The Opioid Epidemic and Primary Headache Disorders: A Nationwide Population-Based Study

Introduction The opioid epidemic has been linked to several other health problems, but its impact on headache disorders has not been well studied. We performed a population-based study looking at the prevalence of opioid use in headache disorders and its impact on outcomes compared to non-abusers with headaches. Methodology We performed a cross-sectional analysis of the Nationwide Inpatient Sample (years 2008-2014) in adults hospitalized for primary headache disorders (migraine, tension-type headache [TTH], and cluster headache [CH]) using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes. We performed weighted analyses using the chi-square test, Student’s t-test, and Cochran-Armitage trend test. Multivariate survey logistic regression analysis with weighted algorithm modelling was performed to evaluate morbidity, disability, and discharge disposition. Among US hospitalizations during 2013-2014, regression analysis was performed to evaluate the odds of having opioid abuse among headache disorders. Results A total of 5,627,936 headache hospitalizations were present between 2008 and 2014 of which 3,098,542 (55.06%), 113,332 (2.01%), 26,572 (0.47%) were related to migraine, TTH, and CH, respectively. Of these headache hospitalizations, 128,383 (2.28%) patients had abused opioids. There was a significant increase in the prevalence trend of opioid abuse among patients with headache disorders from 2008 to 2014. The prevalence of migraine (63.54% vs. 54.86%), TTH (2.29% vs. 2.01%), and CH (0.59% vs. 0.47%) was also higher among opioid abusers than non-abusers (p<0.0001). Opioid abusers with headaches were more likely to be younger (43 years old vs. 50 years old), men (30.17% vs. 24.78%), white (80.83% vs. 73.29%), Medicaid recipients (30.15% vs. 17.03%), and emergency admissions (85.4% vs. 78.51%) as compared to opioid non-abusers with headaches (p<0.0001). Opioid abusers with headaches had higher prevalence and odds of morbidity (4.06% vs. 3.70%; adjusted odds ratio [aOR]: 1.48; 95% CI: 1.39-1.59), severe disability (28.14% vs. 22.43%; aOR: 1.58; 95% CI: 1.53-1.63), and discharge to non-home location (17.13% vs. 18.41%; aOR: 1.35; 95% CI: 1.30-1.40) as compared to non-abusers. US hospitalizations in years 2013-2014 showed the migraine (OR: 1.61; 95% CI: 1.57-1.66), TTH (OR: 1.43; 95% CI: 1.22-1.66), and CH (OR: 1.34; 95% CI: 1.01-1.78) were linked with opioid abuse. Conclusion Through this study, we found that the prevalence of migraine, TTH, and CH was higher in opioid abusers than non-abusers. Opioid abusers with primary headache disorders had higher odds of morbidity, severe disability, and discharge to non-home location as compared to non-abusers.


Introduction
Headache disorder is one of the leading conditions for emergency department visit, accounting for 0.5% to 2.8% of all visits [1,2]. Headaches are also the third highest cause worldwide of years lost due to disability [3]. Migraine, tension-type headache (TTH), and cluster headache (CH) are the most common types of primary headache. Prolonged personal suffering, impaired quality of life, and economic burden are commonly associated with chronic headache disorders [4].
Opioids are often prescribed more than any other non-steroidal anti-inflammatory drugs as a primary level of therapy for headache [5]. Prolonged use of an opioid in patients with headache disorders, such as migraine, carries a high risk of medication overuse headache. According to the American Headache Society, prolonged and overuse of opioids (more than 10 times per month) can lead to medication overuse headaches and cause occasional migraines to transition to chronic migraine [6]. The opioid epidemic has affected a number of other disorders; however, its impact on headache disorders has not been well studied. The relationship is especially important as opioids are an important treatment modality for headache disorders.
Hence, we conducted this study to determine the trend of opioid abuse among patients with primary headache disorders, evaluate whether opioid-dependent and non-dependent opioid abuse associated with headache disorders, and its relationship with outcomes like morbidity, disability, and discharge disposition.

Materials And Methods
Data was obtained from the Agency for Healthcare Research and Quality's Healthcare Cost and Utilization Project (HCUP) Nationwide Inpatient Sample (NIS) between January 2008 and December 2014. The NIS is the largest publicly available all-payer inpatient care database in the United States and contains discharge-level data provided by states that participate in the HCUP (including a total of 46 in 2011). This administrative dataset contains data on approximately eight million hospitalizations in 1,000 hospitals that were chosen to approximate a 20% stratified sample of all US community hospitals, representing more than 95% of the national population. Criteria used for stratified sampling of hospitals into the NIS include hospital ownership, patient volume, teaching status, urban or rural location, and geographic region. Discharge weights are provided for each patient discharge record, which allows extrapolation to obtain national estimates. Each hospitalization is treated as an individual entry in the database and is coded with one principal diagnosis, up to 24 secondary diagnoses, and 15 procedural diagnoses associated with that stay. Detailed information on NIS is available at http://www.hcup-us.ahrq.gov/db/nation/nis/nisdde.jsp. The NIS is a de-identified database, so informed consent or Institutional Review Board approval was not needed for the study. The HCUP Data Use Agreement (HCUP-348L73IZS) for the data utilized in this study was obtained.

Study population
We used the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes to identify adult patients admitted to hospital with a primary diagnosis of migraine, TTH, and CH (ICD-9-CM code Migraine: 346, TTH: 339.1 or 307.81, and CH: 339.0). Similarly, patients with opioid dependence and non-dependent opioid abuse were identified using ICD-9-CM code 304.0 and 305.5, respectively. Age <18 years and admissions with missing data for age, gender, and race were excluded. The sample size was based on the available data. Data from NIS has previously been used to identify and analyze the trends, outcomes, healthcare costs, and disparities of care [7,8]. We have not considered available NIS data from the years 2015 and 2016 due to the lack of literature showing the validity of ICD-10 for identifying headache disorders.

Patient and hospital characteristics
Patient characteristics of interest were age, gender, race, insurance status, and concomitant diagnoses as defined above. Race was defined by white (referent), African American, Hispanic, Asian or Pacific Islander, and Native American. Insurance status was defined by Medicare (referent), Medicaid, private insurance, and other/self-pay/no charge. We defined the severity of comorbid conditions using Deyo's modification of the Charlson comorbidity index (CCI) ( Table 1).

The outcomes
The primary outcome of interest was to determine if opioid abuse among patients hospitalized for migraine, TTH, or CH during 2008-2014 was associated with differences in morbidity, disability, or discharge disposition. Morbidity was defined as patients transferred to a shortterm hospital (STH), or skilled nursing facility (SNF), or intermediate care facility (ICF) and a hospital stay of more than eight days (>90th percentile of mean headache hospitalizations). The comparison of disability/loss of function was investigated by All Patient Refined Diagnosis Related Group (APR-DRG) severity between patients with opioid abuse and patients without opioid abuse. APR-DRGs were assigned using software developed by 3M Health Information Systems (Salt Lake City, UT), where score 0 indicates no loss of function, 1 indicates minor, 2 moderate, 3 major, and 4 indicates extreme loss of function. Detailed information on APR-DRGs is available at https://hcup-us.ahrq.gov/db/vars/aprdrg_severity/nisnote.jsp. Our secondary outcome of interest was to evaluate whether opioid-dependent and non-dependent opioid abuse was associated with headache disorders among patients hospitalized in the US between January 2013 and December 2014. The reason to choose the year 2013-2014 data for secondary outcome was a large number of US hospitalizations (more than 20 million) each year to evaluate patients with and without opioid abuse and headache disorders.

Statistical analysis
All statistical analyses were performed using the weighted survey methods in SAS version 9.4 (SAS Institute Inc., Cary, NC). Weighted values of patient-level observations were generated to produce a nationally representative estimate of the entire US population of hospitalized patients. A p-value of <0.05 was considered significant. Univariate analysis of differences between categorical variables was tested using the chi-square test, and analysis of differences between a continuous variable (age) was tested using paired Student's t-test. Mixed-effects survey logistic regression models with weighted analysis were used for categorical dependent variables to estimate the odds ratio (OR) and 95% confidence intervals for the association between opioid use and outcomes of interest among headache disorders from January 2008 to December 2014 and for the linkage between opioid use and headache disorders from January 2013 to December 2014.
We adjusted models with demographics (age, gender, race), patient-level hospitalization variables (admission day, primary payer, admission type, median household income category), hospital-level variables (hospital region, teaching versus non-teaching hospital, hospital bedsize), and CCI.
For each model, the c-index (a measure of goodness of fit for binary outcomes in a logistic regression model) was calculated. All statistical tests used were two-sided, and p<0.05 was deemed statistically significant. No statistical power calculation was conducted prior to the study.

Data availability
The data that supports the findings of this study is publicly available from the Agency for Healthcare Research and Quality's HCUP-NIS. A raw analysis of the data will be however made available from the authors upon request and with permission from HCUP-NIS.

Trends and prevalence
We analyzed trends of opioid abuse in total headache hospitalizations as well as in hospitalizations due to migraine, TTH, and CH. As shown in Figure 1

Demographics, patient and hospital characteristics, and comorbidities
Opioid abuse was more common in the 18-49 years of age group (p<0.0001

Regression model derivation
We performed the multivariable survey logistic regression models to predict the outcomes of opioid abusers (morbidity, disability, and discharge disposition) among patients with headache disorders after adjusting for basic demographic characteristics with patient and hospital-level variables, and CCI (

Discussion
Headache is the most common neurological disorder presenting to primary care, accounting for 3% of all visits [9]. In 2007, Stovner et al. reported a global prevalence of 46% for headache in general, 11% for migraine, 42% for tension-type headache, and 3% for chronic daily headache [10]. In 2015, the Global Burden of Disease estimated that headache disorders including their subtypes -migraine, TTH, and CH -were the third cause of disability in people less than 50 years of age [11]. Opioid abuse is of particular concern in headache disorders given its role as a treatment modality. Over the last two decades, the opioid epidemic has led to enormous health concerns and it has been triggered to a large extent by prescription opioids: an 80% increase in opioid analgesic prescriptions from the year 2000 through 2010 with incremental use from 7.4% to 11.8% [12]. Similarly, in our study, we found that among the adult American population, opioid use increased from 1.74% in 2008 to 2.71% in 2014.
In recent years, there has been a tremendous increase in opioid prescriptions for acute and chronic pain and this trend is also seen in patients with headache disorders [13]. The American Migraine Prevalence and Prevention Study (AMPP) involving 6008 migraine patients reported that 16.6% of patients met Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria for opioid dependence [14]. A study by Choong et al. reported that the most common medicine prescribed for cluster headache is opioids [15]. Our study found that among 5,627,936 headache hospitalizations, 128,383 (2.28%) were opioid abusers. We also found a significant increasing trend of opioid abuse (years 2008-2014) among headache hospitalizations (1.74% in 2008 to 2.71% in 2014; p-trend<0.0001). This recent trend in overprescribing opioids for various headache disorders despite the lack of strong evidence showing the efficacy of opioid treatment has led to serious consequences both affecting an individual's quality of life and increased burden on society and healthcare that must now be addressed conscientiously.
Opioid use can also lead to significantly higher disability in patients admitted for headaches. Our study also found 1.48 times higher odds of morbidity and 1.58 times higher odds of major/severe disability among opioid abusers compared to opioid non-abusers (p<0.0001). This increased disability can be due to chronic opioid therapy in treating chronic migraine and as prophylaxis for refractory headache. Considering the risks including disability, low quality of life, and higher healthcare costs associated with opioid abuse and lack of opioid efficacy data in the treatment of migraine, it is crucial for providers to evaluate potential benefits of alternate treatment options. Several guidelines have been published for the safe use of opioid medications in the treatment of chronic pain that emphasize screening prior to prescribing opioids [16,17].
To prevent opioid abuse and its associated adverse effects, it is crucial that the patient's medication regimen is reviewed at each visit as well as screening for impaired cognition, use of other illicit or prescription drugs, and concurrent mental illness has been done as all of these factors may increase the risk of opioid overdose [18]. Among patients treated with chronic opioids, a urine toxicology screen should be done at the first clinic visit and then annually to assess for potential polysubstance abuse [19].

Strength and limitations
To our knowledge, this is the first large population-based nationwide study to report prevalence, outcomes, and linkage between opioid abuse and headache disorders. One of the limitations of this study being an observational study is that we cannot prove causation of the temporal association of opioids for headache disorders. Also, our assessment is limited to hospitalized patients with headaches and may not reflect the severity of the issue in outpatients. Long-term outcomes are also not available through this study. In spite of these limitations, we have a very large number of patients in the study, which is possible through the use of a nationwide database such as NIS. The APR-DRG coding system used in this study to assess the severity of illness is an external validated reliable method with accurate and consistent results and is widely used by hospitals [20,21]. A population-wide study with a large number of subjects is ideally suited to understand the impact of the opioid epidemic on headache disorders.

Conclusions
We have found that opioid abusers were associated with the higher prevalence of migraine, TTH, and CH and also had higher odds of morbidity, severe disability, and non-home discharge as compared to non-abusers. The patients with these primary headache disorders were having higher odds of exposure to opioid abuse than patients without these headache disorders. A careful selection of patients for opioid prescription and refill, counselling for recreational use, and identification of such patients might mitigate the risk of opioid abuse-associated poor outcomes among patients with headache disorders.

Additional Information Disclosures
Human subjects: All authors have confirmed that this study did not involve human participants or tissue. Animal subjects: All authors have confirmed that this study did not involve animal subjects or tissue. Conflicts of interest: In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.