Effect of pre-diagnostic NSAID use on ovarian cancer survival

Background
There is some evidence in breast, colorectal and prostate cancer that patients who use nonsteroidal anti-inflammatory (NSAID) drugs have better survival. There is conflicting evidence of benefit in ovarian tumours. Our objective was to determine the association of pre-diagnostic NSAID use with survival in ovarian cancer patients in Ireland.

Method
Women diagnosed with invasive ovarian cancer (ICD-10 code C56) between 2001-2011 were identified in a retrospective cohort study from the National Cancer Registry. Those with continuous eligibility for a (means-tested) medical card in the year immediately prior to diagnosis were identified and linked to community prescription records. Any NSAID prescription in the year prior to diagnosis was determined. Date and cause of death was sourced from linked death-certificates. Association between exposure and cause-specific survival (end of follow-up: 31/12/2012) was estimated using Cox regression (adjusted for: age, smoking, marital status, year of diagnosis, urban/rural, local area deprivation, ovarian cancer stage and grade, surgery at diagnosis). Secondary analysis accounting for competing risks was conducted. Analysis by type of NSAID (aspirin/other) was pre-planned.

Results
Of 3097 invasive ovarian cancers diagnosed 2001-2011, 1823 (59%) had had a medical card history for at least one year prior to diagnosis and, of these, 1123 (62%) had some exposure to NSAIDs in that year. 78% of women in the cohort had died by 31/12/2012 (median follow-up=5.8years). Pre-diagnostic NSAID use was associated with improved ovarian cancer-specific survival (AHR=0.84, 95%CI 0.74, 0.96) as well as other causes (AHR=0.62, 95%CI 0.42, 0.93). Adjusting for competing risks, NSAID use was no longer significant (cancer-specific: 0.90 (0.78, 1.05), other causes: 0.75, (0.5, 1.11)). Effects were similar for aspirin and other NSAIDs.

Conclusion
In this, the largest ever study of NSAID use in ovarian cancer, we observed an association between pre-diagnostic NSAID use and cancer specific survival. The association was no longer observed after adjusting for competing risks.

Acknowledgements
Project funding, Health Research Board Ireland; Registry funding, Department of Health Ireland.