Use of antidepressants and risk of colorectal cancer: a nested case-control study

Summary

Background

Animal studies suggest that selective serotonin reuptake inhibitors (SSRI) retard the growth of colorectal tumours, whereas tricyclic antidepressants increase the risk of colorectal cancer. We aimed to assess whether SSRI use was associated with a decreased risk of colorectal cancer, and tricyclic-antidepressant use with an increased risk of colorectal cancer.

Methods

We did a population-based nested case-control study from Jan 1, 1981, to Dec 31, 2000, of people aged 5–85 years who were registered with Saskatchewan Health and eligible for prescription-drug benefit. Between Jan 1, 1981, and Dec 31, 2000, 6544 cases with colorectal cancer were identified from the Saskatchewan Cancer Agency registry and analysed for use of tricyclic antidepressants; between Jan 1, 1991, and Dec 31, 2000, 3367 cases with colorectal cancer were identified from the Saskatchewan Cancer Agency registry and analysed for SSRI use. For every case, four eligible controls matched for age, sex, and calendar time (ie, free of any cancer in calendar month of case diagnosis) were selected randomly by a statistician who used incidence density sampling. By use of conditional logistic regression, we assessed incidence-rate ratios of having colorectal cancer in association with use of antidepressants, analysing dose and time of use.

Findings

A decreased risk of colorectal cancer was associated with high (ie, >6·0×10⁻⁶ mol per day) daily SSRI dose during 0–5 years before diagnosis (incidence-rate ratio 0·70 [95% CI 0·50–0·96], p for trend=0·0172), adjusted for age, sex, use of non-steroidal anti-inflammatory drugs in the same period, and SSRI use during 6–10 years before index date (ie, date of diagnosis for a case and the same date for matched controls). No consistent relation was recorded for risk of colorectal cancer and use of tricyclic antidepressants.

Interpretation

SSRI use might inhibit the growth of colorectal tumours through an antipromoter effect or direct cytotoxic effect. Further investigation is needed, with more complete assessment of confounders such as lifestyle factors (eg, diet), use of drugs, and comorbidity (eg, diabetes or inflammatory bowel disease) that might affect the occurrence of colorectal cancer.

Introduction

Antidepressants are used widely in the general population1, 2 and in patients with cancer³ to treat various disorders⁴ such as depression, anxiety, insomnia, chronic pain (including cancer pain), and hot flushes in menopausal women and survivors of cancer. In the USA, the annual total number of prescriptions for antidepressants is about 100 million.⁵ In Canada, the total number of prescriptions for antidepressants increased from 3·2 million to 14·5 million between 1981 and 2000.² The possible carcinogenic effect of antidepressants⁶ has received substantial attention; however, evidence remains inconclusive. Some epidemiological studies recorded an increased risk of cancer with use of antidepressants,7, 8, 9 but others10, 11, 12, 13, 14, 15, 16 did not. The main cancer sites investigated include breast, ovarian, or all cancer. To our knowledge, no epidemiological study has assessed the specific association between antidepressant use and colorectal cancer.

Since their introduction in the late 1980s,² selective serotonin reuptake inhibitors (SSRI) have become the most frequently used antidepressants.¹⁷ These inhibitors slow the growth of human colorectal tumours propagated as xenografts in mice, and suppress cell division in chemically induced colon tumours in rats.¹⁸ These results are consistent with previous findings,19, 20, 21 suggesting that serotonin promotes cell proliferation in colorectal carcinomas, and that the inhibition of serotonin uptake inhibits cell proliferation in such tumours. Serotonin was thought previously²² to be a carcinogenic substance with a role in colon-cancer formation. More recently, a study²³ showed that the SSRI drug fluoxetine decreased the viability and growth of a cultured human colon cell, probably through a direct cytotoxic effect. We aimed to test the a-priori hypothesis that SSRI use was associated with a decreased risk of colorectal cancer.

Tricyclic antidepressants are another first-line class of agents. Animal studies24, 25 have suggested that these agents enhance carcinogenesis in the colon by accelerating the proliferation of intestinal epithelial cells. Van Schaik and Graf26, 27 classified tricyclic antidepressants into genotoxic and non-genotoxic using a genotoxicity assay of wing development in Drosophila melanogaster that aimed to identify potential carcinogens. A post-hoc analysis⁹ of a population-based nested case-control study⁹ recorded an incidence-rate ratio of 2·02 (95% CI 1·34–3·04) for breast cancer and use of high-dose genotoxic tricyclic antidepressants during 11–15 years before diagnosis. Although these data were not confirmed in a subsequent study²⁸ of similar design, we aimed to assess the a-priori hypothesis that use of tricyclic antidepressants—especially those that are genotoxic—was associated with increased risk of colorectal cancer.