- Email: [email protected]
Characterization of perioperative infection risk among patients undergoing radical cystectomy: Results from the national surgical quality improvement program
Urologic Oncology: Seminars and Original Investigations ] (2016) ∎∎∎–∎∎∎
Original article
Characterization of perioperative infection risk among patients undergoing radical cystectomy: Results from the national surgical quality improvement program William P. Parker, M.D.a, Matthew K. Tollefson, M.D.a, Courtney N. Heins, B.S.b, Kristine T. Hanson, B.A.b, Elizabeth B. Habermann, Ph.D.b, Harras B. Zaid, M.D.a, Igor Frank, M.D.a, R. Houston Thompson, M.D.a, Stephen A. Boorjian, M.D., FACSa,* b
a Department of Urology, Mayo Clinic, Rochester, MN Department of Health Sciences Research, Mayo Clinic, Rochester, MN
Received 1 March 2016; received in revised form 24 May 2016; accepted 5 July 2016
Abstract Objectives: To evaluate the incidence, risk factors, and timing of infections following radical cystectomy (RC). Methods: The American College of Surgeons National Surgical Quality Improvement Project database was queried to identify patients undergoing RC for bladder cancer from 2006 to 2013. Characteristics including year of surgery, age, sex body mass index, diabetes, smoking, renal function, steroid usage, preoperative albumin, preoperative hematocrit, perioperative blood transfusion (PBT), and operative time were assessed for association with the risk of infection within 30 days of RC using multivariable logistic regression. Results: A total of 3,187 patients who had undergone RC were identified, of whom 766 (24.0%) were diagnosed with a postoperative infection, at a median of 13 days (interquartile ranges 8-19) after RC. Infections included surgical site infection (SSI) (404; 12.7%), sepsis/ septic shock (405; 12.7%), and urinary tract infection (UTI) (309; 9.7%). On multivariable analysis, body mass index Z30 kg/m2 (odds ratios [OR] ¼ 1.52; P o 0.01), receipt of a PBT (OR ¼ 1.27; P o 0.01), and operative time Z480 minutes (OR ¼ 1.72; P o 0.01) were significantly associated with the risk of infection. When the outcomes of UTI, SSI, and sepsis were analyzed separately, operative time Z480 minutes remained independently associated with increased infection risk in each model (OR ¼ 2.11 for UTI, OR ¼ 1.63 for SSI, and OR ¼ 1.80 for sepsis/septic shock; all P o 0.05), whereas PBT was associated with SSI and sepsis/septic shock (OR ¼ 1.33 and OR ¼ 1.29, respectively; both P o 0.05). Conclusions: Approximately 25% of patients undergoing RC experience an infection within 30 days of surgery. Several potentially modifiable risk factors for infection were identified, specifically PBT and prolonged operative time, which may represent opportunities for future care improvement. r 2016 Elsevier Inc. All rights reserved.
Keywords: Bladder cancer; Radical cystectomy; Infection; Risk factors; Urinary diversion; NSQIP
1. Introduction Although radical cystectomy (RC) represents a standard treatment for muscle-invasive bladder cancer as well as highrisk non–muscle-invasive disease [1], the morbidity of the procedure remains substantial. Indeed, complications have been reported to occur in up to 78% of patients undergoing RC [2–6]. In addition to patient-related sequelae, perioperative Corresponding author. Tel.: þ1-507-284-4015; fax: þ1-507-284-4951. E-mail address: [email protected] (Stephen A. Boorjian). *
http://dx.doi.org/10.1016/j.urolonc.2016.07.001 1078-1439/r 2016 Elsevier Inc. All rights reserved.
complications increase the risk of readmission, and the cost of care [7–9]. Infectious complications are of particular concern after RC given the use of intestinal substitution for urinary reconstruction. Perioperative infections, including urinary tract infections (UTI), surgical site infections (SSI), and sepsis, have been noted in 20% to 40% of patients following RC [3–5,10,11]. Importantly, however, limited specific analyses exist regarding infectious events after RC, as most often complications have been analyzed in aggregate. Moreover, most data to date on this topic consist of
2
W.P. Parker et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
reports from single high-volume centers, and as such, the generalizability of these data is uncertain [3,4,11]. Understanding risk factors for infectious events may not only facilitate patient counseling and monitoring, but further may allow for intervention and correction of modifiable variables. Our hypothesis is that infectious complications following RC are associated with specific clinical factors, the identification of which may improve patient care. Herein, therefore, we used a large national dataset to evaluate clinical factors associated with patients' risk of postoperative infection following RC. 2. Materials and methods 2.1. Data source and cohort development The American College of Surgeons National Surgical Quality Improvement Project (ACS-NSQIP) is a prospectively maintained database of patients undergoing surgical procedures at 435 hospitals [12]. Data in the ACS-NSQIP dataset are collected by trained surgical clinical reviewers. Morbidity and mortality outcomes are collected to 30 days postoperatively. To develop the present cohort, the 2006 to 2013 ACS-NSQIP database was queried to identify patients with bladder cancer (International Classification of Diseases, Ninth Revision, codes: 188.0 to 188.9, and 233.7) undergoing RC (Current Procedural Terminology codes: 51570, 51575, 51580, 51585, 51590, 51595, and 51596) (Fig. 1). Patients with a documented infection at the time of RC were excluded from analyses, as were patients undergoing concurrent organ resection, including nephrectomy, nephroureterectomy, and total pelvic exenteration. The final analytic cohort consisted of 3,187 patients.
2.2. Patient characteristics and outcomes Preoperative characteristics including age, sex, smoking status (classified as current [within 1 y of surgery] vs. not current), race/ethnicity, body mass index (BMI), diabetes (yes or no), current steroid use (yes or no), estimated glomerular filtration rate, serum albumin level, and preoperative hematocrit were recorded. Additionally, operative time, receipt of a perioperative blood transfusion (PBT; inclusive of intraoperative and postoperative during hospitalization), and year of surgery were captured. BMI was analyzed as a dichotomous variable (o30 kg/m2 vs. Z30 kg/m2) [13], consistent with previous RC series [3,14]. Hematocrit levels were defined based on sex thresholds as normal (male: Z42%; female: Z38%), with low (male: 32%–41%; female: 28%–37%), and very low (male: o32%; female o28%) considered for each 10% decrease in hematocrit [15]. Operative time was categorized into quartiles and rationalized to the nearest hour (o240 min, 240–359 min, 360–479 min, and Z480 min). The primary outcome of the study was the occurrence of any infection, including UTI, SSI, or sepsis/septic shock, within 30 days of RC. Separate analyses for each infectiontype (UTI, SSI, and sepsis/septic shock) were performed as well. The definitions for each of these events used by ACSNSQIP are provided in Supplemental Table S1. For patients in whom an infection occurred, the timing of the diagnosis of infection relative to RC was documented as well. 2.3. Data analysis Preoperative characteristics of the patient cohort were described using frequencies/percentages and medians/interquartile ranges (IQR). Associations of patient characteristics with the occurrence of infection were assessed using χ2 or
Fig. 1. Cohort selection. Final analytic cohort representative of patients undergoing RC without prior infection or concurrent organ resection.
W.P. Parker et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
3
Table 1 Baseline clinicopathologic characteristics, stratified by perioperative infection status Variable
Total cohort (N ¼ 3,187)
No infection (N ¼ 2,421)
Infection (N ¼ 766)
Median age at RC (IQR) Male sex (%)
70 (62-77) 2,604 (81.8)
70 (62-77) 1,991 (82.3)
70 (62-76) 613 (80.0)
Race/ethnicity (%) Non-Hispanic White Hispanic White Black Other
2,638 45 114 390
2,002 31 92 296
636 14 22 94
Current smoker (%) BMI (in. kg/m2) o30 Z30 Diabetes (%) Preoperative eGFR Z90 60–89 30–59 o30 Steroid use at the time of RC (%)
(82.7) (1.3) (3.8) (12.2)
(83.0) (1.8) (2.9) (12.3)
799 (25.1)
616 (25.4)
183 (23.9)
2,069 (65.4) 1,094 (34.6)
1,630 (67.9) 770 (32.1)
439 (57.5) 324 (42.5)
626 (19.6)
464 (19.2)
162 (21.1)
0.39 o0.01
0.23 0.32
531 1,437 1,027 114
(17.1) (46.2) (33.0) (3.7)
95 (3.0)
399 1,103 781 79
(16.9) (46.7) (33.1) (3.3)
67 (2.8)
132 334 246 35
(17.7) (44.7) (32.9) (4.7)
28 (3.7)
312 (15.6) 1,687 (84.4)
225 (15.1) 1,268 (84.9)
87 (17.2) 419 (82.8)
Preoperative hematocrita Very low Low Normal Receipt of PBT (%)
371 1,755 1,001 1,358
(11.9) (56.1) (32.0) (42.6)
277 1,327 773 985
(74.7%) (75.6%) (77.2%) (40.7)
94 428 228 373
(25.3%) (24.4%) (22.8%) (48.7)
Operative time (min) (%) o240 240–359 360–479 Z480
574 1,264 909 440
(18.0) (39.7) (28.5) (13.8)
458 993 677 293
(18.9) (41.0) (28.0) (12.1)
116 271 232 147
(15.1) (35.4) (30.3) (19.2)
Year of surgery (%) 2006 2007 2008 2009 2010 2011 2012 2013
0.34 0.15 0.45
(82.8) (1.4) (3.6) (12.2)
Albumin o3.5 Z3.5
Median length of stay (d) (IQR)
P value
0.21 0.26
0.52
o0.01 o0.01
8 (6-11)
7 (6-10)
9 (7-16)
o0.01 0.34
6 20 87 170 206 691 876 1131
(0.2) (0.6) (2.7) (5.3) (6.5) (21.7) (27.5) (35.9)
4 12 66 119 163 524 680 853
(0.2) (0.5) (2.7) (4.9) (6.7) (21.6) (28.1) (35.2)
2 8 21 51 43 167 196 278
(0.3) (1.0) (2.7) (6.7) (5.6) (21.8) (25.6) (36.3)
eGFR, estimated glomerular filtration rate. Missing data were observed as follows: sex (N ¼ 2), BMI (N ¼ 24), eGFR (N ¼ 78), albumin (N ¼ 1,188), and hematocrit (N ¼ 60). a Hematocrit levels were defined as follows: normal (male: Z42% and female: Z38%) low (male: 32%–41% and female: 28%–37%), and very low (male: o32%; female: o28%).
Fisher exact tests for categorical variables and t-tests/ Wilcoxon tests or 1-way ANOVA for continuous variables. Multivariable logistic regression was performed, and odds ratios (OR) were calculated to describe the independent
associations between patient characteristics and outcome. All P values are reported for the 2-tailed tests, with P o 0.05 considered significant. Statistical analyses were performed using SAS v9.3 (Cary, NC).
4
W.P. Parker et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
Fig. 2. Temporal distribution of infections within 30 days following RC. Each figure represents the median (dark horizontal bar), mean (diamond), interquartile range (box), and range (thin vertical bars with thin horizontal caps). The actual values are as follows: any infection—median 13, mean 13.9, IQR 8-13, and range 0 to 30 days. Any SSI—median 14, mean 14.3, IQR 9-19, and range 0 to 30 days. Any UTI—median 15, mean 15.0, IQR 9-21, and range 0 to 30 days. Any sepsis/septic shock—median 13, mean 13.9, IQR 9-19, and range 0 to 30 days. Of note, some patients were diagnosed with more than 1 infection and may, therefore, be represented more than once in the figure.
3. Results A total of 3,187 patients identified within the ACSNSQIP database who underwent RC between 2005 and 2013 were included for analysis. Clinicopathologic demographics are provided in Table 1. The median age of the cohort was 70 years (IQR 62-77), while 2,604 (81.8%) were male. A total of 766 (24.0%) patients were diagnosed with an infection within 30 days of RC, including 404 (12.7%) who experienced a SSI, 405 (12.7%) with sepsis/septic shock, and 309 (9.7%) with a UTI. In total, 303 of the 766 patients (39.6%) were diagnosed with more than 1 infection within 30 days of RC. Of patients with a UTI, additional infections occurred as follows: sepsis/septic shock in 111 (35.9%), SSI in 28 (9.1%), and both SSI and sepsis/septic shock in 49 (15.9%). For patients diagnosed with a SSI, 115 (28.5%) were additionally diagnosed with sepsis/septic shock. Infections were diagnosed at a median of 13 days (IQR 8-19) after surgery (Fig. 2), whereas median length of stay was 8 days (IQR 6-10.5) during the study period. Compared with patients without an infection, median length of stay was longer among the patients suffering an infection within 30 days of surgery (median 9 d vs. 7 d; P o 0.01). Compared to patients not diagnosed with infection within 30 days of RC, patients who experienced a postoperative infection were more often obese (BMI Z 30 kg/m2: 42.5% vs. 32.1%; P o 0.01), more likely to have received a PBT (48.7% vs. 40.7%; P o 0.01), and to have had a longer
operative time during RC (median 358 min vs. 329 min; P o 0.01). Indeed, on multivariable analysis (Table 2), BMI Z 30 kg/m2 (OR ¼ 1.52; P o 0.01), receipt of a PBT (OR ¼ 1.27; Po0.01, and prolonged (Z480 min) operative time (OR ¼ 1.72; P o 0.01) remained associated with a significantly increased risk of infection within 30 days of surgery. Additional multivariable models were then created to individually assess the association of clinicopathologic features with the diagnoses of postoperative SSI, UTI, and sepsis/septic shock (Table 3). We found that patient-related factors of younger age (SSI and UTI) and BMI Z 30 kg/m2 (SSI, UTI, and sepsis/septic shock) and operative factors of PBT (SSI and sepsis/septic shock) and operative time Z480 minutes (SSI, UTI, and sepsis/septic shock) remained independently associated with an increased risk of these events.
4. Discussion We found here, in a large national dataset of patients treated with RC for bladder cancer, that approximately 25% were diagnosed with an infection within 30 days of surgery. We further determined that infections occurred at a median of 13 days postoperatively. In addition, we identified risk factors for infection, including the potentially modifiable variables of PBT and operative time.
W.P. Parker et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7 Table 2 Multivariable logistic regression analysis for risk of infection within 30 days of RC Variable
OR
95% CI
P value
Age at RC (ref 4 70) o50 50–69
– 1.50 0.89
– 1.00–2.27 0.74–1.07
– 0.05 0.20
BMI Z 30 kg/m2 (ref o 30) Female Current smoker Diabetes
1.52 1.07 0.95 1.04
1.27–1.81 0.86–1.33 0.77–1.16 0.85–1.28
o0.01 0.54 0.62 0.71
CKD (ref eGFR 4 90) CKD II (eGFR 60–89) CKD III (eGFR 30–59) CKD IV (eGFR o 30)
– 0.94 0.92 1.25
– 0.74–1.20 0.71–1.19 0.77–2.00
– 0.61 0.53 0.34
Steroid use at the time of RC
1.32
0.84–2.09
0.23
Albumin (ref Z 3.5) o3.5
– 1.18
– 0.88–1.58
– 0.26
Hematocrita (ref normal) Low Very low
– 1.01 1.04
– 0.84–1.23 0.76–1.42
– 0.89 0.81
Perioperative transfusion (ref no)
1.27
1.06–1.52
o0.01
Operative time (ref o 240 min) 240–359 min 360–479 min Z480 min
– 1.02 1.25 1.72
– 0.80–1.31 0.96–1.62 1.27–2.32
– 0.87 0.10 o0.01
Year of surgery (ref 2013) 2006 2007 2008 2009 2010 2011 2012
– 1.37 2.04 1.02 1.32 0.87 1.00 0.89
– 0.25–7.70 0.81–5.13 0.61–1.72 0.92–1.89 0.60–1.26 0.80–1.25 0.72–1.10
– 0.72 0.13 0.94 0.14 0.45 0.99 0.27
CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate. a Hematocrit levels were defined as follows: normal (male: Z42%, female: Z38%) low (male: 32%–41%; female: 28%–37%), and very low (male: o32% and female: o28%).
Most series to date focused on perioperative RC outcomes have analyzed complications in aggregate, and as such there has been limited characterization of infectious events, including defining predisposing factors that may guide patient counseling and to offer opportunities for care improvement. For example, a report from the International Robotic Cystectomy Consortium found that BMI and receipt of a PBT were associated with the risk of overall complications within 90 days of surgery among patients undergoing robotic RC [10], whereas a prior analysis of the ACS-NSQIP dataset assessing the risk of any complication after RC identified operative time 46 hours as associated with complication risk, although again, the analysis did not evaluate the risk of infectious complications separately [2].
5
Likewise, BMI has been associated with the risk of wound complications, particularly SSI, following a variety of surgical procedures [16,17]. Recently, moreover, Pariser et al. [11], in an institutional series of 386 patients treated with RC and urinary diversion, demonstrated an association between higher BMI and increased overall postoperative infection risk. We extend these analyses by documenting an association between BMI and the risk of SSI as well as UTI and sepsis/septic shock after RC. The pathophysiology of these associations, particularly regarding UTI, remains to be determined, and may reflect the challenge of performing urinary diversion or differences in microbial distribution among patients with obesity or both. Meanwhile, receipt of a PBT has also been shown to be associated with complications after RC [2,18]. We found here that PBT was independently associated with overall infectious risk after RC, including specifically SSI and sepsis/septic shock. Data linking PBT to infections and complications after RC, together with the cost of PBT, the potential for transfusion reaction, and evidence supporting an association of PBT with adverse cancer outcomes [19,20], including among patients undergoing RC specifically [21–23], support the critical review of PBT as a modifiable risk factor and argue for the use of judicious perioperative blood management strategies [24,25]. Similarly, prolonged operative time has been identified as a risk factor for surgical complications [26,27], including for RC [2,11]. Our data add to these findings to demonstrate an association of prolonged operative time with overall and specific site infectious risk after RC. Interestingly, we noted as well an association between younger patient age and increased risk of postoperative UTI and SSI. Although prior studies have noted an increased risk of complications in older patients undergoing RC [2,3,10], again, these series considered complications in aggregate, thereby potentially obscuring the relationship between age and infection. For UTI specifically, Nazmy et al. [5], in a series of 209 patients treated with robotic RC and urinary diversion, found on multivariable analysis inclusive of both age and smoking status that neither variable was significantly associated with UTI risk. Our data, in 3,187 patients who had undergone RC, may be because of unmeasured confounding factors among younger patients undergoing RC, particularly with urinary diversion status, where increased use of continent diversions among younger patients may explain this association [28]. Similarly, our finding of a decreased risk of UTI among active smokers is likely reflective of unmeasured confounding. We further characterized infectious complications after RC by reporting the time from RC to diagnosis. We demonstrated that the timing of infection after RC is approximately 2 weeks after surgery, at a time when patients have typically been discharged. Indeed, the median length of stay here was 8 days. Based on these data, future targeted follow-up efforts may be
W.P. Parker et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
6
Table 3 Multivariable logistic regression analyses for risks of urinary tract infection, surgical site infection, and sepsis/septic shock within 30 days of RC Variable
UTI
SSI
OR
95% CI
Age (ref 4 70) o50 50–69
– 1.89 1.05
– 1.11–3.21 0.81–1.37
BMI Z 30 kg/m2 (ref o 30) Female Current smoker Diabetes
1.39 0.97 0.69 0.96
CKD (ref eGFR 4 90) CKD II (eGFR 60–89) CKD III (eGFR 30–59) CKD IV (eGFR o30) Steroid use at the time of RC
P value
Sepsis/septic shock
OR
95% CI
– 0.02 0.72
– 1.73 0.82
– 1.07–2.79 0.65–1.05
1.08–1.79 0.71–1.34 0.51–0.94 0.71–1.31
0.01 0.86 0.02 0.81
1.71 1.20 1.07 1.23
– 0.98 0.85 1.06 1.33
– 0.70–1.38 0.58–1.23 0.54–2.07 0.71–2.50
– 0.92 0.38 0.87 0.37
Albumin (ref Z 3.5) o3.5
– 1.24
– 0.82–1.87
Hematocrita (ref normal) Low Very low Receipt of PBT (ref no)
– 1.15 1.17 1.11
Operative time (ref o240 min) 240–359 min 360–479 min Z480 min Year of surgery (ref 2013) 2006 2007 2008 2009 2010 2011 2012
P value
OR
95% CI
P value
– 0.03 0.11
– 1.27 0.88
– 0.75–2.14 0.69–1.11
– 0.38 0.27
1.37–2.14 0.92–1.57 0.82–1.39 0.95–1.59
o0.01 0.19 0.62 0.12
1.39 0.89 0.87 0.99
1.11–1.74 0.67–1.19 0.67–1.13 0.76–1.29
o0.01 0.44 0.30 0.93
– 0.92 0.97 1.53 1.05
– 0.68–1.26 0.69–1.36 0.87–2.69 0.57–1.92
– 0.62 0.85 0.15 0.89
– 0.92 0.94 1.33 1.23
– 0.68–1.26 0.68–1.31 0.76–2.33 0.69–2.21
– 0.61 0.73 0.31 0.48
– 0.31
– 1.08
– 0.74–1.57
– 0.70
– 1.26
– 0.89–1.79
– 0.19
– 0.87–1.51 0.74–1.83 0.86–1.44
– 0.34 0.51 0.41
– 0.90 0.89 1.33
– 0.71–1.15 0.59–1.33 1.06–1.67
– 0.40 0.57 0.02
– 0.96 1.30 1.29
– 0.75–1.23 0.89–1.90 1.03–1.62
– 0.73 0.17 0.03
– 1.20 1.30 2.11
– 0.82–1.75 0.87–1.93 1.37–3.24
– 0.34 0.20 o0.01
– 1.15 1.65 1.63
– 0.82–1.62 1.17–2.34 1.09–2.43
– 0.43 o0.01 0.02
– 0.98 1.09 1.80
– 0.71–1.35 0.78–1.53 1.24–2.60
– 0.91 0.61 o0.01
– 3.90 1.77 1.15 1.17 0.80 1.04 0.76
– 0.68–22.35 0.50–6.28 0.56–2.38 0.70–1.96 0.46–1.39 0.76–1.42 0.55–1.04
– 0.13 0.38 0.71 0.54 0.43 0.81 0.08
– 1.23 1.13 1.02 1.08 0.92 1.07 0.94
– 0.14–11.03 0.32–4.00 0.52–1.99 0.67–1.75 0.57–1.48 0.80–1.43 0.72–1.24
– 0.85 0.85 0.96 0.75 0.72 0.64 0.66
– 1.39 3.35 1.52 1.93 1.07 1.09 0.99
– 0.16–12.33 1.23–9.11 0.83–2.80 1.27–2.94 0.67–1.71 0.82–1.47 0.75–1.30
– 0.77 0.02 0.18 o0.01 0.79 0.55 0.92
CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate. Hematocrit levels were defined as follows: normal (male: Z42% and female: Z38%) Low (male: 32–41% and female: 28%–37%), and very low (male: o32% and female: o28%). a
considered to educate or evaluate patients or both with risk factors for infections to facilitate the early diagnosis and management of these events. We believe that the use of a large national dataset improves the generalizability of our findings. Moreover, in addition to its large sample size and the ability to discern 30-day outcomes, the ACS-NSQIP database is maintained by trained abstractors rather than administrative or index hospital data. Nevertheless, we must acknowledge the limitations of this dataset, and thereby of our analyses. That is, we could not adjust for pathologic variables such as tumor stage or lymph node status that may be important variables to consider with the risk of longer operative times and blood transfusion receipt. Likewise, receipt of neoadjuvant chemotherapy, preoperative radiation therapy, use of bowel preparation, perioperative antibiotic
regimens, and postoperative stent/catheterization were not captured and could, therefore, not be analyzed. Additionally, although we controlled for year of surgery as a means to account for changing surgical practices, the type of urinary diversion and surgical approach (open, minimally invasive) were unable to be specifically identified and evaluated. We, therefore, recognize the unmeasured confounding that these variables may have had on our resultant data. Finally, although we did note a difference in length of stay between patients with and without an infection, we recognize that this finding could be because of either length of stay affecting the incidence of infections or because of the effect of infections on prolonging length of stay. Because of the inability to distinguish these 2 phenomena, we chose not to include length of stay as available in our multivariable models.
W.P. Parker et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
5. Conclusions In conclusion, using a large national dataset, we noted that approximately 25% of patients undergoing RC are diagnosed with an infection within 30 days of surgery. We identified several patient-related risk factors for infection, including age and BMI. Moreover, potentially modifiable risk factors such as receipt of a PBT and prolonged operative time were also found to be associated with infection risk, and may, pending validation in other datasets, thereby represent targets for future quality improvement efforts to reduce the incidence of infection in this setting. Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j. urolonc.2016.07.001.
References [1] Clark PE, Agarwal N, Biagioli MC, et al. Bladder cancer. J Natl Compr Canc Netw 2013;11:446–75. [2] Lavallee LT, Schramm D, Witiuk K, et al. Peri-operative morbidity associated with radical cystectomy in a multicenter database of community and academic hospitals. PLoS One 2014;9:e111281. [3] Shabsigh A, Korets R, Vora KC, et al. Defining early morbidity of radical cystectomy for patients with bladder cancer using a standardized reporting methodology. Eur Urol 2009;55:164–74. [4] Lowrance WT, Rumohr JA, Chang SS, Clark PE, Smith JA Jr., Cookson MS. Contemporary open radical cystectomy: analysis of perioperative outcomes. J Urol 2008;179:1313–8:[discussion 8]. [5] Nazmy M, Yuh B, Kawachi M, et al. Early and late complications of robot-assisted radical cystectomy: a standardized analysis by urinary diversion type. J Urol 2014;191:681–7. [6] Shigemura K, Tanaka K, Matsumoto M, et al. Post-operative infection and prophylactic antibiotic administration after radical cystectomy with orthotopic neobladder urinary diversion. J Infect Chemother 2012;18:479–84. [7] Kim SP, Shah ND, Karnes RJ, et al. The implications of hospital acquired adverse events on mortality, length of stay and costs for patients undergoing radical cystectomy for bladder cancer. J Urol 2012;187:2011–7. [8] Konety BR, Allareddy V. Influence of post-cystectomy complications on cost and subsequent outcome. J Urol 2007;177:280–7:[discussion 7]. [9] Skolarus TA, Jacobs BL, Schroeck FR, et al. Understanding hospital readmission intensity after radical cystectomy. J Urol 2015;193:1500–6. [10] Johar RS, Hayn MH, Stegemann AP, et al. Complications after robotassisted radical cystectomy: results from the International Robotic Cystectomy Consortium. Eur Urol 2013;64:52–7. [11] Pariser JJ, Anderson BB, Pearce SM, et al. The effect of broader, directed antimicrobial prophylaxis including fungal coverage on perioperative
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24] [25]
[26]
[27]
[28]
7
infectious complications after radical cystectomy. Urol Oncol 2015;34:e9121.e14. American College of Surgeons National Surgical Quality Improvement Program. User guide for the 2013 ACS NSQIP Participant Use Data File (PUF); 2013. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: executive summary. Expert panel on the identification, evaluation, and treatment of overweight in adults. Am J Clin Nutr 1998;68:899–917. Chang SS, Jacobs B, Wells N, Smith JA Jr., Cookson MS. Increased body mass index predicts increased blood loss during radical cystectomy. J Urol 2004;171:1077–9. Kratz A, Ferraro M, Sluss PM, Lewandrowski KB. Case records of the Massachusetts General Hospital. Weekly clinicopathological exercises. Laboratory reference values. N Engl J Med. 2004;351: 1548–1563. Mullen JT, Moorman DW, Davenport DL. The obesity paradox: body mass index and outcomes in patients undergoing nonbariatric general surgery. AnnSurg 2009;250:166–72. Shah DK, Vitonis AF, Missmer SA. Association of body mass index and morbidity after abdominal, vaginal, and laparoscopic hysterectomy. Obstet Gynecol 2015;125:589–98. Lee SJ, Seo H, Kim HC, et al. Effect of intraoperative red blood cell transfusion on postoperative complications after open radical cystectomy: old versus fresh stored blood. Clin Genitourin Cancer 2015;13:581–7. Acheson AG, Brookes MJ, Spahn DR. Effects of allogeneic red blood cell transfusions on clinical outcomes in patients undergoing colorectal cancer surgery: a systematic review and meta-analysis. Ann Surg 2012;256:235–44. Al-Refaie WB, Parsons HM, Markin A, Abrams J, Habermann EB. Blood transfusion and cancer surgery outcomes: a continued reason for concern. Surgery 2012;152:344–54. Wang YL, Jiang B, Yin FF, et al. Perioperative blood transfusion promotes worse outcomes of bladder cancer after radical cystectomy: a systematic review and meta-analysis. PLoS One 2015;10: e0130122. Linder BJ, Frank I, Cheville JC, et al. The impact of perioperative blood transfusion on cancer recurrence and survival following radical cystectomy. Eur Urol 2013;63:839–45. Abel EJ, Linder BJ, Bauman TM, et al. Perioperative blood transfusion and radical cystectomy: does timing of transfusion affect bladder cancer mortality? Eur Urol 2014;66:1139–47. Clevenger B, Mallett SV, Klein AA, Richards T. Patient blood management to reduce surgical risk. Br J Surg 2015;102:1325–37. Spahn DR, Moch H, Hofmann A, Isbister JP. Patient blood management: the pragmatic solution for the problems with blood transfusions. Anesthesiology 2008;109:951–3. Daley BJ, Cecil W, Clarke PC, Cofer JB, Guillamondegui OD. How slow is too slowCorrelation of operative time to complications: an analysis from the Tennessee Surgical Quality Collaborative. J Am Coll Surg 2015;220:550–8. Qin C, de Oliveira G, Hackett N, Kim JY. Surgical duration and risk of urinary tract infection: an analysis of 1,452,369 patients using the National Surgical Quality Improvement Program (NSQIP). Int J Surg 2015;20:107–12. Parker WP, Toussi A, Tollefson MK, et al. Risk factors and microbial distribution of urinary tract infections following radical cystectomy. Urology 2016. [Epub ahead of print].
Original article
Characterization of perioperative infection risk among patients undergoing radical cystectomy: Results from the national surgical quality improvement program William P. Parker, M.D.a, Matthew K. Tollefson, M.D.a, Courtney N. Heins, B.S.b, Kristine T. Hanson, B.A.b, Elizabeth B. Habermann, Ph.D.b, Harras B. Zaid, M.D.a, Igor Frank, M.D.a, R. Houston Thompson, M.D.a, Stephen A. Boorjian, M.D., FACSa,* b
a Department of Urology, Mayo Clinic, Rochester, MN Department of Health Sciences Research, Mayo Clinic, Rochester, MN
Received 1 March 2016; received in revised form 24 May 2016; accepted 5 July 2016
Abstract Objectives: To evaluate the incidence, risk factors, and timing of infections following radical cystectomy (RC). Methods: The American College of Surgeons National Surgical Quality Improvement Project database was queried to identify patients undergoing RC for bladder cancer from 2006 to 2013. Characteristics including year of surgery, age, sex body mass index, diabetes, smoking, renal function, steroid usage, preoperative albumin, preoperative hematocrit, perioperative blood transfusion (PBT), and operative time were assessed for association with the risk of infection within 30 days of RC using multivariable logistic regression. Results: A total of 3,187 patients who had undergone RC were identified, of whom 766 (24.0%) were diagnosed with a postoperative infection, at a median of 13 days (interquartile ranges 8-19) after RC. Infections included surgical site infection (SSI) (404; 12.7%), sepsis/ septic shock (405; 12.7%), and urinary tract infection (UTI) (309; 9.7%). On multivariable analysis, body mass index Z30 kg/m2 (odds ratios [OR] ¼ 1.52; P o 0.01), receipt of a PBT (OR ¼ 1.27; P o 0.01), and operative time Z480 minutes (OR ¼ 1.72; P o 0.01) were significantly associated with the risk of infection. When the outcomes of UTI, SSI, and sepsis were analyzed separately, operative time Z480 minutes remained independently associated with increased infection risk in each model (OR ¼ 2.11 for UTI, OR ¼ 1.63 for SSI, and OR ¼ 1.80 for sepsis/septic shock; all P o 0.05), whereas PBT was associated with SSI and sepsis/septic shock (OR ¼ 1.33 and OR ¼ 1.29, respectively; both P o 0.05). Conclusions: Approximately 25% of patients undergoing RC experience an infection within 30 days of surgery. Several potentially modifiable risk factors for infection were identified, specifically PBT and prolonged operative time, which may represent opportunities for future care improvement. r 2016 Elsevier Inc. All rights reserved.
Keywords: Bladder cancer; Radical cystectomy; Infection; Risk factors; Urinary diversion; NSQIP
1. Introduction Although radical cystectomy (RC) represents a standard treatment for muscle-invasive bladder cancer as well as highrisk non–muscle-invasive disease [1], the morbidity of the procedure remains substantial. Indeed, complications have been reported to occur in up to 78% of patients undergoing RC [2–6]. In addition to patient-related sequelae, perioperative Corresponding author. Tel.: þ1-507-284-4015; fax: þ1-507-284-4951. E-mail address: [email protected] (Stephen A. Boorjian). *
http://dx.doi.org/10.1016/j.urolonc.2016.07.001 1078-1439/r 2016 Elsevier Inc. All rights reserved.
complications increase the risk of readmission, and the cost of care [7–9]. Infectious complications are of particular concern after RC given the use of intestinal substitution for urinary reconstruction. Perioperative infections, including urinary tract infections (UTI), surgical site infections (SSI), and sepsis, have been noted in 20% to 40% of patients following RC [3–5,10,11]. Importantly, however, limited specific analyses exist regarding infectious events after RC, as most often complications have been analyzed in aggregate. Moreover, most data to date on this topic consist of
2
W.P. Parker et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
reports from single high-volume centers, and as such, the generalizability of these data is uncertain [3,4,11]. Understanding risk factors for infectious events may not only facilitate patient counseling and monitoring, but further may allow for intervention and correction of modifiable variables. Our hypothesis is that infectious complications following RC are associated with specific clinical factors, the identification of which may improve patient care. Herein, therefore, we used a large national dataset to evaluate clinical factors associated with patients' risk of postoperative infection following RC. 2. Materials and methods 2.1. Data source and cohort development The American College of Surgeons National Surgical Quality Improvement Project (ACS-NSQIP) is a prospectively maintained database of patients undergoing surgical procedures at 435 hospitals [12]. Data in the ACS-NSQIP dataset are collected by trained surgical clinical reviewers. Morbidity and mortality outcomes are collected to 30 days postoperatively. To develop the present cohort, the 2006 to 2013 ACS-NSQIP database was queried to identify patients with bladder cancer (International Classification of Diseases, Ninth Revision, codes: 188.0 to 188.9, and 233.7) undergoing RC (Current Procedural Terminology codes: 51570, 51575, 51580, 51585, 51590, 51595, and 51596) (Fig. 1). Patients with a documented infection at the time of RC were excluded from analyses, as were patients undergoing concurrent organ resection, including nephrectomy, nephroureterectomy, and total pelvic exenteration. The final analytic cohort consisted of 3,187 patients.
2.2. Patient characteristics and outcomes Preoperative characteristics including age, sex, smoking status (classified as current [within 1 y of surgery] vs. not current), race/ethnicity, body mass index (BMI), diabetes (yes or no), current steroid use (yes or no), estimated glomerular filtration rate, serum albumin level, and preoperative hematocrit were recorded. Additionally, operative time, receipt of a perioperative blood transfusion (PBT; inclusive of intraoperative and postoperative during hospitalization), and year of surgery were captured. BMI was analyzed as a dichotomous variable (o30 kg/m2 vs. Z30 kg/m2) [13], consistent with previous RC series [3,14]. Hematocrit levels were defined based on sex thresholds as normal (male: Z42%; female: Z38%), with low (male: 32%–41%; female: 28%–37%), and very low (male: o32%; female o28%) considered for each 10% decrease in hematocrit [15]. Operative time was categorized into quartiles and rationalized to the nearest hour (o240 min, 240–359 min, 360–479 min, and Z480 min). The primary outcome of the study was the occurrence of any infection, including UTI, SSI, or sepsis/septic shock, within 30 days of RC. Separate analyses for each infectiontype (UTI, SSI, and sepsis/septic shock) were performed as well. The definitions for each of these events used by ACSNSQIP are provided in Supplemental Table S1. For patients in whom an infection occurred, the timing of the diagnosis of infection relative to RC was documented as well. 2.3. Data analysis Preoperative characteristics of the patient cohort were described using frequencies/percentages and medians/interquartile ranges (IQR). Associations of patient characteristics with the occurrence of infection were assessed using χ2 or
Fig. 1. Cohort selection. Final analytic cohort representative of patients undergoing RC without prior infection or concurrent organ resection.
W.P. Parker et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
3
Table 1 Baseline clinicopathologic characteristics, stratified by perioperative infection status Variable
Total cohort (N ¼ 3,187)
No infection (N ¼ 2,421)
Infection (N ¼ 766)
Median age at RC (IQR) Male sex (%)
70 (62-77) 2,604 (81.8)
70 (62-77) 1,991 (82.3)
70 (62-76) 613 (80.0)
Race/ethnicity (%) Non-Hispanic White Hispanic White Black Other
2,638 45 114 390
2,002 31 92 296
636 14 22 94
Current smoker (%) BMI (in. kg/m2) o30 Z30 Diabetes (%) Preoperative eGFR Z90 60–89 30–59 o30 Steroid use at the time of RC (%)
(82.7) (1.3) (3.8) (12.2)
(83.0) (1.8) (2.9) (12.3)
799 (25.1)
616 (25.4)
183 (23.9)
2,069 (65.4) 1,094 (34.6)
1,630 (67.9) 770 (32.1)
439 (57.5) 324 (42.5)
626 (19.6)
464 (19.2)
162 (21.1)
0.39 o0.01
0.23 0.32
531 1,437 1,027 114
(17.1) (46.2) (33.0) (3.7)
95 (3.0)
399 1,103 781 79
(16.9) (46.7) (33.1) (3.3)
67 (2.8)
132 334 246 35
(17.7) (44.7) (32.9) (4.7)
28 (3.7)
312 (15.6) 1,687 (84.4)
225 (15.1) 1,268 (84.9)
87 (17.2) 419 (82.8)
Preoperative hematocrita Very low Low Normal Receipt of PBT (%)
371 1,755 1,001 1,358
(11.9) (56.1) (32.0) (42.6)
277 1,327 773 985
(74.7%) (75.6%) (77.2%) (40.7)
94 428 228 373
(25.3%) (24.4%) (22.8%) (48.7)
Operative time (min) (%) o240 240–359 360–479 Z480
574 1,264 909 440
(18.0) (39.7) (28.5) (13.8)
458 993 677 293
(18.9) (41.0) (28.0) (12.1)
116 271 232 147
(15.1) (35.4) (30.3) (19.2)
Year of surgery (%) 2006 2007 2008 2009 2010 2011 2012 2013
0.34 0.15 0.45
(82.8) (1.4) (3.6) (12.2)
Albumin o3.5 Z3.5
Median length of stay (d) (IQR)
P value
0.21 0.26
0.52
o0.01 o0.01
8 (6-11)
7 (6-10)
9 (7-16)
o0.01 0.34
6 20 87 170 206 691 876 1131
(0.2) (0.6) (2.7) (5.3) (6.5) (21.7) (27.5) (35.9)
4 12 66 119 163 524 680 853
(0.2) (0.5) (2.7) (4.9) (6.7) (21.6) (28.1) (35.2)
2 8 21 51 43 167 196 278
(0.3) (1.0) (2.7) (6.7) (5.6) (21.8) (25.6) (36.3)
eGFR, estimated glomerular filtration rate. Missing data were observed as follows: sex (N ¼ 2), BMI (N ¼ 24), eGFR (N ¼ 78), albumin (N ¼ 1,188), and hematocrit (N ¼ 60). a Hematocrit levels were defined as follows: normal (male: Z42% and female: Z38%) low (male: 32%–41% and female: 28%–37%), and very low (male: o32%; female: o28%).
Fisher exact tests for categorical variables and t-tests/ Wilcoxon tests or 1-way ANOVA for continuous variables. Multivariable logistic regression was performed, and odds ratios (OR) were calculated to describe the independent
associations between patient characteristics and outcome. All P values are reported for the 2-tailed tests, with P o 0.05 considered significant. Statistical analyses were performed using SAS v9.3 (Cary, NC).
4
W.P. Parker et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
Fig. 2. Temporal distribution of infections within 30 days following RC. Each figure represents the median (dark horizontal bar), mean (diamond), interquartile range (box), and range (thin vertical bars with thin horizontal caps). The actual values are as follows: any infection—median 13, mean 13.9, IQR 8-13, and range 0 to 30 days. Any SSI—median 14, mean 14.3, IQR 9-19, and range 0 to 30 days. Any UTI—median 15, mean 15.0, IQR 9-21, and range 0 to 30 days. Any sepsis/septic shock—median 13, mean 13.9, IQR 9-19, and range 0 to 30 days. Of note, some patients were diagnosed with more than 1 infection and may, therefore, be represented more than once in the figure.
3. Results A total of 3,187 patients identified within the ACSNSQIP database who underwent RC between 2005 and 2013 were included for analysis. Clinicopathologic demographics are provided in Table 1. The median age of the cohort was 70 years (IQR 62-77), while 2,604 (81.8%) were male. A total of 766 (24.0%) patients were diagnosed with an infection within 30 days of RC, including 404 (12.7%) who experienced a SSI, 405 (12.7%) with sepsis/septic shock, and 309 (9.7%) with a UTI. In total, 303 of the 766 patients (39.6%) were diagnosed with more than 1 infection within 30 days of RC. Of patients with a UTI, additional infections occurred as follows: sepsis/septic shock in 111 (35.9%), SSI in 28 (9.1%), and both SSI and sepsis/septic shock in 49 (15.9%). For patients diagnosed with a SSI, 115 (28.5%) were additionally diagnosed with sepsis/septic shock. Infections were diagnosed at a median of 13 days (IQR 8-19) after surgery (Fig. 2), whereas median length of stay was 8 days (IQR 6-10.5) during the study period. Compared with patients without an infection, median length of stay was longer among the patients suffering an infection within 30 days of surgery (median 9 d vs. 7 d; P o 0.01). Compared to patients not diagnosed with infection within 30 days of RC, patients who experienced a postoperative infection were more often obese (BMI Z 30 kg/m2: 42.5% vs. 32.1%; P o 0.01), more likely to have received a PBT (48.7% vs. 40.7%; P o 0.01), and to have had a longer
operative time during RC (median 358 min vs. 329 min; P o 0.01). Indeed, on multivariable analysis (Table 2), BMI Z 30 kg/m2 (OR ¼ 1.52; P o 0.01), receipt of a PBT (OR ¼ 1.27; Po0.01, and prolonged (Z480 min) operative time (OR ¼ 1.72; P o 0.01) remained associated with a significantly increased risk of infection within 30 days of surgery. Additional multivariable models were then created to individually assess the association of clinicopathologic features with the diagnoses of postoperative SSI, UTI, and sepsis/septic shock (Table 3). We found that patient-related factors of younger age (SSI and UTI) and BMI Z 30 kg/m2 (SSI, UTI, and sepsis/septic shock) and operative factors of PBT (SSI and sepsis/septic shock) and operative time Z480 minutes (SSI, UTI, and sepsis/septic shock) remained independently associated with an increased risk of these events.
4. Discussion We found here, in a large national dataset of patients treated with RC for bladder cancer, that approximately 25% were diagnosed with an infection within 30 days of surgery. We further determined that infections occurred at a median of 13 days postoperatively. In addition, we identified risk factors for infection, including the potentially modifiable variables of PBT and operative time.
W.P. Parker et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7 Table 2 Multivariable logistic regression analysis for risk of infection within 30 days of RC Variable
OR
95% CI
P value
Age at RC (ref 4 70) o50 50–69
– 1.50 0.89
– 1.00–2.27 0.74–1.07
– 0.05 0.20
BMI Z 30 kg/m2 (ref o 30) Female Current smoker Diabetes
1.52 1.07 0.95 1.04
1.27–1.81 0.86–1.33 0.77–1.16 0.85–1.28
o0.01 0.54 0.62 0.71
CKD (ref eGFR 4 90) CKD II (eGFR 60–89) CKD III (eGFR 30–59) CKD IV (eGFR o 30)
– 0.94 0.92 1.25
– 0.74–1.20 0.71–1.19 0.77–2.00
– 0.61 0.53 0.34
Steroid use at the time of RC
1.32
0.84–2.09
0.23
Albumin (ref Z 3.5) o3.5
– 1.18
– 0.88–1.58
– 0.26
Hematocrita (ref normal) Low Very low
– 1.01 1.04
– 0.84–1.23 0.76–1.42
– 0.89 0.81
Perioperative transfusion (ref no)
1.27
1.06–1.52
o0.01
Operative time (ref o 240 min) 240–359 min 360–479 min Z480 min
– 1.02 1.25 1.72
– 0.80–1.31 0.96–1.62 1.27–2.32
– 0.87 0.10 o0.01
Year of surgery (ref 2013) 2006 2007 2008 2009 2010 2011 2012
– 1.37 2.04 1.02 1.32 0.87 1.00 0.89
– 0.25–7.70 0.81–5.13 0.61–1.72 0.92–1.89 0.60–1.26 0.80–1.25 0.72–1.10
– 0.72 0.13 0.94 0.14 0.45 0.99 0.27
CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate. a Hematocrit levels were defined as follows: normal (male: Z42%, female: Z38%) low (male: 32%–41%; female: 28%–37%), and very low (male: o32% and female: o28%).
Most series to date focused on perioperative RC outcomes have analyzed complications in aggregate, and as such there has been limited characterization of infectious events, including defining predisposing factors that may guide patient counseling and to offer opportunities for care improvement. For example, a report from the International Robotic Cystectomy Consortium found that BMI and receipt of a PBT were associated with the risk of overall complications within 90 days of surgery among patients undergoing robotic RC [10], whereas a prior analysis of the ACS-NSQIP dataset assessing the risk of any complication after RC identified operative time 46 hours as associated with complication risk, although again, the analysis did not evaluate the risk of infectious complications separately [2].
5
Likewise, BMI has been associated with the risk of wound complications, particularly SSI, following a variety of surgical procedures [16,17]. Recently, moreover, Pariser et al. [11], in an institutional series of 386 patients treated with RC and urinary diversion, demonstrated an association between higher BMI and increased overall postoperative infection risk. We extend these analyses by documenting an association between BMI and the risk of SSI as well as UTI and sepsis/septic shock after RC. The pathophysiology of these associations, particularly regarding UTI, remains to be determined, and may reflect the challenge of performing urinary diversion or differences in microbial distribution among patients with obesity or both. Meanwhile, receipt of a PBT has also been shown to be associated with complications after RC [2,18]. We found here that PBT was independently associated with overall infectious risk after RC, including specifically SSI and sepsis/septic shock. Data linking PBT to infections and complications after RC, together with the cost of PBT, the potential for transfusion reaction, and evidence supporting an association of PBT with adverse cancer outcomes [19,20], including among patients undergoing RC specifically [21–23], support the critical review of PBT as a modifiable risk factor and argue for the use of judicious perioperative blood management strategies [24,25]. Similarly, prolonged operative time has been identified as a risk factor for surgical complications [26,27], including for RC [2,11]. Our data add to these findings to demonstrate an association of prolonged operative time with overall and specific site infectious risk after RC. Interestingly, we noted as well an association between younger patient age and increased risk of postoperative UTI and SSI. Although prior studies have noted an increased risk of complications in older patients undergoing RC [2,3,10], again, these series considered complications in aggregate, thereby potentially obscuring the relationship between age and infection. For UTI specifically, Nazmy et al. [5], in a series of 209 patients treated with robotic RC and urinary diversion, found on multivariable analysis inclusive of both age and smoking status that neither variable was significantly associated with UTI risk. Our data, in 3,187 patients who had undergone RC, may be because of unmeasured confounding factors among younger patients undergoing RC, particularly with urinary diversion status, where increased use of continent diversions among younger patients may explain this association [28]. Similarly, our finding of a decreased risk of UTI among active smokers is likely reflective of unmeasured confounding. We further characterized infectious complications after RC by reporting the time from RC to diagnosis. We demonstrated that the timing of infection after RC is approximately 2 weeks after surgery, at a time when patients have typically been discharged. Indeed, the median length of stay here was 8 days. Based on these data, future targeted follow-up efforts may be
W.P. Parker et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
6
Table 3 Multivariable logistic regression analyses for risks of urinary tract infection, surgical site infection, and sepsis/septic shock within 30 days of RC Variable
UTI
SSI
OR
95% CI
Age (ref 4 70) o50 50–69
– 1.89 1.05
– 1.11–3.21 0.81–1.37
BMI Z 30 kg/m2 (ref o 30) Female Current smoker Diabetes
1.39 0.97 0.69 0.96
CKD (ref eGFR 4 90) CKD II (eGFR 60–89) CKD III (eGFR 30–59) CKD IV (eGFR o30) Steroid use at the time of RC
P value
Sepsis/septic shock
OR
95% CI
– 0.02 0.72
– 1.73 0.82
– 1.07–2.79 0.65–1.05
1.08–1.79 0.71–1.34 0.51–0.94 0.71–1.31
0.01 0.86 0.02 0.81
1.71 1.20 1.07 1.23
– 0.98 0.85 1.06 1.33
– 0.70–1.38 0.58–1.23 0.54–2.07 0.71–2.50
– 0.92 0.38 0.87 0.37
Albumin (ref Z 3.5) o3.5
– 1.24
– 0.82–1.87
Hematocrita (ref normal) Low Very low Receipt of PBT (ref no)
– 1.15 1.17 1.11
Operative time (ref o240 min) 240–359 min 360–479 min Z480 min Year of surgery (ref 2013) 2006 2007 2008 2009 2010 2011 2012
P value
OR
95% CI
P value
– 0.03 0.11
– 1.27 0.88
– 0.75–2.14 0.69–1.11
– 0.38 0.27
1.37–2.14 0.92–1.57 0.82–1.39 0.95–1.59
o0.01 0.19 0.62 0.12
1.39 0.89 0.87 0.99
1.11–1.74 0.67–1.19 0.67–1.13 0.76–1.29
o0.01 0.44 0.30 0.93
– 0.92 0.97 1.53 1.05
– 0.68–1.26 0.69–1.36 0.87–2.69 0.57–1.92
– 0.62 0.85 0.15 0.89
– 0.92 0.94 1.33 1.23
– 0.68–1.26 0.68–1.31 0.76–2.33 0.69–2.21
– 0.61 0.73 0.31 0.48
– 0.31
– 1.08
– 0.74–1.57
– 0.70
– 1.26
– 0.89–1.79
– 0.19
– 0.87–1.51 0.74–1.83 0.86–1.44
– 0.34 0.51 0.41
– 0.90 0.89 1.33
– 0.71–1.15 0.59–1.33 1.06–1.67
– 0.40 0.57 0.02
– 0.96 1.30 1.29
– 0.75–1.23 0.89–1.90 1.03–1.62
– 0.73 0.17 0.03
– 1.20 1.30 2.11
– 0.82–1.75 0.87–1.93 1.37–3.24
– 0.34 0.20 o0.01
– 1.15 1.65 1.63
– 0.82–1.62 1.17–2.34 1.09–2.43
– 0.43 o0.01 0.02
– 0.98 1.09 1.80
– 0.71–1.35 0.78–1.53 1.24–2.60
– 0.91 0.61 o0.01
– 3.90 1.77 1.15 1.17 0.80 1.04 0.76
– 0.68–22.35 0.50–6.28 0.56–2.38 0.70–1.96 0.46–1.39 0.76–1.42 0.55–1.04
– 0.13 0.38 0.71 0.54 0.43 0.81 0.08
– 1.23 1.13 1.02 1.08 0.92 1.07 0.94
– 0.14–11.03 0.32–4.00 0.52–1.99 0.67–1.75 0.57–1.48 0.80–1.43 0.72–1.24
– 0.85 0.85 0.96 0.75 0.72 0.64 0.66
– 1.39 3.35 1.52 1.93 1.07 1.09 0.99
– 0.16–12.33 1.23–9.11 0.83–2.80 1.27–2.94 0.67–1.71 0.82–1.47 0.75–1.30
– 0.77 0.02 0.18 o0.01 0.79 0.55 0.92
CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate. Hematocrit levels were defined as follows: normal (male: Z42% and female: Z38%) Low (male: 32–41% and female: 28%–37%), and very low (male: o32% and female: o28%). a
considered to educate or evaluate patients or both with risk factors for infections to facilitate the early diagnosis and management of these events. We believe that the use of a large national dataset improves the generalizability of our findings. Moreover, in addition to its large sample size and the ability to discern 30-day outcomes, the ACS-NSQIP database is maintained by trained abstractors rather than administrative or index hospital data. Nevertheless, we must acknowledge the limitations of this dataset, and thereby of our analyses. That is, we could not adjust for pathologic variables such as tumor stage or lymph node status that may be important variables to consider with the risk of longer operative times and blood transfusion receipt. Likewise, receipt of neoadjuvant chemotherapy, preoperative radiation therapy, use of bowel preparation, perioperative antibiotic
regimens, and postoperative stent/catheterization were not captured and could, therefore, not be analyzed. Additionally, although we controlled for year of surgery as a means to account for changing surgical practices, the type of urinary diversion and surgical approach (open, minimally invasive) were unable to be specifically identified and evaluated. We, therefore, recognize the unmeasured confounding that these variables may have had on our resultant data. Finally, although we did note a difference in length of stay between patients with and without an infection, we recognize that this finding could be because of either length of stay affecting the incidence of infections or because of the effect of infections on prolonging length of stay. Because of the inability to distinguish these 2 phenomena, we chose not to include length of stay as available in our multivariable models.
W.P. Parker et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
5. Conclusions In conclusion, using a large national dataset, we noted that approximately 25% of patients undergoing RC are diagnosed with an infection within 30 days of surgery. We identified several patient-related risk factors for infection, including age and BMI. Moreover, potentially modifiable risk factors such as receipt of a PBT and prolonged operative time were also found to be associated with infection risk, and may, pending validation in other datasets, thereby represent targets for future quality improvement efforts to reduce the incidence of infection in this setting. Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j. urolonc.2016.07.001.
References [1] Clark PE, Agarwal N, Biagioli MC, et al. Bladder cancer. J Natl Compr Canc Netw 2013;11:446–75. [2] Lavallee LT, Schramm D, Witiuk K, et al. Peri-operative morbidity associated with radical cystectomy in a multicenter database of community and academic hospitals. PLoS One 2014;9:e111281. [3] Shabsigh A, Korets R, Vora KC, et al. Defining early morbidity of radical cystectomy for patients with bladder cancer using a standardized reporting methodology. Eur Urol 2009;55:164–74. [4] Lowrance WT, Rumohr JA, Chang SS, Clark PE, Smith JA Jr., Cookson MS. Contemporary open radical cystectomy: analysis of perioperative outcomes. J Urol 2008;179:1313–8:[discussion 8]. [5] Nazmy M, Yuh B, Kawachi M, et al. Early and late complications of robot-assisted radical cystectomy: a standardized analysis by urinary diversion type. J Urol 2014;191:681–7. [6] Shigemura K, Tanaka K, Matsumoto M, et al. Post-operative infection and prophylactic antibiotic administration after radical cystectomy with orthotopic neobladder urinary diversion. J Infect Chemother 2012;18:479–84. [7] Kim SP, Shah ND, Karnes RJ, et al. The implications of hospital acquired adverse events on mortality, length of stay and costs for patients undergoing radical cystectomy for bladder cancer. J Urol 2012;187:2011–7. [8] Konety BR, Allareddy V. Influence of post-cystectomy complications on cost and subsequent outcome. J Urol 2007;177:280–7:[discussion 7]. [9] Skolarus TA, Jacobs BL, Schroeck FR, et al. Understanding hospital readmission intensity after radical cystectomy. J Urol 2015;193:1500–6. [10] Johar RS, Hayn MH, Stegemann AP, et al. Complications after robotassisted radical cystectomy: results from the International Robotic Cystectomy Consortium. Eur Urol 2013;64:52–7. [11] Pariser JJ, Anderson BB, Pearce SM, et al. The effect of broader, directed antimicrobial prophylaxis including fungal coverage on perioperative
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24] [25]
[26]
[27]
[28]
7
infectious complications after radical cystectomy. Urol Oncol 2015;34:e9121.e14. American College of Surgeons National Surgical Quality Improvement Program. User guide for the 2013 ACS NSQIP Participant Use Data File (PUF); 2013. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: executive summary. Expert panel on the identification, evaluation, and treatment of overweight in adults. Am J Clin Nutr 1998;68:899–917. Chang SS, Jacobs B, Wells N, Smith JA Jr., Cookson MS. Increased body mass index predicts increased blood loss during radical cystectomy. J Urol 2004;171:1077–9. Kratz A, Ferraro M, Sluss PM, Lewandrowski KB. Case records of the Massachusetts General Hospital. Weekly clinicopathological exercises. Laboratory reference values. N Engl J Med. 2004;351: 1548–1563. Mullen JT, Moorman DW, Davenport DL. The obesity paradox: body mass index and outcomes in patients undergoing nonbariatric general surgery. AnnSurg 2009;250:166–72. Shah DK, Vitonis AF, Missmer SA. Association of body mass index and morbidity after abdominal, vaginal, and laparoscopic hysterectomy. Obstet Gynecol 2015;125:589–98. Lee SJ, Seo H, Kim HC, et al. Effect of intraoperative red blood cell transfusion on postoperative complications after open radical cystectomy: old versus fresh stored blood. Clin Genitourin Cancer 2015;13:581–7. Acheson AG, Brookes MJ, Spahn DR. Effects of allogeneic red blood cell transfusions on clinical outcomes in patients undergoing colorectal cancer surgery: a systematic review and meta-analysis. Ann Surg 2012;256:235–44. Al-Refaie WB, Parsons HM, Markin A, Abrams J, Habermann EB. Blood transfusion and cancer surgery outcomes: a continued reason for concern. Surgery 2012;152:344–54. Wang YL, Jiang B, Yin FF, et al. Perioperative blood transfusion promotes worse outcomes of bladder cancer after radical cystectomy: a systematic review and meta-analysis. PLoS One 2015;10: e0130122. Linder BJ, Frank I, Cheville JC, et al. The impact of perioperative blood transfusion on cancer recurrence and survival following radical cystectomy. Eur Urol 2013;63:839–45. Abel EJ, Linder BJ, Bauman TM, et al. Perioperative blood transfusion and radical cystectomy: does timing of transfusion affect bladder cancer mortality? Eur Urol 2014;66:1139–47. Clevenger B, Mallett SV, Klein AA, Richards T. Patient blood management to reduce surgical risk. Br J Surg 2015;102:1325–37. Spahn DR, Moch H, Hofmann A, Isbister JP. Patient blood management: the pragmatic solution for the problems with blood transfusions. Anesthesiology 2008;109:951–3. Daley BJ, Cecil W, Clarke PC, Cofer JB, Guillamondegui OD. How slow is too slowCorrelation of operative time to complications: an analysis from the Tennessee Surgical Quality Collaborative. J Am Coll Surg 2015;220:550–8. Qin C, de Oliveira G, Hackett N, Kim JY. Surgical duration and risk of urinary tract infection: an analysis of 1,452,369 patients using the National Surgical Quality Improvement Program (NSQIP). Int J Surg 2015;20:107–12. Parker WP, Toussi A, Tollefson MK, et al. Risk factors and microbial distribution of urinary tract infections following radical cystectomy. Urology 2016. [Epub ahead of print].