Clinical Cardiology最新文献

We read with great interest the article by Erbay et al. [1] evaluating the impact of sodium-glucose co-transporter-2 (SGLT2) inhibitors on sleep quality, anxiety, and quality of life in patients with heart failure (HF). The study contributes meaningfully to the dialog surrounding patient-centered outcomes in HF management. However, several important methodological limitations essential for interpretation were not sufficiently discussed in the manuscript. First, the nonrandomized, physician-directed allocation of SGLT2 inhibitor therapy in this study introduces a substantial potential for selection bias and confounding by indication. Physicians may preferentially prescribe SGLT2 inhibitors to patients who are perceived to be more likely to benefit, be clinically stable, or demonstrate better adherence [2]. Such confounding can extend beyond the measured baseline parameters and affect both clinical and subjective outcomes, including sleep and anxiety. Second, the study groups were imbalanced with respect to key baseline characteristics: the SGLT2 cohort was notably younger and had a significantly higher prevalence of diabetes mellitus (DM). Both age and DM status independently alter sleep architecture, anxiety, and overall quality of life, increasing the risk that these factors, rather than the intervention itself, underpin the observed improvements [3]. In small sample sizes, the capacity to statistically adjust for these complex interrelations is inherently limited. Third, sleep quality assessment in the study relied solely on the Pittsburgh Sleep Quality Index (PSQI), a validated but subjective questionnaire, without incorporating objective sleep measures such as actigraphy or polysomnography. Given the high prevalence of undiagnosed sleep-disordered breathing in HF and the recognized limitations of self-reported measures in this population, the absence of objective evaluation may introduce measurement bias and limit causal inference [4]. Fourth, the study is limited by its short follow-up duration (6 months) and the absence of granular data on changes in concurrent HF and psychotropic medications during follow-up. Short-term improvements may not be sustained over time, and undocumented modifications in concomitant therapy can confound the attribution of the observed benefits exclusively to SGLT2 inhibitors. Furthermore, differential attrition, particularly resulting from hospitalization or mortality, exacerbates the potential for survivor bias [5], a critical consideration in the HF population. Taken together, these limitations underscore the need for cautious interpretation of the reported benefits and reinforce the imperative for future randomized, controlled, and objectively assessed studies with longer follow-ups to establish the patient-centered efficacy of SGLT2 inhibitors in HF.

All authors contributed to the writing of the correspondence and have approved the fina

The recent article by Erbay et al. reporting improved sleep quality, anxiety, and quality of life in heart failure patients receiving SGLT2 inhibitors [1] addresses a clinically important but underexplored area. However, several methodological limitations merit attention before these findings are generalized.

First, there was a significant baseline imbalance in Pittsburgh Sleep Quality Index (PSQI) scores between groups (5.0 vs. 6.0, p = 0.036). This difference, favoring the SGLT2 inhibitor group at baseline, may partially explain the magnitude of improvement observed. While within-group analyses were performed, regression models should have included baseline PSQI as a covariate to mitigate this confounding [2].

Second, the multivariate logistic regression did not adjust for several potential confounders that could influence patient-reported outcomes, including changes in diuretic dosing, concurrent initiation of other guideline-directed medical therapies, and intercurrent hospitalizations. These variables are known to impact congestion, sleep, and mood in heart failure [3, 4].

Third, the subgroup analyses by ejection fraction status are based on small sample sizes, limiting statistical power and precision. Without reporting interaction p-values, the assertion of consistent benefit across EF strata is premature [5].

Fourth, multiple SF-36 domains and other secondary outcomes were tested without correction for multiplicity. In this setting, the risk of false-positive findings is high, particularly in small observational cohorts [6].

Lastly, the study's observational, single-center design precludes definitive causal inference, yet several statements in the discussion imply a treatment effect. This language should be tempered to reflect association rather than causation [7].

Given the increasing integration of SGLT2 inhibitors into heart failure care, it is critical that conclusions about novel patient-reported benefits be supported by rigorous methodology. Randomized controlled trials incorporating objective sleep measures (e.g., polysomnography) and adequate adjustment for confounding are needed to validate these intriguing findings.

Use of AI for paraphrasing and in analyzing the statistical model used in the study.

The authors declare no conflicts of interest.

We welcome the recent analysis of acute myocardial infarction (AMI) death in US adults aged ≥ 65 with documented psychoactive substance abuse disorders [1]. Attention by the authors to an elderly population is well-timed, in light of increased co-occurrence of cardiovascular disease and substance abuse in the elderly. Two important limitations—one clinical, the other methodological—are acknowledged, however, because they affect substantially both validity and interpretability of the results.

Epidemiological studies that use death certificate data for employment are fraught with a very high potential for case ascertainment bias in this particular situation. In elderly populations, AMI is frequently listed on death certificates by clinical history or circumstantial information in place of verifying electrocardiography or cardiac biomarker tests [2]. Challenging presentations, cross-competing comorbidities, and silent myocardial infarction are not uncommon in this group and make both under- and overattribution of AMI as a cause of death more likely [3].

Equally problematic is the determination of psychoactive substance use. Toxicology is not the norm for older deaths unless there is obvious suspicion, and when it is done, it can be a restricted panel only [4]. This results in severe underreporting of drug use. Social stigma, clinician refusal to report, and ignorance regarding how to distinguish from prescribed versus nonprescribed use contribute to the risk of misclassification. When both outcome and exposure are tainted by classification error, cause-specific mortality rate estimates can distort changes in reporting patterns rather than capturing epidemiologic trends. Follow-up analyses using mortality registry data here need to be controlled for this dual-source bias to maintain interpretive validity.

Second, operationalization of exposure—reducing all ICD-10 F10–F19 categories into a single combined “psychoactive substance use disorder” category—is clinically restrictive. All of those codes cannot be collapsed into one usefully homogeneous category from the viewpoint of cardiovascular pathophysiology. Risk of AMI due to alcohol is secondary to chronic remodeling and arrhythmogenesis of the myocardium; cocaine and amphetamines via acute coronary vasospasm and proarrhythmic effects; opioids via hypoxia and bradyarrhythmia; sedatives via induction of hypotension and delay in conduction [5].

By combining these mechanistically different exposures, the study conceals substance-specific patterns of mortality and prevents investigators from being able to determine whether trends are a result of stimulant-induced acute events, alcohol-induced chronic damage, or polypharmacy effects. This limitation reduces the value of the study for cardiologists, addiction specialists, and policymakers who need substance-specific data to implement effective interventions. Further

The authors demonstrated a striking and consistent rise in atrial fibrillation (AF)–related mortality involving obstructive sleep apnea (OSA), particularly among elderly, female, rural, and minority populations [1]. This important analysis highlights the growing public health burden of AF–OSA overlap. However, several issues merit further clarification and discussion.

The authors report a steep increase in AF-related mortality involving OSA over the past two decades [1]. However, as AF prevalence and mortality have also generally increased in the U.S. population [2], it is unclear whether the observed trend is specific to patients with comorbid OSA or merely reflects overall AF epidemiology. A comparative analysis of mortality trends in AF patients without OSA would be helpful to determine the additive risk associated with OSA.

During the study period (1999–2020), catheter ablation became increasingly adopted for rhythm control in AF [3]. Recent studies suggest that AF ablation is associated with improved long-term outcomes, particularly in younger male patients [4]. Yet, despite these advances, AF-related mortality in patients with OSA continued to rise. How do the authors interpret this apparent contradiction? Were these procedures less commonly used or less effective in the OSA subgroup?

The study identifies AF (ICD-10 I48.x) as the “underlying” cause of death and OSA (G47.33) as a “contributing” condition [1]. However, the clinical circumstances in which AF is recorded as the primary cause of death—distinct from its role as a comorbidity in stroke, heart failure, or sudden death—are not well defined. It would be informative to clarify how “AF-related mortality” was operationalized in this study and whether misclassification or variation in coding practices may have influenced the observed trends.

The authors correctly point out the rising burden of AF-related mortality in the presence of OSA [1]. Yet over the same period, mortality from stroke and heart failure—two major downstream complications of AF—has generally declined, partly due to improvements in anticoagulation and HF management [5]. This discrepancy raises the question of whether the increase in AF-related deaths reflects actual clinical deterioration or improved documentation of AF on death certificates.

While the use of CDC WONDER provides valuable national-level insights [1], the reliance on death certificate data introduces several limitations. OSA is often underdiagnosed, particularly among women, minorities, and older adults—groups in whom mortality increases were most pronounced. Additionally, data on OSA severity, AF subtype, comorbidities (e.g., heart failure, chronic kidney disease), and continuous positive airway pressure adherence were not available. These unmeasured variables may have played an important role in shaping th