Methods - Within the National Inpatient Sample database, we identified 91,618 RP patients treated between 2008 and 2015. The Johns Hopkins Adjusted Clinical Groups frailty-defining indicator was applied, and we examined the rates of frailty over time, as well as its effect on overall complications, major complications, nonhome-based discharge, length of stay (LOS), and total hospital charges (THCs). Time trends and multivariable logistic, Poisson and linear regression models were applied.
Results - Overall, 12,185 (13.3%) patients were frail. Rates of frail patients increased over time (from 10.3 to 18.2%; p < 0.001). Frail patients had higher rates of overall complications (16.6 vs. 8.6%), major complications (4.9 vs. 2.6%), nonhome-based discharge (5.9 vs. 5%), longer LOS (2 vs. 1), and higher THCs ($37,186 vs. $35,241) (all p < 0.001). Moreover, frailty was an independent predictor of overall complications (OR: 1.95), major complications (OR: 1.76), nonhome-based discharge (OR: 1.20), longer LOS (RR: 1.19), and higher THCs (RR: $3160) (all p < 0.001). Of frail patients, 10,418 (85.5%) neither exhibited body mass index ≥ 30 nor Charlson comorbidity index ≥ 2.
Conclusions - On average, every seventh RP patient is frail and that proportion is on the rise. Frail individuals are at higher risk of adverse short-term postoperative outcomes, that cannot be predicted by other risk factors, such as obesity or comorbidities.
Radical prostatectomy (RP) for localized prostate cancer (PCa) is a surgical procedure not devoid of complications,1-3 and ideal predictors of suboptimal short-term postoperative outcomes after RP have not yet been identified. Nonetheless, body mass index (BMI) ≥ 30, as well as Charlson comorbidity index (CCI) ≥ 2, represents well-established albeit suboptimal indicators of adverse outcomes after major surgery.4-10 In consequence, the search for more accurate and independent indicators of unfavorable outcomes is ongoing. Indeed, preoperative counseling aimed at identifying patients at higher risk of adverse short-term outcomes after surgery may be of particular importance in PCa patients with localized disease, since alternative treatment modalities may be used in this patient population to avoid RP-related complications.11,12 In that regard, frailty13-16 qualifies for being considered as a complementary predictor of adverse outcomes, since it relies on a different definition than either obesity, defined as BMI ≥ 30 and/or comorbidities, defined as CCI ≥ 2. Indeed, frailty is a state of health described by a reduced physical reserve and increased vulnerability to stressors,17 and it accounts for cognitive, functional, and social impairments. However, its ability to predict adverse short-term outcomes after RP has not been properly explored. Moreover, it is unknown what proportion of PCa patients undergoes RP, despite being considered frail. Based on this unmet need, we addressed the ability of frailty to predict adverse short-term outcomes and total hospital charges (THCs) after RP. Specifically, we relied on the Johns Hopkins Adjusted Clinical Groups (ACG) frailty-defining diagnoses indicator18 to identify frail patients.
The current study relied on the National Inpatient Sample (NIS) database19,20 (2008–2015) that is composed of longitudinal hospital inpatient databases from the Healthcare Cost and Utilization Project family including 20% of United States inpatient hospitalizations.21 Institutional ethical board had reviewed the research project, and it meets the requirements for protection of human subjects. Within the NIS database, we focused on patients older than 18 years with the primary diagnosis of nonmetastatic PCa (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes 185) treated with RP (ICD-9 codes 604, 605, 6062) from 2008 to 2015. Secondary procedure codes were used to identify lymph node dissection (ICD-9-CM code 40.3 or 40.5). Overall, 91,618 assessable patients were identified.
Frailty was defined according to the Johns Hopkins ACG18 frailty-defining diagnoses indicator.22 This indicator was designed and validated for the research of frailty-related outcomes and resource use using administrative data.23,24 The ACG frailty indicator is based on ten clusters of frailty-defining diagnoses and allows to account for cognitive, functional, and social impairments. Indeed, it encompasses dementia, vision impairment, malnutrition, urinary and fecal incontinence, difficult in walking, falls and social support needs. Frailty was calculated using patient ICD-9 codes available in the NIS, as previously reported,14,23,25 and categorized as frail vs. non-frail. Obesity was defined as BMI ≥ 30 kg/m2 , according to the World Health Organization definition.26 Diagnostic codes were 278.00 (obesity, unspecified), 278.01 (morbid obesity), V85.30–V85.39 (Obesity, BMI 30–39), and V85.41–V85.45 (Obesity, BMI > 40).4 CCI was defined according to the Deyo definition27 and was categorized as CCI 0–1 vs. CCI ≥ 2. Frailty-defining ICD-9 codes were different from ICD-9 codes defining components of CCI, with the exception of dementia ICD-9 code, that applied to either definition. Adjustment variables included continuously coded age at surgery, ethnicity (Caucasian, African-American, and others), year of surgery, income and insurance status (private insurance, Medicare, Medicaid, and other (self-pay)), hospital region (Northeast, Midwest, South, West), hospital teaching vs. nonteaching status, hospital annual volume (low, medium, and high), surgical approach (open vs. robotic), and lymph node dissection (performed vs. nonperformed). According to previously established methodology,28 overall complications and major complications (Supplementary Table 1) were defined using secondary ICD9-CM diagnostic codes and represented the sum of intraoperative and all postoperative complications (cardiac, respiratory, vascular, operative wound, genitourinary, gastrointestinal, infectious, blood transfusions, miscellaneous medical, and miscellaneous surgical). Nonhome-based discharge was defined according to the final disposition at discharge.29 Length of stay (LOS) was calculated by subtracting the admission date from the discharge date. Finally, inflation-adjusted THCs were defined according to NIS methodology.30
Supp. Table 1: ICD-9 codes from National Inpatient Database according to each complication.
Descriptive statistics included frequencies and proportions for categorical variables. Means, medians, and ranges were reported for continuously coded variables. Four sets of analyses were performed. First, temporal trend analyses examined estimate annual percentage changes (EAPC) in frail patients, patients with BMI ≥ 30, and patients with CCI ≥ 2 who underwent RP during the study span (2008–2015). Second, chi-square test, t-test, and Kruskal–Wallis tests were performed to examine the statistical significance in proportions’, means’, and medians’ differences between the three groups of interest: (1) frail vs. non-frail; (2) BMI ≥ 30 vs. <30; (3) CCI ≥ 2 vs. 0–1. Third, five separate endpoints were addressed in multivariable analyses, namely, overall complications, major complications, nonhome-based discharge, LOS, and THCs. Specifically, multivariable logistic regression models focused on overall complications, major complications, and nonhomebased discharge. Multivariable Poisson regression model focused on LOS. Finally, multivariable linear regression model on THCs. Fourth, all analyses were repeated in the subgroup of patients older than 70 years.
All multivariable models relied on weighting using a generalized estimating equation function to provide more accurate national estimates based on NIS-provided weights.30 All models were then adjusted for clustering at hospital level, as well as for age, CCI, frailty, BMI, year of surgery, ethnicity, insurance status, teaching status, hospital volume, region, income, surgical approach, and lymph node dissection. Within each multivariable model all the three predictors of interest were considered simultaneously along with all the covariates without either forward or backward variable selection.
For all statistical analyses, R software environment for statistical computing and graphics (version 3.4.3) was used. All tests were two-sided with a level of significance set at p < 0.05.
General characteristics of the study populations and trends over time
Overall, 91,618 patients underwent RP between 2008 and 2015. Of these, 12,185 (13.3%) were frail, 7225 (7.9%) had BMI ≥ 30, and 3168 (3.5%) had CCI ≥ 2. Most frail patients did not exhibit other baseline risk factors for adverse short-term postoperative outcomes after RP, namely, BMI ≥ 30 or CCI ≥ 2 (n = 10,418; 85.5%) (Fig. 1).
Fig. 1 Venn diagram. Venn diagram depicting proportions of frail (yellow), body mass index (BMI) ≥ 30 kg/m2 (blue), and Charlson comorbidity index (CCI) ≥ 2 (red) patients with nonmetastatic prostate cancer treated with radical prostatectomy, within the National Inpatient Database (2008–2015). (Color figure online.).
Baseline characteristics of our study cohort are summarized in Table 1. More detailed stratification of the population according to the presence or absence of frailty revealed statistically significant albeit clinically not meaningful age differences (63 vs. 62 years; p < 0.001). Similarly, a very small minority of both frail and non-frail patients exhibited BMI ≥ 30 (7.5 vs. 10.4%; p < 0.001). Finally, an equally marginal proportion of frail and non-frail patients harbored CCI ≥ 2 (5.2 vs. 3.2%; p < 0.001).
Table 1. Descriptive characteristics of 91,618 nonmetastatic prostate cancer patients treated with radical prostatectomy, identified within the nationwide inpatient sample database between 2008 and 2015, stratified according to frailty (frail vs. non-frail).
Temporal trend analyses revealed increasing rates of frail RP patients: 10.3% in 2008 vs. 18.2% in 2015 (EAPC: +7.6%; p < 0.001). Similarly, rates of RP patients with BMI ≥ 30 (from 5.5 to 10.4%; +7.8%; p < 0.001) or with CCI ≥ 2 (from 3.1 to 3.9%; +3.8%; p < 0.001) increased over time (Fig. 2).
Fig. 2 Time trends. Time trends of frail (yellow), body mass index (BMI) ≥ 30 kg/m2 (blue), and Charlson comorbidity index (CCI) ≥ 2 (red) nonmetastatic prostate cancer patients treated with radical prostatectomy, within National Inpatient Database (2008–2015). EAPC estimate annual percentage changes. (Color figure online.).
Unadjusted rates of short-term postoperative complications, length of stay, and THCs according to frailty
Frail patients exhibited higher rates of overall complications (16.6 vs. 8.6%; p < 0.001), major complications (4.9 vs. 2.6%; p < 0.001), and nonhome-based discharge (5.9 vs. 5.0%; p < 0.001). LOS was longer in frail vs. non-frail patients (2 [IQR: 1–3] vs. 1 [IQR: 1–2] days). Finally, THCs were also higher in frail patients ($37,183 [IQR: $27,112–$52,810] vs. $35,241 [IQR: $25,195–$49,411]), compared with non-frail patients. For comparison purposes, the effect of BMI ≥ 30, and that of CCI ≥ 2 also affected all five examined endpoints, in a fashion similar to frailty (Table 2).
Table 2 Crude postoperative outcomes and total hospital charges recorded in 91,618 radical prostatectomy patients, stratified according to frailty status (frail vs. non-frail), body mass index (≥30 vs. <30), and Charlson comorbidity index score categories (≥2 vs. 0–1), identified within the Nationwide Inpatient Sample database between 2008 and 2015.
Multivariable analyses testing the effect of frailty on short-term postoperative complications, length of stay, and THCs
Multivariable analyses testing the effect of frailty, obesity, and CCI ≥ 2 are reported in Table 3. In multivariable logistic regression models, after adjustment for the effect of all covariates, including BMI and CCI, frailty was independently associated with the overall complications (OR: 1.95; CI 1.84–2.06), major complications (OR: 1.76; CI 1.59–1.94), and nonhome-based discharge (OR: 1.20; CI 1.11–1.30) rates (all p < 0.001). In multivariable Poisson regression models, which were also adjusted for the effect of all covariates, including BMI and CCI, frailty was associated with the LOS (RR: 1.20, CI 1.15–1.21; p < 0.01). Finally, in linear regression models, that adjusted for the effect of all covariates, including BMI and CCI, frailty was also associated with higher THCs (RR: 3160, CI 2519–3800; p < 0.001). For comparison purposes, BMI ≥ 30 and CCI ≥ 2 independently predicted worse short-term postoperative complications, as well as longer LOS and higher THCs (Table 3). In this regard, CCI ≥ 2 represented the most important predictor of overall complications (HRs: 2.02 vs. 1.95 vs. 1.58), major complications (HRs: 2.66 vs. 1.76 vs. 1.96), nonhome-based discharge (HRs: 1.54 vs. 1.20 vs. 1.28), longer LOS (RRs: 1.27 vs. 1.19 vs. 1.11), and higher THCs (RRs: 4877 vs. 3160 vs. 3321), compared with, respectively, frailty and BMI ≥ 30.
Table 3 Separate multivariable regression models, predicting overall complications, major complications, nonhome-based discharge, length of stay, and total hospital charges in patients treated with radical prostatectomy for nonmetastatic prostate cancer within the National Inpatient Sample database (2008–2015).
The effect of frailty in patients older than 70 years: unadjusted and multivariable analyses
Of 11,570 (12.6%) patients older than 70 years treated with RP, 1999 (16.4%) were classified as frail. Similar to their younger counterparts, a minority of frail patients older than 70 exhibited a BMI ≥ 30 (8.4%), and a marginal proportion of these patients harbored a CCI ≥ 2 (5.9%). After stratification according to frailty, overall complications (19.9 vs. 11.5%), major complications (5.8 vs. 3.4%), LOS (2 vs. 1), and THCs ($36,389 vs. $37,890) were less favorable in frail patients (all p < 0.001). Moreover, frailty was an independent predictor of overall complications (OR: 1.87; CI 1.64–2.12), major complications (OR: 1.72; CI 1.37–2.15), longer LOS (RR: 1.19; CI 1.10–1.28), and higher THCs (RR: 3085; CI 1500–4670) (all p < 0.001). For comparison purposes, the effect of a BMI ≥ 30, and that of a CCI ≥ 2 also affected all five examined endpoints, in a fashion similar to frailty (data not shown).
Frailty may represent an indicator of adverse short-term postoperative outcomes after RP, but its effect has not been properly explored. Moreover, it is unknown what proportion of patients qualifies as frail according to the established Johns Hopkins definition of frailty.18 Finally, it is also unknown what proportion of frail patients are devoid of other unfavorable characteristics, such as BMI ≥ 30 or multiple comorbidities. We examined these endpoints within a large contemporary cohort of RP patients. Our analyses resulted in several important observations.
First, more than every seventh RP patient was frail. Conversely, only 7.9% of patients exhibited BMI ≥ 30 and 3.5% had CCI ≥ 2. Interestingly, the definition of frailty neither overlapped with BMI ≥ 30 nor with CCI ≥ 2, except for a small or, respectively, marginal proportion of patients. Specifically, only 1.2% of RP patients were frail and also had BMI ≥ 30. Moreover, only 0.5% were frail and also had CCI ≥ 2. Finally, only 0.1% were frail and also had BMI ≥ 30 and CCI ≥ 2. These observations demonstrate that frailty can be qualified as a distinct and unrelated metric, relative to well-established risk factors, namely, BMI ≥ 30 and CCI ≥ 2, in the context of RP. Moreover, it is also noteworthy that the rate of patients harboring CCI ≥ 2 was very similar between frail and non-frail patients (5.2 vs. 3.2%). In consequence, this observation also highlights the role and the importance of preoperative frailty assessment, since frailty and CCI identify different patient populations at higher risk of developing postoperative complications.
Second, it is interesting to note that the rate of frail patients who underwent RP increased over time from 10% in 2008 to 18% in 2015. However, it is also of note that the rates of RP patients with BMI ≥ 30 (5.5–10.4%) and with CCI ≥ 2 (3.1–3.6%) also increased over time but to a less pronounced extent than what was recorded for frailty. These observations indicate that the proportion and the importance of frailty are increasing in the most contemporary RP patients. In consequence, these results may sensitize the urological community regarding increasing rates of this underestimated risk factor, which may be identified during preoperative counseling.
Third, frail individuals exhibited higher rates of overall complications (16.6 vs. 8.6%), major complications (2.6 vs. 4.9%), and longer LOS (2 vs. 1 days) compared with non-frail patients. Moreover, frailty was also associated with higher THCs ($37,183 vs. $35,241). Finally, a higher proportion of frail patients required nonhome-based discharge (5.9 vs. 5.0%) after RP compared with non-frail patients. These differences resulted in an independent predictor status for frailty, even after multivariable adjustments for the effects of standard variables, such as BMI and CCI. However, it is noteworthy that CCI ≥ 2 represented the most important predictor of all examined short-term postoperative outcomes compared with frailty or BMI ≥ 30 kg/m2. Indeed, despite frailty included a higher proportion of patient at higher risk of developing postoperative complications, CCI ≥ 2 exhibited the highest HRs, even after the multivariable adjustment for frailty and BMI. This observation highlights the importance of preoperative comorbidity assessment, which may rely on both frailty and CCI parameters to correctly identify the majority of patients at higher risk of adverse short-term postoperative outcomes and to carefully evaluate their treatment management.
Last but not least, frailty exhibited the very same effects with the same direction and magnitude in the subgroup of patients older than 70 years. These observations imply that frailty is an equally valid indicator of adverse short-term postoperative outcomes even in patients older than 70, where the proportion of frail individuals is higher compared with the younger counterparts (16.4 vs. 12.7%).
Our results are in agreement with Suskind et al.,31 who reported that frailty was strongly associated with both major and minor complications, in patients undergoing urological surgery. Moreover, Levy et al.32 also demonstrated an increased risk of major complications and 30-day mortality in frail patients treated with RP when compared with non-frail patients. However, Levy et al.32 defined frailty according to the Canadian Study of Health and Aging Frailty Index. Although this frailty index is a multidimensional score, it mostly relies on comorbidity domains and, therefore, may less accurately capture the effect of frailty than the Johns Hopkins definition, which is substantially more multifaceted than the Canadian Study of Health and Aging Frailty Index. Indeed, frailty differs from comorbidity since frailty is a state of health described by a reduced physical reserve and increased vulnerability to stressors.17 Conversely, the Johns Hopkins ACG frailtydefining diagnosis indicator-based definition of frailty may allow to capture the different effect of comorbidity and frailty since it accounts for cognitive, functional, and social impairments without overlapping with CCI.
Taken together, our results revealed four novel findings. First, frailty affects an important proportion of RP patients. Second, frailty does not overlap with other well-known risk factors for adverse short-term postoperative outcomes, such as BMI ≥ 30 or CCI ≥ 2. Third, rates of frail RP patients are increasing over time, from one out of ten to one out of five RP patients in the most contemporary study years. Finally, regardless of age and after adjustment for BMI and CCI, frailty represents an important risk factor for adverse shortterm postoperative outcomes, when the Johns Hopkins ACG frailty-defining indicator is applied. In consequence, frailty appears to exhibit all the characteristics worthy of a novel indicator of adverse short-term postoperative outcomes after RP, even in patients that neither have BMI ≥ 30 nor exhibit multiple comorbidities.
Despite its novelty, our study is not devoid of limitations. First, within the NIS database, complications were limited to in-hospital events. In consequence, delayed complications, as well as the readmission rates, could not be examined. Second, information on performance status, ASA score, as well as laboratory values is not available within the NIS database. Moreover, the use of a binary definition of frailty does not allow to evaluate the effect of different degrees of frailty. Third, the NIS is an observational retrospective database relying on ICD-9 codes for assessment of secondary diagnostic codes, which may be subject to potential coding biases. For example, patients with BMI ≥ 30 kg/m2 who did not carry a corresponding ICD-9 code for obesity may have not been identified as obese patients. Moreover, the NIS database is based on ICD-9 codes assigned during admission. In consequence, if ICD-9 codes for frailty assessment were not assigned during admission, frail patients may have been underreported. Finally, the NIS database does not allow us to distinguish between ICD-9 codes assigned present at admission from ICD-9 codes occurred during admission. Nonetheless, the NIS estimates are considered to be precise and accurate.30 Fourth, the cost analyses were based on hospital-related charges that may not be reflective at all of other health-related costs, such as readmissions and post-discharge complication rates. Finally, since the NIS database did not provide tumor characteristics, such as stage and grade, we could not adjust our analyses for these variables.
On average, every seventh RP patient is frail and that proportion is on the rise. Frail individuals are at higher risk of adverse short-term postoperative outcomes that cannot be predicted by other risk factors, such as obesity or comorbidities. For this reason, more attentive preoperative counseling is needed in this patient population to correctly identify patients who may qualify as frail and, in consequence, are at higher risk of developing adverse outcomes after surgery.
Conflict of interest - The authors declare that they have no conflict of interest.
Written by: Giuseppe Rosiello,1,2 Carlotta Palumbo,1,3 Sophie Knipper,1,4 Marina Deuker,1,5 Lara Franziska Stolzenbach,1,4 Zhe Tian,1 Giorgio Gandaglia,2 Nicola Fossati,2 Francesco Montorsi,2 Shahrokh F. Shariat,6,7 Fred Saad,1 Alberto Briganti,2 Pierre I. Karakiewicz1
1. Cancer Prognostics and Health Outcomes Unit, Division of Urology, University of Montreal Health Center, Montreal, QC, Canada
2. Department of Urology and Division of Experimental Oncology, Urological Research Institute (URI), IRCCS San Raffaele Scientific Institute, Milan, Italy
3. Urology Unit, ASST Spedali Civili of Brescia, Department of Medical and Surgical Specialties, Radiological Science and Public Health, University of Brescia, Brescia, Italy
4. Martini-Klinik Prostate Cancer Center, University Hospital Hamburg-Eppendorf, Hamburg, Germany
5. Department of Urology, University Hospital Frankfurt, Frankfurt, Germany
6. Department of Urology, Comprehensive Cancer Center, Medical University of Vienna, Vienna, Austria 7 Institute of Urology and Reproductive Health, I.M. Sechenov First Moscow State Medical University, Moscow, Russia
1. Canda AE, Tilki D, Mottrie A. Rectal injury during radical prostatectomy: focus on robotic surgery. Eur Urol Oncol. 2018;1: 507–9.
2. Johansson E, Steineck G, Holmberg L, Johansson J-E, Nyberg T, Bill-Axelson A, et al. Quality of life after radical prostatectomy or watchful waiting with or without androgen deprivation therapy: the SPCG-4 randomized trial. Eur Urol Oncol. 2018;1:134–42.
3. Gandaglia G, Bravi CA, Dell’Oglio P, Mazzone E, Fossati N, Scuderi S, et al. The impact of implementation of the European Association of Urology Guidelines Panel recommendations on reporting and grading complications on perioperative outcomes after robot-assisted radical prostatectomy. Eur Urol. 2018;74:4–7.
4. Knipper S, Mazzone E, Mistretta FA, Palumbo C, Tian Z, Briganti A, et al. Impact of obesity on perioperative outcomes at roboticassisted and open radical prostatectomy: results from the National Inpatient Sample. Urology. 2019;133:135–44.
5. Monn MF, Jaqua KR, Calaway AC, Mellon MJ, Koch MO, Boris RS. Impact of obesity on wound complications following radical prostatectomy is mitigated by robotic technique. J Endourol. 2016;30:890–5.
6. Vidal AC, Howard LE, Sun SX, Cooperberg MR, Kane CJ, Aronson WJ, et al. Obesity and prostate cancer-specific mortality after radical prostatectomy: results from the Shared Equal Access Regional Cancer Hospital (SEARCH) database. Prostate Cancer Prostatic Dis. 2017;20:72–8.
7. Maj-Hes AB, Mathieu R, Özsoy M, Soria F, Moschini M, Abufaraj M, et al. Obesity is associated with biochemical recurrence after radical prostatectomy: a multi-institutional extended validation study. Urol Oncol. 2017;35:460.e1–8.
8. Sivaraman A, Ordaz Jurado G, Cathelineau X, Barret E, Dell’Oglio P, Joniau S, et al. Older patients with low Charlson score and high-risk prostate cancer benefit from radical prostatectomy. World J Urol. 2016;34:1367–72.
9. Lee JY, Kang HW, Rha KH, Cho NH, Choi YD, Hong SJ, et al. Age-adjusted Charlson comorbidity index is a significant prognostic factor for long-term survival of patients with high-risk prostate cancer after radical prostatectomy: a Bayesian model averaging approach. J Cancer Res Clin Oncol. 2016;142:849–58.
10. Froehner M, Koch R, Litz RJ, Oehlschlaeger S, Twelker L, Hakenberg OW, et al. Detailed analysis of Charlson comorbidity score as predictor of mortality after radical prostatectomy. Urology. 2008;72:1252–7.
11. Briganti A, Fossati N, Catto JWF, Cornford P, Montorsi F, Mottet N, et al. Active surveillance for low-risk prostate cancer: the European Association of Urology position in 2018. Eur Urol. 2018;74:357–68.
12. Wilt TJ, Brawer MK, Jones KM, Barry MJ, Aronson WJ, Fox S, et al. Radical prostatectomy versus observation for localized prostate cancer. N Engl J Med. 2012;367:203–13.
13. Sandini M, Pinotti E, Persico I, Picone D, Bellelli G, Gianotti L. Systematic review and meta-analysis of frailty as a predictor of morbidity and mortality after major abdominal surgery. BJS Open. 2017;1:128–37.
14. McIsaac DI, Bryson GL, van Walraven C. Association of frailty and 1-year postoperative mortality following major elective noncardiac surgery: a population-based cohort study. JAMA Surg. 2016;151:538–45.
15. Palumbo C, Knipper S, Pecoraro A, Rosiello G, Luzzago S, Deuker M, et al. Patient frailty predicts worse perioperative outcomes and higher cost after radical cystectomy worse radical cystectomy outcomes in frails. Surg Oncol. 2019;32:8–13.
16. Taylor BL, Xia L, Guzzo TJ, Scherr DS, Hu JC. Frailty and greater health care resource utilization following major urologic oncology surgery. Eur Urol Oncol. 2019;2:21–7.
17. Clegg A, Young J, Iliffe S, Rikkert MO, Rockwood K. Frailty in elderly people. Lancet Lond Engl. 2013;381:752–62.
18. The Johns Hopkins ACG® system excerpt from version 11.0 technical reference guide. 2015. p. 36. https://www.healthpartners.com/ucm/groups/public/@hp/@public/documents/documents/cntrb_035024.pdf.
19. Barashi NS, Pearce SM, Cohen AJ, Pariser JJ, Packiam VT, Eggener SE. Incidence, risk factors, and outcomes for rectal injury during radical prostatectomy: a population-based study. Eur Urol Oncol. 2018;1:501–6.
20. Zaffuto E, Bandini M, Moschini M, Leyh-Bannurah S-R, Gazdovich S, Dell’Oglio P, et al. Location of metastatic bladder cancer as a determinant of in-hospital mortality after radical cystectomy. Eur Urol Oncol. 2018;1:169–75.
21. National (Nationwide) Inpatient Sample (NIS). 2019. https://www.hcup-us.ahrq.gov/news/exhibit_booth/nis_brochure.jsp.
22. Development and evaluation of the Johns Hopkins University risk adjustment models for Medicare+Choice plan payment. Johns Hopkins ACG® System. 2019. https://www.hopkinsacg.org/ document/development-and-evaluation-of-the-johns-hopkinsuniversity-risk-adjustment-models-for-medicarechoice-plan-payment/.
23. Kim DH, Schneeweiss S. Measuring frailty using claims data for pharmacoepidemiologic studies of mortality in older adults: evidence and recommendations. Pharmacoepidemiol Drug Saf. 2014; 23:891–901.
24. Sternberg SA, Bentur N, Abrams C, Spalter T, Karpati T, Lemberger J, et al. Identifying frail older people using predictive modeling. Am J Manag Care. 2012;18:e392–7.
25. Nieman CL, Pitman KT, Tufaro AP, Eisele DW, Frick KD, Gourin CG. The effect of frailty on short-term outcomes after head and neck cancer surgery: frailty and outcomes in head and neck surgery. Laryngoscope. 2018;128:102–10.
26. WHO | Obesity. 2019. https://www.who.int/topics/obesity/en/.
27. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613–9.
28. Palumbo C, Knipper S, Dzyuba-Negrean C, Pecoraro A, Rosiello G, Tian Z, et al. Complication rates, failure to rescue and in-hospital mortality after cytoreductive nephrectomy in the older patients. J Geriatr Oncol. 2019:S1879406819301961. (in press).
29. Adedayo P, Resnick K, Singh S. Preoperative frailty is a risk factor for non-home discharge in patients undergoing surgery for endometrial cancer. J Geriatr Oncol. 2018;9:513–5.
30. HCUP-US NIS Overview. 2019. https://www.hcup-us.ahrq.gov/ nisoverview.jsp.
31. Suskind AM, Walter LC, Jin C, Boscardin J, Sen S, Cooperberg MR, et al. Impact of frailty on complications in patients undergoing common urological procedures: a study from the American College of Surgeons National Surgical Quality Improvement database. BJU Int. 2016;117:836–42.
32. Levy I, Finkelstein M, Bilal KH, Palese M. Modified frailty index associated with Clavien-Dindo IV complications in robot-assisted radical prostatectomies: a retrospective study. Urol Oncol. 2017; 35:425–31.
Read an Editorial by Henry Woo, MBBS, DMedSc, FRACS: Adverse Short-Term Postoperative Outcomes in Patients Treated with Radical Prostatectomy Predicted by Preoperative Frailty - Editorial