Abstract
Introduction:
Tramadol has been associated with chronic opioid use and emergency room (ER) visits. However, little is known about trends in prescription tramadol use in the US.
Methods:
Optum’s de-identified Clinformatics® Data Mart Database was used to assess trends in monthly incident and prevalent tramadol use from 2005 to 2021, stratified by sex and age (18–64 vs. ≥65 years). State-specific trends following scheduling of tramadol as Class IV controlled substance in August 2014 were analyzed with random effects regression models. Demographics, comorbidities, initiation setting, dose, and co-dispensing with other opioids and central nervous system (CNS) agents were assessed in people initiating tramadol, stratified by age and initiation year (2005–2010, 2011–2015, 2016–2021). Analyses were performed in 2023 and 2024.
Results:
During 2005–2021, the mean percentage using tramadol in a given month was 0.88% of younger females, 0.55% of younger males, 1.97% of older females, and 1.14% of older males; 5,729,652 initiations were identified. Since 2014, estimated relative yearly decrease was 4% (95% CI 3%; 5%) in use and 5% (95% CI 4%; 5%) in initiation, with variation across states. Primary care percentage of tramadol initiations declined from 49.2% in 2005–2010 to 37.2% in 2016–2021. During 2016–2021, co-dispensing with other CNS agents occurred in 37.8% of younger and 32.1% of older adults initiating tramadol.
Conclusions:
Tramadol use was higher in females and older adults, exhibited heterogeneous trends across states, and shifted from primary care to ER and specialist settings over time. Co-dispensing with other CNS agents was common and warrants further monitoring.
Introduction
In the wake of the opioid epidemic, efforts to reduce prescription opioid use, particularly oxycodone and hydrocodone, have led to increased use of other medications for pain management.1–4 Tramadol is a less potent opioid analgesic and often considered safer than more potent opioids, such as oxycodone, morphine, fentanyl, and hydrocodone.5 The perception that tramadol is safer and carries a lower risk of addiction, coupled with tighter regulation of more potent opioids, has led to a rise in tramadol use since its original approval in 1994.1,6,7 However, due to the mounting evidence of tramadol’s potential for harm, regulatory and public health efforts have been implemented to stymie increasing tramadol use around the world.8–10 In August 2014, the US Drug Enforcement Administration (DEA) placed tramadol in Schedule IV of the Comprehensive Drug Abuse Prevention and Control Act of 1970 (“Controlled Substance Act”).11 Since then, tramadol use in the US has declined; however, 30 million tramadol prescriptions were filled in 2021.12
Despite its widespread use, little is known about how tramadol is used in the US. Published tramadol utilization research has been restricted to specific care settings and patient subgroups.1,10,13 Moreover, to the authors’ knowledge, no prior studies have evaluated how recent changes in opioid use and regulations may have affected the type of patients who receive tramadol and whether the changes have been the same across population subgroups. The objective of this study was to evaluate prescription tramadol utilization trends from January 2005 to December 2021 using a nationwide commercial insurance database, conducting age and sex-specific analyses. The secondary objective was to evaluate state-specific trends following the August 2014 change in federal regulation.
Methods
Study Population
Data were sourced from Optum’s de-identified Clinformatics® Data Mart Database, a de-identified US database of commercially insured enrollees, including individuals aged ≥65 years covered through Medicare Advantage. This longitudinal health care claims database includes demographic information of enrollees, linked to their inpatient, outpatient, and pharmacy claims. Pharmacy dispensings are coded using National Drug Codes (NDCs) and include the date of dispensing, quantity dispensed, and days’ supply as recorded by pharmacy. All analyses were restricted to individuals ≥18 years at the time of tramadol dispensing. Since the proportion of individuals ≥65 years of age increased in Clinformatics® over time, all analyses were stratified by age: 18–64 years of age (younger adults), and ≥65 years of age (older adults). Participants could enter the cohort multiple times if all inclusion criteria were satisfied. Ethical approval was provided by Mass General Brigham Institutional Review Board (IRB#: 2020P002865); the study was deemed exempt.
Measures
Trends in use and initiation over time were measured as monthly proportions of eligible enrollees using and initiating tramadol, respectively, during the study period (January 2005 to December 2021). For the prevalent use analyses, eligibility was defined as at least 28 days of insurance enrollment within the month (i.e., alive and insured for most of the month). Monthly prevalent use is presented as percent of eligible enrollees with at least one tramadol pharmacy dispensing that month. For the incident use (initiation) analyses, eligibility was defined as at least 180 days of insurance enrollment prior to the calendar month in question, without a tramadol dispensing during that time. Monthly incident use is presented as the number of individuals with 1 initiation per 1,000 eligible enrollees.
Among people initiating tramadol, the following patient characteristics were measured: age, sex, race, and opioid naivety, defined as no opioids dispensed for at least 6 months prior to tramadol initiation. The prescriber’s specialty was also examined and identified based on information provided in Clinformatics®. Total and daily tramadol morphine milligram equivalents (MME) were measured on the initiation date, using a conversion factor of 0.2 (i.e., 20% of oral morphine’s strength).5 If individuals had more than one tramadol dispensing on the initiation date, total MME included all dispensings. Daily MME was calculated by dividing total MME by total days supplied. For individuals who had other opioids dispensed on the tramadol initiation date, we also estimated total MMEs of the other opioids. Persistent tramadol use was defined as having at least 90 days out of 180 days following tramadol inhiation covered by tramadol pills, based on days’ supply. Additional measures included co-dispensing with other CNS depressants and other opioids (Appendix Table 1). Other CNS depressants included benzodiazepines, gabapentinoids, muscle relaxants, and non-benzodiazepine sleep medications. Other opioids were assessed as a group; hydrocodone and oxycodone, two commonly used opioids, were also evaluated separately. Co-dispensing was defined as a dispensing in the 30 days before or after tramadol initiation. Potential indications for tramadol use and comorbidities (listed in Appendix Table 2) were measured using ICD-9 or ICD-10 diagnosis or procedure codes in the 6 months prior to and including the date of tramadol (specific codes are available as CSV through GitLab: https://gitlab-scm.partners.org/drugepi/tramadol_trends/-/tree/main).
Statistical Analysis
Monthly prevalence (% of eligible enrollees) and incidence (N per 1,000 eligible enrollees) were plotted across calendar time. Averages across the study period or for a particular year were calculated as mean values across all monthly values during that period and are presented with standard deviations (SD).
Overall and state-specific trends after tramadol was designated a Schedule IV drug in August 2014 were evaluated using mixed effects negative binomial regression models.14,15 Trends were estimated annually from September 2014 to December 2021 using monthly counts. The models included the logarithm of the eligible enrollees in each month as an offset (denominator) and adjusted for the fixed effects of age (18–64 vs ≥65 years) and sex. A random intercept for the state variable accounted for correlation between monthly counts within states. A random slope for the year variable allowed its coefficient (log rate ratio) to vary across states, yielding weighted state-specific trend estimates.14,15
To assess changes in patient characteristics and initial tramadol dose over time, tramadol initiations were stratified into three time periods: 2005–2010, 2011–2015, and 2016–2021. Counts are presented as percentages of total N for the specific time period, while continuous variables, such as MMEs and day’s supply, are presented as medians with interquartile range (IQR; 25th – 75th percentile). In line with recent recommendations on statistical significance testing,16,17 no significance testing was done to evaluate changes in characteristics over time. These analyses were also conducted separately in females and males to assess sex differences.
Data management and analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC) and R version 4.3.0.
Results
Averaged over the study period (2005–2021), the mean percentage of enrollees with at least 1 tramadol dispensing in a given month was 0.88% (SD 0.22%) for younger females, 0.55% (SD 0.14%) for younger males, 1.97% (SD 0.54%) for older females and 1.14% (SD 0.37%) for older males. Prevalent tramadol use increased from the mean 0.42% (SD 0.02%) of enrollees in 2005 to 1.17% (SD 0.05%) in 2013 and then decreased to 1.10% (SD 0.04%) in 2021 (Figure 1). In 2021, the mean percentage of individuals using tramadol in any given month was 0.82% among younger females, 0.51% among younger males, 1.98% among older females, and 1.29% among older males. Tramadol initiation increased from the mean 1.58 (SD 0.10) individuals initiating tramadol per 1,000 eligible enrollees in 2005 to 3.50 (SD 0.16) in 2013 and then decreased to 2.96 (SD 0.14) in 2021. In 2021, the mean rate of individuals initiating tramadol in any given month was 2.39 per 1,000 younger females, 1.60 per 1,000 younger males, 4.69 per 1,000 older females, and 3.50 per 1,000 older males (Figure 1). A spike in tramadol initiation in December 2010 was observed following the market withdrawal of propoxyphene in November 2010.18
Figure 1. Tramadol use monthly by age group and sex: United States, 2005–2021.
Note: Monthly counts of people using or initiating tramadol divided by monthly counts of eligible Optum Clinformatics database enrollees. (A) Prevalent tramadol use, defined as at least 1 tramadol dispensing that month. (B) Tramadol initiation, defined as tramadol dispensing following at least180 days without tramadol exposure.
Mixed effects negative binomial regression analyses of tramadol utilization trends after the drug received a Schedule IV designation estimated downward trends in both incident use (RR = 0.95; 95% CI: 0.94, 0.96) and prevalent use (RR=0.96; 95% 0.95, 0.97) across states after age and sex adjustment. However, state-specific trends were highly heterogeneous, with estimated state-specific yearly changes ranging from −15% to +5% for prevalent use and from −13% to +6% for incident use (Figure 2). States with increasing trends in the prevalent use of tramadol included Minnesota, South Dakota, Maine, New Hampshire, Arizona, and West Virginia (Figure 2), while Minnesota and South Dakota also exhibited increasing rates of initiation after August 2014 (Figure 2). Regression results and model diagnostics are provided in Appendix Table 3 and Appendix Figure 1.
Figure 2. State trends in tramadol use following Schedule IV classification: United States, September 2014-December 2021.
Note: (A) Prevalent use. (B) Incident use. Bars represent differences in log rates year-to-year. Bars above 0 indicate increasing annual trends, while bars below 0 indicate decreasing trends.
Exponentiation of these differences in log rates would yield rate ratios (RR) for change year-to-year (i.e., multiply previous year’s rate by RR to get following year’s estimated rate). The error bars represent 95% confidence limits, calculated with standard errors that account for dependence of counts within states.
Overall, there were 4,372,421 individuals who initiated tramadol 5,729,652 times between 2005 and 2021. Patient characteristics are presented in Table 1 (sex-specific characteristics are presented in Appendix Tables 4 and 5). The percentage of females among people initiating tramadol remained relatively constant, averaging 60% over the study period in younger adults, with a slight decline in older adults: from 68.7% in 2005–2010 to 64.5% in 2016–2021 (Table 1). The percentage of individuals classified as Asian, Black, or Hispanic [sic; Latinx] among people initiating tramadol increased over time in both age groups, although the proportion of individuals classified as White remained the same (65.4% of younger adults and 65.4% of older adults initiating tramadol during 2016–2021). Opioid-naïve tramadol initiations increased from 60.6% of tramadol initiations in 2005–2010 to 73.6% in 2016–2021 in younger adults and from 63.2% in 2005–2010 to 77.8% in 2016–2021 in older adults (Table 1).
Table 1.
Characteristics and care settings of tramadol initiations: United States, 2005–2021
| Age 18–64 Years | Age ≥65 Years | |||||
|---|---|---|---|---|---|---|
| 2005–2010 | 2011–2015 | 2016–2021 | 2005–2010 | 2011–2015 | 2016–2021 | |
| Characteristics | (N=904,236) | (N=1,128,015) | (N=1,243,847) | (N=301,167) | (N=700,919) | (N=1,451,468) |
| Female, No. (%) | 556,775 (61.57) | 683,573 (60.60) | 734,874 (59.08) | 206,864 (68.69) | 467,295 (66.67) | 935,916 (64.48) |
| Age group, years, No. (%) | ||||||
| 18–35 | 197,126 (21.80) | 236,754 (20.99) | 209,672 (16.86) | - | - | - |
| 36–49 | 348,671 (38.56) | 379,164 (33.61) | 383,185 (30.81) | - | - | - |
| 50–64 | 358,439 (39.64) | 512,097 (45.40) | 650,990 (52.34) | - | - | - |
| 65–74 | - | - | - | 145,796 (48.41) | 353,230 (50.40) | 742,484 (51.15) |
| ≥75 | - | - | 155,371 (51.59) | 347,689 (49.60) | 708,984 (48.85) | |
| Race and ethnicity, No. (%)a | ||||||
| Asian | 14,820 (1.64) | 23,689 (2.10) | 28,112 (2.26) | 5,039 (1.67) | 15,429 (2.20) | 31,017 (2.14) |
| Black | 106,745 (11.80) | 157,889 (14.00) | 178,984 (14.39) | 36,727 (12.19) | 89,256 (12.37) | 204,500 (14.09) |
| Hispanic | 75,649 (8.37) | 132,114 (11.71) | 165,519 (13.31) | 34,057 (11.31) | 92,539 (13.20) | 191,734 (13.21) |
| Unknown | 110,552 (12.23) | 43,255 (3.83) | 58,230 (4.68) | 17,106 (5.68) | 30,397 (4.34) | 74,423 (5.13) |
| White | 596,470 (65.96) | 771,068 (68.36) | 813,002 (65.36) | 208,238 (69.14) | 473,298 (67.53) | 949,794 (65.44) |
| Opioid naïve, No. (%)b | 547,680 (60.57) | 757,838 (67.18) | 915,689 (73.62) | 190,216 (63.16) | 499,561 (71.27) | 1,128,910 (77.78) |
| Prescribing provider, No. (%) | ||||||
| Primary care | 412,772 (45.65) | 458,524 (40.65) | 386,252 (31.05) | 180,242 (59.85) | 380,156 (54.24) | 616,827 (42.50) |
| Orthopedics | 87,144 (9.64) | 108,603 (9.63) | 127,486 (10.25) | 23,170 (7.69) | 60,642 (8.65) | 147,005 (10.13) |
| Emergency | 56,081 (6.20) | 118,058 (10.47) | 140,615 (11.30) | 9,747 (3.24) | 45,239 (6.45) | 108,099 (7.45) |
| OB/GYN | 14,983 (1.66) | 24,883 (2.21) | 31,507 (2.53) | 765 (0.25) | 2,016 (0.29) | 5,754 (0.40) |
| Dental providers | 5,885 (0.65) | 25,349 (2.25) | 48,146 (3.87) | 751 (0.25) | 8,199 (1.17) | 28,761 (1.98) |
| Tramadol dispensed on index date, median (IQR) | ||||||
| Tramadol MMEc | 300 (225–600) | 300 (200–600) | 240 (150–300) | 500 (300–750) | 400 (300–600) | 300 (200–450) |
| Tramadol days supply | 8 (5–15) | 8 (5–15) | 6 (4–8) | 12 (7–30) | 10 (7–25) | 7 (5–14) |
| Tramadol daily MME | 43 (30–60) | 40 (30–60) | 40 (30–57) | 40 (30–60) | 40 (30–56) | 40 (30–43) |
| Other opioids dispensed on index date, No. (%) | 52,537 (5.81) | 53,717 (4.76) | 50,686 (4.07) | 11,855 (3.94) | 23,408 (3.34) | 54,463 (3.75) |
| Other opioid MME, median (IQR) | 300 (150–600) | 300 (150–570) | 240 (150–450) | 400 (200–864) | 300 (175–600) | 300 (150–450) |
| Persistent tramadol use, No. (%)d | 53,971 (5.97) | 64,648 (5.73) | 52,107 (4.19) | 34,982 (11.62) | 69,453 (9.91) | 96,805 (6.67) |
| Co-dispensed CNS agents, No. (%)e | 348,031 (38.49) | 423,035 (37.50) | 469,835 (37.77) | 85,554 (28.41) | 198,935 (28.38) | 466,282 (32.12) |
| Benzodiazepines | 122,633 (13.56) | 154,750 (13.72) | 172,248 (13.85) | 24,776 (8.23) | 74,368 (10.61) | 176,946 (12.19) |
| Gabapentinoids | 54,525 (6.03) | 86,905 (7.70) | 159,911 (12.86) | 26,546 (8.81) | 76,710 (10.94) | 238,629 (16.44) |
| Muscle relaxants | 196,694 (21.75) | 221,124 (19.60) | 218,313 (17.55) | 31,013 (10.30) | 49,651 (7.08) | 106,031 (7.31) |
| Sleep medications | 62,508 (6.91) | 68,499 (6.07) | 53,141 (4.27) | 18,596 (6.17) | 32,616 (4.65) | 31,764 (2.19) |
| Co-dispensed opioids, No. (%)e | 315,037 (34.84) | 329,226 (29.19) | 291,120 (23.40) | 87,252 (28.97) | 160,101 (22.84) | 276,989 (19.08) |
| Hydrocodone | 220,471 (24.38) | 239,945 (21.27) | 176691 (14.21) | 56,061 (18.61) | 118,215 (16.87) | 166,604 (11.48) |
| Oxycodone | 78,963 (8.73) | 100,805 (8.94) | 118,609 (9.54) | 17,354 (5.76) | 45,108 (6.44) | 111,108 (7.65) |
As provided in Clinformatics database.
No use of any opioids for at least 6 months before tramadol initiation.
MMEs of all pills dispensed on index date.
≥90 days covered of the 180 days following tramadol initiation.
Dispensed within 30 days pre- or post-tramadol initiation. CNS agents included benzodiazepines, gabapentinoids, muscle relaxants, and sleep medications; other opioids included non-tramadol opioids (Appendix Table 1).
CNS, central nervous system; IQR, interquartile range (25th – 75th percentile); MME, morphine milligram equivalents; OB/GYN, obstetrics and gynecology.
The proportion of tramadol initiations prescribed by primary care providers dropped by one-third in both age groups (Table 1), with corresponding increases in ER prescribing (from 6.2% to 11.3% of tramadol initiations in younger adults and from 3.2% to 7.5% of tramadol initiations in older adults). Dental providers prescribed a small but growing share of tramadol initiations, rising 6-fold between 2005–2010 and 2016–2021 in younger adults, and with a nearly 8-fold increase in older adults (Table 1). Median total MME of tramadol dispensed on the initiation date remained constant in younger adults but decreased from 500 (IQR: 300–750) during 2005–2010 to 300 (IQR: 200–450) during 2016–2021 in older adults (Table 1). Daily MME remained relatively constant over time in both age groups at a median of 40 MME/day (IQR: 30–60). Between 3% and 6% of tramadol initiations were accompanied by a dispensing of another opioid (Table 1); in those who had other opioids co-dispensed during 2016–2021, the median total MME of other opioids was 240 (IQR: 150–450) in younger adults and 300 (IQR: 150–450) in older adults. Persistent use of tramadol declined over time, with 4.2% of younger adults and 6.7% of older adults exhibiting persistent use during 2016– 2021 (Table 1).
Other CNS agents were dispensed within 30 days before or after tramadol initiation in over one-third of tramadol initiations (Table 1). Co-dispensing of other CNS agents increased over time in older adults (Table 1), particularly for benzodiazepines, which increased from 8.2% (2005–2010) to 12.2% (2016–2021) of tramadol initiations. Gabapentinoid co-dispensing nearly doubled in both age groups, rising from 6.0% to 12.9% of tramadol initiations in younger adults and from 8.8% to 16.4% in older adults. Other opioids were dispensed within 30 days before or after tramadol initiation in 34.8% of tramadol initiations in younger adults and 29.0% of tramadol initiations in older adults in 2005–2010, decreasing to 23.4% and 19.1% in 2016–2021, respectively. Among co-dispensed opioids, hydrocodone was the most common (Table 1).
The most frequent comorbidities noted prior to tramadol initiation in both age groups were arthropathies other than osteoarthritis, which included rheumatoid arthritis and all other forms of arthropathies. These were recorded in half or more of individuals initiating tramadol (Figure 3). Back and neck pain was the second most prevalent pain condition (36.7% of younger adults and 42.0% of older adults initiating tramadol during 2016–2021). In both age groups, the prevalence of osteoarthritis, abdominal pain, bone fracture, diabetic neuropathy, and stage III-IV chronic kidney disease increased between 2005–2010 and 2016–2021. Only 4.7% of younger adults and 11.2% of older adults initiating tramadol had cancer (Figure 3).
Figure 3. Pain and comorbidities in people initiating tramadol: United States, 2005–2021.
Note: CKD = chronic kidney disease, OA = osteoarthritis. Conditions assessed in 180 days before and including the tramadol initiation date. “Arthropathies other than OA” includes rheumatoid arthritis and all other forms of arthropathy that do not fall under OA. The percentage of tramadol initiations noted in the figures is for 2016–2021 period. (A) Ages 18–64 years at initiation. (B) Ages ≥65 years at initiation.
Sex-specific evaluation of comorbidities is presented in Appendix Figure 2. Females in the 18–64 age group initiating tramadol had higher prevalence of migraine, fibromyalgia, and abdominal pain than males. Older females had higher prevalence of bone fracture, osteoarthritis, and other arthropathies than older males (Appendix Figure 2).
Discussion
Our investigation demonstrates that tramadol use is relatively common in the US, particularly in older adults, with the highest use observed in females ≥65 years of age. While overall tramadol use and initiation declined since Schedule IV designation of tramadol in 2014, changes were heterogeneous across states. Over time, primary care providers initiated fewer episodes of tramadol therapy, while specialists in emergency medicine, orthopedics, and dentistry initiated more episodes of therapy. Co-dispensing of one or more CNS depressants known to interact with opioids occurred in one-third of individuals initiating tramadol.
Higher use of tramadol in older adults is concerning because these individuals are more likely to experience adverse events from opioid therapy.19,20 The updated 2023 American Geriatrics Society Beers Criteria® for potentially inappropriate medications in older adults lists tramadol and other opioids as medications that should be used with caution in this population due to increased risk of CNS adverse events.21 In addition, combining opioids with other CNS depressants, particularly benzodiazepines or gabapentinoids, can lead to clinically important drug-drug interactions.21 A third of older adults initiating tramadol in our study received a benzodiazepine, gabapentinoid, muscle relaxant, or a sleep medication within 30 days before and after tramadol initiation. Moreover, co-dispensing of tramadol with benzodiazepines, as well as gabapentinoids, increased over time. While the risk of adverse events when these medications are taken concomitantly with tramadol is not known, the co-dispensing trends should be further monitored and future studies should evaluate the safety of combining tramadol with other CNS depressants, especially in older adults.
In addition to age differences, sex differences were observed. Females demonstrated higher rates of incident and prevalent tramadol use than males across both age groups and throughout the study period. Our findings are consistent with other studies that reported higher opioid use in females.22 Also in line with previous reports,23,24 females in our study had higher prevalence of certain chronic pain conditions, such as osteoarthritis, fibromyalgia, migraines, and abdominal pain, as compared with men. Further investigation into whether the observed differences in use are driven by a higher burden of pain conditions, difference in how females process pain, differential prescribing patterns, including avoidance of potent opioids, or other factors is needed to ensure safe and optimal prescribing for both sexes.
While, on average, tramadol use decreased in the US since 2014 in all sex and age groups, there was considerable between-state variation in post-2014 trends. Some states displayed no change or an increase in use. This observation is consistent with findings from at least one other investigation that examined state-specific opioid prescribing patterns which found wide variability across states in the quantity, dose, and duration of opioid fills.25 It is also not surprising as states can implement state-specific regulations and interventions targeting both opioid prescribing and monitoring.26,27 Additional studies evaluating state-specific laws and practices are needed to determine the extent to which these are responsible for heterogeneity across states.
Finally, a shift in the setting of tramadol initiation is noteworthy. Across sex and age groups, there was a gradual decrease in the proportion of tramadol initiations prescribed by primary care providers. Notably, that decrease was accompanied by an increase in ER prescribing, which almost doubled. The increase in ER prescribing is consistent with other data from the US. A recent US study observed a 70% increase in ER dispensing of tramadol, from 1.7% to 2.9% of all ER visits, from 2007 to 2018.1 The shift in prescribing from primary care settings to specialized and acute care settings may be due to increased regulatory and public health attention focused on chronic rather than acute opioid therapy.28 However, prescribing patterns require further investigation and monitoring as tramadol effectiveness and safety has not been fully established in acute care settings, yet it has been linked to increased risk of persistent opioid use.19,29,30
Limitations
Our results should be interpreted with several limitations in mind. First, like any study based on health insurance claims data, we were limited to diagnoses and medication fills in medical or pharmacy claims. Pharmacy claims only include outpatient medications paid through prescription drug benefits. Therefore, tramadol use in inpatient setting or outpatient filles not billed through insurance are not recorded in these data, leading to potential underestimation of incident and prevalent use. On the other hand, patients might not have used all or any of the dispensed tramadol. Thus, our trends should be interpreted as dispensing trends, and not strictly as use trends. It is also possible that concomitant medications were used in sequence or kept on hand for as-needed use. Second, while we had information on race and ethnicity, these data were not self-reported and the exact algorithm that the data vendor used to assign enrollees into racial and ethnic groups is unknown. More importantly, our study was not designed to evaluate equitable access to tramadol or pain medications across ethnic and racial groups. While the proportion of non-White individuals initiating tramadol increased over time, that increase was accompanied by a decrease in the proportion of “Unknown” group, suggesting that the change might have been driven by changing algorithms or their ability to estimate race/ethnicity. Future studies should evaluate trends in tramadol use across racial, ethnic, and socioeconomics groups. Finally, since we used a commercial health insurance claims database of employment-based health plans and Medicare Advantage, our findings may not be generalizable to individuals without health insurance or those who are publicly insured (Medicaid or traditional Medicare). A recent study on the utilization of opioids in the US Medicaid programs, however, yielded similar tramadol utilization trends in Medicaid population.31
Conclusions
In conclusion, this investigation contributes to the understanding of evolving trends in tramadol use over the past 20 years in commercially insured adults in the US. Despite decreasing utilization since 2014, prescription tramadol use remains high, particularly among older females. Changing use, especially in older adults with other comorbidities, warrants further pharmacovigilance.
Supplementary Material
Acknowledgements
This study was funded through US Department of Health and Human Services, Agency for Healthcare Research and Quality (AHRQ grant# R01 HS027623). Dr. Bykov was also supported by a grant from National Institute on Aging (K01AG068365). Funders had no role in designing the study; collecting, analyzing, or interpreting data; or decisions related to writing and submitting this manuscript for publication.
CB reports consulting fees from the World Health Organization. KH reports being an investigator on grants to Brigham and Women’s Hospital from Takeda and UCB, unrelated to this work. KB reports consulting fees from Alosa Health, outside the submitted work. HE, CSH, BTB have no financial disclosures to report.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Mullins PM, Mazer-Amirshahi M, Pourmand A, Perrone J, Nelson LS, Pines JM. Tramadol Use in United States Emergency Departments 2007–2018. The Journal of Emergency Medicine. 2022;62(5):668–674. doi: 10.1016/j.jemermed.2022.01.004 [DOI] [PubMed] [Google Scholar]
- 2.Peterman N, Shivdasani K, Pagani N, et al. National Trends in Orthopaedic Pain Management from 2016 to 2020. J Am Acad Orthop Surg. Published online February 28, 2024. doi: 10.5435/JAAOS-D-23-00806 [DOI] [PubMed] [Google Scholar]
- 3.Fauer AJ, Davis MA, Choi SW, Wallner LP, Friese CR. Use of gabapentinoid medications among US adults with cancer, 2005–2015. Support Care Cancer. 2020;28(1):5–8. doi: 10.1007/s00520-019-05100-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Bykov K, He M, Gagne JJ. Trends in Utilization of Prescribed Controlled Substances in US Commercially Insured Adults, 2004–2019. JAMA Intern Med. 2020;180(7):1006. doi: 10.1001/jamainternmed.2020.0989 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Dowell D, Ragan KR, Jones CM, Baldwin GT, Chou R. CDC Clinical Practice Guideline for Prescribing Opioids for Pain — United States, 2022. MMWR Recomm Rep. 2022;71(3):1–95. doi: 10.15585/mmwr.rr7103a1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Curtis JR, Xie F, Smith C, et al. Changing Trends in Opioid Use Among Patients With Rheumatoid Arthritis in the United States. Arthritis & Rheumatology. 2017;69(9):1733–1740. doi: 10.1002/art.40152 [DOI] [PubMed] [Google Scholar]
- 7.Bigal LM, Bibeau K, Dunbar S. Tramadol Prescription over a 4-Year Period in the USA. Curr Pain Headache Rep. 2019;23(10):76. doi: 10.1007/s11916-019-0777-x [DOI] [PubMed] [Google Scholar]
- 8.Sørensen AMS, Rasmussen L, Ernst MT, et al. Use of tramadol and other analgesics following media attention and risk minimization actions from regulators: a Danish nationwide drug utilization study. European Journal of Clinical Pharmacology. 2021;77(4):617–624. doi: 10.1007/s00228-020-03016-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Chen TC, Chen LC, Knaggs RD. A 15-year overview of increasing tramadol utilisation and associated mortality and the impact of tramadol classification in the United Kingdom. Pharmacoepidemiology and Drug Safety. 2018;27(5):487–494. doi: 10.1002/pds.4320 [DOI] [PubMed] [Google Scholar]
- 10.Silverman S, Rice JB, White AG, et al. Clinical and economic burden of prescribing tramadol and other opioids for patients with osteoarthritis in a commercially insured population in the United States. Pain. 2022;163(1):75–82. doi: 10.1097/j.pain.0000000000002300 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Drug Enforcement Administration. Schedule of controlled substances: placement of tramadol into schedule IV. Final rule. Federal Register. 2014;79(127):37623–37630. [PubMed] [Google Scholar]
- 12.Drug Enforcement Administration. Tramadol (Trade Names: Ultram®, Ultracet®).; 2023. https://www.deadiversion.usdoj.gov/drug_chem_info/tramadol.pdf
- 13.Subedi M, Bajaj S, Kumar MS, YC M. An overview of tramadol and its usage in pain management and future perspective. Biomedicine and Pharmacotherapy. 2019;111(December 2018):443–451. doi: 10.1016/j.biopha.2018.12.085 [DOI] [PubMed] [Google Scholar]
- 14.Hilbe JM. Negative Binomial Regression. Cambridge University Press; 2011. [Google Scholar]
- 15.Snijders TA, Bosker RJ. Multilevel Analysis. SAGE; 2012. [Google Scholar]
- 16.Savitz DA, Wise LA, Bond JC, et al. Responding to Reviewers and Editors About Statistical Significance Testing. Ann Intern Med. Published online February 20, 2024:M23–2430. doi: 10.7326/M23-2430 [DOI] [PubMed] [Google Scholar]
- 17.Wasserstein RL, Lazar NA. The ASA Statement on p -Values: Context, Process, and Purpose. The American Statistician. 2016;70(2):129–133. doi: 10.1080/00031305.2016.1154108 [DOI] [Google Scholar]
- 18.Food and Drug Administration. FDA Drug Safety Communication: FDA recommends against the continued use of propoxyphene. Published November 19, 2010. Accessed April 30, 2024. https://www.fda.gov/drugs/drug-safety-and-availability/fda-drug-safety-communication-fda-recommends-against-continued-use-propoxyphene
- 19.Musich S, Wang SS, Schaeffer JA, Slindee L, Kraemer S, Yeh CS. Safety Events Associated with Tramadol Use among Older Adults with Osteoarthritis. Population Health Management. 2021;24(1):122–132. doi: 10.1089/pop.2019.0220 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Makris UE, Abrams RC, Gurland B, Reid MC. Management of Persistent Pain in the Older Patient: A Clinical Review. JAMA. 2014;312(8):825. doi: 10.1001/jama.2014.9405 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.By the 2023 American Geriatrics Society Beers Criteria® Update Expert Panel. American Geriatrics Society 2023 updated AGS Beers Criteria® for potentially inappropriate medication use in older adults. Journal of the American Geriatrics Society. 2023;71(7):2052–2081. doi: 10.1111/jgs.18372 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Schieber LZ, Guy GP, Seth P, Losby JL. Variation in Adult Outpatient Opioid Prescription Dispensing by Age and Sex — United States, 2008–2018. MMWR Morb Mortal Wkly Rep. 2020;69(11):298–302. doi: 10.15585/mmwr.mm6911a5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Bartley EJ, Fillingim RB. Sex differences in pain: a brief review of clinical and experimental findings. British Journal of Anaesthesia. 2013;111(1):52–58. doi: 10.1093/bja/aet127 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Rikard SM, Strahan AE, Schmit KM, Guy GP. Chronic Pain Among Adults — United States, 2019–2021. MMWR Morb Mortal Wkly Rep. 2023;72(15):379–385. doi: 10.15585/mmwr.mm7215a1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Schieber LZ, Guy GP, Seth P, et al. Trends and Patterns of Geographic Variation in Opioid Prescribing Practices by State, United States, 2006–2017. JAMA Netw Open. 2019;2(3):e190665. doi: 10.1001/jamanetworkopen.2019.0665 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Dowell D, Zhang K, Noonan RK, Hockenberry JM. Mandatory Provider Review And Pain Clinic Laws Reduce The Amounts Of Opioids Prescribed And Overdose Death Rates. Health Affairs. 2016;35(10):1876–1883. doi: 10.1377/hlthaff.2016.0448 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Chang H, Murimi I, Faul M, Rutkow L, Alexander GC. Impact of Florida’s prescription drug monitoring program and pill mill law on high-risk patients: A comparative interrupted time series analysis. Pharmacoepidemiology and Drug. 2018;27(4):422–429. doi: 10.1002/pds.4404 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Hwang CS, Kang EM, Ding Y, et al. Patterns of Immediate-Release and Extended-Release Opioid Analgesic Use in the Management of Chronic Pain, 2003–2014. JAMA Netw Open. 2018;1(2):e180216. doi: 10.1001/jamanetworkopen.2018.0216 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Machado-Alba JE, Serna-Echeverri LS, Valladales-Restrepo LF, Machado-Duque ME, Gaviria-Mendoza A. Use of Tramadol or Other Analgesics in Patients Treated in the Emergency Department as a Risk Factor for Opioid Use. Hu L, ed. Pain Research and Management. 2020;2020:1–8. doi: 10.1155/2020/8847777 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Motov SM, Vlasica K, Middlebrook I, LaPietra A. Pain management in the emergency department: a clinical review. Clin Exp Emerg Med. 2021;8(4):268–278. doi: 10.15441/ceem.21.161 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Alsultan MM, Guo JJG. Utilization, Spending, and Price of Opioid Medications in the US Medicaid Programs Between 1991 and 2019. Am Health Drug Benefits. 2022;15(1):31–37. [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.