PDF
Citation
Print
INTRODUCTION
Osteoarthritis (OA) is a prevalent joint disease that causes disability; its burden is increasing with population aging and the growing numbers of obese individuals. Pain is the most common complaint of patients with OA, but there are no safe effective treatments for controlling the pain or improving the joint damage. The only pharmacological treatment recommended for use in the non-surgical management of knee OA in the 2019 Osteoarthritis Research Society International (OARSI) guidelines was that of topical non-steroidal anti-inflammatory drugs (NSAIDs).1 Oral NSAIDs are also strongly recommended in the 2019 American College of Rheumatology (ACR) OA management guidelines, emphasizing clinical considerations aimed at risk mitigation for their safe use, such as use of the lowest dose possible and short-term use.2
Consequently, the use of medications without adequate clinical evidence of their efficacy and safety is widespread. In the USA, OA patients were significantly more likely to receive adjunctive medications for treating depression, anxiety, and insomnia than controls.3 Alarmingly, of five musculoskeletal diseases (OA, gout, fibromyalgia, rheumatoid arthritis, and low back pain), the use of amphetamines, anxiolytics, sedatives, and hypnotics was highest for OA, and increased from 1998 to 2014.4 The increased use of these medications is concerning, given that there is little scientific evidence to support their prescription in patients with OA, due to the side effects of such medications. Even tramadol, a weak opioid agonist strongly recommended by the American Academy of Orthopedic Surgeons guidelines,5 was associated with a significantly higher 1-year mortality rate than NSAIDs among patients aged 50 years and older with OA.6
Many of the strongly recommended treatments are non-pharmacological, including exercise, weight loss, and self-efficacy and self-management programs. However, the adoption of these modalities is time-consuming and labor-intensive. The best outcomes in chronic, aging-associated diseases such as OA are achieved through appropriate patient education regarding these non-pharmacological options and self-help.7 However, the Korean healthcare system is making it harder for clinicians to spend more time with patients, due to the low reimbursement rate for physicians and increased administrative tasks, such as the electronic medical record and insurance certification requirements. In Korea, where short-consultations dubbed “3-minute consultations” are widespread, clinicians may order unnecessary diagnostic tests and prescriptions as a substitute for time spent face-to-face.
Therefore, this study examined the trend in off-label prescription medication use among older patients with OA in Korea, focusing on psychotropic drugs.
METHODS
Dataset
This study used data from the National Health Insurance (NHI) database, a public database containing details such as healthcare utilization, health screening, sociodemographic variables, and mortality for the entire population of South Korea covered by NHI and Medical Aid since 2002. It contains information on inpatient and outpatient medical care (diagnosis, length of hospital stay, and treatment received) and prescription records (drug codes, duration of prescription, and daily dosage).
Study population
Patients with knee OA were identified using a combination of diagnostic, prescription, and radiograph procedure codes. Patients had to have at least 12 months of valid medical claims in the database before study entry to assure complete medical records. Patients who were diagnosed with inflammatory arthritis or who died less than 1 year after study inclusion were excluded. We constructed operational definitions by combining three conditions that were used in a previous report using Korean claims data with some modification as follows: diagnostic codes M15, M16, M17, or M19 and knee or hip x-rays; and diagnostic codes M15, M16, M17, or M19 and a prescription medication approved only for OA (Supplementary Table 1). This definition has respective accuracy, sensitivity, specificity, and positive and negative predictive values of 75%, 81%, 48%, 87%, and 36%.8 Among 4,654,737 patients with knee and hip OA who used healthcare between January 1, 2011, and December 31, 2014, 2,769,311 met the operational definition of OA. From these, a cohort of 1,821,158 patients ≥ 65 years was identified and included in the analysis. We focused on patients older than 65, based on the chronological definition of old age among Koreans, due to their increased risk of adverse effects from using psychotropic drugs.9 Patients were followed until the end of the study period (December 31, 2020) or until 1 year before a first diagnosis of a malignancy or death. In addition, a subset of patients with OA who had comorbidities for which antipsychotics could be prescribed (OA with comorbidities) was identified. The comorbidities for which antipsychotics could be prescribed were mental and behavioral disorders (major depression, psychosis and schizophrenia, bipolar disorder, anxiety disorder, and sleep disorder), neuropathic pain disorders (diabetic neuropathy, neuropathy due to spondylosis, intervertebral disc disorders, sciatica, spinal cord or nerve injury, trigeminal neuralgia, post-infectious neuropathy, hereditary neuropathy, paraneoplastic neuropathy, toxic and metabolic neuropathy, demyelinating diseases, inflammatory neuropathy, neuropathy associated with autoimmune disease, and unspecified neuropathy), complex regional pain syndrome, epilepsy, and fibromyalgia.
Exposure and outcome
The study medications included the following psychotropics (Supplementary Table 2): opioids, barbiturate, benzodiazepines (long- and short-acting), tricyclic antidepressants (TCAs), other antidepressants, Z-drugs, anti-epileptics, selective serotonin reuptake inhibitors (SSRIs), serotonin–norepinephrine reuptake inhibitors (SNRIs), anxiolytics, and typical and atypical antipsychotics.
The prevalent prescription rate was defined as the total number of patients who had at least one prescription for the specific drug divided by the number of persons per calendar year. The long-term prescription rate was calculated as the total number of patients who were prescribed a specific drug for more than 3 months divided by the number of persons per calendar year.
Statistical analyses
Descriptive statistics were used to assess baseline characteristics and to calculate the prevalent and long-term prescription rates of the medications prescribed.
Ethics statement
This study used de-identified data and was conducted without patient and public involvement. The Institutional Review Board (IRB) of the National Evidence-based healthcare Collaborating Agency (NECA) approved the use of National Health Insurance (NHI) data and the study protocol ([NECA] IRB 22-003, NHIS-2022-1-656).
RESULTS
Study population
The study population comprised 1,821,158 patients with OA. The average follow-up time per patient was 110.55 months (3,362.43 days). The mean patient age was 71.7 ± 6.5 years and 65.3% were female (Table 1). Knee OA was more common than hip OA.
Table 1
Characteristics of study subjects
The cohort subset of OA patients with comorbidities for which psychotropics may be indicated comprised 891,874 patients (49.0%). There was no difference in age between the subsets with and without co-morbidities, although the proportion of women was higher in the subset with comorbidities. The most common diagnosis was neuropathic pain disorder, followed by fibromyalgia.
Temporal trends in the prescription of psychotropics
The prevalent prescription rates of psychotropics for all patients decreased from 58.2% at the start of the study period in 2011 to 52.0% in 2018 and then leveled off (Fig. 1). The trend was similar for men and women, for those with and without co-morbidities, and for those ≤ 75 years. However, men, those with comorbidities, and those older than 75 years had slight increases in prescriptions in 2019 and 2020. The long-term prescription rate, defined as more than 90 days of prescription, remained constant until 2018 and then increased slightly (from 26.2% in 2011 to 28.21% in 2020 [data not shown]).
Prescription rates of specific psychotropics
The most commonly prescribed classes of psychotropics were opioids (including tramadol and tramadol/acetaminophen combination), long-acting benzodiazepines, and short-acting benzodiazepines. These were prescribed in 52.4–64.1%, 16.0–28.8%, and 19.5–22.1% of the cases throughout the study period, respectively (Table 2). The prescription rates of opioids and long-acting benzodiazepines decreased from 2011 to 2020. In comparison, the prescription rates for anti-epileptics, SSRIs, SNRIs, other antidepressants, and atypical antipsychotics increased over the study period (Fig. 2). Among those with co-morbidities that needed psychotropic prescriptions, opioids, long- and short-acting benzodiazepines, anti-epileptics, Z-drugs, and TCAs were the most commonly prescribed, with more than 10% of the subjects given prescriptions, while for those without such co-morbidities, only opioids and long- and short-acting benzodiazepines were prescribed in more than 10% of patients (Supplementary Tables 3 and 4). For both groups, the prescription rates of opioids and long-acting benzodiazepines decreased over the study period. Some less commonly prescribed classes had increased prescription rates during the study period. For those with psychiatric co-morbidities, the prescription rates of anti-epileptics, atypical psychotropics, and SNRIs increased from 13.7% to 28.1%, 1.5% to 5.2%, and 1.2% to 4.2%, respectively. For those without such co-morbidities, the prescription rates of anti-epileptics, SSRI, other antidepressants, and atypical psychotropics increased from 2.6% to 6.3%, 2.5% to 5.4%, 1.6% to 3.9%, and 1.2% to 7.6%, respectively. Considering specific drugs (aside from tramadol and tramadol/acetaminophen, which are included in the OA treatment recommendations and approved for OA treatment), the most commonly prescribed psychotropics were diazepam and alprazolam, which were prescribed in 26.8–12.1% and 13.5–11.8% of cases, respectively (Table 3). Other commonly prescribed psychotropics included zolpidem, amitriptyline, etizolam, gabapentin, lorazepam, buspirone, fentanyl, pregabalin, and quetiapine. The prescription rates of diazepam, amitriptyline, and etizolam decreased, while those of gabapentin, pregabalin, fentanyl, and quetiapine increased. For those with psychiatric co-morbidities, the prescription rates of gabapentin, pregabalin, and fentanyl increased from 9.4% to 15.8%, 2.4% to 14.2% and 3.3% to 6.7%, respectively (Supplementary Table 5). In comparison, for those without such co-morbidities, the prescription rates of lorazepam, fentanyl, escitalopram, and quetiapine increased from 2.7% to 4.0%, 1.6% to 3.6%, 1.3% to 4.0%, and 0.7% to 6.6%, respectively (Supplementary Table 6).
Table 2
Prescription rates of specific class of psychotropics
Fig. 2
Prescription rates of specific class of psychotropics.
TCA = tricyclic antidepressants, SSRI = selective serotonin reuptake inhibitor.
Table 3
Prescription rates of 10 most commonly prescribed psychotropics
DISCUSSION
Using data collected from the NHI, we found that more than half of older patients with OA were prescribed psychotropic drugs. Although the prescription rate decreased over the study period, the long-term prescription rate remained constant until 2018 and then increased slightly. The prescription rates of each psychotropic class or individual psychotropic varied, with decreases in opioids and long-acting benzodiazepines and increases in anti-epileptics, SSRIs, SNRIs, and atypical antipsychotics. For both those with and without psychiatric comorbidities, the prescription rate of fentanyl increased more than two-fold.
Guideline recommended medication for the treatment of OA has not changed significantly in recent decades10, and most recommended medications fall short of providing patient satisfaction because of their small effect size and safety issues with long-term use. With such a huge unmet need, it is not surprising that medications are prescribed off-label to relieve OA pain, without proper evidence. For example, long-term opioids are commonly prescribed to patients with OA, and in a recent report from Sweden using data from the Skåne Healthcare Register, the estimated inappropriate dispensing defined as dispensing opioids within the first year of diagnosis or long-term opioid use (> 90 day supply within a 6-month period) attributable to OA was 7.4% for knee OA and 12.8% for hip OA.11 Among persons with prevalent knee or hip OA or both, inappropriate, long-term opioid use attributable to OA was 1.3%, 2.0%, and 2.4%, respectively. Of note, at least 50% of all incident opioid dispensations within the first year of knee OA diagnosis and 62% of those for hip OA diagnosis were deemed inappropriate. In the US, despite the public outcry over the opioid epidemic, one study showed that 17.0% of patients with any joint OA, 13.4% of patients with hip OA, and 15.9% with knee OA were prescribed an opioid for their condition, with the prescription rate being stable from 2007 to 2014.12 It is important to know whether Korean OA patients are prescribed more psychotropic drugs than patients in other countries. Despite differences in the specific medications used and data presentation, the prescription rates of opioids, gabapentinoids, and antidepressants tended to be higher in our patients compared to patients in Sweden, the UK, and the Netherlands.
Although knee OA pain is largely considered to arise from joint damage, recent studies show that as many as one in three patients with knee OA report allodynia and hyperalgesia, hallmarks of neuropathic pain.13 In addition, nociplastic pain, defined as altered nociception without clear evidence of tissue damage or disease of the somatosensory system causing the pain, was found in 15–50% of patients with OA.14 OA patients with non-nociceptive pain may not respond to guideline-recommended OA treatments, such as NSAIDs, corticosteroid injections, or joint replacement surgery. Consequently, researchers have examined the efficacy and safety of psychotropic drugs, which might affect central pain processing, in the treatment of OA pain. The OARSI and ACR recommend the SNRI duloxetine for managing OA pain, especially in patients with concomitant depression or widespread pain.1 A recent meta-analysis reporting the efficacy and safety of any antidepressant drug in participants with hip or knee OA reported low-certainty evidence for SNRIs for OA pain at 3–13 weeks and moderate certainty evidence for disability at ≤ 2 weeks.15 Seven of the eight included trials evaluated duloxetine. A Cochrane review reported analyses from nine trials of antidepressants in OA patients: duloxetine in six and milnacipran, the SSRI fluvoxamine, and the TCA nortriptyline in one each.16 The review found that, compared with placebo, antidepressants provided a small, but clinically important improvement in overall pain, based on high‐certainty evidence. Combined, the studies show that participants receiving an antidepressant were 55% more likely to have a clinically important reduction in pain compared to those receiving placebo. However, the study results showed a moderate degree of variation. The anticonvulsants pregabalin and gabapentin can inhibit centralized pain, and a prospective, randomized clinical trial of 65 participants with hand OA showed that pregabalin improved pain and function compared to placebo.17 Another randomized prospective study of 89 knee OA patients showed that meloxicam in combination with pregabalin improved pain compared to meloxicam alone.18
Despite such limited evidence of efficacy, the prescription of such medications has increased rapidly. A retrospective cohort study of 383,680 newly diagnosed OA cases showed an almost 3-fold increase in first prescriptions of gabapentin and pregabalin between 2005 and 2015.19 A population-based cohort study of 164,292 OA patients in the Integrated Primary Care Information database of the Netherlands showed that the prescription rates of TCAs and anticonvulsants (amitriptyline, gabapentin, and pregabalin) increased significantly over time.20 In another study using the UK Clinical Practice Research Datalink database of subjects aged > 40 years, the incidence of first gabapentinoid prescription and the event rate of gabapentinoid–opioid co-prescription was higher in OA patients than in non-OA subjects.21 Between 2000 and 2017, the rates of gabapentinoid–opioid co-prescriptions roughly trebled, across all gender and age groups and geographical regions, suggesting a general rather than local or isolated phenomenon. Although we found an overall decreased rate of psychotropic use among Korean patients, more than half of OA patients were prescribed psychotropics. The stagnant long-term prescription rate and increases in the prescription of fentanyl and quetiapine are troublesome, considering the lack of evidence and the probable significant adverse effect of such medications in this older population.
Most OA treatment guidelines recommend exercise, patient education, and weight loss (if indicated) as appropriate first-line treatment, instead of pharmaceutical treatment. However, such non-therapeutic approaches are not adequately implemented in real-world situations. In a US study using a random sample of Medicare data from 2016 to 2019, while opioid prescribing decreased from 56.7% to 46.5%, the use of non-pharmaceutical therapies, such as physical therapy, lagged opioids, increasing from 22.8% to only 25.5%.22 In a clinical setting where consultation time is not properly reimbursed, as occurs in Korea, over-reliance on prescription medication may worsen. A Korean survey of rheumatologists reported that more than 80% had inadequate consultation time for their patients and 50% responded that inadequate consultation time leads to unnecessary prescription.2324
Our study has some limitations. First, we excluded younger OA patients and patients with hand OA, which may have introduced biases due to differences in psychotropic drug use in those patient groups. Although our study applied various algorithms to identify OA cases from the administrative database, it was impossible to confirm the diagnosis individually. The choice of the operational definition for the study subjects as ‘those with one or more claims with OA diagnostic codes and knee X-rays, or approved drug prescriptions only for OA’ was arbitrary, although it had high accuracy, sensitivity, and positive predictive value. In addition, coding inaccuracy in diagnosis was plausible because the NHI data are based on payment claims and physicians sometimes use incorrect diagnoses to meet the health insurance coverage requirements. All data are presented descriptively, and statistical analysis for trend changes could not be applied due to limitations of resources. Lastly, because of a funding issue and the early dismissal of our research team, polypharmacy could not be pursued further. Polypharmacy is a serious problem, in which a single patient may use both an approved drug (such as NSAIDs or tramadol) and a psychotropic or multiple psychotropics with an increased risk of adverse effects. This topic needs to be addressed in a future study.
In conclusion, we showed that a significant number of older Korean patients with OA were prescribed psychotropic drugs without comorbidities for such drugs. Some drugs that have little evidence of effectiveness in OA and an unfavorable safety profile in older adults had increased prescription rates. Fentanyl is one such drug that needs to be avoided at all costs for a chronic ailment like OA (its danger to the community led to the Declaration of National Fentanyl Prevention and Awareness Day in the US). Concerns about the misuse of psychotropic drugs have been raised, and caution is necessary when prescribing such medications.
XML Download