The burden of premature ventricular contractions predicts adverse fetal and neonatal outcomes among pregnant women without structural heart disease: A prospective cohort study

Abstract Background Premature ventricular contractions (PVCs) may increase during pregnancy, however, few studies have evaluated the relationship between PVCs and the pregnant outcomes. Hypothesis PVCs may increase the adverse fetal/neonatal outcomes in pregnant women. Methods Six thousand one hundred and forty‐eight pregnant women were prospectively enrolled in our center between 2017 and 2019 in the study. The average PVC burden was determined by calculating the number of PVCs in total beats. Those who had a PVC burden >0.5% were divided into two groups based on the presence or absence of adverse fetal or neonatal events. The adverse outcomes were compared between the groups to assess the impact of PVCs on pregnancy. Results A total of 103 (1.68%) women with a PVC burden >0.5% were recorded. Among them, 17 adverse events (12 cases) were documented, which was significantly higher than that among women without PVCs (11.65% vs. 2.93%, p < .01). The median PVC burden among pregnant women with PVCs was 2.84% (1.02%–6.1%). Furthermore, compared with that of the women without adverse events, the median PVC burden of women with adverse fetal or neonatal outcomes was significantly higher (9.02% vs. 2.30%, p < .01). Multivariate logistic regression analysis demonstrated that not the LVEF, heart rate and bigeminy, but only the PVC burden was associated with adverse fetal or neonatal outcomes among pregnant women with PVCs (OR: 1.34, 95% CI [1.11–1.61], p < .01). Conclusions Frequent PVCs have adverse effects on pregnancy, and the PVC burden might be an important factor associated with adverse fetal and neonatal outcomes among pregnant women with PVCs.

analysis demonstrated that not the LVEF, heart rate and bigeminy, but only the PVC burden was associated with adverse fetal or neonatal outcomes among pregnant women with PVCs (OR: 1.34, 95% CI [1.11-1.61], p < .01).
Conclusions: Frequent PVCs have adverse effects on pregnancy, and the PVC burden might be an important factor associated with adverse fetal and neonatal outcomes among pregnant women with PVCs.
K E Y W O R D S fetal/neonatal outcomes, pregnancy, premature ventricular contractions (PVCs)

| INTRODUCTION
Idiopathic premature ventricular contractions (PVCs) are relatively benign in cases without structural heart diseases but may signal an increased risk of sudden death in cases with structural heart disease (SHD) and may be markers of underlying pathology. An estimated prevalence of 1%-4% is found in the general population on standard 12-lead electrocardiography (ECG). 1 Frequent PVCs, defined as more than 10% of all QRS complexes on standard 24 h Holter monitoring, have been found to be associated with subsequent development of left ventricular dilatation and PVC-induced cardiomyopathy. Treatment of frequent PVCs in patients with impaired ventricular function can reverse this pattern. [2][3][4][5][6][7] PVCs may increase during pregnancy, which leads to an additional challenge, as pregnancy outcome implications are not known for this already cardiac overloaded state. 8,9 Relatively few studies have evaluated the relationship between PVCs and the adverse outcomes of fetuses or neonates. 10 Therefore, we aimed to determine the relationship between the adverse outcomes of the fetus and neonate among pregnant women with PVCs.

| Study design and patient population
This is a single-center prospective cohort study of consecutive pregnancies referred to The First Affiliated Hospital of Nanjing Medical University from July 2017 to July 2019. The inclusion criteria were pregnant women with normal cardiac structure and function based on echocardiographic examination. Pregnant women with PVC burdens greater than 0.5% on Holter examination each time during the whole pregnancy period were classified as the PVC group. The exclusion criteria were as follows: Pregnant women with (1) hyperthyroidism, (2) hypothyroidism, (3) gestational hypertension or pregnancy with hypertension, (4) Type I or II diabetes, (5) chronic nephritis or impaired renal function (creatinine clearance rate, Ccr <60 ml/min), (6) autoimmune disease, (7) congenital heart diseases, (8)

| Baseline characteristics
Clinical and 12-lead ECG data were collected at the time of the initial clinic visit or for routine obstetric examinations during the first trimester of pregnancy. If the standard 12-lead ECG recorded the PVC events, Holter monitoring was then performed and reexamined in each subsequent trimester to assess the burden of ventricular premature contractions and analyze the total number of QRS complexes, total number of PVCs and total runs of nonsustained ventricular tachycardia (NSVT) (run with >3 consecutive PVCs but <30 s). Furthermore, if the women had self-reported symptoms such as palpitation during pregnancy, the 12-lead ECG examination was then performed. If no arrhythmia was recorded, a wearable singlelead ECG recorder (Shuweikang company, Nanjing, China) was then weared for three continuous days to record the possible PVC events in those pregnant women. The single-lead wearable ECG recorder was designed to recognize the arrhythmia based on the convolutional neural network technology. It can easily identify the PVCs when it happens. And if the PVCs was documented by 12-lead ECG or the wearable single-lead ECG recorder and confirmed by the ECG experts in our center (Lin J & Chen QS), the Holter was examined thereafter once a trimester for the remainder of the pregnancy period. The average PVC burden was determined by calculating the total number of PVCs in the total QRS complexes. PVC morphological patterns were analyzed to determine the PVC origin. PVCs with inferior axis and left bundle branch block (LBBB) patterns were determined to most likely originate from the right ventricular outflow tract (RVOT). Specific diagnostic procedures were carried out with reference to published diagnostic criteria. 11 Specifically, the burden, morphology and the origin of PVCs were calculated not by the wearable recorder but by the Holter and 12-lead ECG results due to its single lead characteristics.
Echocardiography was performed in the first and third trimesters in all pregnant women. Experienced echocardiographers performed echocardiography to ensure that there was no evidence of SHD in the pregnant women and to ensure that the cardiac function of the enrolled pregnant women was within the normal range (left ventricular ejection function [LVEF] >60%).
All pregnant women were followed, and the data on characteristics including age, comorbid medical conditions and family history were recorded. All women were followed from preconception to 1 week after delivery. If the exclusion criteria were met, they were excluded from the study.

| Neonatal/fetal outcome events
Two physicians blinded to the women's baseline characteristics independently verified adverse events. Adverse fetal and neonatal events were defined as previously described 10 : Premature birth (<37 weeks gestation), small-for-gestational-age birth weight (<10th percentile for gestational age or < 2500 g), respiratory distress syndrome, intraventricular hemorrhage and fetal death (after 20 weeks gestation and before birth). Only one event was counted if multiple events occurred in the fetus or newborn simultaneously.

| Statistical analysis
Statistical analysis was performed using the SPSS (Version 22.0). Data are presented as the means ± SDs, medians (25th-75th percentiles), or proportions. Student's t-test or the Wilcoxon rank-sum test was used to compare continuous variables between the women with and without adverse fetal/neonatal events who had PVCs >0.5%. Possible risk factors of adverse events were analyzed using univariable logistic regression, followed by multivariable logistic regression. A p value <.05 was considered statistically significant.

| Baseline characteristics and fetal or neonatal outcomes
Six thousand one hundred and forty-eight pregnant women who fulfilled the inclusion criteria were enrolled consecutively, of which 103 (1.68%) had symptomatic or asymptomatic PVCs. A total of 17 adverse events occurred in 103 pregnant women with PVCs, including five cases of respiratory distress syndrome, five preterm births, and seven small-for-gestational-age births. Finally, only 12 women with adverse events were counted, as multiple events occurred in some of them simultaneously. The remaining defined adverse events did not occur. The other 91 pregnant women with PVCs delivered safely without adverse events. A total of 177 fetal and neonatal adverse cases were counted in the cohort of 6045 pregnant women without PVCs. The incidence of adverse events was significantly higher in PVC cases (11.65%) than in those without PVC (2.93%) (Figures 1 and 2).

| Cardiac rhythm of pregnant women with PVCs
The pregnant women with PVCs were divided into two groups on the basis of the presence or absence of adverse events. The baseline characteristics are listed in Table 1. The proportion of bigeminy PVCs was significantly higher in pregnant women with adverse fetal or neonatal outcomes (50% vs. 19.8%, p < .05), as was the median PVC burden, than in women without adverse events (9.02% vs. 2.30%, p < .01).
Eight pregnancies had a maximal PVC burden >10%, and of these, two pregnancies had a PVC burden >20%. Although the LVEF of the adverse outcome group was within the normal range, it remained slightly lower than that of the control group (64.16% ± 1.56% F I G U R E 1 Summarized flow chart for the study findings and the fetal/neonatal outcomes of pregnancies F I G U R E 2 The incidence of adverse events in pregnant women with or without premature ventricular contraction (PVC). The incidence of adverse events is higher in PVC cases (11.65%) than normal controls without PVC (2.93%). * indicates p < .05 vs. 65.69% ± 2.54%, p < .05). There were no other significant differences in the remaining baseline data between the two groups.

| Predictors of fetal and neonatal outcomes
The baseline characteristics of the study cohort suggested that the average burden of PVCs was related to adverse outcomes, so the burden of PVCs was further divided into low (<33rd percentile), middle (33rd-67th percentile) and high groups (>67th percentile) according to the percentile of the average PVC burden (Figure 3). The incidence of adverse events was significantly higher in the high-burden PVC group than in the low and middle group.
Univariable logistic regression analysis demonstrated that the LVEF, heart rate, bigeminy and burden of PVCs were associated with adverse fetal or neonatal outcomes among pregnant women with PVCs. Variables which were statistically significant in univariate analysis were further analyzed in multivariate logistic regression analysis; however, statistical significance was only evident for PVC burden in the multivariate logistic regression analysis (OR: 1.34, 95% CI [1.11-1.61], p < .05, Table 2).

| DISCUSSION
It is widely recognized that idiopathic PVCs are relatively benign in structurally normal hearts; 16 however, relatively few studies have evaluated the relationship between PVCs and the adverse outcomes of fetuses or neonates among pregnant women. 10,17 In this F I G U R E 3 Incidence of adverse fetal events in low premature ventricular contraction (PVC) burden (low), medium PVC burden (medium) and high PVC burden group (high). low PVC burden: < The 33rd percentile (<1.69%), medium PVC burden: The 33rd percentilethe 67th percentile (1.69%-4.77%), high PVC burden: > The 67th percentile (>4.77%), * indicates p < .05 study, we found that PVCs were associated with a higher frequency fetal/neonatal adverse events among pregnant women with a structurally normal heart. In fact, we further demonstrated that 'high' PVC burden is an important factor for predicting the incidence of adverse fetal/neonatal outcomes.
In our study, 1.68% pregnancies with structurally normal heart were found to have high PVC burden in 6148 consecutive enrolled pregnant women. And the rate of adverse fetal/neonatal events in women with PVCs was 11.65%. Similar results were reported in Tong C et al., 10 who enrolled 53 consecutive pregnancies with PVCs and found 13% adverse fetal and/or neonatal events. However, in the group of women without PVCs, the incidence rate was only 2.93% in our study, which was significantly lower than that previously reported in the group of pregnant women with or without structural heart disease. 10

| Study limitations
Although this study is a prospective, single-center study, the number of participants included is still relatively small; therefore, a large-scale and multicenter study should be performed in the future. Second, PVC origin was classified by a standard 12-lead ECG pattern in the study. Undeniably, changes in diaphragm position during pregnancy and cardiac rotation may cause misjudgment of the origin of PVCs. Third, because the PVC burden may fluctuate throughout the day and throughout stages of pregnancy, we applied the average PVC burden screened by Holter in each trimester to represent the PVC burden during pregnancy.
It is therefore suggested to use long-term wearable monitoring equipment to screen the PVC burden in the future. Finally, all patients underwent only two-dimensional echocardiography in the study, and cardiac magnetic resonance might be suitable for further study to eliminate the possibility of latent cardiomyopathy.
In conclusion, our study revealed that the prevalence of PVCs in pregnancy is higher than that in the normal population. Frequent PVCs have adverse effects on pregnancy, and the PVC burden might be the main factor associated with adverse pregnancy outcomes.  Abbreviations: CI, confidence interval; HR, heart rate; LVEF, left ventricular ejection fraction; OR, odds ratio; PVC, premature ventricular contraction.