Background
Literature on the use of bowel preparation in gynecologic surgery is scarce and limited to minimally invasive gynecologic surgery. The decision on the use of bowel preparation before benign or malignant hysterectomies is mostly driven by extrapolating data from the colorectal literature.
Objective
Bowel preparation is a controversial element within enhanced recovery protocols, and literature investigating its efficacy in gynecologic surgery is scarce. Our aim was to determine if mechanical bowel preparation alone, oral antibiotics alone, or a combination are associated with decreased rates of surgical site infections or anastomotic leaks compared to no bowel preparation following benign or malignant hysterectomy.
Study Design
We identified women who underwent hysterectomy between January 2006 and July 2017 using OptumLabs, a large US commercial health plan database. Inverse propensity score weighting was used separately for benign and malignant groups to balance baseline characteristics. Primary outcomes of 30-day surgical site infection, anastomotic leaks, and major morbidity were assessed using multivariate logistic regression that adjusted for race, census region, household income, diabetes, and other unbalanced variables following propensity score weighting.
Results
A total of 224,687 hysterectomies (benign, 186,148; malignant, 38,539) were identified. Median age was 45 years for the benign and 54 years for the malignant cohort. Surgical approach was as follows: benign: laparoscopic/robotic, 27.2%; laparotomy, 32.6%; vaginal, 40.2%; malignant: laparoscopic/robotic, 28.8%; laparotomy, 47.7%; vaginal, 23.5%. Bowel resection was performed in 0.4% of the benign and 2.8% of the malignant cohort. Type of bowel preparation was as follows: benign: none, 93.8%; mechanical bowel preparation only, 4.6%; oral antibiotics only, 1.1%; mechanical bowel preparation with oral antibiotics, 0.5%; malignant: none, 87.2%; mechanical bowel preparation only, 9.6%; oral antibiotics only, 1.8%; mechanical bowel preparation with oral antibiotics, 1.4%. Use of bowel preparation did not decrease rates of surgical site infections, anastomotic leaks, or major morbidity following benign or malignant hysterectomy. Among malignant abdominal hysterectomies, there was no difference in the rates of infectious morbidity between mechanical bowel preparation alone, oral antibiotics alone, or mechanical bowel preparation with oral antibiotics, compared to no preparation.
Conclusion
Bowel preparation does not protect against surgical site infections or major morbidity following benign or malignant hysterectomy, regardless of surgical approach, and may be safely omitted.
The practice of preoperative evacuation of the bowel with the use of mechanical bowel preparation (MBP) has been a longstanding surgical dogma. First described by Halsted in the 19th century, it was originally based on perceived but unproven advantages related mostly to a reduction in intraluminal stool burden, risk of spillage, and bacterial load, theoretically leading to improved ease of bowel manipulation and a decreased risk of postoperative infectious morbidity and anastomotic leaks (AL). In contrast, the use of MBP is known to result in prolonged fasting, leading to preoperative dehydration and electrolyte abnormalities in addition to significant patient dissatisfaction. , Owing to these untoward effects coupled with efforts to decrease surgical site infections (SSI), use of bowel preparation has come under scrutiny. Randomized controlled trials (RCTs) investigated the use of MBP following colorectal surgery (CRS) as compared to no MBP and found that their use is not associated with decreased overall morbidity, SSI, AL rates, or need for reoperation following CRS. These data resulted in a trend over the last 5–7 years towards foregoing routine use of MBP before scheduled bowel surgery.
Why was the study conducted?
- •
There is paucity of data regarding use of bowel preparation in gynecologic surgery.
- •
This study was conducted to determine whether use of bowel preparation decreases rates of postoperative infectious or major morbidity following benign or malignant hysterectomies.
Key findings
- •
Use of any type of bowel preparation did not result in decreased rates of postoperative infectious or major morbidity compared to no bowel preparation following benign or malignant hysterectomies, irrespective of surgical approach or need for bowel surgery.
What does this add to what is known?
- •
Bowel preparation may be safely omitted in gynecologic surgery, especially in the context of well-established Enhanced Recovery After Surgery pathways.
- •
Practices with high baseline rates of surgical site infections may still consider using bowel preparation in the form of oral antibiotics alone or in combination with mechanical bowel preparation rather than mechanical bowel preparation alone.
While the concept of foregoing MBP before bowel surgery slowly started to garner increasing popularity among surgical practices, the discussion surrounding the use of oral antibiotics (OA) in combination with MBP resurfaced in the literature. The use of nonabsorbable luminal OA in combination with MBP was investigated as early as the 1970s. In an RCT, Rosenberg et al showed that combination of MBP with OA significantly decreased abdominal wound infection (23% vs 40%, P value not reported), sepsis (37.3% vs 64.4%, P < .05), and AL rates (24% vs 52%, P value not reported) compared to MBP alone following large bowel surgery. However, their use was slowly abandoned in favor of the use of MBP alone, as the addition of OA was associated with greater risk for preoperative dehydration and electrolyte abnormalities requiring prolonged preoperative hospital admissions and with decreased compliance over time. Modern studies suggest that preoperative use of OA as bowel preparation may reduce hospital length of stay and readmissions after CRS. Furthermore, a meta-analysis of RCTs found that the combination of OA with MBP as compared to MBP alone was associated with a decrease in the overall SSI rate (7.2% vs 16%, P < .001) as well as incisional SSI (4.6% vs 12.1%, P < .001) but no change in the deep organ space SSI (4% vs 4.8%, P = .056) following scheduled bowel surgery. These data resulted in a trend toward reintroduction of bowel preparation in the form of combined OA with MBP in CRS.
Data regarding bowel preparation are scarce in gynecologic surgery and are mostly limited to minimally invasive gynecologic surgery. These studies concluded that its use is not associated with easier bowel handling or superior visualization during laparoscopic surgery and thus its use should be abandoned. There are no studies examining the safety and efficacy of bowel preparation (MBP alone, OA alone, or combination of both) following laparotomy for gynecologic surgery with or without bowel resection. The practice is guided by data extrapolated from the CRS literature, which, as discussed above, has changed over the years. The practice among gynecologic surgery programs is thus highly variable and not standardized across the specialty. The objective of our study was thus to determine if MBP alone, OA alone, or a combination of the 2 are associated with decreased rates of SSI or AL compared to no bowel preparation following hysterectomies performed for benign or malignant indications.
Materials and Methods
Data Source
We conducted a retrospective, inverse propensity score–weighted cohort study using deidentified administrative claims data from the OptumLabs Data Warehouse, which includes medical and pharmacy claims, laboratory data, and enrollment records for commercial and Medicare Advantage enrollees. The database contains longitudinal health information on enrollees and patients, representing a diverse mixture of ages, ethnicities, and geographic regions across the United States. Because this study involved analysis of preexisting, deidentified data, it was exempt from the Institutional Review Board.
Study population
Women aged 18 years and older who underwent hysterectomy between January 2006 and July 2017 were identified. Inclusion criteria were hysterectomies for benign or malignant indications irrespective of surgical approach (laparotomy, laparoscopic, robotic, vaginal) with at least 12 months of medical and pharmacy coverage prior to surgery and 30 days of coverage after surgery. Patients were excluded if the surgical approach was not specified or if the preoperative bowel preparation status was unknown. Patients were categorized for comparisons based on the type of bowel preparation they received as MBP alone, OA alone, combination of MBP plus OA, and no bowel preparation.
The International Classification of Diseases, Ninth and Tenth Revision (ICD-9 and ICD-10, respectively) diagnosis and procedure codes as well as the Current Procedural Terminology codes were used in order to identify the study cohort and specifically the procedures of “hysterectomy” and “bowel resection,” the surgical approach of these procedures (laparotomy/open, minimally invasive laparoscopic/robotic, vaginal, laparoscopic-assisted vaginal), and the diagnosis of “cancer/malignancy” vs “noncancer” ( Appendix , Supplemental Table 1 ). The primary and secondary outcomes of this study were also determined by ICD-9 and ICD-10 codes ( Appendix , Supplemental Table 2 ). In order to identify information regarding preoperative use of bowel preparation, we reviewed pharmacy claims data for 45 days prior to and including the date of surgery. The list of MBPs and OA that were considered for inclusion in our study are presented in the Appendix ( Supplemental Table 3 ). In order to minimize the risk of capturing bowel preparations and antibiotics used for indications other than preoperative bowel preparation, we restricted the bowel preparation claims to 3-day supply only for the preparations that can be used for other indications such as chronic constipation and all antibiotics claims to OA and 3-day supply only. Patient characteristics and information pertaining to surgery, such as year and place of surgery (inpatient, outpatient), were available from the OptumLabs database.
Outcomes
The primary outcome was the rate of SSI. Secondary outcomes included rates of AL, postoperative ileus, and major morbidity. SSI included superficial, deep, and organ-space infections. Major morbidity was defined as cardiac arrest, deep vein thrombosis, myocardial infarction, pneumonia, acute renal failure, septic shock, sepsis, urinary tract infection, or wound dehiscence. All rates were measured at 30 days postoperatively. Outcomes were assessed independently based on the indication of the surgery (benign vs malignant).
Statistical analysis
We used inverse propensity score weighting separately for benign and malignant cohorts to balance baseline patient characteristics between those who received any type of bowel preparation (MBP alone, OA alone, or combination of MBP with OA) and those with no bowel preparation. We analyzed outcomes among all hysterectomies, abdominal hysterectomies only, and abdominal hysterectomies with and without bowel resection. Patient characteristics in the propensity score model included age at time of hysterectomy, year of surgery, type of surgery (open, laparotomy/open, minimally invasive laparoscopic/robotic, vaginal, laparoscopic-assisted vaginal), bowel resection (yes/no), and place of service (inpatient/outpatient). Standardized differences of patients’ baseline characteristics were used to assess the balance between the weighted groups ( Appendix , Supplemental Table 4 ). We assessed primary outcomes using multivariate logistic regression analyses that adjusted for race, census region, household income, diabetes, and other variables that were included in propensity score weighting but were not balanced between groups following weighting. Adjusted odds ratios (OR) and 95% confidence intervals (CI) are reported. A P value of <.05 was used for statistical significance. All statistical analyses were performed using SAS software version 9.4 (SAS Institute Inc, Cary, NC).
Results
Patient characteristics
A total of 224,687 patients undergoing hysterectomy between January 2006 and July 2017 were identified using the OptumLabs database. Patient characteristics are presented in Table 1 . Of the total, 186,148 surgeries (82.9%) were performed for benign and 38,539 (17.1%) for malignant indications. The patients undergoing surgery for malignant indications tended to be older compared to their benign counterparts with a median age of 54 vs 45 years of age. The most common surgical approach was vaginal for the benign cohort as compared to open laparotomy for the malignant cohort. As expected, there were more cases of bowel resection when hysterectomies were performed for malignant indications compared to benign. The majority of the patients did not receive bowel preparation before hysterectomy; however, among those who received some type of bowel preparation, MBP was the most frequently used in both the benign as well as the malignant cohort.
| Indication for surgery | ||
|---|---|---|
| Benign (n = 186,148) | Malignant (n = 38,539) | |
| Age in years (median, IQR) | 45 (40, 50) | 54 (46, 63) |
| Race | ||
| Asian | 3971 (2.1%) | 927 (2.4%) |
| Black | 26,133 (14%) | 4094 (10.6%) |
| Hispanic | 18,566 (10%) | 3181 (8.3%) |
| White | 130,263 (70%) | 28,551 (74.1%) |
| Unknown | 7215 (3.9%) | 1786 (4.6%) |
| Diabetes | 7632 (4.1%) | 3627 (9.4%) |
| Surgery type | ||
| Laparoscopy/robotic | 50,616 (27.2%) | 11,105 (28.8%) |
| Laparotomy | 60,748 (32.6%) | 18,370 (47.7%) |
| Vaginal | 74,784 (40.2%) | 9064 (23.5%) |
| Bowel resection | 816 (0.4%) | 1071 (2.8%) |
| Type of bowel preparation | ||
| None | 174,613 (93.8%) | 33,593 (87.2%) |
| MBP only | 8557 (4.6%) | 3714 (9.6%) |
| OA only | 2082 (1.1%) | 707 (1.8%) |
| MPB+OA only | 896 (0.5%) | 525 (1.4%) |
Bowel preparation
Among the entire cohort of hysterectomies (vaginal, laparoscopic/robotic, and open), use of bowel preparation did not result in decreased rates of SSI, AL, postoperative ileus, or major morbidity irrespective of benign or malignant indication for surgery ( Table 2 ) Similarly, when focusing the analysis on the subgroup of abdominal hysterectomies, use of bowel preparation did not improve outcomes among the benign or malignant cohort ( Table 2 ). Interestingly, there was a trend toward higher rates of postoperative ileus when bowel preparation was used following malignant hysterectomies (no bowel preparation vs any type of bowel preparation: 1.6% vs 3.1%; OR, 1.45; 95% CI, 0.98–2.16, P = .06; Table 2 ), which was found to be statistically significant when limiting the analysis to the subgroup of abdominal hysterectomies for malignant indications (3.2% vs 5.2%; OR, 1.64; 95% CI, 1.13–2.37, P = .01; Table 2 ).
| Any type of bowel preparation (MBP and/or OA) | Adjusted | |||
|---|---|---|---|---|
| No | Yes | OR (95% CI) | P value | |
| ALL HYSTERECTOMY TYPES | ||||
| Benign cohort | 174,613 | 11,535 | ||
| SSI | 5923 (3.4%) | 521 (4.5%) | 1.07 (0.77, 1.50) | .68 |
| AL | 747 (0.4%) | 70 (0.6%) | 0.64 (0.27, 1.50) | .30 |
| Postoperative ileus | 1193 (0.7%) | 154 (1.3%) | 1.46 (0.81, 2.63) | .20 |
| Major morbidity | 4184 (2.4%) | 368 (3.2%) | 1.03 (0.70, 1.52) | .88 |
| Malignant cohort | 33,593 | 4946 | ||
| SSI | 1781 (5.3%) | 324 (6.6%) | 1.05 (0.79, 1.39) | .74 |
| AL | 324 (1.0%) | 75 (1.5%) | 1.16 (0.66, 2.03) | .61 |
| Postoperative ileus | 554 (1.6%) | 155 (3.1%) | 1.45 (0.98, 2.16) | .06 |
| Major morbidity | 1775 (5.3%) | 330 (6.7%) | 1.04 (0.79, 1.38) | .78 |
| ABDOMINAL HYSTERECTOMIES | ||||
| Benign cohort | 56,558 | 4190 | ||
| SSI | 2958 (5.2%) | 303 (7.2%) | 1.01 (0.66, 1.53) | .98 |
| AL | 391 (0.7%) | 48 (1.1%) | 0.55 (0.21, 1.44) | .22 |
| Postoperative ileus | 854 (1.5%) | 122 (2.9%) | 1.44 (0.77, 2.70) | .25 |
| Major morbidity | 2182 (3.9%) | 227 (5.4%) | 1.01 (0.63, 1.62) | .96 |
| Malignant cohort | 15,528 | 2842 | ||
| SSI | 1231 (7.9%) | 245 (8.6%) | 1.11 (0.83, 1.48) | .49 |
| AL | 244 (1.6%) | 64 (2.3%) | 1.43 (0.82, 2.51) | .21 |
| Postoperative ileus | 494 (3.2%) | 148 (5.2%) | 1.64 (1.13, 2.37) | .01 |
| Major morbidity | 1329 (8.6%) | 271 (9.5%) | 1.12 (0.85, 1.48) | .43 |
Individual types of bowel preparation
When comparing outcomes between individual types of bowel preparation vs no bowel preparation, there was no difference in the rates of infectious morbidity and postoperative ileus between MBP alone, OA alone, or the combination of MBP with OA compared to no bowel preparation among both cohorts of benign as well as malignant abdominal hysterectomies ( Tables 3 and 4 , respectively). Importantly, even within the subgroup of malignant abdominal hysterectomies combined with bowel resection at time of hysterectomy, which represents the highest risk for postoperative infectious morbidity subgroup of hysterectomies, use of individual types of bowel preparation was not associated with decreased rates of SSI, AL, postoperative ileus, and major morbidity compared to no bowel preparation ( Table 4 ).
| Overall (± bowel resection) | Bowel preparation | Adjusted | ||
|---|---|---|---|---|
| None | Yes | OR (95% CI) | P value | |
| MBP vs no preparation | 56,558 | 2923 | ||
| SSI | 2958 (5.2%) | 215 (7.4%) | 1.18 (0.65, 2.17) | .59 |
| AL | 391 (0.7%) | 32 (1.1%) | 0.75 (0.17, 3.34) | .70 |
| Postoperative ileus | 854 (1.5%) | 84 (2.9%) | 1.61 (0.64, 4.05) | .31 |
| Major morbidity | 2182 (3.9%) | 161 (5.5%) | 1.18 (0.60, 2.34) | .63 |
| OA vs no preparation | 56,558 | 771 | ||
| SSI | 2958 (5.2%) | 48 (6.2%) | 0.85 (0.07, 10.75) | .90 |
| AL | 391 (0.7%) | — a | — b | — b |
| Postoperative ileus | 854 (1.5%) | 12 (1.6%) | 0.75 (0.01, 93.91 | .91 |
| Major morbidity | 2182 (3.9%) | 38 (4.9%) | 0.89 (0.05, 15.16) | .93 |
| MBP+OA vs no preparation | 56,558 | 496 | ||
| SSI | 2958 (5.2%) | 40 (8.1%) | 0.83 (0.03, 26.80) | .92 |
| AL | 391 (0.7%) | — a | — b | — b |
| Postoperative ileus | 854 (1.5%) | 26 (5.2%) | 1.92 (0.03, 128.78) | .76 |
| Major morbidity | 2182 (3.9%) | 28 (5.6%) | 0.75 (0.01, 43.07) | .89 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
