Transitions in Prescription Benzodiazepine Use and Misuse and in Substance Use Disorder Symptoms Through Age 50
Abstract
Objective:
Prescription benzodiazepines are among the most commonly used and misused controlled medications. The authors aimed to examine transitions from medical use of prescription benzodiazepines to prescription benzodiazepine misuse, prescription opioid misuse, and substance use disorder symptoms during adulthood.
Methods:
Eleven national cohorts of U.S. 12th graders (N=26,575) were followed up from ages 18 (1976–1986) to 50 (2008–2018). Prescription benzodiazepine misuse, prescription opioid misuse, and substance use disorder symptoms were examined with prevalence estimates and multivariable logistic regression.
Results:
By age 35, 70.9% of respondents had not used or misused prescription benzodiazepines, 11.3% reported medical use only, 9.8% indicated both medical use and misuse, and 14.1% reported misuse only. In analyses adjusted for demographic and other characteristics, adults reporting only medical use of prescription benzodiazepines by age 35 had higher odds of later prescription benzodiazepine misuse (adjusted OR [AOR]=2.17, 95% CI=1.72–2.75) and prescription opioid misuse (AOR=1.40, 95% CI=1.05–1.86) than respondents ages 35–50 who never used prescription benzodiazepines. More frequent medical use of prescription benzodiazepines by age 35 was associated with increased risk for substance use disorder symptoms at ages 40–50. Any history of prescription benzodiazepine misuse by age 35 was associated with higher odds of later prescription benzodiazepine misuse, prescription opioid misuse, and substance use disorder symptoms, compared with no misuse.
Conclusions:
Prescription benzodiazepine use or misuse may signal later prescription drug misuse or substance use disorders. Medical use of prescription benzodiazepines by age 35 requires monitoring for prescription drug misuse, and any prescription benzodiazepine misuse warrants an assessment for substance use disorder.
HIGHLIGHTS
•
Nearly three in 10 (29%) U.S. adults reported medical use (11%), misuse (14%), or both medical use and misuse (10%) of prescription benzodiazepines by age 35.
•
Medical use of prescription benzodiazepines by age 35 (with or without substance use disorder symptoms) was associated with later misuse of prescription opioids or prescription benzodiazepines.
•
More frequent medical use of prescription benzodiazepines by age 35 was associated with later symptoms of substance use disorder.
•
Approximately 40% of adults who reported prescription benzodiazepine misuse by age 35 had multiple subsequent substance use disorder symptoms; therefore, any history of prescription benzodiazepine misuse warrants a comprehensive assessment for substance use disorder.
Prescription benzodiazepines are among the most commonly used and misused medications in the United States (1–3), and their use has significantly increased over the past two decades (4, 5). Long-term use of these controlled medications is associated with serious health risks (6, 7), and several studies have shown that the highest prevalence of prescription benzodiazepine use occurs during middle to later adulthood (4, 5, 8, 9). Prescription benzodiazepines can be highly effective for managing anxiety disorders, insomnia, seizures, and acute alcohol withdrawal when they are prescribed and used properly. However, unintended consequences of increased and chronic use of prescription benzodiazepines include the concomitant increase in prescription benzodiazepine misuse, overdose, toxicity, and withdrawal (5–7). Studies rarely account for the heterogeneity of exposure to prescription benzodiazepines by assessing lifetime history of prescription benzodiazepine use and misuse for the same individuals over time.
When prescription benzodiazepines are misused, they are often consumed with other substances (10–12). In a general population sample, polysubstance use was identified among approximately half of U.S. adults who misused prescription benzodiazepines (12). Furthermore, most emergency department visits that involved prescription benzodiazepines also involved other substances, including alcohol and prescription opioids (11).
The progression of medical use of prescription benzodiazepines to prescription drug misuse and substance use disorder symptoms during middle adulthood remains largely unexplored, mostly because of a lack of long-term prospective studies. To address this gap, we used national panel data from the Monitoring the Future (MTF) study (3) and examined the following research question: What are the transitions over time from medical use and misuse of prescription benzodiazepines by age 35 to subsequent prescription benzodiazepine misuse, prescription opioid misuse, and substance use disorder symptoms in middle adulthood?
Methods
Study Design
Approximately 2,400 high school seniors were randomly selected each year for biennial follow-ups and were surveyed via mailed questionnaires through age 30 years. They were then followed up (also via mailed questionnaires) every 5 years, at ages 35, 40, 45, and 50 years. The analytic sample (N=26,575) included data from 11 cohorts (baseline 1976–1986; modal age 18) that were followed up until age 50 (in the 2008–2018 period). Response rates at baseline over the study period ranged from 77% to 84%. The overall weighted retention for the longitudinal sample over a 32-year period from baseline to age 50 was 50%. Attrition weights were used in analyses to help correct for potential attrition bias (13–15) by accounting for respondent characteristics associated with nonresponse at the age 50 follow-up. This secondary analysis study received exemption status from the University of Michigan Institutional Review Board.
Measures
Key independent variables.
Lifetime medical use of prescription benzodiazepines by age 35 was assessed by asking respondents whether they had taken benzodiazepines because they had received a prescription for the medication from a physician in their lifetime, the past year, or the past month. Respondents were informed that prescription benzodiazepines are sometimes prescribed by physicians to help calm people down, relax them, or quiet their nerves, among other reasons, and were provided a list of several benzodiazepines, such as Librium, Valium, and Xanax. The response scale ranged from 1, no occasions, to 7, ≥40 occasions, consistent with previous research (2, 3, 16).
Key dependent variables.
Substance use disorder symptoms at ages 40, 45, and 50 were measured for the past 5 years by using questions based on DSM-IV and DSM-5 criteria for alcohol use disorder, cannabis use disorder, and other drug use disorder. Other drugs included prescription drugs (e.g., benzodiazepines, opioids, and stimulants) used without a physician’s orders and illicit drugs (e.g., hallucinogens and cocaine). Although these symptoms do not yield a clinical diagnosis, they are consistent with measures used in other large-scale surveys (17–19) and reflect DSM-IV and DSM-5 criteria (13, 14, 20). Respondents were asked to report symptoms for each substance over the past 5 years. Fifteen items were used to measure the following eight of the 11 DSM-5 criteria that were consistent with alcohol, cannabis, and other drug use disorder symptoms: substance use resulting in a failure to fulfill major role obligations, continued substance use when it is physically hazardous, continued substance use despite persistent or recurrent interpersonal or social problems, tolerance, withdrawal, persistent desire or unsuccessful efforts to cut down substance use, health-related issues due to substance use, and craving. We followed recommended practice by using two or more of the criteria for each substance class, resulting in estimates closely resembling other national estimates for similar age groups (21–23). In addition to reporting symptoms of individual drug classes (i.e., alcohol, cannabis, and other drugs), we created a measure of any substance use disorder symptoms on the basis of two or more symptoms of one substance class. These measures were treated as binary variables (e.g., no or one symptom vs. two or more symptoms).
Prescription benzodiazepine or opioid misuse at ages 40, 45, and 50 was assessed separately by asking respondents on how many occasions, if any, in the past 5 years they used prescription benzodiazepines or prescription opioids on their own—that is, without receiving a physician’s prescription. Prescription benzodiazepine and opioid misuse were assessed with separate questions and examples, which were identical to the questions and examples mentioned above for the age 35 assessment. The response scale ranged from 1, no occasions, to 7, ≥40 occasions (2, 3, 16). These measures were coded to indicate whether respondents misused prescription benzodiazepines or opioids during the past 5 years (i.e., no occasions vs. one or more occasions), consistent with previous research (2, 3, 16).
Covariates.
Covariates at baseline (12th grade) included sex, race-ethnicity, U.S. Census region, urbanicity, parental education, college expectation, GPA, 12th-grade cohort year, past-month cigarette smoking, past–2-week binge drinking, and past-month marijuana use. College graduation by age 50 and time-varying covariates that assessed both sleeping problems (past 30 days: none vs. ≥1 days) and emotional, nervous, or psychiatric problems (lifetime determination by a physician or other medical professional) between ages 35 and 50 were included to help adjust for potential confounding in the full analytic models.
Data Analysis
First, we examined the descriptive results for the four subgroups of lifetime medical use and misuse of prescription benzodiazepines by age 35 in relation to the past–5-year prevalence of substance use disorder symptoms, prescription benzodiazepine misuse, and prescription opioid misuse from ages 40 to 50. Figures were created to show the past–5-year prevalence (along with 95% CIs based on standard errors by using Taylor series linearization) for each of these substance-related outcomes from ages 40 to 50.
Second, logistic regression models were fitted by using the generalized estimating equations (GEE) methodology (24, 25), with an autoregressive correlation structure to account for the panel design. These GEE models assessed the association between the four subgroups of lifetime medical use and misuse of prescription benzodiazepines from the age 35 survey and past–5-year prevalence of prescription drug misuse and substance use disorder symptoms from ages 40 to 50. Additional GEE logistic regression analyses were used to also assess an ordinal measure (0, never used to 6, ≥40 occasions) for each of the four subgroups of lifetime medical use and misuse of prescription benzodiazepines from the age 35 survey and past–5-year prevalence of prescription drug misuse and substance use disorder symptoms at ages 40–50. Analyses accounted for the key covariate variables summarized above. Both unadjusted and adjusted ORs (AORs) and 95% CIs were reported in the GEE models. All descriptive and GEE logistic regression analyses were conducted with Stata, version 17.0.
Results
Table 1 provides the baseline sociodemographic characteristics of the sample. The sample was 50.8% female, 79.3% White, 10.7% Black, 3.6% Hispanic, and 6.4% other race-ethnicity. Most respondents (70.9%; weighted percentages are reported) never used or misused prescription benzodiazepines by age 35. One in nine respondents (11.3%, N=1,683 of 14,932) indicated only medical use of prescription benzodiazepines, 9.8% (N=1,460 of 14,932) reported both medical use and misuse, and 14.1% (N=2,112 of 14,932) indicated only misuse by age 35. Among respondents who misused benzodiazepines, 43.6% (N=1,558 of 3,572) initiated use at or before age 18, and 56.4% (N=2,014 of 3,572) initiated use later in life.
| Total (N=26,575) | No medical use or misuse (N=9,677) | Medical use only (N=1,683) | Misuse only (N=2,112) | Medical use and misuse (N=1,460) | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Characteristic | N | % | N | % | N | % | N | % | N | % |
| Sex | ||||||||||
| Male | 13,073 | 49.2 | 4,354 | 45.0 | 546 | 32.5 | 1,116 | 52.8 | 622 | 42.6 |
| Female | 13,496 | 50.8 | 5,322 | 55.0 | 1,136 | 67.5 | 996 | 47.2 | 838 | 57.4 |
| Race-ethnicity | ||||||||||
| White | 21,082 | 79.3 | 8,133 | 84.0 | 1,491 | 88.6 | 1,928 | 91.3 | 1,314 | 90.0 |
| Black | 2,846 | 10.7 | 817 | 8.4 | 80 | 4.8 | 50 | 2.4 | 51 | 3.5 |
| Hispanic | 946 | 3.6 | 226 | 2.3 | 37 | 2.2 | 38 | 1.8 | 23 | 1.6 |
| Other | 1,701 | 6.4 | 501 | 5.2 | 75 | 4.5 | 96 | 4.5 | 72 | 4.9 |
| GPA | ||||||||||
| B– or higher | 17,917 | 69.8 | 7,349 | 77.8 | 1,263 | 76.2 | 1,462 | 71.2 | 982 | 69.4 |
| C+ or lower | 7,735 | 30.2 | 2,100 | 22.2 | 394 | 23.8 | 592 | 28.8 | 432 | 30.6 |
| College plans | ||||||||||
| No | 11,314 | 45.6 | 3,804 | 41.2 | 706 | 43.7 | 921 | 46.1 | 677 | 49.7 |
| Yes | 13,522 | 54.4 | 5,431 | 58.8 | 910 | 56.3 | 1,075 | 53.9 | 685 | 50.3 |
| Parental education level | ||||||||||
| Less than college degree | 16,458 | 64.7 | 5,891 | 63.2 | 1,019 | 63.1 | 1,274 | 62.8 | 883 | 62.9 |
| College degree or higher | 8,964 | 35.3 | 3,428 | 36.8 | 596 | 36.9 | 756 | 37.2 | 520 | 37.1 |
| U.S. metropolitan statistical area | ||||||||||
| Large (urban) | 7,072 | 26.6 | 2,361 | 24.4 | 411 | 24.4 | 542 | 25.7 | 331 | 22.7 |
| Other (suburban) | 11,206 | 42.2 | 3,996 | 41.3 | 721 | 42.8 | 917 | 43.4 | 682 | 46.7 |
| None (rural) | 8,297 | 31.2 | 3,320 | 34.3 | 551 | 32.7 | 653 | 30.9 | 447 | 30.6 |
| U.S. region | ||||||||||
| Northeast | 6,441 | 24.2 | 2,289 | 23.7 | 380 | 22.6 | 530 | 25.1 | 323 | 22.1 |
| Midwest | 7,882 | 29.7 | 3,325 | 34.4 | 536 | 31.8 | 653 | 30.9 | 414 | 28.4 |
| South | 7,937 | 29.9 | 2,578 | 26.6 | 472 | 28.0 | 622 | 29.5 | 471 | 32.3 |
| West | 4,315 | 16.2 | 1,485 | 15.6 | 295 | 17.5 | 307 | 14.5 | 252 | 17.3 |
| Cohort year | ||||||||||
| 1976–1978 | 6,993 | 26.3 | 2,580 | 26.7 | 508 | 30.2 | 780 | 36.9 | 531 | 36.4 |
| 1979–1981 | 7,353 | 27.7 | 2,889 | 29.9 | 446 | 26.5 | 678 | 32.1 | 409 | 28.0 |
| 1982–1984 | 7,301 | 27.5 | 2,569 | 26.5 | 444 | 26.4 | 438 | 20.7 | 331 | 22.7 |
| 1985–1986 | 4,928 | 18.5 | 1,639 | 16.9 | 285 | 16.9 | 216 | 10.2 | 189 | 12.9 |
| Cigarette smoking | ||||||||||
| No | 15,761 | 60.5 | 6,745 | 71.1 | 1,034 | 62.2 | 879 | 42.0 | 606 | 42.2 |
| Yes | 10,295 | 39.5 | 2,741 | 28.9 | 628 | 37.8 | 1,212 | 58.0 | 829 | 57.8 |
| Binge drinking | ||||||||||
| No | 13,098 | 52.3 | 5,674 | 61.8 | 956 | 59.0 | 665 | 32.6 | 531 | 37.8 |
| Yes | 11,952 | 47.7 | 3,506 | 38.2 | 663 | 41.0 | 1,377 | 67.4 | 872 | 62.2 |
| Marijuana use | ||||||||||
| No | 14,733 | 57.4 | 6,771 | 72.1 | 1,098 | 66.6 | 571 | 27.6 | 446 | 31.3 |
| Yes | 10,947 | 42.6 | 2,624 | 27.9 | 551 | 33.4 | 1,495 | 72.4 | 981 | 68.7 |
a
Weighted percentages are reported; history of benzodiazepine use at age 35 was based on a smaller sample size (N=14,932) because of attrition at age 35. Data were missing for some categories; specifically, at baseline, the percentage of missing data ranged from 0% for race-ethnicity, urbanicity, region, and cohort year to <1% for sex; <4% for GPA, cigarette smoking, and marijuana use; and <7% for college plans and binge drinking. Sample sizes for the four groups by history of benzodiazepine use and misuse were smaller because of the attrition by age 35. Females; non-White respondents; respondents from more recent cohorts; respondents with lower GPAs in 12th grade; respondents with no intentions to go to college during 12th grade; and respondents who indicated smoking, binge drinking, or marijuana use during 12th grade were more likely to drop out of the longitudinal sample.
Table 2 provides descriptive results assessing the history of prescription benzodiazepine use and misuse by age 35 and prevalence of two or more substance use disorder symptoms at ages 40, 45, and 50. Across all three ages, the highest rates of alcohol, cannabis, and other drug use disorder symptoms were among respondents who indicated a history either of both medical use and misuse of prescription benzodiazepines or of only misuse in the age 35 survey. Approximately 40% of adults who reported prescription benzodiazepine misuse by age 35 had multiple substance use disorder symptoms at later ages. Respondents who never used or misused prescription benzodiazepines by age 35 and those who reported only medical use of prescription benzodiazepines by age 35 had the lowest past–5-year prevalence of substance use disorder symptoms.
| Past–5-year alcohol use disorder (≥2 symptoms) (N=12,723) | Past–5-year cannabis use disorder (≥2 symptoms) (N=12,763) | Past–5-year other drug use disorder (≥2 symptoms) (N=12,499) | Past–5-year any substance use disorder (≥2 symptoms) (N=12,822) | Past–5-year prescription benzodiazepine misuse (N=12,666) | Past–5-year prescription opioid misuse (N=12,148) | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Benzodiazepine use by age 35 | % | 95% CI | % | 95% CI | % | 95% CI | % | 95% CI | % | 95% CI | % | 95% CI |
| Age 40 | ||||||||||||
| No medical use or misuse | 18.1 | 17.3–19.1 | 2.4 | 2.1–2.8 | 1.4 | 1.1–1.8 | 19.1 | 18.1–20.0 | 1.8 | 1.4–2.1 | 2.5 | 2.1–2.9 |
| Medical use only | 20.0 | 17.8–22.6 | 2.7 | 1.9–3.9 | 1.7 | 1.1–2.6 | 21.1 | 18.8–23.6 | 5.5 | 4.2–7.1 | 4.5 | 3.3–5.9 |
| Misuse only | 38.3 | 35.8–41.2 | 13.8 | 12.0–15.9 | 10.6 | 8.9–12.4 | 44.5 | 41.7–47.2 | 17.4 | 15.3–19.7 | 11.7 | 10.0–13.7 |
| Medical use and misuse | 39.6 | 36.4–43.0 | 12.5 | 10.5–14.9 | 10.8 | 8.9–12.9 | 45.1 | 41.8–48.4 | 20.7 | 18.1–23.6 | 15.6 | 13.3–18.2 |
| Age 45 | ||||||||||||
| No medical use or misuse | 17.6 | 16.6–18.5 | 1.8 | 1.4–2.1 | 1.2 | 1.0–1.5 | 18.3 | 17.4–19.3 | 2.4 | 2.1–2.8 | 3.2 | 2.8–3.7 |
| Medical use only | 20.4 | 18.1–22.9 | 1.8 | 1.2–2.7 | 2.0 | 1.3–2.9 | 22.0 | 19.6–24.6 | 5.1 | 3.9–6.6 | 6.0 | 4.6–7.7 |
| Misuse only | 37.1 | 34.4–39.8 | 11.4 | 9.7–13.3 | 8.7 | 7.1–10.4 | 42.2 | 39.5–45.0 | 16.3 | 14.2–18.6 | 14.2 | 12.3–16.3 |
| Medical use and misuse | 36.4 | 33.2–39.7 | 9.5 | 7.8–11.5 | 10.0 | 8.2–12.2 | 40.6 | 37.4–44.0 | 19.6 | 17.1–22.5 | 17.3 | 14.9–20.0 |
| Age 50 | ||||||||||||
| No medical use or misuse | 15.6 | 14.6–16.6 | 1.5 | 1.2–1.9 | 1.0 | 0.7–1.3 | 16.5 | 15.5–17.5 | 2.6 | 2.2–2.9 | 3.2 | 2.7–3.6 |
| Medical use only | 18.9 | 16.3–21.8 | 1.9 | 1.1–3.2 | 1.7 | 1.1–2.7 | 20.3 | 17.7–23.3 | 6.6 | 5.2–8.2 | 5.3 | 4.1–6.9 |
| Misuse only | 35.2 | 32.4–38.2 | 8.1 | 6.5–10.1 | 6.9 | 5.5–8.7 | 39.5 | 36.6–42.4 | 14.6 | 12.6–16.8 | 14.7 | 12.6–17.1 |
| Medical use and misuse | 31.2 | 27.8–34.7 | 10.2 | 8.3–12.5 | 9.1 | 7.3–11.1 | 38.0 | 34.5–41.6 | 19.9 | 17.2–22.9 | 17.2 | 14.8–19.9 |
a
Weighted percentages are reported.
Table 2 also shows the association between prescription benzodiazepine use and misuse by age 35 and subsequent 5-year prevalence of prescription opioid misuse or benzodiazepine misuse at ages 40–50. Prescription opioid and benzodiazepine misuse were most prevalent between ages 40 and 50 among respondents who by age 35 had indicated a history of both medical use and misuse of prescription benzodiazepines or a history of only misuse.
Table 3 shows the unadjusted (ORs) and adjusted (AORs) results of multivariable logistic regression analyses based on GEE models to identify associations between histories of medical use and misuse of prescription benzodiazepines by age 35 and substance-related problems from ages 40 to 50; analyses controlled for potentially confounding factors, including sociodemographic characteristics, substance use at age 18, and sleep and emotional problems and symptoms. The results indicated that respondents who by age 35 reported a history of both medical use and misuse or a history of only misuse of prescription benzodiazepines had significantly higher odds of substance use disorder symptoms, prescription opioid misuse, and prescription benzodiazepine misuse over the next 15 years than those who reported no use of prescription benzodiazepines by age 35. For instance, respondents who indicated a history of both medical use and misuse of prescription benzodiazepines by age 35 had roughly four times higher odds (AOR=4.09) of prescription opioid misuse between ages 35 and 50 compared with those who never used or misused prescription benzodiazepines by age 35. Although respondents reporting prescription benzodiazepine misuse had the highest odds of later prescription drug misuse and substance use disorder symptoms, the multivariable logistic regression analyses also indicated that respondents who engaged in only medical use of prescription benzodiazepines by age 35 also had significantly higher odds of prescription benzodiazepine misuse (AOR=2.17) and prescription opioid misuse (AOR=1.40) between ages 35 and 50, compared with respondents who never used prescription benzodiazepines.
| Past–5-year alcohol use disorder (≥2 symptoms) (N=11,643)b | Past–5-year cannabis use disorder (≥2 symptoms) (N=11,701)b | Past–5-year other drug use disorder (≥2 symptoms) (N=11,417)b | Past–5-year any substance use disorder (≥2 symptoms) (N=11,789)b | Past–5-year prescription benzodiazepine misuse (N=11,656)b | Past–5-year prescription opioid misuse (N=11,132)b | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Characteristic | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI |
| Benzodiazepine use by age 35 (reference: no medical use or misuse)c | ||||||||||||
| Medical use only | 1.16* | 1.01–1.34 | 1.06 | .71–1.59 | 1.58* | 1.07–2.33 | 1.18* | 1.03–1.35 | 2.67*** | 2.15–3.31 | 1.74*** | 1.39–2.25 |
| Misuse only | 2.84*** | 2.53–3.19 | 6.53*** | 5.24–8.14 | 7.65*** | 5.92–9.88 | 3.32*** | 2.97–3.72 | 8.56*** | 7.18–10.20 | 4.98*** | 4.18–5.94 |
| Medical use and misuse | 2.73*** | 2.39–3.12 | 6.30*** | 4.99–7.95 | 8.74*** | 6.73–11.40 | 3.22*** | 2.83–3.66 | 10.90*** | 9.18–13.10 | 6.32*** | 5.29–7.56 |
| Age (in years; reference: 40)d | ||||||||||||
| 45 | .95* | .90–.93 | .74*** | .66–.82 | .90 | .79–1.00 | .93** | .89–.98 | 1.07 | .93–1.17 | 1.31*** | 1.15–1.49 |
| 50 | .76*** | .71–.80 | .65*** | .58–.74 | .81* | .68–.96 | .77*** | .73–.82 | 1.08 | .96–1.22 | 1.29*** | 1.13–1.48 |
| Past–5-year alcohol use disorder (≥2 symptoms) (N=10,041)b | Past–5-year cannabis use disorder (≥2 symptoms) (N=10,080)b | Past–5-year other drug use disorder (≥2 symptoms) (N=9,901)b | Past–5-year any substance use disorder (≥2 symptoms) (N=10,142)b | Past–5-year prescription benzodiazepine misuse (N=10,053)b | Past–5-year prescription opioid misuse (N=9,631)b | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| AOR | 95% CI | AOR | 95% CI | AOR | 95% CI | AOR | 95% CI | AOR | 95% CI | AOR | 95% CI | |
| Benzodiazepine use by age 35 (reference: no medical use or misuse)c | ||||||||||||
| Medical use only | 1.03 | .88–1.21 | 1.00 | .64–1.56 | 1.33 | .88–2.04 | 1.04 | .89–1.21 | 2.17*** | 1.72–2.75 | 1.40* | 1.05–1.86 |
| Misuse only | 1.77*** | 1.54–2.02 | 3.55*** | 2.72–4.64 | 3.89*** | 2.87–5.28 | 2.06*** | 1.81–2.35 | 6.31*** | 5.14–7.74 | 3.42*** | 2.78–4.22 |
| Medical use and misuse | 1.69*** | 1.45–1.99 | 3.46*** | 2.58–4.64 | 4.19*** | 3.04–5.78 | 1.97*** | 1.69–2.30 | 7.10*** | 5.72–8.81 | 4.09*** | 3.29–5.08 |
| Age (in years; reference: 40)d | ||||||||||||
| 45 | .95 | .89–1.00 | .74*** | .65–.83 | .88 | .75–1.03 | .93* | .88–.98 | 1.02 | .89–1.15 | 1.22** | 1.06–1.40 |
| 50 | .84*** | .78–.90 | .69*** | .59–.79 | .78** | .64–.94 | .84*** | .78–.89 | 1.05 | .91–1.20 | 1.22** | 1.05–1.42 |
a
Results from unadjusted (ORs) and adjusted (adjusted ORs [AORs]) analyses are shown; unweighted sample sizes are provided. Attrition weights by age 50 were used for all estimates. Adjusted models controlled for 12th-grade measures of sex, race-ethnicity, parents’ level of education, urbanicity, U.S. region, cohort year, GPA during high school, college expectations, past–30-day cigarette use, past–2-week binge drinking, past–30-day marijuana use, college graduation (yes or no), sleeping problems (time varying; ages 40–50), and medical diagnosis of an emotional or nervous problem (time varying; ages 40–50). Sample sizes varied because of missing data across items used in the analyses.
b
All dependent measures were time varying.
c
Time invariant.
d
Time varying.
*p<0.05, **p<0.01, ***p<0.001.
Table 3 also shows overall associations between age of the respondents and each of the substance-related outcome variables. As respondents aged, their odds of reporting any substance use disorder symptoms decreased, regardless of their history of medical use or misuse of prescription benzodiazepines. However, with increasing age, the odds of reporting prescription opioid misuse modestly increased after age 40, whereas the odds of reporting prescription benzodiazepine misuse remained relatively stable.
Finally, we used GEE logistic regression in additional analyses to examine the frequency of lifetime medical use or misuse of prescription benzodiazepines based on an ordinal measure (0, never used to 6, ≥40 occasions) for each subgroup from the age 35 survey and past–5-year prevalence of prescription drug misuse and substance use disorder symptoms at ages 40–50. As shown in Table 4, results examining medical use and misuse as ordinal variables were similar to the results treating these variables as dichotomous (as shown in Table 3), with respondents who reported prescription benzodiazepine misuse having higher adjusted odds of later prescription drug misuse or substance use disorder symptoms than respondents with only medical use. As shown in Table 4, multivariable logistic regression analyses indicated that compared with respondents who never used prescription benzodiazepines, respondents with more frequent medical use of prescription benzodiazepines by age 35 had significantly higher odds of having multiple symptoms of alcohol, other drug, and any substance use disorder; prescription benzodiazepine misuse; and prescription opioid misuse at ages 40–50.
| Past–5-year alcohol use disorder (≥2 symptoms) (N=10,055)b | Past–5-year cannabis use disorder (≥2 symptoms) (N=10,095)b | Past–5-year other drug use disorder (≥2 symptoms) (N=9,875)b | Past–5-year any substance use disorder (≥2 symptoms) (N=10,160)b | Past–5-year prescription benzodiazepine misuse (N=10,062)b | Past–5-year prescription opioid misuse (N=9,650)b | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Characteristic | AOR | 95% CI | AOR | 95% CI | AOR | 95% CI | AOR | 95% CI | AOR | 95% CI | AOR | 95% CI |
| Benzodiazepine use by age 35 (reference: no medical use or misuse)c | ||||||||||||
| Medical use only | 1.06* | 1.00–1.12 | 1.01 | .890–1.14 | 1.21** | 1.06–1.38 | 1.07** | 1.01–1.13 | 1.28*** | 1.19–1.37 | 1.11** | 1.02–1.20 |
| Misuse only | 1.18*** | 1.13–1.22 | 1.34*** | 1.26–1.43 | 1.43*** | 1.34–1.52 | 1.25** | 1.19–1.30 | 1.63*** | 1.55–1.70 | 1.42*** | 1.35–1.49 |
| Medical use and misuse | 1.16*** | 1.12–1.21 | 1.32*** | 1.25–1.39 | 1.47*** | 1.38–1.55 | 1.22*** | 1.18–1.27 | 1.58*** | 1.51–1.64 | 1.43*** | 1.37–1.49 |
| Age (in years; reference: 40)d | ||||||||||||
| 45 | .966 | .914–1.02 | .733*** | .649–.829 | .914 | .782–1.06 | .951 | .900–1.00 | 1.07 | .942–1.21 | 1.30*** | 1.13–1.48 |
| 50 | .767*** | .715–.822 | .686*** | .595–.790 | .789*** | .654–.953 | .775*** | .725–.829 | 1.11 | .963–1.27 | 1.28*** | 1.11–1.48 |
a
Unweighted sample sizes are provided. Attrition weights at age 50 were used for all estimates. Adjusted models controlled for 12th-grade measures of sex, race-ethnicity, parents’ level of education, urbanicity, U.S. region, cohort year, GPA during high school, college expectations, past–30-day cigarette use, past–2-week binge drinking, past–30-day marijuana use, college graduation (yes or no), sleeping problems (time varying; ages 40–50), and medical diagnosis of an emotional or nervous problem (time varying; ages 40–50). Sample sizes varied because of missing data across items used in the analyses. AOR, adjusted OR.
b
All dependent measures were time varying.
c
Time invariant (0, none; 1, 1–2 occasions; 2, 3–5 occasions; 3, 6–9 occasions; 4, 10–19 occasions; 5, 20–39 occasions; and 6, ≥40 occasions).
d
Time varying.
*p<0.05, **p<0.01, ***p<0.001.
Supplemental analyses were performed that included substance use disorder symptoms at age 35 as a covariate, yielding results very similar to those of the original analysis that did not adjust for substance use disorder symptoms at age 35 (see the online supplement to this article). Because adults who use or misuse other drugs might also have increased medical use of benzodiazepines, thereby potentially confounding the results, a sensitivity analysis was conducted by excluding individuals who had two or more substance use disorder symptoms at age 35. The association between medical use of benzodiazepines by age 35 and later benzodiazepine misuse at ages 40–50 remained statistically significant even after exclusion of individuals who had two or more substance use disorder symptoms at age 35 (AOR=1.38, 95% CI=1.30–1.47), further strengthening the link between medical use of benzodiazepines by age 35 and later misuse of benzodiazepines.
Discussion
To our knowledge, this is the first prospective national study to examine transitions in prescription benzodiazepine use, misuse, and substance use disorder symptoms from young adulthood through age 50. Notably, by accounting for sociodemographic characteristics, previous substance use, and other covariates, we found that adults who reported only medical use of prescription benzodiazepines by age 35 (compared with those who never used or misused prescription benzodiazepines by age 35) were over two times more likely to transition to prescription benzodiazepine misuse at ages 40–50 and nearly 1.5 times more likely to transition to prescription opioid misuse at ages 40–50. Additionally, more frequent medical use of benzodiazepines was associated with significantly increased risk for later alcohol, other drug, or overall substance use disorder symptoms. Our findings suggest that medical or nonmedical exposure to prescription benzodiazepines by age 35 is associated with increased risk for engaging in substance misuse in middle adulthood. A history of being prescribed benzodiazepines by age 35 should be considered a risk factor for prescription drug misuse and substance use disorder symptoms during middle adulthood, especially when medical use of benzodiazepines is frequent or co-occurs with benzodiazepine misuse.
Our results on the medical use of prescription benzodiazepines are clinically noteworthy and suggest that adults reporting more frequent medical use of benzodiazepines should be monitored for prescription drug misuse and substance use disorder symptoms during middle adulthood. Our findings are timely given the shift toward e-prescribing of controlled substances (26) and the U.S. Food and Drug Administration’s pending updates to the boxed warning on all benzodiazepines regarding associated risks (27, 28).
It is increasingly recognized that benzodiazepines are widely prescribed, often over long periods, and that these medications are frequently misused and ingested with other substances, which may result in dangerous drug-drug interactions (28). We found that the prevalence of medical use of prescription benzodiazepines remained relatively unchanged through middle adulthood, and our results were comparable to those of other national studies (29). Nearly one in six adults ages ≥50 used prescription benzodiazepines in the past year despite guidelines discouraging such use in this age group (30), and clinicians should continue to avoid coprescribing opioids and benzodiazepines whenever possible (31, 32). Initiatives that inform adults about the risks of long-term benzodiazepine use and suggest alternative treatments have been effective, leading to patient-initiated tapering or deprescribing (33). Several nonbenzodiazepine interventions for insomnia and anxiety, such as cognitive-behavioral therapy (34), have been proven effective and can be considered.
When clinicians regularly examine the need for ongoing benzodiazepine therapy, prescribe the appropriate dose, instruct medication adherence, and monitor use, these practices not only enhance appropriate medical use of the medication, but also lessen the potential for misuse and reduce the risk that reservoirs of leftover medication are used for diversion to friends and family, the leading diversion source across all ages (2, 35, 36). Evidence suggests that medical use of benzodiazepines is associated with greater odds of later medical use of opioids (37). The findings of this study show that medical use of benzodiazepines is associated with greater odds of later misuse of prescription benzodiazepines and opioids. The increased risk for later prescription opioid misuse found in this study among adults who reported more frequent medical use of prescription benzodiazepines could be influenced by an increased access of these adults to prescription opioids, which warrants more attention. A thoughtful discussion about a patient’s substance use history and about the potential that treatment of a condition or disorder may increase a patient’s risk for substance use disorder can help prescribers select the most appropriate treatment and medication management strategies. Finally, future research that provides a risk-benefit analysis at the individual or population level is warranted to guide decisions about who should be treated with benzodiazepines.
Findings from the analyses controlled for demographic and other characteristics indicated that individuals with a history of prescription benzodiazepine misuse by age 35 had at least two times greater odds of reporting two or more symptoms of alcohol, cannabis, or other drug use disorder as well as of prescription opioid misuse and prescription benzodiazepine misuse over the 15-year period between ages 35 and 50, compared with those who reported only medical use of prescription benzodiazepines by age 35. Previous research has found that, among U.S. adults, prescription opioid misuse and benzodiazepine misuse greatly contribute to emergency department visits and overdoses (10, 11) and have been linked to past-year suicidal ideation (38). Our findings and previous research support screening for prior prescription benzodiazepine misuse before a physician may prescribe benzodiazepines and highlight the need for more comprehensive surveillance for potential prescription drug misuse, medication diversion, and substance use disorder (31). To detect prescription drug misuse, prescribers can check prescription drug monitoring programs, conduct point-of-care urine drug testing, administer brief screening instruments, and identify other signs of misuse (e.g., early refill requests). Nonetheless, the course of benzodiazepine misuse is discontinuous among many adults, reinforcing the need to discuss past or current misuse with patients rather than simply discontinuing the prescription. Given that benzodiazepine misuse often covaries with symptoms of psychiatric disorders, removing a potential treatment for these symptoms may be counterproductive.
The MTF panel data had several limitations inherent in population-based national longitudinal survey research, such as having methods different from those used in clinical settings for assessing substance use disorder symptoms, self-report bias, measurement bias, and attrition. First, the MTF study could not establish formal clinical substance use disorder diagnoses. Although the prevalence of benzodiazepine use, benzodiazepine misuse, and substance use disorder symptoms in the MTF study closely resemble other national estimates (29), we note the need to replicate these findings with more detailed clinical measures, such as indications for benzodiazepine use; to validate the frequency of benzodiazepine use measures; and to assess different types of prescription benzodiazepine misuse (e.g., using more than prescribed) and other screening tools (e.g., the Drug Abuse Screening Test and Alcohol Use Disorders Identification Test). Second, the MTF study does not assess some variables (e.g., family history of substance use disorders) related to problems with substance use over time. Third, self-reported substance use data have been found to be reliable and valid, but studies with youths indicate misclassification (typically underreporting) if substance use does occur. Fourth, the MTF study began with representative U.S. samples of high school seniors, excluding those who dropped out of high school, institutionalized individuals, and home-schooled students; the generalizability of our findings is therefore somewhat limited. Fifth, as is typical in longitudinal studies regarding substance use, individuals who use drugs are less likely to remain in longitudinal samples. Although this factor was addressed to some extent by our use of attrition weights, our findings did not fully account for unobservable confounders and likely underrepresent heavy drug use, suggesting that our findings reflect conservative estimates of prescription benzodiazepine misuse and substance-related problems.
Conclusions
Our investigation represents one of the longest prospective studies and used a large multicohort U.S. panel of research participants. Our study fills an important research gap by revealing that medical use and misuse of prescription benzodiazepines by age 35 are associated with prescription drug misuse and substance use disorder symptoms later in life. These findings highlight several areas of concern regarding prescription benzodiazepines and reinforce the importance of physicians’ screening and monitoring for misuse when prescribing these medications to reduce prescription benzodiazepine misuse, prescription opioid misuse, and the prevalence of substance use disorders.
Acknowledgments
The authors thank Kathryn Lundquist, A.B.A., University of Michigan, Center for the Study of Drugs, Alcohol, Smoking and Health, for assisting in proofreading and formatting the manuscript and the respondents, school personnel, and research staff for their participation in the study.
Footnote
Dr. Schulenberg died on February 9, 2023.
Supplementary Material
File (appi.ps.20220247.ds001.pdf)
- View/Download
- 84.72 KB
References
1.
Blanco C, Han B, Jones CM, et al: Prevalence and correlates of benzodiazepine use, misuse, and use disorders among adults in the United States. J Clin Psychiatry 2018; 79:18m12174
2.
Miech RA, Johnston LD, O’Malley PM, et al: Monitoring the Future National Survey Results on Drug Use, 1975–2019: Volume I, Secondary School Students. Ann Arbor, Institute for Social Research, University of Michigan, 2020
3.
Schulenberg JE, Johnston LD, O’Malley PM, et al: Monitoring the Future National Survey Results on Drug Use, 1975–2019: Volume II, College Students and Adults Ages 19–60. Ann Arbor, Institute for Social Research, University of Michigan, 2020
4.
Agarwal SD, Landon BE: Patterns in outpatient benzodiazepine prescribing in the United States. JAMA Netw Open 2019; 2:e187399
5.
Bachhuber MA, Hennessy S, Cunningham CO, et al: Increasing benzodiazepine prescriptions and overdose mortality in the United States, 1996–2013. Am J Public Health 2016; 106:686–688
6.
Lader M: Benzodiazepines revisited—will we ever learn? Addiction 2011; 106:2086–2109
7.
Davies SJ, Rudoler D, de Oliveira C, et al: Comparative safety of chronic versus intermittent benzodiazepine prescribing in older adults: a population-based cohort study. J Psychopharmacol 2022; 36:460–469
8.
Olfson M, King M, Schoenbaum M: Benzodiazepine use in the United States. JAMA Psychiatry 2015; 72:136–142
9.
Maust DT, Lin LA, Blow FC: Benzodiazepine use and misuse among adults in the United States. Psychiatr Serv 2019; 70:97–106
10.
Geller AI, Dowell D, Lovegrove MC, et al: US emergency department visits resulting from nonmedical use of pharmaceuticals. Am J Prev Med 2019; 56:639–647
11.
Moro RN, Geller AI, Weidle NJ, et al: Emergency department visits attributed to adverse events involving benzodiazepines, 2016–2017. Am J Prev Med 2020; 58:526–535
12.
Votaw VR, McHugh RK, Vowles KE, et al: Patterns of polysubstance use among adults with tranquilizer misuse. Subst Use Misuse 2020; 55:861–870
13.
Merline A, Jager J, Schulenberg JE: Adolescent risk factors for adult alcohol use and abuse: stability and change of predictive value across early and middle adulthood. Addiction 2008; 103:84–99
14.
Schulenberg JE, Patrick ME, Kloska DD, et al: Substance use disorder in early midlife: a national prospective study on health and well-being correlates and long-term predictors. Subst Abuse 2015; 9:41–57
15.
Bachman JG, Johnston LD, O’Malley PM, et al: The Monitoring the Future Project After Four Decades: Design and Procedures. Monitoring the Future Occasional Paper no 82. Ann Arbor, Institute for Social Research, University of Michigan, 2015
16.
Hamilton CM, Strader LC, Pratt JG, et al: The PhenX Toolkit: get the most from your measures. Am J Epidemiol 2011; 174:253–260
17.
Harford TC, Muthén BO: The dimensionality of alcohol abuse and dependence: a multivariate analysis of DSM-IV symptom items in the national longitudinal survey of youth. J Stud Alcohol 2001; 62:150–157
18.
Muthén BO: Psychometric evaluation of diagnostic criteria: application to a two-dimensional model of alcohol abuse and dependence. Drug Alcohol Depend 1996; 41:101–112
19.
Nelson CB, Heath AC, Kessler RC: Temporal progression of alcohol dependence symptoms in the US household population: results from the National Comorbidity Survey. J Consult Clin Psychol 1998; 66:474–483
20.
Patrick ME, Schulenberg JE, O’Malley PM, et al: Adolescents’ reported reasons for alcohol and marijuana use as predictors of substance use and problems in adulthood. J Stud Alcohol Drugs 2011; 72:106–116
21.
Compton WM, Thomas YF, Stinson FS, et al: Prevalence, correlates, disability, and comorbidity of DSM-IV drug abuse and dependence in the United States: results from the National Epidemiologic Survey on Alcohol and Related Conditions. Arch Gen Psychiatry 2007; 64:566–576
22.
Grant BF, Goldstein RB, Saha TD, et al: Epidemiology of DSM-5 alcohol use disorder: results from the National Epidemiologic Survey on Alcohol and Related Conditions III. JAMA Psychiatry 2015; 72:757–766
23.
Hasin DS, Kerridge BT, Saha TD, et al: Prevalence and correlates of DSM-5 cannabis use disorder, 2012–2013: findings from the National Epidemiologic Survey on Alcohol and Related Conditions–III. Am J Psychiatry 2016; 173:588–599
24.
Hanley JA, Negassa A, Edwardes MD, et al: Statistical analysis of correlated data using generalized estimating equations: an orientation. Am J Epidemiol 2003; 157:364–375
25.
Zeger SL, Liang KY, Albert PS: Models for longitudinal data: a generalized estimating equation approach. Biometrics 1988; 44:1049–1060
26.
Medicare program: electronic prescribing of controlled substances; request for information (RFI). Fed Regist 2020; 85:47151–47157. https://www.govinfo.gov/content/pkg/FR-2020-08-04/pdf/2020-16897.pdf
27.
Hirschtritt ME, Olfson M, Kroenke K: Balancing the risks and benefits of benzodiazepines. JAMA 2021; 325:347–348
28.
FDA Requiring Boxed Warning Updated to Improve Safe Use of Benzodiazepine Drug Class. Silver Spring, MD, US Food and Drug Administration, 2020. https://www.fda.gov/drugs/drug-safety-and-availability/fda-requiring-boxed-warning-updated-improve-safe-use-benzodiazepine-drug-class. Accessed Oct 11, 2021
29.
McCabe SE, Wilens TE, Boyd CJ, et al: Age-specific risk of substance use disorders associated with controlled medication use and misuse subtypes in the United States. Addict Behav 2019; 90:285–293
30.
2019 American Geriatrics Society Beers Criteria Update Expert Panel: American Geriatrics Society 2019 updated AGS Beers Criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc 2019; 67:674–694
31.
US Preventive Services Task Force, Krist AH, Davidson KW, et al: Screening for unhealthy drug use: US Preventive Services Task Force recommendation statement. JAMA 2020; 323:2301–2309
32.
Dowell D, Haegerich TM, Chou R: CDC guideline for prescribing opioids for chronic pain—United States, 2016. MMWR Recomm Rep 2016; 65:1–49
33.
Tannenbaum C, Martin P, Tamblyn R, et al: Reduction of inappropriate benzodiazepine prescriptions among older adults through direct patient education: the EMPOWER cluster randomized trial. JAMA Intern Med 2014; 174:890–898
34.
Soong C, Burry L, Greco M, et al: Advise non-pharmacological therapy as first line treatment for chronic insomnia. BMJ 2021; 372:n680
35.
McCabe SE, Teter CJ, Boyd CJ, et al: Sources of prescription medication misuse among young adults in the United States: the role of educational status. J Clin Psychiatry 2018; 79:17m11958
36.
Schepis TS, McCabe SE: Prescription tranquilizer/sedative sources for misuse in older adults. Subst Use Misuse 2019; 54:1908–1912
37.
Skurtveit S, Furu K, Bramness J, et al: Benzodiazepines predict use of opioids—a follow-up study of 17,074 men and women. Pain Med 2010; 11:805–814
38.
Schepis TS, Simoni-Wastila L, McCabe SE: Prescription opioid and benzodiazepine misuse is associated with suicidal ideation in older adults. Int J Geriatr Psychiatry 2019; 34:122–129
Information & Authors
Information
Published In
History
Received: 3 May 2022
Revision received: 28 November 2022
Revision received: 3 March 2023
Accepted: 22 March 2023
Published online: 5 May 2023
Published in print: November 01, 2023
Keywords
Authors
Competing Interests
Dr. Wilens has been a consultant to 3D Therapeutics, Bay Cove Human Services, Gavin Foundation, and Major and Minor League Baseball; has a licensing agreement with Ironshore for the Before School Functioning Questionnaire; and receives book royalties from Cambridge University Press, Elsevier, and Guilford Press. The other authors report no financial relationships with commercial interests.
Funding Information
The work reported in this article was supported by the National Institute on Drug Abuse (NIDA) (research grants R01 DA-001411, R01 DA-016575, R01 DA-031160, R01 DA-036541, and R01 DA-042146).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. The content is solely the responsibility of the authors and does not necessarily represent the views of NIDA, the NIH, or the U.S. government.
Metrics & Citations
Metrics
Citations
Export Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
For more information or tips please see 'Downloading to a citation manager' in the Help menu.
View Options
View options
PDF/EPUB
View PDF/EPUBLogin options
Already a subscriber? Access your subscription through your login credentials or your institution for full access to this article.
Personal login Institutional Login Open Athens loginNot a subscriber?
PsychiatryOnline subscription options offer access to the DSM-5-TR® library, books, journals, CME, and patient resources. This all-in-one virtual library provides psychiatrists and mental health professionals with key resources for diagnosis, treatment, research, and professional development.
Need more help? PsychiatryOnline Customer Service may be reached by emailing [email protected] or by calling 800-368-5777 (in the U.S.) or 703-907-7322 (outside the U.S.).