Antibiotic use after cefuroxime prophylaxis in hip and linee joint replacem& L
d
The amount of additional antibiotics measured by defined daily dose (DDD) methods after 2651 hip and 362 knee replacements was assessed after prophylaxis with one or three doses (1502'1511 patients) of cefuroxime. No differences were observed between the two regimens with respect to total amount, type, indication, and duration of additional antibiotics. The incidence of joint sepsis did not differ significantly between the two trial arms, but the sample was too small for definite conclusions. There were 11.4 DDDI100 bed days of additional antibiotics used in 21% of patients after hip replacement and 15.7 DDD1100 bed days in 31% after knee replacement. For wound problems, 3.8 and 6.9 DDD1100 bed days were given in the hip- and knee-replacement groups. For distant infection, 6.5 DDD1100 bed days was administered in both groups. Duration of therapy varied only in relation to indication. Prescribed were penicillins (43% to 50%), sulfonamides (18%), cephalosporins (10% to 16%), and nitrofurantoin THER1991;50:215-20.) (8% to 13%); drug use was related to the type of infection. (CLINPHARMACOL
Ate B. Wymenga, MD, PhD, Yechiel A. Hekster, PhD, Ad Theeuwes, MSc, Harrie L. Muytjens, PhD, Jim R. van Horn, MD, PhD, and Tom J. J. H. Slooff, MD, PhD N+neg.en, The Netherlands During the past decades, various trials have demonstrated that a short perioperative course of antibiotics for the prevention of infection is effective in several surgical procedures, and even a single dose has been advocated in selected cases.'32 In joint-replacement surgery, no difference was found between 14, 7, and 1 day3x4and 5 and 2 days of prophylaxis.5 Recently a prospective, randomized, controlled trial was performed to compare the efficacy of a single perioperative dose of cefuroxime in hip- and knee-joint replacement surgery with that of a three-dose regimen that also served as a c ~ n t r o l . ~ " Cefuroxime, a second-generation cephalosporin, was chosen because this antibiotic provides a broad spectrum of activity against the microorganisms that cause joint sepsis in orthopedic implants.' The purpose of this study was to uncover details concerning the total amount, type, indication, and du-
From the Departments of Orthopaedic Surgery, Clinical Pharmacy, Medical Statistics, and Medical Microbiology, University Hospital Nijmegen. Supported by Glaxo B.V., Nieuwegein, The Netherlands. Received for publication Oct. 19, 1990; accepted April 3, 1991. Reprint requests: Ate B. Wymenga, MD, Institute of Orthopaedics, University Hospital Nijmegen, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. 1311129977
ration of additional antimicrobial treatment used after a single dose or three doses of perioperative cefuroxime prophylaxis in hip- and knee-replacement surgery.
PATIENTS AND METHODS From July 1986 to July 1988, a prospective, randomized, controlled trial was performed at 27 Dutch hospitals with conventionally ventilated operating rooms. In the single-dose group and the three-dose group, 1500 mg cefuroxime was administered intravenously at the induction of anesthesia. In the three-dose group the patients also received 750 mg cefuroxime intravenously after 8 and 16 hours, respectively. Patients who underwent total hip replacement, hemiarthroplasty of the hip, or total knee arthroplasty were eligible for participation in the study. The most important exclusion criteria were the use of antibiotics less than 48 hours before the operation, the administration of perioperative antibiotics other than cefuroxime, former or current infections in the joint, and the use of gentamicin-impregnated bone cement; the latter was often used in high-risk patients. The study was approved by the ethics committees at all the hospitals. Definitions. Joint sepsis was defined as a positive culture from the joint at reoperation or a draining sinus. Strong evidence of joint sepsis was defined as four or more possible signs of sepsis present at follow-up (pain on weight bearing or at rest, tenderness
2 15
CI.IN PHARMACOL THER AUGUSI' 1991
Table I. Summary of patient characteristics and trial results (joint replacements) -
Knee replacement ( n )
H I replacement ~ (n)
-
Character1 \tzc s und results
I Dote ( n = 1327)
3 Do~es ( n = 1324)
Totul ( n = 2651) (70)
l Dose ( n = 175)
3 dose^ ( n = 187)
Total ( n = 362) (%)
Mean age (yr) Malelfemale Diagnosis Arthrosis Rheumatoid arthritis Fracture (fresh) Other Previous operations Joint sepsis Wound infection Minor wound problems Hematoma (LIMIS) Wound drainage Urinary tract infection Skin infection Pulmonary infection Septicemia Bed days (mean) Total
69 1 28711040
69 1 26611058
69 1 55312098
70 6 231 152
71 1 231164
70 9 4613 16
-
LIMIS. L~ghtlmoderateisevcrc. *One othcr ~nfectionwas d~agnosedin a withdrawn patient who received three doses instead of a single dose.
of the wound, fever, an x-ray film showing periostial reaction or progressive bone resorption, an increased erythrocyte sedimentation rate [20 mm above the preoperative value or >35 mm], a positive joint aspiration, a positive arthrogram, and a bone scan showing typical signs of infection or an increased C-reactive protein). Wound infection was defined simply as an area of erythema of more than 1 cm measured from the incision. Minor wound-healing problems were defined as an area of erythema of less than 1 cm measured from the incision, pus in the suture, small wound dehiscence, necrosis of the wound edge, and blisters. Distant infections were diagnosed on the basis of positive cultures. In the urine, more than 10"acterialml was defined as a positive culture. Blood cultures of Staphylococcus epidermidis were regarded as positive only if cultured twice or more. Only a few skin, pulmonary, and other infections were diagnosed on a clinical basis. Patients. A total of 3074 hip and 455 knee operations entered the trial, of which 278 hip and 58 knee operations were excluded. The most frequent reason for exclusion was the use of gentamicin-impregnated bone cement. From the remaining 2796 hip and 397 knee replacements, 145 and 35 were withdrawn, respectively. The major reasons were the administration
of the wrong type or dose of antibiotics, death (not sepsis related) within 7 days of the operation, or a second operation within the same period of hospitalization. This left 265 1 hip replacements (2547 patients) and 362 knee replacements (345 patients) for analysis. This series therefore represented a rather homogeneous group of patients who underwent replacements, from which a number of high-risk patients were eliminated because of the use of gentamicin bone cement. The hip and knee replacements were analyzed separately; the patient characteristics and trial results are shown in Table I. The incidence of joint sepsis in hip replacements was 0.83% in the single-dose group (95% confidence limits, 0.33% to 1.32%) and 0.45% in the three-dose group (95% confidence limits, 0.08% to 0.81%). In the knee replacements the incidence was 1.71% in the single-dose group (95% confidence limits, 0.09% to 3.33%) and 3.20% in the three-dose group (95% confidence limits, 0.63% to 5.77%). These incidences are comparable with those reported in other large series on hip and knee replacement^.^'^-^^ The incidence of joint sepsis was quite similar for the various participating hospitals. The fact that all patients were operated on in the same type of conventionally ventilated operating rooms with the same type of antibiotic prophylaxis has certainly contributed to
Antibiotic use after joint replacement 2 17 Table 11. Number of replacements with no therapy, monotherapy, and multiple antibiotic therapy I Dose No antibiotics
Monotherapy Multiple therapy TOTAL
n
%
1047 224 56
78.9 16.9 4.2
1327
Hrp replacement 3 Dosec n %
1054 222 48
Total
79.6 16.8 3.6
1324
1 Dose
11
7c
n
C/c
2101 446 104
79.3 16.8 3.9
123 40 12
70.3 22.9 6.9
265 1
175
Wound Distantinfection TemperatureECI Other TOTAL
1 Dose (n=1327) n %
3.56 5.84 1.19 0.47 11.05
32.2 52.9 10.8 4.3 100
3 Doses (n=1324) n 5%
Total (n=2651) n %
3.81 6.51 1.12 0.18
3.77 6.15 1.15 0.33
11.82
32.3 55.1 9.5 1.5 100
11.43
33.0 53.9 10.1 2.9 100
126 49 12
----
I Dose (n=175) n 5% 6.88 6.81 1.76 0.11 15.56
674 262 64
187
Table 111. Amount of additional antibiotics prescribed per indication Hip replacement Indication
K w e replac emerlt 3 Dose\ n r4
44.2 43.8 11.3 0.7 I00
16.05
C/c
239 89 24
68.8 24.6 6.6
362
-
Knee replacement 3 Doses ( 1 1 = 187) n 56.
6.94 6.34 2.50 0.12
Totnl n
43.2 395 15.6 0.8 100
--
Tottrl 362)
(11 = 11
6.91 6.57 2.12 0.13 15.71
%
44.0 41.8 13.5 0.7 100
Ant~bioticsin DDD per 100 bed day>.
this.' The surgical approach (anterolateral or posterolateral) and operation time did not influence the rate of joint sepsis. Dejned daily dose (DDD). To enable comparisons between the two groups of patients regarding their overall antibiotic drug use, the methods of the DDD were used.'' The level of the DDD was based on the main indication for the drug and expressed, where possible, as the weight of active substance. For each individual antibiotic drug, the total amount in grams was calculated and divided by its DDD value (depending on the route of administration) to give the total amount of DDDs. By adding the DDD values for the drugs, group DDDs and the total amount of DDDs were assessed. The DDD values were taken from the Nordic Statistics on Medicines. DDDIIUU bed days. To make comparisons possible, the total value of DDDs was analyzed in relation to the total number of bed days. For the hospitals this relationship was expressed as DDD per 100 bed days. This unit provides a rough estimate of the proportion of patients who would receive antibiotic drug treatment. The total number of bed days was calculated from the data of individual patients. The number of bed days in the hospital was missing in 45 patients, and a correction was performed by adding the mean
''
number of bed days from the patient group concerned for each missing value (Table I).
RESULTS Additional antibiotics. Additional antibiotics were prescribed in 20.7% of the hip replacements versus 31.2% of the knee replacements. The percentage of patients receiving multiple therapy was also somewhat higher in the knee-replacement group than in the hipreplacement group. There was no difference between the use of antibiotics in the single- and three-dose groups for either operation (Table 11). The mean duration of hospital stay, without additional antibiotics prescribed, was 22.6 days after a hip replacement and 27.3 days after a knee replacement. With additional antibiotics prescribed the mean duration was 35.5 and 40.3 days, respectively. Indication for additional antibiotics. The amount of additional antibiotics, with the indication for which they were prescribed, is shown in Table 111, expressed as DDD per 100 bed days. No large differences were observed between the single- and three-dose groups in hip and knee replacements. The total amount of antibiotics prescribed was higher in the knee-replacement group: nearly twice the amount of antibiotics was prescribed for wound prob-
CLIN PHARMACOI. TWER AUGUST 1991
2 18 Wymenga et al.
Table IV. Mean duration of additional antibiotic therapy per indication in combination with the number of patients receiving antibiotics
Indication
Wound Distant infection Temperature ECI Other
I Dose ( n = 1327) Duys n 16.9 9.7 12.7 3.3
48 196 18 35
Hip replacement 3 Doses ( n = 1324) Days n 18.7 1 1 .O 12.6 2.3
40 198 18 24
Total ( n = 2651) Days n 88 394 36 59
17.7 9.6 12.7 3.0
I Dose ( n = 175) Duys n 23.8 11.0 9.6 1.0
11 30 8 7
Knee replacement 3 Doses Total ( n = 187) ( n = 362) Days n Days n 15.3 8.6 14.3 1.0
15 38 7 5
18.8 9.7 11.8 1.0
26 68 15 12
Table V. Types of additional antibiotic prescribed - - - ~ ~
Antibiotic Penicillin Sulfonamides Cefalosporins Nitrofurantoin Aminoglycosides Quinolones Tetracyclines Metronidazole Miconazole Erythromycin Isoniazid Lincosamides Fusidic acid TOTAL
Hip replacement (DDDIIOO bed days) I Dose 3 Doses Total ( n = 1327) ( n = 1324) ( n = 2651) 4.767 1.990 1.787 1.638 0.424 0.136 0.195 0.060 0.014 0.043
Knee replacement (DDDIIOO bed days) 1 Dose 3 Doses % ( n = 175) ( n = 187)
4.931 2.101 1.873 1.562 0.329 0.199 0.193 0.078 0.021 0.013 0.010
43.1 18.4 16.4 13.7 2.9 1.7 1.7 0.7 0.2 0. I 0. I
7.499 2.916 1.095 0.871 0.320 0.402 0.089 0.024
-
5.115 2.224 1.958 1.480 0.224 0.260 0.191 0.099 0.024 0.220 0.020
-
-
-
-
-
-
11.05
11.82
11.43
( 100)
lems and temperature e causa ignota. The amount of antibiotics administered for distant infections in hip and knee replacement was fairly similar (around 6.5 DDDJ100 bed days). In the hip-replacement group the largest amount of antibiotics was administered for distant infections (53%) and somewhat less for wound problems (33%). In knee replacement the amount of antibiotics administered for wound problems (44%) and distant infections (42%) was similar. Duration of treatment. The length of antibiotic treatment in days is shown in Table IV. In hip replacement no large differences were observed between the single and three-dose groups, although the number of patients in the single-dose group with wound problems was somewhat higher than in the hipreplacement group. The differences observed in the knee-replacement group can be explained by the smaller number of patients and the large range of treatment days between the patients. The overall length of antibiotic therapy for hip and
7.747 2.754 2.060 1.640 0. 184 0.378 0.535 -
(n
Total = 362)
%
7.890 2.835 1.582 1.255 0.249 0.381 0.316 0.012
50.2 18.0 10.1 8.0 1.6 2.4 2.0 0.1 -
-
-
-
-
0.285
0.142
0.9
-
-
-
-
0.551 1.280 15.56
0.281
0.416 0.641 15.71
2.7 4.1 (100)
-
16.05
knee replacement was fairly similar and depended mainly on the indication for antibiotic therapy and not on the type of operation. In this trial, orthopedic surgeons prescribed antibiotics for wound problems for an average of 18 days, for distant infections 10 days, and for temperature ECI 12 days. Types of antibiotic. In Table V the DDD1100 bed days for the several groups of antibiotics is shown. The antibiotics prescribed most often were (in descending order) penicillin, the sulfonamides, the cephalosporins, and nitrofurantoin. The only striking difference between the two types of operation was a higher DDDilOO bed days for the penicillin group in knee replacements. Within the cephalosporin group, first-generation cephalosporins, such as cefazolin, were prescribed in 85%, second-generation cephalosporins in 14%, and third-generation cephalosporins in only 1% of cases. Type of antibiotic and indication. More detailed information on the type of antibiotic prescribed for a
VOLUME SO NUMRER 2
Antibiotic use a$ev joint replacement 2 19
Table VI. Types of additional antibiotic prescribed for wound problems and distant infection ---
Antibiotic
Hip replucement Wound problems Distunt infection n % n %
Knee replucernent Wound problems Distant infection n 8 II %
Sulfonamides Nitrofurantoin Penicillin Cephalosporins Tetracyclines Quinolones Aminoglycosides Metronidazole Erythromycin Miconazole Lincosamides Isoniazid Fusidic acid TOTAL
Antibiol~cain DDD per I00 bed days
particular indication is presented in Table VI. Penicillin was the drug of choice for wound problems in both hip and knee replacement in more than two thirds of the cases. Within this group, floxacillin was prescribed in 89% of cases. For distant infections (the majority being urinary tract infection; Table I), sulfonamides were the most popular (33% to 35%); the amount of nitrofurantoin (19% to 25%) and penicillin (25% to 28%) was fairly similar. For distant infections in the penicillin group, amoxicillin was prescribed in 54% and floxacillin in 29% of cases.
DISCUSSION Joint sepsis. In hip replacement the difference in joint sepsis between the single- and three-dose group was 0.38%, with an upper confidence limit of 0.90%. We could therefore not confirm that a single dose of cefuroxime was as effective as a three-dose regimen. The trial sample in the knee-replacement group was too small to draw definite conclusions. A longer follow-up (with probably more cases of joint sepsis) is presently underway. Use of additional antibiotics. In this study, data are presented on the use of additional antibiotics after a short course of perioperative prophylaxis, with a single dose or three doses of cefuroxime. In 20% of the hip and 30% of the knee replacements, additional antimicrobial therapy was given. This use of antibiotics could partly be seen as a consequence of hospitalacquired infections. The prevalence of such infections
varied widely between the hospitals: six hospitals had less than 10% of urinary tract infections, 13 hospitals had 10% to 20%, and six more than 20%. Others have confirmed such findings.'' The variation might be caused by differences in maleifemale distribution and the percentage of the patients receiving a urinary catheter. The incidence of urinary tract infection increased with the length of hospital stay from 8% to 12% after 3 weeks to around 35% after 8 weeks or more. Monotherapy was used in the majority of cases requiring antibiotic treatment. With the DDD per 100 bed days method, it was possible to compare the use of antibiotics between groups. The analysis showed that the prescription of antibiotics was parallel in both groups and did not detect any difference per indication for antibiotic treatment. The use of antibiotics in knee replacement was higher than in hip replacement. This phenomenon can be explained by the higher incidence of postoperative wound problems and joint sepsis (Table I), as has been found by several other authors.''-'" Therefore orthopedic surgeons are inclined to administer antibiotics more often after knee-replacement surgery, even for less serious wound-healing problems. The duration of treatment for the different kinds of infection did not differ between the knee and the hip groups. This confirms the idea that irrespective of the type of operation, wound problems and distant infections are treated in a similar way. Wound infections were caused by the Staphylococcus species in 51% of our patients.I8 It therefore seems worthwhile to choose a drug that covers these
220 Wymengp et al. strains. For wound problems, penicillin and cephalosporins were prescribed for more than 60% of the cases (Table VI). Within the penicillin group, floxacillin was indeed chosen, as were the first-generation cephalosporins such as cefazolin. These drugs have broad coverage against staphylococci. Penicillin and cephalosporins were used mainly for wound infections. For distant infections, mostly composed of urinary tract infections, sulfonamides in combination with trimethoprim were the drugs of choice. Nitrofurantoin was also popular. These drugs provide good coverage against the common causative microorganisms that can be found in these infections. The frequency of resistance to cefuroxime was low and did not differ much between the hospitals. All strains of S. aureus were sensitive; only 15% of S. epidermidis and 3% of Escherichia coli were resistant against cefuroxime. This study shows that with the DDD method, no relevant differences were found between a single dose and three doses of perioperative cefuroxime in hip and knee replacements, with respect to the amount, type, indication, and duration of additional antibiotic therapy. The number of patients who received additional antibiotics was considerable and the main indications were distant infection (usually urinary tract infections) and wound-healing problems. In the knee-replacement group, more antibiotics were used than in the hip group, mainly for wound-healing problems. We thank all participating orthopedic surgeons who made this study possible. We also thank Dr. 0 . M. Lidwell (MRC Common Cold Unit, Harvard Hospital, Salisbury, Wiltshire, England) for advice on the interpretation of the trial results, Glaxo B.V., The Netherlands, who supplied cefuroxime, and Mrs. J. Abma-Hill, who corrected the English.
References 1. Dipiro J, Cheung R, Bowden TA, Mansberger JA. Single-dose systemic antibiotic prophylaxis of surgical wound infection. Am J Surg 1986;152:552-9. 2. Pollock AV. Surgical prophylaxis: the emerging picture. Lancet 1988;1:225-30 3. Heydemann JS, Nelson CL. Short-term preventive antibiotics. Clin Orthop 1986;205:184-7. 4. Pollard JP, Hughes SPF, Scott JE, Evans MJ, Benson MKD. Antibiotic prophylaxis in total hip replacement. Br Med J 1979;1:707-9.
CLIN PHARMACOL THER AUGUST 1991
5. Evrard J, Doyon F, Aacar JF, Salord JC, Mazas F, Flamant R. Two-day cefamandole versus five-day cephalzolin prophylaxis in 965 total hip replacements. lnt Orthop 1988;12:69-73. 6. Wymenga AB, van Horn JR, Theeuwes A, Muytjens HL, Slooff TJJH. Joint sepsis after infection prophylaxis with one and three doses cefuroxime in 2651 hip replacements [Thesis]. Nijmegen, The Netherlands: University of Nijmegen, 1991:33-43. 7. Wymenga AB, van Horn JR, Theeuwes A, Muytjens HL, Slooff TJJH. Joint sepsis after infection prophylaxis with one and three doses cefuroxime in 362 knee replacements. Int Orthop [In press]. 8. Sanderson PJ. The choice between prophylactic agents for orthopaedic surgery. J Hosp Infect 1988;ll (suppl): 57-67. 9. Lidwell OM, Lowbury EJL, Whyte W, Blowers R, Stanley SJ, Lowe D. Infection and sepsis after operation for total hip or knee-joint replacement: influence of clean air, prophylactic antibiotics and other factors. J Hyg Cambridge 1984;93:505-29. 10. Hill C, Mazas F, Flamant R, Evrard J. Prophylactic cefazolin versus placebo in total hip replacement. Lancet 1981;i:795-7. I I. Johnson DP, Bannister GC. The outcome of infected arthroplasty of the knee. J Bone Joint Surg 1986;68B:795-803. 12. Knutson K, Lindstrand A, Lidgren L. Survival of knee arthroplasties: a nation-wide multicentre investigation of 8000 cases. J Bone Joint Surg 1986;68B:289-91. 13. Grogan TJ, Dorey F, Rollins J, Amstutz HC. Deep sepsis following total knee arthroplasty. J Bone Joint Surg 1986;68A:226-34. 14. Rand JA, Fitzgerald RH. Diagnosis and management of the infected total knee arthroplasty. Orthop Clin North Am I989;20:2Ol- 10. 15. Wertheimer AI. The defined daily dose system (DDD) for drug utilization review. Hosp Pharm 1986;21:233-58. 16. Anonymous. Nordic statistics on medicines I98 1 - 1983. Part 11. Nordic drug index with classification and defined daily doses. Uppsala: Nordic Council on Medicines, 1985. 17. Tikhomirov E. WHO programme for the control of hospital infection. Chemioterapia 1987;6:148-51. 18. Wymenga AB, van Horn JR, Theeuwes A , Muytjens HL, Slooff TJJH. The relation of wound and urine cultures and joint sepsis after hip and knee replacement [Thesis]. Nijmegen, The Netherlands: University of Nijmegen, 1991:83-95.
d
The amount of additional antibiotics measured by defined daily dose (DDD) methods after 2651 hip and 362 knee replacements was assessed after prophylaxis with one or three doses (1502'1511 patients) of cefuroxime. No differences were observed between the two regimens with respect to total amount, type, indication, and duration of additional antibiotics. The incidence of joint sepsis did not differ significantly between the two trial arms, but the sample was too small for definite conclusions. There were 11.4 DDDI100 bed days of additional antibiotics used in 21% of patients after hip replacement and 15.7 DDD1100 bed days in 31% after knee replacement. For wound problems, 3.8 and 6.9 DDD1100 bed days were given in the hip- and knee-replacement groups. For distant infection, 6.5 DDD1100 bed days was administered in both groups. Duration of therapy varied only in relation to indication. Prescribed were penicillins (43% to 50%), sulfonamides (18%), cephalosporins (10% to 16%), and nitrofurantoin THER1991;50:215-20.) (8% to 13%); drug use was related to the type of infection. (CLINPHARMACOL
Ate B. Wymenga, MD, PhD, Yechiel A. Hekster, PhD, Ad Theeuwes, MSc, Harrie L. Muytjens, PhD, Jim R. van Horn, MD, PhD, and Tom J. J. H. Slooff, MD, PhD N+neg.en, The Netherlands During the past decades, various trials have demonstrated that a short perioperative course of antibiotics for the prevention of infection is effective in several surgical procedures, and even a single dose has been advocated in selected cases.'32 In joint-replacement surgery, no difference was found between 14, 7, and 1 day3x4and 5 and 2 days of prophylaxis.5 Recently a prospective, randomized, controlled trial was performed to compare the efficacy of a single perioperative dose of cefuroxime in hip- and knee-joint replacement surgery with that of a three-dose regimen that also served as a c ~ n t r o l . ~ " Cefuroxime, a second-generation cephalosporin, was chosen because this antibiotic provides a broad spectrum of activity against the microorganisms that cause joint sepsis in orthopedic implants.' The purpose of this study was to uncover details concerning the total amount, type, indication, and du-
From the Departments of Orthopaedic Surgery, Clinical Pharmacy, Medical Statistics, and Medical Microbiology, University Hospital Nijmegen. Supported by Glaxo B.V., Nieuwegein, The Netherlands. Received for publication Oct. 19, 1990; accepted April 3, 1991. Reprint requests: Ate B. Wymenga, MD, Institute of Orthopaedics, University Hospital Nijmegen, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. 1311129977
ration of additional antimicrobial treatment used after a single dose or three doses of perioperative cefuroxime prophylaxis in hip- and knee-replacement surgery.
PATIENTS AND METHODS From July 1986 to July 1988, a prospective, randomized, controlled trial was performed at 27 Dutch hospitals with conventionally ventilated operating rooms. In the single-dose group and the three-dose group, 1500 mg cefuroxime was administered intravenously at the induction of anesthesia. In the three-dose group the patients also received 750 mg cefuroxime intravenously after 8 and 16 hours, respectively. Patients who underwent total hip replacement, hemiarthroplasty of the hip, or total knee arthroplasty were eligible for participation in the study. The most important exclusion criteria were the use of antibiotics less than 48 hours before the operation, the administration of perioperative antibiotics other than cefuroxime, former or current infections in the joint, and the use of gentamicin-impregnated bone cement; the latter was often used in high-risk patients. The study was approved by the ethics committees at all the hospitals. Definitions. Joint sepsis was defined as a positive culture from the joint at reoperation or a draining sinus. Strong evidence of joint sepsis was defined as four or more possible signs of sepsis present at follow-up (pain on weight bearing or at rest, tenderness
2 15
CI.IN PHARMACOL THER AUGUSI' 1991
Table I. Summary of patient characteristics and trial results (joint replacements) -
Knee replacement ( n )
H I replacement ~ (n)
-
Character1 \tzc s und results
I Dote ( n = 1327)
3 Do~es ( n = 1324)
Totul ( n = 2651) (70)
l Dose ( n = 175)
3 dose^ ( n = 187)
Total ( n = 362) (%)
Mean age (yr) Malelfemale Diagnosis Arthrosis Rheumatoid arthritis Fracture (fresh) Other Previous operations Joint sepsis Wound infection Minor wound problems Hematoma (LIMIS) Wound drainage Urinary tract infection Skin infection Pulmonary infection Septicemia Bed days (mean) Total
69 1 28711040
69 1 26611058
69 1 55312098
70 6 231 152
71 1 231164
70 9 4613 16
-
LIMIS. L~ghtlmoderateisevcrc. *One othcr ~nfectionwas d~agnosedin a withdrawn patient who received three doses instead of a single dose.
of the wound, fever, an x-ray film showing periostial reaction or progressive bone resorption, an increased erythrocyte sedimentation rate [20 mm above the preoperative value or >35 mm], a positive joint aspiration, a positive arthrogram, and a bone scan showing typical signs of infection or an increased C-reactive protein). Wound infection was defined simply as an area of erythema of more than 1 cm measured from the incision. Minor wound-healing problems were defined as an area of erythema of less than 1 cm measured from the incision, pus in the suture, small wound dehiscence, necrosis of the wound edge, and blisters. Distant infections were diagnosed on the basis of positive cultures. In the urine, more than 10"acterialml was defined as a positive culture. Blood cultures of Staphylococcus epidermidis were regarded as positive only if cultured twice or more. Only a few skin, pulmonary, and other infections were diagnosed on a clinical basis. Patients. A total of 3074 hip and 455 knee operations entered the trial, of which 278 hip and 58 knee operations were excluded. The most frequent reason for exclusion was the use of gentamicin-impregnated bone cement. From the remaining 2796 hip and 397 knee replacements, 145 and 35 were withdrawn, respectively. The major reasons were the administration
of the wrong type or dose of antibiotics, death (not sepsis related) within 7 days of the operation, or a second operation within the same period of hospitalization. This left 265 1 hip replacements (2547 patients) and 362 knee replacements (345 patients) for analysis. This series therefore represented a rather homogeneous group of patients who underwent replacements, from which a number of high-risk patients were eliminated because of the use of gentamicin bone cement. The hip and knee replacements were analyzed separately; the patient characteristics and trial results are shown in Table I. The incidence of joint sepsis in hip replacements was 0.83% in the single-dose group (95% confidence limits, 0.33% to 1.32%) and 0.45% in the three-dose group (95% confidence limits, 0.08% to 0.81%). In the knee replacements the incidence was 1.71% in the single-dose group (95% confidence limits, 0.09% to 3.33%) and 3.20% in the three-dose group (95% confidence limits, 0.63% to 5.77%). These incidences are comparable with those reported in other large series on hip and knee replacement^.^'^-^^ The incidence of joint sepsis was quite similar for the various participating hospitals. The fact that all patients were operated on in the same type of conventionally ventilated operating rooms with the same type of antibiotic prophylaxis has certainly contributed to
Antibiotic use after joint replacement 2 17 Table 11. Number of replacements with no therapy, monotherapy, and multiple antibiotic therapy I Dose No antibiotics
Monotherapy Multiple therapy TOTAL
n
%
1047 224 56
78.9 16.9 4.2
1327
Hrp replacement 3 Dosec n %
1054 222 48
Total
79.6 16.8 3.6
1324
1 Dose
11
7c
n
C/c
2101 446 104
79.3 16.8 3.9
123 40 12
70.3 22.9 6.9
265 1
175
Wound Distantinfection TemperatureECI Other TOTAL
1 Dose (n=1327) n %
3.56 5.84 1.19 0.47 11.05
32.2 52.9 10.8 4.3 100
3 Doses (n=1324) n 5%
Total (n=2651) n %
3.81 6.51 1.12 0.18
3.77 6.15 1.15 0.33
11.82
32.3 55.1 9.5 1.5 100
11.43
33.0 53.9 10.1 2.9 100
126 49 12
----
I Dose (n=175) n 5% 6.88 6.81 1.76 0.11 15.56
674 262 64
187
Table 111. Amount of additional antibiotics prescribed per indication Hip replacement Indication
K w e replac emerlt 3 Dose\ n r4
44.2 43.8 11.3 0.7 I00
16.05
C/c
239 89 24
68.8 24.6 6.6
362
-
Knee replacement 3 Doses ( 1 1 = 187) n 56.
6.94 6.34 2.50 0.12
Totnl n
43.2 395 15.6 0.8 100
--
Tottrl 362)
(11 = 11
6.91 6.57 2.12 0.13 15.71
%
44.0 41.8 13.5 0.7 100
Ant~bioticsin DDD per 100 bed day>.
this.' The surgical approach (anterolateral or posterolateral) and operation time did not influence the rate of joint sepsis. Dejned daily dose (DDD). To enable comparisons between the two groups of patients regarding their overall antibiotic drug use, the methods of the DDD were used.'' The level of the DDD was based on the main indication for the drug and expressed, where possible, as the weight of active substance. For each individual antibiotic drug, the total amount in grams was calculated and divided by its DDD value (depending on the route of administration) to give the total amount of DDDs. By adding the DDD values for the drugs, group DDDs and the total amount of DDDs were assessed. The DDD values were taken from the Nordic Statistics on Medicines. DDDIIUU bed days. To make comparisons possible, the total value of DDDs was analyzed in relation to the total number of bed days. For the hospitals this relationship was expressed as DDD per 100 bed days. This unit provides a rough estimate of the proportion of patients who would receive antibiotic drug treatment. The total number of bed days was calculated from the data of individual patients. The number of bed days in the hospital was missing in 45 patients, and a correction was performed by adding the mean
''
number of bed days from the patient group concerned for each missing value (Table I).
RESULTS Additional antibiotics. Additional antibiotics were prescribed in 20.7% of the hip replacements versus 31.2% of the knee replacements. The percentage of patients receiving multiple therapy was also somewhat higher in the knee-replacement group than in the hipreplacement group. There was no difference between the use of antibiotics in the single- and three-dose groups for either operation (Table 11). The mean duration of hospital stay, without additional antibiotics prescribed, was 22.6 days after a hip replacement and 27.3 days after a knee replacement. With additional antibiotics prescribed the mean duration was 35.5 and 40.3 days, respectively. Indication for additional antibiotics. The amount of additional antibiotics, with the indication for which they were prescribed, is shown in Table 111, expressed as DDD per 100 bed days. No large differences were observed between the single- and three-dose groups in hip and knee replacements. The total amount of antibiotics prescribed was higher in the knee-replacement group: nearly twice the amount of antibiotics was prescribed for wound prob-
CLIN PHARMACOI. TWER AUGUST 1991
2 18 Wymenga et al.
Table IV. Mean duration of additional antibiotic therapy per indication in combination with the number of patients receiving antibiotics
Indication
Wound Distant infection Temperature ECI Other
I Dose ( n = 1327) Duys n 16.9 9.7 12.7 3.3
48 196 18 35
Hip replacement 3 Doses ( n = 1324) Days n 18.7 1 1 .O 12.6 2.3
40 198 18 24
Total ( n = 2651) Days n 88 394 36 59
17.7 9.6 12.7 3.0
I Dose ( n = 175) Duys n 23.8 11.0 9.6 1.0
11 30 8 7
Knee replacement 3 Doses Total ( n = 187) ( n = 362) Days n Days n 15.3 8.6 14.3 1.0
15 38 7 5
18.8 9.7 11.8 1.0
26 68 15 12
Table V. Types of additional antibiotic prescribed - - - ~ ~
Antibiotic Penicillin Sulfonamides Cefalosporins Nitrofurantoin Aminoglycosides Quinolones Tetracyclines Metronidazole Miconazole Erythromycin Isoniazid Lincosamides Fusidic acid TOTAL
Hip replacement (DDDIIOO bed days) I Dose 3 Doses Total ( n = 1327) ( n = 1324) ( n = 2651) 4.767 1.990 1.787 1.638 0.424 0.136 0.195 0.060 0.014 0.043
Knee replacement (DDDIIOO bed days) 1 Dose 3 Doses % ( n = 175) ( n = 187)
4.931 2.101 1.873 1.562 0.329 0.199 0.193 0.078 0.021 0.013 0.010
43.1 18.4 16.4 13.7 2.9 1.7 1.7 0.7 0.2 0. I 0. I
7.499 2.916 1.095 0.871 0.320 0.402 0.089 0.024
-
5.115 2.224 1.958 1.480 0.224 0.260 0.191 0.099 0.024 0.220 0.020
-
-
-
-
-
-
11.05
11.82
11.43
( 100)
lems and temperature e causa ignota. The amount of antibiotics administered for distant infections in hip and knee replacement was fairly similar (around 6.5 DDDJ100 bed days). In the hip-replacement group the largest amount of antibiotics was administered for distant infections (53%) and somewhat less for wound problems (33%). In knee replacement the amount of antibiotics administered for wound problems (44%) and distant infections (42%) was similar. Duration of treatment. The length of antibiotic treatment in days is shown in Table IV. In hip replacement no large differences were observed between the single and three-dose groups, although the number of patients in the single-dose group with wound problems was somewhat higher than in the hipreplacement group. The differences observed in the knee-replacement group can be explained by the smaller number of patients and the large range of treatment days between the patients. The overall length of antibiotic therapy for hip and
7.747 2.754 2.060 1.640 0. 184 0.378 0.535 -
(n
Total = 362)
%
7.890 2.835 1.582 1.255 0.249 0.381 0.316 0.012
50.2 18.0 10.1 8.0 1.6 2.4 2.0 0.1 -
-
-
-
-
0.285
0.142
0.9
-
-
-
-
0.551 1.280 15.56
0.281
0.416 0.641 15.71
2.7 4.1 (100)
-
16.05
knee replacement was fairly similar and depended mainly on the indication for antibiotic therapy and not on the type of operation. In this trial, orthopedic surgeons prescribed antibiotics for wound problems for an average of 18 days, for distant infections 10 days, and for temperature ECI 12 days. Types of antibiotic. In Table V the DDD1100 bed days for the several groups of antibiotics is shown. The antibiotics prescribed most often were (in descending order) penicillin, the sulfonamides, the cephalosporins, and nitrofurantoin. The only striking difference between the two types of operation was a higher DDDilOO bed days for the penicillin group in knee replacements. Within the cephalosporin group, first-generation cephalosporins, such as cefazolin, were prescribed in 85%, second-generation cephalosporins in 14%, and third-generation cephalosporins in only 1% of cases. Type of antibiotic and indication. More detailed information on the type of antibiotic prescribed for a
VOLUME SO NUMRER 2
Antibiotic use a$ev joint replacement 2 19
Table VI. Types of additional antibiotic prescribed for wound problems and distant infection ---
Antibiotic
Hip replucement Wound problems Distunt infection n % n %
Knee replucernent Wound problems Distant infection n 8 II %
Sulfonamides Nitrofurantoin Penicillin Cephalosporins Tetracyclines Quinolones Aminoglycosides Metronidazole Erythromycin Miconazole Lincosamides Isoniazid Fusidic acid TOTAL
Antibiol~cain DDD per I00 bed days
particular indication is presented in Table VI. Penicillin was the drug of choice for wound problems in both hip and knee replacement in more than two thirds of the cases. Within this group, floxacillin was prescribed in 89% of cases. For distant infections (the majority being urinary tract infection; Table I), sulfonamides were the most popular (33% to 35%); the amount of nitrofurantoin (19% to 25%) and penicillin (25% to 28%) was fairly similar. For distant infections in the penicillin group, amoxicillin was prescribed in 54% and floxacillin in 29% of cases.
DISCUSSION Joint sepsis. In hip replacement the difference in joint sepsis between the single- and three-dose group was 0.38%, with an upper confidence limit of 0.90%. We could therefore not confirm that a single dose of cefuroxime was as effective as a three-dose regimen. The trial sample in the knee-replacement group was too small to draw definite conclusions. A longer follow-up (with probably more cases of joint sepsis) is presently underway. Use of additional antibiotics. In this study, data are presented on the use of additional antibiotics after a short course of perioperative prophylaxis, with a single dose or three doses of cefuroxime. In 20% of the hip and 30% of the knee replacements, additional antimicrobial therapy was given. This use of antibiotics could partly be seen as a consequence of hospitalacquired infections. The prevalence of such infections
varied widely between the hospitals: six hospitals had less than 10% of urinary tract infections, 13 hospitals had 10% to 20%, and six more than 20%. Others have confirmed such findings.'' The variation might be caused by differences in maleifemale distribution and the percentage of the patients receiving a urinary catheter. The incidence of urinary tract infection increased with the length of hospital stay from 8% to 12% after 3 weeks to around 35% after 8 weeks or more. Monotherapy was used in the majority of cases requiring antibiotic treatment. With the DDD per 100 bed days method, it was possible to compare the use of antibiotics between groups. The analysis showed that the prescription of antibiotics was parallel in both groups and did not detect any difference per indication for antibiotic treatment. The use of antibiotics in knee replacement was higher than in hip replacement. This phenomenon can be explained by the higher incidence of postoperative wound problems and joint sepsis (Table I), as has been found by several other authors.''-'" Therefore orthopedic surgeons are inclined to administer antibiotics more often after knee-replacement surgery, even for less serious wound-healing problems. The duration of treatment for the different kinds of infection did not differ between the knee and the hip groups. This confirms the idea that irrespective of the type of operation, wound problems and distant infections are treated in a similar way. Wound infections were caused by the Staphylococcus species in 51% of our patients.I8 It therefore seems worthwhile to choose a drug that covers these
220 Wymengp et al. strains. For wound problems, penicillin and cephalosporins were prescribed for more than 60% of the cases (Table VI). Within the penicillin group, floxacillin was indeed chosen, as were the first-generation cephalosporins such as cefazolin. These drugs have broad coverage against staphylococci. Penicillin and cephalosporins were used mainly for wound infections. For distant infections, mostly composed of urinary tract infections, sulfonamides in combination with trimethoprim were the drugs of choice. Nitrofurantoin was also popular. These drugs provide good coverage against the common causative microorganisms that can be found in these infections. The frequency of resistance to cefuroxime was low and did not differ much between the hospitals. All strains of S. aureus were sensitive; only 15% of S. epidermidis and 3% of Escherichia coli were resistant against cefuroxime. This study shows that with the DDD method, no relevant differences were found between a single dose and three doses of perioperative cefuroxime in hip and knee replacements, with respect to the amount, type, indication, and duration of additional antibiotic therapy. The number of patients who received additional antibiotics was considerable and the main indications were distant infection (usually urinary tract infections) and wound-healing problems. In the knee-replacement group, more antibiotics were used than in the hip group, mainly for wound-healing problems. We thank all participating orthopedic surgeons who made this study possible. We also thank Dr. 0 . M. Lidwell (MRC Common Cold Unit, Harvard Hospital, Salisbury, Wiltshire, England) for advice on the interpretation of the trial results, Glaxo B.V., The Netherlands, who supplied cefuroxime, and Mrs. J. Abma-Hill, who corrected the English.
References 1. Dipiro J, Cheung R, Bowden TA, Mansberger JA. Single-dose systemic antibiotic prophylaxis of surgical wound infection. Am J Surg 1986;152:552-9. 2. Pollock AV. Surgical prophylaxis: the emerging picture. Lancet 1988;1:225-30 3. Heydemann JS, Nelson CL. Short-term preventive antibiotics. Clin Orthop 1986;205:184-7. 4. Pollard JP, Hughes SPF, Scott JE, Evans MJ, Benson MKD. Antibiotic prophylaxis in total hip replacement. Br Med J 1979;1:707-9.
CLIN PHARMACOL THER AUGUST 1991
5. Evrard J, Doyon F, Aacar JF, Salord JC, Mazas F, Flamant R. Two-day cefamandole versus five-day cephalzolin prophylaxis in 965 total hip replacements. lnt Orthop 1988;12:69-73. 6. Wymenga AB, van Horn JR, Theeuwes A, Muytjens HL, Slooff TJJH. Joint sepsis after infection prophylaxis with one and three doses cefuroxime in 2651 hip replacements [Thesis]. Nijmegen, The Netherlands: University of Nijmegen, 1991:33-43. 7. Wymenga AB, van Horn JR, Theeuwes A, Muytjens HL, Slooff TJJH. Joint sepsis after infection prophylaxis with one and three doses cefuroxime in 362 knee replacements. Int Orthop [In press]. 8. Sanderson PJ. The choice between prophylactic agents for orthopaedic surgery. J Hosp Infect 1988;ll (suppl): 57-67. 9. Lidwell OM, Lowbury EJL, Whyte W, Blowers R, Stanley SJ, Lowe D. Infection and sepsis after operation for total hip or knee-joint replacement: influence of clean air, prophylactic antibiotics and other factors. J Hyg Cambridge 1984;93:505-29. 10. Hill C, Mazas F, Flamant R, Evrard J. Prophylactic cefazolin versus placebo in total hip replacement. Lancet 1981;i:795-7. I I. Johnson DP, Bannister GC. The outcome of infected arthroplasty of the knee. J Bone Joint Surg 1986;68B:795-803. 12. Knutson K, Lindstrand A, Lidgren L. Survival of knee arthroplasties: a nation-wide multicentre investigation of 8000 cases. J Bone Joint Surg 1986;68B:289-91. 13. Grogan TJ, Dorey F, Rollins J, Amstutz HC. Deep sepsis following total knee arthroplasty. J Bone Joint Surg 1986;68A:226-34. 14. Rand JA, Fitzgerald RH. Diagnosis and management of the infected total knee arthroplasty. Orthop Clin North Am I989;20:2Ol- 10. 15. Wertheimer AI. The defined daily dose system (DDD) for drug utilization review. Hosp Pharm 1986;21:233-58. 16. Anonymous. Nordic statistics on medicines I98 1 - 1983. Part 11. Nordic drug index with classification and defined daily doses. Uppsala: Nordic Council on Medicines, 1985. 17. Tikhomirov E. WHO programme for the control of hospital infection. Chemioterapia 1987;6:148-51. 18. Wymenga AB, van Horn JR, Theeuwes A , Muytjens HL, Slooff TJJH. The relation of wound and urine cultures and joint sepsis after hip and knee replacement [Thesis]. Nijmegen, The Netherlands: University of Nijmegen, 1991:83-95.
