Objectives
To identify risk factors for and outcomes of surgical site infections and cellulitis after abdominal hysterectomies.
Study Design
We used logistic regression analysis to analyze data from a case-control study of 1104 patients undergoing abdominal hysterectomies at a university hospital between Jan. 1, 2007 and Dec. 30, 2010.
Results
Factors significantly associated with surgical site infections and with cellulitis were: pulmonary disease, operations done in Main Operating Room East, and seroma. Body mass index >35, no private insurance, and fluid and electrolyte disorders were risk factors for surgical site infections. The mean prophylactic dose of cefazolin was significantly higher for controls than for patients with surgical site infections. Preoperative showers with Hibiclens (Molnlycke Health Care US, LLC, Norcross, GA) and cefazolin prophylaxis were associated with a significantly decreased cellulitis risk. Surgical site infections and cellulitis were significantly associated with readmissions and return visits and surgical site infections were associated with reoperations.
Conclusion
Preoperative showers, antimicrobial prophylaxis, surgical techniques preventing seromas, and the operating room environment may affect the risk of surgical site infections and cellulitis after abdominal hysterectomies.
Surgical site infections (SSIs) are the second most common healthcare-associated infection (HAI) reported to the Centers for Disease Control and Prevention’s National Health Safety Network (NHSN) and these infections adversely affect patient outcomes. Patient-related risk factors for SSIs after hysterectomies have included: diabetes, obesity, weight, a thick subcutaneous tissue layer, neoplasm, low preoperative serum albumin, and proliferative endometrium. Age <40 years and advanced age have been associated with increased risk of SSIs. Procedure-related risk factors have included: abdominal procedures, total hysterectomies compared with subtotal hysterectomies, procedure duration, preoperative hospital stay greater than 2 days, hematoma, nongynecological procedures done with the hysterectomy, excessive intraoperative bleeding, transfusions, and low postoperative hemoglobin. Perioperative antimicrobial prophylaxis has been associated with decreased risk of SSIs.
The mean rate of SSIs after hysterectomies done at the University of Iowa Hospitals and Clinics (UIHC) was above the NHSN 75th percentile. A preliminary assessment suggested that obesity, diabetes, and asthma were common among patients with SSIs. In addition, rates of cellulitis after these procedures seemed to be high. Thus, we conducted a case-control study to identify potentially modifiable risk factors for SSIs and for cellulitis. We also sought to identify outcomes of these infections. During the study, the surgeons became concerned that SSI rates were higher among patients whose operations were done in Main Operating Room (MOR) East than among patients whose operations were done in MOR. To address this concern, we extended the study period 18 months and we collected data on the rooms in which procedures were done.
Materials and Methods
We defined an SSI case as any patient who had an abdominal hysterectomy ( International Classification of Diseases , Ninth Revision Clinical Modification procedure codes: 68.31, 68.39, 68.41, 68.49, 68.61, 68.69) at the UIHC between Jan. 1, 2007 and Dec. 31, 2010, and who was identified by the UIHC’s Program of Hospital Epidemiology as meeting the NHSN definition of SSI, including superficial incisional, deep incisional, and organ/space infections. We defined a cellulitis case as any patient who had an abdominal hysterectomy at the UIHC between Jan. 1, 2007 and June 30, 2009, (the original study period) and who had erythema around the incision that the clinician described as cellulitis and treated with antimicrobial agents. We used a computerized screening algorithm to identify patients who were likely to meet the SSI definition. An experienced infection preventionist reviewed the medical records of these patients and determined whether they met criteria for SSI or cellulitis. We randomly selected controls (average 3 controls per SSI case and 4 controls per cellulitis case) from among patients who underwent abdominal hysterectomies during the study period but who did not acquire a SSI or cellulitis. We excluded patients whose procedures were started by general surgeons and involved the bowel and patients who were discharged on the day of their procedures.
We collected demographic data and data on patients’ underlying diseases, medications, operations, hospital stays, and postoperative visits from medical records. To determine whether specific operating rooms were associated with increased risk of SSIs, we extended the study through Dec. 31, 2010. For procedures done during the extension, we evaluated risk factors for SSIs but not for cellulitis. On the basis of the initial data analyses, we streamlined data collection and did not collect data on some possible risk factors (intrafascial/extrafascial procedure, suture used to reapproximate the subcutaneous tissue, and Hibiclens [Molnlycke Health Care US, LLC, Norcross, GA] used for preoperative shower) and on outcomes.
We entered data into Excel (Microsoft, Redmond, CA) and we validated data entry. We calculated frequencies, percentages, means, medians, and 95% confidence intervals (CIs) to evaluate the distributions of potential risk factors for cases and controls. We used the χ 2 test and the Fisher exact test to compare the results of categorical variables for cases and controls. To compare the results of continuous variables, we used the Student’s t test for normally distributed variables or the median analysis for nonnormally distributed variables. We used logistic regression to assess the association of either SSI or cellulitis with variables that had P < .1 in the bivariable analysis. We used a stepwise method ( P entry = .15, P stay = .15) to select variables. P values < .05 were considered significant. A c statistic close to 1 indicated the model had good predictive efficacy. A Hosmer-Lemeshow Goodness-of-fit test P value > .05 indicated that the model fit the data well. We conducted all analyses using SAS software program, version 9.2 (SAS Institute, Inc., Cary, NC).
The University of Iowa’s Institutional Review Board approved the study.
Results
During the initial study period (Jan. 1, 2007–June 30, 2009), 590 women had abdominal hysterectomies, of whom 66 acquired SSIs and 60 acquired cellulitis. During the additional 18 months (July 1, 2009–Dec. 31, 2010), 514 patients had abdominal hysterectomies, of whom 50 acquired SSIs. Thus, 116 patients acquired SSIs during the entire study period. The overall SSI rate was 10.2% during the initial study period and 10.5% for the entire period. Three hundred ninety controls were randomly selected from patients who underwent abdominal hysterectomy between Jan. 1, 2007 and Dec. 31, 2010, and who did not have SSIs or cellulitis. Of the 390 control patients, 240 had operations during the initial period; these patients served as controls for patients with cellulitis and for the patients with SSIs who underwent hysterectomies during the initial study period. The remaining 150 patients served as controls for patients with SSIs whose operations occurred between July 1, 2009 and Dec. 31, 2010.
The most common signs of cellulitis and SSI were localized swelling, erythema, or warmth (93.2% cellulitis; 66.4% SSIs). Patients with SSIs also had pain or tenderness (62.9%), and temperature >38°C (60.3%). Surgeons opened 61.2% of the infected surgical incisions. The median time from the operation to identification of SSIs was 12 days (range, 4–29 days); 84.4% of SSIs were identified after hospital discharge. The median time from the operation to identification of cellulitis was 11 days (range, 1–26 days); 71.7% of cellulitis cases were diagnosed after discharge. Culture results were available for 46 (39.7%) patients with SSIs. The most common organisms causing SSI included Staphylococcus aureus alone (n = 9) or in combination with other organisms (n = 6), Enterococcus species alone (n = 2) or in combination with other organisms (n = 9), and Escherichia coli alone (n = 3) or in combination with other organisms (n = 6).
Patients with SSIs and those with cellulitis were fairly similar ( Table 1 ). However, patients with cellulitis were significantly older than patients with SSIs and they were less likely to have used Hibiclens for preoperative showers. In contrast, patients with SSIs were more likely to be on oral hypoglycemics and to have longer procedures.
| Variable | Control (n = 240) | SSI (n = 66) | Cellulitis (n = 60) | Cellulitis vs control | SSI vs cellulitis | |
|---|---|---|---|---|---|---|
| Preoperative factor | ||||||
| Age, y | (Mean ± SD) | 54.9 ± 14.1 | 53.2 ± 11.9 | 57.6 ± 12.9 | P = .19 | P = .05 |
| (Median) | 54.0 | 53.0 | 60.0 | P = .06 | P = .18 | |
| Weight, kg | (Mean ± SD) | 86.9 ± 24.8 | 101.7 ± 33.5 | 100.3 ± 28.3 | P = .0003 | P = .80 |
| (Median) | 82.8 | 97.2 | 98.1 | P = .02 | P = .10 | |
| BMI | (Mean ± SD) | 32.8 ± 9.0 | 38.2 ± 11.5 | 38.5 ± 10.7 | P < .0001 | P = .89 |
| (Median) | 31.3 | 36.0 | 36.4 | P = .004 | P = .72 | |
| BMI <18.5 | 3 (1.3) | 0 (0) | 0 (0) | P = .03 | P = .44 | |
| ≤18.5 BMI <25 | 47 (19.6) | 10 (15.2) | 5 (8.3) | |||
| ≤25 BMI <30 | 55 (22.9) | 7 (10.6) | 9 (15.0) | |||
| BMI ≥30 | 135 (56.3) | 49 (74.2) | 46 (76.7) | |||
| No private insurance | 57 (23.8) | 27 (40.9) | 22 (36.7) | P = .05 OR, 1.9 (95% CI, 1.0–3.4) | P = .72 OR, 1.2 (95% CI, 0.6–2.5) | |
| Comorbidities | Peripheral vascular disease | 1 (0.4) | 0 (0) | 3 (5.0) | P = .03 OR, 12.6 (95% CI, 1.3–123) | P = .11 |
| Chronic pulmonary disease | 20 (8.3) | 17 (25.8) | 11 (18.3) | P = .03 OR, 2.5 (95% CI, 1.1–5.5) | P = .39 OR, 1.5 (95% CI, 0.7–3.6) | |
| Hypertension | 96 (40.0) | 27 (40.9) | 35 (58.3) | P = .01 OR, 2.1 (95% CI, 1.2–3.7) | P = .07 OR, 0.5 (95% CI, 0.2–1.0) | |
| Obesity | 66 (27.5) | 28 (42.4) | 28 (46.7) | P = .005 OR, 2.3 (95% CI, 1.3–4.1) | P = .72 OR, 0.8 (95% CI, 0.4–1.7) | |
| Preoperative oral hypoglycemic agent | 33 (13.8) | 14 (21.2) | 5 (8.3) | P = .38 OR, 0.6 (95% CI, 0.2–1.5) | P = .05 OR, 3.0 (95% CI, 1.0–8.8) | |
| Operative or immediate postoperative factor | ||||||
| Reason for hysterectomy: cancer | 150 (62.5) | 51 (77.3) | 45 (75.0) | P = .07 OR, 1.8 (95% CI, 0.9–3.4) | P = .84 OR, 1.1 (95% CI, 0.5–2.6) | |
| MOR East | 25 (10.4) | 19 (28.8) | 12 (20.0) | P = .05 OR, 2.2 (95% CI, 1.0–4.6) | P = .30 OR, 1.6 (95% CI, 0.7–3.7) | |
| Preoperative shower with Hibiclens | 40 (16.7) | 17 (27.0) | 4 (6.7) | P = 0.06 OR, 0.4 (95% CI, 0.1–1.0) | P = .004 OR, 5.2 (95% CI, 1.6–16.5) | |
| Procedure | Total | 99 (41.3) | 20 (30.3) | 24 (40.0) | P = .61 | P = .24 |
| Subtotal | 12 (5.0) | 5 (7.6) | 1 (1.7) | |||
| Radical | 129 (53.7) | 41 (62.1) | 35 (58.3) | |||
| Intraoperative blood loss, mL | (Median) | 250.0 | 300.0 | 250.0 | P = .83 | P = .26 |
| Cefazolin as prophylactic antimicrobial agent | 88.6% (n = 236) | 80.3% | 78.3% | P = .05 OR, 0.5 (95% CI, 0.2–1.0) | P = .83 OR, 1.1 (95% CI, 0.5–2.7) | |
| Dose/weight, mg/kg (Mean ± SD) | 23.9 ± 7.4 (n = 209) | 21.8 ± 7.2 (n = 53) | 20.8 ± 6.0 (n = 47) | P = .008 | P = .48 | |
| Operation duration, min | (Mean ± SD) | 212.8 ± 85.2 | 238.8 ± 74.6 | 209.7 ± 76.7 | P = .80 | P = .03 |
| (Median) | 200.0 | 249.0 | 198.5 | P = .77 | P = .08 | |
| Intraoperative temperature, °C | (Mean ± SD) | 35.7 ± 0.5 (n = 236) | 35.7 ± 0.7 (n = 64) | 35.7 ± 0.5 (n = 58) | P = .27 | P = .43 |
| NHSN risk index ≥2 | 66 (27.5) | 31 (47.0) | 24 (40.0) | P = .08 OR, 1.8 (95% CI, 1.0–3.2) | P = .47 OR, 1.3 (95% CI, 0.7–2.7) | |
| Postoperative hematoma | 2 (0.8) | 2 (3.0) | 4 (6.7) | P = .02 OR, 8.5 (95% CI, 1.5–47.6) | P = .42 OR, 0.4 (95% CI, 0.1–2.5) | |
| Postoperative seroma | 13 (5.4) | 18 (27.3) | 14 (23.3) | P = .0001 OR, 5.3 (95% CI, 2.3–12.1) | P = .68 OR, 1.2 (95% CI, 0.5–2.8) | |
Risk factors for SSIs
Patients with SSIs and controls were similar with respect to age, cigarette smoking, preoperative medications (eg, insulin, steroids, chemotherapy, antimicrobial agents), Charlson comorbidity score, scheduling (scheduled, urgent, or emergent), cancer as the reason for the procedure, total, subtotal, or radical hysterectomy, proliferative endometrium, subcutaneous drains, and perioperative blood transfusion. Patients with SSIs and controls differed with respect to a number of factors ( Table 2 ). Multivariable analyses of perioperative factors ( Table 3 ; preoperative model, Table 4 ; overall model) determined that the following factors were significantly associated with SSI: BMI >35, no private insurance, chronic pulmonary disease, fluid and electrolyte disorders, an operation done in MOR East, and postoperative seroma. SSIs were not clustered in time. One surgeon had an increased risk of SSI (odds ratio [OR], 1.8; 95% confidence interval [CI], 1.1–2.9) and cellulitis (OR, 2.6; 95% CI, 1.4–4.7) in the bivariable analysis, but the surgeon variable did not remain in the multivariable model.
| Variable | Control (n = 390) | SSI (n = 116) | P value | OR (95% CI) | |
|---|---|---|---|---|---|
| Preoperative factor | |||||
| Age, y | (Mean ± SD) | 55.0 ± 13.6 | 52.8 ± 12.4 | .12 | n/a |
| (Median) | 54.0 | 51.0 | .11 | n/a | |
| Weight, kg | (Mean ± SD) | 87.2 ± 25.3 | 99.1 ± 29.0 | < .0001 | n/a |
| (Median) | 83.5 | 97.3 | .0001 | n/a | |
| BMI, kg/m 2 | (Mean ± SD) | 32.8 ± 9.1 | 37.5 ± 10.3 | < .0001 | n/a |
| (Median) | 31.5 | 36.8 | < .0001 | n/a | |
| BMI <18.5 | 4 (1.0) | 0 (0) | .015 | n/a | |
| ≤18.5 BMI <25 | 81 (20.8) | 14 (12.1) | |||
| ≤25 BMI <30 | 83 (21.3) | 17 (14.7) | |||
| BMI ≥30 | 222 (56.9) | 85 (73.3) | |||
| No private insurance | 102 (26.2) | 50 (43.1) | .0008 | 2.1 (1.4–3.3) | |
| Preoperative medications | Oral hypoglycemic agent | 40 (10.3) | 20 (17.2) | .05 | 1.8 (1.0–3.3) |
| Anxiolytics | 67 (17.2) | 29 (25.0) | .08 | 1.6 (1.0–2.6) | |
| Comorbidities | Chronic pulmonary disease | 32 (8.2) | 21 (18.1) | .005 | 2.5 (1.4–4.5) |
| Mild liver disease | 3 (0.8) | 4 (3.5) | .05 | 4.6 (1.0–20.9) | |
| Diabetes with or without chronic complications | 49 (12.6) | 26 (22.4) | .01 | 2.0 (1.2–3.4) | |
| Metastatic cancer | 35 (9.0) | 4 (3.5) | .05 | 0.4 (0.1–1.0) | |
| Obesity | 96 (24.6) | 47 (40.5) | .001 | 2.1 (1.3–3.2) | |
| Fluid and electrolyte disorders | 93 (23.9) | 38 (32.8) | .07 | 1.6 (1.0–2.4) | |
| Operative or immediate postoperative factor | |||||
| Proliferative endometrium | 73 (18.7) | 31 (26.7) | .07 | 1.6 (1.0, 2.6) | |
| ASA score ≥3 | 126 (32.3) | 55 (47.4) | .004 | 1.9 (1.2–2.9) | |
| Antimicrobial prophylaxis | Cefazolin dose/weight, mg/kg (Mean ± SD) | 24.1 ± 7.3 (n = 328) | 21.8 ± 6.6 (n = 94) | .005 | n/a |
| Clindamycin vs others | 41 (10.6) (n = 386) | 19 (16.4) (n = 116) | .10 | 1.7 (0.9–3.0) | |
| MOR East | 97 (24.9) | 53 (45.7) | .00003 | 2.5 (1.7–3.9) | |
| Procedure: open vs laparoscopic | 345 (88.5) | 112 (96.6) | .007 | 3.7 (1.3–10.4) | |
| Procedure | Total | 166 (42.5) | 43 (37.1) | .44 | n/a |
| Subtotal | 26 (6.7) | 6 (5.2) | |||
| Radical | 198 (50.8) | 67 (57.7) | |||
| ICD-9 procedure code 68.31 | Laparoscopic supracervical hysterectomy | 15 (3.9) | 0 (0) | .03 | n/a |
| Operation duration, min | (Mean ± SD) | 220.0 ± 89.1 | 236.9 ± 78.0 | .07 | n/a |
| (Median) | 203.0 | 233.0 | .02 | n/a | |
| Intraoperative temperature, °C | (Mean ± SD) | 35.7 ± 0.6 (n = 386) | 35.7 ± 0.7 (n = 114) | .90 | n/a |
| NHSN risk index ≥2 | 109 (28.0) | 51 (44.0) | .001 | 2.0 (1.3–3.1) | |
| No. of postoperative antimicrobial doses | 0 | 343 (88.2) | 106 (91.4) | .02 | n/a |
| 1 | 8 (2.1) | 6 (5.2) | |||
| ≥2 | 38 (9.8) | 4 (3.5) | |||
| Postoperative seroma | 22 (5.6) | 35 (30.2) | < .0001 | 7.2 (4.0–13.0) | |
| Variable | P value | Odds ratio (95% CI) | Estimates | Risk score assigned b |
|---|---|---|---|---|
| Diabetes with or without chronic complications | .13 | 1.6 (0.9–2.8) | 0.45 | 5 |
| Mild liver disease | .08 | 4.0 (0.8–19.1) | 1.39 | 14 |
| Metastatic cancer | .05 | 0.3 (0.1–1.0) | –1.13 | Not assigned |
| Fluid and electrolyte disorders | .01 | 1.9 (1.2–3.1) | 0.64 | 6 |
| Chronic pulmonary disease | .005 | 2.5 (1.3–4.6) | 0.90 | 9 |
| No private insurance | .004 | 2.0 (1.2–3.1) | 0.67 | 7 |
| Body mass index >35 | .0005 | 2.3 (1.4–3.6) | 0.81 | 8 |
a c statistic = 0.69; Hosmer-Lemeshow Goodness-of-fit test P = .865
b We multiplied regression coefficient estimates from the preoperative model by 10 and rounded to establish the points assigned for each risk factor.
| Variable | P value | Odds ratio (95% CI) |
|---|---|---|
| Mild liver disease | .09 | 4.5 (0.8–25.2) |
| Preoperative use of anxiolytics | .07 | 1.7 (1.0–3.1) |
| Body mass index >35 | .02 | 1.7 (1.1–2.8) |
| Fluid and electrolytes disorders | .02 | 1.9 (1.1–3.2) |
| Chronic pulmonary disease | .02 | 2.3 (1.2–4.4) |
| No private insurance | .0005 | 2.4 (1.5–3.9) |
| MOR East | < .0001 | 2.9 (1.7–4.7) |
| Postoperative seroma | < .0001 | 6.7 (3.5–12.8) |
a c statistic = 0.765; Hosmer-Lemeshow Goodness-of-fit test P = .52.
All patients received antimicrobial prophylaxis, most commonly with a single dose of cephazolin immediately before surgery. Patients with beta-lactam allergies received alternate agents, usually gentamicin and clindamycin. Additional antimicrobial doses were given if procedures lasted ≥3 hours or estimated blood loss was ≥1500 mL. Among patients who received cefazolin for perioperative prophylaxis, controls received a significantly higher dose (24.1 ± 7.3 mg/kg) than did cases with SSIs (21.8 ± 6.6 mg/kg; P = .005). Cefazolin dose was not included in the multivariable modeling because 16.6% of the patients did not receive cefazolin.
Risk factors for cellulitis (Jan. 1, 2007–June 30, 2009)
Patients with cellulitis and control patients were similar with respect to: age, use of oral hypoglycemics, intraoperative blood loss, and operative duration ( Table 1 ). Perioperative factors significantly associated with cellulitis ( Table 5 ; preoperative model, Table 6 ; overall model), were cancer as the reason for hysterectomy, chronic pulmonary disease, procedure done in MOR East, and postoperative seroma. Preoperative showers with Hibiclens and prophylaxis with cefazolin were associated with a decreased risk of cellulitis.
| Variable | P value | Odds ratio (95% CI) |
|---|---|---|
| No private insurance | .12 | 1.7 (0.9–3.1) |
| Hypertension | .09 | 1.7 (0.9–3.2) |
| Body mass index >35 | .09 | 1.7 (0.9–3.2) |
| Peripheral vascular disease | .03 | 13.5 (1.3–143.6) |
| Chronic pulmonary disease | .02 | 2.7 (1.2–6.2) |
a c statistic = 0.675; Hosmer-Lemeshow Goodness-of-fit test P = .704.
| Variable | P value | Odds ratio (95% CI) |
|---|---|---|
| No private insurance | .09 | 1.8 (0.9–3.6) |
| Hypertension | .09 | 1.8 (0.9–3.5) |
| Peripheral vascular disease | .07 | 9.7 (0.8–112.6) |
| Cancer reason for hysterectomy | .03 | 2.3 (1.1–5.0) |
| Chronic pulmonary disease | .02 | 2.9 (1.2–7.3) |
| Preoperative shower with Hibiclens | .01 | 0.2 (0.06–0.7) |
| Cefazolin as prophylactic antimicrobial agent | .05 | 0.4 (0.2–1.0) |
| MOR East | .03 | 2.7 (1.1–6.5) |
| Postoperative seroma | < .0001 | 7.5 (2.9–19.6) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
