Rectal Cancer Survivors Follow Three Distinct Bowel Symptom Paths After Sphincter-Preserving Surgery

For the hundreds of thousands of people who undergo sphincter-preserving surgery for rectal cancer each year, surviving the disease is only the beginning of a long and often invisible struggle. The operation, which removes the tumor while carefully preserving the anal sphincter complex, spares patients a permanent colostomy but frequently leaves behind a constellation of bowel symptoms known collectively as low anterior resection syndrome. Urgency, frequency, fragmentation of bowel movements, incontinence, and unpredictable clustering of evacuations can persist for months or years, eroding physical comfort, psychological resilience, and social confidence. A new prospective longitudinal study from West China Hospital of Sichuan University, published in BMC Nursing, has now tracked how these symptoms unfold over the first twelve months after surgery in one of the largest cohorts assembled to date, and its findings suggest that recovery is not a single road but three very different journeys.

The research team, led by Yaolan Wang and Mingjun Huang of the West China School of Nursing, enrolled 480 patients who underwent sphincter-preserving surgery for rectal cancer at a tertiary hospital in China. Bowel symptom scores were assessed at one, three, six, nine, and twelve months after tumor resection or, in patients who received a temporary defunctioning stoma, after stoma reversal. Of the original cohort, 464 patients provided complete data for analysis. Rather than simply averaging symptom scores across the group, the investigators applied latent class growth modeling, a statistical technique designed to uncover hidden subpopulations within longitudinal data. The method searches for distinct trajectory shapes that, when combined, best explain the observed pattern of symptom change over time, using fit indices such as the Akaike Information Criterion and the Bayesian Information Criterion alongside likelihood ratio tests to determine how many trajectory classes the data genuinely support.

The answer was three. Half of the patients, 50.2 percent, fell into a low symptom–rapid decline group, starting with relatively mild bowel dysfunction and improving quickly over the first year. A third, 33.0 percent, belonged to a moderate symptom–slow decline group, beginning with more substantial symptoms that eased only gradually. Most strikingly, 16.8 percent of patients formed a high symptom–persistent disorder group, in which severe bowel dysfunction showed little meaningful improvement across the entire twelve-month observation window. For this minority, roughly one in six patients, the first year after sphincter-preserving surgery was dominated by symptoms that refused to fade, a pattern the authors say carries serious implications for physical, psychological, and social functioning.

Identifying the trajectories was only the first step. The team then asked which factors, measured before or during treatment, predicted which path a patient would follow. Because trajectory membership in latent class growth modeling is itself an estimated quantity rather than a directly observed label, the researchers used a classification-error-adjusted three-step multinomial logistic regression, a method that accounts for the uncertainty in assigning patients to classes and thereby produces more honest odds ratios than naive approaches. The analysis revealed that patients with low or ultra-low rectal tumors and those with TNM stage III disease were significantly more likely to belong to the moderate symptom–slow decline trajectory. Meanwhile, tumor length and tumor width were significantly associated with membership in the high symptom–persistent disorder group, the trajectory with the worst outlook.

Subgroup comparisons sharpened the picture further. Patients who had received a defunctioning stoma, a temporary diverting stoma created to protect the surgical anastomosis while it heals, were more than twice as likely to end up in the high symptom–persistent disorder group compared with those who did not, at 22.2 percent versus 10.4 percent. Similarly, patients who had undergone neoadjuvant therapy, radiation or chemotherapy delivered before surgery to shrink the tumor, showed higher proportions in the worst trajectory than those who had not, at 25.0 percent versus 15.6 percent. These figures align with the established biology of low anterior resection syndrome: neoadjuvant radiation damages the sphincter apparatus and pelvic nerves, ultra-low anastomoses disrupt the rectal reservoir function that normally buffers stool delivery to the anal canal, and prolonged stoma diversion delays the natural adaptation of the reconstructed bowel.

The study went beyond regression, adding an exploratory machine learning layer using Chi-squared Automatic Interaction Detection, or CHAID, a decision-tree algorithm that recursively splits patients into branches based on the variables most strongly associated with symptom outcomes. Because the dataset contained missing values, the researchers conducted the CHAID analyses separately in ten multiply imputed datasets and internally evaluated each tree using ten-fold cross-validation, a technique that repeatedly holds out portions of the data to test whether the derived structure generalizes. Across the ten imputed datasets, tumor location and defunctioning stoma emerged as the most stable splitting variables, appearing in all ten decision trees, whereas tumor width and TNM stage were retained less consistently. This reproducible upper-level hierarchy suggests that where the tumor sits in the rectum and whether a temporary stoma was created are the two most informative early questions for stratifying bowel symptom risk.

Yet the authors are careful not to oversell the machine learning component. Cross-validated classification accuracy ranged from 46.1 percent to 55.2 percent, a modest performance that falls well short of the reliability needed for clinical deployment. In practical terms, the decision trees could reproduce a broad hierarchical pattern involving tumor location and defunctioning stoma, but their data-derived split points and lower-level branches were not stable enough to serve as a bedside risk calculator. The researchers explicitly state that the tree structures require further validation before any clinical application, a caveat that reflects a growing awareness in the symptom science community that internally derived models often overfit the cohorts from which they emerge.

What makes the study valuable despite these limitations is its framing of bowel dysfunction as a heterogeneous, trajectory-based phenomenon rather than a uniform complication. The authors argue that understanding these distinct symptom paths is essential for early nursing risk stratification, individualized symptom management, and survivorship care in rectal cancer patients. In the current clinical landscape, follow-up after sphincter-preserving surgery tends to be organized around cancer surveillance, with bowel function assessed opportunistically if at all. A trajectory-based framework flips that logic: if a patient’s tumor position, stage, tumor dimensions, stoma status, and neoadjuvant history can flag them at the outset as likely to follow the high symptom–persistent disorder path, targeted bowel rehabilitation, pelvic floor physiotherapy, dietary counseling, and structured nursing support could be deployed in the first weeks after surgery rather than after a year of preventable suffering.

The implications reach beyond nursing. Low anterior resection syndrome has been shown in prior research to affect quality of life as profoundly as some cancer recurrences, yet it remains under-recognized by both patients and clinicians, many of whom assume that symptoms will simply resolve with time. This study demonstrates that for one in six patients, that assumption is wrong: twelve months after surgery, their symptoms remain severe. The finding that defunctioning stoma and neoadjuvant therapy roughly double or increase the odds of the worst trajectory gives surgeons and oncologists concrete information for preoperative conversations, allowing patients to anticipate the realistic range of functional outcomes before consenting to a particular treatment sequence.

The work, registered prospectively in the Chinese Clinical Trial Registry in July 2021 and funded by the National Natural Science Foundation of China, arrives at a moment when survivorship care for colorectal cancer is expanding worldwide. As more patients live longer after sphincter-preserving surgery, the burden of functional sequelae will grow in absolute terms even if surgical techniques continue to improve. The three trajectories identified here, low and rapidly declining, moderate and slowly improving, and high and persistently disordered, offer a vocabulary for that burden and a starting point for the risk-assessment strategies the authors hope to develop. The decision trees are not yet ready for the clinic, the authors caution, but the underlying message is already clear: bowel recovery after rectal cancer surgery follows predictable paths, and the factors that determine which path a patient takes are largely known before the first incision is made.

Subject of Research: Health

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

Article References: Wang, Y., Lin, Y., Wang, L., Li, L., Wang, S., Wang, X., & Huang, M. (2026). Trajectories and factors associated with bowel symptoms after sphincter-preserving surgery for rectal cancer: a longitudinal study. BMC Nursing. https://doi.org/10.1186/s12912-026-05520-z

Image Credits: AI Generated

DOI: 10.1186/s12912-026-05520-z

Keywords: rectal cancer, bowel symptoms, low anterior resection syndrome, sphincter-preserving surgery, longitudinal study, latent class growth model, survivorship, nursing research, Trajectories, factors, associated, bowel