Introduction

Lower urinary tract injury (LUTI) at the time of hysterectomy is a rare but highly morbid complication. Multiple risk factors for LUTI at the time of hysterectomy have been identified, including route of surgery, indication for hysterectomy, increased blood loss, concomitant surgical procedures, and surgical technique. Moreover, immediate intraoperative identification and treatment are associated with improved clinical outcomes and decreased litigation. However, observational studies have shown that 54% to 86% of injuries are recognized in the postoperative period.

Postoperative diagnosis of LUTI often results from 2 sources: failure to recognize injury at the time of hysterectomy or a delayed mechanism of injury. Although sharp genitourinary injury or complete ureteral transection can be diagnosed at the time of surgery, LUTI from thermal injury is more insidious. Undiagnosed injuries are associated with greater risk of serious sequelae, such as bladder and ureteral fistulae, renal failure, infection, and death. ^{, ,}

Limited research exists describing risk factors specific to delayed diagnosis of LUTI and temporal trends regarding the interval between surgery and diagnosis. The objective of this study was to describe the type of and time until diagnosis of LUTI in the 30-day postoperative period using a large prospective national surgical database; we also aimed to identify factors associated with differences in these outcomes.

Methods

This was a secondary analysis of the 2014 to 2018 National Surgical Quality Improvement Program (NSQIP) hysterectomy target files. The American College of Surgeons (ACS) NSQIP database is a prospective surgical registry that hospitals use to track their risk-adjusted outcomes after surgery. Hospital participation is voluntary and not designed or stratified to be a representative sampling. Data collected include preoperative characteristics and health status, intraoperative events and outcomes, and postoperative outcomes for 30 days following the incident surgery. Data entry is periodically audited to ensure accuracy and quality of the data.

Since 2014, the ACS has maintained a hysterectomy target file with additional information regarding patient and surgeon characteristics of interest, such as parity, surgeon specialty, or history of abdominal surgery. All women from the hysterectomy target file were included in this analysis; individuals with a gynecologic malignancy were excluded.

Demographic and patient-related preoperative variables extracted included age, race and ethnicity, body mass index, baseline functional status, the American Society of Anesthesiologists classification, smoking within the previous year, hypertension requiring medication, diabetes mellitus requiring insulin or oral therapy, parity, history of previous abdominal surgery, and history of previous pelvic surgery. Major medical comorbidities were categorized into cardiac, pulmonary, and neurologic categories. Major cardiac comorbidities include congestive heart failure with exacerbation within 1 month of surgery, myocardial infarction within 6 months of surgery, history of peripheral vascular disease requiring treatment, history of percutaneous coronary intervention, and previous cardiac surgery. Major pulmonary comorbidities include history of severe chronic obstructive pulmonary disease. Major neurologic comorbidities were defined as a history of stroke with or without residual neurologic deficit or history of transient ischemic attack. Definitions were drawn from the NSQIP participant use file and coded as yes, no, or unknown.

Operative characteristics extracted included surgeon specialty, history or intraoperative evidence of endometriosis, and surgical approach to hysterectomy, as defined by the current procedural terminology (CPT) code ( Supplemental Table ). Individuals with concomitant cystoscopy at the time of hysterectomy and concomitant apical suspension were also identified using CPT coding.

LUTI is identified in the hysterectomy target file by the following classification: ureteral obstruction, ureteral fistula, or bladder fistula per NSQIP coding guidelines. We report these as complex LUTI (cLUTI) for the remainder of the document as these data exclude simple cystotomy, which is the most common form of LUTI diagnosed at the time of surgery. The presence of these injuries and the number of days from surgery to diagnosis were extracted. For the purposes of this study, “delayed diagnosis” was defined as diagnosis at any time following the day of surgery. Our primary outcome was the presence of any cLUTI, defined as presence of at least 1 of the above injuries. We also performed a subgroup analysis to assess risk factors specific to ureteral obstruction or LUT fistula, defined as presence of ureteral or bladder fistula. Of note, the subgroup analysis outcomes were not mutually exclusive, and when multiple injuries were present, the shortest time to diagnosis was utilized in the combined cLUTI measure. Patients with missing uterine weight data (4.4%) were excluded from multivariable analysis.

Descriptive analysis was performed for each of the variables and comparisons made for all intraoperative characteristics between individuals with and without cLUTI and time until cLUTI. These analyses were performed using Kruskal-Wallis, Pearson χ ² , or Fisher exact tests where appropriate. To better assess differences in time to diagnosis of cLUTI, we generated, Kaplan-Meier curves for each variable with a significant relationship to cLUTI on bivariate analysis. Individuals without a cLUTI were assumed to have complete data and were censored at 31 days. Violations of proportional hazards among groups were assessed visually and with log-rank tests.

Next, multivariable logistic regression was used to assess for independent relationships between intraoperative factors and the odds of any cLUTI. Nonhomogeneity between ureteral obstruction and LUT fistula was confirmed and subgroup multivariable analysis performed. Cox proportional hazards analysis was utilized to assess for differences in time until recognition of injury, controlling for demographic and other intraoperative factors. Proportional hazards assumptions of the model were tested using cumulative martingale residuals and the supremum test. If the supremum test was significant, an interaction term between days from surgery and the variable of interest was included in the model to assess for a significant temporal trend.

Given there may be unobserved factors affecting likelihood of cLUTI diagnosis on the day of surgery or during initial hospitalization, a second set of sensitivity analyses were performed using only those individuals with a delayed cLUTI, defined as ≥1 day from surgery to diagnosis as defined by the NSQIP, to confirm the findings of the logistic regression models. All analyses were performed using SAS (version 4.6; SAS Institute, Cary, NC). The study was determined to be institutional review board (IRB) exempt, as reviewed by the Northwestern University IRB (STU00212227).

Results

In this study, 100,823 women in the 2014 to 2018 NSQIP hysterectomy target files met the inclusion criteria. Most women were white (58%) and functionally independent (99%), with low rates of medical comorbidities ( Table 1 ). More than half of the women had previous pelvic surgery (55.8%), and 13.4% of the women had a postoperative diagnosis of endometriosis; only 20.3% of the women had concomitant cystoscopy coded at the time of hysterectomy. The mode of hysterectomy differed by surgeon specialty ( P <.01). Gynecologic oncologists were more likely to perform cases via total abdominal hysterectomy (TAH) (34.6%) or total laparoscopic hysterectomy (TLH) (55.6%), whereas urogynecologists were more likely to perform a total vaginal hysterectomy (TVH) (52.5%). Obstetrician-gynecologists were the primary surgeon in 72.5% of cases and had proportional utilization of all forms of hysterectomy except for laparoscopic assisted vaginal hysterectomy (LAVH), where they performed a disproportionally high number (85.2%) of all cases. Intraoperative characteristics are presented in Table 2 , stratified by injury type.

Table 2

Cumulative incidence of and median time until diagnosis of complex cLUTI after benign hysterectomy, stratified by intraoperative characteristics

Variable	cLUTI				Ureteral obstruction				LUT fistula
	30-day incidence	P value	Days to diagnosis	P value	30-day incidence	P value	Days to diagnosis	P value	30-day incidence	P value	Days to diagnosis	P value
Total	208 (0.206)	—	—	—	111 (0.110)	—	—	—	105 (0.104)	—	—	—
Surgeon specialty	—	.276	—	.086	—	.03 b	—	.049 b	—	.391	—	.477
Gynecologic oncology	26 (0.141)	—	16.5 (5–24)	—	12 (0.065)	—	16.5 (3.0–25.5)	—	14 (0.076)	—	16.5 (6.0–23.0)	—
Generalist	163 (0.223)	—	7.0 (2.0–16.0)	—	86 (0.118)	—	3.5 (2.0–11.0)	—	84 (0.115)	—	13.5 (5.0–22.0)	—
Urogynecologist	18 (0.224)	—	6.0 (1.0–19.0)	—	13 (0.162)	—	7.0 (1.0–19.0)	—	6 (0.075)	—	9.5 (5.0–11.0)	—
Other a	1 (0.078)	—	4	—	0 (0.000)	—	—	—	1 (0.078)	—	4	—
Method of hysterectomy	—	.032 b	—	.014 b	—	.003 b	—	.381	—	.001 b	—	.457
TLH	78 (0.165)	—	8.5 (3.0–15.0)	—	37 (0.078)	—	7.0 (1.0–14.0)	—	44 (0.093)	—	11.0 (4.5–20.5)	—
LAVH	34 (0.264)	—	14.0 (5.0–21.0)	—	10 (0.078)	—	11.0 (4.0–16.0)	—	26 (0.202)	—	15.5 (8.0–23.0)	—
TAH	63 (0.255)	—	6.0 (2.0–20.0)	—	39 (0.158)	—	3.0 (2.0–11.0)	—	25 (0.101)	—	18.0 (5.0–22.0)	—
TVH	33 (0.206)	—	3.0 (1.0–11.0)	—	25 (0.156)	—	3.0 (1.0–8.0)	—	10 (0.062)	—	12.5 (1.0–22.0)	—
Uterine weight (g)	—	.365	—	.230	—	.578	—	.853	—	.779	—	.211
<100	57 (0.183)	—	7.0 (2.0–20.0)	—	31 (0.099)	—	6.0 (1.0–16.0)	—	28 (0.090)	—	10.5 (4.0–23.5)	—
100–249	83 (0.213)	—	10.0 (2.0–16.0)	—	44 (0.113)	—	5.0 (1.5–13.5)	—	43 (0.110)	—	14.0 (6.0–21.0)	—
250–499	26 (0.191)	—	14.0 (4.0–22.0)	—	13 (0.095)	—	5.0 (2.0–14.0)	—	14 (0.103)	—	21.0 (12.0–24.0)	—
≥500	33 (0.265)	—	5.0 (2.0–14.0)	—	18 (0.145)	—	2.5 (2.0–10.0)	—	15 (0.120)	—	11.0 (4.0–15.0)	—
Diagnosis of endometriosis	—	<.001 b	—	.827	—	<.001 b	—	.502	—	.043 b	—	—
No	161 (0.184)	—	8.0 (2.0–18.0)	—	82 (0.094)	—	4.0 (2.0–11.0)	—	84 (0.096)	—	13.0 (5.0–22.0)	—
Yes	47 (0.350)	—	7.0 (2.0–23.0)	—	29 (0.216)	—	7.0 (1.0–16.0)	—	21 (0.157)	—	14.0 (5.0–23.0)	—
Concomitant apical suspension	—	.068	—	.011 b	—	<.001 b	—	.655	—	.37	—	.085
No	184 (0.198)	—	9.0 (3.0–20.0)	—	91 (0.098)	—	5.0 (2.0–13.0)	—	99 (0.107)	—	14.0 (5.0–22.0)	—
Yes	24 (0.293)	—	3.0 (0.5–12.5)	—	20 (0.245)	—	4.5 (1.0–11.5)	—	6 (0.073)	—	4.5 (1.0–11.0)	—

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Professionally responsible coronavirus disease 2019 vaccination counseling of obstetrical and gynecologic patients
2. Balancing risks: making decisions for maternal treatment without data on fetal safety
3. The coronavirus disease 2019 vaccine in pregnancy: risks, benefits, and recommendations
4. Quantitative and qualitative assessment of an all-inclusive postpartum human papillomavirus vaccination program

Stay updated, free articles. Join our Telegram channel

Join