Association of statin use with a pathologic complete response to neoadjuvant chemoradiation for rectal cancer

To assess whether 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors, or statins, might enhance the efficacy of neoadjuvant chemoradiation in rectal cancer.

Between 1996 and 2001, 358 patients with clinically resectable, nonmetastatic rectal cancer underwent surgery at Memorial Sloan-Kettering Cancer Center after neoadjuvant chemoradiation for either locally advanced tumors or low-lying tumors that would require abdominoperineal resection. We excluded 9 patients for radiation therapy dose <45 Gy or if statin use was unknown, leaving 349 evaluable patients. Median radiation therapy dose was 50.4 Gy (range, 45-55.8 Gy), and 308 patients (88%) received 5-fluorouracil-based chemotherapy. Medication use, comorbid illnesses, clinical stage as assessed by digital rectal examination and ultrasound, and type of chemotherapy were analyzed for associations with pathologic complete response (pCR), defined as no microscopic evidence of tumor. Fisher's exact test was used for categoric variables, Mantel-Haenszel test for ordered categoric variables, and logistic regression for multivariate analysis.

Thirty-three patients (9%) used a statin, with no differences in clinical stage according to digital rectal examination or ultrasound compared with the other 324 patients. At the time of surgery, 23 nonstatin patients (7%) were found to have metastatic disease, compared with 0% for statin patients. The unadjusted pCR rates with and without statin use were 30% and 17%, respectively (p = 0.10). Variables significant univariately at the p = 0.15 level were entered into a multivariate model, as were nonsteroidal anti-inflammatory drugs (NSAIDs), which were strongly associated with statin use. The odds ratio for statin use on pCR was 4.2 (95% confidence interval, 1.7-12.1; p = 0.003) after adjusting for NSAID use, clinical stage, and type of chemotherapy.

In multivariate analysis, statin use is associated with an improved pCR rate after neoadjuvant chemoradiation for rectal cancer. The low prevalence of statin use limits the power to detect a significant difference at a type I error threshold of p = 0.05 in this analysis. Although no definitive conclusions can be drawn on the basis of this retrospective study, the unusually high incidence of pCR after chemoradiation suggests that the use of statins in the treatment of rectal cancer warrants further evaluation.