Objective
We sought to determine risk factors associated with surgical site infection (SSI) among a cohort of human immunodeficiency virus (HIV)-infected women undergoing hysterectomy during the era of highly active antiretroviral therapy.
Study Design
This is a retrospective study of HIV-infected women who underwent a hysterectomy for benign indications at a tertiary care center. Electronic medical records were reviewed from January 1999 through December 2012. SSI was defined using Centers for Disease Control and Prevention criteria.
Results
There were 77 HIV-infected women who underwent a hysterectomy: 47 (61%) were abdominal; 16 (21%) were laparoscopic or robot-assisted; and 14 (18%) were vaginal. Acquired immune deficiency syndrome was diagnosed in 58% of patients, and 75% of patients self-reported use of highly active antiretroviral therapy at the time of surgery. There were 17 (22%) SSIs; 5 (29%) superficial incisional wound infections, 3 (18%) vaginal cuff cellulitis, and 9 (53%) pelvic abscesses were diagnosed. After multivariable logistic regression, preoperative albumin level (adjusted odds ratio [aOR], 0.14; 95% confidence interval [CI], 0.02–0.86) and minimally invasive hysterectomy (aOR, 0.16; 95% CI, 0.03–0.84) were associated with decreased SSI. Preoperative absolute CD4 count was not associated with SSI (aOR, 0.99; 95% CI, 0.99–1).
Conclusion
Low preoperative serum albumin levels and abdominal hysterectomy are associated with increased risk of SSIs in HIV-infected women.
According to 2011 Centers for Disease Control and Prevention (CDC) surveillance data, women represent 21% of all diagnoses of human immunodeficiency virus (HIV) infection among US adults and account for 25% of the acquired immune deficiency syndrome (AIDS) diagnoses. Expected longevity for HIV-infected patients is estimated at 73 years with advancements in HIV treatment and clinical care. Gynecologists may soon be presented with increasing numbers of women living with HIV/AIDS who require a hysterectomy for benign indications such as dysfunctional uterine bleeding, pelvic pain, symptomatic fibroids, or persistent cervical dysplasia.
Hysterectomy is the leading major nonobstetric operation among women in the United States with approximately 575,000 procedures performed annually and rates increase as a woman ages. While advances have been made in infection control practices, including appropriate selection and dosing of perioperative antibiotics, surgical site infection (SSI) remain the most common surgical complication. It is estimated up to 5% of patients undergoing operative procedures will develop an SSI leading to a longer hospital stay and increased cost.
Hysterectomy operations are classified as clean-contaminated according to the Surgical Wound Classification System as they are operative wounds in which the genital tract is entered under controlled conditions and without unusual contamination. Polymicrobial flora of aerobes and anaerobes contaminate the surgical site after entry into the vagina during a hysterectomy. Besides modern infection control procedures, prevention of SSI depends on immune function and wound healing. Prevention of SSI has been targeted by the Department of Health and Human Services as a national priority especially given an estimated 4.3% (95% confidence interval [CI], 3.6–5.1) of patients will develop an SSI after hysterectomy.
Previous investigations of postoperative infectious morbidity among HIV-infected patients include a heterogeneous group of surgeries in both genders ; few studies report specifically on the gynecology population. Clinically, there are ongoing concerns about potential increased risk of SSI in this population due to underlying immune system defects; however, there are few data on modifiable risk factors among cohorts of HIV-infected women that could potentially lead to decreased SSIs. For these reasons, we sought to determine risk factors associated with SSI at our institution during the era of highly active antiretroviral therapy (HAART).
Materials and Methods
This is a retrospective cohort study of HIV-infected women who underwent a hysterectomy performed by benign gynecology surgeons from Jan. 1, 1999, through Dec. 31, 2012, at Johns Hopkins Medical Institution. Patients were identified through the billing and coding departments. A majority of the hysterectomies were performed by 3 attending surgeons who specialized in HIV care with the assistance of a resident physician. Hysterectomies performed by gynecology-oncology surgeons were excluded, even if they were performed for benign indications. Hysterectomy types included total abdominal, supracervical abdominal, laparoscopic-assisted vaginal, total laparoscopic, robotic-assisted total, and vaginal. We combined laparoscopic-assisted vaginal, total laparoscopic, robotic-assisted total, and vaginal hysterectomies into 1 variable named “minimally invasive hysterectomy.” Per protocol, antimicrobial prophylaxis was administered for all procedures, which consisted of preoperative single-dose intravenous administration of cefazolin (2 g); clindamycin (600 mg) alone or with gentamicin was administered in patients who reported a penicillin allergy. The institutional review board of Johns Hopkins University School of Medicine determined that the project met criteria for exempt review.
Data abstraction
Electronic medical records, including inpatient and outpatient charts, were reviewed and demographic and clinical data were collected. Abstracted variables included age, gravidity, body mass index (BMI), race, medical comorbidities, prior laparotomies, smoking history, number of years living with HIV, AIDS diagnosis, and antiretroviral therapy. Surgical data included indication for hysterectomy, type of hysterectomy, estimated uterine size by pelvic examination (as uterine weights were not routinely recorded on the pathology report before 2010), estimated blood loss, perioperative blood transfusion, complications, length of hospital stay, readmission (including intensive care unit [ICU]), and any reoperations. Preoperative laboratory blood tests (within 60 days of the procedure) included complete blood cell count and serum albumin concentration. Absolute and percent CD4 count and plasma viral load data within 6 months of surgery were collected. Postoperative laboratory blood tests (day 1) included complete blood cell count.
SSI definition
SSI occurring within 30 days of the operation was categorized into superficial incisional, deep incisional, or organ space. CDC criteria were used to define SSI : (1) purulent drainage from the surgical incision; (2) organisms isolated from an aseptically obtained culture of fluid or tissue from the surgical incision; (3) superficial incision that is deliberately opened by a surgeon and is culture-positive or not cultured and patient has at least one of the following signs or symptoms: pain or tenderness, localized swelling, redness, or heat; (4) deep incision that spontaneously dehisces or is deliberately opened by a surgeon and is culture-positive or not cultured and patient has at least one of the following signs or symptoms: fever (>38°C), localized pain, or tenderness; (5) an abscess or other evidence of infection involving the deep incision that is found on direct examination, during invasive procedure, or by histopathologic examination or imaging test; and (6) diagnosis of an SSI by the surgeon or attending. A diagnosis of skin cellulitis was not included.
Statistical analysis
Continuous variables were summarized by mean ± SD or median, interquartile range. The χ 2 or Fisher exact test was used for dichotomous variables. Kruskal-Wallis test was used for nominal variables and if there was a significant difference within the group ( P < .05), then Dunn test was performed for pairwise comparisons. Student t test with unequal variance was used for evaluation of continuous variables that followed a normal distribution. Nonnormal continuous variables were analyzed using Wilcoxon rank sum test. All reported P values are 2-tailed and P values < .05 were considered significant. A priori, age, and preoperative CD4 count (absolute) were included in a multivariable logic regression model. Variables that were significant after univariate analysis were added to the model. The results of the logistic regression are expressed as odds ratios (ORs) with corresponding 95% CI. All analyses were performed using statistical software (STATA 12; StataCorp, College Station, TX).
Results
There were 77 HIV-infected women, aged 27-58 years, who underwent nonemergent hysterectomies over the 13-year time frame. Eighty-three percent of women were black, 58.4% had a diagnosis of AIDS, 75% self-reported use of HAART, and 48.1% reported heterosexual sex as the main mode of HIV transmission. Forty percent of patients had a BMI ≥30. Other patient characteristics are listed in Table 1 .
| Characteristic | No SSI | SSI | P value |
|---|---|---|---|
| Age, y (mean, SD) | 43 ± 5.8 | 41 ± 6.3 | .09 |
| Race | .06 | ||
| Black | 47 (78.3) | 17 (100) | |
| White | 13 (21.7) | 0 (0) | |
| BMI (median, IQR) | 26.9 (24.1–31.4) | 27.3 (20.5–37.3) | .86 |
| No. of pregnancies (median, IQR) | 3 (2–4) | 2 (1–4) | .42 |
| No. of laparotomies (median, IQR) | 0 (0–1) | 1 (0–1) | .04 |
| HIV transmission risk factor | .03 | ||
| Heterosexual sex | 32 (53.3) | 5 (29.4) | |
| Injection drug use | 13 (21.7) | 6 (35.3) | |
| Blood transfusion | 1 (1.7) | 1 (5.9) | |
| ≥2 risk factors | 14 (23.3) | 3 (17.7) | |
| Unknown | 0 (0) | 2 (11.8) | |
| HAART | 46 (76.7) | 12 (70.6) | .75 |
| AIDS diagnosis | 37 (61.7) | 8 (47.1) | .28 |
| Tobacco use (current or past year) | 41 (68.3) | 14 (82.4) | .37 |
| Comorbid conditions | |||
| Hypertension | 18 (30) | 6 (35.3) | .68 |
| Diabetes | 3 (5) | 3 (17.7) | .12 |
| MRSA | 12 (20) | l2 (11.8) | .72 |
There were 17 (22%) patients diagnosed with an SSI, of which 8 (47%) occurred during the immediate postoperative period and 9 (53%) occurred after hospital discharge. There were 5 (29%) superficial incisional wound infections, 3 (18%) vaginal cuff cellulitis cases, and 9 (53%) pelvic abscesses. Nineteen (25%) patients had at least 1 temperature ≥38°C during the immediate postoperative period, and atelectasis was recorded as the etiology in 9 (47%) of these patients.
Specific details regarding the 17 SSIs in the cohort are presented in Table 2 . Among women who developed an SSI, 82% occurred after abdominal hysterectomy, 12% occurred after laparoscopic/robotic-assisted hysterectomy, and 6% occurred after vaginal hysterectomy ( P = .12). Women who underwent a minimally invasive hysterectomy were 74% less likely to develop an SSI compared to those who underwent an abdominal hysterectomy ( P = .05). The number of prior laparotomies was not associated with a diagnosis of an SSI ( P = .07).
| Patient | Age, y | Hysterectomy | BMI | HIV, y | AIDS | HAART | Preoperative | Infection | Reoperation | Readmitted | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Indication | Type | Uterine size, wk | Hb, g/dL | Albumin, g/dL | CD4 (cells/μL) | HIV RNA | |||||||||
| 1 | 45 | Dysplasia | Abd | 6 | 20 | 18 | No | Yes | 14.9 | 4.2 | 280 | UND | Pelvic abscess | No | Yes |
| 2 | 48 | Heavy bleeding | LAVH | 12 | 20 | 14 | Yes | Yes | 7.6 | 3.2 | 46 | 187,809 | Superficial wound, pelvic abscess | Yes | Yes |
| 3 | 43 | Pelvic pain | Abd | 6 | 37 | 4 | No | No | 11.4 | 4 | 834 | 4988 | Vaginal cuff cellulitis | No | No |
| 4 | 27 | Heavy bleeding | Abd | 10 | 20 | 3 | Yes | No | 12.4 | 3.4 | 173 | 59,982 | Superficial wound, UTI | No | No |
| 5 | 41 | Dysplasia | Abd | 7 | 32 | 16 | Yes | Yes | 12.8 | 4.1 | 597 | UND | Superficial wound | No | No |
| 6 | 36 | Dysplasia | Abd | 5 | 27 | 19 | Yes | Yes | 12.1 | 4.2 | 337 | UND | Superficial wound, pelvic abscess | Yes | No |
| 7 | 36 | Heavy bleeding | LAVH | 10 | a | 4 | No | Yes | 11.3 | a | 664 | UND | Vaginal cuff cellulitis | No | No |
| 8 | 38 | Dysplasia | Abd | 12 | 24 | 13 | Yes | Yes | 11.7 | 3.8 | 85 | 2483 | Pelvic abscess, UTI | No | Yes |
| 9 | 49 | Dysplasia | Abd | 14 | 41 | 0 | No | Yes | 14.8 | 3.8 | 412 | UND | Superficial wound | No | No |
| 10 | 43 | Dysplasia | Abd | 12 | 21 | 4 | Yes | Yes | 11 | 4.1 | 356 | UND | Vaginal cuff cellulitis | No | No |
| 11 | 47 | Heavy bleeding | Abd | 10 | 27 | a | Yes | Yes | 12.1 | 3.7 | 86 | 1170 | Superficial wound, pelvic abscess | No | Yes |
| 12 | 47 | Heavy bleeding | Abd | 24 | 27 | 4 | Yes | Yes | 12.5 | 4.0 | 184 | UND | Superficial wound | No | No |
| 13 | 44 | Heavy bleeding | Abd | 6 | a | 7 | No | Yes | 13.8 | 4.4 | 947 | 126 | Superficial wound, pelvic abscess | Yes | Yes |
| 14 | 34 | Heavy bleeding | Abd | 6 | 45 | 4 | No | No | 13.3 | 4 | 591 | 6700 | Superficial wound, pelvic abscess | No | Yes |
| 15 | 36 | Dysplasia | TVH | 8 | 23 | 4 | No | Yes | 15.3 | 4.4 | 589 | 2281 | Pelvic abscess | Yes | Yes |
| 16 | 41 | Heavy bleeding | Abd | a | 37 | 3 | No | No | 10.8 | 3.1 | 646 | 22,000 | Superficial wound | No | Yes |
| 17 | 31 | Heavy bleeding | Abd | 10 | a | 1 | No | No | 11.1 | 4 | 352 | a | Pelvic abscess | No | No |
For patients diagnosed with an SSI, preoperative serum albumin concentration median value was 4.0 g/dL (3.75-4.15), whereas for patients without an SSI, albumin concentration median value was 4.2 g/dL (3.95-4.4). In unadjusted analysis, for every 1-g/dL increase in serum albumin concentration, the odds of an SSI decreased by 84% (OR 0.16; 95% CI, 0.03–0.76). In univariate analyses, SSI was not associated with preoperative absolute CD4 count (OR, 0.99; 95% CI, 0.99–1), percent CD4 count (OR, 0.99; 95% CI, 0.95–1.04), detectable plasma viral load (OR, 2.1; 95% CI, 0.7–6.2), or preoperative hemoglobin concentration (OR, 1.13; 95% CI, 0.78–1.64) ( Table 3 ). Estimated blood loss (>250 mL) was similar between patients with and without a diagnosis of SSI ( P = .88). After adjusting for age and preoperative absolute CD4 count, hysterectomy type ( P = .03) and preoperative serum albumin concentration remained significant ( P = .03).
