Data source

This study used the InGef (Institute for Applied Health Research Berlin) database which is an anonymized healthcare claims database with longitudinal data over a look-back period of up to six years from approximately 6.7 million Germans insured in one of approximately 70 German social health insurances (SHIs) currently contributing data to the database (mainly company or guild health insurances). Claims data are transferred directly from the healthcare providers to a specialized data center owned by the SHIs, which provides data warehouse and information technology services. In the data center, data are anonymized before entering the InGef database. Data are anonymized with respect to individual insurant, healthcare providers (e.g. physicians, practices, hospitals and pharmacies) and the respective SHI. The estimated delay of data being available in the database is approximately 3 to 9 months (e.g. healthcare data until 31 December 2016 can be expected to be available in the database at September of 2017 at the latest). The most important data elements included in the database are demographic information (including the date of death if applicable); ambulatory services and ambulatory diagnoses; hospitalization information including date of admission and discharge, diagnoses and procedures.

Diagnoses were coded via the International Statistical Classification of Diseases and Related Health Problems (ICD-10-GM [German Modification]).

Study population

The InGef research database was used to conduct yearly retrospective cohort studies to estimate the population-based HCM prevalence per 100,000 persons for each year between 2011 and 2015.

In order to estimate the HCM prevalence, yearly population-based cohorts were created between 2011 and 2015. Subjects were eligible to enter the cohort in the respective year, if they fulfilled all of the following criteria in the respective year: (1) continuous enrolment in the SHI in the respective year or from birth onwards or until death (base population at risk = denominator); (2) no concomitant diagnosis of HCM phenocopies (mitochondrial disease (G31.81), Danon disease (E74.0x), Friedreich’s ataxia (G11.3x), Leopard syndrome (Q87.8x), Noonan syndrome (Q87.1x), Anderson–Fabry disease (E75.2x), Amyloidosis (E85x). Patients with at least one ambulatory verified or a hospital main- or secondary discharge diagnosis of HCM (I42.1 or I42.2) between 01.01.and 31.12. of the respective year were considered HCM cases.

Cohort entry was the 1st January of the respective study year, if all inclusion criteria were fulfilled, or the date of birth for newborns. Cohort exit was defined as the occurrence of HCM, death, or the 31st December of the respective study year, whichever occurred first.

Main findings

This study is–to the best of our knowledge–the first to systematically explore HCM prevalence of clinically apparent cases in Germany. For this purpose, a claims database of more than 5 million patients was analyzed and revealed (1) a lower prevalence of diagnosed HCM than in cohort studies based on echocardiographic diagnosis; (2) gradual HCM prevalence increases with advancing age and (3) from the year 2011 to 2015.

Comparison with previous studies

Several studies have established a HCM prevalence of about 0.2% (1:500 cases) in adults [2–5]. Importantly, those studies used population screening by echocardiography that also allowed detection of clinically silent cases. Consequently, it is not surprising that the number of clinically diagnosed cases was lower in U.S. claims data [6]. While in the U.S. study HCM was found in 0.03%, we detected clinically apparent cases in 0.07%, which has been considered to represent “the tip of the iceberg of the disease spectrum” [6]. This is also in accordance with a similar smaller study from Sweden that reported a prevalence of 0.04% [8]. The discrepancies become even more prominent when genetic population studies are being factored in. Based on these data, the minimal prevalence of HCM gene carriers could be estimated at 0.5% (1 in 200 people) [9]. Although not all gene carriers may develop clinical HCM, the high frequency of HCM-causing pathogenic mutations strongly suggests a prevalence exceeding that reported in echocardiographic screening studies and by far exceeding that found when conducting claims-based analysis.

Traditionally, HCM has been regarded as a disease of the young, but is nowadays identified with increasing frequency in older patients. For instance, 25% of HCM patients in a community-based cohort not subject to tertiary center referral bias were > 75 years [10] which is in agreement with our observations. The notion that HCM may remain clinically silent for long periods of time with symptoms and initial diagnosis deferred until late in life may at least in part be explained by age-dependent penetrance of disease causing mutations and the presence of additional clinical and genetic modifiers. For instance, sarcomere protein mutations such as mutations in cardiac myosin binding protein-C, troponin I, and α-cardiac myosin heavy chain have been shown to cause elderly-onset HCM. This mutation spectrum is strikingly different from that of familial, early-onset HCM [11]. The presence of an additional risk factor for LV hypertrophy has been identified as modulator for the phenotypic expression of HCM [12]. Interestingly, hypertension–which was also found in the majority of our HCM cases–has been associated with later HCM diagnosis [13].

While HCM is most frequently transmitted as an autosomal-dominant trait most studies have reported a male preponderance that was also found in our study. This observation remains largely unexplained but might also reflect the impact of genetic and hormonal modifiers [14]. In addition, hypertension as modulating factor is more prevalent in males which may result on more frequent echocardiographic examinations and subsequent HCM diagnosis.

Recent advances such as better understanding of the underlying molecular and genetic pathophysiology, implementation of contemporary family screening, more sensitive diagnostic cardiac imaging, better risk stratification with improved outcomes may also contribute to the increasing prevalence of this disease.

Limitations

Limitations of this analysis include those inherent to claims database studies, such as the reliance on accurate coding and diagnosis. Consequently, results need to be interpreted cautiously. While in one study, accuracy of HCM coding was high [15], there were some misclassifications found in another study [8]. Since there is lack of clinical, ECG and imaging information database diagnoses could not be verified. Furthermore anonymized claims data cannot be combined with clinical records to confirm HCM diagnoses documented in claims data. In particular, the high prevalence of hypertension could have led to misdiagnosis, acting not only as modulator, but also as cause for secondary form of hypertrophy which would result in even lower overall HCM prevalence. Nevertheless, the discrepancy between claims data analysis revealing clinically diagnosed cases and prevalence studies based on outreach programs is striking.

Approximately 10% of the German population with private medical insurance is not covered in this database, which limits the extrapolation of the results to this population. Nevertheless, this database has been shown to be representative of the German population demographics and the directly standardized prevalence was calculated to account for any differences in age and gender distribution between the InGef database and the German population [16].