- Email: [email protected]
Efficacy of Prophylactic Antibiotics in Simple Knee Arthroscopy
Efficacy of Prophylactic Antibiotics in Simple Knee Arthroscopy Ronald W. B. Wyatt, M.D., Gregory B. Maletis, M.D., Liisa L. Lyon, M.S., Joan Schwalbe, M.S., and Andrew L. Avins, M.D., M.P.H.
Purpose: To determine the association between the use of preoperative antibiotics and the risk of postoperative infection after simple knee arthroscopy. Methods: The electronic medical records of a large integrated health care organization were used to identify patients who underwent simple knee arthroscopy between 2007 and 2012. Patient demographics, potential infection risk factors, and antibiotic administration data were extracted. Simple knee arthroscopy included debridement, meniscectomy, meniscus repair, synovectomy, microfracture, and lateral release. Complex knee arthroscopy, septic knees, and cases involving fractures were excluded. Deep infection was defined as a positive synovial fluid culture or signs and symptoms of infection and gross pus in the knee. Superficial infection was defined as clinical signs of infection localized to a portal site and treatment with an antibiotic. Results: Of 40,810 simple knee arthroscopies, 32,836 (80.5%) received preoperative antibiotics and 7,974 (19.5%) did not. There were 25 deep infections in the antibiotic group (0.08%) and 11 in the no-antibiotics group (0.14%) (risk ratio ¼ 0.55, 95% confidence interval: 0.27 to 1.12, P ¼ .10). There were 134 superficial infections in the antibiotic group (0.41%) and 32 in the no-antibiotics group (0.40%) (risk ratio ¼ 1.01, 95% confidence interval: 0.29 to 1.49, P ¼ .93). Conclusions: In our large sample of patients who underwent simple knee arthroscopy, there was no association between preoperative antibiotic use and postoperative deep or superficial infection rates at the 95% confidence level (P ¼ .05). There was an association between preoperative antibiotic use and a decreased deep infection rate at the P ¼ .10 level. Level of Evidence: Level IV, case series.
K
nee arthroscopy is the most commonly performed orthopaedic procedure in the United States, with approximately 1 million procedures performed annually.1 Studies have reported postoperative infection rates after knee arthroscopy that range from 0.1% to 3.4%.2-6 Because of the low incidence of infection, studies to determine the effect of preoperative antibiotics on the postoperative infection rate require large sample sizes to achieve sufficient statistical power and most studies to date have been underpowered.2-7
From the Department of Orthopedic Surgery, Kaiser-Permanente Walnut Creek (R.W.B.W.), Walnut Creek; Department of Orthopedic Surgery, KaiserPermanente Baldwin Park (G.B.M.), Baldwin Park; and Division of Research, Kaiser Permanente Northern California, (L.L.L., J.S., A.L.A.), Oakland, California, U.S.A. The author reports the following conflict of interest or source of funding: Kaiser Permanente Northern California (KPNC) Community Benefit Grant, CN-12RWyatt-01-H. Received July 2, 2015; accepted May 10, 2016. Address correspondence to Ronald W.B. Wyatt, M.D., Department of Orthopedic Surgery, Kaiser-Permanente, Walnut Creek, CA 94596, U.S.A. E-mail: [email protected] Ó 2016 by the Arthroscopy Association of North America 0749-8063/15611/$36.00 http://dx.doi.org/10.1016/j.arthro.2016.05.020
Given the lack of medical evidence regarding the effectiveness of preoperative antibiotics before knee arthroscopy, surgeons have had to decide whether to administer preoperative antibiotics without the benefit of a sufficient evidence base. A survey of orthopaedic surgeons on practices associated with knee arthroscopy reported that the primary reason why surgeons gave preoperative antibiotics was medicolegal concerns.4,8 This practice of “defensive” medicine, with tests and treatments of unproven value being administered to avoid potential future lawsuits, is a known and serious problem in the provision of low-value medical care.9-11 The purpose of this study was to determine the effect of administering a preoperative antibiotic to patients undergoing knee arthroscopy on the incidence of postoperative infection. We hypothesized that there would be an association between the administration of a preoperative antibiotic and the postoperative infection rate.
Methods Kaiser Permanente in California is a large, integrated health care delivery system caring for more than 7.4 million persons and is broadly representative of the statewide population.12 This study was a retrospective
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol
-,
No
-
(Month), 2016: pp 1-6
1
2
R. W. B. WYATT ET AL.
cohort study of California Kaiser Permanente patients. We identified patients who underwent 1 or more simple knee arthroscopy procedures (defined below) between January 1, 2007, and December 31, 2012, and had 3 months of membership after the procedure. Using health plan databases, data on age, sex, selfreported race, or ethnic group, preoperative diagnosis, operative procedure, and the use of preoperative antibiotics were collected. Clinical characteristics studied included American Society of Anesthesiology (ASA) scores, diagnosis of diabetes, and body mass index (BMI). Procedure variables included whether a preoperative antibiotic was administered and the type of antibiotic, anesthesia type, postoperative diagnosis (based on operative report text string search), procedure performed (based on the International Classification of Diseases [9th revision] codes), and operative time. Whether or not a preoperative antibiotic was administered was at the discretion of the individual surgeon. Patients with the following diagnoses or conditions were excluded: knee sepsis, previous ipsilateral knee replacement, patients who had taken therapeutic (not prophylactic) antibiotics within 24 hours of the knee arthroscopy, patients undergoing another surgical procedure at the time of the knee arthroscopy, and patients who received their preoperative antibiotic more than 2 hours before the surgical incision. For patients who had more than 1 knee arthroscopy during the study period only the first procedure was included. Simple knee arthroscopy included the following procedures: diagnostic arthroscopy, joint debridement, synovectomy, partial or complete meniscectomy, meniscus repair, microfracture, and lateral retinacular release. Meniscus repairs included implants and often a nonportal incision. Complex knee arthroscopy procedures were excluded, including ligament reconstruction, meniscus transplant, cartilage restoration procedures (except microfracture), procedures associated with fractures (such as tibial plateau fracture), and arthroscopic procedures that proceeded to arthrotomy. We also excluded cases lasting longer than 120 minutes. A data mining algorithm, which has previously been described,13 was used to determine which patients developed a deep or superficial infection within 90 days of the simple knee arthroscopy. All of the infections that were identified by data mining were confirmed by manual review of the electronic medical record. Deep infection was defined as (1) a positive culture of the knee synovial fluid or (2) clinical diagnosis based on the patient’s signs or symptoms (severe knee pain with range of motion, fever, effusion, erythema, increased warmth), positive laboratory findings (increased serum white blood count, increased erythrocyte sedimentation rate, knee fluid white cell count greater than 50,000 per cubic millimeter,
positive gram stain), and a finding of gross pus in the knee at therapeutic arthroscopy. A superficial infection was defined as clinical signs of localized infection at a portal site (erythema, tenderness, swelling, drainage) that was treated with an oral antibiotic, with or without a single dose of an intravenous antibiotic. Statistical Analysis Demographic and clinical characteristics were examined in bivariate analysis comparing the groups that were administered antibiotics and those that were not. Categorical variables were evaluated using frequencies and proportions and associations were tested with c2 tests. Continuous variables were evaluated using means and t-tests. Medians and Mann-Whitney U-tests were used to compare continuous but non-normally distributed data. Multiple logistic regression was used to identify independent risk factors associated with infection. The primary predictor was administration of a preoperative antibiotic. Potential confounders included were BMI, gender, race, incision duration, age, diagnosis of diabetes, and ASA rating. SAS 9.3 was used for all analysis. A 2-sided alpha level of 0.05 was considered significant. An a priori power calculation was performed before conducting this study. The assumptions used in and the results of these calculations are as follows. Hypothesis: There is no association between receipt of preoperative prophylactic antibiotics and risk of postoperative infection among patients undergoing simple knee arthroscopy. Alpha: 0.05 (2-tailed). Beta: 0.2 (Power ¼ 80%). Clinically relevant change in relative risk ¼ 0.5. Baseline (control) risk of postoperative infection ¼ 0.003. Ratio of antibiotic-treated to untreated patients: 2.33 (estimate based on an initial chart review). The statistical test on which calculations are based: z-test (pooled variance). The results of our initial power calculations indicated that the number of subjects required to answer our research question was 39,105. However, after data collection, we found that the ratio of antibiotic-treated patients to noneantibiotic-treated patients was 4.1 (not 2.33), which reduced our power, but the baseline risk of infection was found to be 0.54% (not 0.3%), which had the effect of increasing our power. Repeating the power calculation for the actual observed values resulted in a requirement for 29,529 patients, a value well below the actual 40,810 we analyzed. This indicates that our study was well powered to address the research question.
Results During the study period, there were 40,810 patients who underwent a simple knee arthroscopy. Of these patients, 32,836 (80.5%) received a preoperative
3
ANTIBIOTICS FOR KNEE ARTHROSCOPY
antibiotic and 7,974 (19.5%) did not. The 2 groups were similar with respect to patient demographics and clinical characteristics (Table 1). There were 202 postoperative infections in the study group for an overall infection rate of 0.50% (Table 2). Of the patients who received a preoperative antibiotic, 159 (0.48%) developed an infection; among those patients who did not receive a preoperative antibiotic, 43 (0.54%) developed an infection (risk ratio [RR] ¼ 0.90, 95% confidence interval [CI]: 0.64 to 1.26, P ¼ .53). A deep infection occurred in 36 patients (0.09%), and of these patients, 25 (0.08%) were in the antibiotic group and 11 (0.14%) were in the no-antibiotics group. This difference (RR ¼ 0.55, 95% CI: 0.27 to 1.12, P ¼ .10) did not reach conventional levels of
statistical significance (P ¼ .05) but was significant at the 0.10 level. A superficial infection occurred in 166 patients (0.41%). Of the patients with superficial infection, 134 were in the antibiotic group (0.41%) and 32 were in the no-antibiotics group (0.40%). There was no significant association between administration of a preoperative antibiotic and the risk of superficial postoperative infection (RR ¼ 1.01, 95% CI: 0.69 to 1.49, P ¼ .93). All of the superficial infections responded to antibiotics and local wound care and none progressed to deep infection. Of the patients who received an antibiotic, cefazolin was administered to 29,662 (90.3%), clindamycin to 2,742 (8.4%), vancomycin to 472 (1.4%),
Table 1. Comparison of Patient Demographics and Procedure Characteristics by Antibiotic Status* Categoryy No. of study patients Age Mean Less than 40 yr 40 yr and above Sex Male Female ASA classz ASA 1 ASA 2 ASA 3/4 Race White Hispanic Black Asian Other BMI Mean Less than 25 25 to less than 30 30 and above Diagnosis of diabetes Postoperative diagnosisx Meniscus tear Osteoarthritis Patella disorder Loose body Chondral injury Ligament injury Other Operative timejj Mean Less than 60 min 60 min and above
Antibiotics 32,836 (80.5)
No Antibiotics 7,974 (19.5)
Total 40,810 (100.0)
P Value
47.6 yr 8,654 (26.4) 24,182 (73.6)
48.7 yr 1,954 (24.5) 6,020 (75.5)
47.8 yr
<.0001 .0007
18,199 (55.4) 14,637 (44.6)
4,230 (53.0) 3,744 (47.0)
22,429 (55.0) 18,381 (45.0)
7,118 (21.7) 17,906 (54.5) 3,240 (9.9)
1,442 (18.1) 4,047 (50.8) 801 (10.0)
8,560 (24.8) 21,953 (63.5) 4,041 (11.7)
18,075 8,843 2,845 1,788 1,285
4,430 2,051 669 434 390
22,505 10,894 3,514 2,222 1,675
.0001
<.0001
.0007 (55.0) (26.9) (8.7) (5.5) (3.9)
(55.6) (25.7) (8.4) (19.9) (4.9)
(55.1) (26.7) (8.6) (5.5) (4.1)
29.9 6,948 (21.2) 11,696 (35.6) 14,173 (43.2) 3,845 (11.7)
29.8 1,706 (21.4) 2,878 (36.1) 3,387 (42.5) 865 (10.8)
29.9 8,654 (21.2) 14,574 (35.7) 17,560 (43.0) 4,710 (100.0)
25,881 3,635 2,848 916 634 469 1,138
5,849 1,205 761 202 127 74 227
31,730 4,840 3,609 1,118 761 543 1,365
(78.8) (11.1) (8.7) (2.8) (1.9) (1.4) (3.5)
29.2 min 30,166 (91.9) 1,343 (4.1)
(73.4) (15.1) (9.5) (2.5) (1.6) (0.9) (2.8)
26.5 min 7,412 (93.0) 179 (2.2)
.1965 .6235
.0307
(77.8) (11.9) (8.8) (2.7) (1.9) (1.3) (3.3)
<.0001 <.0001 .0141 .2083 .0543 .0005 .0058
28.7 min 37,578 (92.1) 1,522 (3.7)
<.0001 <.0001
ASA, American Society of Anesthesiology; BMI, body mass index. *Values are expressed as n (%) unless otherwise indicated. y Associations and P values for age, ASA class, race, and BMI were determined for the entire category by c2 tests. z ASA class missing: n ¼ 6,256 (antibiotic groups 4,572, no-antibiotics group 1,684). x Some knee arthroscopies had multiple postoperative diagnoses (total ¼ 43,966); therefore individual c2 tests were performed for each item in this category. jj Operative time missing: n ¼ 1,710 (antibiotic groups: 1,327; no-antibiotics group: 383).
4
R. W. B. WYATT ET AL.
Table 2. Infections by Antibiotic Status Category No. of study patients (%) Infection (%) Deep Superficial Antibiotic* (%) Cefazolin Clindamycin Vancomycin Other
Antibiotics 32,836 (80.5) 159 (0.48) 25 (0.08) 134 (0.41) 29,662 2,742 472 17
No Antibiotics 7,974 (19.5) 43 (0.54) 11 (0.14) 32 (0.40)
Total 40,810 (100) 202 (0.50) 36 (0.09) 166 (0.41)
Risk Ratio
95% Confidence Interval
P Value
0.90 0.55 1.01
0.64-1.26 0.27-1.12 0.69-1.49
.53 .10 .93
(90.3) (8.4) (1.4) (0.1)
*57 knee arthroscopies had more than 1 antibiotic given (total ¼ 32,893).
Univariate, but not multivariate, analysis showed a decreased likelihood of infection among Hispanics compared with white patients. The only variable associated with the likelihood of an infection in the multivariate analysis was age; we found that age more than 40 years was associated with an increased risk of infection (RR ¼ 1.57, 95% CI 1.10 to 2.24, P ¼ .01). However, there was no interaction between age and the effect of antibiotics on the risk of any postoperative infection (P ¼ .85). The other variables studied (race, gender, BMI, ASA, diagnosis of diabetes, operative time) were not associated risk of postoperative infection.
and 17 (.04%) received another antibiotic (Table 2). More than 1 antibiotic was administered to 57 (0.14%) patients. For patients with deep infection, coagulase-negative Staphylococcus (S.) was the most common organism (n ¼ 13, [36%]) followed by S. aureus (n ¼ 10, [27%]), methicillin-resistant S. aureus (n ¼ 4, [11%]), Enterobacteria (n ¼ 2, [6%]), Streptococcal species (n ¼ 2, [6%]), and Serratia marcescens (n ¼ 1, [3%]). In 4 patients (11%), the synovial fluid cultures were negative. During the study period, 2,649 patients had a total of 3,016 additional simple knee arthroscopies performed on the ipsilateral knee. However, only data from the first knee arthroscopy were included in this study. Of these 3,016 surgeries, there were 5 deep and 20 superficial postoperative infections. A comparison of the infection rates for these additional surgeries versus the infection rates for the study population was as follows: overalld0.83% versus 0.50%, deepd0.16% versus .09%, and superficiald0.66% versus 0.41%. Because of the small number patients who had additional surgeries and the number of dependent variables in this group, a statistical analysis was not performed. Both univariate and multivariate regression analyses for postoperative infections were performed (Table 3).
Discussion In this study of 40,810 patients who underwent simple knee arthroscopy, we found no statistical association between the administration of a preoperative antibiotic and the postoperative infection rate at the P ¼ .05 level. We did find a statistical association between giving a preoperative antibiotic and the deep infection rate at the P ¼ .10 level. Whether prophylactic antibiotics should be given before simple knee arthroscopy is controversial. Because of the rare occurrence of postoperative infection after simple knee arthroscopy, previous studies
Table 3. Logistic Regression Predicting Postoperative Infection Univariate Models Category Antibiotic (yes v no) BMI Less than 25 v less than 30 Less than 25 v 30 and greater Sex (male v female) Operative time (greater than or equal to 60 min v less than 60) Age (greater than or equal to 40 yr v less than 40 yr) Race Hispanic v white Black v white Asian v White Other v white Diabetes (yes v no) ASA class (2, 3/4 v 1)
Odds Ratio 0.897
95% CI 0.64-1.26
1.000 0.887 0.886 0.939 1.569
0.69-1.45 0.62-1.28 0.67-1.17 0.44-2.00 1.1-2.24
0.666 0.907 0.717 0.739 0.842 1.289
0.47-0.95 0.55-1.49 0.36-1.41 0.35-1.59 0.53-1.34 0.88-1.89
ASA, American Society of Anesthesiology; BMI, body mass index; CI, confidence interval.
Multivariate Model P Value .53 .70 .99 .52 .39 .87 .01 .21 .02 .70 .34 .44 .47 .19
Odds Ratio 0.904
95% CI 0.65-1.27
P Value .56
1.567
1.10-2.24
.01
ANTIBIOTICS FOR KNEE ARTHROSCOPY
have not had a sufficient number of patients to reach meaningful conclusions. Bert et al.4 performed a retrospective study on the incidence of infection after routine arthroscopic meniscectomy with and without prophylactic antibiotics. Over a 3-year period, 2,780 arthroscopic cases were collected. Of the 933 patients who received an antibiotic, 1 (0.11%) had a deep infection. Of the 1,847 who did not receive an antibiotic, 3 (0.16%) had a deep infection. Although this difference was not statistically significant (P ¼ .59), this study had an insufficient sample size to rule out a clinically significant difference. Two other studies had similar low numbers of study patients. Wieck et al.5 reported on 437 patients who had arthroscopy without receiving an antibiotic. No postoperative infections occurred, and they concluded that “the routine use of antibiotics is not indicated.” However, the 95% CI for the risk of postoperative infection in this sample included an upper limit of 0.84%, so the validity of this statement is questionable, given the low number of patients and lack of controls. Rose et al.6 reported on 302 patients undergoing knee arthroscopy, with 14 (4.6%) receiving a prophylactic antibiotic and 288 (95.3%) receiving no antibiotic. There was only 1 deep infection in the group receiving no antibiotic. Again, no conclusions are possible given the low number of patients. Kurzwell,10 in a Level V evidence paper on antibiotic prophylaxis for arthroscopic surgery, opined that “antibiotics should be routinely prescribed in patients who undergo arthroscopic surgery.” However, he also reported that “a review of the literature yields no studies with statistical power adequate to resolve the controversy regarding the use of prophylactic antibiotics in arthroscopy.” Kurzwell cites the medicolegal concern, stating that “it may be easier for a surgeon to argue that everything was done to prevent infection if antibiotics were given than to explain the reasons for not giving them.” However, surgeons may also be responsible for the harms of indiscriminate use of antibiotics in the absence of evidence that they provide meaningful clinical benefit. The “harm” or potential negative consequences of antibiotic use include allergic reaction, drugadministration errors, and the development of resistant organisms. Allergic reactions from antibiotics can range from a mild rash to anaphylaxis. Although most cases of anaphylaxis during surgical procedures are from neuromuscular agents or latex, reactions from antibiotics are also a potential cause.14-18 With the administration of any drug, there is the possibility of medication error. A study by Treiber and Jone19 on perioperative medication administration reported that preoperative medication errors were the most frequently reported. Another potential adverse
5
effect of the administration of antibiotics is the development of resistant organisms. Campbell et al.,11 in a paper on “antibiotic stewardship,” encouraged the “elimination of extraneous and inappropriate antibiotic use.” Their guidelines recommended no antimicrobial prophylaxis for arthroscopy without an implant. However, they did not include supporting evidence for this recommendation. Although the actual cost of antibiotics routinely given preoperatively for knee arthroscopy is relatively low, there are a number of other “costs” associated with drug administration. Kerr et al.20 defined several other categories than the cost of the drug itself, including drug delivery, drug monitoring, dosage adjustments, general monitoring, and sharps disposal. Like any clinical decision, an explicit balancing of benefits and risks is required to formulate a rational approach to the issue of preprocedure antibiotics for simple knee arthroscopy. In this regard, it is important to appreciate the distinction between absolute and relative risks. We found the point estimate for the relative risk reduction of deep infection for patients given antibiotics was 0.55 (95% CI: 0.27 to 1.12), though this association was not statistically significant. However, the absolute risks were low, so that the absolute risk reduction was only 0.05%. Viewed another way, this means that the number needed to treat ¼ 2,000; that is, 2,000 patients would need to be treated with preoperative antibiotics to prevent 1 deep joint infection. The strength of our study is the large number of patients, with data collected from a single health care entity. Limitations A weakness of this study is the number of variables that were not studied that may influence the infection rate. Variables that were not studied included tourniquet use, the use of a high volume fluid pump versus gravity, the perioperative injection of anesthetics, steroids, or other medications, portal closure technique (suture, tape, other), surgery setting (hospital, surgery center, procedure room, other), physician case volume or training, and weight-adjusted antibiotic dosing. Another weakness is that we were unable to identify which study patients may have had ipsilateral knee arthroscopy before the study period. Our data did show that for patients who had additional ipsilateral knee arthroscopies performed during the study period, there was a higher infection rate for subsequent surgeries. Because of the relatively small number of patients, and the difficulties in evaluating the dependent variables associated with multiple knee arthroscopies, we included only the data from the first knee arthroscopy. Furthermore, the patient’s socioeconomic status, smoking status, recent illnesses, skin vitality, and breaks
6
R. W. B. WYATT ET AL.
in surgical technique were not available in our databases. We were also unable to obtain meaningful data on any adverse events that may have resulted from the administration of the preoperative antibiotics. Should preoperative antibiotics be given in simple knee arthroscopy? Our data showed that the risk of postoperative knee infection after simple knee arthroscopy within our health care system was very low. We did not find a statistically significant association between administration of a preoperative antibiotic and a decreased infection rate at 95% confidence level (P < .5%). However, there was a trend toward a decreased risk of deep postoperative infections with the use of preoperative antibiotics (P ¼ .10). That is, under the null hypothesis of no effect, and in the absence of bias or confounding, there is an approximate 10% chance that an effect at least as strong as that observed in this study could have arisen by chance alone. This study provides data for knee arthroscopists on the potential benefits of antibiotic use, as well as a literature view and discussion of the potential negative consequences. Hopefully, this information will help the practicing orthopaedic surgeon to decide when a preoperative antibiotic is appropriate for a knee arthroscopy patient.
Conclusions In our large sample of patients who underwent simple knee arthroscopy, there was no association between preoperative antibiotic use and postoperative deep or superficial infection rates at the 95% confidence level (P ¼ .05). There was an association between preoperative antibiotic use and a decreased deep infection rate at the P ¼ .10 level.
Acknowledgment Oversight of this study was provided by the Kaiser Permanente Northern California Institutional Review Board.
References 1. Kim S, Bosque J, Meehan J, Jamali A, Marder R. Increase in outpatient knee arthroscopy in the United States: A comparison of National Surveys of Ambulatory Surgery, 1996 and 2006. J Bone Joint Surg Am 2011;93: 994-1000. 2. Sherman OH, Fox JM, Snyder SJ, et al. Arthroscopyd“no problem surgery” an analysis of complications in two thousand six hundred and forty cases. J Bone Joint Surg Am 1986;68:256-265. 3. Armstrong RW, Bolding F, Joseph R. Septic arthritis following arthroscopy: Clinical syndromes and analysis of risk factors. Arthroscopy 1992;8:213-223.
4. Bert JM, Giannini D, Nace L. Antibiotic prophylaxis for arthroscopy of the knee: Is it necessary? Arthroscopy 2007;23:4-6. 5. Wieck JA, Jackson JK, O’Brien TJ, Lurate RB, Russell JM, Dorchak JD. Efficacy of prophylactic antibiotics in arthroscopic surgery. Orthopedics 1997;20:133-134. 6. Rose R, Ameerally A, Frankson M, Henry H. Knee arthroscopy: Surgical site infections and the need for prophylactic antibiotics. Internet J Orthop Surg. 2007;10. http://ispub.com/IJOS/10/2/4080. Accessed March 3, 2014. 7. D’Angelo GL, Olgilvie-Harris DJ. Septic arthritis following arthroscopy, with cost/benefit analysis of antibiotic prophylaxis. Arthroscopy 1988;4:10-14. 8. No benefit from prophylactic antibiotics in routine arthroscopy. Orthopedics Today. http://www.healio.com/ orthopedics/arthroscopy/news/online/%7Be6606748-ab574055-867a-081bf8b2624d%7D/no-benefit-from-prophylacticantibiotics-in-routine-arthroscopy. Published May 17, 2005. Accessed April 19, 2013. 9. Sethi MK, Obremsky WT, Natividad H, Mir HR, Jahangir AA. Incidence and costs of defensive medicine among orthopedic surgeons in the United States: A national survey study. Am J Orthop 2012;41:69-73. 10. Kurzwell PR. Antibiotic prophylaxis for arthroscopic surgery. Arthroscopy 2006;22:452-454. 11. Campbell KA, Stein S, Looze C, Bosco JA. Antibiotic stewardship in orthopedic surgery: Principles and practice. J Am Acad Orthop Surg 2014;22:772-781. 12. Koebnick C, Langer-Gould AM, Gould MK, et al. Sociodemographic characteristics of members of a large, integrated health care system: Comparison with US census bureau data. Perm J 2012;16:37-41. 13. Inacio MC, Paxton EW, Chen Y, et al. Leveraging electronic medical records for surveillance of surgical site infection in a total joint replacement population. Infect Control Hosp Epidemiol 2011;32:351-359. 14. Gruchalla RS, Pirmohamed M. Clinical practice. Antibiotic allergy. N Engl J Med 2006;354:601-609. 15. Levy JH, Castells MC. Perioperative anaphylaxis and the United States perspective. Anesth Analg 2011;113: 979-981. 16. Mertes PM, Tajima K, Regnier-Kimmoun MA, et al. Perioperative anaphylaxis. Med Clin North Am 2010;94: 761-789. 17. Liberman P. Anaphylactic reactions during surgical and medical procedures. J Allergy Clin Immunol 2002;110: 64-69. 18. Laxenair MC, Mertes PM. Groupe d’Etudes des Réactions Anaphylactoïdes Peranesthésiques. Anaphylaxis during anesthesia. Results of a two-year survey in France. Br J Anaesth 2001;87:549-558. 19. Treiber LA, Jone JH. Medication errors, routines, differences between perioperative and non-perioperative nurses. AORN J 2005;96:285-294. 20. Kerr JR, Barr JG, Smyth ET, O’Hare J, Bell PM, Callender ME. Antibiotic pharmacoeconomics: An attempt to find the real cost of hospital antibiotic prescribing. Ulster Med J 1993;62:50-57.
Purpose: To determine the association between the use of preoperative antibiotics and the risk of postoperative infection after simple knee arthroscopy. Methods: The electronic medical records of a large integrated health care organization were used to identify patients who underwent simple knee arthroscopy between 2007 and 2012. Patient demographics, potential infection risk factors, and antibiotic administration data were extracted. Simple knee arthroscopy included debridement, meniscectomy, meniscus repair, synovectomy, microfracture, and lateral release. Complex knee arthroscopy, septic knees, and cases involving fractures were excluded. Deep infection was defined as a positive synovial fluid culture or signs and symptoms of infection and gross pus in the knee. Superficial infection was defined as clinical signs of infection localized to a portal site and treatment with an antibiotic. Results: Of 40,810 simple knee arthroscopies, 32,836 (80.5%) received preoperative antibiotics and 7,974 (19.5%) did not. There were 25 deep infections in the antibiotic group (0.08%) and 11 in the no-antibiotics group (0.14%) (risk ratio ¼ 0.55, 95% confidence interval: 0.27 to 1.12, P ¼ .10). There were 134 superficial infections in the antibiotic group (0.41%) and 32 in the no-antibiotics group (0.40%) (risk ratio ¼ 1.01, 95% confidence interval: 0.29 to 1.49, P ¼ .93). Conclusions: In our large sample of patients who underwent simple knee arthroscopy, there was no association between preoperative antibiotic use and postoperative deep or superficial infection rates at the 95% confidence level (P ¼ .05). There was an association between preoperative antibiotic use and a decreased deep infection rate at the P ¼ .10 level. Level of Evidence: Level IV, case series.
K
nee arthroscopy is the most commonly performed orthopaedic procedure in the United States, with approximately 1 million procedures performed annually.1 Studies have reported postoperative infection rates after knee arthroscopy that range from 0.1% to 3.4%.2-6 Because of the low incidence of infection, studies to determine the effect of preoperative antibiotics on the postoperative infection rate require large sample sizes to achieve sufficient statistical power and most studies to date have been underpowered.2-7
From the Department of Orthopedic Surgery, Kaiser-Permanente Walnut Creek (R.W.B.W.), Walnut Creek; Department of Orthopedic Surgery, KaiserPermanente Baldwin Park (G.B.M.), Baldwin Park; and Division of Research, Kaiser Permanente Northern California, (L.L.L., J.S., A.L.A.), Oakland, California, U.S.A. The author reports the following conflict of interest or source of funding: Kaiser Permanente Northern California (KPNC) Community Benefit Grant, CN-12RWyatt-01-H. Received July 2, 2015; accepted May 10, 2016. Address correspondence to Ronald W.B. Wyatt, M.D., Department of Orthopedic Surgery, Kaiser-Permanente, Walnut Creek, CA 94596, U.S.A. E-mail: [email protected] Ó 2016 by the Arthroscopy Association of North America 0749-8063/15611/$36.00 http://dx.doi.org/10.1016/j.arthro.2016.05.020
Given the lack of medical evidence regarding the effectiveness of preoperative antibiotics before knee arthroscopy, surgeons have had to decide whether to administer preoperative antibiotics without the benefit of a sufficient evidence base. A survey of orthopaedic surgeons on practices associated with knee arthroscopy reported that the primary reason why surgeons gave preoperative antibiotics was medicolegal concerns.4,8 This practice of “defensive” medicine, with tests and treatments of unproven value being administered to avoid potential future lawsuits, is a known and serious problem in the provision of low-value medical care.9-11 The purpose of this study was to determine the effect of administering a preoperative antibiotic to patients undergoing knee arthroscopy on the incidence of postoperative infection. We hypothesized that there would be an association between the administration of a preoperative antibiotic and the postoperative infection rate.
Methods Kaiser Permanente in California is a large, integrated health care delivery system caring for more than 7.4 million persons and is broadly representative of the statewide population.12 This study was a retrospective
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol
-,
No
-
(Month), 2016: pp 1-6
1
2
R. W. B. WYATT ET AL.
cohort study of California Kaiser Permanente patients. We identified patients who underwent 1 or more simple knee arthroscopy procedures (defined below) between January 1, 2007, and December 31, 2012, and had 3 months of membership after the procedure. Using health plan databases, data on age, sex, selfreported race, or ethnic group, preoperative diagnosis, operative procedure, and the use of preoperative antibiotics were collected. Clinical characteristics studied included American Society of Anesthesiology (ASA) scores, diagnosis of diabetes, and body mass index (BMI). Procedure variables included whether a preoperative antibiotic was administered and the type of antibiotic, anesthesia type, postoperative diagnosis (based on operative report text string search), procedure performed (based on the International Classification of Diseases [9th revision] codes), and operative time. Whether or not a preoperative antibiotic was administered was at the discretion of the individual surgeon. Patients with the following diagnoses or conditions were excluded: knee sepsis, previous ipsilateral knee replacement, patients who had taken therapeutic (not prophylactic) antibiotics within 24 hours of the knee arthroscopy, patients undergoing another surgical procedure at the time of the knee arthroscopy, and patients who received their preoperative antibiotic more than 2 hours before the surgical incision. For patients who had more than 1 knee arthroscopy during the study period only the first procedure was included. Simple knee arthroscopy included the following procedures: diagnostic arthroscopy, joint debridement, synovectomy, partial or complete meniscectomy, meniscus repair, microfracture, and lateral retinacular release. Meniscus repairs included implants and often a nonportal incision. Complex knee arthroscopy procedures were excluded, including ligament reconstruction, meniscus transplant, cartilage restoration procedures (except microfracture), procedures associated with fractures (such as tibial plateau fracture), and arthroscopic procedures that proceeded to arthrotomy. We also excluded cases lasting longer than 120 minutes. A data mining algorithm, which has previously been described,13 was used to determine which patients developed a deep or superficial infection within 90 days of the simple knee arthroscopy. All of the infections that were identified by data mining were confirmed by manual review of the electronic medical record. Deep infection was defined as (1) a positive culture of the knee synovial fluid or (2) clinical diagnosis based on the patient’s signs or symptoms (severe knee pain with range of motion, fever, effusion, erythema, increased warmth), positive laboratory findings (increased serum white blood count, increased erythrocyte sedimentation rate, knee fluid white cell count greater than 50,000 per cubic millimeter,
positive gram stain), and a finding of gross pus in the knee at therapeutic arthroscopy. A superficial infection was defined as clinical signs of localized infection at a portal site (erythema, tenderness, swelling, drainage) that was treated with an oral antibiotic, with or without a single dose of an intravenous antibiotic. Statistical Analysis Demographic and clinical characteristics were examined in bivariate analysis comparing the groups that were administered antibiotics and those that were not. Categorical variables were evaluated using frequencies and proportions and associations were tested with c2 tests. Continuous variables were evaluated using means and t-tests. Medians and Mann-Whitney U-tests were used to compare continuous but non-normally distributed data. Multiple logistic regression was used to identify independent risk factors associated with infection. The primary predictor was administration of a preoperative antibiotic. Potential confounders included were BMI, gender, race, incision duration, age, diagnosis of diabetes, and ASA rating. SAS 9.3 was used for all analysis. A 2-sided alpha level of 0.05 was considered significant. An a priori power calculation was performed before conducting this study. The assumptions used in and the results of these calculations are as follows. Hypothesis: There is no association between receipt of preoperative prophylactic antibiotics and risk of postoperative infection among patients undergoing simple knee arthroscopy. Alpha: 0.05 (2-tailed). Beta: 0.2 (Power ¼ 80%). Clinically relevant change in relative risk ¼ 0.5. Baseline (control) risk of postoperative infection ¼ 0.003. Ratio of antibiotic-treated to untreated patients: 2.33 (estimate based on an initial chart review). The statistical test on which calculations are based: z-test (pooled variance). The results of our initial power calculations indicated that the number of subjects required to answer our research question was 39,105. However, after data collection, we found that the ratio of antibiotic-treated patients to noneantibiotic-treated patients was 4.1 (not 2.33), which reduced our power, but the baseline risk of infection was found to be 0.54% (not 0.3%), which had the effect of increasing our power. Repeating the power calculation for the actual observed values resulted in a requirement for 29,529 patients, a value well below the actual 40,810 we analyzed. This indicates that our study was well powered to address the research question.
Results During the study period, there were 40,810 patients who underwent a simple knee arthroscopy. Of these patients, 32,836 (80.5%) received a preoperative
3
ANTIBIOTICS FOR KNEE ARTHROSCOPY
antibiotic and 7,974 (19.5%) did not. The 2 groups were similar with respect to patient demographics and clinical characteristics (Table 1). There were 202 postoperative infections in the study group for an overall infection rate of 0.50% (Table 2). Of the patients who received a preoperative antibiotic, 159 (0.48%) developed an infection; among those patients who did not receive a preoperative antibiotic, 43 (0.54%) developed an infection (risk ratio [RR] ¼ 0.90, 95% confidence interval [CI]: 0.64 to 1.26, P ¼ .53). A deep infection occurred in 36 patients (0.09%), and of these patients, 25 (0.08%) were in the antibiotic group and 11 (0.14%) were in the no-antibiotics group. This difference (RR ¼ 0.55, 95% CI: 0.27 to 1.12, P ¼ .10) did not reach conventional levels of
statistical significance (P ¼ .05) but was significant at the 0.10 level. A superficial infection occurred in 166 patients (0.41%). Of the patients with superficial infection, 134 were in the antibiotic group (0.41%) and 32 were in the no-antibiotics group (0.40%). There was no significant association between administration of a preoperative antibiotic and the risk of superficial postoperative infection (RR ¼ 1.01, 95% CI: 0.69 to 1.49, P ¼ .93). All of the superficial infections responded to antibiotics and local wound care and none progressed to deep infection. Of the patients who received an antibiotic, cefazolin was administered to 29,662 (90.3%), clindamycin to 2,742 (8.4%), vancomycin to 472 (1.4%),
Table 1. Comparison of Patient Demographics and Procedure Characteristics by Antibiotic Status* Categoryy No. of study patients Age Mean Less than 40 yr 40 yr and above Sex Male Female ASA classz ASA 1 ASA 2 ASA 3/4 Race White Hispanic Black Asian Other BMI Mean Less than 25 25 to less than 30 30 and above Diagnosis of diabetes Postoperative diagnosisx Meniscus tear Osteoarthritis Patella disorder Loose body Chondral injury Ligament injury Other Operative timejj Mean Less than 60 min 60 min and above
Antibiotics 32,836 (80.5)
No Antibiotics 7,974 (19.5)
Total 40,810 (100.0)
P Value
47.6 yr 8,654 (26.4) 24,182 (73.6)
48.7 yr 1,954 (24.5) 6,020 (75.5)
47.8 yr
<.0001 .0007
18,199 (55.4) 14,637 (44.6)
4,230 (53.0) 3,744 (47.0)
22,429 (55.0) 18,381 (45.0)
7,118 (21.7) 17,906 (54.5) 3,240 (9.9)
1,442 (18.1) 4,047 (50.8) 801 (10.0)
8,560 (24.8) 21,953 (63.5) 4,041 (11.7)
18,075 8,843 2,845 1,788 1,285
4,430 2,051 669 434 390
22,505 10,894 3,514 2,222 1,675
.0001
<.0001
.0007 (55.0) (26.9) (8.7) (5.5) (3.9)
(55.6) (25.7) (8.4) (19.9) (4.9)
(55.1) (26.7) (8.6) (5.5) (4.1)
29.9 6,948 (21.2) 11,696 (35.6) 14,173 (43.2) 3,845 (11.7)
29.8 1,706 (21.4) 2,878 (36.1) 3,387 (42.5) 865 (10.8)
29.9 8,654 (21.2) 14,574 (35.7) 17,560 (43.0) 4,710 (100.0)
25,881 3,635 2,848 916 634 469 1,138
5,849 1,205 761 202 127 74 227
31,730 4,840 3,609 1,118 761 543 1,365
(78.8) (11.1) (8.7) (2.8) (1.9) (1.4) (3.5)
29.2 min 30,166 (91.9) 1,343 (4.1)
(73.4) (15.1) (9.5) (2.5) (1.6) (0.9) (2.8)
26.5 min 7,412 (93.0) 179 (2.2)
.1965 .6235
.0307
(77.8) (11.9) (8.8) (2.7) (1.9) (1.3) (3.3)
<.0001 <.0001 .0141 .2083 .0543 .0005 .0058
28.7 min 37,578 (92.1) 1,522 (3.7)
<.0001 <.0001
ASA, American Society of Anesthesiology; BMI, body mass index. *Values are expressed as n (%) unless otherwise indicated. y Associations and P values for age, ASA class, race, and BMI were determined for the entire category by c2 tests. z ASA class missing: n ¼ 6,256 (antibiotic groups 4,572, no-antibiotics group 1,684). x Some knee arthroscopies had multiple postoperative diagnoses (total ¼ 43,966); therefore individual c2 tests were performed for each item in this category. jj Operative time missing: n ¼ 1,710 (antibiotic groups: 1,327; no-antibiotics group: 383).
4
R. W. B. WYATT ET AL.
Table 2. Infections by Antibiotic Status Category No. of study patients (%) Infection (%) Deep Superficial Antibiotic* (%) Cefazolin Clindamycin Vancomycin Other
Antibiotics 32,836 (80.5) 159 (0.48) 25 (0.08) 134 (0.41) 29,662 2,742 472 17
No Antibiotics 7,974 (19.5) 43 (0.54) 11 (0.14) 32 (0.40)
Total 40,810 (100) 202 (0.50) 36 (0.09) 166 (0.41)
Risk Ratio
95% Confidence Interval
P Value
0.90 0.55 1.01
0.64-1.26 0.27-1.12 0.69-1.49
.53 .10 .93
(90.3) (8.4) (1.4) (0.1)
*57 knee arthroscopies had more than 1 antibiotic given (total ¼ 32,893).
Univariate, but not multivariate, analysis showed a decreased likelihood of infection among Hispanics compared with white patients. The only variable associated with the likelihood of an infection in the multivariate analysis was age; we found that age more than 40 years was associated with an increased risk of infection (RR ¼ 1.57, 95% CI 1.10 to 2.24, P ¼ .01). However, there was no interaction between age and the effect of antibiotics on the risk of any postoperative infection (P ¼ .85). The other variables studied (race, gender, BMI, ASA, diagnosis of diabetes, operative time) were not associated risk of postoperative infection.
and 17 (.04%) received another antibiotic (Table 2). More than 1 antibiotic was administered to 57 (0.14%) patients. For patients with deep infection, coagulase-negative Staphylococcus (S.) was the most common organism (n ¼ 13, [36%]) followed by S. aureus (n ¼ 10, [27%]), methicillin-resistant S. aureus (n ¼ 4, [11%]), Enterobacteria (n ¼ 2, [6%]), Streptococcal species (n ¼ 2, [6%]), and Serratia marcescens (n ¼ 1, [3%]). In 4 patients (11%), the synovial fluid cultures were negative. During the study period, 2,649 patients had a total of 3,016 additional simple knee arthroscopies performed on the ipsilateral knee. However, only data from the first knee arthroscopy were included in this study. Of these 3,016 surgeries, there were 5 deep and 20 superficial postoperative infections. A comparison of the infection rates for these additional surgeries versus the infection rates for the study population was as follows: overalld0.83% versus 0.50%, deepd0.16% versus .09%, and superficiald0.66% versus 0.41%. Because of the small number patients who had additional surgeries and the number of dependent variables in this group, a statistical analysis was not performed. Both univariate and multivariate regression analyses for postoperative infections were performed (Table 3).
Discussion In this study of 40,810 patients who underwent simple knee arthroscopy, we found no statistical association between the administration of a preoperative antibiotic and the postoperative infection rate at the P ¼ .05 level. We did find a statistical association between giving a preoperative antibiotic and the deep infection rate at the P ¼ .10 level. Whether prophylactic antibiotics should be given before simple knee arthroscopy is controversial. Because of the rare occurrence of postoperative infection after simple knee arthroscopy, previous studies
Table 3. Logistic Regression Predicting Postoperative Infection Univariate Models Category Antibiotic (yes v no) BMI Less than 25 v less than 30 Less than 25 v 30 and greater Sex (male v female) Operative time (greater than or equal to 60 min v less than 60) Age (greater than or equal to 40 yr v less than 40 yr) Race Hispanic v white Black v white Asian v White Other v white Diabetes (yes v no) ASA class (2, 3/4 v 1)
Odds Ratio 0.897
95% CI 0.64-1.26
1.000 0.887 0.886 0.939 1.569
0.69-1.45 0.62-1.28 0.67-1.17 0.44-2.00 1.1-2.24
0.666 0.907 0.717 0.739 0.842 1.289
0.47-0.95 0.55-1.49 0.36-1.41 0.35-1.59 0.53-1.34 0.88-1.89
ASA, American Society of Anesthesiology; BMI, body mass index; CI, confidence interval.
Multivariate Model P Value .53 .70 .99 .52 .39 .87 .01 .21 .02 .70 .34 .44 .47 .19
Odds Ratio 0.904
95% CI 0.65-1.27
P Value .56
1.567
1.10-2.24
.01
ANTIBIOTICS FOR KNEE ARTHROSCOPY
have not had a sufficient number of patients to reach meaningful conclusions. Bert et al.4 performed a retrospective study on the incidence of infection after routine arthroscopic meniscectomy with and without prophylactic antibiotics. Over a 3-year period, 2,780 arthroscopic cases were collected. Of the 933 patients who received an antibiotic, 1 (0.11%) had a deep infection. Of the 1,847 who did not receive an antibiotic, 3 (0.16%) had a deep infection. Although this difference was not statistically significant (P ¼ .59), this study had an insufficient sample size to rule out a clinically significant difference. Two other studies had similar low numbers of study patients. Wieck et al.5 reported on 437 patients who had arthroscopy without receiving an antibiotic. No postoperative infections occurred, and they concluded that “the routine use of antibiotics is not indicated.” However, the 95% CI for the risk of postoperative infection in this sample included an upper limit of 0.84%, so the validity of this statement is questionable, given the low number of patients and lack of controls. Rose et al.6 reported on 302 patients undergoing knee arthroscopy, with 14 (4.6%) receiving a prophylactic antibiotic and 288 (95.3%) receiving no antibiotic. There was only 1 deep infection in the group receiving no antibiotic. Again, no conclusions are possible given the low number of patients. Kurzwell,10 in a Level V evidence paper on antibiotic prophylaxis for arthroscopic surgery, opined that “antibiotics should be routinely prescribed in patients who undergo arthroscopic surgery.” However, he also reported that “a review of the literature yields no studies with statistical power adequate to resolve the controversy regarding the use of prophylactic antibiotics in arthroscopy.” Kurzwell cites the medicolegal concern, stating that “it may be easier for a surgeon to argue that everything was done to prevent infection if antibiotics were given than to explain the reasons for not giving them.” However, surgeons may also be responsible for the harms of indiscriminate use of antibiotics in the absence of evidence that they provide meaningful clinical benefit. The “harm” or potential negative consequences of antibiotic use include allergic reaction, drugadministration errors, and the development of resistant organisms. Allergic reactions from antibiotics can range from a mild rash to anaphylaxis. Although most cases of anaphylaxis during surgical procedures are from neuromuscular agents or latex, reactions from antibiotics are also a potential cause.14-18 With the administration of any drug, there is the possibility of medication error. A study by Treiber and Jone19 on perioperative medication administration reported that preoperative medication errors were the most frequently reported. Another potential adverse
5
effect of the administration of antibiotics is the development of resistant organisms. Campbell et al.,11 in a paper on “antibiotic stewardship,” encouraged the “elimination of extraneous and inappropriate antibiotic use.” Their guidelines recommended no antimicrobial prophylaxis for arthroscopy without an implant. However, they did not include supporting evidence for this recommendation. Although the actual cost of antibiotics routinely given preoperatively for knee arthroscopy is relatively low, there are a number of other “costs” associated with drug administration. Kerr et al.20 defined several other categories than the cost of the drug itself, including drug delivery, drug monitoring, dosage adjustments, general monitoring, and sharps disposal. Like any clinical decision, an explicit balancing of benefits and risks is required to formulate a rational approach to the issue of preprocedure antibiotics for simple knee arthroscopy. In this regard, it is important to appreciate the distinction between absolute and relative risks. We found the point estimate for the relative risk reduction of deep infection for patients given antibiotics was 0.55 (95% CI: 0.27 to 1.12), though this association was not statistically significant. However, the absolute risks were low, so that the absolute risk reduction was only 0.05%. Viewed another way, this means that the number needed to treat ¼ 2,000; that is, 2,000 patients would need to be treated with preoperative antibiotics to prevent 1 deep joint infection. The strength of our study is the large number of patients, with data collected from a single health care entity. Limitations A weakness of this study is the number of variables that were not studied that may influence the infection rate. Variables that were not studied included tourniquet use, the use of a high volume fluid pump versus gravity, the perioperative injection of anesthetics, steroids, or other medications, portal closure technique (suture, tape, other), surgery setting (hospital, surgery center, procedure room, other), physician case volume or training, and weight-adjusted antibiotic dosing. Another weakness is that we were unable to identify which study patients may have had ipsilateral knee arthroscopy before the study period. Our data did show that for patients who had additional ipsilateral knee arthroscopies performed during the study period, there was a higher infection rate for subsequent surgeries. Because of the relatively small number of patients, and the difficulties in evaluating the dependent variables associated with multiple knee arthroscopies, we included only the data from the first knee arthroscopy. Furthermore, the patient’s socioeconomic status, smoking status, recent illnesses, skin vitality, and breaks
6
R. W. B. WYATT ET AL.
in surgical technique were not available in our databases. We were also unable to obtain meaningful data on any adverse events that may have resulted from the administration of the preoperative antibiotics. Should preoperative antibiotics be given in simple knee arthroscopy? Our data showed that the risk of postoperative knee infection after simple knee arthroscopy within our health care system was very low. We did not find a statistically significant association between administration of a preoperative antibiotic and a decreased infection rate at 95% confidence level (P < .5%). However, there was a trend toward a decreased risk of deep postoperative infections with the use of preoperative antibiotics (P ¼ .10). That is, under the null hypothesis of no effect, and in the absence of bias or confounding, there is an approximate 10% chance that an effect at least as strong as that observed in this study could have arisen by chance alone. This study provides data for knee arthroscopists on the potential benefits of antibiotic use, as well as a literature view and discussion of the potential negative consequences. Hopefully, this information will help the practicing orthopaedic surgeon to decide when a preoperative antibiotic is appropriate for a knee arthroscopy patient.
Conclusions In our large sample of patients who underwent simple knee arthroscopy, there was no association between preoperative antibiotic use and postoperative deep or superficial infection rates at the 95% confidence level (P ¼ .05). There was an association between preoperative antibiotic use and a decreased deep infection rate at the P ¼ .10 level.
Acknowledgment Oversight of this study was provided by the Kaiser Permanente Northern California Institutional Review Board.
References 1. Kim S, Bosque J, Meehan J, Jamali A, Marder R. Increase in outpatient knee arthroscopy in the United States: A comparison of National Surveys of Ambulatory Surgery, 1996 and 2006. J Bone Joint Surg Am 2011;93: 994-1000. 2. Sherman OH, Fox JM, Snyder SJ, et al. Arthroscopyd“no problem surgery” an analysis of complications in two thousand six hundred and forty cases. J Bone Joint Surg Am 1986;68:256-265. 3. Armstrong RW, Bolding F, Joseph R. Septic arthritis following arthroscopy: Clinical syndromes and analysis of risk factors. Arthroscopy 1992;8:213-223.
4. Bert JM, Giannini D, Nace L. Antibiotic prophylaxis for arthroscopy of the knee: Is it necessary? Arthroscopy 2007;23:4-6. 5. Wieck JA, Jackson JK, O’Brien TJ, Lurate RB, Russell JM, Dorchak JD. Efficacy of prophylactic antibiotics in arthroscopic surgery. Orthopedics 1997;20:133-134. 6. Rose R, Ameerally A, Frankson M, Henry H. Knee arthroscopy: Surgical site infections and the need for prophylactic antibiotics. Internet J Orthop Surg. 2007;10. http://ispub.com/IJOS/10/2/4080. Accessed March 3, 2014. 7. D’Angelo GL, Olgilvie-Harris DJ. Septic arthritis following arthroscopy, with cost/benefit analysis of antibiotic prophylaxis. Arthroscopy 1988;4:10-14. 8. No benefit from prophylactic antibiotics in routine arthroscopy. Orthopedics Today. http://www.healio.com/ orthopedics/arthroscopy/news/online/%7Be6606748-ab574055-867a-081bf8b2624d%7D/no-benefit-from-prophylacticantibiotics-in-routine-arthroscopy. Published May 17, 2005. Accessed April 19, 2013. 9. Sethi MK, Obremsky WT, Natividad H, Mir HR, Jahangir AA. Incidence and costs of defensive medicine among orthopedic surgeons in the United States: A national survey study. Am J Orthop 2012;41:69-73. 10. Kurzwell PR. Antibiotic prophylaxis for arthroscopic surgery. Arthroscopy 2006;22:452-454. 11. Campbell KA, Stein S, Looze C, Bosco JA. Antibiotic stewardship in orthopedic surgery: Principles and practice. J Am Acad Orthop Surg 2014;22:772-781. 12. Koebnick C, Langer-Gould AM, Gould MK, et al. Sociodemographic characteristics of members of a large, integrated health care system: Comparison with US census bureau data. Perm J 2012;16:37-41. 13. Inacio MC, Paxton EW, Chen Y, et al. Leveraging electronic medical records for surveillance of surgical site infection in a total joint replacement population. Infect Control Hosp Epidemiol 2011;32:351-359. 14. Gruchalla RS, Pirmohamed M. Clinical practice. Antibiotic allergy. N Engl J Med 2006;354:601-609. 15. Levy JH, Castells MC. Perioperative anaphylaxis and the United States perspective. Anesth Analg 2011;113: 979-981. 16. Mertes PM, Tajima K, Regnier-Kimmoun MA, et al. Perioperative anaphylaxis. Med Clin North Am 2010;94: 761-789. 17. Liberman P. Anaphylactic reactions during surgical and medical procedures. J Allergy Clin Immunol 2002;110: 64-69. 18. Laxenair MC, Mertes PM. Groupe d’Etudes des Réactions Anaphylactoïdes Peranesthésiques. Anaphylaxis during anesthesia. Results of a two-year survey in France. Br J Anaesth 2001;87:549-558. 19. Treiber LA, Jone JH. Medication errors, routines, differences between perioperative and non-perioperative nurses. AORN J 2005;96:285-294. 20. Kerr JR, Barr JG, Smyth ET, O’Hare J, Bell PM, Callender ME. Antibiotic pharmacoeconomics: An attempt to find the real cost of hospital antibiotic prescribing. Ulster Med J 1993;62:50-57.