To the best of our knowledge, this is the first study to evaluate HF hospitalization among patients with HCM in Japan using a large, hospital-based claims database. Using data from a large sample of patients with HCM from the Japanese MDV claims database, we identified both the clinical and economic burden of hospitalization with HF among patients with HCM, including incidence, associated factors, hospitalization outcomes, and HCRU. No clear differences or trends were observed among HF phenotypes with regard to hospital stay duration, medical cost, and clinical outcomes, possibly due to the sample size. Compared with HF hospitalization in the general population, the proportion of hospitalizations due to HFpEF is higher in patients with HCM [21].

In this study, the incidence rate of HF hospitalization among patients with HCM was 17.2 events per 1000 patient-years. The Kochi RYOMA study [14], undertaken by Miyamoto et al., reported that the 5-year rate of HF events among patients with HCM was 9.6% in a cohort of patients with HCM from an unselected regional Japanese population [14]. Our study reinforced the findings of the Kochi RYOMA study by demonstrating the incidence rate of HF hospitalization in a large cohort of newly diagnosed patients with HCM in Japan. It should be noted that HF event rates reported in studies from countries outside Japan were comparatively lower than those observed in the present study and another study (5.3–14.0 per 1000 patient-years) [22,23,24,25]. High rates of HF hospitalization in this study may be attributed to the older mean age of patients (71.4 years) with HCM in the overall cohort. In the Kochi RYOMA study and other studies, the average age of Japanese patients with HCM was 65 years and < 40 years, respectively [9, 14]. Possible reasons for the lower incidence of HF hospitalization reported in this study compared with the RYOMA study were, firstly, that this study evaluated newly diagnosed patients with HCM; secondly, this study was designed to capture only the first hospitalization due to HF; and thirdly, data obtained from patients hospitalized for HF in other hospitals were not captured in this study. This implies that the burden of HCM in Japan may be different from that in other countries, mainly due to age, and further investigation is needed to explore the burden of HCM, including AF, stroke, and other complications.

In the nested case–control analysis, patients in the HF hospitalization group had higher rates of HF, diabetes mellitus, hypertension, AF, MI, arrhythmia except AF, and dyslipidemia than those in the control group. Several other studies have reported that these comorbidities are more prevalent among patients with HF and individuals hospitalized for HF [26,27,28,29]. In addition, patients in the HF hospitalization group were more frequently prescribed beta-blockers, non-DHP Ca channel blockers, DHP channel blockers, ACE inhibitors, ARBs, ARNIs, loop diuretics, thiazide, vasopressin 2 receptor agonists, MCRA, digitalis, and SGLT2i than those in the control group; this may reflect the necessity of effectively managing the multiple comorbidities seen in patients who have been hospitalized for HF. In addition, patients in the HF hospitalization group were prescribed medication to manage multiple comorbidities more frequently than those in the control group. In the present study, AF, HF, MI, COPD, and diuretics were identified as significant predictors of hospitalization with HF using adjusted logistic regression analysis. In the RYOMA study, AF, fractional shortening, and BNP were identified as significant predictors of HF events using a multivariate Cox proportional hazards model [14]. As a limitation of our study, NYHA classification, echocardiographic data, and laboratory testing data were not available; hence, fractional shortening and BNP levels were not assessed. Conversely, we identified new risk factors such as HF, MI, COPD, and loop diuretics by virtue of the large sample size. Regarding the use of loop diuretics, there is a possibility that some patients may have been treated for “worsening HF” on an outpatient basis before being admitted to hospital for HF. In addition, it is reported that residual congestion at discharge is associated with an 88% increased risk in readmissions rate for HF and a 54% increase in all-cause mortality risk, thereby being highly associated with prognosis [30]. Further research is needed to explore the disease–pathology relationship between HCM and comorbidities.

This is the first study describing HCRU due to hospitalization with HF in patients with HCM. The overall mean hospitalization costs in this study were 1035 kJPY (~ 156,750 USD) on average, and the median length of hospital stay was 20.0 days. These hospitalization costs and lengths of stay were comparable to those reported in previous studies evaluating HCRU of hospitalization due to HF [31,32,33]. Given that the incidence rate of first HF hospitalization is considerably high among patients with HCM and the prognosis after an HF hospitalization can be poor, with potential for repetitive hospitalizations, the economic burden of HF hospitalization is thought to be significant [34, 35].

Another important finding in the present study is the mortality rate observed during hospitalization with HF. Around 8.8% (46 of 521) of patients with HCM hospitalized with HF died during the hospitalization period, which is comparable to that reported in previous studies [34,35,36]. The Japanese Cardiac Registry of Heart Failure in Cardiology (JCARE-CARD) study [34,35,36] and an MDV claims-based study reported mortality rates of 7–8% and 10%, respectively [15]. Given that readmission rates for HF reported at 6 months and 1 year were 27% and 35%, respectively, in the JCARE-CARD study [34, 35] and the mortality rate is comparable to that after the first HF hospitalization, the mortality rate after the first hospitalization is expected to be significant.

Although current treatment options are targeted only at relieving symptoms, the long-term impact of HCM should be considered in its management. Disease-modifying treatments are expected to improve the burden of comorbidities such as HF.

The strengths of the present study were, firstly, that the MDV database covers a broad range of patients and hospitals in Japan; secondly, the study population included a large sample of patients diagnosed with HCM who were hospitalized with HF and treated in a real-world clinical setting in Japan; and finally, the MDV database contains valuable information about patients during hospitalization for HF, including diagnoses, administered medications and prescriptions, medical procedures, and outcomes of the hospitalization.

There are also some limitations to this study. First, the MDV is not a closed system; thus, while information is expected to be complete for the HF hospitalization, services received outside the MDV hospitals have not been captured, such as some hospitalizations post-discharge and outpatient visits. Second, given that the 12-month lookback period was established before HCM diagnosis, patients had been visiting DPC hospitals for the treatment or observation of other diseases. Therefore, patients might have had more comorbidities than the general population. Additionally, because the majority of patients in this study had received a diagnosis of HCM of an unspecified subtype, we could not fully assess differences among HCM subtypes. Third, as mentioned earlier, NYHA classification, echocardiographic data, and laboratory testing data were not available; therefore, important clinical factors associated with cardiac function and HF events were not included in the analysis. Additionally, the absence of data on the number of patients on implantable cardioverter-defibrillator (ICD) therapy and the number of ICD shocks is a limitation of our study. Fourth, only in-hospital deaths are captured in the MDV; thus, post-discharge mortality could not be analyzed. Fifth, due to the limited availability of laboratory data in the database, including biomarkers for HF such as BNP/N-terminal pro B-type natriuretic peptide (NT-proBNP), we were unable to assess their potential as prognostic indicators or treatment targets, nor analyze their relationship with patient outcomes. Given this lack of laboratory data, it is difficult to evaluate the impact on the length of stay and the costs of treatments. Sixth, we acknowledge that understanding the interplay between HF phenotypes or severity and other comorbidities is critical for approaching the prognosis of patients with HCM more comprehensively. Lastly, the retrospective study design is a limitation that may affect the findings. Such analyses would require prospective studies designed to collect robust datasets capturing information on HF subtypes, severity, and relevant comorbid conditions. These factors should be examined in future investigations to enhance the understanding of prognostic determinants in patients with HCM suffering from HF. Results of the analyses using the data from HCM registries in Japan are eagerly anticipated in this regard [14, 37,38,39].