Prescription Opioid Duration, Dose, and Increased Risk of Depression in 3 Large Patient Populations

Jeffrey F. Scherrer; Joanne Salas; Laurel A. Copeland; Eileen M. Stock; Brian K. Ahmedani; Mark D. Sullivan; Thomas Burroughs; F. David Schneider; Kathleen K. Bucholz; Patrick J. Lustman

doi:10.1370/afm.1885

Article Figures & Data

Figures

Tables
Additional Files

Download figure
Open in new tab
Download powerpoint
Figure 1
Eligibility criteria for Veterans Hospital Administration patient population.
HIV = human immunodeficiency virus.
Download figure
Open in new tab
Download powerpoint
Figure 2
Eligibility criteria for Baylor Scott & White patient population.
HIV = human immunodeficiency virus.
Download figure
Open in new tab
Download powerpoint
Figure 3
Eligibility criteria Henry Ford Health System patient population.
HIV = human immunodeficiency virus.

View popup

Table 1

Distribution of Opioid Exposure and Covariates by Health Care Organization

Variable	VHA (n = 70,997)	BSWH (n = 13,777)	HFHS (n = 22,981)
Age, mean (SD), y	55.4 (13.0)	44.6 (15.1)	47.7 (14.4)
Sex, male, No. (%)	66,697 (93.9)	5,163 (37.5)	9,167 (39.9)
Race, white, No. (%)	53,790 (75.8)	9,542 (69.3)	10,696 (46.5)
Insurance
VHA only, No. (%)	46,196 (65.1)	…	…
Medicare, No. (%)	…	1,846 (13.4)	170 (0.7)
Married, No. (%)	39,076 (55.0)	…	…
Opioid use duration
1 to 30 d, No. (%)	55,306 (77.9)	12,817 (93.0)	20,487 (89.1)
31 to 90 d, No. (%)	8,640 (12.2)	765 (5.6)	1,685 (7.3)
>90 d, No. (%)	7,051 (9.9)	195 (1.4)	809 (3.5)
Last MED dose^a
1 to 50 mg/d, No. (%)	66,601 (93.8)	12,893 (93.6)	20,006 (87.1)
51 to 100 mg/d, No. (%)	3,613 (5.1)	706 (5.1)	2,628 (11.4)
>100 mg/d, No. (%)	783 (1.1)	178 (1.3)	347 (1.5)
Psychiatric comorbidity
Posttraumatic stress disorder, No. (%)	10,357 (14.6)	35 (0.3)	123 (0.5)
Other anxiety, No. (%)^c	8,383 (11.8)	771 (5.6)	2,057 (9.0)
Nicotine dependence/history of smoking, No. (%)	30,175 (42.5)	1,070 (7.8)	5,274 (22.9)
Alcohol abuse/dependence, No. (%)	14,313 (20.2)	234 (1.7)	1,010 (4.4)
Any illicit drug abuse/dependence, No. (%)	8,727 (12.3)	101 (0.7)	625 (2.7)
Metabolic/cardiovascular comorbidity^b
Diabetes type 2, No. (%)	29,946 (42.2)	2,732 (19.8)	7,101 (30.9)
Hypertension, No. (%)	58,616 (82.6)	6,845 (49.7)	15,857 (69.0)
Cardiovascular disease, No. (%)	61,380 (86.5)	8,161 (59.2)	17,444 (75.9)
Cerebrovascular disease, No. (%)	14,938 (21.0)	1,370 (9.9)	3,560 (15.5)
Obesity diagnosis, No. (%)	27,621 (38.9)	2,640 (19.2)	6,522 (28.4)
Other comorbidity^b
Low testosterone, No. (%)	2,171 (3.1)	219 (1.6)	313 (1.4)
Sleep apnea, No. (%)	7,563 (10.6)	1,196 (8.7)	1,583 (6.9)
Painful condition^b
Arthritis, No. (%)	58,504 (82.4)	8,539 (62.0)	18,381 (80.0)
Back pain, No. (%)	46,592 (65.6)	7,323 (53.2)	15,915 (69.3)
Headache, No. (%)	14,389 (20.3)	3,631 (26.4)	8391 (36.5)
Musculoskeletal pain, No. (%)	44,817 (63.1)	9,149 (66.4)	19,273 (83.9)
Neuropathic pain, No. (%)	22,995 (32.4)	2,391 (17.4)	5,212 (22.7)
Maximum pain score, mean (SD)	8.4 (2.2)	…	…
Health care utilization
≤25th percentile, No. (%)	6,937 (9.8)	3,134 (22.8)	4,379 (19.1)
26 to 50th percentile, No. (%)	19,928 (28.1)	4,394 (31.9)	…
51 to 75th percentile, No. (%)	24,540 (34.6)	2,758 (20.0)	13,620 (59.3)
>75th percentile, No. (%)	19,592 (27.6)	3,491 (25.3)	4,982 (21.7)

BSWH = Baylor Scott & White Health; GAD = generalized anxiety disorder; HFHS = Henry Ford Health System; MED = morphine equivalent dose; VHA = Veterans Health Administration.
↵a Opioid MED at end of follow-up period defined by new-onset depression, end, or incident opioid prescription, or last available observation.
↵b Comorbidities occurring before new-onset depression.
↵c Other anxiety disorders include panic disorder, obsessive-compulsive disorder, social phobia, generalized anxiety disorder, anxiety not otherwise specified.
d Cardiovascular diseases include hyperlipidemia, ischemic heart disease, diseases of pulmonary circulation, other heart disease, hypertensive heart disease, myocardial infarction.

View popup

Table 2

Association Between Duration of Incident Opioid Use and New-Onset Depression, Unweighted and Weighted by Inverse Probability Duration Exposure

Variable	VHA Patients			BSWH Patients			HFHS Patients

	Model 1^a HR (95% CI)	Model 2^b HR (95% CI)	Model 3^c HR (95% CI)	Model 1^a HR (95% CI)	Model 2^b HR (95% CI)	Model 3^c HR (95% CI)	Model 1^a HR (95% CI)	Model 2^b HR (95% CI)	Model 3^c HR (95% CI)
Opioid duration
1 to 30 d	1.00 [Reference]	1.00 [Reference]	1.00 [Reference]	1.00 [Reference]	1.00 [Reference]	1.00 [Reference]	1.00 [Reference]	1.00 [Reference]	1.00 [Reference]
31 to 90 d	1.23 (1.16–1.31)	1.20 (1.13–1.28)	1.18 (1.10–1.25)	1.22 (0.97–1.52)	1.31 (1.05–1.65)	1.29 (1.03–1.62)	1.39 (1.22–1.59)	1.33 (1.16–1.52)	1.33 (1.16–1.52)
>90 d	1.31 (1.22–1.40)	1.39 (1.30–1.48)	1.35 (1.26–1.44)	2.26 (1.63–3.14)	2.04 (1.38–3.02)	1.88 (1.27–2.78)	1.90 (1.62–2.24)	2.19 (1.87–2.56)	2.05 (1.75–2.40)
Arthritis	…	…	1.18 (1.12–1.25)	…	…	1.35 (1.17–1.56)	…	…	1.36 (1.22–1.51)
Back pain	…	…	1.48 (1.41–1.56)	…	…	1.43 (1.27–1.61)	…	…	1.43 (1.31–1.57)
Headache	…	…	1.59 (1.52–1.67)	…	…	1.95 (1.73–2.20)	…	…	1.52 (1.41–1.65)
Musculoskeletal pain	…	…	1.32 (1.26–1.39)	…	…	0.76 (0.66–0.88)	…	…	1.24 (1.11–1.38)
Neuropathy	…	…	1.12 (1.07–1.17)	…	…	1.41 (1.21–1.63)	…	…	1.32 (1.20–1.44)
Pain score	…	…	1.24 (1.22–1.26)	…	…	…	…	…	…

BSWH = Baylor Scott & White Health; HFHS = Henry Ford Health System; HR = hazard ratio; VHA = Veterans Health Administration.
↵a Unweighted data.
↵b Weighted data.
↵c Weighted data plus pain.

View popup

Table 3

Association Between Last Daily Morphine Equivalent Dose of Incident Opioid Use and New-Onset Depression, Unweighted and Weighted By Inverse Probability Dose Exposure

Variable	VHA Patients			BSWH Patients			HFHS Patients

	Model 1^a HR (95% CI)	Model 2^b HR (95% CI)	Model 3^c HR (95% CI)	Model 1^a HR (95% CI)	Model 2^b HR (95% CI)	Model 3^c HR (95% CI)	Model 1^a HR (95% CI)	Model 2^b HR (95% CI)	Model 3^c HR (95% CI)
Opioid dose
1 to 50 mg/d	1.00 [Reference]	1.00 [Reference]	1.00 [Reference]	1.00 [Reference]	1.00 [Reference]	1.00 [Reference]	1.00 [Reference]	1.00 [Reference]	1.00 [Reference]
51 to 100 mg/d	1.20 (1.10–1.31)	1.02 (0.93–1.12)	1.02 (0.93–1.12)	0.78 (0.60–1.03)	0.92 (0.71–1.20)	0.94 (0.72–1.21)	0.93 (0.83–1.06)	0.97 (0.86–1.10)	0.98 (0.86–1.10)
>100 mg/d	1.74 (1.49–2.04)	1.13 (0.94–1.37)	1.14 (0.94–1.39)	1.15 (0.72–1.86)	0.73 (0.41–1.31)	0.71 (0.40–1.28)	1.13 (0.84–1.53)	1.24 (0.92–1.65)	1.24 (0.92–1.65)
Arthritis	…	…	1.18 (1.11–1.25)	…	…	1.36 (1.18–1.57)	…	…	1.41 (1.27–1.56)
Back pain	…	…	1.49 (1.42–1.57)	…	…	1.47 (1.30–1.66)	…	…	1.47 (1.35–1.61)
Headache	…	…	1.58 (1.51–1.66)	…	…	1.93 (1.71–2.18)	…	…	1.53 (1.42–1.66)
Musculoskeletal pain	…	…	1.32 (1.26–1.38)	…	…	0.74 (0.64–0.85)	…	…	1.25 (1.12–1.40)
Neuropathy	…	…	1.12 (1.07–1.18)	…	…	1.43 (1.23–1.65)	…	…	1.33 (1.21–1.46)
Pain score	…	…	1.24 (1.22–1.26)	…	…	–	…	…	–

BSWH = Baylor Scott & White Health; HFHS = Henry Ford Health System; HR = hazard ratio; VHA = Veterans Health Administration.
↵a Unweighted data.
↵b Weighted data.
↵c Weighted data plus pain.

Additional Files

Figures
Tables

Supplemental Appendix

Files in this Data Supplement:

Supplemental data: Appendix - PDF file
The Article in Brief

Prescription Opioid Duration, Dose, and Increased Risk of Depression in 3 Large Patient Populations

Jeffrey F. Scherrer , and colleagues

Background Recent research suggests that the risk of new onset of depression increases with longer use of opioid medication. It is unclear, however, whether opioid-related depression is a result of the dose prescribed, the length of time the medication is used, or both. With more than 200 million prescriptions for opioids written in the United States annually, researchers investigate the association between opioid use and the risk of depression

What This Study Found Opioid-related new onset of depression is associated with duration of use but not dose.The risk of new-onset depression with 31 to 90 days of opioid use ranges from 1.18 to 1.33; and in opioid use more than 90 days, ranges from 1.35 to 2.05. Dose is not significantly associated with a new onset of depression.

Implications

According to the authors, opioids may cause short-term improvement in mood, but long-term use of more than 30 days carries a risk of new-onset depression.
The authors call for further research to identify which patients are most vulnerable to opioid-related depression.