Prescribing sleep, provoking harm? rethinking Z-drugs in heart failure

This editorial refers to ‘Z-drugs are associated with cardiovascular outcomes among patients with heart failure: a population-based cohort study in Hong Kong’, by H.-Y. Yang et al. https://doi.org/10.1093/eurjpc/zwaf128. Despite therapeutic improvement over the last three decades, overall heart failure (HF) prevalence, regardless of ejection fraction, is increasing worldwide, and is associated with adverse outcomes such as reduced quality of life, unscheduled hospitalizations, and mortality.1 Indeed, the one-year HF hospitalization and five-year mortality rates are each around 50%. Non-cardiovascular comorbidities become more and more prevalent, especially among patients with HF and preserved ejection fraction and/or in elderly patients.1 Moreover, although cardiovascular death remains predominant in HF, there are growing evidence that we are facing a shift towards non-cardiovascular death over the last two decades. Among others, mental health and sleep disorders such as depression and insomnia incidence increased significantly.2 In chronic HF and after an episode of acute HF, insomnia is associated with a higher risk of mortality.2,3 Furthermore, insomnia symptoms themselves are associated with incident HF.4 Even if trials specifically designed to assess the effect of anti-depressor agents such as selective serotonin reuptake inhibitors in patients with HF and depression have failed to reach their primary endpoint,5,6 there are no existing trials and a lack of data in the context of insomnia. In this issue of the European Journal of Preventive Cardiology, Yang et al.7 showed, in a retrospective cohort study including 202 585 adults with HF (regardless of ejection fraction) from Hong Kong, that (i) the use of non-benzodiazepine GABA receptor agonists (Z-drugs), hypnotics commonly used to treat insomnia, was independently associated with a 21% higher risk of HF hospitalization or cardiovascular death, and (ii) the relationship between Z-drugs use and the risk of unscheduled hospitalization for HF and cardiovascular death was duration-dependent since patients with a longer-term Z-drugs use (>180 days) had worse outcome (risk of hospitalization for HF or cardiovascular death increased by 24%) compared to those with a shorter use. These results were consistent when HF hospitalization and cardiovascular death were considered separately. The deleterious association between benzodiazepine and HF was already documented through several studies but data regarding Z-drug prescription in the field of HF remained sparse. In a previous study,8 Sato et al. reported that, compared to Z-drug prescriptions in patient with HF, benzodiazepine prescription was independently associated with rehospitalizations for HF (HR = 1.53, 95% CI [1.03–2.28], P = 0.038), even after propensity match score. Thus, and as reported by the authors, Z-drugs, especially zolpidem and zopiclone, are often perceived (and thus prescribed) by physicians to have a favourable safety profile compared to benzodiazepines and many countries have seen a rise in the prescription of Z-drugs.9,10 Yang et al. managed to analyse in detail the safety profile of these drugs in the HF population and raised several issues, which could have important implications on clinical practice. Indeed, they showed that Z-drug prescription was not only independently associated with HF hospitalizations and cardiovascular death but also with a higher risk of HF rehospitalizations (HR = 1.21, 95% CI [1.17–1.24]), a higher risk of cardiovascular diseases (HR = 1.11, 95% CI [1.01–1.15]), and a 15% increased risk of all-cause mortality (HR = 1.15, 95% CI [1.11–1.18]). This is in line with a previous study from Saito et al.11 in Japan, showing that both benzodiazepines and Z-drugs prescriptions were independently associated with a higher risk of HF-related hospitalizations. These findings have several implications: (i) healthcare professionals (including general practitioners) and patients with HF must be aware of the consequences of such a prescription, and patient education is of importance, and (ii) pharmaceutical companies should extend the post market clinical follow-up and provide appropriate warning in the short product characteristics. Another important result raised by this study is the duration-dependent relationship that exists between Z-drugs prescription and the risk of the composite endpoint including hospitalization and cardiovascular death. Patients with a prescription of Z-drug for more than 30 days were at higher risk of HF hospitalization and cardiovascular disease but also all-cause mortality and the longer the prescription, the higher the risk. As far as we know, this is the first study that emphasized this relationship in patients with HF. Similarly, but in a population of patients with recent myocardial infarction, Liu et al.12 showed a dose-dependent relationship between benzodiazepines and Z-drugs prescriptions with all-cause and cardiovascular death. Once again, this study underlines that (i) if a prescription for Z-drug is necessary for particularly disabling insomnia, it should be for as short as possible a period, ideally <30 days, and (ii) caring physicians (and nurses) should regularly reassess all medical therapy, not just HF-specific therapies, at each consultation. Despite the important observations and conclusions by Yang et al., several aspects require further considerations. First, as mentioned by the authors in the method section, the documentation of HF as the cause of first hospitalization was based on ICD-9 and ICD-10 code, which introduced a selection bias. There was a relative low prescription rate of beta-blockers, angiotensin-converting enzyme inhibitor/angiotensin receptor blockers, and diuretics provided in Table 1 (there were no data regarding mineralocorticoid receptor antagonists, angiotensin receptor–neprilysin inhibitor, and sodium–glucose co-transporter 2 inhibitors), questioning the optimal therapy of HF patients. Secondly, it would have been of interest to investigate if this relationship remained true for the whole spectrum of HF (i.e. HF with preserved vs. mildly reduced vs. reduced ejection fraction). Prospective studies, ideally international registries, are still needed to explore the impact of these treatments as well as randomized control trials to search for the best therapeutic option to treat mental issues such as insomnia in this context. Thirdly, the authors investigated only Z-drugs in this study. Yet, anti-depressants and/or neuroleptics are often prescribed in patients with HF and mental health disorders. The potential impact on sleep quality and insomnia symptoms may also interfere with the aforementioned results. Fourthly, sleep disorders per se affect the prognosis of patients with HF. That is why HF patients with insomnia already have worse prognosis compared to those without sleep disorders.2,3 However, this study does not allow us to say whether Z-drugs are the culprits or mere innocent witnesses of this aggravation. Consequently, the logical questions raised by this study are: do Z-drugs prescribed for sleep disorders in patients with HF further worsen their already worse prognosis? Is it possible that Z-drugs prescriptions do not affect the already worse prognosis in HF patients with insomnia? Could they not even improve the worse outcome in HF insomnia patients targeting the sleep quality and quantity (maybe not to the extent of the HF population without sleep disorders, but still an improvement compared to HF insomnia patients that are not treated for insomnia)? Correct comparisons in pursue of the answers to these questions would be: (i) HF patients with sleep disorders treated with Z-drugs vs. HF patients with sleep disorders that are not treated; and (ii) HF patients with sleep disorders treated with Z-drugs with improved sleep quality and quantity vs. HF patients with sleep disorders treated with Z-drugs without improvement in sleep quality and quantity. Fifthly, the authors do not provide data regarding the progression in time of insomnia symptoms after Z-drugs prescription: were the symptoms improved and in what proportion? Do these patients (with improvement of insomnia) fare better, or worse, or equal to patients with Z-drugs prescription and persistence of symptoms? These important questions will need to be addressed in the future. Finally, the mechanisms explaining the relationship between Z-drugs prescription and adverse outcomes remains elusive and requires further studies In conclusion, this study reinforces the fact that HF management is complex and continuously evolving, mainly based on guidelines-directed medical therapy and treatment of comorbidities, including sleep disorders. Mental health, like depression and insomnia for instance, is certainly an under recognized but important risk factor contributing to the high morbidity and mortality rate in these patients. There are many uncertainties regarding the therapy of sleep disorders and insomnia in the HF population that further complicate the therapeutic strategy.