Right Colectomy and Abdominal Perineal Resection for Cancer: Do Urinary Tract Infections Impact Outcomes?

Abstract

Background:

We aim to assess the patient factors and concomitant infectious outcomes associated with urinary tract infection (UTI) occurrence and the impact of UTI on length of stay (LOS), re-admission, and death in a colorectal surgical population.

Patients and Methods:

National Surgical Quality Improvement Program User Data for right colectomy and abdominal perineal resection (APR) procedures for cancer between 2006 and 2012 were analyzed. Concomitant infectious complications and timing of UTI diagnosis, inpatient versus outpatient, were considered.

Results:

We identified 7,615 right colectomies with 107 (1.4%) UTIs and 2,493 APRs with 88 (3.5%) UTIs (p < 0.001). On multivariable analysis and correction for other post-operative complications, UTI remained statistically correlated with prolonged LOS for right colectomy and APR (LOS increases of 59.0% and 37.4%, respectively, p < 0.001) but not death. Patients with a diagnosis of UTI after discharge showed significantly increased re-admission rates compared with UTI diagnosis before discharge (37.7% vs. 9.7%, p < 0.001).

Conclusions:

After excluding deaths, outpatient UTI occurrences, and correcting for other infectious complications, UTI is associated with increased LOS but is not correlated with re-admission or death. Outpatient occurrence of UTI after hospital discharge is associated with a dramatic re-admission rate of 37.7%.

Urinary tract infections (UTI) are a frequent complication of colectomy and proctectomy. The incidence has been reported at 0.4%–5%, and UTI has been reported as the fifth most common complication after colorectal resection [1 –4]. The significance of nosocomial UTI has been described at length in the medical literature, where its occurrence profoundly impacts death and hospital costs. In the colorectal literature, UTI is a proposed “pay for performance” outcome measure [1,5 –7]. While UTI has been associated with prolonged length of stay (LOS) in colorectal resection procedures, it has not been associated with death, and a specific analysis of the risks and outcomes of UTI after colorectal resection has not been performed [1,2]. Further, the impact of the underlying colorectal disease process on UTI occurrence is still being determined in the colorectal literature [8].

Two studies in the literature have isolated the patient factors correlated with UTI in colorectal resection and the impact of these infections on outcomes. Both studies focused on pre-operative patient risk factors and the increased risk of colorectal resection procedures in comparison with other surgical specialties [1,2]. In these studies, the use of National Surgical Quality Improvement Program (NSQIP) UTI outcomes has not accounted for the occurrence of other post-operative infectious outcomes and outpatient UTI events. In addition, the relationship between UTI, as either an inpatient or outpatient occurrence, and re-admission has not been examined.

We aim to assess the patient factors and concomitant infectious outcomes associated with UTI occurrence and the impact of UTI on LOS, re-admission, and death in a homogenous population of patients undergoing colectomies and proctectomies performed for cancer. In addition, we aim to determine the impact on re-admission of UTIs diagnosed during the post-operative hospital stay compared with those diagnosed after hospital discharge.

Patients and Methods

The University of Iowa Hospitals and Clinics Institutional Review Board waived the review of this study. Data were obtained from NSQIP Participant User Files 2006–2012. The primary outcomes measured were differences in death, LOS, and re-admission in patients with and without UTI. Additional analyses were performed to assess for factors associated with UTI occurrence and for correlates in re-admission for patients who had a diagnosis of UTI after hospital discharge. Our patient population was defined by the diagnosis of colorectal cancer and by the procedure performed—right colectomy or abdominal perineal resection (APR). The diagnosis of colorectal cancer was chosen for its low propensity for infectious complications compared with inflammatory diagnoses. The procedures of right colectomy and APR were chosen to provide a homogeneous procedural population.

Right colectomy was defined by Current Procedural Terminology (CPT) codes 44160 and 44205 and APR by codes 45395 and 45110. Colorectal cancer diagnoses included were identified by the International Classification of Diseases, 9th Revision (ICD9) code 153.0–153.9, 154.0, and 154.1. Initially 14,601 patient cases were identified. Duplicated observations (n = 211), emergency procedures (n = 523), and procedures in which the operation was not performed on the admission date (n = 2,719) were excluded. Patients having UTI present on admission (n = 8) and patients with more than one UTI (n = 2) were excluded. The 1,101 patients with missing demographic and operative details, or outlier values considered likely to be entered in error, were also excluded. After exclusions, 10,108 unique patient cases were available for analysis.

Pre-operative characteristics and outcomes selected for analysis are shown in the presented tables. Standard definitions for all variables are provided by NSQIP [9]. For our primary analysis including LOS, re-admission, death, and factors associated with UTI, patients with UTI that was diagnosed as an outpatient were excluded. Outpatient UTI was defined by the variable days from operation to UTI occurrence being greater than LOS. A second analysis was performed for risk of re-admission for the patients with UTI diagnosed as an outpatient. Because re-admission data are available only for 2011–2012 in our data set, the analysis of re-admission was limited to those years.

Statistical analyses were performed using Stata 13.1 (StataCorp, College Station, TX). Analysis of variance (ANOVA) was used for comparing numeric variables. For comparisons of categoric variables, the chi square and Fisher exact tests were used. Outcome variables suspected of log-normal distribution were log transformed for regression analysis. Cases with missing values in regression variables were excluded from individual models. A p value of <0.05 was considered statistically significant, and all tests were two-tailed.

Results

Of the 10,108 cases analyzed, 7,615 were right colectomies and 2,493 were APR. In the right colectomy group, there were 107 UTIs (UTI rate of 1.4%), and in the APR group, there were 88 UTIs (UTI rate of 3.6%, p < 0.001).

Table 1 illustrates patient characteristics, demographics, and co-morbidities analyzed for patients with and without UTI in the right colectomy and APR groups. For right colectomy, patients with UTI were older and more commonly had a history of chronic obstructive pulmonary disease, transient ischemic attack, cardiac surgery, and dyspnea. Open right colectomy was more commonly performed in patients with UTI (57.9%) compared with the laparoscopic approach (42.1%, p = 0.025). For APR, female gender, patient history of diabetes, and disseminated cancer were more common in patients with UTI. For both right colectomy and APR, higher body mass index and American Society of Anesthesiologists score, lower hematocrit, and patients with dependent functional status were more common in patients with UTI.

Table 1.

Patient Demographics, Co-Morbidity, and Operative Characteristics by Operation Performed and Urinary Tract Infection Occurrence

	Right colectomy			APR
Characteristic	UTI n = 107	No UTI n = 7,508	p	UTI n = 88	No UTI n = 2,405	p
Age (years)	72.32 ± 10.98	68.61 ± 12.45	0.003	62.21 ± 10.96	59.61 ± 12.92	0.068
Body mass index	29.52 ± 6.69	28.27 ± 6.52	0.049	29.24 ± 8.18	27.42 ± 6.30	0.009
Male gender	41 (38.3%)	3492 (46.5%)	0.104	44 (50.0%)	1528 (63.5%)	0.009
White	80 (74.8%)	6019 (80.2%)	0.608	61 (69.3%)	1923 (80.0%)	0.136
ASA classification			0.014			0.003
1–2	35 (32.7%)	3350 (44.6%)		26 (29.5%)	1093 (45.4%)
3,4,5	70 (65.4%)	4158 (55.4%)		62 (70.5%)	1312 (54.6%)
Pre-operative laboratories
Albumin	3.80 ± 0.50	3.86 ± 0.52	0.265	3.78 ± 0.57	3.89 ± 0.53	0.132
Hematocrit	34.85 ± 5.84	36.38 ± 5.44	0.004	37.00 ± 4.56	38.30 ± 4.52	0.010
Creatinine	1.07 ± 0.56	1.00 ± 0.53	0.193	0.98 ± 0.33	0.95 ± 0.58	0.676
INR	1.12 ± 0.32	1.07 ± 0.23	0.072	1.03 ± 0.12	1.04 ± 0.16	0.681
Co-morbidity
Smoking history	9 (8.4%)	931 (12.4%)	0.213	18 (20.5%)	523 (21.7%)	0.773
Alcoholic	5 (4.7%)	164 (2.2%)	0.186	2 (2.3%)	107 (4.4%)	0.582
Chronic heart failure	1 (0.9%)	45 (0.6%)	0.479	0 (0%)	5 (0.2%)	1.000
COPD	14 (13.1%)	444 (5.9%)	0.002	5 (5.7%)	105 (4.4%)	0.591
MI history	2 (1.9%)	45 (0.6%)	0.161	0 (0%)	4 (0.2%)	1.000
History of angina	2 (1.9%)	34 (0.5%)	0.104	0 (0%)	7 (0.3%)	1.000
CVA	5 (4.7%)	241 (3.2%)	0.398	4 (4.5%)	71 (3.0%)	0.336
TIA	6 (5.6%)	173 (2.3%)	0.043	2 (2.3%)	51 (2.1%)	0.700
Diabetes	18 (16.8%)	1426 (19.0%)	0.570	21 (23.9%)	329 (13.7%)	0.007
Cardiac surgery	11 (10.3%)	378 (5.0%)	0.030	2 (2.3%)	80 (3.3%)	1.000
Dyspnea	25 (23.4%)	1092 (14.5%)	0.010	10 (11.4%)	205 (8.5%)	0.351
Transfusion	0 (0%)	28 (0.4%)	1.000	1 (1.1%)	1 (0.04%)	0.069
Weight loss >10%	3 (2.8%)	333 (4.4%)	0.632	8 (9.1%)	159 (6.6%)	0.361
Steroid use	4 (3.7%)	178 (2.4%)	0.327	2 (2.3%)	40 (1.7%)	0.659
Chemotherapy	0 (0%)	54 (0.7%)	1.000	7 (8.0%)	262 (10.9%)	0.447
Radiation therapy	0 (0%)	30 (0.4%)	1.000	24 (27.3%)	837 (34.8%)	0.190
Procedure
Open	62 (57.9%)	3533 (47.1%)	0.025	71 (80.7%)	1809 (75.2%)	0.242
Laparoscopic	45 (42.1%)	3975 (52.9%)		17 (19.3%)	596 (24.8%)
Wound class
Clean contaminated	100 (93.5%)	7063 (94.1%)	0.789	76 (86.4%)	2136 (88.8%)	0.475
Contaminated	6 (5.6%)	311 (4.1%)	0.451	7 (8.0%)	217 (1.1%)	0.731
Dirty infected	1 (0.9%)	66 (0.9%)	0.614	4 (4.5%)	46 (1.9%)	0.098
Other characteristics
Pre-operative sepsis	0 (0%)	59 (0.8%)	1.000	0 (0%)	18 (0.7%)	1.000
Pre-operative functional status			0.010			0.001
Total dependence	6 (5.6%)	188 (2.5%)		8 (9.1%)	43 (1.8%)
Partial dependence	1 (0.9%)	6 (0.1%)		0 (0%)	3 (0.1%)
Pre-operative ascites	1 (0.9%)	35 (0.5%)	0.400	1 (1.1%)	2 (0.1%)	0.102
Disseminated cancer	6 (5.6%)	439 (5.8%)	0.916	16 (18.2%)	167 (6.9%)	<0.001
Pre-operative need for dialysis	1 (0.9%)	26 (0.3%)	0.318	0 (0%)	6 (0.2%)	1.000

APR = abdominal perineal resection; UTI = urinary tract infection; ASA = American Society of Anesthesiologists Physical Status Classification; INR = international normalized ratio; COPD = chronic obstructive pulmonary disease; MI = myocardial infarction; CVA = cerebrovascular accident; TIA = transient ischemic attack.

Table 2 shows concomitant outcomes for the right colectomy and APR groups with and without UTI. The LOS was noticeably longer in both surgical groups with UTI occurrences (right colectomy 13.9 d vs. 6.1 d, p < 0.001; APR 13.7 d vs. 8.5 d, p < 0.001). Right colectomy patients with UTI also had increased rates of numerous complications including surgical site infections, wound dehiscence, pneumonia, renal failure, bleeding, cardiac complications, sepsis, and septic shock. For APR patients, the concomitant complications were less numerous but included deep surgical site infection, renal failure, sepsis, and septic shock. Inpatient UTI was present in two (5.7%) patients in the right colectomy group and four (14.8%) patients in the APR group who were re-admitted and did not correlate with re-admission in either group (right colectomy p = 0.699; APR p = 0.446). Death was significantly higher in right colectomy patients with UTI (3.7% vs. 1.1%, p = 0.035).

Table 2.

Concomitant Outcomes by Procedure Performed and Urinary Tract Infection Occurrence

	Right colectomy			APR
Outcome	UTI n = 107	No UTI n = 7,508	p	UTI n = 88	No UTI n = 2,405	p
Procedure time (min)	140.39 ± 60.90	136.62 ± 57.75	0.503	241.28 ± 60.10	225.90 ± 66.31	0.032
Length of stay (d)	13.91 ± 13.02	6.06 ± 4.83	<0.001	13.67 ± 9.18	8.50 ± 6.63	<0.001
Superficial SSI	18 (16.8%)	442 (5.9%)	<0.001	13 (14.77%)	281 (11.7%)	0.378
Deep SSI	3 (2.8%)	65 (0.9%)	0.070	7 (8.0%)	77 (3.2%)	0.015
Organ space SSI	8 (7.5%)	173 (2.3%)	<0.001	6 (6.8%)	101 (4.2%)	0.234
Any SSI	28 (26.2%)	662 (8.8%)	<0.001	26 (29.6%)	441 (18.3%)	0.008
Wound dehiscence	5 (4.7%)	62 (0.8%)	0.002	3 (3.4%)	58 (2.4%)	0.474
Pneumonia	10 (9.4%)	161 (2.1%)	<0.001	5 (5.7%)	55 (2.3%)	0.058
Failure to wean ventilator within 48 h	8 (7.5%)	91 (1.2%)	<0.001	3 (3.4%)	28 (1.2%)	0.094
Acute renal failure	6 (5.6%)	17 (0.2%)	<0.001	2 (2.3%)	9 (0.4%)	0.055
Progressive renal failure	4 (3.7%)	41 (0.6%)	0.004	4 (4.6%)	10 (0.4%)	0.001
Bleeding	15 (14.0%)	414 (5.5%)	<0.001	8 (9.1%)	165 (6.9%)	0.419
Stroke/coma	0 (0%)	23 (0.3%)	1.000	1 (1.1%)	8 (0.3%)	0.277
MI/cardiac arrest	4 (3.7%)	83 (1.1%)	0.034	1 (1.1%)	25 (1.0%)	0.609
Sepsis	15 (14.0%)	168 (2.2%)	<0.001	13 (14.8%)	88 (3.7%)	<0.001
Septic shock	11 (10.3%)	78 (1.0%)	<0.001	4 (4.6%)	28 (1.2%)	0.025
Sepsis + septic shock	26 (24.3%)	243 (3.2%)	<0.001	17 (19.3%)	115 (4.8%)	<0.001
30 day death	4 (3.7%)	84 (1.1%)	0.035	3 (3.4%)	22 (0.9%)	0.056
For 2011–2012	n = 35	n = 3023		n = 27	n = 703
Re-admission	2 (5.7%)	149 (4.9%)	0.699	4 (14.8%)	70 (10.0%)	0.446

APR = abdominal perineal resection; UTI = urinary tract infection; SSI = surgical site infection; Any SSI = any occurrence of superficial, deep, or organ space SSI; MI = myocardial infarction.

Clinically and statistically significant pre-operative and operative characteristics and post-operative complications from our univariable analysis of patients with and without UTI were used in a multivariable model to look for correlations with longer LOS (Table 3). For this LOS analysis, deaths and outpatient UTI occurrences were excluded. For right colectomy and APR, UTI remained statistically correlated with prolonged LOS with increases of 59.0% and 37.4%, respectively (p < 0.001). Other characteristics associated with LOS are shown in Table 3.

Table 3.

Multivariable Analysis of Characteristics Associated with Length of Stay ^*

	Right colectomy				APR
	LOS increase	95% CI		p	LOS increase	95% CI		p
Age (10 y)	4.9%	4.0%	5.8%	<0.001	4.2%	2.8%	5.8%	<0.001
ASA (3,4,5 vs. 1,2)	8.9%	6.6%	11.3%	<0.001	6.1%	2.2%	10.1%	0.002
Dyspnea	4.8%	1.7%	8.0%	0.002	11.9%	4.8%	19.6%	0.001
CHF	32.1%	15.4%	51.3%	<0.001	29.9%	−16.7%	102.7%	0.249
COPD	9.9%	5.1%	15.0%	<0.001	8.7%	−0.8%	19.2%	0.073
Laparoscopic	−21.1%	−22.7%	−19.5%	<0.001	−22.3%	−25.5%	−19.0%	<0.001
Any SSI	25.2%	20.7%	30.0%	<0.001	22.8%	17.0%	28.8%	<0.001
Pneumonia	95.4%	81.3%	110.6%	<0.001	66.9%	47.1%	89.4%	<0.001
Sepsis + septic shock	64.4%	54.2%	75.2%	<0.001	44.7%	32.6%	57.9%	<0.001
UTI	59.0%	45.5%	73.7%	<0.001	37.4%	24.2%	52.0%	<0.001

Adjusted R²				43.94%				43.74%

Patients who died within 30 days are excluded.

APR = abdominal perineal resection; LOS = length of stay; ACI = confidence interval; ASA = American Society of Anesthesiologists Physical Status Classification; CHF = congestive heart failure, COPD = chronic obstructive pulmonary disease; SSI = surgical site infection; UTI = urinary tract infection.

Table 4 illustrates factors associated with the occurrence of UTI in the right colectomy and APR population after multivariable analysis, again with models including clinically significant pre-operative, operative, and post-operative outcomes. For both procedures, male gender was associated with less UTI occurrence, and sepsis or septic shock was associated with UTI occurrence. For right colectomy, age and any surgical site infection was associated with UTI occurrence. For APR, a poorer pre-operative functional status was also seen to be associated with UTI occurrence.

Table 4.

Multivariable Analysis of Characteristics Associated Urinary Tract Infection Occurrence

	Right colectomy n = 7,615				APR n = 2,493
	OR	95% CI		p	OR	95% CI		p
Age (10 y)	1.22	1.02	1.46	0.027	1.08	0.90	1.30	0.0384
Male	0.63	0.41	0.95	0.027	0.52	0.33	0.82	0.005
ASA (3,4,5 vs. 1,2)	1.26	0.81	1.96	0.306	1.56	0.96	2.56	0.076
Pre-operative functional status as partial/total dependence	1.63	0.71	3.73	0.248	3.19	1.32	7.71	0.010
Laparoscopic	0.78	0.52	1.16	0.216	0.86	0.49	1.49	0.586
Any SSI	1.92	1.12	3.26	0.017	1.17	0.68	2.02	0.579
Pneumonia	1.65	0.77	3.56	0.197	1.61	0.59	4.44	0.355
Sepsis or septic shock	6.27	3.47	11.31	<0.001	4.14	2.15	7.94	<0.001
Pseudo R²				8.3%				6.2%

APR = abdominal perineal resection; OR = odds ratio; CI = confidence interval; ASA = American Society of Anesthesiologists Physical Status Classification; SSI = surgical site infection.

In the univariable analysis, patient death was seen more commonly in the UTI patient group for right colectomy patients only. We conducted a multivariable analysis for factors associated with death in the right colectomy patients. As shown in Table 5, after correction for other patient characteristics and significant post-operative occurrences, UTI was not statistically associated with death (odds ratio [OR] 0.55, p = 0.3346).

Table 5.

Multivariable Analysis of Characteristics Associated with 30-day Death

	Right colectomy
	OR	95% CI		p
Age (10 y)	2.00	1.53	2.60	<0.001
ASA (3,4,5 vs. 1,2)	1.94	1.06	3.57	0.032
Pre-operative functional status as partial/total dependence	3.56	1.81	7.00	<0.001
Any SSI	0.73	0.37	1.44	0.362
Pneumonia	1.68	0.78	3.62	0.185
Failure to wean ventilator within 48 h	1.11	0.48	2.57	0.809
Acute renal failure	8.26	2.60	26.25	<0.001
Sepsis + septic shock	14.85	7.85	28.06	<0.001
UTI	0.55	0.16	1.91	0.346
R²	22.8%

OR = odds ratio; CI = confidence interval; ASA = American Society of Anesthesiologists Physical Status Classification; SSI = surgical site infection; UTI = urinary tract infection.

We performed a subanalysis for both right colectomy and APR patients for the impact of timing of UTI diagnosis on re-admission. We identified a re-admission rate of 9.7% (n = 6/62) in right colectomy and APR patients having UTI diagnosed before discharge and 37.7% (n = 20/53) for patients having outpatient diagnosis of UTI (p < 0.001). Table 6 analyzes other patient characteristics and outcomes that may potentially contribute to re-admission. Surgical site infections and septic shock were also more commonly identified in the re-admitted outpatient UTI diagnosis group.

Table 6.

Demographics, co-morbidity, operative characteristics, and outcomes for patients with urinary tract infection diagnosis after discharge by re-admission occurrence

	Having UTI after discharge
	Re-admission n = 20	No re-admission n = 33	p
Age	71.52 ± 10.71	69.37 ± 14.21	0.571
Body mass index	26.24 ± 5.28	29.39 ± 6.52	0.073
Male gender	6 (30.0%)	15 (45.5%)	0.228
White	17 (85.0%)	24 (72.7%)	1.000
ASA (3,4,5 vs. 1,2)	15 (75.0%)	20 (60.6%)	0.375
Co-morbidity
Smoking history	3 (15.0%)	4 (12.1%)	1.000
Alcoholic	0 (0%)	0 (0%)	1.000
Chronic heart failure	0 (0%)	0 (0%)	1.000
COPD	1 (5.0%)	2 (6.1%)	1.000
CVA	1 (5.0%)	1 (3.0%)	1.000
TIA	0 (0%)	1 (3.0%)	1.000
Diabetes	3 (15.0%)	11 (33.3%)	0.203
Dyspnea	2 (10.0%)	1 (3.0%)	0.549
Steroid use	2 (10.0%)	1 (3.0%)	0.549
Procedure
Laparoscopic	9 (45.0%)	17 (51.5%)	0.646
Other characteristics
Pre-operative functional status as partial/total dependence	0 (0%)	3 (9.1%)	0.282
Disseminated cancer	1 (5.0%)	2 (6.1%)	1.000
Outcome
Operative time (min)	184.15 ± 87.91	190.96 ± 80.48	0.774
Any SSI	9 (45.0%)	4 (12.1%)	0.010
Wound dehiscence	0 (0%)	0 (0%)	1.000
Pneumonia	1 (5.0%)	1 (3.0%)	1.000
Failure to wean ventilator within 48 h	0 (0%)	0 (0%)	1.000
Acute renal failure	0 (0%)	0 (0%)	1.000
Progressive renal failure	0 (0%)	1 (3.0%)	1.000
Bleeding	6 (30.0%)	5 (15.2%)	0.296
Stroke/coma	0 (0%)	0 (0%)	1.000
MI/cardiac arrest	0 (0%)	0 (0%)	1.000
Sepsis + septic shock	8 (40.0%)	4 (12.1%)	0.039

UTI = urinary tract infection; ASA = American Society of Anesthesiologists Physical Status Classification; COPD = chronic obstructive pulmonary disease; CVA = cerebrovascular accident; TIA = transient ischemic attack; SSI = surgical site infection; MI = myocardial infarction.

Last, for our entire cohort of 10,108 patients, 54 (0.5%) had both a UTI and a surgical site infection. These 54 patients represent 4.7% of all patients with surgical site infection (n = 54/1,157) and 27.7% of patients with UTI (n = 54/195). For these 54 patients, 75.9% (n = 41/54) of the UTIs were diagnosed before or on the same day as the surgical site infection.

Discussion

We performed an analysis of NSQIP Participant User Files for the years 2006–2012 for a homogenous patient population with colorectal cancer undergoing right colectomy or APR assessing for pre-operative characteristics, operative details, and post-operative complications associated with UTI. By focusing our analysis on specific patient diagnoses and procedures, while correcting for inpatient and outpatient UTI occurrence and post-operative complications, we provide a clear assessment of factors associated with UTI occurrence and its impact on patient LOS, re-admission, and death.

Although a retrospective analysis cannot show causation, UTI was associated with a 59.0% increase in LOS for right colectomy (p < 0.001) and a 37.4% increase for APR patients (p < 0.001). When diagnosed in an inpatient, UTI did not correlate with re-admission for either procedure group. When diagnosed in an outpatient, however, UTI was associated with a 37.7% re-admission rate. Inpatient UTI was associated with death for right colectomy patients with colorectal cancer on univariable analysis only.

Colorectal resection is well known for its propensity for infectious complications, including UTI. These increased infection rates have made colorectal surgery a landscape for extensive study regarding the impact of infections on other traditional surgical outcomes such has LOS, re-admission, and death. Largely, UTI has been associated with death in the medical population but not in the surgical population specifically [4,5,7]. In our patients, the technically less demanding and less morbid patients with right colectomy did exhibit an increase in death for UTI patients with UTI on univariable analysis.

There is support for this in the literature where Attaluri et al. [1] showed that the morbidity profile of the procedure performed did impact UTI-related outcomes. Ultimately, after correction for other statistically and clinically significant patient characteristics and complications, UTI did not predict death in patients with right colectomy keeping in line with the majority of the colorectal literature [4,10]. Of note, many of the complications occurring more commonly in the right colectomy UTI group, as well as the increased death for right colectomy, were found to approach statistical significance in the APR group (Table 2). The larger sample size in the right colectomy group compared with the APR patients may suggest that with a larger APR cohort, many of the complications could become statistically significant.

The literature regarding re-admission in colorectal surgical procedures has been variable in investigating the relative importance of pre-operative, non-modifiable patient factors on adverse outcomes [11 –14]. One of the main aims of this study was to determine whether UTI as an individual, adverse post-operative outcome correlated with re-admission. On univariable analysis the readmission rates for UTI patients were not different and a more detailed analysis of predictors was abandoned. We did find, however, a striking re-admission rate (37.7%) for patients with UTI diagnosed as an outpatient. While the sample size of patients with outpatient UTI occurrence limited multivariable analysis, there was also an increase in surgical site infection, sepsis, and septic shock for patients re-admitted in our cohort. We were unable to determine UTI as the indication for re-admission or the timing of diagnosis, meaning while still an outpatient or on re-admission, because of limitations of our analysis and dataset.

The study of LOS in the literature on colorectal surgical procedures has investigated minimally invasive operations, the benefits of enhanced recovery protocols, and the impact of post-operative complications [15 –17]. Many of our identified LOS correlates (Table 3) confirm this previous literature where complications and patient co-morbidity are both associated with prolonged LOS [17,18]. The correlation of UTI with LOS in the colorectal literature has been shown in NSQIP-based literature [4], and our results confirm this relationship for a homogenous population and after correction for concomitant complications.

Predictors of UTI in the colorectal literature have focused on pre-operative patient characteristics where age, female gender, overall medical co-morbidity, diagnosis, and procedure performed have been associated with an increased UTI rate [1,2,8]. In our analysis, we incorporated other post-operative infections into our model allowing for assessment of the importance of events such as surgical site infection on the occurrence of UTI.

In agreement with the literature, we found age, gender, and pre-operative functional status to impact UTI occurrence. Interestingly, the impact of these variables was different between procedures as the importance of pre-operative functional status was significant in the higher morbidity APR but not lower morbidity right hemicolectomy. Unique to our analysis, the concomitant occurrence of other post-operative infectious complications, such as surgical site infection and sepsis, appeared correlated with UTI. While sepsis was correlated with UTI in both procedure groups, surgical site infection was correlated with UTI for right hemicolectomy but not APR.

Our results are subject to the limitations of the NSQIP data collection methodology and the use of a de-identified, multi-institutional, and retrospective database. Limitations particularly applicable to our results include the lack of information pertaining to the UTI cultures and the pathophysiologic etiology of sepsis and septic shock recorded in the NSQIP database. It is not known whether the organisms responsible for the surgical site infections, UTIs, and sepsis are singular or if multiple organisms and etiologies are at play. In addition, NSQIP does not collect information on duration of urinary catheter placement. It is assumed urinary catheters were in place post-operatively but it is not known whether patients had indwelling catheters pre-operatively for other medical reasons, had prolonged post-operative catheterizations, or definable catheter-associated UTIs.

In a retrospective review, no causality is confirmed for any individual infectious outcome as the reason for an observed prolonged LOS, re-admission, or death. It is of interest to the authors, however, to try to analyze the importance of UTI on outcomes of the highly morbid colorectal surgical population where other infectious complications may heavily influence patient outcomes. We believe that our multivariable analysis allows some clarification of the individual impact of UTI on surgical outcomes, as opposed to UTI being a marker for other infectious outcomes or a highly co-morbid patient population. Because we identified that 27.7% of patients having a diagnosis of post-operative UTI also had a surgical site infection documented, and that for these patients, 75.9% of the UTIs occurred first or on the same day, the relationships of concomitant UTI and surgical site infection may be an avenue of important further study.

Conclusion

After excluding deaths, outpatient UTI occurrences, and correcting for other infectious complications, our review of NSQIP data revealed that UTI is associated with increased LOS but is not correlated with re-admission or death. Outpatient occurrence of UTI after hospital discharge is associated with a dramatic re-admission rate of 37.7%. In patients having both a surgical site infection and UTI, the UTI was diagnosed on the same day or before the surgical site infection in 75.9% of patients, and this may be a hypothesis-generating finding for future study.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Attaluri

, Kiran

, Vogel

, et al. Risk factors for urinary tract infections in colorectal compared with vascular surgery: A need to review current present-on-admission policy. J Am Coll Surg, 2011; 212:356–361.

Regenbogen

, Read

, Roberts

, et al. Urinary tract infection after colon and rectal resections: More common than predicted by risk-adjustment models. J Am Coll Surg, 2011; 213:784–792.

Ingraham

, Cohen

, Bilimoria

, et al. Comparison of hospital performance in nonemergency versus emergency colorectal operations at 142 hospitals. J Am Coll Surg, 2010; 210:155–165.

Ricciardi

, Roberts

. Read

, et al. Which adverse events are associated with mortality and prolonged length of stay following colorectal surgery. J Gastrointest Surg, 2013; 17:1485–1493.

Platt

, Polk

, Murdock

, Rosner

. Mortality associated with nosocomial urinary-tract infection. N Engl J Med, 1982; 307:637–642.

Wald

, Kramer

. Nonpayment for harms resulting from medical care: Catheter-associated urinary tract infections. JAMA, 2007; 298:2782–2784.

Saint

. Clinical and economic consequences of nosocomial catheter-related bacteriuria. Am J Infect Control, 2000; 28:68–75.

Wideroff

, Xing

, Liao

, Byrn

. Crohn's disease but not diverticulitis is an independent risk factor for surgical site infections in colectomy. J Gastrointest Surg, 2014; 18:1817–1823.

American College of Surgeons National Surgical Quality Improvement Program Data User Guide. September 2011.

10.

Longo

, Virgo

, Johnson

, et al. Risk factors for morbidity and mortality after colectomy for colon cancer. Dis Colon Rectum, 2000; 43:83–91.

11.

Damie

, Cherng

, Flahive

, et al. Clinical and financial impact of hospital readmissions after colorectal resection: Predictors, outcomes, and costs. Dis Colon Rectum, 2014; 57:1421–1429.

12.

Wick

, Shore

, Hirose

, et al. Readmission rates and cost following colorectal surgery. Dis Colon Rectum, 2011; 54:147–1479.

13.

Kelly

, Iannuzzi

, Rickles

, et al. Risk factors associated with 30-day postoperative readmission in major gastrointestinal resections. J Gastrointest Surg, 2014; 18:35–43.

14.

Kerwel

, Leichtle

, Asgeirsson

, et al. Risk factors for readmission after elective colectomy: Postoperative complications are more important than patient and operative factors. Dis Colon Rectum, 2014; 57:98–104.

15.

Zhuang

, Ye

, Zhang

, et al. Enhanced recovery after surgery programs versus traditional care for colorectal surgery: A meta-analysis of randomized controlled trials. Dis Colon Rectum, 2013; 56:667–678.

16.

Lee

, Fleming

, Deeb

, et al. A laparoscopic approach reduces short-term complications and length of stay following ileocolic resection in Crohn's disease: An analysis of outcomes from the NSQIP database. Colorectal Dis, 2012; 14:572–577.

17.

Dimick

, Chen

, Taheri

, et al. Hospital costs associated with surgical complications: A report from the private-sector National Surgical Quality Improvement Program. J Am Coll Surg, 2004; 199:531–537.

18.

Faiz

, Haji

, Burns

, et al. Hospital stay amongst patients undergoing major elective colorectal surgery: Predicting prolonged stay and readmissions in NHS hospitals. Colorectal Dis, 2011; 13:816–822.