Trajectories and factors associated with bowel symptoms after sphincter-preserving surgery for rectal cancer: a longitudinal study

Patients undergoing sphincter-preserving surgery for rectal cancer experience persistent bowel symptoms that negatively affect physical, psychological, and social functioning. Understanding heterogeneous trajectories of bowel symptoms and their predictors is essential for early nursing risk stratification, individualized symptom management, and survivorship care in rectal cancer patients. We conducted a prospective longitudinal study of 480 patients who underwent sphincter-preserving surgery for rectal cancer at a tertiary hospital in China. Bowel symptom scores were assessed at 1, 3, 6, 9, and 12 months after tumor resection or stoma reversal. Latent class growth modeling (LCGM) was applied to identify trajectory subgroups. Factors associated with trajectory membership were examined using classification-error-adjusted three-step multinomial logistic regression. Exploratory CHAID analyses were conducted separately in ten multiply imputed datasets and internally evaluated using 10-fold cross-validation. A total of 464 patients were included in the analysis. Three distinct trajectories of bowel symptoms were identified: high symptom-persistent disorder group (16.8%), moderate symptom-slow decline group (33.0%), and low symptom-rapid decline group (50.2%). Classification-error-adjusted multinomial regression showed that low/ultra-low rectal position and TNM stage III were significantly associated with membership in the moderate symptom–slow decline trajectory, whereas tumor length and tumor width were significantly associated with membership in the high symptom–persistent disorder trajectory. Across the ten imputed datasets, tumor location and defunctioning stoma were the most stable CHAID splitting variables, appearing in all ten trees, whereas tumor width and TNM stage were retained less consistently. Ten-fold cross-validated classification accuracy ranged from 46.1% to 55.2%, indicating modest classification performance. Subgroup analysis showed that patients with a defunctioning stoma or neoadjuvant therapy had higher proportions in the high symptom–persistent disorder trajectory than those without a defunctioning stoma or neoadjuvant therapy (22.2% vs. 10.4% and 25.0% vs. 15.6%, respectively). Three distinct trajectories of bowel symptoms were identified in patients after sphincter-preserving surgery for rectal cancer. Several disease- and treatment-related factors were associated with trajectory membership. Exploratory CHAID analyses showed a reproducible upper-level hierarchical pattern involving tumor location and defunctioning stoma, although cross-validated classification performance was modest. These findings may inform future development of postoperative nursing risk-assessment strategies, but the decision-tree structures and data-derived split points require further validation before clinical application. The clinical trial registry is Chinese Clinical Trial Registry. Registration Date: July 9, 2021. Registration number: ChiCTR2100048467. Website: https://www.chictr.org.cn/searchproj.html .