Association of the Flemish Triage Risk Screening Tool with postoperative outcomes in older adults after colorectal cancer surgery: retrospective cohort study

Abstract

Background

Older adults undergoing colorectal cancer (CRC) surgery are at increased risk of postoperative complications, mortality, and loss of independence. Frailty screening tools may support perioperative risk stratification. This study evaluated the association between the Flemish Triage Risk Screening Tool (fTRST) and postoperative outcomes.

Methods

This retrospective multicentre cohort study analysed the Emilia-Romagna Surgical Colorectal Audit. Patients aged ≥ 70 years undergoing CRC surgery (2019–2023) were included. Patients with fTRST scores ≥ 2 and < 2 were matched 1 : 1 using propensity score matching. Primary outcomes were 30-, 90-, and 180-day mortality and complications. Secondary outcomes included prolonged hospital stay and discharge to a non-independent setting. Logistic regression and receiver operating characteristic (ROC) curve analyses were performed.

Results

In all, 1539 patients were included in the study. Mortality at 30, 90, and 180 days was 4.2%, 7.1%, and 10.5%, respectively. In the unmatched cohort, each 1-point increase in fTRST score was significantly associated with 30-day mortality (odds ratio (OR) 1.36; 95% confidence interval (c.i.) 1.12 to 1.66), complications (OR 1.12; 95% c.i. 1.02 to 1.23), prolonged hospital stay (OR 1.13; 95% c.i. 1.03 to 1.24), and discharge to a non-independent setting (OR 1.35; 95% c.i. 1.21 to 1.51). After matching 566 pairs, higher fTRST score remained significantly associated with 30-day mortality (OR 1.33; 95% c.i. 1.09 to 1.63) and discharge to a non-independent setting (OR 1.40; 95% c.i. 1.25 to 1.59). ROC analysis using an fTRST cut-off of ≥2 showed moderate discrimination for 30-day mortality and discharge to a non-independent setting, with areas under the ROC curve of 72.4% (negative predictive value 98.2%) and 67.2% (positive predictive value 66.3%), respectively.

Conclusions

The fTRST is a simple screening tool for identifying older adults at risk of adverse postoperative and functional outcomes after CRC surgery, supporting perioperative planning.

Using a large regional registry, this multicentre retrospective study evaluated the predictive value of the Flemish Triage Risk Screening Tool (fTRST) in older adults (age ≥ 70 years) undergoing colorectal cancer surgery. An fTRST score ≥ 2 was independently associated with higher 30-day mortality, complications, prolonged hospital stay, and discharge to non-independent settings, with good discriminative ability, particularly a high negative predictive value for mortality. Overall, the fTRST emerged as a simple and effective screening tool for identifying low-risk patients and supporting tailored perioperative management in geriatric colorectal surgery.

Introduction

Colorectal cancer (CRC) is predominantly a disease of older adults, with most patients undergoing surgery aged 70 years or older¹. Surgical resection remains the cornerstone for stage I–III CRC, and it plays a critical role even in the palliative/urgent settings for symptom control. Older patients present unique challenges due to age-related physiological decline, co-morbidities, and increased vulnerability to postoperative complications. It is known that the presence of co-morbidities doubles the risk of postoperative complications and that one in five older patients undergoing an operation for CRC will experience postoperative loss of independence^2,3. In this context, frailty has emerged as a key determinant of perioperative outcomes and a predictor of functional recovery.

Although several preoperative assessment tools have been proposed to evaluate surgical risk, most of these instruments focus primarily on predicting medical or surgical complications⁴. Some other tools, such as the Clinical Frailty Scale or the Geriatric 8 screening tool, have been shown to be effective in predicting outcomes in the oncogeriatric population, but they often cannot simultaneously capture the risks of both postoperative morbidity and functional decline^5,6. Moreover, many of these tools are either time-consuming, require specialized training, or are not routinely integrated into surgical workflows. To address these gaps, the Flemish version of the Triage Risk Screening Tool (fTRST) has been proposed as a rapid, easy-to-administer screening tool to identify frailty in older adults⁷. The fTRST includes a limited number of items and has demonstrated good prognostic utility for functional decline and overall survival in oncological populations⁸. The fTRST comprises five domains: the presence of cognitive decline (2 points), living alone or no help from partner/family available (1 point), reduced mobility or falls in the past 6 months (1 point), hospital admission within 3 months (1 point), and ≥ 5 different medications (1 point). A total score of ≥ 2 must be interpreted as a proxy for patient frailty. Originally, fTRST ≥ 2 identified senior adults with cancer in need of comprehensive geriatric assessment⁸. However, fTRST was recently found to also be effective in predicting the risk of postoperative adverse outcomes in specific subgroups of older patients^9,10.

However, despite its simplicity and growing use, evidence validating the fTRST in the context of surgical outcomes, particularly in CRC surgery, is still limited. Exploiting the Emilia-Romagna Surgical Colorectal Audit (ESCA) database, this study aimed to evaluate the association between the fTRST and postoperative morbidity, mortality and functional outcomes in older adults undergoing CRC surgery.

Methods

This study was a multicentre retrospective analysis of a prospectively maintained database. Data were derived from the ESCA, an Italian regional registry that collects perioperative data of all patients aged > 18 years undergoing CRC surgery¹¹. The study was approved by the Romagna Ethics Committee (CEROM; Approval no. 2278). The study has been registered with ClinicalTrials.gov (ID: NCT03982641).

Data from centres are regularly audited by the Outcome Research Unit of the coordinating centre (IRCCS Istituto Romagnolo per lo Studio dei Tumori ‘Dino Amadori’). Written informed consent was obtained whenever feasible. When not possible due to the urgent clinical setting, patient privacy was safeguarded in accordance with applicable regulations, and a substitute privacy declaration was provided. On a yearly basis, the number of patients included in the ESCA is cross-checked against administrative data via regional hospital discharge cards to ensure the highest possible enrolment rate, currently at 93%. The full list of variables collected in the ESCA database, together with details of the study protocol and methodology, has been published previously¹¹. The present study followed the TRIPOD reporting guidelines (Table S1)¹².

Patients

All patients aged ≥ 70 years who underwent CRC surgery and were registered in the ESCA database between 1 May 2019 and 31 December 2023 were selected for analysis. Exclusion criteria were the absence of or incomplete fTRST screening and/or incomplete 30-day postoperative follow-up.

Data collection

Patient demographics and clinical baseline characteristics (sex, age, tumour localization, surgical intent, setting and body mass index) were collected. To obtain a multimodal and multidimensional evaluation of preoperative frailty, five screening tools were used: the Eastern cooperative Oncology Group Performance Status (ECOG PS)¹³, the American Society of Anaesthesiologists (ASA) grade¹⁴, Nutritional Risk Screening (NRS)¹⁵, the Charlson age co-morbidity index (CACI)¹⁶, and the fTRST⁷. The final pathology report, according to the seventh edition of the tumour node metastasis (TNM) cancer staging system¹⁷, was included. Thirty-day postoperative complications according to the Clavien–Dindo grade¹⁸ and the Comprehensive Complication Index¹⁹, 30-, 90-, and 180-day mortality, unplanned readmission, and changes in living situation after discharge complemented the analysis data set. Length of hospital stay (LOS) was considered prolonged when > 5 days²⁰.

Statistical analysis

Descriptive statistics were used to summarize baseline characteristics of the study population and patients’ postsurgical outcomes. Continuous variables are presented as the median with interquartile range (i.q.r.), and categorical variables are presented as absolute frequencies and percentages. The objective of the analysis was to assess the association between fTRST and postoperative outcomes: 30-day mortality, 30-day complications, LOS > 5 days, 30-day readmissions, and discharge to a non-independent setting. The analysis was conducted in three sequential stages. In the first stage, the association between the fTRST score as a continuous variable and each outcome was assessed using univariate logistic regression models. Odds ratios (ORs), with their standard error (s.e.), 95% confidence interval (c.i.), and P values were reported to assess the unadjusted effect of the fTRST score on outcomes. Second, to evaluate the association between the fTRST and postoperative outcomes, multivariable logistic regression models that included fTRST as the main independent variable and adjusted for a set of clinically relevant covariates (age, sex, tumour location, CACI, ASA grade, ECOG PS, nutritional status, pathological (p) TNM, surgical intent, and setting) were developed. Multivariable logistic regression models were built using backward stepwise selection based on the Akaike information criterion, with covariates selected on the basis of both clinical relevance and statistical significance (P < 0.100 in univariate analyses). Results are reported as adjusted ORs with their 95% c.i. for the fTRST and other covariates included in the models. Finally, to further control for potential confounding, propensity score matching (PSM) was performed based on the probability of having an fTRST score ≥ 2. Specifically, the propensity score was estimated using a logistic regression model developed starting from all baseline covariates selected through a stepwise approach based on the Akaike information criterion. In practice, 1 : 1 matching was done using nearest-neighbour matching without replacement. Multivariable logistic regression analyses were then repeated in the matched cohorts to confirm the robustness of the associations observed in the full data set.

In addition, to complement the analysis, the diagnostic performance of the fTRST score (< 2 versus ≥ 2) was evaluated. Sensitivity, specificity, the positive predictive value (PPV), the negative predictive value (NPV), and the area under the receiver operating characteristic (ROC) curve (AUC) are reported, along with the 95% c.i. for each outcome.

Statistical significance was set at a two-sided P  < 0.050. Analyses were conducted using R version 4.4.2 (R Foundation for Statistical Computing, Vienna, Austria; https://www.r-project.org/). The MatchIt package was used to perform PSM²¹; reportROC was used to assess the diagnostic performance of the fTRST score²².

Results

In all, 1719 patients undergoing CRC surgery between May 2019 and December 2023 were screened, of whom 1539 were eligible for analysis (pre-PSM population). Among these patients, 566 with an fTRST score ≥ 2 were propensity score matched 1 : 1 with patients with a lower score, resulting in 1132 matched patients (post-PSM population; 566 in both the fTRST ≥ 2 and < 2 groups). Demographic characteristics and postoperative outcomes of the unmatched population are summarized in Table 1. Overall, 764 patients (49.6%) were female, with a median age of 80 (i.q.r. 75–84) years. Most tumours were in the colon (83.2%), and 91.7% of patients were classified as M0. The median body mass index was 25.5 (i.q.r. 23.0–28.4) kg/m². The median CACI was 7 (i.q.r. 6–8), with 72.8% of patients having an ASA grade ≥ III. Nutritional impairment and ECOG PS ≥ 2 were observed in 39.8% and 19.3% of patients, respectively. Palliative surgery was performed in 8.8% of patients, urgent procedures were performed in 15.5% of patients, and open surgery was performed in 14.1% of patients. According to the fTRST, 63.3% of patients were non-frail and 36.7% were frail. Overall, the mean postoperative LOS was 8.2 days; 49.8% of colon cancer patients had a postoperative LOS > 5 days, compared with 57.0% of rectal cancer patients.

Table 1.

Baseline characteristics and postoperative outcomes of the overall population before propensity score matching (n = 1539)

Age (years), median (i.q.r.)	80 (75–84)
Sex
Female	764 (49.6%)
Male	775 (50.4%)
Tumour localization
Colon	1281 (83.2%)
Rectum	258 (16.8%)
Pathological metastatic stage
M0	1411 (91.7%)
M1	128 (8.3%)
BMI (kg/m2), median (i.q.r.)	25.5 (23.0–28.4)
fTRST ≥ 2	566 (36.8%)
CACI, median (i.q.r.)	7 (6–8)
ASA grade ≥ III	1121 (72.8%)
ECOG PS ≥ 2	297 (19.3%)
Impaired nutritional risk score	612 (39.8%)
Palliative surgical intent	136 (8.8%)
Urgent surgery	238 (15.5%)
Open surgery	217 (14.1%)
Mean LOS (days)
Overall	8.2
Colon cancer patients	8.1
Rectal cancer patients	8.5
LOS > 5 days
Overall	785 (51.0%)
Colon cancer patients	638 (49.8%)
Rectal cancer patients	147 (57.0%)
Overall postoperative complications	642 (42.1%)
Severe complications (CD ≥ III)	138 (9.0%)
Severe complications (CD > III)	107 (7.0%)
Discharge to home (independent)	819 (53.2%)
ER access due to surgical complications	61 (4.0%)
Readmission from index surgery	107 (7.0%)
Reintervention within 30 days of index surgery	96 (6.2%)
Mortality
30 day	65 (4.2%)
90 day	110 (7.1%)
180 day	162 (10.5%)

Values are n (%) unless otherwise stated. i.q.r., interquartile range; BMI, body mass index; fTRST, Flemish version of the Triage Risk Screening Tool; ASA, American Society of Anaesthesiologists; CACI, Charlson age co-morbidity index; ECOG PS, Eastern cooperative Oncology Group performance status; LOS, length of hospital stay; CD, Clavien–Dindo; ER, emergency room.

Postoperative complications occurred in 42.1% of patients, with 9.0% experiencing severe complications (Clavien–Dindo ≥ III). Emergency room visits occurred in 4.0% of patients, 7.0% of patients required readmission, and 6.2% of patients underwent reintervention within 30 days. Thirty-day mortality was 4.2%, increasing to 7.1% at 90 days and 10.5% at 180 days. Overall, 53.2% of patients were discharged home (independent living).

Thirty-day mortality

In the unmatched cohort, 30-day mortality was 4.2% (Table 1). From the univariate analysis, a 1-point increase in the fTRST was associated with an increased risk of 30-day mortality (OR 1.76; 95% c.i. 1.50 to 2.08; P < 0.001; Table 2). The association between fTRST and 30-day mortality was also evaluated after controlling for additional baseline covariates in a multivariable logistic regression analysis (Table 3). Interestingly, even adjusting for other clinically relevant factors, the fTRST was significantly associated with increased 30-day mortality (OR 1.36 for a 1-point increase in fTRST; 95% c.i. 1.12 to 1.66; P = 0.002). Other significant predictors of 30-day mortality included palliative surgery (OR 2.60; 95% c.i. 1.37 to 4.80; P = 0.003) and urgent surgery (OR 3.35; 95% c.i. 1.87 to 5.98; P < 0.001), while normal nutritional status was associated with a lower risk of 30-day mortality (OR 0.48; 95% c.i. 0.26 to 0.85; P = 0.014).

Table 2.

Univariate model before PSM for surgical outcomes

	OR*	P
30-day mortality	1.76 (1.50, 2.08)	< 0.001
30-day complications	1.27 (1.17, 1.37)	< 0.001
LOS > 5 days	1.23 (1.13, 1.33)	< 0.001
30-day readmission	0.98 (0.83, 1.14)	0.793
Discharge to non-independent setting	1.78 (1.62, 1.96)	< 0.001

^*Values in parentheses are 95% confidence intervals. Statistical significance was set at P < 0.050. PSM, propensity score matching; OR, odds ratio; LOS, length of hospital stay.

Table 3.

Multivariable model before and after PSM for 30-day mortality and 30-day complications

Outcome	Variable	Before PSM		After PSM
		OR*	P	OR*	P
30-day mortality	fTRST (per 1-point increase)	1.36 (1.12, 1.66)	0.002	1.33 (1.09, 1.63)	0.005
	Sex (male versus female)	1.61 (0.93, 2.82)	0.094	1.77 (1.01, 3.15)	0.047
	Age (per 1-year increase)	1.05 (1.00, 1.10)	0.061	1.05 (1.00, 1.11)	0.049
	Palliative surgery	2.60 (1.37, 4.80)	0.003	2.01 (1.03, 3.83)	0.036
	Urgent surgery	3.35 (1.87, 5.98)	< 0.001	3.77 (2.07, 6.86)	< 0.001
	ECOG-PS (≥ 2 versus < 2)	1.63 (0.86, 3.07)	0.130	1.89 (1.00, 3.60)	0.050
	ASA grade (≥ III versus < III)	4.24 (0.99, 18.20)	0.052
	Nutritional risk (normal versus impaired)	0.48 (0.26, 0.85)	0.014	0.50 (0.27, 0.90)	0.023
30-day complications	fTRST (per 1-point increase)	1.12 (1.02, 1.23)	0.022	1.11 (1.00, 1.23)	0.061
	Tumour localization (rectum versus colon)	1.64 (1.24, 2.17)	< 0.001	1.75 (1.24, 2.48)	0.002
	Sex (male versus female)	1.29 (1.04, 1.59)	0.020	1.33 (1.04, 1.71)	0.024
	CACI	1.10 (1.03, 1.17)	0.003	1.12 (1.04, 1.20)	0.002
	Urgent surgery	1.93 (1.43, 2.60)	< 0.001	1.83 (1.31, 2.55)	< 0.001
	ECOG PS (≥ 2 versus < 2)	1.81 (1.35, 2.43)	< 0.001	1.75 (1.28, 2.39)	< 0.001

^*Values in parentheses are 95% confidence intervals. Statistical significance was set at P < 0.050. PSM, propensity score matching; OR, odds ratio; fTRST, Flemish version of the Triage Risk Screening Tool; ECOG PS, Eastern cooperative Oncology Group performance status; ASA, American Society of Anaesthesiologists; CACI, Charlson age co-morbidity index.

To further test the impact of the fTRST while accounting for possible differences in baseline characteristics between patients with low and high fTRST scores, a logistic regression model was developed on the post-PSM cohorts. After PSM, the fTRST remained significantly associated with 30-day mortality (OR 1.33; 95% c.i. 1.09 to 1.63; P = 0.005). Beyond the factors identified as statistically significant in the multivariable model of the unmatched cohorts, two additional predictors exhibited statistical significance: male sex (OR 1.77; 95% c.i. 1.01 to 3.15; P = 0.047) and age (OR, 1.05; 95% c.i. 1.00 to 1.11; P = 0.049).

Thirty-day complications

Postoperative complications within 30 days occurred in 42.1% of patients (Table 1). In the unmatched cohort, univariate analysis showed that each 1-point increment in the fTRST was associated with a higher risk of complications (OR 1.27; 95% c.i. 1.17 to 1.37; P < 0.001; Table 2). This association remained significant in the multivariable model after adjusting for relevant baseline covariates (OR 1.12; 95% c.i. 1.02 to 1.23; P = 0.022). After PSM, the association was not significant (OR 1.11; 95% c.i. 1.00 to 1.23; P = 0.061; Table 3). Among the covariates showing statistical significance, urgent surgery appeared to be the strongest risk factor for 30-day complications (OR 1.93; 95% c.i. 1.43 to 2.60; P < 0.001; Table 3).

LOS > 5 days

A prolonged hospital stay (> 5 days) was observed in 51.0% of patients (Table 1). It was significantly associated with each 1-point increase in the fTRST in the univariate analysis (OR 1.23; 95% c.i. 1.13 to 1.33 P < 0.001; Table 2). This positive association was also confirmed in multivariable analyses in both the unmatched cohort (OR 1.13; 95% c.i. 1.03 to 1.24; P = 0.010) and matched cohort (OR 1.12; 95% c.i. 1.01 to 1.24; P = 0.027; Table 4). As indicated in Table 4, rectal cancer (OR 1.65; 95% c.i. 1.24 to 2.18; P < 0.001) and urgent surgery (OR 2.55; 95% c.i. 1.87 to 3.51; P < 0.001) were identified as the strongest risk factors for a prolonged hospital stay.

Table 4.

Multivariable model before and after PSM for LOS > 5 days, 30-day readmission, and discharge to a non-independent setting

Outcome	Variable	Before PSM		After PSM
		OR*	P	OR*	P
LOS > 5 days	fTRST (per 1-point increase)	1.13 (1.03, 1.24)	0.010	1.12 (1.01, 1.24)	0.027
	Tumour localization (rectum versus colon)	1.65 (1.24, 2.18)	< 0.001	1.50 (1.06, 2.14)	0.022
	Sex (male versus female)	0.79 (0.64, 0.97)	0.026	0.73 (0.57, 0.93)	0.011
	CACI	1.14 (1.07, 1.22)	< 0.001	1.14 (1.06, 1.23)	< 0.001
	Urgent surgery	2.55 (1.87, 3.51)	< 0.001	2.49 (1.76, 3.56)	< 0.001
	Nutritional risk (normal versus impaired)	0.85 (0.69, 1.06)	0.148
30-day readmission	fTRST (per 1-point increase)	1.07 (0.91, 1.26)	0.393	1.08 (0.90, 1.30)	0.401
	Tumour localization (rectum versus colon)	1.95 (1.22, 3.03)	0.004	2.08 (1.16, 3.59)	0.010
	Sex (male versus female)	2.02 (1.33, 3.10)	0.001	2.16 (1.33, 3.62)	0.002
	Age (per 1-year increase)	0.96 (0.92, 0.99)	0.021	0.96 (0.92, 1.01)	0.111
Discharge to non-independent setting	fTRST (per 1-point increase)	1.34 (1.21, 1.51)	< 0.001	1.40 (1.25, 1.59)	< 0.001
	Tumour localization (rectum versus colon)	1.37 (1.01, 1.86)	0.040	1.55 (1.08, 2.26)	0.019
	Sex (male versus female)	0.78 (0.62, 0.98)	0.033	0.78 (0.59, 1.02)	0.066
	Age (years)	1.05 (1.03, 1.08)	< 0.001	1.05 (1.02, 1.08)	< 0.001
	CACI	1.18 (1.09, 1.27)	< 0.001	1.18 (1.09, 1.28)	< 0.001
	Urgent surgery	2.24 (1.60, 3.17)	< 0.001	2.43 (1.65, 3.63)	< 0.001
	ECOG PS (≥ 2 versus < 2)	2.46 (1.75, 3.48)	< 0.001	2.51 (1.74, 3.65)	< 0.001
	ASA grade (≥ III versus < III)	1.52 (1.14, 2.02)	0.004	1.62 (1.08, 2.44)	0.020
	Nutritional risk (normal versus impaired)	0.81 (0.64, 1.02)	0.079	0.82 (0.63, 1.08)	0.155

^*Values in parentheses are 95% confidence intervals. Statistical significance was set at P < 0.050. PSM, propensity score matching; LOS, length of hospital stay; OR, odds ratio; fTRST, Flemish version of the Triage Risk Screening Tool; CACI, Charlson age co-morbidity index; ECOG PS, Eastern Cooperative Oncology Group performance status; ASA, American Society of Anaesthesiologists.

Thirty-day readmission

Readmission within 30 days from the index surgery was observed in 7.0% of patients (Table 1). No significant association was found between the fTRST and readmission, either in the univariate model (OR 0.98; 95% c.i. 0.83 to 1.14; P = 0.793) or multivariable models (unmatched: OR 1.07 (95% c.i., 0.91 to 1.26; P = 0.393); matched cohort: OR 1.08 (95% c.i. 0.90 to 1.30; P = 0.401); Tables 2 and 4). In the multivariable models, rectal cancer and sex were significant predictors of readmission within 30 days from the index surgery.

Discharge to a non-independent setting

Discharge to a non-independent setting occurred in 46.8% of patients (Table 1). A 1-point increase in the fTRST score was strongly associated with this outcome in both the univariate analysis (OR 1.78; 95% c.i. 1.62 to 1.96; P < 0.001) and multivariable analyses in the unmatched (OR 1.35; 95% c.i. 1.21 to 1.51; P < 0.001) and matched (OR 1.40; 95% c.i. 1.25 to 1.59; P < 0.001) cohorts (Tables 2 and 4). Among the covariates, age, CACI, and urgent surgery were the strongest predictors of discharge to a non-independent setting.

Prognostic performance of the fTRST

Across the entire cohort and in the propensity score-matched analyses, an fTRST score ≥ 2 consistently emerged as a significant predictor of adverse postoperative outcomes. The strongest associations were observed for 30-day mortality (OR 1.36; 95% c.i. 1.12 to 1.66; P = 0.002), discharge to a non-independent setting (OR 1.35; 95% c.i. 1.21 to 1.51; P < 0.001), and prolonged LOS (OR 1.13; 95% c.i. 1.03 to 1.24; P = 0.010; Tables 3 and 4). The predictive performance of the fTRST was further supported by ROC curve analysis, with an AUC of 72.4% for 30-day mortality (Fig. 1). This result was accompanied by a high NPV of 98.2%, indicating a strong ability to rule out mortality in low-risk patients. Similarly, for discharge to a non-independent setting, the highest specificity (76.7%) across the considered outcomes was observed together with a moderate discriminatory ability, with an AUC of 67.2%. Furthermore, a PPV of 66.3% and an NPV of 64.5% were estimated, indicating a relevant clinical utility of the fTRST in anticipating postdischarge care needs. Interestingly, the fTRST demonstrated the lowest predictive performance for 30-day readmissions, with a sensitivity of 34.6% and an AUC of 48.8%, suggesting limited discriminatory ability. Nevertheless, a high NPV (92.8%) still highlights its potential role in excluding patients with low fTRST scores from high risk of rehospitalization. Although the PPVs were modest across outcomes, the consistently high NPVs reinforce the utility of the fTRST as a screening tool to identify patients at low risk for poor surgical and functional outcomes.

Figure 1.

Receiver operating characteristic curves for fTRST in predicting postoperative outcomes in older adults undergoing colorectal cancer surgery

Performance of the fTRST in predicting a 30–day mortality, b discharge to a non–independent setting, c LOS >5 days, d 30–day rehospitalization, and e 30–day postoperative complications. For each outcome, sensitivity, specificity, PPV, NPV, and the AUC are reported, with 95% confidence intervals in parentheses. An fTRST score ≥ 2 was considered indicative of frailty. ROC, receiver operating characteristic; fTRST, Flemish version of the Triage Risk Screening Tool; LOS, length of hospital stay; PPV, positive predictive value; NPV, negative predictive value; AUC, area under the receiver operating characteristic curve.

Discussion

Population ageing is increasingly reshaping healthcare systems. Because a large proportion of CRC patients are aged 70 years and over, it is no longer chronological age alone that determines suitability for surgery, but rather the ability to tailor treatment and perioperative pathways to the individual patient²³. With personalized treatment derived from preoperative assessment and individualized care plans, senior adults can achieve excellent surgical outcomes²⁴. As stated by Scardino et al.²⁵, the future of geriatric CRC surgery lies in systematic implementation of patient-centred frameworks, combining precise preoperative frailty assessment with clinical excellence and a deep respect for individual patient goals.

The value of holistic preoperative geriatric assessment in surgical oncology has been previously demonstrated²⁶. However, although comprehensive geriatric assessment offers detailed insights, it is often impractical for widespread use in time-sensitive surgical settings. Recently, the American Society of Colon and Rectal Surgeons reinforced the concept that treatment recommendations should consider a patient’s degree of frailty rather than chronological age²⁷. Compared with other frailty screening tools, the fTRST stands out for its brevity, ease of use, and ability to identify patients at risk of both medical and functional adverse outcomes^28–31. Moreover, the fTRST can be administered rapidly, even by non-specialist staff, and has demonstrated comparable or superior predictive performance in several settings^32,33.

In this study, the fTRST demonstrated strong and consistent prognostic value for both surgical and functional outcomes. A score ≥ 2 was independently associated with increased 30-day mortality, postoperative complications, prolonged LOS, and discharge to a non-independent setting. The fTRST also showed strong NPVs for 30-day mortality (98.2%) and discharge to a non-independent setting (64.5%), suggesting that it may be particularly useful in identifying low-risk patients who may benefit from standard perioperative pathways without the need for additional geriatric interventions. Conversely, patients with an fTRST score ≥ 2 may warrant more comprehensive perioperative care. This includes preoperative optimization, closer postoperative monitoring, and possible early involvement of multidisciplinary care teams that include geriatricians³⁴. After discharge, patients with an fTRST score ≥ 2 may benefit from postoperative clinical follow-up, which can result in improved survivorship: a recent retrospective cohort study showed that oncogeriatric patients who received a primary care physician visit after discharge had lower 90-day mortality compared with those who did not³⁵.

Interestingly, the fTRST was not significantly associated with 30-day readmission in the present study. This could reflect the fact that surgical factors could represent a stronger element for this specific outcome. Across studies, several predictors of 30-day readmission after CRC surgery have been consistently identified: a recent meta-analysis³⁶ highlights factors such as advanced disease, the presence of a stoma, multiple co-morbidities, and open surgery as increasing the risk of readmission. The present study confirmed that rectal cancer patients have a higher risk of unplanned readmission, probably due to stoma-related issues and pelvic complications, as widely reported in the literature³⁷. Socioeconomic status also plays a substantial role in unplanned readmission: a retrospective cohort study³⁸ of 207 299 CRC surgery patients found a significant association between socioeconomic factors and 30-day readmission. This suggests that social frailty may contribute to readmissions independent of clinical fitness.

The brevity, ease of use, and strong NPV of the fTRST make it particularly suitable for routine preoperative assessment in geriatric CRC surgery. Its implementation may support more personalized perioperative care strategies, including early geriatric referral, prehabilitation, and tailored discharge planning. Future prospective studies should explore the integration of the fTRST into surgical decision-making algorithms and assess its impact on patient-reported outcomes.

The strengths of this study include the use of a large, population-based data set, which enhances the generalizability of the findings. The use of PSM further strengthens the validity of the observed associations. However, this study still has limitations. First, its retrospective design may introduce residual confounding despite statistical adjustment. Second, functional outcomes were inferred from discharge disposition rather than direct measures of independence or quality of life. Finally, social determinants were not captured by the ESCA.

Collaborators

Members of the ESCA Collaborative Group are: G. Frascaroli (‘Santa Maria delle Croci’ Hospital, Ravenna, Italy); D. Zattoni (‘Santa Maria delle Croci’ Hospital, Ravenna; Department of Medical and Surgical Sciences (DIMEC), Bologna, Italy); A. Romboli, M. Giuffrida (‘Guglielmo Da Saliceto’ Hospital, Piacenza, Italy); G. Stacchini, R. Bova (‘Bufalini’ Hospital, Cesena, Italy); F. Pasini (‘Degli Infermi’ Hospital, Faenza, Italy); L. Solaini (‘Morgagni – Pierantoni’ Hospital, Forlì; Department of Medical and Surgical Sciences (DIMEC), Bologna, Italy); A. Avanzolini (‘Morgagni – Pierantoni’ Hospital, Forlì, Italy); B. Pirrera, M. Portinari (‘Infermi’ Hospital, Rimini, Italy); A. Cacurri, G. Vitali (‘Ceccarini’ Hospital, Riccione, Italy); M. T. Montella (Healthcare Administration, ‘Istituto Nazionale Tumori’, Milan, Italy); M. Altini (Healthcare Administration, Modena Local Health Authority, Modena, Italy)

Funding

This work was supported, in part, by the Italian Ministry of Health ‘Ricerca Corrente’.

Author contributions

Federico Ghignone (Conceptualization, Writing—original draft, Writing—review & editing), Giampaolo Ugolini (Conceptualization, Writing—review & editing), Isacco Montroni (Conceptualization, Writing—review & editing), Filippo Banchini (Supervision, Writing—review & editing), Fausto Catena (Supervision, Writing—review & editing), Davide Cavaliere (Supervision, Writing—review & editing), Giorgio Ercolani (Supervision, Writing—review & editing), Gianluca Garulli (Supervision, Writing—review & editing), Andrea Lucchi (Supervision, Writing—review & editing), Valentina Danesi (Writing—original draft, Writing—review & editing), Ilaria Massa (Conceptualization, Supervision, Writing—review & editing), Parisa Soleimani (Writing—original draft, Writing—review & editing), William Balzi (Resources, Software, Writing—review & editing), Alice Andalò (Resources, Software, Writing—review & editing), Martina Cavallucci (Resources, Software, Writing—review & editing), Nicola Gentili (Resources, Software, Writing—review & editing), and Andrea Roncadori (Conceptualization, Formal analysis, Methodology, Visualization, Writing—original draft, Writing—review & editing)

Disclosures

The authors declare no conflict of interest.

Supplementary material

Supplementary material is available at BJS Open online.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.