Implications of fetal premature atrial contractions: systematic review

ABSTRACT Objective Fetal heart‐rate irregularities occur in 1–2% of pregnancies and are usually caused by premature atrial contractions (PAC). Although PAC are considered benign, they may be associated with cardiac defects and tachyarrhythmia. We aimed to determine the incidence of congenital heart defects (CHDs) and complications in fetuses with PAC. Methods This was a systematic review and meta‐analysis conducted in accordance with the PRISMA statement for reporting items for systematic reviews and meta‐analyses. MEDLINE and EMBASE were searched from 1990 to June 2021 to identify studies on fetuses with PAC. The primary outcome was CHD; secondary outcomes were complications using the endpoints supraventricular tachyarrhythmia (SVT), cardiac failure and intrauterine fetal demise. Meta‐analysis of proportions was performed, subdivided into high‐risk and low‐risk populations based on reason for referral. Pooled incidences with 95% CIs were calculated. Results Of 2443 unique articles identified, 19 cohort studies including 2260 fetuses were included. The pooled incidence of CHD in fetuses with PAC was 2.8% (95% CI, 1.5–4.1%), when 0.6% is the incidence expected in the general population. The pooled incidence of CHD was 7.2% (95% CI, 3.5–10.9%) in the high‐risk population and 0.9% (95% CI, 0.0–2.0%) in the low‐risk population. SVT occurred in 1.4% (95% CI, 0.6–3.4%) of fetuses diagnosed with PAC. Cardiac failure was described in 16 fetuses (1.4% (95% CI, 0.5–3.5%)), of which eight were CHD‐related. Intrauterine fetal demise occurred in four fetuses (0.9% (95% CI, 0.5–1.7%)) and was related to CHD in two cases. Conclusions Our findings suggest that the risk of CHD in fetuses with PAC is 4–5 times higher than that in the general population. CHD was present more frequently in the high‐risk population. Consequently, an advanced ultrasound examination to diagnose PAC correctly and exclude CHD is recommended. Complications of PAC are rare but can result in fetal demise, thus weekly fetal heart‐rate monitoring remains advisable to enable early detection of SVT and to prevent cardiac failure. © 2022 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.

PAC are usually benign and are thought to be the result of the physiological immaturity of the fetal conduction system 9 . As the majority of cases of PAC resolve during pregnancy or the first year after birth, they are of minor clinical consequence. Despite this, current guidelines still advise closely monitoring affected fetuses and their mothers 5,11,12 . The diagnosis of PAC is based on pregnant women undergoing advanced ultrasonography and weekly follow-up of the fetal heart rate by their obstetrician thereafter 11,13 . This protocol is based on the small possibility that PAC are associated with an underlying congenital heart defect (CHD), and that complications such as supraventricular tachyarrhythmia (SVT), cardiac failure or intrauterine fetal demise can occur 5,9,13,14 .
In the current era, one may wonder if this monitoring protocol is appropriate. Pregnant women undergo a standard second-trimester anomaly scan in many countries, and most fetuses have already been screened for CHD by the time the arrhythmia is detected 1,2,5 . Therefore, the added value of an additional ultrasound scan to detect CHD is unclear. Furthermore, the same generic monitoring protocol is applied in every mother and fetus with an irregular heart beat. It is not known whether the risk of complications depends upon the frequency or pattern of the PAC (solitary, bigeminy, trigeminy or blocked).
This systematic review and meta-analysis addressed the following research questions: What is the incidence of CHD in pregnancies affected by fetal PAC? What is the incidence of complications in these pregnancies, using SVT, cardiac failure and intrauterine fetal demise as endpoints? Finally, is the pattern of PAC related to complications observed?

METHODS
This systematic review was performed according to the preferred reporting items for systematic reviews and meta-analyses (PRISMA) statement 15 . The research protocol was prepublished in the PROSPERO international database for systematic reviews (CRD42020187873).

Search strategy and eligibility criteria
A medical information specialist (J.L.) conducted a search in MEDLINE (Ovid) and EMBASE from 1990 to 24 June 2021 to identify studies involving fetuses with PAC. The search strategy consisted of controlled terms (i.e. MeSH-terms in MEDLINE) and free-text terms for: (1) irregular heart rhythm including premature contractions and (2) the fetal period (antenatal/prenatal). To minimize the risk of missing studies, the search did not include pregnancy outcomes or study types. Animal studies, reviews and conference abstracts were excluded. No language restrictions were applied. The retrieved records were imported into EndNote X9.3.3 (Clarivate, Philadelphia, PA, USA) and duplicate records were removed. Cited and citing references of the included studies were screened for additional relevant publications. The complete search strategy is presented in Appendix S1.
We included prospective and retrospective cohort studies reporting on CHD or SVT as a complication of PAC. The focus of the article could vary (i.e. outcome of fetal arrhythmia or detection of CHD) but the outcomes for fetuses with PAC had to be mentioned separately from other described populations. Furthermore, the minimum cohort size was 25 subjects, the diagnosis had to be made with M-mode or pulsed-Doppler ultrasound and structural fetal echocardiography had to have been performed. We excluded studies published before 1990, those with an unclear method for diagnosing PAC and those with unclear follow-up.

Study selection
All abstracts were reviewed independently by two authors (B.B.B. and J.M.d.V.). The selection process was noted in Rayyan together with the reason for exclusion 16 . Disagreement over selection was resolved by consensus. The full text of potentially relevant articles was subsequently reviewed by the same two authors. Disagreement over inclusion was again resolved by consensus and, if necessary, by consultation with a third reviewer (E.P.). Articles published in a language other than English were translated by native speakers.
The quality of the included studies and the risk of bias were assessed independently by two authors (B.B.B. and J.M.d.V.), using the methodological characteristics of the Newcastle-Ottawa scale (NOS) for cohort studies 17 . Studies were scored on the three domains of NOS: selection, comparability and outcome. Additionally, we assessed eligible articles on the quality of described methods, cohort size and presentation of results.

Data extraction
After selection of the eligible articles, data extraction was completed by one reviewer (B.B.B.) using a predesigned data extraction form. The extracted study variables were: first author's name, country, year of publication, inclusion method, study period, PAC diagnostic method, cohort Fetal premature atrial contractions 723 size, mean gestational age at diagnosis, PAC pattern, incidence of foramen ovale aneurysm, resolution rate, gestational age at resolution, frequency of CHD, type of CHD, incidence of tachyarrhythmia transformation, incidence of cardiac failure, incidence of intrauterine fetal demise, prenatal cardiac drug treatment and other congenital anomalies.
We included subjects with PAC or extrasystoles not further specified, as the majority of extrasystoles are of supraventricular origin 3,6,14 . The pattern of PAC was also noted as intermittent, bigeminy, trigeminy or blocked.
The reason for referral and the baseline population from which the subjects were selected were evaluated. If a study selected only subjects referred for an irregular heart rate, or when the reason for referral was not mentioned, the population was characterized as low risk. When additional referral reasons were present, such as maternal medical conditions, use of medication or unfavorable obstetric or family history, the population was characterized as high risk.
The primary outcome was the incidence of CHD, defined as an abnormality in the structure of the fetal heart, in fetuses with PAC. An aneurysm of the foramen ovale was not classified as a heart defect and was described separately. Persistent ductus arteriosus and Type II atrial septal defects were not considered as CHDs. The secondary outcome was the incidence of fetal complications, using SVT, cardiac failure and intrauterine fetal demise as endpoints. SVT was defined as a heart rate above 180 bpm 2 . Cardiac failure was defined as the inability of the fetal heart to deliver adequate blood flow to organs with the accompanying sonographic findings of cardiomegaly and cardiac effusion and/or fetal hydrops 18 . Intrauterine fetal demise was defined as demise from 16 weeks' gestation until birth.
If the methods or results remained unclear, we approached the authors by e-mail and/or discussed the diagnosis or complication with our pediatric cardiologist (S.A.C.).

Statistical analysis
When at least 10 studies had data on primary or secondary outcomes, we performed a meta-analysis of proportions, using the meta and metafor packages in R version 3.6.1 (R Foundation for Statistical Computing, Vienna, Austria) 19,20 . For each included study we calculated the incidence of primary and secondary outcomes with the corresponding 95% CIs. An average estimate of the incidence was calculated using a random-effects model with an exact binomial CI. A continuity correction was applied if a study had either zero or all events 20 . The I 2 test of heterogeneity was expressed as a percentage ranging from 0% to 100%, with higher values indicating more heterogeneity. Heterogeneity was considered high if I 2 exceeded 50%.
To put the incidence of CHD into perspective, we compared the pooled incidence in fetuses with PAC to that in the general population, using the large registry database of EUROCAT 1999-2008 21 .

Study selection and characteristics
A total of 2443 articles were identified after duplicates were removed, of which 2350 were excluded based on the title and abstract. Of the remaining 93 articles, 57 were excluded following full-text assessment because they did not cover the correct population, the method or follow-up was unclear, the cohort contained fewer than 25 subjects, the cohorts overlapped or the desired outcome was not presented. For a further 17 articles, the full text was not available and the authors could not be reached. Therefore, 19 studies were eventually included in the systematic review and meta-analysis ( Figure 1).
All studies scored four or five stars for risk of bias according to NOS 17 , with two out of four stars for the selection criteria and two or three out of three stars for the outcome criteria (Table S1). The comparability section was not included in our quality assessment since none of the studies used a control group.
The results of the 19 included studies involving 2260 fetuses are presented in Table 2. Sample size ranged from 43 to 299 fetuses. Gestational age at diagnosis was described for fetuses with PAC in six studies, with a mean in the third trimester of between 28 and 33 weeks 22,27,29,36,37,39 . Additionally, six studies described the pattern of PAC 22,27,34,36,38,40 . Of the 733 fetuses in these studies, the pattern was bigeminal or trigeminal in 44 (6.0%). The studies of Maragnes et al. 38 and Wloch et al. 40 accounted for 20 and 15 of these fetuses, respectively. Blocked PAC were noted in 42/733 (5.7%) fetuses, the majority of which were described by Boldt et al. 22 and Oberhänsli et al. 27 , with 21 and 15 fetuses, respectively.
The types of CHD detected are shown in Table 3. Among 108 fetuses with PAC and CHD, isolated ventricular septal defect was the most commonly described anomaly (36 cases (33.3%)). One heart defect remained unspecified despite contacting the author 24 .
The incidence of CHD in the general population lies between 0.6% and 1.1% 21,41,42 . We used the large cohort of EUROCAT 1999-2008 as a historical benchmark for the incidence of CHD in the general population 21 . Compared with the general population, the incidence of CHD was significantly higher in the whole PAC population (2.8% vs 0.6%) and high-risk population (7.2% vs 0.6%), but approximately equal in the low-risk population (0.9% vs 0.6%).

Figure 2
Forest plots showing incidence of congenital heart defect in fetuses with premature atrial contractions, subdivided into high-risk and low-risk populations based on reason for referral. Only first author is given for each study. Table 3 Types of congenital heart defect (CHD) reported by 12 studies in which CHD was associated with premature atrial contractions

Figure 3
Forest plot showing incidence of supraventricular tachyarrhythmia (SVT) in fetuses with premature atrial contractions. Studies that did not report on SVT were excluded from this meta-analysis. Only first author is given for each study.
extracardiac anomalies. Importantly, only one neonate died as a result of tachyarrhythmia and cardiac failure 28 .

DISCUSSION
This systematic review and meta-analysis found that the estimated pooled incidence of CHD in pregnancies complicated by fetal PAC was 2.8%, which is 4-5 times higher than the expected incidence in the general population (0.6%) 21 . In the low-risk subgroup, the incidence of CHD was 0.9% and comparable with that of the general population. Complications of fetal PAC were rare, with an incidence of 1.4% for SVT, 1.4% for cardiac failure and 0.9% for intrauterine fetal demise.

Strengths and limitations
To the best of our knowledge, this is the first review describing the incidence of CHD in fetuses with PAC. Our intentionally broad search resulted in articles with different focuses but which all included fetuses with PAC as a (sub)group of their study population. The absence of language restriction in the search strategy made it unlikely that important literature was missed. The exclusion of publications before 1990 contributed to greater comparability and reliability of the ultrasound methods used. Additionally, we analyzed complications separately so that we could review specific parts of the current guidelines for the management of fetal PAC. The study has several limitations which should be acknowledged. First, PAC are an intermittent phenomenon, therefore the true incidence will be higher than the reported incidence, and by inference the true complication rate will also be lower than the reported rate. Second, the study populations differed over time, as CHD screening was initially available only to high-risk women, so earlier studies included predominantly higher-risk populations. This selection bias could explain in part the heterogeneity of the results that was demonstrated in the subgroup analysis. In the low-risk population, the incidence of CHD in fetuses with PAC was found to be comparable with that of the general population. This might suggest that pregnancies that are uneventful except for the occurrence of fetal PAC do not have a higher risk for CHD. As our allocation of fetuses to the low-risk group when the population type was not defined may have influenced these findings, we performed a sensitivity analysis with the exclusion of those four studies and found that the results were unaffected.
Third, the timing of the diagnosis of CHD is important when interpreting our results. The most interesting cases are those in which CHD was diagnosed after the PAC, as PAC may be the first sign of CHD, thus additional ultrasonography would be indicated. This is particularly relevant for defects observed later in pregnancy (valvular lesions, cardiomyopathies) 5 . Unfortunately, all the studies had a retrospective design and did not report the timing of the CHD diagnosis relative to that of PAC. Consequently, we were unable to assess the value of additional ultrasonography for the detection of CHD in fetuses with PAC. However, in the studies that mentioned the timing of the diagnosis of PAC, it occurred between 28 and 33 weeks' gestation. As first-and second-trimester routine scans are currently offered to pregnant women in many countries, most fetuses would already have been screened for CHD by the time PAC were diagnosed.
Fourth, some cohorts were too small to reliably detect complications of PAC. The incidence of CHD was highest in the small cohort of Oberhänsli et al. 27 (36.5%), and the three largest cohorts reported incidences between 0% and 1.3% 23,32,34 . As no individual study was large enough to reflect a reliable incidence of CHD, we made estimates in our meta-analysis.
Finally, subtypes of PAC that may be associated with a higher incidence of CHD and SVT (blocked or bigeminal) 14,[43][44][45] were not mentioned in most studies. Reporting bias complicated interpretation of the subtypes risk analysis. In studies with a relatively large number of blocked PAC, the incidence of CHD was also high 22,27 ; this requires further investigation. Association of blocked PAC with a higher risk of SVT could not be confirmed owing to the low number of reported cases [43][44][45] . Additionally, PAC were not differentiated from ventricular extrasystoles in all studies, which could have introduced further bias, as ventricular extrasystoles have a higher complication rate than do PAC 2,3,9 .

Interpretation
In this review, fetal PAC were mostly benign and the majority resolved spontaneously during pregnancy. This is consistent with current fetal medicine guidelines 5,46 . We endorse the recommendation of performing advanced ultrasonography to distinguish PAC correctly from arrhythmias with a higher complication rate and to also (re)assess cardiac anatomy, as we found a higher incidence of CHD among fetuses with PAC. Our findings also support the theory that an aneurysmal foramen ovale may mechanically trigger rhythm disturbances, as it was described in 11.2% of fetuses with PAC 47,48 . Unfortunately, the findings of this review do not help us to determine whether ultrasound assessment is still required in cases in which the first-/second-trimester scan was favorable. A specific recommendation for ultrasonography in fetuses with PAC that are frequent, persistent or follow a specific pattern cannot be issued based on our results.
We found the complication rate of fetal PAC to be low. Transformation of PAC to SVT was observed in 1.4% (pooled) of the fetuses in this review, which is comparable with the 0.5-1.0% rate cited in the statement of the American Heart Association 5 , and the advice to perform regular auscultation therefore seems reasonable. The suggestion that certain patterns of PAC increase the risk of SVT and should be monitored more intensively 2,5 cannot be confirmed or denied by our results, therefore physicians should continue monitoring all fetuses with PAC 3 . Our current protocol to perform weekly monitoring until resolution of PAC is extensive but justified, as it may prevent serious complications 11 . A large prospective cohort study should be conducted in order to determine which PAC patterns warrant close monitoring and to evaluate the additional value of echocardiography after a normal second-trimester anomaly scan. This would provide the opportunity to truly assess the value of additional echocardiography and to identify risk factors for complications.

Conclusions
This systematic review supports the current guidelines recommending advanced ultrasonography when fetal PAC are detected. Moreover, follow-up fetal monitoring should be performed locally for all cases, with clear instructions on when to refer patients back to a fetal medicine unit. However, the low incidence of CHD in a low-risk population and low complication rate of fetal PAC should be communicated positively to expectant parents and referring physicians.

SUPPORTING INFORMATION ON THE INTERNET
The following supporting information may be found in the online version of this article: Appendix S1 Complete search strategy Table S1 Quality assessment of the included studies using the Newcastle-Ottawa scale for cohort studies Figure S1 Forest plot showing incidence of cardiac failure in fetuses with premature atrial contractions. Studies that did not report on cardiac failure were excluded from this meta-analysis. Only first author is given for each study.

Figure S2
Forest plot showing incidence of intrauterine fetal demise (IUFD) in fetuses with premature atrial contractions. Studies that did not report on IUFD were excluded from this meta-analysis. Only first author is given for each study.
A video abstract of this article is available online.