A recent study investigated the long-term continuation of opioid therapy among previously opioid-naïve patients following ICU treatment. Using a retrospective, nationwide cohort drawn from South Korea’s National Health Insurance Service (NHIS) database, the authors aimed to estimate the six-month incidence of new persistent opioid use and identify associated demographic, clinical, and socioeconomic predictors.
The study included adults admitted to any ICU in South Korea between January 2020 and December 2022. Eligible participants survived to hospital discharge and remained alive for at least six months thereafter, with no opioid prescriptions in the year preceding admission. New persistent opioid use was defined as having at least one outpatient opioid prescription within 90 days after discharge and another between 91 and 180 days post-discharge. Opioids were categorised as less-potent (e.g. tramadol, dihydrocodeine) or potent (e.g. morphine, fentanyl, oxycodone, hydromorphone, methadone).
From a total of 1,245,251 ICU admissions, 567,260 opioid-naïve ICU survivors met the inclusion criteria. Of these, 23,945 patients (4.2%) developed new persistent opioid use within six months of discharge. The majority, 22,643 patients (4.0%), were prescribed less-potent opioids, while 1,643 (0.3%) received potent opioids.
Several demographic and clinical characteristics were significantly associated with persistent use. Older age slightly increased the risk, as did female sex. Patients in lower socioeconomic groups were more vulnerable, particularly those enrolled in the Medical Aid programme, who had a 30% higher risk compared with the lowest income quartile. Clinically, malignancy and metastatic tumours were important predictors, alongside invasive ICU treatments such as extracorporeal membrane oxygenation (ECMO) and continuous renal replacement therapy (CRRT). The most powerful predictor by far was an early opioid prescription within 30 days after discharge, which increased the likelihood of persistent use almost twenty-fold. COVID-19 diagnosis and ICU readmission were inversely associated with persistent opioid use, possibly reflecting closer medical supervision or limited outpatient prescribing opportunities in these patients.
Sociodemographic and treatment-related disparities emerged as key contributors. Women, older adults, and economically disadvantaged individuals were more likely to continue opioid therapy, echoing trends seen in surgical cohorts. These findings suggest that socioeconomic inequities may influence access to non-opioid alternatives, pain management literacy, and follow-up care. Clinically, the association of ECMO and CRRT with persistent use underscores the lasting sequelae of critical illness and the high analgesic burden associated with such invasive treatments. The authors note that ECMO survivors often experience chronic pain and neuromuscular weakness, potentially increasing long-term opioid dependence.
Mechanical ventilation and respiratory failure were inversely related to persistent opioid use. The authors propose that many ventilated patients underwent short-term postoperative care with structured tapering, while those with recurrent complications may have been subject to stricter prescribing oversight. This contrasts with earlier assumptions that mechanical ventilation predicts prolonged use.
The research provides valuable national-level insight into post-ICU opioid use. The results underscore that approximately one in twenty-five ICU survivors becomes a new persistent opioid user, and that early post-discharge prescribing is the critical transition point. The authors advocate for targeted stewardship strategies, including early tapering, multimodal non-opioid analgesia, patient education, and enhanced prescription monitoring. They also recommend tailoring interventions to opioid potency and patient subgroups, particularly those with malignancy or high-intensity ICU treatments, to balance effective pain management with the prevention of long-term dependence.
Source: Critical Care
Image Credit: iStock
