- Email: [email protected]
A prospective study on surgical-site infections in thyroid operation
A prospective study on surgical-site infections in thyroid operation Claudia Bures, MD,a,b Tobias Klatte, MD,a,c Monika Gilhofer, BS,d,e Michael Behnke, MD,f Ann-Christin Breier, MD,f Nikolaus Neuhold, MD,d and Michael Hermann, MD,a,b Vienna, Austria, and Berlin, Germany
Background. To evaluate the incidence and the microbe spectrum of surgical-site infections (SSIs) in patients undergoing elective thyroid operation and to develop a risk factor based predictive model. Methods. This prospective study included 6,778 consecutive patients who underwent thyroid operation at a single institution between 2007 and 2012. SSI was defined according to the Centers for Disease Control and Prevention. Regression models were fitted to evaluate risk factors for SSI. A predictive nomogram was constructed from relevant variables in the multivariable analysis. Discrimination and calibration of the nomogram were assessed. Results. The cumulative incidence of SSI after 30 days was 0.49%. The median time from operation to SSI was 7 days (interquartile range, 4–10.5 days). SSI was classified as superficial incisional in 30 cases (93.8%), deep incisional in 1 case (3.1%), and organ/space in 1 case (3.1%). Staphylococcus aureus was the most common isolate. In multivariable analysis, duration of operation (P = .004) and American Society of Anesthesiologists’ score (P = .031) were identified as independent risk factors for SSI. These variables formed the basis of a nomogram, which was validated internally by bootstrapping and reached a predictive accuracy of 70.1%. The calibration curve showed a good agreement between predicted probability and actual observation. Conclusion. The cumulative incidence of SSI in thyroid operation is <0.5%. American Society of Anesthesiologists’ score and the duration of operation are independent risk factors for SSI. Antibiotic prophylaxis may be considered for selected patients based on the individual risk profile. (Surgery 2014;155:675-81.) From the Second Department of Surgery,a Krankenanstalt Rudolfstiftung, Department of Surgery,b Kaiserin-Elisabeth-Spital, Department of Urology,c Medical University of Vienna, Department of Pathology and Microbiology with Hospital Hygiene Team,d Kaiserin-Elisabeth-Spital, and Department of Pathology and Microbiology,e Krankenanstalt Rudolfstiftung, Vienna, Austria; and Nationales Referenzzentrum f€ u r Surveillance von nosokomialen Infektionen,f Charite -Universit€ a tsmedizin, Berlin, Germany
SURGICAL-SITE INFECTIONS (SSIs) are on the most common nosocomial infections1 and cause substantial morbidity to patients and costs to the health care system.2 The Centers for Disease Control and Prevention (CDC) have developed standard criteria to define SSI,3,4 which are regarded the contemporary, international standard. A SSI is defined as infection related to an operative procedure, which occurs at or near the surgical incision within a 30-day period. Thyroid operation is considered a clean procedure in a well-vascularized area. The incidence of SSI is between 0.3 and 2.9%.5-12 Routine Accepted for publication December 6, 2013. Reprint requests: Tobias Klatte, MD, Second Department of Surgery, Krankenanstalt Rudolfstiftung, Juchgasse 25, 1030 Vienna, Austria. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2014 Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2013.12.002
perioperative antibiotic prophylaxis is not recommended by guidelines but should be considered for selected patients.13 A better understanding of the incidence of SSI and its risk factors is helpful in guiding surveillance after hospital discharge and patient selection for perioperative antibiotic prophylaxis. So far, there are relatively few studies on this topic, which are all hampered by inconsistent definitions of SSI, different durations of postoperative surveillance, retrospective designs, and the relatively small sample sizes that did not permit multivariable analysis.5-12 The goals of this study were (1) to determine the incidence of SSI with the use of a prospective standard protocol; (2) to evaluate the isolated pathogens and management of SSI; and (3) to define clinical and surgical risk factors for SSI, which may facilitate patient selection for perioperative antibiotic prophylaxis. To achieve these goals, we prospectively studied a large cohort from a single institution. SURGERY 675
676 Bures et al
PATIENTS AND METHODS Study design. This study accrued 6,784 consecutive patients who underwent thyroid surgery at the Department of Surgery at the KaiserinElisabeth-Hospital, Vienna, Austria, between March 1, 2007, and October 31, 2012. During this period, the hospital participated in the German Nosocomial Infection Surveillance System (KISS) and collected prospective data on all patients undergoing surgery. Institutional review board approval was obtained. The current study includes the patients who underwent thyroid operation for benign or malignant disease. We excluded two patients who had an artificial heart valve and four who were immunosuppressed with cortisone because of chronic obstructive pulmonary disease, leaving 6,778 as principal study cohort. Details of KISS have been outlined previously.14-16 In brief, the method by KISS is almost identical to the surveillance method of the National Nosocomial Infections Surveillance System.17 SSI are defined, classified and recorded according to the CDC by an interdisciplinary team of surgeons and microbiologists.3,4 Patients are under surveillance for 30 days after operation. If a repeat operation is performed at the same site, surveillance of the primary operation is stopped. Patients’ demographic, clinical and pathologic characteristics, operative details, perioperative complications, and follow-up data are prospectively recorded in an computerized documentation system (MediCare ChirDok; Micom, Wiener Neustadt, Austria), as described previously.18,19 In brief, data are entered by the surgeon and subsequently double-checked by independent physicians. For the current study, the ChirDok database was linked to the KISS system, which allowed us to evaluate all possible risk factors for SSI. All patients underwent standard open thyroid operation without antibiotic prophylaxis. Preoperative skin cleansing was performed with chlorhexidine-alcohol. Dexamethasone was not used before the operative incision for prevention of nausea and vomiting. An active 8-F Redon suction drain was inserted and usually removed on the first postoperative day. After discharge, patients were monitored on an outpatient basis for a total of 30 days Patients were contacted weekly via telephone. Site visits were performed on demand. Measures. Collected variables from the databases included age, gender, duration of hospital stay, duration of operation, surgical details, isolate details, type of SSI, and management of SSI. Category of SSI was assigned according to CDC
Surgery April 2014
case definitions: A1 (superficial incisional SSI) was assigned if there was either (1) purulent discharge from the superficial incision; (2) if a microorganism was isolated from an aseptically obtained culture of fluid or tissue from the superficial incision; (3) if local pain or tenderness, localized swelling, redness, or heat was present at the surgical site and the superficial incision was opened by the surgeon and the culture was positive; or (4) if diagnosis of A1 SSI was made by the attending surgeon. A2 (deep incisional SSI) was assigned if there was (1) purulent discharge from the deep incision but not from the organ or cavity; (2) the deep incision spontaneously dehisced or was deliberately opened by a surgeon because the patient was symptomatic; (3) an abscess or other evidence of infection involving the deep layers was found on direct examination, during reoperation, or by histological or radiological examination; or (4) if diagnosis of A2 SSI was made by the attending surgeon. A3 (organ and cavity SSI) was assigned if there was (1) purulent discharge from a drain that is placed into the organ or cavity; (2) a microorganism isolated from an aseptically obtained culture of fluid or tissue from the organ or cavity; (3) an abscess or other evidence of infection involving the organ or cavity was found on direct examination, during reoperation, or by histologic or radiologic examination; or (4) if diagnosis of A3 SSI was made by the attending surgeon. Comorbidity status was graded with the American Society of Anesthesiologists’ (ASA) score. The operative procedures were classified according to the size/weight of the lobe remnant, which was based on surgeons’ estimates and measurements during the operation. We distinguished subtotal (remnant mass 6–8 g), near-total (unilateral capsular remnant of <2 g) and total lobectomies (no remnant). With these definitions, procedures were stratified in five groups19: 1, bilateral total and near-total lobectomy (LE); 2, Dunhill procedure, ie, total LE plus contralateral subtotal LE (SLE); 3, bilateral SLE or SLE plus near-total LE; 4, SLE plus enucleation (ENUC) or bilateral ENUC or unilateral SLE or unilateral ENUC; 5, total LE plus ENUC or near-total LE plus ENUS or unilateral total LE or unilateral near-total LE. Duration of operation was evaluated as continuous variable. A total of 12 attending surgeons were involved in this study, which performed between 52 and 1,354 of the 6,778 operations. Statistical analysis. Categorical variables are presented as numbers and proportions and were
Bures et al 677
Surgery Volume 155, Number 4
Table I. Cohort characteristics and risk factors for SSI Variable Sex, n (%) Female Male Age, y, median (IQR) ASA score, n (%) 1 2–3 Duration of operation, min, median (IQR) Surgery for recurrent disease, n (%) No Yes Surgical procedure, n (%) Total or near-total thyroidectomy Dunhill procedure Bilateral SLE or SLE plus NTL Bilateral SLE, SLE plus ENUC, bilateral ENUC, unilateral SLE, unilateral ENUC LE/NTL with or without ENUC Laterality of procedure, n (%) Unilateral Bilateral Indication for surgery, n (%) Benign Malignant Node dissection, n (%) No Yes
SSI
Total, n = 6,778
Yes, n = 32
No, n = 6,746
OR (95% CI)
5,203 (76.8) 1,575 (23.2) 54 (45–64)
20 (0.4) 12 (0.8) 51 (45–59)
5,183 (99.6) 1,563 (99.2) 54 (45–64)
1.00 1.99 (0.97–4.08) 1.02 (0.99–1.05)
3,242 (47.8) 3,536 (52.2) 93 (73–119)
9 (0.3) 23 (0.7) 125 (65–141)
3,233 (99.7) 3,513 (99.4) 93 (73–119)
1.00 2.35 (1.09–5.09) 1.16 (1.06–1.29)*
P value .061
.170 .032
.004 .416
6,412 (94.6) 366 (5.4)
32 (0.5) 0 (0.0)
6,380 (99.5) 366 (100.0)
4,572 384 193 178
23 2 1 0
4,549 382 192 178
1.00 0 (0 to N) .962
(67.5) (5.7) (2.9) (2.6)
(0.5) (0.5) (0.5) (0.0)
(99.5) (99.5) (99.5) (100.0)
1.00 1.03 (0.24–4.41) 1.03 (0.14–7.67) 0 (0 to N)
1,451 (21.4)
6 (0.4)
1,445 (99.6)
0.82 (0.33–2.02)
1,444 (21.3) 5,334 (78.7)
5 (0.3) 27 (0.5)
1,439 (99.7) 5,307 (99.5)
1.00 1.46 (0.56–3.81)
.522
5,792 (85.5) 986 (14.5)
26 (0.4) 6 (0.6)
5,766 (99.6) 980 (99.4)
1.00 1.36 (0.56–3.31)
.454
5,893 (86.9) 885 (13.1)
25 (0.4) 7 (0.8)
5,868 (99.6) 878 (99.2)
1.00 1.87 (0.81–4.34)
.180
*per 30 minutes. P values were generated with Fisher exact tests. CI, Confidence interval; ENUC, enucleation; IQR, interquartile range; LE, lobectomy; NTL, near-total lobectomy; OR, odds ratio; SLE, subtotal lobectomy; SSI, surgical-site infection.
compared by the use of the Fisher exact test. Continuous data (age, duration of operation, duration of hospital stay) were tested for normal distribution using Kolmogorov-Smirnov tests and were found to be not-normally distributed (each P < .05). Thus, continuous variables are presented as median and interquartile range and are compared using the Mann-Whitney U-tests and Kruskal-Wallis tests, as appropriate. The cumulative incidence of SSI during the surveillance period was estimated with the Kaplan-Meier method. Univariable and multivariable logistic regression models were fitted to identify variables that are associated with SSI. To exclude variables with limited prognostic ability from the multivariable model, a stepwise backward variable selection with the likelihood ratio criterion was used, resulting in a model that contained only clinically relevant variables. The rank of elimination
was assigned when a variable was removed from the model. A nomogram predicting the probability of SSI was constructed upon the final multivariable model. The nomogram was validated internally by bootstrapping (200 re-samples). The bootstrap-corrected predictive accuracy of the nomogram was evaluated by the area under the receiver-operating characteristics curve. Calibration was assessed graphically. All statistical testing was two-sided. Data analyses were performed with STATA 12.0 (StataCorp, College Station, TX). RESULTS Incidence of SSI. Patient characteristics are shown in Table I. In all, SSI occurred in 32 cases. The cumulative incidence of SSI after 5, 10, 20, and 30 days was 0.18%, 0.36%, 0.46%, and 0.49%, respectively (Fig 1). The median time
678 Bures et al
Surgery April 2014
Table II. Microbiology of SSI Isolate Staphylococcus aureus Gemella morbillorum Gemella morbillorum, Capnocytophaga Peptostreptococcus anaerobius Prevotella bivia Streptococcus pneumonia Streptococcus viridans, Lactobacillus, Bifidobacterium, Propionibacterium Streptococcus group C (beta-hemolytic) No pathogenic isolate No smear Fig 1. Cumulative incidence of SSI. The cumulative incidence of SSI after 5, 10, 20, and 30 days was 0.18%, 0.36%, 0.46%, and 0.49%, respectively. The median time from surgery to SSI was 7 days.
from surgery to SSI was 7 days (IQR 4–10.5 days). Of the 32 SSI, 5 (15.6%) occurred during the hospital stay for surgery and 27 occurred after hospital discharge (84.4%). Of the latter group, 10 patients (37.0%) required re-admission to the hospital for a median of 6 days (IQR 3–11 days). Classification, management, and isolates. SSI was classified as superficial incisional in 30 (93.8%), deep incisional in 1 (3.1%), and organ/ space in 1 (3.1%). Patients received systemic antibiotics with an aminopenicillin (n = 27) or clindamycin (n = 5). Three patients (9.3%) required cervical re-exploration in general anesthesia. The microbiological spectrum is shown in Table II. Staphylococcus aureus was the most common isolate, accounting for more than 50% of cases. The median time to SSI was similar in patients with versus without S. aureus infection (P = .489) and in those with anaerobic versus aerobic bacteria (P = .214). Risk factors and the development of a predictive nomogram. Table I shows risk factors associated with SSI in univariable analysis. Increasing duration of operation and an ASA score $2 were associated with an increased risk of SSI (each P < .05). In addition, men had a 2-fold greater incidence of SSI, although this difference did not reach statistical significance (P = .061). Among the different surgeons, the SSI rate ranged between 0 and 1.3%, but there was no clinically relevant difference (P = .243). To identify independent risk factors for SSI, multivariable logistic regression analysis was performed. Using the backward selection approach, the final multivariable model contained the variables duration of operation (odds ratio 1.17, P = .004) and ASA score (odds ratio 2.34, P = .031, Table III). From these variables, a nomogram
n
%*
14 1 1 1 1 1 1
51.8 3.7 3.7 3.7 3.7 3.7 3.7
1 6 5
3.7 22.2 —
*Calculated from 27 SSI cases with smear. SSI, Surgical-site infection.
Table III. Multivariable backward logistic regression analysis predicting occurrence of SSI Variable Sex Age ASA score Duration of operation Operation for recurrence Operative procedure Bilateral vs unilateral Node dissection
Rank OR 6 5 — — 1 2 3 4
1.62 1.01 2.34 1.17 0.00 1.14 1.14 1.35
95% CI
P value
0.77–3.39 0.98–1.04 1.08–5.07 1.05–1.30 0.00 to N 0.74–1.77 0.43–3.04 0.52–3.50
.205 .537 .031 .004 .994 .554 .789 .537
ASA score and the duration of operation were identified as independent predictors of SSI. The OR for the duration of operation is given per 30 minutes. The rank of elimination was given for insignificant variables and indicates the step of removal. ASA, American Society of Anesthesiologists; CI, confidence interval; OR, odds ratio; SSI, surgical-site infection.
predicting the risk of SSI was constructed (Fig 2). The nomogram was internally validated by bootstrapping and reached a predictive accuracy of 70.1%. The calibration curve showed a good agreement between predicted probability and actual observation (Fig 3). DISCUSSION This study shows that the cumulative incidence of SSI is <0.5%. ASA score and the duration of operation are independent risk factors for SSI. The low incidence of SSI does not justify routine antibiotic prophylaxis, but it may be considered for selected patients based on the individual risk profile. In our study, the incidence of SSI was 0.49% at 30 days. This rate is somewhat lower than the average rate reported in the literature (reviewed in Table IV). This may in part be attributed to the relatively high numbers of thyroid surgeries performed annually, but also to differing definitions
Bures et al 679
Surgery Volume 155, Number 4
Fig 2. Nomogram predicting the probability of SSI within 30 days. Instructions: Locate the expected duration of operation and the patient’s ASA score on the respective axis. Draw a straight line up to the Points axis to determine how many points are assigned for each single variable. Sum the points and locate this number on the Total Points axis. Draw a straight line down to find the patient’s probability of SSI within 30 days after thyroid surgery.
Table IV. Details of studies during the past 10 years that reported the incidence of SSI in thyroid operation Reference 12
Fig 3. Calibration plot comparing the actual rate of SSI and the probability predicted by the nomogram: Both the apparent and the bias-corrected plot are very close to the dashed line (‘‘ideal’’), indicating well calibration. The ideal line represents perfect calibration, ie, that the actual rate always equals the predicated probability..
of SSI. For this study, we followed the definition and surveillance protocol of the Centers of Disease Control and Prevention and followed our patients for 30 days. This extended surveillance period may also explain why our median duration to SSI was 7 days and thus greater than reported previously.8,9 Thus, surveillance on an outpatient basis is necessary to detect relevant SSI. Although the risk of SSI decreases after 2 weeks, there is still a small chance to develop SSI. Patients should be aware of this and counseled accordingly. Nonetheless, this and previous studies clearly show a low incidence of SSI, which does not justify routine antibiotic prophylaxis. This underscores current recommendations by guidelines. However, implementation of these guidelines appears to be
Zambudio et al Avenia et al22 Alvarado et al5 Serpell and Phan10 Suslu et al11 Bergenfelz et al7 Dionigi et al8 Dionigi et al9 Barbaros et al6 Total
Year
n SSI
n total
% SSI
2004 2009 2009 2007 2006 2008 2006 2008 2008 —
1 3 2 5 2 57 6 3 7 86
301 500 193 336 135 3,660 241 112 239 5,717
0.3 0.6 1.0 1.5 1.5 1.6 2.5 2.6 2.9 1.5
SSI, Surgical-site infection.
problematic among surgeons. A recent study from the Italian Endocrine Surgery Units Association showed that antibiotic prophylaxis was applied in 38.7% of thyroid surgeries and thus overused.20 Development of antibiotic resistance may be a serious complication of this approach. The goal of antibiotic prophylaxis is to prevent SSI. A first/second-generation cephalosporin or aminopenicillin-sulbactam is effective against the most common pathogen (S. aureus) and thus represents a first-choice agent, but local resistance patterns should be considered. A single dose should be administered 30–60 minutes before incision and repeated if the procedure lasts longer than 4 hours.21 Because there are no appropriately powered prospective randomized trials in thyroid surgery, the level of evidence for this recommendation is 4 and the concept of antibiotic prophylaxis is unproven.
680 Bures et al
The only randomized multicenter study on antibiotic prophylaxis in thyroid surgery reported preliminary data from 500 patients.22 Three patients (0.6%) developed SSI, two of whom had received antibiotic prophylaxis (P = .57). The study, however, was underpowered to detect any statistically significant difference. In our study without antibiotic prophylaxis, the incidence of SSI was 0.5%. Hypothesizing that antibiotic prophylaxis lowers this incidence by 50% (which is an unproven concept), more than 10,000 patients need to be randomized in each group to detect a significant difference with a statistical power of 0.80. Risk factors can be helpful in identifying patients with an increased odds of perioperative complications. Careful preoperative evaluation of these risk factors may minimize perioperative SSI by administration of antibiotic prophylaxis. In few studies, however, have authors evaluated these risk factors in thyroid procedures. It has been shown that insertion of a drain does not impact the incidence of SSI23 but is helpful in identifying other complications such as postoperative bleeding.19 Bergenfelz et al7 found that the incidence of SSI increases approximately 5-fold (5% vs 1%) when a lymph node dissection is performed. In contrast, Alvarado et al5 reported an SSI rate of 1.0% in patients undergoing thyroidectomy with lymph node dissection for papillary thyroid carcinoma, which is comparable with the rate seen in surgery for benign goiter. In our study, lymph node dissection was associated with an increased risk of SSI, but this difference did not reach statistical significance in multivariable analysis (P = .5). It is therefore possible that not the operative procedure itself but the duration of operation drives the risk of developing SSI. Similarly, men tended to have an increased risk of SSI (P = .061) in univariable analysis. As men had greater ASA scores than women (ASA 2–3: 60% in men, 50% in women, P < .0001), this trend towards statistical significance was lost (P = .21), when the ASA score was simultaneously analyzed in the multivariable model. We performed multivariable analysis to account for potential intercorrelation of variables. From the significant factors in multivariable analysis, a predictive nomogram was generated. The nomogram represents a simple and individual tool to assess the probability of SSI and thus help identifying patients at an increased risk for SSI. It shows that operation time is by far the most important predictor of SSI, but the ASA score should be considered. The nomogram may therefore guide
Surgery April 2014
the clinician in the use of antibiotic prophylaxis in selected cases or represent a tool to identify a highrisk population that should be included in randomized prospective trials. In patients with a nomogram-derived SSI probability of >1%, prophylaxis may be indicated. The nomogram, however, should be validated on external datasets before its implementation in clinical routine. Several additional limitations merit to be mentioned. Specifically, we did not record individual diseases, such as obesity, smoking history, or presence of diabetes but summarized these comorbidities with the ASA score. Future studies should focus on these individual comorbidities. The authors hypothesize that inclusion of these parameters (instead of the overall ASA score) will refine the nomogram and improve its accuracy. Furthermore, the impact of the patient’s sex on SSI warrants further investigation. The reported pathogens may represent the local situation at our center, but others reported similar data.8,9 The nomogram contained the variables ASA score and duration of operation. We recognize that the duration of operation may be difficult to estimate preoperatively and is related to intraoperative findings, which represents an issue, because the antibiotic prophylaxis should be administered 30–60 minutes before the operation. Finally, these data are derived from a tertiary care center with more than 1,000 open thyroid surgeries performed annually and thus may not be applicable to centers with a lower case volume or different surgical approaches. CONCLUSIONS The cumulative incidence of SSI in thyroid surgery is <0.5%. ASA score and the duration of operation are independent risk factors for SSI. Routine antibiotic prophylaxis is not recommended but may be considered for selected patients on the basis of our predictive model. External validation of the nomogram is advocated. REFERENCES 1. Burke JP. Infection control---a problem for patient safety. N Engl J Med 2003;348:651-6. 2. Perencevich EN, Sands KE, Cosgrove SE, Guadagnoli E, Meara E, Platt R. Health and economic impact of surgical site infections diagnosed after hospital discharge. Emerg Infect Dis 2003;9:196-203. 3. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC definitions for nosocomial infections, 1988. Am J Infect Control 1988;16:128-40. 4. Horan TC, Gaynes RP, Martone WJ, Jarvis WR, Emori TG. CDC definitions of nosocomial surgical site infections, 1992: a modification of CDC definitions of surgical wound infections. Infect Control Hosp Epidemiol 1992;13:606-8.
Surgery Volume 155, Number 4
5. Alvarado R, Sywak MS, Delbridge L, Sidhu SB. Central lymph node dissection as a secondary procedure for papillary thyroid cancer: is there added morbidity? Surgery 2009;145:514-8. 6. Barbaros U, Erbil Y, Aksakal N, Citlak G, Issever H, Bozbora A, et al. Electrocautery for cutaneous flap creation during thyroidectomy: a randomised, controlled study. J Laryngol Otol 2008;122:1343-8. 7. Bergenfelz A, Jansson S, Kristoffersson A, Martensson H, Reihner E, Wallin G, et al. Complications to thyroid surgery: results as reported in a database from a multicenter audit comprising 3,660 patients. Langenbecks Arch Surg 2008; 393:667-73. 8. Dionigi G, Rovera F, Boni L, Castano P, Dionigi R. Surgical site infections after thyroidectomy. Surg Infect (Larchmt) 2006;7(Suppl 2):S117-20. 9. Dionigi G, Rovera F, Boni L, Dionigi R. Surveillance of surgical site infections after thyroidectomy in a one-day surgery setting. Int J Surg 2008;6(Suppl 1):S13-5. 10. Serpell JW, Phan D. Safety of total thyroidectomy. ANZ J Surg 2007;77:15-9. 11. Suslu N, Vural S, Oncel M, Demirca B, Gezen FC, Tuzun B, et al. Is the insertion of drains after uncomplicated thyroid surgery always necessary? Surg Today 2006;36:215-8. 12. Zambudio AR, Rodriguez J, Riquelme J, Soria T, Canteras M, Parrilla P. Prospective study of postoperative complications after total thyroidectomy for multinodular goiters by surgeons with experience in endocrine surgery. Ann Surg 2004;240:18-25. 13. Association of Operating Room Nurses. Recommended practices for skin preparation of patients. AORN J 2002;75:184-7. 14. Gastmeier P, Sohr D, Geffers C, Behnke M, Daschner F, Ruden H. Mortality risk factors with nosocomial Staphylococcus aureus infections in intensive care units: results from the German Nosocomial Infection Surveillance System (KISS). Infection 2005;33:50-5.
Bures et al 681
15. Geffers C, Gastmeier P. Nosocomial infections and multidrug-resistant organisms in Germany: epidemiological data from KISS (the Hospital Infection Surveillance System). Dtsch Arztebl Int 2011;108:87-93. 16. Manni€en J, van den Hof S, Brandt C, Behnke M, Wille JC, Gastmeier P. Comparison of the National Surgical Site Infection surveillance data between The Netherlands and Germany: PREZIES versus KISS. J Hosp Infect 2007;66: 224-31. 17. Emori TG, Culver DH, Horan TC, Jarvis WR, White JW, Olson DR, et al. National nosocomial infections surveillance system (NNIS): description of surveillance methods. Am J Infect Control 1991;19:19-35. 18. Ott J, Meusel M, Schultheis A, Promberger R, Pallikunnel SJ, Neuhold N, et al. The incidence of lymphocytic thyroid infiltration and Hashimoto’s thyroiditis increased in patients operated for benign goiter over a 31-year period. Virchows Arch 2011;459:277-81. 19. Promberger R, Ott J, Kober F, Koppitsch C, Seemann R, Freissmuth M, et al. Risk factors for postoperative bleeding after thyroid surgery. Br J Surg 2012;99:373-9. 20. Gentile I, Rosato L, Avenia N, Testini M, D’Ajello M, Antonino A, et al. Do Italian surgeons use antibiotic prophylaxis in thyroid surgery? Results from a national study (UEC - Italian Endocrine Surgery Units Association). Ann Ital Chir doi: S0003469X12020416 (epub ahead of print). 21. Weber WP, Marti WR, Zwahlen M, Misteli H, Rosenthal R, Reck S, et al. The timing of surgical antimicrobial prophylaxis. Ann Surg 2008;247:918-26. 22. Avenia N, Sanguinetti A, Cirocchi R, Docimo G, Ragusa M, Ruggiero R, et al. Antibiotic prophylaxis in thyroid surgery: a preliminary multicentric Italian experience. Ann Surg Innov Res 2009;3:10. 23. Samraj K, Gurusamy KS. Wound drains following thyroid surgery. Cochrane Database Syst Rev 2007:CD006099.
Background. To evaluate the incidence and the microbe spectrum of surgical-site infections (SSIs) in patients undergoing elective thyroid operation and to develop a risk factor based predictive model. Methods. This prospective study included 6,778 consecutive patients who underwent thyroid operation at a single institution between 2007 and 2012. SSI was defined according to the Centers for Disease Control and Prevention. Regression models were fitted to evaluate risk factors for SSI. A predictive nomogram was constructed from relevant variables in the multivariable analysis. Discrimination and calibration of the nomogram were assessed. Results. The cumulative incidence of SSI after 30 days was 0.49%. The median time from operation to SSI was 7 days (interquartile range, 4–10.5 days). SSI was classified as superficial incisional in 30 cases (93.8%), deep incisional in 1 case (3.1%), and organ/space in 1 case (3.1%). Staphylococcus aureus was the most common isolate. In multivariable analysis, duration of operation (P = .004) and American Society of Anesthesiologists’ score (P = .031) were identified as independent risk factors for SSI. These variables formed the basis of a nomogram, which was validated internally by bootstrapping and reached a predictive accuracy of 70.1%. The calibration curve showed a good agreement between predicted probability and actual observation. Conclusion. The cumulative incidence of SSI in thyroid operation is <0.5%. American Society of Anesthesiologists’ score and the duration of operation are independent risk factors for SSI. Antibiotic prophylaxis may be considered for selected patients based on the individual risk profile. (Surgery 2014;155:675-81.) From the Second Department of Surgery,a Krankenanstalt Rudolfstiftung, Department of Surgery,b Kaiserin-Elisabeth-Spital, Department of Urology,c Medical University of Vienna, Department of Pathology and Microbiology with Hospital Hygiene Team,d Kaiserin-Elisabeth-Spital, and Department of Pathology and Microbiology,e Krankenanstalt Rudolfstiftung, Vienna, Austria; and Nationales Referenzzentrum f€ u r Surveillance von nosokomialen Infektionen,f Charite -Universit€ a tsmedizin, Berlin, Germany
SURGICAL-SITE INFECTIONS (SSIs) are on the most common nosocomial infections1 and cause substantial morbidity to patients and costs to the health care system.2 The Centers for Disease Control and Prevention (CDC) have developed standard criteria to define SSI,3,4 which are regarded the contemporary, international standard. A SSI is defined as infection related to an operative procedure, which occurs at or near the surgical incision within a 30-day period. Thyroid operation is considered a clean procedure in a well-vascularized area. The incidence of SSI is between 0.3 and 2.9%.5-12 Routine Accepted for publication December 6, 2013. Reprint requests: Tobias Klatte, MD, Second Department of Surgery, Krankenanstalt Rudolfstiftung, Juchgasse 25, 1030 Vienna, Austria. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2014 Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2013.12.002
perioperative antibiotic prophylaxis is not recommended by guidelines but should be considered for selected patients.13 A better understanding of the incidence of SSI and its risk factors is helpful in guiding surveillance after hospital discharge and patient selection for perioperative antibiotic prophylaxis. So far, there are relatively few studies on this topic, which are all hampered by inconsistent definitions of SSI, different durations of postoperative surveillance, retrospective designs, and the relatively small sample sizes that did not permit multivariable analysis.5-12 The goals of this study were (1) to determine the incidence of SSI with the use of a prospective standard protocol; (2) to evaluate the isolated pathogens and management of SSI; and (3) to define clinical and surgical risk factors for SSI, which may facilitate patient selection for perioperative antibiotic prophylaxis. To achieve these goals, we prospectively studied a large cohort from a single institution. SURGERY 675
676 Bures et al
PATIENTS AND METHODS Study design. This study accrued 6,784 consecutive patients who underwent thyroid surgery at the Department of Surgery at the KaiserinElisabeth-Hospital, Vienna, Austria, between March 1, 2007, and October 31, 2012. During this period, the hospital participated in the German Nosocomial Infection Surveillance System (KISS) and collected prospective data on all patients undergoing surgery. Institutional review board approval was obtained. The current study includes the patients who underwent thyroid operation for benign or malignant disease. We excluded two patients who had an artificial heart valve and four who were immunosuppressed with cortisone because of chronic obstructive pulmonary disease, leaving 6,778 as principal study cohort. Details of KISS have been outlined previously.14-16 In brief, the method by KISS is almost identical to the surveillance method of the National Nosocomial Infections Surveillance System.17 SSI are defined, classified and recorded according to the CDC by an interdisciplinary team of surgeons and microbiologists.3,4 Patients are under surveillance for 30 days after operation. If a repeat operation is performed at the same site, surveillance of the primary operation is stopped. Patients’ demographic, clinical and pathologic characteristics, operative details, perioperative complications, and follow-up data are prospectively recorded in an computerized documentation system (MediCare ChirDok; Micom, Wiener Neustadt, Austria), as described previously.18,19 In brief, data are entered by the surgeon and subsequently double-checked by independent physicians. For the current study, the ChirDok database was linked to the KISS system, which allowed us to evaluate all possible risk factors for SSI. All patients underwent standard open thyroid operation without antibiotic prophylaxis. Preoperative skin cleansing was performed with chlorhexidine-alcohol. Dexamethasone was not used before the operative incision for prevention of nausea and vomiting. An active 8-F Redon suction drain was inserted and usually removed on the first postoperative day. After discharge, patients were monitored on an outpatient basis for a total of 30 days Patients were contacted weekly via telephone. Site visits were performed on demand. Measures. Collected variables from the databases included age, gender, duration of hospital stay, duration of operation, surgical details, isolate details, type of SSI, and management of SSI. Category of SSI was assigned according to CDC
Surgery April 2014
case definitions: A1 (superficial incisional SSI) was assigned if there was either (1) purulent discharge from the superficial incision; (2) if a microorganism was isolated from an aseptically obtained culture of fluid or tissue from the superficial incision; (3) if local pain or tenderness, localized swelling, redness, or heat was present at the surgical site and the superficial incision was opened by the surgeon and the culture was positive; or (4) if diagnosis of A1 SSI was made by the attending surgeon. A2 (deep incisional SSI) was assigned if there was (1) purulent discharge from the deep incision but not from the organ or cavity; (2) the deep incision spontaneously dehisced or was deliberately opened by a surgeon because the patient was symptomatic; (3) an abscess or other evidence of infection involving the deep layers was found on direct examination, during reoperation, or by histological or radiological examination; or (4) if diagnosis of A2 SSI was made by the attending surgeon. A3 (organ and cavity SSI) was assigned if there was (1) purulent discharge from a drain that is placed into the organ or cavity; (2) a microorganism isolated from an aseptically obtained culture of fluid or tissue from the organ or cavity; (3) an abscess or other evidence of infection involving the organ or cavity was found on direct examination, during reoperation, or by histologic or radiologic examination; or (4) if diagnosis of A3 SSI was made by the attending surgeon. Comorbidity status was graded with the American Society of Anesthesiologists’ (ASA) score. The operative procedures were classified according to the size/weight of the lobe remnant, which was based on surgeons’ estimates and measurements during the operation. We distinguished subtotal (remnant mass 6–8 g), near-total (unilateral capsular remnant of <2 g) and total lobectomies (no remnant). With these definitions, procedures were stratified in five groups19: 1, bilateral total and near-total lobectomy (LE); 2, Dunhill procedure, ie, total LE plus contralateral subtotal LE (SLE); 3, bilateral SLE or SLE plus near-total LE; 4, SLE plus enucleation (ENUC) or bilateral ENUC or unilateral SLE or unilateral ENUC; 5, total LE plus ENUC or near-total LE plus ENUS or unilateral total LE or unilateral near-total LE. Duration of operation was evaluated as continuous variable. A total of 12 attending surgeons were involved in this study, which performed between 52 and 1,354 of the 6,778 operations. Statistical analysis. Categorical variables are presented as numbers and proportions and were
Bures et al 677
Surgery Volume 155, Number 4
Table I. Cohort characteristics and risk factors for SSI Variable Sex, n (%) Female Male Age, y, median (IQR) ASA score, n (%) 1 2–3 Duration of operation, min, median (IQR) Surgery for recurrent disease, n (%) No Yes Surgical procedure, n (%) Total or near-total thyroidectomy Dunhill procedure Bilateral SLE or SLE plus NTL Bilateral SLE, SLE plus ENUC, bilateral ENUC, unilateral SLE, unilateral ENUC LE/NTL with or without ENUC Laterality of procedure, n (%) Unilateral Bilateral Indication for surgery, n (%) Benign Malignant Node dissection, n (%) No Yes
SSI
Total, n = 6,778
Yes, n = 32
No, n = 6,746
OR (95% CI)
5,203 (76.8) 1,575 (23.2) 54 (45–64)
20 (0.4) 12 (0.8) 51 (45–59)
5,183 (99.6) 1,563 (99.2) 54 (45–64)
1.00 1.99 (0.97–4.08) 1.02 (0.99–1.05)
3,242 (47.8) 3,536 (52.2) 93 (73–119)
9 (0.3) 23 (0.7) 125 (65–141)
3,233 (99.7) 3,513 (99.4) 93 (73–119)
1.00 2.35 (1.09–5.09) 1.16 (1.06–1.29)*
P value .061
.170 .032
.004 .416
6,412 (94.6) 366 (5.4)
32 (0.5) 0 (0.0)
6,380 (99.5) 366 (100.0)
4,572 384 193 178
23 2 1 0
4,549 382 192 178
1.00 0 (0 to N) .962
(67.5) (5.7) (2.9) (2.6)
(0.5) (0.5) (0.5) (0.0)
(99.5) (99.5) (99.5) (100.0)
1.00 1.03 (0.24–4.41) 1.03 (0.14–7.67) 0 (0 to N)
1,451 (21.4)
6 (0.4)
1,445 (99.6)
0.82 (0.33–2.02)
1,444 (21.3) 5,334 (78.7)
5 (0.3) 27 (0.5)
1,439 (99.7) 5,307 (99.5)
1.00 1.46 (0.56–3.81)
.522
5,792 (85.5) 986 (14.5)
26 (0.4) 6 (0.6)
5,766 (99.6) 980 (99.4)
1.00 1.36 (0.56–3.31)
.454
5,893 (86.9) 885 (13.1)
25 (0.4) 7 (0.8)
5,868 (99.6) 878 (99.2)
1.00 1.87 (0.81–4.34)
.180
*per 30 minutes. P values were generated with Fisher exact tests. CI, Confidence interval; ENUC, enucleation; IQR, interquartile range; LE, lobectomy; NTL, near-total lobectomy; OR, odds ratio; SLE, subtotal lobectomy; SSI, surgical-site infection.
compared by the use of the Fisher exact test. Continuous data (age, duration of operation, duration of hospital stay) were tested for normal distribution using Kolmogorov-Smirnov tests and were found to be not-normally distributed (each P < .05). Thus, continuous variables are presented as median and interquartile range and are compared using the Mann-Whitney U-tests and Kruskal-Wallis tests, as appropriate. The cumulative incidence of SSI during the surveillance period was estimated with the Kaplan-Meier method. Univariable and multivariable logistic regression models were fitted to identify variables that are associated with SSI. To exclude variables with limited prognostic ability from the multivariable model, a stepwise backward variable selection with the likelihood ratio criterion was used, resulting in a model that contained only clinically relevant variables. The rank of elimination
was assigned when a variable was removed from the model. A nomogram predicting the probability of SSI was constructed upon the final multivariable model. The nomogram was validated internally by bootstrapping (200 re-samples). The bootstrap-corrected predictive accuracy of the nomogram was evaluated by the area under the receiver-operating characteristics curve. Calibration was assessed graphically. All statistical testing was two-sided. Data analyses were performed with STATA 12.0 (StataCorp, College Station, TX). RESULTS Incidence of SSI. Patient characteristics are shown in Table I. In all, SSI occurred in 32 cases. The cumulative incidence of SSI after 5, 10, 20, and 30 days was 0.18%, 0.36%, 0.46%, and 0.49%, respectively (Fig 1). The median time
678 Bures et al
Surgery April 2014
Table II. Microbiology of SSI Isolate Staphylococcus aureus Gemella morbillorum Gemella morbillorum, Capnocytophaga Peptostreptococcus anaerobius Prevotella bivia Streptococcus pneumonia Streptococcus viridans, Lactobacillus, Bifidobacterium, Propionibacterium Streptococcus group C (beta-hemolytic) No pathogenic isolate No smear Fig 1. Cumulative incidence of SSI. The cumulative incidence of SSI after 5, 10, 20, and 30 days was 0.18%, 0.36%, 0.46%, and 0.49%, respectively. The median time from surgery to SSI was 7 days.
from surgery to SSI was 7 days (IQR 4–10.5 days). Of the 32 SSI, 5 (15.6%) occurred during the hospital stay for surgery and 27 occurred after hospital discharge (84.4%). Of the latter group, 10 patients (37.0%) required re-admission to the hospital for a median of 6 days (IQR 3–11 days). Classification, management, and isolates. SSI was classified as superficial incisional in 30 (93.8%), deep incisional in 1 (3.1%), and organ/ space in 1 (3.1%). Patients received systemic antibiotics with an aminopenicillin (n = 27) or clindamycin (n = 5). Three patients (9.3%) required cervical re-exploration in general anesthesia. The microbiological spectrum is shown in Table II. Staphylococcus aureus was the most common isolate, accounting for more than 50% of cases. The median time to SSI was similar in patients with versus without S. aureus infection (P = .489) and in those with anaerobic versus aerobic bacteria (P = .214). Risk factors and the development of a predictive nomogram. Table I shows risk factors associated with SSI in univariable analysis. Increasing duration of operation and an ASA score $2 were associated with an increased risk of SSI (each P < .05). In addition, men had a 2-fold greater incidence of SSI, although this difference did not reach statistical significance (P = .061). Among the different surgeons, the SSI rate ranged between 0 and 1.3%, but there was no clinically relevant difference (P = .243). To identify independent risk factors for SSI, multivariable logistic regression analysis was performed. Using the backward selection approach, the final multivariable model contained the variables duration of operation (odds ratio 1.17, P = .004) and ASA score (odds ratio 2.34, P = .031, Table III). From these variables, a nomogram
n
%*
14 1 1 1 1 1 1
51.8 3.7 3.7 3.7 3.7 3.7 3.7
1 6 5
3.7 22.2 —
*Calculated from 27 SSI cases with smear. SSI, Surgical-site infection.
Table III. Multivariable backward logistic regression analysis predicting occurrence of SSI Variable Sex Age ASA score Duration of operation Operation for recurrence Operative procedure Bilateral vs unilateral Node dissection
Rank OR 6 5 — — 1 2 3 4
1.62 1.01 2.34 1.17 0.00 1.14 1.14 1.35
95% CI
P value
0.77–3.39 0.98–1.04 1.08–5.07 1.05–1.30 0.00 to N 0.74–1.77 0.43–3.04 0.52–3.50
.205 .537 .031 .004 .994 .554 .789 .537
ASA score and the duration of operation were identified as independent predictors of SSI. The OR for the duration of operation is given per 30 minutes. The rank of elimination was given for insignificant variables and indicates the step of removal. ASA, American Society of Anesthesiologists; CI, confidence interval; OR, odds ratio; SSI, surgical-site infection.
predicting the risk of SSI was constructed (Fig 2). The nomogram was internally validated by bootstrapping and reached a predictive accuracy of 70.1%. The calibration curve showed a good agreement between predicted probability and actual observation (Fig 3). DISCUSSION This study shows that the cumulative incidence of SSI is <0.5%. ASA score and the duration of operation are independent risk factors for SSI. The low incidence of SSI does not justify routine antibiotic prophylaxis, but it may be considered for selected patients based on the individual risk profile. In our study, the incidence of SSI was 0.49% at 30 days. This rate is somewhat lower than the average rate reported in the literature (reviewed in Table IV). This may in part be attributed to the relatively high numbers of thyroid surgeries performed annually, but also to differing definitions
Bures et al 679
Surgery Volume 155, Number 4
Fig 2. Nomogram predicting the probability of SSI within 30 days. Instructions: Locate the expected duration of operation and the patient’s ASA score on the respective axis. Draw a straight line up to the Points axis to determine how many points are assigned for each single variable. Sum the points and locate this number on the Total Points axis. Draw a straight line down to find the patient’s probability of SSI within 30 days after thyroid surgery.
Table IV. Details of studies during the past 10 years that reported the incidence of SSI in thyroid operation Reference 12
Fig 3. Calibration plot comparing the actual rate of SSI and the probability predicted by the nomogram: Both the apparent and the bias-corrected plot are very close to the dashed line (‘‘ideal’’), indicating well calibration. The ideal line represents perfect calibration, ie, that the actual rate always equals the predicated probability..
of SSI. For this study, we followed the definition and surveillance protocol of the Centers of Disease Control and Prevention and followed our patients for 30 days. This extended surveillance period may also explain why our median duration to SSI was 7 days and thus greater than reported previously.8,9 Thus, surveillance on an outpatient basis is necessary to detect relevant SSI. Although the risk of SSI decreases after 2 weeks, there is still a small chance to develop SSI. Patients should be aware of this and counseled accordingly. Nonetheless, this and previous studies clearly show a low incidence of SSI, which does not justify routine antibiotic prophylaxis. This underscores current recommendations by guidelines. However, implementation of these guidelines appears to be
Zambudio et al Avenia et al22 Alvarado et al5 Serpell and Phan10 Suslu et al11 Bergenfelz et al7 Dionigi et al8 Dionigi et al9 Barbaros et al6 Total
Year
n SSI
n total
% SSI
2004 2009 2009 2007 2006 2008 2006 2008 2008 —
1 3 2 5 2 57 6 3 7 86
301 500 193 336 135 3,660 241 112 239 5,717
0.3 0.6 1.0 1.5 1.5 1.6 2.5 2.6 2.9 1.5
SSI, Surgical-site infection.
problematic among surgeons. A recent study from the Italian Endocrine Surgery Units Association showed that antibiotic prophylaxis was applied in 38.7% of thyroid surgeries and thus overused.20 Development of antibiotic resistance may be a serious complication of this approach. The goal of antibiotic prophylaxis is to prevent SSI. A first/second-generation cephalosporin or aminopenicillin-sulbactam is effective against the most common pathogen (S. aureus) and thus represents a first-choice agent, but local resistance patterns should be considered. A single dose should be administered 30–60 minutes before incision and repeated if the procedure lasts longer than 4 hours.21 Because there are no appropriately powered prospective randomized trials in thyroid surgery, the level of evidence for this recommendation is 4 and the concept of antibiotic prophylaxis is unproven.
680 Bures et al
The only randomized multicenter study on antibiotic prophylaxis in thyroid surgery reported preliminary data from 500 patients.22 Three patients (0.6%) developed SSI, two of whom had received antibiotic prophylaxis (P = .57). The study, however, was underpowered to detect any statistically significant difference. In our study without antibiotic prophylaxis, the incidence of SSI was 0.5%. Hypothesizing that antibiotic prophylaxis lowers this incidence by 50% (which is an unproven concept), more than 10,000 patients need to be randomized in each group to detect a significant difference with a statistical power of 0.80. Risk factors can be helpful in identifying patients with an increased odds of perioperative complications. Careful preoperative evaluation of these risk factors may minimize perioperative SSI by administration of antibiotic prophylaxis. In few studies, however, have authors evaluated these risk factors in thyroid procedures. It has been shown that insertion of a drain does not impact the incidence of SSI23 but is helpful in identifying other complications such as postoperative bleeding.19 Bergenfelz et al7 found that the incidence of SSI increases approximately 5-fold (5% vs 1%) when a lymph node dissection is performed. In contrast, Alvarado et al5 reported an SSI rate of 1.0% in patients undergoing thyroidectomy with lymph node dissection for papillary thyroid carcinoma, which is comparable with the rate seen in surgery for benign goiter. In our study, lymph node dissection was associated with an increased risk of SSI, but this difference did not reach statistical significance in multivariable analysis (P = .5). It is therefore possible that not the operative procedure itself but the duration of operation drives the risk of developing SSI. Similarly, men tended to have an increased risk of SSI (P = .061) in univariable analysis. As men had greater ASA scores than women (ASA 2–3: 60% in men, 50% in women, P < .0001), this trend towards statistical significance was lost (P = .21), when the ASA score was simultaneously analyzed in the multivariable model. We performed multivariable analysis to account for potential intercorrelation of variables. From the significant factors in multivariable analysis, a predictive nomogram was generated. The nomogram represents a simple and individual tool to assess the probability of SSI and thus help identifying patients at an increased risk for SSI. It shows that operation time is by far the most important predictor of SSI, but the ASA score should be considered. The nomogram may therefore guide
Surgery April 2014
the clinician in the use of antibiotic prophylaxis in selected cases or represent a tool to identify a highrisk population that should be included in randomized prospective trials. In patients with a nomogram-derived SSI probability of >1%, prophylaxis may be indicated. The nomogram, however, should be validated on external datasets before its implementation in clinical routine. Several additional limitations merit to be mentioned. Specifically, we did not record individual diseases, such as obesity, smoking history, or presence of diabetes but summarized these comorbidities with the ASA score. Future studies should focus on these individual comorbidities. The authors hypothesize that inclusion of these parameters (instead of the overall ASA score) will refine the nomogram and improve its accuracy. Furthermore, the impact of the patient’s sex on SSI warrants further investigation. The reported pathogens may represent the local situation at our center, but others reported similar data.8,9 The nomogram contained the variables ASA score and duration of operation. We recognize that the duration of operation may be difficult to estimate preoperatively and is related to intraoperative findings, which represents an issue, because the antibiotic prophylaxis should be administered 30–60 minutes before the operation. Finally, these data are derived from a tertiary care center with more than 1,000 open thyroid surgeries performed annually and thus may not be applicable to centers with a lower case volume or different surgical approaches. CONCLUSIONS The cumulative incidence of SSI in thyroid surgery is <0.5%. ASA score and the duration of operation are independent risk factors for SSI. Routine antibiotic prophylaxis is not recommended but may be considered for selected patients on the basis of our predictive model. External validation of the nomogram is advocated. REFERENCES 1. Burke JP. Infection control---a problem for patient safety. N Engl J Med 2003;348:651-6. 2. Perencevich EN, Sands KE, Cosgrove SE, Guadagnoli E, Meara E, Platt R. Health and economic impact of surgical site infections diagnosed after hospital discharge. Emerg Infect Dis 2003;9:196-203. 3. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC definitions for nosocomial infections, 1988. Am J Infect Control 1988;16:128-40. 4. Horan TC, Gaynes RP, Martone WJ, Jarvis WR, Emori TG. CDC definitions of nosocomial surgical site infections, 1992: a modification of CDC definitions of surgical wound infections. Infect Control Hosp Epidemiol 1992;13:606-8.
Surgery Volume 155, Number 4
5. Alvarado R, Sywak MS, Delbridge L, Sidhu SB. Central lymph node dissection as a secondary procedure for papillary thyroid cancer: is there added morbidity? Surgery 2009;145:514-8. 6. Barbaros U, Erbil Y, Aksakal N, Citlak G, Issever H, Bozbora A, et al. Electrocautery for cutaneous flap creation during thyroidectomy: a randomised, controlled study. J Laryngol Otol 2008;122:1343-8. 7. Bergenfelz A, Jansson S, Kristoffersson A, Martensson H, Reihner E, Wallin G, et al. Complications to thyroid surgery: results as reported in a database from a multicenter audit comprising 3,660 patients. Langenbecks Arch Surg 2008; 393:667-73. 8. Dionigi G, Rovera F, Boni L, Castano P, Dionigi R. Surgical site infections after thyroidectomy. Surg Infect (Larchmt) 2006;7(Suppl 2):S117-20. 9. Dionigi G, Rovera F, Boni L, Dionigi R. Surveillance of surgical site infections after thyroidectomy in a one-day surgery setting. Int J Surg 2008;6(Suppl 1):S13-5. 10. Serpell JW, Phan D. Safety of total thyroidectomy. ANZ J Surg 2007;77:15-9. 11. Suslu N, Vural S, Oncel M, Demirca B, Gezen FC, Tuzun B, et al. Is the insertion of drains after uncomplicated thyroid surgery always necessary? Surg Today 2006;36:215-8. 12. Zambudio AR, Rodriguez J, Riquelme J, Soria T, Canteras M, Parrilla P. Prospective study of postoperative complications after total thyroidectomy for multinodular goiters by surgeons with experience in endocrine surgery. Ann Surg 2004;240:18-25. 13. Association of Operating Room Nurses. Recommended practices for skin preparation of patients. AORN J 2002;75:184-7. 14. Gastmeier P, Sohr D, Geffers C, Behnke M, Daschner F, Ruden H. Mortality risk factors with nosocomial Staphylococcus aureus infections in intensive care units: results from the German Nosocomial Infection Surveillance System (KISS). Infection 2005;33:50-5.
Bures et al 681
15. Geffers C, Gastmeier P. Nosocomial infections and multidrug-resistant organisms in Germany: epidemiological data from KISS (the Hospital Infection Surveillance System). Dtsch Arztebl Int 2011;108:87-93. 16. Manni€en J, van den Hof S, Brandt C, Behnke M, Wille JC, Gastmeier P. Comparison of the National Surgical Site Infection surveillance data between The Netherlands and Germany: PREZIES versus KISS. J Hosp Infect 2007;66: 224-31. 17. Emori TG, Culver DH, Horan TC, Jarvis WR, White JW, Olson DR, et al. National nosocomial infections surveillance system (NNIS): description of surveillance methods. Am J Infect Control 1991;19:19-35. 18. Ott J, Meusel M, Schultheis A, Promberger R, Pallikunnel SJ, Neuhold N, et al. The incidence of lymphocytic thyroid infiltration and Hashimoto’s thyroiditis increased in patients operated for benign goiter over a 31-year period. Virchows Arch 2011;459:277-81. 19. Promberger R, Ott J, Kober F, Koppitsch C, Seemann R, Freissmuth M, et al. Risk factors for postoperative bleeding after thyroid surgery. Br J Surg 2012;99:373-9. 20. Gentile I, Rosato L, Avenia N, Testini M, D’Ajello M, Antonino A, et al. Do Italian surgeons use antibiotic prophylaxis in thyroid surgery? Results from a national study (UEC - Italian Endocrine Surgery Units Association). Ann Ital Chir doi: S0003469X12020416 (epub ahead of print). 21. Weber WP, Marti WR, Zwahlen M, Misteli H, Rosenthal R, Reck S, et al. The timing of surgical antimicrobial prophylaxis. Ann Surg 2008;247:918-26. 22. Avenia N, Sanguinetti A, Cirocchi R, Docimo G, Ragusa M, Ruggiero R, et al. Antibiotic prophylaxis in thyroid surgery: a preliminary multicentric Italian experience. Ann Surg Innov Res 2009;3:10. 23. Samraj K, Gurusamy KS. Wound drains following thyroid surgery. Cochrane Database Syst Rev 2007:CD006099.