Typically a chronic opioid user has developed high levels of tolerance. Importantly, chronic opioid users can also develop tolerance to the life-threatening effects such as suppression of breathing. Physical dependence occurs after chronic opioid use. It manifests itself in the form of readily observed withdrawal symptoms when the drug is not available. For opioids, withdrawal symptoms include nausea diarrhea, vomiting, muscle aches, and can last from days to weeks depending on the type and duration of opioid use.
Both tolerance and physical dependence are temporary, lasting only weeks to months after the drug is withdrawn. Tolerance may be easier to understand in the context of alcohol. Most of us cannot consume the same amount of alcohol as our glory days in college; we have lost tolerance for the drug.
The same is true for physical dependence on opioids; once a person is detoxified, the withdrawal symptoms fade. However, a person who is no longer physically dependent can still be addicted. Addiction occurs in a small percentage of opioid users and can manifest itself long after drug use has stopped, when tolerance is no longer present and physical dependence has subsided.
It is thought to have distinct molecular mechanisms leading to long lasting functional changes in the regions of the brain that control reward, emotion and cognition.
It has a strong genetic component, and adolescents are at higher risk because of their developing brain. Addiction is a chronic illness that can last a lifetime. Philip Seymour Hoffman was reported to have relapsed to opioid abuse after 23 years of abstinence. These include compulsive drug taking, obsessive thinking about the drug and escalation of drug use. Simply put, addicts take drugs despite negative consequences. Coming back to the original question — why are people who re-initiate opiate use after quitting at higher risk of overdose?
It is because they have lost their tolerance for the drug but are still addicts. Addicts will relapse re-initiate drug use ; this is one of the characteristics of addiction. But if the person consumes the same dose of the drug as when he or she last used opiates chronically, a time when that person had high levels of tolerance, there is a serious risk of overdose.
This is because they have lost their tolerance to the respiratory suppressant effects of opiates. The insurance claims data are from a large, nationwide, commercial payer, and the EHR data are derived from a large network of hospitals, clinicals, and other health care organizations throughout the US. The databases are linked by unique identifiers shared across both data sets. In total, OLDW includes more than million individuals. This retrospective study used preexisting, deidentified data; therefore, it did not require institutional review board approval or informed consent per 45 CFR Three types of data were used in this study: insurance claims, structured EHR data, and unstructured clinical notes ie, free-text EHR data.
Key elements captured by insurance claims data include individual level characteristics eg, sex, year of birth, geography , insurance type eg, commercial or Medicare Advantage , enrollment periods, and health care use, including inpatient stays and opioid analgesic prescription fills.
Structured EHR data are collected by clinicians at the point of care and include encounter details, such as contact dates and types eg, telephone, emergency, outpatient , clinical procedures, diagnoses including patient concerns ; clinical observations eg, blood pressures, pulse rates, respiratory rates ; and treatment details, such as prescriptions written and medications administered.
Unstructured clinical notes are data collected in the EHR that have no text entry restrictions. They may include dictated or typed notes from clinicians or automated text generated by the EHR system.
To minimize the risk of identifying individual patients, the full text of clinical notes is not available to OLDW researchers.
Notes data were accessed in a deidentified format with specific terms, sentiments ie, opinions, feelings, or mood , and dates recorded. Three cohorts were used in this study: one created using only claims data, a second using structured EHR data linked to claims, and a third using a subset of the structured EHR data plus unstructured clinical notes linked to claims.
The goal in all 3 cohorts was to select patients initiating use of one of the medications included in our analysis between January 1, , and December 31, We refer to these medications as OTO medications because all are labeled only for use in people who are opioid-tolerant. Inclusion criteria are summarized in Table 1 , with further detail and cohort flow tables available in the eAppendix and eTables in the Supplement.
The primary outcome was opioid tolerance at the beginning of OTO episodes. We defined opioid tolerance based on the labeling requirements for the medications: at least 30 mg oxycodone equivalents on each of 7 days prior to the start of the OTO episode Table 1.
Sensitivity analyses used 3 additional opioid tolerance definitions from Larochelle et al 6 that are less stringent than the primary tolerance definition eAppendix in the Supplement. We also report opioid tolerance rates by dosage strength for transdermal fentanyl. We used claims data to estimate the prevalence of opioid tolerance at the beginning of OTO episodes; we stratified the analysis by OTO medication, insurance type, year, and medication strength transdermal fentanyl only.
To determine whether structured EHR data provided additional evidence of opioid tolerance beyond that available in claims data, we first assessed the subset of patients with both structured EHR and claims data. We selected OTO episodes from the claims data analysis for which we had any evidence of structured EHR use in the prior days.
We then divided the episodes into 4 groups in a 2-by-2 table: evidence of tolerance in claims data yes or no vs evidence of tolerance in structured EHR data yes or no. The goal was to determine the proportion of episodes with no evidence of prior opioid tolerance in claims data but with evidence of prior opioid tolerance in the structured EHR data. These potential cases of opioid tolerance would be missed in a claims-only analysis.
We repeated the analysis comparing evidence of tolerance in claims and structured EHR data in a smaller cohort for whom the initial OTO prescription was recorded in both claims data as a prescription fill and structured EHR data as a written prescription; in these cases, we hypothesized that any prior prescriptions that could demonstrate tolerance might be more likely to be present in the EHR for this cohort because there was a record of a prescribed opioid in the EHR ie, the initial OTO prescription.
To account for a possible lag between when the prescription was written and when it was filled, we allowed a difference of 14 days between the date in claims data and the date in structured EHR data. Finally, we repeated the analysis comparing evidence of tolerance in claims data, structured EHR data, and unstructured notes.
In this group, we required the note with the OTO prescription to match within 7 days of the claims fill date. Unstructured clinical notes were analyzed using natural language processing NLP techniques to summarize information in unstructured EHR fields that might provide evidence of prior opioid tolerance. We present a brief description of the approach here, with more detail in the eAppendix and eTables in the Supplement.
We combined all notes from the 30 days before each OTO episode into a single document per episode. The goal was to identify terms that were specific and important—that is, they appeared in relatively few documents ie, they were specific to those documents but frequently in those documents ie, when they appeared in a document, they represented an important aspect of the document.
Notes were further filtered to remove repetition from cut-and-paste notes and common template language, such as discharge instructions and templated review-of-systems text. A vector-space model was used to classify each document according to the frequency with which terms appeared in each document. Following this step, members of a technical expert panel including R. Topics with no obvious interpretation were dropped from the final model.
The technical expert panel included physicians and researchers with expertise in pain medicine, addiction medicine, critical care, primary care, psychiatry, and data science. We also consulted the technical expert panel to understand potential reasons a physician might prescribe an OTO medication to a patient who is not opioid-tolerant.
Analyses were completed using SAS statistical software version 9. A cohort flow table is provided in eTable 1 in the Supplement. Between and evidence of tolerance increased for patients who received extended-release oxycodone change, The prevalence of prior tolerance for extended-release hydromorphone between , when the drug first appeared in the OLDW database, and did not change significantly change, 8.
Complete details on prevalence of tolerance by year are presented in eFigures 4, 5, and 6 and eTable 7 in the Supplement. Because the evidence of prior tolerance for newly prescribed transdermal fentanyl was particularly low, we calculated tolerance prevalence stratified by initial dosage strength and insurance group. At each dosage strength, patients with commercial insurance were more likely to show prior tolerance than patients with Medicare Advantage, and prior tolerance rates increased with higher dosage strengths Figure ; eTable 24 in the Supplement.
As expected, evidence of prior tolerance was higher with the less stringent definitions and differed by OTO medication type. The goal of this analysis was to identify additional evidence of prior opioid tolerance not captured in claims data. A detailed cohort definition is provided in eTable 2 in the Supplement. Structured EHR data contributed minimal additional evidence of prior tolerance Table 3.
This evidence was contributed by records of opioid analgesic prescriptions written in the EHR data that did not appear as prescription fills in the claims data. These prescriptions may not have been filled, or they may have been filled but paid for with cash or by another source of insurance coverage. When we further limited the sample to OTO episodes identified in claims with a matching OTO prescription within 14 days in the structured EHR data, we again found little additional evidence of tolerance contributed by structured EHR data Table 3.
Only 40 of episodes 4. Finally, to search for evidence of prior tolerance not present in either claims or structured EHR data, we used natural language processing techniques to summarize the clinical notes from the 30 days preceding an OTO episode Table 1.
Topics were reviewed by the technical expert panel and given a descriptive name. After reviewing the topics with our technical expert panel, we suspected the topics were unlikely to identify additional evidence of tolerance because most topics were related to specific illnesses and health services unrelated to opioid tolerance eg, laboratory tests, lower extremity ailments.
We asked the technical expert panel to help us understand why a patient who was not opioid-tolerant might receive an OTO product.
From that discussion, we identified additional terms, which we called whitelist terms , that might suggest a reason these patients received an OTO medication Box ; eTables 8, 20, 21, 22, and 23 in the Supplement. Patient using many medications daily or attempt to fit opioid into existing medication regimen eg, morning and evening. Clinician taught that extended-release opioids are better than immediate-release for some patients.
Terms and topics from the unstructured clinical notes were linked to claims data for episodes in which the initial OTO prescription was recorded in the claims data as a prescription fill and in the structured EHR data as a written prescription along with a clinical note within 7 days of the claim.
To determine whether textual analysis provided evidence of opioid tolerance, we looked for topics and terms that seemed to represent opioid tolerance or conversely, prescribing for individuals who were not opioid-tolerant. None of the topics identified by natural language processing analysis directly represented opioid use in individuals who were not opioid-tolerant.
We also used bivariate and multivariate analysis to assess the associations of terms or topics with opioid tolerance at the start of the episode.
The prespecified analytic plan was to use 5-fold cross-validation to solve a lasso logistic regression model. However, the lack of clearly written reasons for prescribing and low term counts made this approach unfruitful. To address the infrequency with which any 1 term was observed, we grouped the whitelist words developed by our technical expert panel to identify possible reasons for prescribing OTO medications to individuals who were not opioid-tolerant.
Adult cancer pain: part 2—the latest guidelines for pain management. Accessed February 10, Analgesic management of acute pain in the opioid-tolerant patient. Curr Opin Anaesthesiol. Prince V. Pain management in patients with substance-use disorders.
Schaible HG, Richter F. Pathophysiology of pain. Langenbecks Arch Surg. Fine PG. Long-term consequences of chronic pain: mounting evidence for pain as a neurological disease and parallels with other chronic disease states. Pain Med. The molecular basis of pain and its clinical implications in rheumatology. Nat Clin Pract Rheumatol. Clinical consequences of inadequate pain relief: barriers to optimal pain management.
Plast Reconstr Surg. Comorbid depression, chronic pain, and disability in primary care. Psychosom Med. Pain symptoms in depression: definition and clinical significance. Clin Med Lond. Pain and depression: an integrative review of neurobiological and psychological factors. Curr Psychiatry Rep. Wilson PR. Clinical practice guideline: acute pain management.
Clin J Pain. Schug SA. Acute pain management in the opioid-tolerant patient. Pain Manag. Management of acute pain in patients on treatment with opioids. Acute pain management in patients with prior opioid consumption: a case-controlled retrospective review. DSM-5 criteria for substance use disorders: recommendations and rationale. Am J Psychiatry. Pain management in the elderly population: a review. Ochsner J. Cavalieri TA.
Management of pain in older adults. J Am Osteopath Assoc. American Geriatrics Society updated Beers Criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc. Increases in drug and opioid-involved overdose deaths—United States, Risk factors for drug dependence among out-patients on opioid therapy in a large US health-care system. Abuse-deterrent opioids—evaluation and labeling. Guidance for industry. Silver Spring, MD: U. Department of Health and Human Services; Acute pain management for patients receiving maintenance methadone or buprenorphine therapy.
Ann Intern Med.
0コメント