Heart rate as an early predictor of severe cardiomyopathy and increased mortality in peripartum cardiomyopathy

Abstract Background Delays in diagnosis of peripartum cardiomyopathy (PPCM) are common and are associated with worse outcomes; however, few studies have addressed methods for improving early detection. Hypothesis We hypothesized that easily accessible data (heart rate [HR] and electrocardiograms [ECGs]) could identify women with more severe PPCM and at increased risk of adverse outcomes. Methods Clinical data, including HR and ECG, from patients diagnosed with PPCM between January 1998 and July 2016 at our institution were collected and analyzed. Linear and logistic regression were used to analyze the relationship between HR at diagnosis and the left ventricular ejection fraction (LVEF) at diagnosis. Outcomes included overall mortality, recovery status, and major adverse cardiac events. Results Among 82 patients meeting inclusion criteria, the overall mean LVEF at diagnosis was 26 ± 11.1%. Sinus tachycardia (HR > 100) was present in a total of 50 patients (60.9%) at the time of diagnosis. In linear regression, HR significantly predicted lower LVEF (F = 30.00, p < .0001). With age‐adjusted logistic regression, elevated HR at diagnosis was associated with a fivefold higher risk of overall mortality when initial HR was >110 beats per minute (adjusted odds ratio 5.35, confidence interval 1.23–23.28), p = .025). Conclusion In this study, sinus tachycardia in women with PPCM was associated with lower LVEF at the time of diagnosis. Tachycardia in the peripartum period should raise concern for cardiomyopathy and may be an early indicator of adverse prognosis.


| INTRODUCTION
Pregnancy-related mortality remains a leading cause of death among women across the world. 1 Nearly two-thirds of pregnancy-related deaths occur around the time of delivery or within the first 6 weeks postpartum, 2 and as many as two-thirds of these deaths are considered to be preventable. [3][4][5] Cardiovascular disease is the leading cause of pregnancy-related deaths in the United States. 6,7 Between 2014 and 2017, cardiovascular conditions comprised 15.5% of maternal deaths in the United States. 6,7 Cardiomyopathy was responsible for 11.5% of deaths 7 and is currently the leading cause of death between 6 weeks and 1 year following delivery. 8 Peripartum cardiomyopathy (PPCM) is an idiopathic form of left ventricular systolic dysfunction that causes clinical heart failure in women during the last trimester of pregnancy or the early postpartum period. 9,10 Although most patients recover, mortality averages 5%-15% in the United States 9-11 and can be even higher in other areas of the world. 10,12 Survivors may suffer significant morbidity related to persistently decreased cardiac function, including the need for multiple medications, frequent hospital readmissions, and poor quality of life. 13 Numerous studies have shown that left ventricular ejection fraction (LVEF) less than 30% at the time of diagnosis portends a worse prognosis, with lower rates of recovery and increased risk of adverse events. 9,11 Delays in diagnosis are associated with lower LVEF, higher rates of adverse outcomes, and lower rates of recovery. 14,15 Early detection of PPCM is challenging because symptoms of heart failure mimic those of normal pregnancy, especially late in the third trimester and early after delivery. 9,10 Despite the clear association between late diagnosis and worse outcomes, few studies have addressed methods for improving early detection. One risk prediction tool was recently validated but includes several demographic and comorbidity variables that would be most effective if integrated into an electronic medical record. 16 We, therefore, sought to determine whether simple bedside clinical data (heart rate [HR]) and electrocardiogram [ECG]) could effectively identify women with more severe forms of PPCM and increased risk for adverse outcomes. We utilized a well-defined clinical cohort of women with PPCM to test our hypothesis that significant elevations in HR or abnormal ECG findings in peripartum women would be indicators of more severe cardiomyopathy and, consequently, increased risk of adverse outcomes.

| Study population
Approval for the study and waiver of written informed consent was obtained from the University of Michigan Institutional Review Board.
Patients diagnosed with PPCM (January 1, 1998-July 31, 2016) were identified using the University of Michigan Electronic Medical Record Search Engine (EMERSE). Details of the search engine function and performance have been previously described elsewhere. 17 EMERSE was used to search for all patient encounters with the following terms: pregnancy-associated cardiomyopathy, post-partum cardiomyopathy, postpartum cardiomyopathy, peri-partum cardiomyopathy, peripartum cardiomyopathy, and PPCM. Each chart was manually reviewed in detail. ECG and HR data were extracted by a physician reviewer initially blinded to associated outcomes and LVEF.
Patients were included if they met the diagnostic criteria for PPCM defined according to the currently accepted definition as an idiopathic cardiomyopathy with initial LVEF < 45% presenting towards the end of pregnancy or in the months following delivery. 9,10,18 Only patients diagnosed at our institution with sufficient imaging data to confirm the diagnosis and with a minimum of 6 months of follow-up were included (Figure 1). Exclusion criteria included prior chemotherapy, chest radiation, congenital heart disease, valvular disease, family history suggestive of hereditary dilated cardiomyopathy, preceding febrile illnesses, and recent or recurrent illicit drug use.
Race was self-identified in the medical record.

| Data collection
Demographic information, HR, LVEF, and outcomes data were collected for all patients. HR at the time of diagnosis (baseline HR) was determined by ECG on presentation if it was available, or by initial HR obtained on examination and documented in the electronic medical record at the visit during which PPCM was diagnosed. Sinus tachycardia was defined as HR ≥ 100 beats per minute (bpm); this was further specified to identify patients with HR >110 and >120 bpm. For the subset of patients who had ECGs available (n = 68), conventional parameters including ventricular rate, QRS axis, PR interval, QRS duration, QTc interval, and T-wave amplitude in lead aVR were calculated. ECGs were also analyzed for the presence of arrhythmia, pathologic Q waves (defined as a Q wave >2 mm deep or >40 ms wide), left or right atrial enlargement, right or left bundle branch block, T-wave inversion, left ventricular hypertrophy, or poor R-wave progression (defined as R < S in V4). QTc interval was considered prolonged if it was >460 ms. A positive T wave in lead aVR was defined as a wave with a positive deflection >0 mV. An ECG was considered abnormal if any interval was prolonged or if any of the above findings were present. Both sinus tachycardia (ventricular rate >100 bpm) and sinus bradycardia (ventricular rate < 60 bpm) were considered abnormal findings. All ECG findings were confirmed by a board-certified cardiologist and also reviewed by a physician investigator blinded to the clinical outcomes.

| Statistical analysis
Categorical variables were compared using the Pearson χ 2 test and reported as frequency (n) ± percentage (%). Linear regression was used to assess the associations between continuous variables.
Continuous outcome results are reported as mean ± SD or SE. Odds ratios (OR) for dichotomous outcomes were calculated using binary logistic regression and reported as 95% confidence interval (CI). In addition, logistic regression was performed to control for patient age and is presented as an adjusted odds ratio (AOR). All p values are two-sided, with p < .05 considered significant. Statistical analyses were performed using SPSS 27 (IBM Corp).

| Comparison of HR with LVEF recovery, MACE, and mortality
In the analysis of outcomes, sinus tachycardia at diagnosis was significantly associated with recovery at 6 months (

| ECG characteristics and subgroup analysis
Within the study population, there were 68 patients with complete ECG data available from the time of diagnosis. The majority of ECGs Note: χ 2 test was performed; the exact test was performed for variables with groups n < 5. Age at diagnosis is reported as age (mean ± SD). All other data are reported as n (%), unless otherwise indicated. Hypertensive disorders of pregnancy include pre-existing hypertension, gestational hypertension, and/or pre-eclampsia.

| DISCUSSION
This study demonstrates that tachycardia in the peripartum period is a simple, early, independent predictor of severe cardiac dysfunction among women with PPCM. In this well-defined cohort, sinus tachycardia at the time of diagnosis was associated with a lower initial LVEF, a lower likelihood of recovery, and a significantly increased risk of overall mortality, especially if the HR was >110 bpm (Central Illustration). While other ECG findings such as prolonged QTc, T-wave inversion, and pathologic Q waves were common in our PPCM cohort, only sinus tachycardia was significantly associated with outcomes. Although tachycardia is common among women at the end of pregnancy or early after delivery, we found that among women diagnosed with PPCM, sinus tachycardia (HR > 110 bpm) was associated with a fivefold increased risk of mortality.
Sinus tachycardia is common during and after pregnancy, but F I G U R E 2 Relationship between ejection fraction and heart rate at diagnosis. This figure shows the relationship between HR (independent variable) and LVEF (dependent variable). Using linear regression, higher HR was associated with lower LVEF. HR, heart rate; LVEF, left ventricular ejection fraction  (24) .687 12 (25) .925 12 (20) .316

T A B L E 3 Association between elevated HR and recovery
Note: Values are n (%), unless otherwise indicated.
Similar to the sepsis initiative of recent years, there has been a focus within obstetrics on identifying early warning signs that may help differentiate these healthy women from those with critical illness in the peripartum period to decrease maternal mortality related to delays in diagnosis and treatment. Mhyre and colleagues proposed the Maternal Early Warning Criteria, a set of vital sign parameters and clinical signs/symptoms that should prompt physician evaluation of a pregnant patient deemed to be at risk of a critical illness such as hemorrhage, sepsis, hypertensive crisis, or heart failure. 20 One of the criteria included is HR > 120 bpm, which the authors argue should raise concern for the possibility of a lifethreatening cardiac problem among other diagnoses. Our finding that sinus tachycardia (especially HR > 110-120 bpm) predict severe cardiomyopathy and increased mortality in PPCM further supports the recommendation that elevated HR should draw attention in a pregnant or postpartum woman.
Our study also supports prior findings that lower LVEF at the time of diagnosis is strongly associated with a worse prognosis. 11,[21][22][23] One of the most prominent studies in this regard is the

Investigations of Pregnancy-Associated Cardiomyopathy study,
which demonstrated that an initial LVEF < 30% was significantly associated with a persistently low LVEF and lack of recovery at 12-month follow-up. 21 Similarly, we found that women with an initial LVEF < 25% were 75% less likely to fully recover. In a study by Goland et al., 22 24 found that a positive T wave in lead aVR was a strong predictor of adverse cardiac outcomes. A recent study by Hoevelmann et al. 25 found that prolonged QTc at presentation predicted poor 6-month outcomes, while sinus tachycardia predicted poor outcomes at 12 months. Although our study investigated the relevance of each of these ECG findings, only sinus tachycardia was found to be significantly associated with lower initial LVEF and increased mortality. Overall, our findings and those of others show that various ECG abnormalities are common in women with PPCM; however, a normal ECG is not sufficiently sensitive to rule out the diagnosis.
While the above study by Hoevelmann et al. 25 and another recent study by Mbakwem et al. 29 also investigated the relationship between tachycardia and outcomes, both designated sinus tachycardia as HR of 100 bpm or greater and used a cutoff of LVEF < 35% to evaluate systolic function. This is in contrast to our study in which we further evaluated HR >110 and >120 bpm to characterize a linear relationship between increasing degrees of tachycardia and increased severity of cardiomyopathy ( Figure 2). Another prior study, from a registry in Nigeria, reported that sinus tachycardia was associated with increased mortality; however, most patients in this study were diagnosed late at 3-6 months postpartum, likely representing those who had failed to recover and therefore already at higher risk of death. 30 Because such delays in diagnosis are common and are associated with worse outcomes, 14 and echocardiographic data that may have further described the baseline characteristics of this cohort. It is worth noting that this cohort may represent a particularly high-risk patient population as evidenced by the high proportion of patients with baseline LVEF < 25% (49%), mortality rate of 13.4%, six patients requiring LVAD, and two patients undergoing heart transplantation. Additionally, our study focused on the use of EKG and HR, but future studies could assess the implication of other clinical variables that could also predict adverse outcomes.
In conclusion, we suggest that elevated HR in a pregnant woman, especially above 110 bpm, should prompt health professionals to assess for signs and symptoms of heart failure. When there is clinical concern, additional testing such as BNP and echocardiogram are required to confirm the diagnosis. Future investigation is needed to determine whether tachycardia would reliably identify women with heart failure in a general population of peripartum women, potentially leading to reductions in maternal morbidity and mortality.

| Competencies in medical knowledge
Sinus tachycardia is common during pregnancy and after delivery; however, elevated HR may be an early predictor of severe cardiomyopathy and early mortality. Elevated HR in a pregnant woman, especially above 110 bpm, should prompt health professionals to assess for signs and symptoms of heart failure and perform additional testing such as BNP and echocardiogram when warranted.