Mitral annular calcification (MAC) and heart failure with preserved ejection fraction (HFpEF) share risk factors but links between the 2 have not been well studied. Using data from the TriNetX US Collaborative Network various cohorts were created which were propensity score matched (by demographics and comorbidities) and used to perform 3 studies: 1) All hospitalized patients between 2020 and 2023, without prior heart failure, were stratified by presence or absence of MAC to assess risk of developing HFpEF at 1 year. 2) Patients with a HFpEF diagnosis were stratified by presence or absence of a new diagnosis of MAC between 2020 and 2023 and evaluated for cardiovascular outcomes at 1 year. 3) Data from study-1 were further explored for interactions between MAC and body-mass index (BMI) in development of HFpEF at 1 year. In study-1, MAC patients were at higher risk of developing HFpEF versus no-MAC (HR 3.80, p < 0.0001). In study-2, HFpEF patients with vs without MAC were at higher risk of HF hospitalization (HR 1.24, p < 0.0001), all-cause hospitalization (HR 1.34, p < 0.0001), and various cardiovascular outcomes. In study-3, incident HFpEF in the no-MAC group was low with a “J-shaped” curve noted across BMI quintiles. MAC patients had a much higher incidence of HFpEF with little influence of BMI except in the highest quintile. In conclusion: MAC is an independent risk factor for development of HFpEF. When present in HFpEF patients, there is increased risk of re-hospitalization and cardiovascular events. MAC has much greater impact on development of HFpEF than does BMI.
Nonrheumatic mitral annular calcification (MAC) is a degenerative process that is typically associated with biological aging and chronic inflammation. It is a common finding on echocardiography and other types of cardiac imaging. , Risk factors for development of MAC include obesity, diabetes, hyperlipidemia, and hypertension . Additionally, MAC has been shown to be associated with a variety of major adverse cardiac events. ,,, Despite these associations, MAC is generally an underappreciated condition. Heart failure with preserved ejection fraction (HFpEF) is a prevalent condition which shares risk factors with MAC. Interactions between MAC and HFpEF have only recently been investigated. , In this series of studies, we aimed to examine the impact of MAC on development of HFpEF and on cardiovascular outcomes in those with pre-existing HFpEF.
Methods
This is a series of observational studies utilizing data from the TriNetX US Collaborative Network which collects information from 71 Healthcare Organizations (HCOs) covering approximately 120 million patients. The data in TriNetX is deidentified and thus our analysis was exempt from institutional review board approval. The TriNetX application also provides integrated analytical tools which allow for patient selection based on diagnoses, treatments, and medications. Pooled characteristics of chosen patient cohorts, incidence and prevalence analysis, cohort propensity score matching, and statistical analysis of outcomes between cohorts can all be conducted on their platform.
For our studies, we focused on patients over the age of 50 years with data collected between January 1, 2020 to December 31, 2023 to allow for a full 1-year follow-up period. 50 years of age was selected due to a substantial drop-off in prevalence of MAC at younger ages. We additionally limited our data query to 2020 as MAC codes were infrequently used prior to that time. Our goals for these analyses were to determine: (1) the impact of MAC on the incidence of HFpEF, (2) the impact of MAC on cardiovascular outcomes in patients with pre-existing HFpEF, and (3) possible interactions between MAC and BMI in development of HFpEF. Patient diagnoses were determined using International Classification of Diseases, Tenth Revision (ICD-10). We additionally used TriNetX curated patient data allowing us to query for specific patient lab and vital sign data provided to TriNetX by some HCOs. Exclusion criteria differed by cohort but patients with rheumatic mitral valve disease were excluded in all analyses; heart failure with reduced ejection fraction (HFrEF) and combined diastolic and systolic heart failure were excluded, as appropriate, by using their associated ICD-10 codes.
In our first study, we investigated the role of MAC in the development of HFpEF. Our primary outcome was the development of HFpEF with secondary outcomes comprising all heart failure and HFrEF. We queried for patients ≥50-years of age who were hospitalized between 2020 and 2023. They could not have any history of rheumatic mitral valve disease, any history of heart failure, or end stage renal disease (ESRD) prior to this hospitalization. The patients must have had a recorded BMI during this time frame. Patients in the MAC group needed to have newly diagnosed MAC during this time frame while those in the no-MAC group had MAC excluded. To ensure accurate results, patients with the outcomes of interest prior to the outcome window were excluded from analysis.
In our second analysis, we looked at MAC in patients with pre-existing HFpEF. Our primary outcome was heart failure hospitalization at 1 year, with secondary outcomes including all-cause hospitalization, intensive care unit (ICU) admission, ischemic heart disease, cerebrovascular disease, atrial fibrillation or flutter, ventricular tachyarrhythmias, aortic valve insufficiency or stenosis, mitral valve insufficiency or stenosis, heart valve replacement, acute renal failure, ESRD or dialysis initiation, cardiac arrest, insertion of a pacemaker or ICD, and cardiogenic shock. We queried for patients ≥50-years of age who had a diagnosis of HFpEF between 2020 and 2023. History of HFrEF, occurrence of ESRD before the outcome window, or any history of rheumatic mitral valve disease were excluded while prior HFpEF was not. Patients in the MAC group must have received a new diagnosis of MAC between 2020 and 2023 while those in the no-MAC group had no prior diagnosis of MAC. The patients must have had a recorded BMI during this time frame.
In our third analysis, we explored the impact of BMI on the development of heart failure for patients with and without MAC with the primary outcome being new onset HFpEF and secondary outcomes being new onset HFrEF and all-cause heart failure. Using our dataset from Study-1, BMI quintiles were created with BMI of <20, 20–24.9, 25–29.9, 30–34.9, and ≥35 for the hospitalized cohorts with and without MAC. We used ICD-10 BMI codes as well as TriNetX curated codes to create our groups. The normal BMI group (20–24.9) served as the comparator for each of the other quintiles. Analyses were done separately for the MAC and no-MAC groups.
Baseline characteristics are represented as mean ± SD for continuous variables and compared with independent-sample Student’s t-test. Additional categorical variables are shown as a number and percentage and analyzed with chi-square testing. A 1:1 propensity match technique by the nearest-neighbor algorithm was used to limit bias. A standard mean difference (SMD) of ≤0.1 was taken as indicating proper balance based on Cohen’s Criterion which states an SMD of 0.2 denotes a small difference, 0.5 a medium difference, and 0.8 a large difference between the groups. Propensity matching had slight variation between our 3 different testing groups. Throughout all 3 studies age, racial demographics, sex, diabetes, smoking history, ischemic heart disease, cerebrovascular disease, hypertensive disease, hyperlipidemia, chronic kidney disease, atrial fibrillation/flutter, and sleep apnea were matched. BMI was also accounted for in our analyses except in our study on the impact of BMI on MAC outcomes.
We used TriNetX’s integrated analytic tools to run our statistical analyses. Hazard Ratios (HRs), 95% confidence intervals, and p-values are reported for outcomes of interest. The relevant tests of significance were 2-tailed, with a p – value of <0.05 considered significant.
Results
Analysis 1: Impact of MAC on development of HFpEF
The first analysis explored the impact of MAC on the development of HFpEF. 10,629 patients were included in the MAC group and 3,473,162 in the no-MAC group. Prior to propensity matching the MAC group was significantly older (73.3 ± 9.9 years vs 65.7 ± 11.2 years), more white (78.7% vs 68.5%), more female (57.1% vs 50.8%), and had significantly higher prevalence of each comorbidity that was matched. BMI, however, was not significantly different (29.2 ± 7.0 vs 29.4 ± 7.0, SMD 0.025). See Table 1 for full details.
Table 1
Baseline characteristics of the population before and after propensity match scoring
| Study-1 | ||||||
|---|---|---|---|---|---|---|
| Before match | After match | |||||
| MAC cohort | No MAC cohort | Standard difference | MAC cohort | No MAC cohort | Standard difference | |
| n = 10,629 | n = 3,473,162 | <0.1 indicates adequate balance | n = 10,628 | n = 10,628 | <0.1 indicates adequate balance | |
| Demographics | ||||||
| Age at Index | 73.3 ± 9.9 | 65.7 ± 11.2 | 0.722 | 73.3 ± 9.9 | 73.3 ± 9.9 | 0.004 |
| Sex (% Female) | 57.1 | 50.8 | 0.126 | 57.1 | 57.0 | 0.001 |
| White (%) | 78.7 | 68.5 | 0.232 | 78.7 | 78.8 | 0.003 |
| Black/African-American (%) | 10.6 | 15.2 | 0.135 | 10.6 | 10.5 | 0.002 |
| Comorbidities | ||||||
| Hypertensive diseases (%) | 81.3 | 44.4 | 0.824 | 81.3 | 81.6 | 0.007 |
| Atrial fibrillation and flutter (%) | 22.1 | 6.2 | 0.821 | 22.1 | 21.8 | 0.007 |
| Diabetes mellitus (%) | 39.0 | 17.8 | 0.484 | 39.0 | 39.0 | < 0.001 |
| Hyperlipidemia (%) | 71.8 | 38.3 | 0.713 | 71.8 | 72.1 | 0.008 |
| Chronic kidney disease (%) | 23.5 | 6.4 | 0.496 | 23.5 | 23.3 | 0.005 |
| Sleep apnea (%) | 19.7 | 10.1 | 0.273 | 19.7 | 19.9 | 0.006 |
| Personal history of nicotine dependence (%) | 29.9 | 12.1 | 0.445 | 29.8 | 29.5 | 0.007 |
| Body mass index | 29.2 ± 7.0 | 29.4 ± 7.0 | 0.025 | 29.2 ± 7.0 | 29.2 ± 6.9 | 0.005 |
| Study-2 | ||||||
|---|---|---|---|---|---|---|
| Before match | After match | |||||
| MAC and HFpEF cohort | HFpEF without MAC cohort | Standard difference | MAC and HFpEF cohort | HFpEF without MAC cohort | Standard difference | |
| n = 3,117 | n = 158,548 | n = 3,117 | n = 3,117 | |||
| Demographics | ||||||
| Age at index | 77.2 ± 9.0 | 74.2 ± 10.8 | 0.311 | 77.2 ± 9.0 | 77.3 ± 9.6 | 0.007 |
| Sex (% Female) | 63.8 | 57.1 | 0.138 | 63.8 | 63.4 | 0.008 |
| White (%) | 78.6 | 74.6 | 0.094 | 78.6 | 79.9 | 0.033 |
| Black/African-American (%) | 12.8 | 14.7 | 0.056 | 12.8 | 11.8 | 0.029 |
| Comorbidities | ||||||
| Hypertensive diseases (%) | 94.0 | 79.3 | 0.443 | 94.0 | 93.6 | 0.015 |
| Hypertensive heart disease (%) | 32.3 | 11.9 | 0.507 | 32.3 | 32.0 | 0.007 |
| Ischemic heart disease (%) | 69.8 | 45.2 | 0.515 | 69.8 | 70.7 | 0.020 |
| Cerebrovascular disease (%) | 39.1 | 23.5 | 0.341 | 39.1 | 39.3 | 0.004 |
| Chronic rheumatic heart disease (%) | 25.3 | 12.4 | 0.333 | 25.3 | 24.3 | 0.024 |
| Atrial fibrillation and flutter (%) | 47.1 | 33.9 | 0.271 | 47.1 | 46.3 | 0.016 |
| Diabetes mellitus (%) | 50.7 | 39.9 | 0.218 | 50.7 | 50.1 | 0.011 |
| Hyperlipidemia (%) | 77.1 | 55.6 | 0.467 | 77.1 | 78.4 | 0.031 |
| Chronic kidney disease (%) | 46.6 | 29.6 | 0.356 | 46.6 | 45.9 | 0.014 |
| Personal history of nicotine dependence (%) | 39.7 | 24.0 | 0.342 | 39.7 | 38.1 | 0.031 |
| Medications | ||||||
| Loop diuretics (%) | 72.6 | 59.0 | 0.289 | 72.6 | 72.2 | 0.009 |
| Thiazide diuretics (%) | 40.3 | 35.9 | 0.091 | 40.3 | 39.2 | 0.022 |
| Potassium sparing diuretics (%) | 24.4 | 16.8 | 0.19 | 24.4 | 23.3 | 0.027 |
| Calcium channel blockers (%) | 67.8 | 50.1 | 0.366 | 67.8 | 68.4 | 0.014 |
| Angiotensin II inhibitors (%) | 43.1 | 31.2 | 0.249 | 43.1 | 43.6 | 0.010 |
| ACE inhibitors (%) | 40.7 | 36.4 | 0.088 | 40.7 | 41.1 | 0.008 |
| Aspirin (%) | 72.0 | 54.2 | 0.375 | 72.0 | 72.2 | 0.006 |
| Body mass index | 30.8 ± 8.2 | 32.0 ± 8.8 | 0.141 | 30.8 ± 8.2 | 31.3 ± 8.3 | 0.059 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
