Abstract
Study Design:
Retrospective review
Objective:
Compare post-operative infection rates and 30-day outcomes in spine surgery patients with and without a pre-operative urinary tract infection (UTI).
Summary of Background Data:
There is mixed evidence regarding safety and risks when operating on spine patients with a pre-operative UTI.
Methods:
Using data from the American College of Surgeons National Surgical Quality Improvement Program, we identified all adult patients undergoing spine surgery between 2012–2017 with a pre-operative UTI. Patients with other pre-operative infections were excluded. Our primary outcome was any post-operative infection (pneumonia, sepsis, surgical site infection, and organ space infection). Our secondary outcomes included surgical site infections, non-infectious complications, return to operating room, and 30-day readmission and mortality. We used univariate, then multivariate Poisson regression models adjusted for demographics, comorbidities, laboratory values, and case details to investigate the association between pre-operative UTI status and post-operative outcomes.
Results:
270,371 patients who underwent spine surgery were analyzed. The most common procedure was laminectomy (41.9%), followed by spinal fusion (31.7%) and laminectomy/fusion (25.6%). 341 patients had a pre-operative UTI (0.14%). Patients with a pre-operative UTI were more likely to be older, female, inpatients, emergency cases, with a higher ASA score, and a longer operating time (for all, P<0.001). Patients with a pre-operative UTI had higher rates of infectious and non-infectious complications, return to OR, and unplanned readmissions (for all, P<0.001). However, there was no significant difference in mortality (0.6% vs. 0.2%, P=0.108). Even after controlling for demographics, comorbidities, labs, and case details, pre-operative UTI status was significantly associated with more post-operative infectious complications (IRR: 2.88, 95% CI: 2.25–3.70, P<0.001).
Conclusion:
Pre-operative UTI status is significantly associated with post-operative infections and worse 30-day outcomes. Spine surgeons should consider delaying or cancelling surgery in patients with a UTI until the infection has cleared to reduce adverse outcomes.
Keywords: urinary tract infection, spine surgery, outcomes, infection, post-operative complications, NSQIP, surgical safety, laminectomy, spinal fusion, quality improvement, surgical risk
INTRODUCTION
With over a million cases annually, spine surgery is one of the most common procedures in the United States, and the rate of spine surgery continues to increase substantially. A recent study examining elective lumbar fusion cases documented a 62.3% increase in utilization from 122,679 cases in 2004 to 199,140 cases in 2015, with the greatest growth seen among the elderly (age ≥65), where the volume increased almost 140%.1 We observed similar trends in increased utilization among spinal deformity,2 and degenerative cervical spine operations.3 Among spine surgeons and patients, post-operative infections, particularly urinary tract infections (UTIs), persist as a major source of concern.4
UTIs are among the most common type of community- and hospital-acquired bacterial infections,5 with more than half of all women experiencing at least one symptomatic UTI by the age of 35.6 Population-adjusted incidence rates of these infections have grown 52% between 1998 and 2011,7 and are only expected to grow as life expectancy lengthens and antimicrobial-resistant uropathogens become more prevalent. In the realm of spine surgery, Bohl et al. reported the incidence of a post-operative UTI among patients undergoing posterior lumbar fusion at 1.8%.8 They also found that post-operative UTI increased the risk for systemic sepsis 14-fold and readmission 6-fold. In addition to the higher morbidity and mortality, longer length of stay, and higher re-operation rates,9 post-operative UTIs place significant financial burden on hospitals, as the Centers for Medicare and Medicaid Services no longer reimburses for healthcare-acquired UTIs.10
Although the consequences of a post-operative UTI have been extensively studied, the impact of a pre-operative UTI on spine surgery outcomes has not been well characterized. In fact, there is mixed evidence in the surgical literature debating the safety of operating on patients with a pre-operative UTI. While a study of total joint arthroplasty patients found no association between pre-operative UTI and deep joint infection (Odds Ratio: 0.34; 95% CI: 0.09–1.36; P=0.127),11 an analysis of elective general surgery patients found higher risk of post-operative infectious complications (Odds Ratio: 1.52; 95% CI: 1.00–2.30; P=0.050) for the pre-operative UTI group.12 The goal of our study was to understand the association between pre-operative UTIs and 30-day outcomes in spine surgery patients using a large, national database.
MATERIALS AND METHODS
Data Source and Study Population
This study is a secondary cross-sectional analysis of the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP), a national validated, risk-adjusted, clinical database of surgical procedures from over 700 hospital across the United States.13 Certified clinical reviewers at each participating hospital collect data on over 270 variables that include patient demographics, pre-operative risk factors, laboratory values, intraoperative variables, and 30-day post-operative outcomes. This database is particularly well suited for our study because starting in 2011, ACS-NSQIP began to collect a “present at time of surgery” (PATOS) modifier to identify patients who enter the operating room (OR) with a pre-existing infection. Patients with a pre-operative UTI must have met both of the following criteria to be assigned the PATOS modifier: 1) Preoperative evidence highly suggestive of a UTI (symptomatic or asymptomatic) with or without treatment and 2) Patient has a post-operative symptomatic UTI documented.13
The inclusion criteria for our study are all adult patients (≥18 years old) who underwent spine surgery between 2012 and 2017 in ACS-NSQIP. These patients were identified using the Clinical Classifications Software from the US Agency for Healthcare Research and Quality, which groups Current Procedural Terminology (CPT) codes into 244 clinically meaningful categories.14 The following CPT-CCS codes were used to select our patient population: 3 – laminectomy [including partial facetectomy, foraminotomy, and discectomy; open or minimally invasive technique], 5 – insertion of catheter or spinal stimulator and injection into spinal canal, and 158 – spinal fusion [including instrumentation; open or minimally invasive technique]. Patients with other suspected pre-operative infections (wound infection, pneumonia, and sepsis) were excluded. This study was assigned exempt review status by the Yale University institutional review board, as ACS-NSQIP data is publicly available and de-identified.
Outcomes and Statistical Analysis
The primary study outcome was the rate of post-operative infectious complications, which include deep and superficial surgical site infection (SSI), pneumonia, sepsis, and deep organ space infection. We also examined an array of secondary outcome variables: surgical site infections, non-infectious complications (acute renal failure, cardiac arrest, stroke, deep vein thrombosis, myocardial infarction, progressive renal insufficiency, pulmonary embolism, transfusion, unplanned intubation, ventilator support for > 48 hours, and wound dehiscence), return to OR, 30-day readmission, and 30-day mortality.
We first used univariate Poisson regression models to examine the unadjusted associations between the following demographic and peri-operative factors and each of our outcome variables: age, sex, race, body mass index (BMI), American Society of Anesthesiologists (ASA) score, inpatient status, medical comorbidities, pre-operative labs (hemoglobin and albumin) less than 90 days before surgery, surgeon specialty (neurosurgery or orthopedic surgery), emergency case status, length of operation, and type of spine surgery. Continuous variables were categorized at clinically meaningful cut points: age (<65, ≥65), BMI (underweight: <18.5 kg/m2, normal weight: 18.5–24.9 kg/m2, overweight: 25–29.9 kg/m2, obese: 30–34.9 kg/m2, morbidly obese: ≥35 kg/m2), anemia (hemoglobin [Hgb] <10 g/dL), hypoalbuminemia (albumin <3.5 g/dL).15 We used a Poisson regression to account for the fact that patients could have multiple post-operative complications. We report incidence rate ratios (IRRs), which show the ratio of event rates.
Variables that were statistically significant in the univariate regressions (P <0.05) were then included in our multivariate models. We used the medical comorbidities (bleeding disorders, CHF, COPD, diabetes, dialysis, functional dependence, hypertension, smoking, chronic steroid use, transfusion <72 hours pre-op, and >10% weight loss in last 6 months) as adjustment variables in this final multivariate model. The data met the assumptions of a Poisson regression model (goodness-of-fit χ2=53,531; P=1.000). We then created multivariable regression models with the same adjustment and predictor variables with our secondary outcome measures.
All data aggregation and statistical analyses were performed using Stata, version 14 (Stata Corp., College Station, TX). There were several demographic variables and laboratory data with missing values in our database. As excluding patients with missing fields and performing a complete case analysis may limit our analysis and bias the results,16 we used multiple imputation to impute variables with missing data. We created an imputation model that included all predictor and outcome variables and produced 10 imputed data sets. The mi estimate command suite in Stata was then used to adjust the regression coefficients and standard errors for the variability among the 10 imputed data sets according to Rubin’s combination rules.17 Statistical significance was assessed at P<0.05.
RESULTS
A total of 270,371 patients who underwent spine surgery in ACS-NSQIP from 2012–2017 met our inclusion/exclusion criteria and were included in the analysis. The median age of the patients was 58 years old and 47.7% were female (Table 1). The most frequent procedure performed was laminectomy (n=113,235; 41.9%), followed by spinal fusion (n=35,721; 31.7%) and combined laminectomy and fusion (n=69,198; 25.6%). The overall post-operative infection rate was 2.4% (n=6,604), with surgical site infections (1.3%) and pneumonia (0.7%) being the most common infectious complications.
Table 1.
Patient demographics and peri-operative characteristics
| Total | Pre-op UTI | No Pre-op UTI | P-value | |
|---|---|---|---|---|
| N | 270,371 | 341 | 270,030 | |
| Age, median (IQR), years | 58 (47–68) | 69 (58–77) | 58 (47–68) | <0.001 |
| ≥65 | 92,571 (34.2%) | 222 (65.1%) | 92,349 (34.2%) | <0.001 |
| Female sex | 129,009 (47.7%) | 249 (73.0%) | 128,760 (47.7%) | <0.001 |
| Race | ||||
| White | 237,036 (87.7%) | 298 (87.3%) | 236,738 (87.7%) | 0.975 |
| Black | 24,695 (9.1%) | 32 (9.4%) | 24,663 (9.1%) | |
| Other | 8,640 (3.2%) | 11 (3.4%) | 8,629 (3.2%) | |
| BMI, kg/m2 | ||||
| Underweight: <18.5 | 1,541 (0.6%) | 6 (1.6%) | 1,536 (0.6%) | 0.884 |
| Overweight: 25–29.9 | 51,659 (19.1%) | 78 (22.8%) | 51,581 (19.1%) | |
| Normal: 18.5–24.9 | 90,632 (33.5%) | 89 (26.3%) | 90,542 (33.5%) | |
| Obese: 30–34.9 | 70,038 (25.9%) | 90 (26.4%) | 69,948 (25.9%) | |
| Morbidly obese: ≥35 | 56,501 (20.9%) | 78 (22.9%) | 56,423 (20.9%) | |
| ASA Class | ||||
| 1 or 2 | 146,654 (54.2%) | 89 (26.1%) | 146,565 (54.3%) | <0.001 |
| 3 | 116,467 (43.1%) | 213 (62.5%) | 116,254 (43.1%) | |
| 4 or 5 | 7,250 (2.7%) | 39 (11.4%) | 7,211 (2.7%) | |
| Inpatient | 192,201 (71.1%) | 311 (91.2%) | 191,890 (71.1%) | <0.001 |
| Comorbidities | ||||
| Bleeding disorders | 4,211 (1.6%) | 12 (3.5%) | 4,199 (1.6%) | 0.005 |
| CHF | 806 (0.3%) | 6 (1.8%) | 800 (0.3%) | <0.001 |
| COPD | 11,189 (4.1%) | 27 (7.9%) | 11,162 (4.1%) | 0.001 |
| Diabetes | 45,371 (16.8%) | 92 (27.0%) | 44,279 (16.8%) | <0.001 |
| Dialysis | 741 (0.3%) | 2 (0.6%) | 739 (0.3%) | 0.281 |
| Functional dependence | 6,035 (2.2%) | 55 (16.2%) | 5,980 (2.2%) | <0.001 |
| Hypertension | 134,589 (49.8%) | 226 (66.3%) | 134,363 (49.8%) | <0.001 |
| Smoking | 60,003 (22.2%) | 49 (14.4%) | 59,954 (22.2%) | 0.001 |
| Steroid use | 10,785 (4.0%) | 37 (10.9%) | 10,748 (4.0%) | <0.001 |
| Transfusion 72 hrs prior to surgery | 458 (0.2%) | 4 (1.2%) | 454 (0.2%) | <0.001 |
| >10% weight loss in last 6 months | 870 (0.3%) | 8 (2.3%) | 862 (0.3%) | <0.001 |
| Pre-op Labs | ||||
| Hemoglobin <10 g/dL | 32,640 (12.1%) | 119 (35.0%) | 32,521 (12.0%) | <0.001 |
| Albumin <3.5 g/dL | 22,237 (8.2%) | 99 (29.0%) | 22,138 (8.2%) | <0.001 |
| Surgical specialty | ||||
| Orthopedic surgery | 86,168 (31.9%) | 96 (28.2%) | 86,072 (31.9%) | 0.141 |
| Neurosurgery | 184,203 (68.1%) | 245 (71.8%) | 183,958 (68.1%) | |
| Emergency case | 4,674 (1.7%) | 29 (8.5%) | 4,645 (1.7%) | <0.001 |
| Operative time | ||||
| <4hrs | 232,116 (85.9%) | 258 (75.7%) | 231,858 (85.9%) | <0.001 |
| ≥4hrs | 38,255 (14.1%) | 83 (24.3%) | 38,172 (14.1%) | |
| Spine surgery type | ||||
| Laminectomy | 113,235 (41.9%) | 131 (38.4%) | 113,104 (41.9%) | 0.009 |
| Spinal fusion | 85,721 (31.7%) | 95 (27.9%) | 85,626 (31.7%) | |
| Laminectomy & Fusion | 69,198 (25.6%) | 114 (33.4%) | 69,084 (25.6%) | |
| Spinal stimulator | 2,217 (0.8%) | 1 (0.3%) | 2,216 (0.8%) |
Abbreviations: UTI = Urinary Tract Infection, IQR= Interquartile Range, BMI = Body Mass Index, ASA = American Society of Anesthesiologists, CHF = Congestive Heart Failure, COPD = Chronic Obstructive Pulmonary Disease
Of the patients included in our study, 341 patients had a pre-operative UTI (0.14%). Patients with a pre-operative UTI were more likely to be older, female, and inpatients with a higher ASA score than those without a UTI (for all, P<0.001; Table 1). Furthermore, pre-operative UTI patients presented with more medical comorbidities (e.g., diabetes, hypertension, and COPD), and poorer pre-operative laboratory values (for all, P<0.001). Operations on these patients were more likely to be an emergency and had a longer operating time (both P<0.001).
Patients with a pre-operative UTI had a six-fold higher rate of infectious complications (14.7% vs. 2.4%; P<0.001), and were more likely to have multiple infections than the uninfected patients (26.0% vs. 16.5%; P<0.001; Table 2). Pre-operative UTI patients were more likely to experience other poor outcome measures, including higher rates of non-infectious complications, 30-day readmissions, and unplanned returns to the OR (for all, P<0.001). However, there were no statistically significant differences in mortality (0.6% vs. 0.2%, P=0.108).
Table 2.
Post-operative complications and 30-day outcomes
| Total | Pre-op UTI | No Pre-op UTI | P-value | |
|---|---|---|---|---|
| N | 270,371 | 341 | 270,030 | |
| Infectious complication | 6,604 (2.4%) | 50 (14.7%) | 6,554 (2.4%) | <0.001 |
| Pneumonia | 1,933 (0.7%) | 14 (4.1%) | 1,919 (0.7%) | <0.001 |
| Sepsis | 1,280 (0.5%) | 20 (5.9%) | 1,260 (0.5%) | <0.001 |
| Septic Shock | 379 (0.1%) | 9 (2.6%) | 370 (0.1%) | <0.001 |
| Surgical Site Infection (SSI) | 3,464 (1.3%) | 16 (4.7%) | 3,448 (1.3%) | <0.001 |
| Superficial SSI | 2,201 (0.8%) | 11 (3.2%) | 2,190 (0.8%) | <0.001 |
| Deep Incisional SSI | 1,282 (0.5%) | 5 (1.5%) | 1,277 (0.5%) | <0.001 |
| Organ Space Infection | 700 (0.3%) | 5 (1.5%) | 695 (0.3%) | 0.001 |
| No. of infectious complications | ||||
| 1 | 5,510 (83.4%) | 37 (74.0%) | 5,473 (83.5%) | <0.001 |
| 2 | 1,020 (15.5%) | 12 (24.0%) | 1,008 (15.4%) | |
| 3 | 71 (1.1%) | 1 (2.0%) | 70 (1.1%) | |
| 4 | 3 (0.1%) | 0 (0.0%) | 3 (0.0%) | |
| Non-infectious complication | 18,662 (6.9%) | 74 (21.7%) | 18,588 (6.9%) | <0.001 |
| Acute renal failure | 217 (0.1%) | 3 (0.9%) | 214 (0.1%) | <0.001 |
| Cardiac arrest | 409 (0.2%) | 1 (0.3%) | 408 (0.2%) | 0.508 |
| CVA | 343 (0.1%) | 5 (1.5%) | 338 (0.1%) | <0.001 |
| DVT | 1,611 (0.6%) | 11 (3.2%) | 1,600 (0.6%) | <0.001 |
| Myocardial Infarction | 656 (0.2%) | 6 (1.8%) | 650 (0.2%) | <0.001 |
| Progressive Renal Insufficiency | 253 (0.1%) | 2 (0.6%) | 251 (0.1%) | 0.010 |
| Pulmonary Embolism | 1,117 (0.4%) | 5 (1.5%) | 1,112 (0.4%) | 0.005 |
| Transfusion | 14,670 (5.4%) | 53 (15.5%) | 14,617 (5.4%) | <0.001 |
| Unplanned Intubation | 1,152 (0.4%) | 4 (1.2%) | 1,148 (0.4%) | 0.042 |
| Ventilator >48 hours | 1,057 (0.4%) | 11 (3.2%) | 1,046 (0.4%) | <0.001 |
| Wound Dehiscence | 639 (0.2%) | 4 (1.2%) | 635 (0.2%) | 0.001 |
| No. of non-infectious complications | ||||
| 1 | 16,372 (87.7%) | 56 (75.7%) | 16316 (87.8%) | <0.001 |
| 2 | 1,686 (9.0%) | 13 (17.6%) | 1673 (9.0%) | |
| 3 | 425 (2.3%) | 2 (2.7%) | 423 (2.3%) | |
| 4 | 137 (0.7%) | 3 (4.1%) | 134 (0.7%) | |
| 5 | 38 (0.2%) | 0 (0.0%) | 38 (0.2%) | |
| 6 | 4 (0.0%) | 0 (0.0%) | 4 (0.0%) | |
| Return to OR | 7,398 (2.7%) | 25 (7.3%) | 7,373 (2.7%) | <0.001 |
| 30-day readmission | 12,526 (4.6%) | 67 (19.8%) | 12,459 (4.6%) | <0.001 |
| 30-day mortality | 509 (0.2%) | 2 (0.6%) | 507 (0.2%) | 0.108 |
Abbreviations: UTI = Urinary Tract Infection, SSI = Surgical Site Infection, CVA = Cerebrovascular Accident, DVT = Deep Vein Thrombosis, OR = operating room
Univariate Poisson regression analyses showed that the following factors were associated with higher rates of infectious complications: older age (≥65 years), female sex, Black race, normal BMI, higher ASA class, inpatient status, anemia, hypoalbuminemia, surgeon specialty (neurosurgery), emergency case status, and operative time of greater than 4 hours (for all, P<0.001; Table 3). Furthermore, all medical comorbidities we examined, most notably dialysis and pre-operative transfusions, were significantly associated with post-operative infections (for all, P<0.001). Compared to laminectomies, spinal fusions (IRR: 1.19; 95% CI: 1.12–1.26) and combined laminectomy and fusion (IRR: 1.84; 95% CI: 1.74–1.94) had higher incidence rates of infections (both P<0.001). The largest predictor of a post-operative infectious complication among our univariate models was having a pre-operative UTI, which increased the incidence by a factor of 6.57 (95% CI: 5.14–8.41; P<0.001).
Table 3.
Univariate Poisson regressions modeling post-operative infectious complications
| IRR (95% CI) | P-value | |
|---|---|---|
| Pre-operative UTI | 6.57 (5.14–8.41) | <0.001 |
| Age, years | ||
| <65 | 1 [ref.] | |
| ≥65 | 1.59 (1.52–1.66) | <0.001 |
| Female sex | 1.05 (1.00–1.09) | 0.043 |
| Race | ||
| White | 1 [ref.] | |
| Black | 1.29 (1.20–1.39) | <0.001 |
| Other | 1.13 (0.98–1.30) | 0.089 |
| BMI, kg/m2 | ||
| Normal: 18.5–24.9 | 1 [ref.] | |
| Underweight: <18.5 | 0.58 (0.45–0.75) | <0.001 |
| Overweight: 25–29.9 | 0.52 (0.40–0.67) | <0.001 |
| Obese: 30–34.9 | 0.58 (0.46–0.75) | <0.001 |
| Morbidly obese: ≥35 | 0.91 (0.71–1.17) | 0.469 |
| ASA Class | ||
| 1 or 2 | 1 [ref.] | |
| 3 | 2.56 (2.44–2.69) | <0.001 |
| 4 or 5 | 5.90 (5.42–6.42) | <0.001 |
| Inpatient | 2.74 (2.56–2.93) | <0.001 |
| Comorbidities | ||
| Bleeding disorders | 2.76 (2.47–3.08) | <0.001 |
| CHF | 3.39 (2.71–4.24) | <0.001 |
| COPD | 2.51 (2.33–2.70) | <0.001 |
| Diabetes | 1.77 (1.68–1.86) | <0.001 |
| Dialysis | 4.31 (3.50–5.30) | <0.001 |
| Functional dependence | 3.66 (3.37–3.98) | <0.001 |
| Hypertension | 1.69 (1.62–1.77) | <0.001 |
| Smoking | 1.15 (1.10–1.21) | <0.001 |
| Steroid use | 1.96 (1.80–2.13) | <0.001 |
| Transfusion 72 hrs prior to surgery | 4.66 (3.62–6.00) | <0.001 |
| >10% weight loss in last 6 months | 3.79 (3.09–4.65) | <0.001 |
| Pre-op Labs | ||
| Hemoglobin <10 g/dL | 2.27 (2.15–2.40) | <0.001 |
| Albumin <3.5 g/dL | 2.62 (2.44–2.81) | <0.001 |
| Surgical specialty | ||
| Orthopedic surgery | 1 [ref.] | |
| Neurosurgery | 1.08 (1.03–1.14) | 0.001 |
| Emergency case | 2.81 (2.53–3.12) | <0.001 |
| Operative time ≥4 hrs | 2.47 (2.35–2.59) | <0.001 |
| Spine surgery type | ||
| Laminectomy | 1 [ref.] | |
| Spinal fusion | 1.19 (1.12–1.26) | <0.001 |
| Laminectomy & Fusion | 1.84 (1.74–1.94) | <0.001 |
| Spinal stimulator | 0.66 (0.47–0.93) | 0.017 |
Abbreviations: IRR= Incidence Rate Ratio, CI = Confidence Interval, UTI = Urinary Tract Infection, BMI = Body Mass Index, ASA = American Society of Anesthesiologists, CHF = Congestive Heart Failure, COPD = Chronic Obstructive Pulmonary Disease
To control for the differences in baseline clinical characteristics, we then created a multivariate Poisson regression modeling post-operative infections with statistically significant variables from the univariate regressions (Table 4). Our multivariate model confirmed many of the associations found in the univariate analyses: older age, higher ASA class, inpatient status, anemia, bleeding disorders, COPD, diabetes, dialysis, functional dependence, hypertension, smoking, steroid use, hypoalbuminemia, pre-op transfusion, severe weight loss, emergency case status, and longer operative time were predictive of higher infection rates (for all, P<0.001). Differences among sex, race, specialties, surgery types were no longer significant. Even after controlling for patient demographics and medical comorbidities, pre-operative UTI status was significantly associated with more post-operative infectious complications (IRR: 2.88, 95% CI: 2.25–3.70, P<0.001).
Table 4.
Adjusted multivariate Poisson regression modeling post-operative infectious complications
| IRR (95% CI) | P-value | |
|---|---|---|
| Pre-operative UTI | 2.88 (2.25–3.70) | <0.001 |
| Age, years | ||
| <65 | 1 [ref.] | |
| ≥65 | 1.15 (1.10–1.21) | <0.001 |
| Female sex | 1.03 (0.98–1.08) | 0.208 |
| Race | ||
| White | 1 [ref.] | |
| Black | 1.02 (0.94–1.10) | 0.704 |
| Other | 1.11 (0.96–1.28) | 0.144 |
| BMI, kg/m2 | ||
| Normal: 18.5–24.9 | 1 [ref.] | |
| Underweight: <18.5 | 1.06 (0.82–1.37) | 0.657 |
| Overweight: 25–29.9 | 0.95 (0.89–1.02) | 0.152 |
| Obese: 30–34.9 | 1.01 (0.94–1.08) | 0.838 |
| Morbidly obese: ≥35 | 1.39 (1.29–1.49) | <0.001 |
| ASA Class | ||
| 1 or 2 | 1 [ref.] | |
| 3 | 1.64 (1.55–1.73) | <0.001 |
| 4 or 5 | 2.20 (1.99–2.43) | <0.001 |
| Inpatient | 1.84 (1.71–1.98) | <0.001 |
| Pre-op Labs | ||
| Hemoglobin <10 g/dL | 1.38 (1.29–1.47) | <0.001 |
| Albumin <3.5 g/dL | 1.49 (1.37–1.62) | <0.001 |
| Surgical specialty | ||
| Orthopedic surgery | 1 [ref.] | |
| Neurosurgery | 1.05 (1.00–1.10) | 0.067 |
| Emergency case | 1.62 (1.45–1.81) | <0.001 |
| Operative time ≥4 hrs | 1.82 (1.73–1.92) | <0.001 |
| Spine surgery type | ||
| Laminectomy | 1 [ref.] | |
| Spinal fusion | 0.86 (0.81–0.92) | <0.001 |
| Laminectomy & Fusion | 1.07 (1.01–1.13) | 0.032 |
| Spinal stimulator | 0.69 (0.49–0.97) | 0.034 |
Abbreviations: IRR= Incidence Rate Ratio, CI = Confidence Interval, UTI = Urinary Tract Infection, BMI = Body Mass Index, ASA = American Society of Anesthesiologists
We continued to examine the relationships between the presence of a pre-operative UTI and our secondary outcome variables using a multivariate Poisson model with the same variables. Full models can be found in our Supplementary Table 1. A forest plot of the multivariable model results illustrates that pre-operative UTI increases a risk of surgical site infections (IRR: 2.22, 95% CI: 1.35–3.64), non-infectious complications (IRR: 1.36, 95% CI: 1.12–1.66), return to OR (IRR: 1.52, 95% CI: 1.02–2.25), and 30-day readmission (IRR: 2.32, 95% CI: 1.82–2.96) (for all, P<0.05; Figure 1). There was no statistically significant association between the presence of a pre-operative UTI and 30-day mortality (IRR: 0.70, 95% CI: 0.17–2.83; P=0.618).
Figure 1.
Relationships between pre-operative UTI and 30-day postoperative mortality and morbidity in spine surgery patients: Forest plot of adjusted multivariable Poisson regression results.
DISCUSSION
Post-operative infections are a devastating complication that increases length of stay, hospital costs, risk of chronic pain, and mortality. The field of spine surgery presents a particularly large opportunity for controlling infections due to its large case volume and high expenditure. Estimated at approximately $24,000, the cost of a surgical site infection following a neurosurgical or orthopedic procedure is higher than that of any other surgical specialty.18 Despite surgeon awareness of this issue and multidisciplinary efforts to develop preventative bundles with aseptic techniques and antimicrobial prophylaxis,19 post-operative infections remain a problem.
In this study, we analyzed 270,371 patients who underwent spine surgery in ACS-NSQIP between 2012 and 2017 and observed that patients with a pre-operative UTI have a six-fold higher rate of post-operative infectious complications. A major strength of our study is the granularity and depth of the database, which allowed us to control for over a dozen different medical comorbidities across organ systems. Even after multivariate adjustment for the differences in patient demographics, comorbidities, laboratory values, and case details, UTI remained a strong, independent predictor of a post-operative infection. Additionally, we found statistically significant associations between pre-operative UTI status and higher rates of surgical site infections, non-infectious complications, return to OR, and 30-day readmission. Pre-operative UTI is a modifiable risk factor of infection, unlike other predisposing factors such as age, diabetes, and hypertension. We did not observe a statistically significant difference in 30-day mortality, likely due to the scarcity of the event given that most spine cases performed are elective procedures. These relationships have not been clearly demonstrated in spine surgery previously, aside from a handful of single-institution, small case series studies.
We also found several other independent associations with post-operative infections from our multivariate model: older age, higher ASA class, inpatient status, anemia, hypoalbuminemia, emergency case status, and longer operative time were predictive of higher infection rates. Many of these factors were previously identified as risk factors for a post-operative infection.20 Interestingly, after multivariate adjustment, we found that spinal fusion or combined laminectomy and fusion did not have higher infectious complication rates compared to laminectomy. Previous studies in the literature have quoted infection rates of 1% for simple discectomy or laminectomy, 2–5% for non-instrumented spinal fusions, and 6–18% for instrumented spinal fusions.21–23 Results from our multivariate model suggest that other variables (e.g., increased procedure time for fusion and instrumentation cases) account for the higher infection rates rather than the performance of a fusion or instrumentation.
Based on the ACS-NSQIP definition of UTI present at the time of surgery, our study population included a combination of those who presented with symptomatic UTI and asymptomatic bacteriuria (ASB). For symptomatic UTIs, studies have postulated hematogenous spread of uropathogens,24 distant seeding of bacteria from the urinary tract,25 and UTI-associated systemic immunosuppression26 as mechanisms in which pre-operative UTI status adversely affects post-operative infection rates. On the other hand, there is mixed evidence if such relationship persists between ASB and postoperative complications. A multi-center cohort study of 2,497 total hip or total knee arthroplasty patients reported a three-fold increase in the risk of prosthetic joint infection among those with an ASB.27 Gallegos Salazar et al. recently examined 68,000 patients who underwent cardiac, orthopedic, or vascular surgery in the Veterans Affairs health care system to study this association. After controlling for age, sex, ASA class, smoking, race, and diabetes, they found no statistically significant differences in rates of SSI in patients with or without ASB (Odds Ratio: 1.58; 95% CI: 0.93–2.70; P=0.08).28 The study authors concluded that there is no value in obtaining urine cultures or providing prophylactic antibiotic treatment to sterilize the urine in the peri-operative period. A major limitation of this study is the limited generalizability of their findings due to its study cohort which was comprised of 96% male veterans, adjustment for few medical comorbidities, and exclusion of spine procedures. Based on this limited evidence, the 2019 clinical practice guidelines from the Infectious Diseases Society of America (IDSA) cautiously recommends against screening for or treating ASB in patients undergoing elective, non-urological surgery.29 The society claims that concerns over increased costs, antimicrobial resistance and stewardship, and adverse drug effects of screening and treating for ASB outweigh the potentially increased risks of ASB on post-operative outcomes. Future well-controlled, prospective clinical trials evaluating screening and treatment of ASB in patients, especially considering the impact on spinal implants, are necessary in the current absence of high-quality evidence.
A previous single-institution, retrospective review investigated the relationship between pre-operative ASB and post-operative infection specifically in spine surgery. In this study of 355 South Korean elderly women, study authors initially found no increased risk of infection for those with ASB.30 However, when they added Foley catheterization as a variable in their multivariate analysis, pre-operative ASB patients who received a urinary catheter were at a statistically significantly greater risk for a spine infection. The vast majority of patients received an indwelling catheter, which was removed two days post-operatively. The authors conclude that in patients with ASB, the use of a catheter predisposes them to developing a UTI. Potential mechanisms for this include damage and inflammation of urethral or bladder tissue upon catheter insertion, and creation of biofilms, which foster colonization in a matter of hours.31 In a study of total joint arthroplasty cases, David and Vrahas found similar increased risk of post-operative infections in asymptomatic pre-operative UTI with Foley catheterization.32 Given that the overwhelming majority of spine surgery patients receive an indwelling catheter, the interaction between ASB and catheterization may be the causative link to higher risk of post-operative infection.
Limitations
This study has several limitations. First, it is limited by the retrospective and observational design. In our analysis of 270,371 patients over the five most recently available years of ACS-NSQIP, there were only 341 patients who had a pre-operative UTI. Given the rarity of a pre-operative UTI, a large randomized clinical trial to achieve statistical power would be impractical. Second, the database is unable to provide more granular details regarding the patient’s pre-operative UTI status. We unfortunately do not have any data on urine cultures to match organisms between the pre-operative UTI and other post-operative infections, whether the patient had begun prophylactic antibiotic treatment, evidence for suspicion of a UTI, or documentation of pre-operative UTI symptoms. The ACS-NSQIP also lacks information on the discontinuation or duration of peri-operative urinary catheterization, which limits our ability to analyze the role of catheterization in post-operative infection rates in our data set. Although we thoughtfully selected clinically relevant comorbidities and known risk factors to adjust in our multivariate analysis, there may be factors, such as hospital and individual physician factors, associated with post-operative infections that are not captured within ACS-NSQIP. Lastly, our outcomes were limited to 30 days after surgery, and unable to capture the long-term effects.
Supplementary Material
Acknowledgments
The manuscript submitted does not contain information about medical device(s)/drug(s). The National Heart, Lung and Blood Institute of the National Institutes of Health under Award Number T35HL007649 funds were received in support of this work. No relevant financial activities outside the submitted work.
Abbreviations:
- UTI
Urinary Tract Infection
- ACS-NSQIP
American College of Surgeons-National Surgical Quality Improvement Program
- BMI
Body Mass Index
- CHF
Congestive Heart Failure
- COPD
Chronic Obstructive Pulmonary Disease
- ASA
American Society of Anesthesiologists
- CPT
Current Procedural Terminology
- IRR
Incidence Rate Ratio
- OR
Operating Room
- CVA
Cerebrovascular Accident
- IQR
Interquartile Range
- ASB
Asymptomatic Bacteriuria
- SSI
Surgical Site Infection
REFERENCES
- 1.Martin BI, Mirza SK, Spina N, et al. Trends in Lumbar Fusion Procedure Rates and Associated Hospital Costs for Degenerative Spinal Diseases in the United States, 2004 to 2015. Spine (Phila Pa 1976) 2019;44:369–76. [DOI] [PubMed] [Google Scholar]
- 2.Zygourakis CC, Liu CY, Keefe M, et al. Analysis of National Rates, Cost, and Sources of Cost Variation in Adult Spinal Deformity. Neurosurgery 2018;82:378–87. [DOI] [PubMed] [Google Scholar]
- 3.Patil PG, Turner DA, Pietrobon R. National trends in surgical procedures for degenerative cervical spine disease: 1990–2000. Neurosurgery 2005;57:753–8; discussion −8. [PubMed] [Google Scholar]
- 4.Deyo RA, Mirza SK, Martin BI, et al. Trends, major medical complications, and charges associated with surgery for lumbar spinal stenosis in older adults. JAMA 2010;303:1259–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Magill SS, Edwards JR, Bamberg W, et al. Multistate point-prevalence survey of health care-associated infections. N Engl J Med 2014;370:1198–208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Foxman B, Barlow R, D’Arcy H, et al. Urinary tract infection: self-reported incidence and associated costs. Ann Epidemiol 2000;10:509–15. [DOI] [PubMed] [Google Scholar]
- 7.Simmering JE, Tang F, Cavanaugh JE, et al. The Increase in Hospitalizations for Urinary Tract Infections and the Associated Costs in the United States, 1998–2011. Open Forum Infect Dis 2017;4:281. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bohl DD, Ahn J, Tabaraee E, et al. Urinary Tract Infection Following Posterior Lumbar Fusion Procedures: An American College of Surgeons National Surgical Quality Improvement Program Study. Spine (Phila Pa 1976) 2015;40:1785–91. [DOI] [PubMed] [Google Scholar]
- 9.Platt R, Polk BF, Murdock B, et al. Mortality associated with nosocomial urinary-tract infection. N Engl J Med 1982;307:637–42. [DOI] [PubMed] [Google Scholar]
- 10.Saint S, Meddings JA, Calfee D, et al. Catheter-associated urinary tract infection and the Medicare rule changes. Ann Intern Med 2009;150:877–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Koulouvaris P, Sculco P, Finerty E, et al. Relationship between perioperative urinary tract infection and deep infection after joint arthroplasty. Clin Orthop Relat Res 2009;467:1859–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Pokrzywa CJ, Papageorge CM, Kennedy GD. Preoperative urinary tract infection increases postoperative morbidity. J Surg Res 2016;205:213–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ko CY, Hall BL, Hart AJ, et al. The American College of Surgeons National Surgical Quality Improvement Program: achieving better and safer surgery. Jt Comm J Qual Patient Saf 2015;41:199–204. [DOI] [PubMed] [Google Scholar]
- 14.CCS Services and Procedures. Healthcare Cost and Utilization Project (HCUP). May 2018. Agency for Healthcare Research and Quality, Rockville, MD: https://www.hcup-us.ahrq.gov/toolssoftware/ccs_svcsproc/ccssvcproc.jsp. Accessed September 10, 2019. [Google Scholar]
- 15.King JT Jr, Gordon AJ, Perkal MF, et al. Disparities in rates of spine surgery for degenerative spine disease between HIV-infected and uninfected veterans. Spine (Phila Pa 1976) 2012;37:612–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Glance LG, Osler TM, Mukamel DB, et al. Impact of statistical approaches for handling missing data on trauma center quality. Ann Surg 2009;249:143–8. [DOI] [PubMed] [Google Scholar]
- 17.Rubin DB. Multiple imputation for nonresponse in surveys. Hoboken, N.J. Wiley-Interscience, 2004. [Google Scholar]
- 18.Schweizer ML, Cullen JJ, Perencevich EN, et al. Costs Associated with Surgical Site Infections in Veterans Affairs Hospitals. JAMA Surg 2014;149:575–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Featherall J, Miller JA, Bennett EE, et al. Implementation of an Infection Prevention Bundle to Reduce Surgical Site Infections and Cost Following Spine Surgery. JAMA Surg 2016;151:988–90. [DOI] [PubMed] [Google Scholar]
- 20.Nasser R, Kosty JA, Shah S, et al. Risk Factors and Prevention of Surgical Site Infections Following Spinal Procedures. Global Spine J 2018;8:44S–8S. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Schimmel JJ, Horsting PP, de Kleuver M, et al. Risk factors for deep surgical site infections after spinal fusion. Eur Spine J 2010;19:1711–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Beiner JM, Grauer J, Kwon BK, et al. Postoperative wound infections of the spine. Neurosurg Focus 2003;15:E14. [DOI] [PubMed] [Google Scholar]
- 23.Dowdell J, Brochin R, Kim J, et al. Postoperative Spine Infection: Diagnosis and Management. Global Spine J 2018;8:37S–43S. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Stinchfield FE, Bigliani LU, Neu HC, et al. Late hematogenous infection of total joint replacement. J Bone Joint Surg Am 1980;62:1345–50. [PubMed] [Google Scholar]
- 25.D’Ambrosia RD, Shoji H, Heater R. Secondarily infected total joint replacements by hematogenous spread. J Bone Joint Surg Am 1976;58:450–3. [PubMed] [Google Scholar]
- 26.Yassa RR, Khalfaoui MY, Veravalli K, et al. Pre-operative urinary tract infection: is it a risk factor for early surgical site infection with hip fracture surgery? A retrospective analysis. JRSM Open 2017;8:1–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Sousa R, Munoz-Mahamud E, Quayle J, et al. Is asymptomatic bacteriuria a risk factor for prosthetic joint infection? Clin Infect Dis 2014;59:41–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Gallegos Salazar J, O’Brien W, Strymish JM, et al. Association of Screening and Treatment for Preoperative Asymptomatic Bacteriuria with Postoperative Outcomes Among US Veterans. JAMA Surg 2019;154:241–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Nicolle LE, Gupta K, Bradley SF, et al. Clinical Practice Guideline for the Management of Asymptomatic Bacteriuria: 2019 Update by the Infectious Diseases Society of America. Clin Infect Dis 2019;68:e83–e110. [DOI] [PubMed] [Google Scholar]
- 30.Lee SE, Kim KT, Park YS, et al. Association between Asymptomatic Urinary Tract Infection and Postoperative Spine Infection in Elderly Women : A Retrospective Analysis Study. J Korean Neurosurg Soc 2010;47:265–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Donlan RM. Biofilms: microbial life on surfaces. Emerg Infect Dis 2002;8:881–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.David TS, Vrahas MS. Perioperative lower urinary tract infections and deep sepsis in patients undergoing total joint arthroplasty. J Am Acad Orthop Surg 2000;8:66–74. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.