[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Views 107
Citations 0
Comment & Response
March 19, 2020

Questionable Survival Benefit of Aspirin Use in Patients With Biliary Tract Cancer—Reply

Author Affiliations
  • 1Division of Cancer Epidemiology and Genetics, National Cancer Institute, Rockville, Maryland
JAMA Oncol. Published online March 19, 2020. doi:10.1001/jamaoncol.2020.0125

In Reply We thank Bergquist and colleagues for their interest in our article1 on the association of aspirin use with biliary tract cancer (BTC) survival in the UK’s Clinical Practice Research Datalink (CPRD). We agree that the results cannot be the sole justification to incorporate aspirin therapy into the standard management of BTCs. For that, careful replication of the findings and, if possible, a randomized clinical trial are needed to provide further evidence. We also agree that the analysis has limitations that are typical with the use of electronic databases for observational research. We appreciate the opportunity to respond to these limitations beyond what was discussed in our article.

Although all observational studies are subject to error, not all potential sources of error will nullify the results. For instance, surgical status is not a strong enough confounder to completely explain our findings.1 Klatskin tumors are very rare, making up approximately 3% of cholangiocarcinomas2; therefore, the benefit of surgical resection for a small subset of these BTC cases is unlikely to confound the results. Furthermore, failure to account for surgery in patients with cholangiocarcinoma would not explain the findings for the other BTCs.

The survival curves we presented are a graphical presentation of the differences between 2 groups accounting for time-varying aspirin use.1 We agree that fewer events occur after 12 months than within the first year. However, this does not affect the associations represented by the hazard ratios, which should be interpreted as averages over the investigated time period. Clearly, there is no information contributing to those estimates when there are no more events the aspirin group.

To reduce misclassification bias, we required 2 or more prescriptions recorded in the CPRD to classify patients as aspirin users, which is a valid measure of adherence for chronic medications.3 Because aspirin is available over the counter in the UK, some true aspirin users may have been misclassified as nonusers. This misclassification would dilute the findings by making the 2 groups more alike, in which case the true survival benefit would be higher than what we reported.

We chose not to link to the National Cancer Data Repository or Office of National Statistics data because they only cover a percentage of the practices in England (approximately 57% of the CPRD).4 Case ascertainment is not usually a concern in studies of survival wherein both comparison groups are patients with cancer. Validation studies show high levels of concordance between the National Cancer Data Repository and the CPRD for both cancer incidence and mortality.4 On the other hand, a linked analysis would have reduced the sample size, power, and generalizability.4 Although lack of data on disease stage is a limitation, BTCs are nearly always fatal with short survival time after diagnosis, largely owing to the late stage at diagnosis; it is likely that mortality was attributable to BTC in this analysis. Although our study1 has limitations that hamper any analysis of electronic medical data, the big advantage is that the results can truly be thought of as representative of the case population and not a selected subset.

Back to top
Article Information

Corresponding Author: Sarah S. Jackson, PhD, MPH, Division of Cancer Epidemiology and Genetics, National Cancer Institute, 9609 Medical Center Dr, Room 6-E210, Bethesda, MD 20892 (sarah.jackson@nih.gov).

Published Online: March 19, 2020. doi:10.1001/jamaoncol.2020.0125

Conflict of Interest Disclosures: None reported.

References
1.
Jackson  SS, Pfeiffer  RM, Liu  Z,  et al.  Association between aspirin use and biliary tract cancer survival.  JAMA Oncol. 2019;5(12):1802-1804. doi:10.1001/jamaoncol.2019.4328PubMedGoogle ScholarCrossref
2.
Surveillance, Epidemiology and End Results Program. SEER Stat Database: incidence—SEER 9 regs research data, Nov 2018 sub (1975-2016) Katrina/Rita population adjustment—linked to county attributes—total US, 1969-2017 counties. Published April 2019. Accessed January 16, 2020. https://seer.cancer.gov/data-software/
3.
Strom  BL, Kimmel  SE, Hennessy  S, eds.  Pharmacoepidemiology. 6th ed. Wiley-Blackwell; 2020.
4.
McDonald  L, Schultze  A, Carroll  R, Ramagopalan  SV.  Performing studies using the UK Clinical Practice Research Datalink: to link or not to link?  Eur J Epidemiol. 2018;33(6):601-605. doi:10.1007/s10654-018-0389-5PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    ×