Review of Neurologic Comorbidities in Hospitalized Patients With Opioid Abuse

Abstract

Purpose of Review

To determine the prevalence and burden of neurologic comorbidities in hospitalized patients with opioid abuse.

Recent Findings

From 1 year of hospital discharges, 2,182 patients with opioid abuse were identified (prevalence 6.3%), with abuse greater among younger patients (p < 0.0001), women (p < 0.0001), Whites (p < 0.0001), and urban population (p = 0.028). Matching for age, sex, race, and urban-rural residence, 347 patients were reviewed, and 179 (52%) had a neurologic comorbidity. The comorbidities frequently overlapped and included encephalopathy (130), neuromuscular disorders (42), seizures (23), spine disorders (23), strokes (20), CNS infections (3), and movement disorders (2). Abuse patients with neurologic comorbidities experienced substantially greater number of hospital and intensive care unit days and mortality, independent of overdose.

Summary

Neurologic comorbidities are a frequent and heretofore underappreciated contributor to the disease burden of hospitalized patients with opioid abuse. The importance of neurologic comorbidities should be included in the public health discussions surrounding the opioid epidemic.

Central Appalachia has one of the highest opioid abuse rates in the United States.¹ The region draws from counties within Southwest West Virginia, Southwest Virginia, Southwest Ohio, Eastern Tennessee and Western North Carolina, and Eastern Kentucky. High rates of opioid abuse are rooted in several socioeconomic factors, among them a legacy of geographic isolation and poverty. Currently, important contributions include a reliance on coal mining with a high rate of injury and the financial hardship prompted by recent collapse of the coal industry. Consequently, Kentucky comprising a large portion of Central Appalachia stands among the top 5 states in overdose deaths,² despite ranking 26th in the US population.

Although ravages born of the nationwide opioid epidemic are well chronicled, little is known about the neurologic comorbidities of this epidemic. Most reports of neurologic conditions with opioid abuse consist of small series or case reports describing myelopathy,^3,4 unusual infections,^5,6 muscle and nerve infarction,^7,8 pyomyositis,^9,10 and other afflictions. Rhabdomyolysis in heroin users has been the subject of earlier series.^11-13 Some neurologic conditions occur as regional sporadic outbreaks such as spongioform leukoencephalopathy,¹⁴ botulinism,^15-18 or parkinsonism.¹⁹ Common neurologic disorders such as stroke²⁰ and seizures²¹ are infrequently associated with opioid abuse in large surveys. Other conditions reflect sequelae such as peripheral nerve compression with coma²² or the consequences of HIV infection.

The literature lacks an analysis of neurologic conditions from a large cohort of opioid abusers during the current era of opioid abuse.

This study investigated the prevalence and clinical impact of opioid abuse in 1 year of hospital discharges from a referral center serving central Appalachia. In addition, the hospital records of a representative set of patients with opioid abuse underwent review with the aim of categorizing the neurologic comorbidity and determining the impact on mortality and length of hospitalization.

Methods

We analyzed University of Kentucky Healthcare Enterprise discharge data from October 1, 2016, to September 30, 2017, for 34,414 patients aged ≥18 years. The University of Kentucky medical center is a regional quaternary care center serving Lexington and surrounding central and eastern Kentucky encompassing rural central Appalachia. The medical center includes medical-surgical, obstetrics, and behavioral health units in 2 hospitals on the University of Kentucky campus. Patients with opioid abuse were identified by from electronic claims coding abstraction using International Classification of Diseases, Tenth Revision: F11. Mental and behavioral disorders due to use of opioids. The University of Kentucky Institutional Review Board approved this study.

Patients with opioid abuse were compared with both the general hospital population and a control group of hospitalized patients (1:3) matched for age, sex, race, and residence (Figure). From this pool of opioid abuse patients, 360 patients were randomly selected for chart review from the entire opioid abuse cohort matching for age, sex, race, and urban-rural residence. Using a standardized instrument captured in REDcap database, 4 investigators each reviewed the medical records of 90 randomly assigned patients.

Figure. Diagram of Study Design.

Blue boxes: sets of hospital discharges, yellow boxes: sets of patients whose medical record underwent expert review, numbers: set size, black arrows: derivation of subsets, green arrows: derivation of subsets matched to pool of opioid abuse, and red lines: subsets of hospital discharges or patients compared.

Opioid abuse was diagnosed if opioids were used without indication or above prescribed dosages. Data were aggregated for opioid patients with multiple hospitalizations in the study year. Urban residence was considered Lexington and Fayette County, a merged city/county entity. Rural residence included surrounding central and eastern Kentucky counties. Mortality and number of admissions were considered only for the University of Kentucky hospitals during the study year. Neurologic comorbidities were identified by medical record review and then determined if the comorbidity received active diagnostic or therapeutic consideration during the hospitalization. Encephalopathy was considered multifactorial in origin if 2 or more contributors were identified. Rhabdomyositis was noted if creatine kinase exceeded 10 times normal. Designation of stroke and spinal infection required confirmation by MRI or CT imaging. No statistically significant interrater disagreement was identified.

Statistical analysis was performed using SAS 9.4 (Cary, NC). Significant differences in Tables 1–3 and the study population were tested using chi-square or t test analysis based on different type data (i.e., continuous data using the t test and categorical data using the χ² test).

Table 1.

Demographics and Clinical Measures of Hospitalized Patients Without and With Opioid Abuse

Table 2.

Demographics and Clinical Measures of 347 Hospitalized Opioid Abuse Patients With and Without Neurologic Comorbidity

Table 3.

Neurologic Comorbidities in 179 Hospitalized Opioid Abuse Patients, 152 With Active Comorbidities Requiring Management

Data Availability

Aggregated deidentified data can be shared on request. Individual electronic medical record data cannot be deidentified and therefore are not subject to sharing for legal and ethical reasons.

Results

For the year period, 34,414 patients were discharged, and 2,182 (6.3%) with a diagnosis of opioid abuse identified. Those with opioid abuse were younger, more often female, overwhelmingly White, and slightly more likely to be urban (Table 1). Of note, 32% of opioid abusers were 46 years or older. Although patients with opioid abuse experienced greater length of stays (LOSs), there was no difference in the mean number of intensive care unit (ICU) days used, yet mortality was greater in those without opioid abuse. Patients with opioid abuse were more often discharged against medical advice and experience 2 or more admissions in the study year.

Comparing patients with opioid abuse with matched controls, the LOS for opioid abuse was longer (Table 1). The mean ICU days and mortality were no different. Compared with matched controls, those with opioid abuse were more likely discharged against medical advice and be readmitted within the study year.

Of the 360 patients selected for review, 347 (96%) had clinical documentation of opioid abuse. Neurologic comorbidity was present in 179 (52%) (Table 2). Those with a neurologic comorbidity were slightly older and more likely male, with no difference in race or residence. Mean LOS was 73% greater for those with neurologic comorbidities. Patients with neurologic comorbidities also had 5 times the mean number of ICU days and 9 times the mortality rate. Those with neurologic comorbidities were less likely to be discharged alive, more likely to elope against advice or expire, and experience readmission in the study year.

A total of 243 comorbidities (Table 3) were identified in the 179 patients and frequently comorbidities overlapped (1.4 comorbidity per subject). In 152 patients, 198 active comorbidities were found. An active neurologic comorbidity directly contributed to the admission decision in 139 patients.

Encephalopathy was the most prevalent active comorbidity with 114 total and of multifactorial origin 68% of the time. All the 18 deaths in patients with neurologic comorbidities occurred with encephalopathy that was often multifactorial origin in (13) and accompanied by coma (16).

Neuromuscular disorders were the second most prevalent active comorbidity in 33, yet represented only 3 deaths. Rhabdomyolysis, the most prevalent neuromuscular comorbidity, occurred in 6 patients with coma but more often without coma (10). Nonspinal pyomyositis was present in 12, all abused IV opioids. Pyomyositis involved a limb in 7, pelvis-psoas region in 4, and the trunk in a single instance. The infectious agent was Staphylococcus aureus in 7, Streptococcus pyogenes in 2, and 1 each of vancomycin-resistant Enterococcus faecium (VRE), Streptococcus anginosus, and indeterminate. In the 4 with neuropathy, diabetes was present in 1 and the cause indeterminate in 3.

Seizures, an active comorbidity in 17, were most often generalized (12) but focal in 2. Status epilepticus occurred in 2, including the single death. Seizure type was indeterminate in 1.

Paravertebral infections dominated spinal conditions. The infection could be diffuse or localized, involving paraspinous, vertebral, and intraspinal tissues, alone or in combination. Accordingly, treatment ranged from surgical debridement to antibiotic therapy alone. The infection site was cervical in 1, thoracic in 3, and lumbar in 7. The infectious agent S aureus was confirmed by culture in 6, presumed S aureus in 4 with negative cultures, and Streptococcus parasanguinis in 1. All 11 paravertebral infections were seen in patients with spinal pain and abused IV opioids.

Stroke, the 5th leading neurologic comorbidity was seen with the second highest number of deaths. In 1 patient, the stroke was of small vessel origin and indeterminate in another. More than a single stroke type occurred in 5 patients.

Three CNS infections were noted: Cryptococcus, herpes simplex, and HIV in the single mortality. Restless legs in 2 patients were the only movement disorder found.

Active neurologic comorbidities in patients with and without opioid overdose are also compared in Table 3. Of the 18 deaths, 11 occurred with and 7 without overdose. The mortality rate for overdose was 17% overall and 40% with coma. In the overdose group of 66, as well as encephalopathy, there were 23 neurologic comorbidities spanning each neurologic category except movement disorder. In 152 patients with active neurologic comorbidity, there was no difference in demographics, LOS, ICU days, mortality, or neurologic comorbidity category between patients with or without opioid overdose.

In contrast, nonoverdose patients with neurologic comorbidities (n = 86) were different in demographics and several clinical measures from opioid abuse patients without neurologic comorbidities (n = 168). The neurologic comorbidity subgroup was slightly older (mean 41 vs 36 years, p < 0.01), more often male (65% vs 33%, p = 0.001), with no differences in race or residence. Clinically, nonoverdose patients with neurologic comorbidity experienced longer LOS (mean 12.7 vs 6.7 days, p < 0.0001), more ICU days (mean 2.2 vs 0.4, p = 0.0001), and greater mortality (8.1% vs 0.6%, p < 0.01). Neurologic comorbidities without overdose contributed to the admission decision in 75 of the 254 opioid abusers without overdose.

Discussion

This observational study of patients with opioid abuse drew on a medical center serving a large referral base from central Appalachia. The 6.3% of hospital discharges represented by patients with opioid abuse constitutes a substantial segment of the hospitalized population. Patients with opioid abuse were younger, more often female, White, and slightly more likely of urban residence than the general hospital population. Although opioid abusers dominated younger age categories, opioid abuse is not confined to the young. The portions of abusers vs nonabusers were nearly equal in the ages of 46–60 years (24 vs 29%). The observation that 32% of opioid abuse are 46 years or older bespeaks of the pervasiveness of the opioid epidemic and raises the likelihood of opioid abuse adding to comorbidities unrelated to opioid abuse, further complicating medical care. Yet only 8% of opioid abusers were 60 years or greater. Whether the oldest patients possess greater social or biologic resistance to opioid abuse remains unanswered by this study.

The length of hospital stays for opioid abusers was 37% longer than matched controls. Despite longer hospital stays, opioid patients were also more likely to leave against advice and to experience readmission, compounding the management difficulties often facing clinicians caring for opioid abuse patients, and potentially escalating the costs of care. Patients with opioid abuse used the same number of ICU days and faced mortality similar to controls.

Neurologic comorbidities were found in 52% of hospitalized patients with opioid abuse and medically distinguished an important group of opioid abuse patients. Patients with neurologic comorbidities had longer hospitalizations, a 5-fold greater number of ICU days, and a nine-fold increase in mortality. An active comorbidity was present in 44% and contributed to admission in 40% of those hospitalized with opioid abuse. Observations that emphasize the influence of neurologic comorbidities.

Moreover, patients with neurologic comorbidities were slightly older and more often male than opioid abuse patients without comorbidities. Those with neurologic comorbidities had a greater likelihood of leaving against advice and being readmitted during the study year.

Encephalopathy was by far the most common neurologic comorbidity, occurring in 37% of all hospitalized opioid abuse patients. Patients with or without overdose had a similarly high prevalence of encephalopathy, often multifactorial in origin, and frequently culminating in coma.

Although 6 patients had rhabdomyolysis with coma, the 10 cases of rhabdomyolysis without coma imply a myotoxic factor yet uncovered. This observation reinforces an earlier study that found rhabdomyolysis may arise with abuse following a period of abstinence.²³ The extent of rhabdomyolysis varies but can result in acute renal injury and subsequent hemodialysis.¹² Rhabdomyolysis follows several different opioid formulations including methadone.¹³

Pyomyositis has received little previous attention. The causative organism was most often S aureus in our series and other case reports,⁹ but may include a wide variety of other pathogens. Our cases occurred only in those abusing opioids IV.

The similar prevalence of seizures with or without overdose suggests a limited direct effect of opioid abuse and seizure activity.

The number of spinal infections found is alarming and a growing concern.^24,25 Vigilance for this condition is crucial in the patient with septic opioid abuse who is often encephalopathic because timely recognition can mitigate lasting neurologic sequelae. All patients had vertebral pain, and spinal infection should be considered in the opioid abuse patient with recent onset of back pain regardless of neurologic symptoms or systemic sepsis. In our patients, spinal infection occurred exclusively among IV drug users. The infection could be diffuse or localized, involving paraspinous, vertebral, and intraspinal elements, alone or in combination. Symptomatic intraspinal spinal infection did not always coalesce into abscess, and treatment ranged from surgical debridement to antibiotics alone. Most often, the infection site was lumbar, and S aureus the infectious agent. As in this investigation, others have found most infections located in the lumbar region, and the majority isolated S aureus.²⁴

Strokes in this series of 13 patients were of multiple types and not exclusively embolic from cardiac valvular lesions whose incidence appears to be rising.²⁶ Single cases or small series of stroke have describe ischemic,²⁷ borderzone,²⁸ potentially vasculitic,²⁹ and hemorrhagic²⁹ types. Of interest, 5 patients in this study had more than a single stroke type suggesting a complicated relationship with opioid abuse deserving further scrutiny.

This review found no instances of several neurologic disorders previously observed with drug abuse including leukoencephalopathy, botulism, myelopathy, mycotic aneurysms, and brain abscess.

Opioid overdose failed to clearly define 2 patient populations with neurologic comorbidities. Although opioid overdose accounted for a large portion of neurologic comorbidities, an even greater number of patients with opioid abuse without overdose were identified with a broad spectrum of neurologic comorbidities.

Several observations underscore the importance of neurologic comorbidities independent of overdose. Neurologic comorbidities were present in 34% of all nonoverdose opioid abuse patients reviewed and directly influenced admission in 30%. Excluding overdose patients, mortality, ICU usage, and LOS were all substantially greater in opioid abuse patients with neurologic comorbidities than other opioid abuse patients. These same clinical measures were not different between overdose or nonoverdose opioid abuse patients with neurologic comorbidities. Additional neurologic comorbidities were seen in 30% of overdose patients spanning several neurologic categories over and above encephalopathy.

The considerable mortality with neurologic comorbidities deserves emphasis. The mortality rate for all hospitalized opioid abuse patients with neurologic comorbidity of 10% rose to 17% with overdose and encephalopathy, and to 40% when coma accompanied overdose, and compared with 8% without overdose. A focus on overdose or mortality alone ignores a large segment of nonoverdose opioid abuse patients with neurologic comorbidities with comparable hospital and ICU lengths of stay.

This is not a population-based investigation of opioid abuse in eastern and central Kentucky. Rather, this retrospective review was confined to a single medical center and its referral network. Furthermore, mortality and readmission measures were limited to 1 year and a lone medical center. Nonetheless, observations from this study likely apply to other groups with a high prevalence of opioid abuse.

The frequent and substantial contribution of neurologic comorbidity to the disease burden of hospitalized patients with opioid abuse must be considered in the public health discussions surrounding the opioid epidemic.

Appendix. Authors

Study Funding

No targeted funding reported.

Disclosure

The authors report no disclosures relevant to the manuscript. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.

TAKE-HOME POINTS

• → Thirty-two percent of patients with opioid abuse were aged 46 years or older.

• → Neurologic comorbidities occurred in 52% (179) who had nine-fold greater mortality.

• → The most common comorbidity encephalopathy was typically multifactorial.

• → Sixteen patients had rhabdomyolysis, 10 without coma.

• → Spinal infections occurred in 11, the lumbar region in 7.