negative staphylococcus is a much more frequent skin commensal. It has the capacity to acquire multiple resistance and to colonise prosthetic heart valves when the bacterial ecology has been disturbed. Resistance to antibiotics is by no means the sole province of cardiothoracic surgery and paediatric intensive care units. In 1981 Whyte reported the bacteriology of 51 cases of infected joint replacements in Glasgow of which only 17% yielded organisms normally regarded as pathogenic and 82% those considered to be commensal. 2 Of the total, 45% were staphylococcus epidermidis of which 78% were resistant to penicillin.
The Organisms Failure of a joint replacement as a result of infection has declined from 9% in the early '70s to approximately 1% at ~ present. The majority of cases present within two years of surgery appearing sporadically thereafter. The infection involves the bone-cement interface. A broad spectrum of bacteria are isolated from deep joint infections (Table 1). They comprise predominantly staphylococci, the remainder being skin and bowel organisms. Whilst the proportion of most skin and faecal flora has remained constant there has been a gradual decline of staphylococcus aureus and an emergence of coagulase negative organisms in cultures over the last 15 years. The staphylococcus aureus has a predilection for rheumatoids and for causing major wound infections
Source of Organisms Organisms infecting joint replacements come from the operating theatre, the wound and distant foci via the blood system in that order. The most accurate method of localising the source of bacteria is the correlation of cultures taken from the area under suspicion with the isolates from subsequent deep joint infection. In the Medical Research Council (MRC) trial on ultraclean air, it was possible to identify similar organisms to those cultured from the theatre where the operation was performed in 86% of staphylococcal infections by phage typing. 3 In that study some 24% of deep joint infections were associated with prior wound infection. 4 Thirty percent of major wound infections culturing staphylococci subsequently manifested as deep infection. Of staphylococcal joint infections 79% demonstrated the same strain as found in the wound and 86% of these organisms could be traced to a source in theatre. Whilst there was a strong association between theatre staphylococcal wound infection and joint sepsis, overall less than half of all organisms isolated from wounds could be correlated with a theatre source
1--% Organisms Isolated from Medical Research Council Trial on Ultraclean Air Table
94 cultures from 86 joints
% 30
Staphylococcus Aureus Other Skin Organisms
Faecal Organisms
Coagulase -ve staphylococci Pripionibacter spp Corynebacteria Gram positive cocci Anaerobic cocci Group B streptococci Mixed Strep. E. Colifaecalis Pseudomonas spp Alkaligenes faecalis Proteus spp Mixed Flora
111 29 iJ
i1
13 8
Gordon C. Bannister, MCh.Orth, FRCS, FRCSEd.(Orth) Lecturer in Orthopaedics, Department of Orthopaedics, Bristol Royal Infirmary, Bristol. Current Orthopaedics (1986)1,35-39 © 1986LongmanGroupUKLtd
35
36
CAUSESAND PREVENTION--(ii) BACTERIOLOGY
and thus were acquired between leaving the operating room and wound healing. Haematogeneous spread from other septic foci is well described. A wide range of organisms have been isolated but the single most potent source appears to be from the _urinary tract in post-operative retention. Wroblewski reported a 6.2% infection rate from such patients at Wrightington when the concurrent incidence of deep sepsis was 0.5%. Transient asymptomatic bacteraemia has been reported in 6.8% of blood donors but this is a theoretical mechanism.
Culture Techniques Once infected, the septic joint replacement can elude accurate culture if sampling is limited to a superficial swab whilst the patient is currently taking antibiotics. The best returns from the laboratory come when antimicrobial chemotherapy has been stopped for 2 days, several biopsies of l g of tissue are taken from the area of active infection, and these are rapidly transported to the laboratory in culture medium. The samples should then be homogenised and set up for aerobic and anaerobic culture. Gram staining gives returns of only 20% in the most optimistic series and anaerobic cultures continue to grow and should be left for 5 days.
Virulence of Organisms The relative virulence of infecting organisms can be inferred from Fitzgerald's studies at the Mayo Clinic 6'7 (Table 2). He was able to obtain positive cultures from the operation sites of 30% of patients undergoing joint replacement in conventional theatres with 12-30 air exchanges per hour. The staphylococcus aureus was not isolated from theatre air and in only 1% of wound cultures, yet was responsible for 24% of infections in a subsequent series of hip replacements from the same unit. Gram negative organisms expressed at a slightly higher rate of deep infection than their appearance in theatre whilst skin flora were less aggressive. The large numbers of coagulase negative staphylococci in the theatre appear to have governed their prevalence in joint infections.
Inoculum and Host Resistance The capacity to infect a joint replacement is a function of the inoculum of bacteria and the ability of the host to resist them. From bacterial counts in operating Table 2--Relationship of Theatre and Wound Organisms to Deep Joint Infection (%)
Table ~Inoculum of Bacteria Required to Infect a Cemented Arthroplasty
Species
Type of Inoculation
InoculumSize
Man Rabbit Rabbit Rabbit
Intraoperative Intraoperative Immediately Postoperative After wound healing
10 50 100000 380 000
theatre air Charnley and Lidwell separately concluded that fewer than 10 organisms are required to infect a human joint replacement 3'8 (Table 3). In rabbit experimental arthroplasty, Southwood found that less than 50 bacteria would infect a joint if given locally intraoperatively. 9 An intravenous bolus of 100000 organisms was required before sepsis ensued in the immediate post-operative period, whilst after wound healing had taken place 380 000, a near lethal dose, was needed. The cemented arthroplasty is the one constant to which the majority of patients are subjected. Charnley noted a wide variation in infection rates between hip arthrodeses and arthroplasties. 8 Although he did not incriminate acrylic cement, Elson demonstrated its potent adjuvant effect in experimental burr holes in rat tibiae. ~° Salvage of infected arthroplasties has been more effective in both knee and hip arthroplasties when they have been uncemented. Other variables that appear to depress the capacity of the host to resist infection are rheumatoid arthritis and obesity severe enough to be recorded in hospital records (Table 4). Wound infection, particularly with Table 4~-Factors Increasingthe Risk of Deep Joint Infection
Rheumatoid Arthritis Obesity Non-staphylococcalmajor wound sepsis Wound haematoma Staphylococcalminor wound sepsis Diabetes mellitus Revision surgery Constrained knee arthroplasty Concurrent urinary retention Staphylococcalmajor wound sepsis
Factor 2 2 2 2.5 3 3.5 4 5 6 10
staphylococci and haematomata, increase sepsis by a factor of at least two. Hip replacement in diabetics is more hazardous and previous surgery carries a fourfold greater risk of infection in a knee arthroplasty. A series from a secondary referral clinic in Boston records a twenty-fold increased risk of infection after constrained knee arthroplasty but the literature would suggest considerably less. Methods of Prevention
Wound Joint Air Culture Culture Staphylococcusaureus
0
1
24
Other Coagulase -re staphylococci 42 Skin Corynebacteria 30 Organisms Streptococcus viridans 12
36 7 4
20 0 12
Faecal Gram -re organisms Organisms Fungi
4 18
18 0
12 3
Infection of joint replacements can be prevented by reducing the inoculum of bacteria, creating an inhospitable environment with antibiotics, or both.
Theatre Air Whilst spores measure 1/, the vast majority of bacteria are carried by dust or scales of skin measuring 10t*. The
CURRENT ORTHOPAEDICS scales are spread by theatre personnel and five members of staff in a conventional operating theatre increase colony counts by a factor of 34. The most potent spreaders of scales and bacteria are excessive skin 'shedders'. Thirteen percent of these are male, 5% post-menopausal and 1% pre-menopausal women. Skin disorders such as eczema result in high shedding rates but screening requires time-consuming enclosure techniques. Shedders can reduce their output by application of alcohol-based lotions. The organisms dispersed are representative of body flora. These are located in the outer 0.3mm of skin, and can be reduced by regular bathing with antiseptic soap. The staphylococcus aureus forms a small percentage of skin organisms and can normally only be isolated from the perineum. However 33% of theatre personnel are nasal carriers of the organism. They are no more frequent amongst surgeons than patients or other theatre staff and surgeons are no more likely to infect patients than anyone else in the operating theatre. Dispersal of organisms can be reduced by wearing theatre gowns and nasopharyngeal bacteria redirected away from the wound by wearing a mask. Charnley reduced the infection rates in hip arthroplasty at Wrightington from 7-0.5% by a variety of measures to limit bacteria in theatre air. 8 The large numbers required to compare his techniques were amassed in the M R C trial on ultraclean air. ~ In this study conventional opeating theatres had a mean 164 colony forming particles per m 3 of air, body exhaust suits in a conventional theatre 50 and vertical laminar flow and plastic isolates less than 10. The study incorporated a number of variables but the randomised comparison was of ultraclean air and conventional operating theatres. O f over 8000 joint replacements 3.4% performed in conventional theatres became septic compared with 1.2% of those implanted in ultraclean air. The additions of other measures to reduce colonisation were additive and a body exhaust suit further reduced infection rates to less than 0.5% (Table 5). Table
~Measures Reducing the Risk of Deep Joint Infection
Gentamicin cement Systemic antibiotics Ultraclean air Plastic isolators Body exhaust systems
Factor 11 4.8 2.6 2.2 2.2
A cautionary note concerning the use of horizontal laminar flow is made by Salvati wh~o\recorded double the number of infected knee i m p l a n t s ~ m p a r e d with those performed in conventional theatre~suggesting that the interposition of operating personnel between the air flow and the wound may have caused an increase in bacterial contamination. 11
Antibiotic Prophylaxis Surgical incisions undergo a period during which they are particularly vulnerable to bacterial colonisation.
37
Fitzgerald's studies demonstrate that this occurs whilst the wound is open 7 and Elson, in the rat tibia, l° and Southwood, in experimental rabbit arthroplasty, confirm that the period extends for some time after wound closure. The principle that antibiotics given before this period could reduce infection was established by Burke in experimental skin lesions. 12 He found that preoperative antibiotics reduced the area of inflammation whereas when agents were given 3 h after the lesion was fashioned no benefit accrued. W h e n this principle was applied to orthopaedic surgery Burke reduced the infection rate in hip fractures by a factor of 6 and similar results were achieved in total hip replacement. 13 It appears that the vulnerable period in human wounds lasts for about 24 h and that antibiotic cover should continue for this period. The evidence for this derives from a number of randomised prospective controlled trials with variable lengths of follow-up. Using cefamandole the Barcelona group found one dose to carry 10 times the risk of wound infection of five doses in internally fixed fractures. Pollard demonstrated that 12 h of cephaloridine was as effective as 14 days of flucloxacillin, 14 whilst Nelson confirmed that 1 day of nafcitlin and cefazolin was as effective as 7 days. is The Paris group compared 2 days of cefamandole with 5 of cefazolin with no infections recorded in 966 arthroplasties after 2 years. The two studies that demonstrated the benefit of prophylactic antibiotics in total hip replacement were multicentre based on Lund 16 and the H6pital Cochin, Paris 17 (Table 6). The Swedish study used cloxacillin in the early days of high infection rates. They demonstrated the increasing benefit of perioperative antibiotics with the passage of time reducing infection by a factor of 8. The Paris study used the broader spectrum cefazolin with a m u c h greater number of patients. Their treated group retained a constant infection rate of 0.9%, whilst the control group manifested increasing numbers of infections over the 6-year follow-up period. The second Swedish study compared gentamicin cement with cloxacillin and cephalaxin, demonstrating a fourfold reduction in deep infection rates although there was a slight but significantly higher incidence of superficial wound infection in the gentamicin group, is They achieved infection rates of 0.4% in conventional operating theatres after a 2-year followup. The difference was highly significant. This finding is not isolated. The infection rate of patients receiving gentamicin cement was 0.4% in the M R C trials whilst Murray achieved similar results using erythromycin and colistin mixed with methylmethacrylate.19 There are three essential differences between these methods of antibiotic delivery. The antibiotic impregnated cements are broad spectrum, they achieve high concentrations at the bone-cement interface where infection occurs and give prolonged action at this site. The addition of 1 g of antibiotic powder to a 40 g packet of cement has a negligible effect on the strength of the methylmethacrylate cement.
38
CAUSES AND PREVENTION
(ii) BACTERIOLOGY
Table ~-Controlled Trials on Prophylactic Antibiotics in Hip Replacement
Centre
Number of cases
Lurid
118
Paris
2137
Length of Administration
Drug Cloxacillin vs Control
14 days
Cefazolin
5 days
% Infected Mean Follow-up 2 yrs 6 yrs 0
3
12
23
0.9
0.9
Control
3.4
4.6
Gentamicin Cement
0.4
Ratio 8
5
VS
Lund
1685
4
VS
Cloxacillin or Cephalexin
Choice of Antibiotic T h e present practice is to use a systemic a gen t that has g o o d activity against the s t a p h y l o c o c c u s aureus. This m a y be a methicillin related penicillin or a 1st or 2nd g e n e r a t i o n c e p h a l o s p o r i n . T h i r d g e n e r a t i o n c e p h a l o s p o r i n s have a c q u i r e d activity against g r a m negative o r g a n i s m s at the expense o f s t a p h y lo c o cci and are n o t generally used. T h e evidence indicates that the s p e c t r u m o f infecting o r g a n i s m s is b r o a d a n d that the c e m e n t v e c t o r is superior to systemic a d m i n i s t r a t i o n . Systemic and c e m e n t - i n c o r p o r a t e d antibiotics are m i n i m a l l y additive. A n y a n t i b i o t i c a d d e d to c e m e n t m u s t be heat stable. It w o u l d a p p e a r logical to use antibiotic l o a d e d c e m e n t with an a n t i s t a p h y l o c o c c a l systemic agent to reduce w o u n d infection. T h e period o f a d m i n i s t r a t i o n o f the latter need n o t exceed 24 h.
Cost of Prevention Lidwell gives an a c c o u n t o f the cost o f p r e v e n t i t i v e measures using Swedish records o f 90 extra hospital days an d an a v e r a g e o f 1.6 extra o p e r a t i o n s per infected case. z° These are c r u d e ly calculated to cost £14 000 per case a n d L i d w e l l' s figures are modified to include g e n t a m i c i n c e m e n t (Table 7). It is a s s u m e d that antibiotics cost an extra £30 per patient, that ultraclean air can be p r o v i d e d by a plastic isolator at a p p r o x i m a t e l y £50 per case and that disposable suits costing £20 per case can be used in theatres p e r f o r m i n g fewer than 50 a r t h r o p l a s t i e s per a n n u m . In o p e r a t i n g r o o m s with a greater l o a d the installation o f ultraclean air w o u l d cost a b o u t £2500 a n d b o d y exhaust suits a b o u t £1000 per a n n u m , b o t h including m a i n t e n a n c e and w e a r an d tear. T h e costs o f clean air w o u l d fall in units with a higher v o l u m e o f cases.
1.6
It can be seen that the single m o s t cost-effective m e a s u r e is antibiotic prophylaxis. T h e isolated cost o f clean air w o u l d be repaid in high v o l u m e units b u t even in small centres the overall saving o f a p p l y i n g all these m e a s u r e s w o u l d a m o u n t to over £3000 per case. W i t h the additive effects o f antibiotics and clean air infection rates well u n d e r 0.2% should be achieved.
References i. Lidwell O M, Lowbury E J L, Whyte W, Blowers R, Stanley S J, Lowe D 1982 Effect of ultraclean air in operating rooms on deep sepsis of the joint after total hip or knee replacement: A randomised study. British Medical Journal. 285:1(~14 2. Whyte W, Hodgson R, Tinkler J, Graham J 1981 The isolation of bacteria of low pathogenicity from faulty orthopaedic implants. Journal of Hospital Infection. 2:219-230 3. Lidwell O M, Lowbury E J L, Whyte W, Blowers R, Stanley S J, Lowe D 1983 Bacteria isolated from deep joint sepsis after operation for total hip or knee replacement and the sources of the infections with staphylococcus aureus. Journal of Hospital Infection. 4:19-29 4. Lidwell O M, Lowbury E J L, Whyte W, Blowers R, Stanley S J, Lowe D 1984 Infection and sepsis after operations for total hip or knee-joint replacement: influence of ultraclean air, prophylactic antibiotics and other factors. Journal of Hygiene.
93:505-529 5. Wroblewski B M, De1 Sel H J 1980 Urethral instrumentation and deep sepsis in total hip replacement. Clinical Orthopaedics. 146:209-212 6. Fitzgerald R H, Nolan D R, Ilstrup D M, Van Scoy R E, Washington J AII, Coventry M B 1977 Deep wound sepsis following total hip arthroplasty. Journal of Bone and Joint Surgery. 59A: 847-855 7. Fitzgerald R H, Peterson L F A, Washington J A, Van Scoy R E, Coventry M B 1973 Bacterial colonisation of wounds and sepsis in total hip arthroplasty. Journal of Bone and Joint Surgery. 55A: 1242-1250 8. Charnley J, Eftekhar N 1969 Post-operative infection in total prosthetic replacement arthroplasty of the hip joint. British Journal of Surgery. 56:641-649
Table 7 Costings of Infected Joint Replacement per 100 Cases in £
Type of Theatre Conventional Gentamicin cement and antibiotics + Ultraclean air + Body exhaust suits
Infection Rate %
Cost of Treatment
Cost of Prevention
Financial Savings of Prevention
3.4 0.4
44 200 5 200
Nil 3 000
Nil 36000
0.15
1950
5 000
34 250
0.06
780
2 000
33420
CURRENT ORTHOPAEDICS 9. Southwood R T, Rice J L, McDonald P J, Hakendorf P H, Rozenbilds M A 1985 Infection in experimental hip arthroplasties. Journal o f Bone and Joint Surgery. 67B: 229-231 10. Elson R A, Jephcott A E, McGechie D B, Veretta D 1977 Bacterial infection and acrylic cement in the rat. Journal of Bone and Joint Surgery. 59B: 452-457 11. Salvati E A,. Robinson R P, Zeno S M, Koslin B L, Brause B D, Wilson P D 1982 Infection rates after 3175 total hip and total knee replacements performed with and without a horizontal unidirectional air-flow system. Journal of Bone and Joint Surgery. 64A: 525-535 12. Burke J F 1961 The effective period ofpreventitive antibiotic action in experimental incisions and dermal lesions. Surgery. 50:161-168
13. Boyd R J, Burke J F, Colton T 1973 A double-blind clinical trial of prophylactic antibiotics in hip fractures. Journal of Bone and Joint Surgery. 55A: 1251-1258 14. Pollard J P, Hughes S P F, Scott J E, Evans M J, Benson M K D 1979 Antibiotic prophylaxis in total hip replacement. British Medical Journal. 707-709 15. Nelson C L, Green T G, Porter R A, Warren R D 1983 One day
16.
17.
18.
19. 20.
39
versus 7 days of preventitive antibiotic therapy in orthopaedic surgery. Clinical Orthopaedics and related research. 176: 258-263 Ericson C, Lidgren L, Lindberg L 1973 Cloxacillin in prophylaxis of post-operative infection of the hip. J ournal of Bone and Joint Surgery. 55A: 808-813 Hill C, Mazas F, Flamant R, Evrard J 1981 Prophylactic Cefazolin versus placebo in total hip replacement. Lancet. 795-797 Josefsson G, Lindberg L, Wiklander B 1981 Systemic antibiotics and gentamicin-containing bone cement in the prophylaxis of post-operative infections in total hip arthroplasty. Clinical Orthopaedics and related research. 159:[94-200 Murray W R 1984 Use of antibiotic-containing bone cement. Clinical Orthopaedics and related research. 190:89-95 Lidwell O M 1984 The cost implications of clean air systems and antibiotic prophylaxis in operations for total joint replacement. Infection Control. 5:36-37
Address for reprints: Gordon C. Bannister, Lecturer in Orthopaedics, Bristol Royal Infirmary, Bristol BS2 8HW.