Association Between Regulatory Advisories and Codeine Prescribing to Postpartum Women

Some patients are ultrarapid metabolizers of codeine, with prevalence ranging from 2% to 40%.¹ Nursing mothers who take codeine may be putting their infant at risk if they carry the polymorphisms for elevated activity of CYP2D6, an enzyme that metabolizes codeine to morphine. High levels of morphine in breast milk may lead to infant death from drug-induced respiratory depression.²

The US Food and Drug Administration (FDA) released a public health advisory in August 2007 warning about the potentially life-threatening adverse effects in infants of breastfeeding mothers taking codeine.³ Health Canada published a similar advisory in October 2008.⁴ We evaluated the rate of codeine dispensations among postpartum mothers before and after these 2 regulatory advisories.

We examined postpartum codeine use among all women with live births between January 1, 2002, and December 31, 2011, in British Columbia, Canada (population: 4.5 million). We used deidentified linked health data sets provided by Population Data BC⁵ with approval from relevant data stewards and the University of British Columbia’s Behavioural Research Ethics Board. Information about maternal health came from the BC Perinatal Data Registry, which includes information for 99% of all births in British Columbia. Information about prescription dispensations came from BC PharmaNet, a comprehensive database of every prescription filled outside of acute care hospitals, regardless of patient age or insurance status.

We performed an interrupted time series analysis⁶ to measure changes in dispensation rates of codeine and other opioid drugs during the first 6 months of the postpartum period. We used data for cephalexin and hydrocortisone cream (2 commonly prescribed drugs during the postpartum period) as controls. We tested for changes in monthly levels of and trends in postpartum use of codeine, other opioids, and controls. The best fitting models were seasonally adjusted, autoregressive-moving average processes. We used 2-sided tests for significance at the .05 confidence level and performed all analyses using Stata version 13 (StataCorp) and EViews version 8 (IHS Global Inc).

There were 320 351 live births to 225 532 women. Of these, new mothers filled at least 1 codeine prescription during 47 095 postpartum periods. Before the FDA advisory, the monthly average for the proportion of postpartum mothers filling at least 1 codeine prescription was 16.7% (95% CI, 16.3%-17.1%). It declined to a monthly average of 9.1% (95% CI, 8.8%-9.4%) from September-December 2011, which is a 45% relative reduction over 4 years (Figure 1).

Time series analysis showed a rapid, consistent, and statistically significant change in trend (P < .001) following the FDA announcement. The Health Canada announcement was not associated with any further significant reductions in codeine dispensations; however, the US and Canadian warnings were too close in time to separate effects. The rates of prescription filling for cephalexin (P = .56 for trend) and hydrocortisone (P = .87 for trend) did not significantly change throughout the study period.

At the same time, postpartum use of tramadol, hydromorphone, oxycodone, and morphine increased (P < .001 for trend) but the magnitude of the increases was not enough to offset the reduction in codeine use (Figure 2).

The public health advisories from the FDA and Health Canada were associated with significant reductions in the codeine dispensation rate to postpartum women, possibly reflecting the adoption of the recommendations by clinicians. Some of the reduction could also reflect changes in patient behavior, including not filling written prescriptions, not asking for codeine, or both. The speed and the magnitude of the response suggests that regulatory warnings regarding postpartum drug safety can have a strong influence on prescribing patterns. An increase in postpartum use of other opioids may indicate substitution.

Because an association between health advisories and codeine prescribing has not been documented previously, the generalizability of our findings to other jurisdictions is uncertain. Important study limitations include not capturing drugs dispensed in hospitals. We could not ascertain if the women were actually breastfeeding at the time of codeine dispensation. In addition, we used data on drugs dispensed, which is not the same as drugs used. However, there is no reason to believe that the proportion of filled prescriptions that are not taken would systematically differ over time.

Corresponding Author: Kate Smolina, PhD, Centre for Health Services and Policy Research, School of Population and Public Health, University of British Columbia, 215-2206 E Mall, Vancouver, BC V6T 1Z3, Canada (kate.smolina@ubc.ca).

Author Contributions: Dr Smolina had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Smolina, Morgan, Ross, Carleton.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Smolina, Weymann.

Critical revision of the manuscript for important intellectual content: Morgan, Ross, Carleton.

Statistical analysis: Smolina, Weymann, Morgan.

Obtained funding: Smolina, Morgan.

Administrative, technical, or material support: Weymann, Morgan.

Study supervision: Morgan, Carleton.

Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.

Funding/Support: This study was supported by Canadian Institutes for Health Research (CIHR) grant DC0190GP. Dr Smolina is funded in part by a CIHR Banting postdoctoral fellowship.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.