Patient-Reported Outcomes and Return to Intended Oncologic Therapy After Colorectal Enhanced Recovery Pathway

Objective: To evaluate the influence of enhanced recovery pathway (ERP) on patient-reported outcome measures (PROMs) and return to intended oncologic therapy (RIOT) after colorectal surgery. Background: ERP improves early outcomes after colorectal surgery; however, little is known about its influence on PROMs and on RIOT. Methods: Prospective multicenter enrollment of patients who underwent colorectal resection with anastomosis was performed, recording variables related to patient-, institution-, procedure-level data, adherence to the ERP, and outcomes. The primary endpoints were PROMs (administered before surgery, at discharge, and 6 to 8 weeks after surgery) and RIOT after surgery for malignancy, defined as the intended oncologic treatment according to national guidelines and disease stage, administered within 8 weeks from the index operation, evaluated through multivariate regression models. Results: The study included 4529 patients, analyzed for PROMs, 1467 of which were analyzed for RIOT. Compared to their baseline preoperative values, all PROMs showed significant worsening at discharge and improvement at late evaluation. PROMs values at discharge and 6 to 8 weeks after surgery, adjusted through a generalized mixed regression model according to preoperative status and other variables, showed no association with ERP adherence rates. RIOT rates (overall 54.5%) were independently lower by aged > 69 years, ASA Class III, open surgery, and presence of major morbidity; conversely, they were independently higher after surgery performed in an institutional ERP center and by ERP adherence rates > median (69.2%). Conclusions: Adherence to the ERP had no effect on PROMs, whereas it independently influenced RIOT rates after surgery for colorectal cancer.


The iCral3 Prospective Study
The Italian ColoRectal Anastomotic Leakage (iCral3) study group BACKGROUND Enhanced recovery pathway (ERP) is a multimodal and multifactorial approach to the optimization of perioperative management, 1,2 designed to modify and improve the response to surgery-induced trauma based on a series of evidence-based items related to perioperative care. 3 Several meta-analyses have shown a significant reduction in overall morbidity (OM) rates and length of stay after colorectal surgery. [4][5][6] A significant dose-effect curve between adherence rate to ERP items and early outcomes has been demonstrated, [7][8][9][10][11][12] and recent evidence derived from retrospective studies suggest that ERP may also offer a definite advantage over long-term survival after colorectal resection for malignancy. [13][14][15] Apart from early outcomes, little is evident about patients' experiences during the perioperative period. Measuring patient-reported outcomes address the gap in enhanced recovery assessment by incorporating patient-centered quality of life (QoL) into the global assessment of outcomes. 16 At the same time, it is still unclear whether the ERP could improve the return to intended oncologic therapy (RIOT) 17 after cancer surgery through the reduction in complications and poor performance status, both of which are associated with worse long-term oncologic outcomes. 18,19 Therefore, the Italian ColoRectal Anastomotic Leakage study group planned this study (iCral3) to prospectively evaluate the impact of adherence to ERP items after colorectal resections on patient-reported outcome measures (PROMs) and RIOT after surgery for malignant disease.

Study Design, Participants, and Setting
Prospective enrollment from November 2020 to October 2021 was carried out among 76 Italian surgical centers that voluntarily participated in iCral3. All patients undergoing elective or delayed urgency (>24 hours from admission) colorectal surgery with anastomosis were assessed for inclusion in a prospective database after obtaining written informed consent. Inclusion criteria were as follows: a) patients who underwent laparoscopic/robotic/open/converted colorectal resection with anastomosis, including planned Hartmann's reversals; b) American Society of Anesthesiologists' (ASA) class I, II, or III; c) elective or delayed urgency (>24 hours from admission) surgery; and d) patients' written acceptance to be included in the study. The exclusion criteria were as follows: a) ASA class IV-V, b) emergent (≤24 hours from admission) surgery, c) pregnancy, and d) hyperthermic intraperitoneal chemotherapy for carcinomatosis.
According to the median number of cases enrolled during recruitment, each center was defined as high-volume (>median) or low-volume (≤median). The existence of an institutional ERP (having a locally implemented ERP team and protocol, supported by a specific resolution of the hospital/company strategic management) was declared by 48 out of 76 (63.1%) participating centers. All data of the included patients were prospectively uploaded into a web-based database via an electronic case report form, specifically designed for iCral3, protected by access credentials for each center/investigator.

Clinical and Adherence Data
Continuous and discrete variables related to biometric data, patient-related and institution-related variables, indication and type of surgical procedure, and outcomes were recorded. Adherence to the 26 items of the ERP was measured for each single enrolled case upon criteria adapted from the 2018 ERAS Society 20 and 2019 national 21 guidelines (Table 1). Quality control of the data for consistency, plausibility, and completeness was performed on every single record by local investigators and subsequently validated by the study coordinator, resolving any discrepancies through strict cooperation.
During the perioperative period, patients were examined daily by local investigators, who were left free to decide on any complementary imaging and any further action according to their local criteria. Patients were followed up for a minimum of 8 weeks, during which data on all outcome measures and other study variables were collected.

Outcomes
The primary endpoints were PROMs and RIOT. 17 PROMs were administered to all enrolled patients 4 weeks to 1 day before the planned operation (preoperative), on the day of discharge (discharge), and 6 to 8 weeks after the operation (late), using the Euro-Quality of Life Group EQ-5D-5L, 22 the MD Anderson Symptom Inventory for Gastrointestinal Surgery patients (MDASI-GI 23 ), and the Functional Assessment of Cancer Therapy -Colorectal (FACT-C 24 ) questionnaires. The EQ-5D-5L is a generic questionnaire on QoL divided into 2 sections: the EQ-5D index and the EQ-5D visual analog scale (VAS); the EQ-5D index assesses health status across 5 domains: Patient with Hb concentration < 130 g/L for men and <120 g/L for nonpregnant women receive the correction of anemia before surgery preferably through intravenous iron preparations (ferric carboxymaltose) and blood transfusion(s) in strictly necessary cases Antithrombotic prophylaxis Patient receives graduate compression stockings and/or pneumatic compression device, together with prophylaxis with low molecular weight heparin during the perioperative period, to be extended up to 28 d after surgery in case of malignancy Antibiotic prophylaxis Patient is administered i.v. antibiotic 30 to 60 mins before incision, according to local protocols No bowel preparation No routine bowel preparation is used, except in case of the anticipated need for covering stoma Oral carbohydrates load Carbohydrates-rich beverage (12.5% maltodextrins) is given preoperatively (800 mL on the evening before surgery and 400 mL 2 to 3 hrs before surgery) Preoperative fasting Preoperative fasting is limited to 2 hrs for clear liquids (water, coffee, tea) and to 6 hrs for milk and solid food No premedication No long-or medium-action sedatives. Short and ultra-short-acting sedatives (eg, Lorazepam, Midazolam, Methohexital, Dexmedetomidine, Ketamine) are allowed before performing spinal, epidural, or loco-regional anesthesia PONV prophylaxis Postoperative nausea/vomiting prophylaxis is administered according to individual risk assessment (Apfel score) through a multimodal approach Normothermia Body temperature is monitored during surgery, utilizing fluid warmers and/or thermic blankets as necessary Standard anesthetic protocol General anesthesia through short-acting anesthetics, cerebral activity monitoring to enhance recovery and to reduce postoperative delirium, anesthesia level monitoring, and complete reversal of neuromuscular blockade Intraoperative fluid management Restrictive fluid therapy (defined as maintenance fluids at <2 mL/kg/h) or goal-oriented fluid therapy (stroke volume) Multimodal analgesia Use of more than 2 drugs or analgesia strategies ( anxiety/depression, mobility, pain/discomfort, self-care and usual activities, each scoring from 1 (worst) to 5 (best); EQ-5D VAS is a single 20 cm VAS with a range of 0 to 100, where 0 is the worst and 100 is the best imaginable health status. The EQ-5D-5L total scores ranged from 5 (worst) to 125 (best). The MDASI-GI is a specific questionnaire designed for evaluating digestive symptoms (18 questions) and the extent to which such symptoms interfere with daily activities (6 questions), each of which is presented as a VAS ranging from 0 (not present and/or not interfering at all) to 10 (as bad as you can imagine and/or completely interfering), with total scores ranging from 0 (best) to 240 (worst). The FACT-C is a specific colorectal cancer health questionnaire investigating physical, social/family, emotional, and functional well-being, with scores ranging from 0 (worst) to 136 (best). RIOT rates, defined as the intended oncologic treatment according to the national guidelines for colorectal cancer and the disease stage 25 administered within 8 weeks from the index operation, were recorded in all patients who underwent surgery for malignancy.
Secondary endpoints were anastomotic leakage (AL), defined and graded according to international 26 and national 27 consensus, OM (any adverse event, graded according to 29 and the Japanese Clinical Oncology Group-JCOG extended criteria 30 ), major morbidity (MM, any adverse event grade > II), readmission, reoperation, and mortality rates. The overall postoperative length of stay (LOS) was calculated including any eventual readmission. All patients with a proximal diverting stoma at index operation underwent a routine check of anastomotic integrity through an intraluminal contrast exam (standard X-rays or CT scan), MRI, or direct endoscopic evaluation 3 to 8 weeks after the operation.

Ethics and Dissemination
The study was conducted in accordance with the Declaration of Helsinki and the Guideline for Good Clinical Practice E6 (R2) principles. The study protocol was approved by the coordinating center ethics committee (Comitato Etico Regionale delle Marche -C.E.R.M. #2020/192, approved on 07/30/2020) and registered at ClinicalTrials.gov (NCT04397627). Thereafter, all participating centers obtained authorization from the local institutional review board. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology reporting guidelines for cohort studies. 31 Individual participant-level anonymized datasets were made available upon reasonable request by contacting the study coordinator.

Statistical Analysis
All quantitative values are expressed as mean ± SD and 95% confidence intervals (95% CIs), median and interquartile range (IQR), and categorical data with percentage frequencies. For categorical data, the analysis included the use of cross-tabulation, chi-square, or Fisher's exact test where indicated. Continuous or discrete variables were analyzed using Student's 2-sided t test (allowing for heterogeneity of variances) or a non-parametric test (Mann-Whitney U test or Kruskal-Wallis test as indicated).
With the double intent of adjusting mean values for variables and identifying factors associated with PROMs, a generalized linear mixed regression model was used according to the "Setting International Standards in Analyzing Patient-Reported Outcomes and Quality of Life Endpoints Data Consortium" recommendations, 32 where the dependent variables were PROMs at 2 different time-points (discharge and late) and independent items were all other variables, including baseline preoperative values for each PROM, AL, OM, MM, and reoperation. A generalized estimating equation 33 was used to calculate model parameters to take into account the clustered, multicenter, nature of data. Each Beta coefficient measures the mean increment or decrement in PROMs values according to the presence of each independent item. To measure variable multicollinearity, 34 the variance inflation factor was calculated using multiple linear regression for all the primary endpoints. No missing-value imputations were performed.
Quantitative variables such as age (years), operation length (minutes), and adherence rates (%) to the ERP were categorized according to their median values and/or their fourth centile. Other variables were categorized according to accepted predefined ranges: Mini Nutritional Assessment -Short Form (MNA-SF 35 ) < 12, indicating potential malnutrition, or ≥12, indicating normal nutritional status; body mass index (Kg/m 2 ) ≤25.0, 25.1 to 30.0, and >30.0. Surgical procedures were categorized as standard (anterior resection, right colectomy, left colectomy) versus non-standard (splenic flexure resection, transverse colectomy, Hartmann's reversal, subtotal and total colectomy, and other) resections. 12 The location of the tumor in case of malignancy was categorized as "right" (up to the transverse colon) or "left" (from the splenic flexure to the lower rectum).
A logistic regression analysis, excluding any variable with variance inflation factor ≥4, was used to assess associations between the examined variables and RIOT, presenting odds ratio (OR) and 95% CI.
For all statistical tests, the significance level was set at P < 0.05. All analyses were conducted using StatsDirect statistical software (StatsDirect Ltd., United Kingdom) and IBM SPSS Statistics for Windows, v.28.0 (Armonk, NY).

Sample Size
Adherence to at least 70% of the ERP items was identified as a cutoff for significant improvement in outcomes, 9 with a 2:1 expected ratio below:above this cutoff. Estimating a reduction of postoperative PROMs from preoperative baseline (1.0) at 0.7 for adherence above the cutoff and at 0.64 for adherence below the cutoff, 36 alpha 0.04, beta 0.8, the required sample size was 2406 (802 cases above and 1604 cases below 70% adherence). Reported rates of failure to RIOT and ERP items adherence below or above 70% were 13% and 6.5%, respectively 19 ; the required sample size for evaluation of RIOT was 885 (295 cases above and 590 cases below 70% adherence). Based on previous iCral observational studies on colorectal surgery in Italy, 12,37,38 the expected ratio of malignant:benign indications to surgery was 70:30 (2100 resections for malignancy and 900 resections for benign disease based on 3000 expected cases).

RESULTS
A total of 6174 potentially eligible cases were assessed, of which 4529 (73.3%) were included in the study and analyzed for PROMs (Fig. 1); surgery for malignancy was performed in 3283 cases (72.5%); indications for an adjuvant therapy according to disease stage and national guidelines 30  There were 62 deaths (mortality rate 1.4%). Median overall LOS (IQR; range) was 6 (4 to 8; 0 to 91) days, with 174 re-admissions (3.8%) and 232 re-operations (5.1%).

PROMs
Compliance with complete preoperative, discharge, and late evaluation QoL questionnaires varied from 88.1% to 95.5%; unadjusted PROMs values (Supplemental Table 2, http://links. lww.com/AOSO/A219) showed a significant worsening at discharge and a significant improvement beyond the preoperative baseline at late evaluation (Fig. 2), while their unadjusted values according to the median or fourth quartile of ERP adherence rates are shown in Table 3. After adjustment, no significant difference was detected in PROMs values according to the median or fourth quartile of ERP adherence rates ( Table 4). The regression coefficients for the generalized mixed linear regression model considering ERP adherence rates below or above the median cutoff are reported in Table 5 (in Supplemental Table 3, http://links.lww.com/AOSO/A220, considering ERP adherence rates first to third vs fourth quartile). EQ-5D-5L adjusted values, both at discharge and at late evaluation, showed a significant negative association with ASA class III, presence of chronic renal failure, neoadjuvant therapy, OM, AL, length of procedure > 180 min, and, at discharge only, standard procedures. MDASI-GI adjusted values, both at discharge and at late evaluation, showed a significant negative association with female sex, neoadjuvant therapy, OM, AL, and length of procedure > 180 min. Similarly, FACT-C adjusted values at discharge and at late evaluation showed a significant negative association with ASA class III, presence of chronic renal failure, neoadjuvant therapy, elective admission, and AL. At the late evaluation, surgery for malignancy was independently linked to worst values for all PROMs.

DISCUSSION
This prospective multicenter observational study investigated the effects of the ERAS program on PROMs and RIOT after elective colorectal surgery in more than 4000 patients enrolled over a 12-months period in 76 Italian surgical centers, without any limitation concerning the presence of an institutional ERP or center caseload. During the design of the current study, we carefully screened the discriminant properties of existing PROMs, being aware that there is still very limited evidence supporting the measurement properties of existing PROMs used to evaluate recovery  after abdominal surgery, 39 many of which fail to detect even obvious differences, such as those related to the type of surgical approach. 40 Moreover, a previous prospective series of 100 cases using a different set of PROMs 36 failed to detect any influence of ERP adherence on late postoperative recovery, hypothesizing either a lack of association or a lack of PROMs' ability to detect any difference. Therefore, we decided to use 3 different questionnaires: EQ-5D-5L was chosen as a generic QoL index because of its worldwide availability (>130 languages), low (<5 minutes) time requirement, and great discriminatory power; MDASI-GI as a digestive disease-specific instrument because it is concise, stable and comprehensive, using simple and familiar 0 to 10 scales; and FACT-C as a colorectal cancer-specific index because of its disease-specific discriminatory power. 41,42 As a matter of fact, the raw values of all these 3 QoL instruments showed a parallel and significant worsening at discharge and improvement at late (6 to 8 weeks after surgery) evaluation compared to their preoperative values (Fig. 2), with high compliance rates ranging from 88.1 to 95.5% of cases (Supplemental Table 2, http://links.lww.com/AOSO/A219). This finding confirms the reliability of the PROMs set used in the current study in recording the strong influence of major abdominal surgery on patients' perceived QoL and functional status, and that postoperative recovery takes time beyond hospital discharge. 43 At the same time, raw PROMs values showed little or no significant association with ERP adherence rates, with only MDASI-GI scores significantly improved both at discharge and at late evaluation according to the fourth centile (Table 3). Once raw data were adjusted through a generalized linear mixed regression model accounting for baseline PROMs values and all other variables considered in the study, 32 it appeared that ERP adherence rates had no significant effect on patient-reported outcomes (Tables 4  and 5). This finding may stem, as already recorded, 36,39 from the inability of the current PROMs set to detect any difference. Considering that neoadjuvant treatments, surgery for malignancy, and the occurrence of morbidity and AL showed an independent negative association with nearly all PROMs and time spans used in the current analysis (Table 5), this finding is more likely related to the independent detrimental effect of cancer and postoperative complications on self-reported recovery after colorectal surgery. Failure to detect any independent effect of ERP adherence on patients' perceived QoL may appear as a negative result of the study; on the other hand, it could be intended as a starting point or as the end of the beginning in the clinical research area dealing with the measurement of patient-reported outcomes in colorectal surgery; all the clinicians (surgeons as well) involved in the perioperative journey of the patient should, on the one hand, continue to struggle to improve time-honored outcome measures such as morbidity, mortality and postoperative LOS, and, at the same time, move away from them and forward to the individuation of an entirely new set of PROMs, able to address patients' still unmet needs. 44,45 RIOT rates were defined and calculated as the actual number of patients receiving the indicated adjuvant therapy within 8 weeks from the index operation, as timely initiation is linked to improved long-term results. 46 The overall rate of 54.5% recorded in this study is within the literature range, 47 and it was independently lower in older and comorbid patients, open surgery, and in presence of major morbidity (Table 6 and Fig. 3). This is not surprising, since these factors are commonly reported as the main determinants of failure of RIOT after surgery for colorectal cancer. 48,49 Both the presence of an institutional ERP and its adherence rate beyond the median cutoff (69.2%) resulted in independent protective factors. While a previous multivariate analysis of a smaller retrospective cohort comparing pre-and post-ERP implementation identified significantly higher RIOT rates in the post-implementation group, 49 the current study is, to our knowledge, the first to identify an independent role of ERP adherence and institutionalization on RIOT rates in a larger and multicenter prospective cohort. This effect could partially explain the impact of ERP adherence on long-term oncologic outcomes recorded in previous retrospective series. [13][14][15] For this purpose, a 3-and 5-year follow-up of all oncologic patients enrolled in the present cohort has already been planned.
The median number of enrolled patients per center in the present study (No.= 44) was significantly lower than that (No.= 82) recorded in the previous iCral2 study, 12 and while OM (26.8%) and AL (4.5%) rates recorded in the present study were similar to those recorded in the previous study, MM (7.5%) and mortality (1.4%) rates were somewhat higher. The lower accrual rate in the present study is probably due to 2 main factors. First, the world coronavirus pandemic had a deep impact on the reduction of elective surgical activities 50 ; second, the inclusion of PROMs in the investigation protocol led to a higher number of exclusions due to consent denial and/or incomplete data (Fig. 1). Anyway, this prospective cohort was more representative (inclusion rate 73.3%) than that of the previous study (inclusion rate 57.8%), probably because proximal diverting stoma and delayed urgency cases were included. These cases may be responsible for the higher MM and mortality rates compared to those in the previous study.
This study has several strengths: it represents, by far, the largest multicenter prospective investigation on some of the currently available PROMs after colorectal surgery; it was performed in a well-defined time-lapse in a large number of centers representing a wide sample of surgical units performing colorectal surgery in Italy; the prospective design of the study allowed the measurement of outcomes through adherence to ERP items in all the enrolled cases, responding to clear and sheer compliance criteria. Its main limitation is the potential for residual, measured, and unmeasured confounding intrinsic to any observational study. Moreover, although data quality control was performed and repeated at various levels, we cannot exclude any measurement error from the participating investigators.
In conclusion, this study brings 2 important findings: on one side, the confirmation that every effort should be made to improve ERP adherence rates since this could improve long-term oncologic outcomes through its independent boosting effect on the timely RIOT; on the other side, the consciousness that this is still not sufficient to significantly improve patients' self-perceived QoL during their perioperative journey. Further clinical research to cope with this caregiving shortcoming should be strongly encouraged.