Frailty is a predictor of moderate to severe pain after robot‐assisted laparoscopic prostatectomy: A case‐control study (FRAP study)

Abstract Objective To investigate the association of pain with frailty in patients with localized prostate cancer (PC) who underwent robot‐assisted laparoscopic radical prostatectomy (RARP). Materials and Methods Between January 2017 and June 2019, we prospectively evaluated the geriatric 8 (G8) score, simplified frailty index (sFI), and numerical rating scale (NRS) of 154 patients with localized PC who underwent RARP at our institution. NRS was measured on preoperative day 0, postoperative days 1, 2, 3, and at discharge. Moderate to severe pain was defined as NRS ≥ 5, whereas frailty was defined as G8 ≤ 14. The primary objectives of this study were to investigate the effects of moderate to severe pain (NRS ≥ 5) on frailty, postoperative complications, and the use of analgesics after RARP. Our secondary objectives were the effect of frailty on postoperative complications and the use of analgesics. Results The median age of participants was 69 years. Of 154 patients, 37 (24%) and 61 (40%) were classified to have NRS ≥ 5 and G8 ≤ 14, respectively. Patients with NRS > 5 presented significantly association with G8 < 14, whereas they did not show the association with sFI, complication, or analgesics. Multivariate logistic regression analysis showed that G8 ≤ 14 was significantly associated with NRS ≥ 5. Frailty was not significantly associated with postoperative complications and analgesics. Conclusions Frailty was significantly associated with moderate to severe pain after RARP, and might be a potential predictor of postoperative pain. Frail patients require individual care to avoid painful experiences.


| INTRODUC TI ON
Reflecting an increased focus on elderly patients with urological cancers, interest in frailty has also gained increasing prominence. [1][2][3][4][5][6] Although frailty screening in patients with cancer has clinical value, there is not enough evidence available for the optimal screening tool and clinical implication of frailty on postoperative outcomes in prostate cancer (PC). [7][8][9] PC is the most frequently diagnosed male cancer in the United States, Europe, and Japan, [10][11][12][13][14][15] and the treatment requires concerns of frailty because health status including frailty should influence the treatment outcomes. 16,17 Several simple assessment tools are available, including simplified frailty index (sFI) 18 and the geriatric 8 (G8) screening tool. 19 A recent systematic review found that although G8 was one of the most robust screening tools currently available, it was also associated with treatment-related complications. 20 As frailty was suggested to be associated with unfavorable outcomes in urologic surgeries, [7][8][9]18 we hypothesized that patients with frailty might experience greater pain compared with patients without frailty.
However, no evidence is available for the effect of frailty on postoperative pain in patients with localized PC who have undergone robot-assisted laparoscopic radical prostatectomy (RARP).
Therefore, we performed a prospective observational study investigating the effect of G8 on postoperative pain in this particular group of patients.

| Ethics statement
This study was performed in accordance with the ethical standards established by the Declaration of Helsinki and was approved by the ethics review board (authorization number: 2014-297). All participants provided written or verbal informed consent. The study was registered on the UMIN-CTR (UMIN000038969).

| Study population and treatment protocol
This prospective observational study planned to enroll 180 patients within 30 months based on the annual number of RARPs performed in our hospital, accounting for a 15% withdrawal/refusal rate. No statistical sample size calculation was performed, because no previous evidence has been reported on the effect of frailty on postoperative pain. We enrolled 181 patients with localized PC who were treated at the Hirosaki University Hospital between January 2017 and June 2019. The inclusion criteria called for patients with localized PC who were treated with RARP. Procedures of RARP and standard pelvic lymph node dissection were performed using the basic technique described previously. 21,22 Extended pelvic lymph node dissection was not performed in this study. The exclusion criteria were as follows: (1) patients who could not be evaluated for frailty and postoperative pain and (2) those treated with open radical prostatectomy.
All RARPs were mainly performed five expert surgeons who had enough experience of all urological surgeries and patients were equally distributed among surgeons.

| Patient variables
The patient variables evaluated at diagnosis were age, sex, Eastern

| Assessment of frailty
We assessed frailty using the G8 screening tool and sFI ( Figure S1).
G8, which includes eight items in multiple geriatric assessment domains, was administered at the initial outpatient clinic visit. The G8 score ranges from 0 to 17, with a frailty cutoff of ≤14. 19 We evaluated sFI including comorbidities (HTN, DM, CVD, CRD) and IADL. Nonfrail patients in sFI displayed none or one characteristic (score 0-1), whereas frail patients showed two or more characteristics (score 2-5). IADL was assessed using the Tokyo Metropolitan Institute of Gerontology Index of Competence (TMIG) index, 23 with TMIG scores ranging from 1 to 13. Functional capacity impairment (IADLlow) was assumed when a patient scored <11 (an IADL of < 80%). 4

| Evaluation of pain and use of analgesics
We used a numerical rating scale (NRS) for the quantitative evaluation of pain. NRS was measured on preoperative day 0, postoperative days 1, 2, 3, and at discharge as a self-reported survey. 24 The baseline NRS score was reported was prior to admission of any additional pain relief. Moderate to severe pain was defined as NRS ≥ 5. 25, 26 We developed our original analgesic score for the postoperative use of pain relievers. We used acetaminophen (1000 mg, intravenously), loxoprofen sodium hydrate (60 mg, orally), and diclofenac (50 mg, suppository) as a standard initial dose. The number of analgesics was scored according to the amount and type of agent (standard initial dose of antianalgesic agent = score 1) used during postoperative periods. When a patient used acetaminophen 1000 mg twice, half dose of diclofenac 50 mg once, and loxoprofen sodium hydrate 60 mg once, the analgesics score was 3.5 (= 2 + 0.5 + 1) points. This scoring system was developed according to our previous study that evaluated the use of antihypertensive medications. 27

| Anesthetic procedure and intraoperative period
The anesthetic management and postoperative analgesia of patients was consistent and was not modified during this study. RARP was performed using general anesthesia, which consisted of remifentanil, ketamine, and propofol. At the end of the surgery, intravenous morphine (5-10 mg) or fentanyl (2-4 μg/kg) was administered as boluses for an opioid rotation. Thereafter, intravenous patient-controlled analgesia using morphine (20-30 mg/day) or fentanyl (400-500 μg/day) with ketamine (20 mg/day) was administered to manage postoperative pain within 2 days. All patients used the same composition of patientcontrolled analgesia with a total volume of 50 ml for 48 hours.

| Outcomes
The primary objectives of this study were to assess the effect of mod- (3) the optimal cutoff value evaluation of G8 for NRS ≥ 5 using the receiver operating characteristic (ROC) curve and area under the curve (AUC), and (4) the association between the G8 and sFI.

| Statistical analysis
Statistical analysis was performed with GraphPad Prism 7.00 (GraphPad Software, Inc., San Diego, CA, USA), Bell Curve for Excel (Social Survey Research Information Co., Ltd., Tokyo, Japan), and R 3.3.3 (The R Foundation for Statistical Computing, Vienna, Austria). Categorical variables were compared using Fisher's exact test or the χ 2 test.
Quantitative variables were expressed as means ± standard deviation.
The significance of between-group differences was determined using Student's t test for normally distributed data or the Mann-Whitney U test for nonnormally distributed data. A P-value < .05 was considered statistically significant. Odds ratio (OR) with 95% confidence intervals (95% CI) were calculated using the multivariate logistic regression model for NRS ≥ 5 after controlling for potential confounders including age, frailty (G8 ≤ 14), analgesic score, and any complications.

| Patient selection and characteristics
Of 181 patients, we excluded six patients who were treated with open radical prostatectomy and 21 patients who did not provide any response to the NRS questionnaire. This left us with a total of 154 patients for the analysis. The median age, PSA, G8, length of hospital-stay, ASA score, and analgesic score were 69 years, 8.74 ng/mL, 14 days, 2, and 2, respectively. The number of patients who experienced grade 1, 2, and 3 complications were 19 (12%), 4 (2.6%), and 1 (0.7%), respectively. Of the 154 study participants, 37 (24%) and 61 (40%) patients were classified to have NRS ≥ 5 and G8 ≤ 14, respectively. The median NRS on days 0, 1, 2, 3, and at discharge were 0, 3, 2, 1, and 0, respectively. Meanwhile, the median value of total pain (sum of NRS) was 6, and the number of patients with moderate to severe pain (NRS ≥ 5) was 37 (24%). The maximum value of NRS was observed on postoperative day 1 in 153 (99.3%) patients. There was no significant difference in background between patients with NRS < 5 and those with NRS ≥ 5 ( Table 1). The number of patients who underwent standard lymph node dissection was not significantly different between the NRS < 5 and NRS ≥ 5 (P = .487).

| Secondary outcomes
The G8 score was not significantly associated with postoperative complications (Figure 2A) and the use of analgesics ( Figure 2B). The sFI 2-5 was not significantly associated with postoperative complications ( Figure 2C) and the use of analgesics ( Figure 2D).

| Exploratory outcome
The time-dependent change in NRS scores between G8 > 14 and ≤ 14 and between sFI 0-1 and 2-5 was shown in Figure 3A,B, respectively.
Regarding frailty, patients with G8 ≤ 14 had significantly greater total pain (total NRS) than those with G8 > 14 ( Figure 3A); meanwhile, the frailty of sFI was not significantly associated with total NRS ( Figure 3B). Maximum NRS was significantly correlated with sum of NRS (R 2 = .705, P < .001) ( Figure 3C). The linear association between the G8 and maximum NRS ( Figure 3D) and between the sFI   . The association between maximum NRS (maximum pain) and total NRS (total pain) was also evaluated (C). The linear association between the total NRS and maximum NRS (D), between maximum NRS and G8 (E), between maximum NRS and sFI (F) were evaluated. The optimal cutoff value of G8 for NRS ≥ 5 was evaluated using receiver operating characteristic (ROC) curve and the area under the curve (AUC) (E) and maximum NRS ( Figure 3E) was very weak (correlation coefficient: R, −.2 to.2). The optimal cutoff value of G8 for the NRS of ≥5 was defined as 14 with an AUC of 0.625 ( Figure 3F). A G8 score of ≤14 was not significantly associated with sFI ( Figure 3G). The linear association between the G8 and sFI was a very weak association ( Figure S1C; correlation coefficient: R, −.2 to .2).

| D ISCUSS I ON
In this prospective observational study, we found that G8 ≤ 14 was an independent risk factor for postoperative moderate to severe The key finding of our study was that patients with frailty might be sensitive to pain than those without frailty. We speculate that patients with painful experiences such as complications may require greater amounts of analgesics compared with patients without severe pain or complications. However, total analgesic score and complication rate were not significantly different between patients with NRS < 5 and those with NRS ≥ 5. These results may suggest that NRS reveals not only physical pain, but also psychological pain. However, our analgesic score was not validated and further study is necessary to address this issue. A cross-sectional study suggested that a higher level of frailty is associated with a higher risk of adverse physical and psychological health situations. 28 Another cross-sectional study, which involved a total of 252 community-dwelling elderly individuals (mean age, 79.2 ± 7.3 years), examined the relationship between frailty and pain. 29 In that study, Coelho et al 29  suggested that there was a significant association between frailty (modified frailty index [mFI]) and postoperative Clavien-Dindo IV complications after RARP. 9 The authors suggested that patients with frailty had increased rates of (any) complications, wound disruptions, bleeding transfusions, and 30-day mortality. Patients with the highest mFI scores (≥3) had significantly higher rate of Clavien-Dindo IV complications (OR = 12.107; 95% CI, 2.800-52.351, P < .005) in comparison with patients without frailty. Therefore, RARP was not recommended for patients with frailty because of worsening outcomes. These findings suggested that frailty might be a useful tool in identifying the optimal patients for surgical interventions. Our previous study revealed that frailty was significantly higher in patients with muscle-invasive bladder cancer who underwent nonsurgical therapy than in those who underwent radical cystectomy (frailty discriminant score: 3.27 vs 2.06, respectively; P < .001). 30  The reason for the no significant association of sFI with moderate to severe pain (NRS ≥ 5) after RARP needs to debate. As the sFI was the comorbidity-based frailty, it was suggested that the significant association of comorbidity-based frailty tools with severe (Clavien-Dindo grade IV or V) postoperative complications. 8,31 As no grade IV or V complication was observed in this study, sFI might be not significantly associated with moderate to severe pain (NRS ≥ 5) after RARP.
We observed no significant association between frailty in G8 and sFI. Although both tools were developed for frailty measure- They concluded that the use of a single frailty screening tool is not sufficient in identifying patients with frailty because of multidimensionality factors. 20 As each tool evaluated different aspects of frailty, it might be not easy to find the "one size fits all" tool for frailty evaluation. Furthermore, optimal frailty tools may change depending on the type of disease. Therefore, a single frailty test might be not enough to evaluate the multidimensional phenomenon of frailty.
Further studies are warranted to identify a suitable combination of frailty tools and diseases.
As several studies suggested G8 of ≤ 14 is the useful cutoff of frailty in many cancers, 20,32 we selected 14 as a cutoff in this study.
However, the definition of the optimal cutoff of the G8 score remains unelucidated. Our recent study evaluating the association between the G8 and treatment by disease stages in a patient with PC suggests that the G8 ≤ 14 may be the useful cutoff of frailty between the surgical treatment (robot-assisted radical prostatectomy) and non-surgical therapies (radiotherapy and/or standard of care including androgen-deprivation therapy) in patients with localized PC. 33 Also, the poor prognosis was associated with G8 < 12 and <13 in the patients with metastatic castration-sensitive and castration-resistant prostate cancers, respectively. 33 Therefore, the cutoff value of G8 may need to be modified depending on the disease status.
Further studies are necessary to identify the optimal cutoff of G8 in the different disease stages.
The study limitations include the small sample size, selection bias, and unmeasurable confounding factors. We could not address the association of frailty with the type of pain because of data unavailability. The results may not be generalized to other countries because of racial and regional differences. For example, the median duration of hospital stay was not significantly different between patients with NRS < 5 and those with NRS ≥ 5 (15 vs 14 days, P = .395), because our national insurance system covers hospitalization costs.
We could not select hospital stay as a primary endpoint because of the universal insurance system used in Japan. Also, the sample size was not enough to evaluate longitudinal changes over time as well as the aspect of confounding introduced by repeated measurements.
Despite these limitations, the study demonstrates the clinical value of frailty on postoperative pain in patients with RARP. Our results suggested that frailty is a key factor for pain management in clinical practice. Further study is necessary to validate our findings.
In conclusion, frailty was significantly associated with moderate to severe pain after RARP and might be a potential predictor of postoperative pain. Frail patients require individual care to avoid painful postoperative experiences.

ACK N OWLED G M ENTS
We thank Yuki Fujita and Atsuko Sakuraba for their invaluable help with the data collection.

CO N FLI C T O F I NTE R E S T S
The authors have no conflicts of interest to declare.