Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Commentary on: OpenUrlAbstract/FREE Full Text
Benzodiazepines are used with relatively high frequency and in potentially concerning ways among individuals with chronic obstructive pulmonary disease (COPD).1 The scope and pattern of opioid use in COPD is less well-known. While studies evaluating benzodiazepines and opioids for refractory dyspnoea in advanced COPD are available, none were specifically designed to examine for potential drug-related side effects, as they involved small numbers of participants, low-drug doses and relatively short follow-up durations.2 The purpose of this study was to prospectively evaluate all-cause hospitalisation and mortality in relation to use and non-use of benzodiazepines and opioids among individuals with COPD requiring supplemental oxygen.
This was a Swedish population-based, prospective cohort study using health administrative and clinical data. Individuals with doctor-diagnosed COPD newly receiving supplemental oxygen were identified through health administrative data. The day of supplemental oxygen receipt was the index date. Prevalent outpatient benzodiazepine and opioid dispensation within 3 months preindex was identified using health administrative data. Drug use was categorised as ‘lower’ or ‘higher’ dose, with ‘lower’ dose defined as ≤0.3 defined daily doses per day (ie, ≤3 mg of oral diazepam equivalents per day; ≤30 mg oral morphine equivalents per day). Outcomes included first all-cause hospitalisation and all-cause mortality. Hospitalisation and mortality outcomes were expressed as HRs using Fine-Grey and Cox regression analysis, respectively. HRs were adjusted for a broad range of relevant covariates at index including, but not limited to, WHO performance status, forced expiratory volume in 1 s, arterial blood gas oxygen and carbon dioxide tensions, smoking status, number of previous hospitalisations and comorbidities.
Neither benzodiazepines (HR=0.98, 95% CI 0.87 to 1.10) nor opioids (HR=0.98, 95% CI 0.86 to 1.10) were associated with significantly increased adjusted all-cause hospitalisation, even after stratifying by dose. Benzodiazepines were associated with significantly increased adjusted all-cause mortality (HR=1.21, 95% CI 1.05 to 1.39), especially the ‘higher’ dose subgroup (‘higher’ dose HR=1.23, 95% CI 1.02 to 1.48 vs ‘lower’ dose HR=1.18, 95% CI 0.99 to 1.40). The adjusted HR for mortality for all opioid users was not reported, but ‘higher’ opioid dose was associated with significantly increased adjusted all-cause mortality (HR=1.21, 95% CI 1.02 to 1.44), whereas ‘lower’ dose was not (HR=1.03, 95% CI 0.84 to 1.26).
After adjusting for confounders, there was significantly increased all-cause mortality risk with ‘higher’ benzodiazepine and opioid doses among individuals with COPD receiving oxygen therapy. Although increased mortality was observed only among ‘higher’ dose medication users, it should be noted that what the authors define as ‘higher’ dose is really not all that high (ie, >3 mg of oral diazepam equivalents per day and >30 mg oral morphine equivalents per day). Many individuals with oxygen-dependent COPD receiving these medications are likely receiving higher doses than these; in fact the authors’ own data support this speculation, with ≥50% of study patients receiving the ‘higher’ drug doses.
Although ‘lower’ benzodiazepine and opioid doses were not associated with increased mortality, the accuracy of the approach used to quantify drug dose is somewhat questionable. Drug dose was based on the mean dispensed doses per day during the preindex period. However, benzodiazepines and opioids can be used by patients on an as-needed, rather than a regular basis, and drug dose can be lowered over time in response to other adverse events. Therefore, a patient categorised as receiving ‘higher’ dose based on the mean daily dispensed dose preindex, may have in fact been taking a ‘lower’ dose, suffered an adverse event and been misclassified.
The authors focus much of their discussion on the safety of benzodiazepines and opioids from a respiratory perspective, yet it should be emphasised that neither of the study outcomes were respiratory specific. A recent Ontario population-based retrospective cohort study of older adults with COPD that included respiratory-specific outcomes found that incident benzodiazepine use was associated with significantly increased risks of outpatient respiratory exacerbations and emergency room visits for COPD or pneumonia.3 Other relevant and clinically important drug-related side effects were also not examined, including cognitive, psychomotor and gastrointestinal side effects.
Although significant associations between benzodiazepine and opioid drug receipt and mortality were found, causal links cannot be inferred in observational-type studies such as this, and confounding by indication is still possible, despite the impressive list of confounders controlled for. Given the significantly increased mortality risk associated with benzodiazepine and opioid receipt, physicians should exercise caution and monitor for side effects when prescribing these drugs to vulnerable advanced patients with COPD.
Competing interests None.