Random vs Targeted Urine Drug Testing Among Patients Undergoing Long-term Opioid Treatment for Cancer Pain

There is limited information on best practices regarding urine drug testing (UDT) during long-term opioid therapy for cancer pain.¹ Practices involve random UDT¹ or UDT based on risk for nonmedical opioid use.² The main objective of this study was to report the frequency of abnormal UDT findings among patients who underwent random vs targeted testing.

Between April 1, 2017, and January 31, 2018, electronic medical records of consecutive patients attending an outpatient supportive care clinic who underwent random UDT were reviewed. Eligible patients were 18 years or older, had a diagnosis of cancer, and were receiving opioid therapy. The results were compared with a cohort of consecutive patients who underwent targeted UDT wherein the test was ordered based on clinical suspicion of nonmedical opioid use. The institutional review board at the University of Texas MD Anderson Cancer Center approved this study and granted waiver of informed consent because individuals were at no more than minimal risk because of the retrospective nature of this study. All patients provided information regarding their opioid intake prior to UDT to assist in result interpretation. The Mayo Medical Laboratories test used in this study consisted of screening immunoassay confirmed by gas chromatography–mass spectrometry whenever a positive screening result was obtained.

Analysis began June 2018. χ² Test or Fisher exact and Wilcoxon rank sum test were used to compare categorical or continuous variables, respectively. Regression analyses were used to explore factors associated with abnormal UDT findings. All computations were carried out in SAS, version 9.4 (SAS Institute) and TIBCO Spotfire S+ 8.2 (TIBCO). A 2-sided P of .05 was considered statistically significant.

Overall, 59 of 212 individuals (28%) in the random group and 38 of 88 individuals (43%) in the targeted UDT had abnormal findings (P = .01; χ² = 6.699). The median (interquartile range) age was 60 (48-63) years in the random cohort and 53 (40-61) in the targeted cohort. The random cohort included 112 women (53%), and the targeted cohort included 45 women (51%). When marijuana was excluded from the list of abnormal results, 33 (16%) in the random group and 29 (33%) in the targeted group had abnormal UDT findings (P < .001; χ² = 11.468). The Table summarizes the regression analysis of factors associated with abnormal UDT findings among the random cohort. There was no significant association between time from cancer diagnosis and abnormal UDT. A summary of the types and distribution of UDT abnormalities in both cohorts is presented in the Figure.

In this study, more than 1 in every 4 patients with cancer receiving long-term opioid therapy had an abnormal random UDT finding, consistent with nonmedical use of opioids. When marijuana was excluded from the list of urine abnormalities, the rate of nonmedical opioid use remained considerably high at 1 in 6 patients. To our knowledge, our study is the first to look at the frequency of UDT abnormality among a random sample of patients who were receiving opioids for cancer pain. All other studies were conducted among a selected sample of patients who already had a known elevated risk for nonmedical opioid use,³ raising questions about potential bias. Currently, there are no standard guidelines regarding the frequency and timing of UDT ordering in patients with cancer.⁴ Although the frequency of abnormal UDT in the random group was significantly lower than in the targeted group, it was still quite high, suggesting that routine random monitoring may be justified among patients with cancer receiving opioids.

In this study, we found that younger age, male sex, and higher anxiety levels were independently associated with nonmedical opioid use. Other studies have reported similar findings.^5,6

Overall, 13 (19%) random and 14 (28%) targeted abnormal UDT findings had no prescribed opioids, indicating possible opioid diversion. This is particularly concerning because such opioids may end up being used by individuals other than patients and contribute to the risk for unintentional overdose or death. The presence of marijuana in urine may be of limited importance mainly because its classification as an illicit drug is currently debatable and general perception continues to evolve. Limitations of the study include its retrospective and single-center design. Ultimately, further studies are needed to guide clinical practice regarding the use of UDT among patients treated with opioids for cancer pain.

Corresponding Author: Eduardo Bruera, MD, Department of Palliative, Rehabilitation and Integrative Medicine, Unit 1414, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77030 (ebruera@mdanderson.org).

Accepted for Publication: December 10, 2019.

Published Online: February 6, 2020. doi:10.1001/jamaoncol.2019.6756

Author Contributions: Dr Arthur 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.

Concept and design: Arthur, Nguyen, Hui, Bruera.

Acquisition, analysis, or interpretation of data: Arthur, Lu, Nguyen, Prado, Hui, Edwards.

Drafting of the manuscript: Arthur, Edwards, Bruera.

Critical revision of the manuscript for important intellectual content: Arthur, Lu, Nguyen, Prado, Hui, Bruera.

Statistical analysis: Arthur, Lu.

Administrative, technical, or material support: Arthur, Lu, Nguyen, Prado, Edwards, Bruera.

Supervision: Arthur, Bruera.

Conflict of Interest Disclosures: Dr Hui reports grants from Teva Pharmaceutical, Insys Therapeutics, and Helsinn Healthcare outside the submitted work. Dr Bruera reports grants from Helsinn Healthcare outside the submitted work. No other disclosures were reported.

Additional Contributions: We acknowledge Diane Liu, MS (Department of Biostatistics, the University of Texas MD Anderson Cancer Center, Houston), for her assistance with data analysis and interpretation. No compensation was received.