We Only Find What We Look For: Fetal Heart Rate and the Diagnosis of Long QT Syndrome

Abstract

Long QT syndrome (LQTS), an inherited channelopathy, is a common cause of arrhythmic death in infants, children and young adults. Although many LQTS genes have been identified, most (~75%) of LQTS mutations are found in KCNQ1, KCNH2 or SCN5A. In most cases, treatment for LQTS is successful and modifies the risk of life-threatening arrhythmias; thus, making the correct diagnosis is important. The diagnosis of LQTS is made by the measurement of a prolonged QT interval on the standard ECG; family history or characteristic arrhythmia features are used to strengthen the diagnosis and genetic testing confirms the diagnosis.

Before birth, however, LQTS is difficult to diagnose because fetal echocardiography, the stalwart in the diagnosis of fetal arrhythmias, cannot measurer the QT interval or detect T-wave alternans and fetal electrocardiography (ECG) is not feasible. Although fetal magnetocardiography (fMCG) has successfully recorded fetal ECG and identified fetal LQTS¹ it is available at only a few centers world-wide. Recent studies have identified torsades de pointes (TdP) and 2° atrioventricular (AV) block as signature fetal LQTS arrhythmias^{1, 2}: These arrhythmias are easily recognized as abnormal but occur in only about 25% of fetuses with LQTS. The arrhythmia most commonly associated with fetal LQTS is sinus bradycardia. Unfortunately, fetuses with sinus bradycardia often escape recognition because their heart rate and rhythm are not appreciated to be abnormal since the obstetrical definition of sinus bradycardia is limited to a fetal heart rate (FHR) of ≤110 bpm at every gestational age (GA)³. This GA independent definition of fetal bradycardia has been difficult to challenge despite the known inverse relationship between FHR and gestational age in normal fetuses⁴ and a previous study of in which only 15% of LQTS FHRs were ≤110 bpm².

Why is it important to suspect the diagnosis of LQTS before birth? One reason is to decrease prenatal morbidity. Fetuses with LQTS are often delivered prematurely because the LQTS rhtyhm phenotype of sinus bradycardia or ventricular bradycardia due to 2° AVB is misinterpreted as fetal distress⁵. Alternatively, the treating physician may be unaware even the hydropic fetus with ventricular tachycardia can be successfully treated in utero⁶. The LQTS fetus may be at increased risk of ventricular tachycardia if low maternal magnesium levels remain untreated or the mother is treated with QT prolonging medications common in obstetrical practice such as oxytocin or ondansetron. Other important reasons to suspect a prenatal diagnosis of LQTS are the genetic autopsy findings in unexplained fetal demise and sudden infant death (SIDS) victim: genes associated with ion channel dysfunction or de novo channelopathy gene mutations have been found in 8.8% of the former and about 10% of the latter^{7, 8}. Finally, if LQTS is confirmed after birth in a fetal proband, cascade screening can identify affected but unsuspecting family members².

Dr. Annika Winbo and her colleagues have published extensively on the genotype and postnatal phenotype from Swedish populations with 2 founder KCNQ1 mutations^9–13. The two KCNQ1 mutations, R518X and Y111C, are thought to be loss-of-function mutations. The annual incidences of cardiac arrhythmia events for postnatal subjects heterozygous for R518X and Y111C is only 0.04% and 0.05%⁹, which is much lower than in other heterozygous nonsense mutations of KCNQ1¹⁴, other KCNQ1 founder populations¹⁵ and mutations in other LQTS genes. In this edition of Circulation Arrhythmia and Electrophysiology, Dr. Winbo and colleagues extend our knowledge of the 2 founder populations by defining the prenatal arrhythmia phenotype of the KCNQ1 R518X and Y111C mutations¹⁶.

There are 2 important findings in Winbo’s current study¹⁶. First, existing obstetrical FHR criteria for sinus bradycardia, that is a FHR ≤110 bpm at any time during gestation, did not define the rhythm phenotype of the KCNQ1 study cohort. Based on the accepted gestational age independent criteria for bradycardia, only 3/110 mutation positive fetuses would have been suspected to have LQTS. The second important finding is that fetal LQTS demonstrates a mutation dose response for disease severity, including the degree of bradycardia and the likelihood of adverse cardiac events. Thus, not only does FHR differ between mutation carriers and non-carriers, but FHR correlates inversely with the number of KCNQ1 mutations and the postnatal QTc duration, and directly with mutation functionality¹⁶.

Rather than the accepted definition of bradycardia (FHR of ≤110 bpm), the average 3rd trimester FHR of KCNQ1 R518X and Y111C mutation-carriers was 2 standard deviations below the mean for mutation negative fetuses from the same population, that is ≤133 bpm¹⁶. These higher than expected FHRs for LQTS subjects are similar to the FHRs of LQTS subjects in a previous study which found that 67% of the heart rates of LQTS fetuses were ≤ 3^rd percentile for GA and the cut-off for the 3^rd trimester was a ≤128 bpm². The reason that the FHR cut-off was lower in Mitchell’s study may be because the cohort included KCNH2, SCN5A and uncharacterized LQTS mutations in addition to KCNQ1 mutations².

Only 15% of LQTS FHRs are reported as ≤ 110 pm², but a FHR≤ 110 bpm predicts a severe LQTS phenotype with subjects at high risk for adverse cardiac events¹⁶. Fetuses with LQTS and sinus bradycardia of ≤110 bpm in the 3^rd trimester are more likely to have SCN5A R1623Q mutations, uncharacterized mutations or double mutations of KCNQ1^2,17,18,16, very prolonged pre and postnatal QTc intervals^1,16,18 and cardiac arrest. In SCN5A R1623Q and some uncharacterized mutations, FHR of ≤ 110 bpm are seen even earlier than the 3^rd trimester^2,18. Fetuses with double mutations in the Swedish founder KCNQ1 population may also have FHR ≤110 earlier in gestation, but his has not yet been evaluated. Whether other mutations in the known LQTS genes may incur a similar high risk profile based on FHRs ≤110 bpm is unknown and will require further research.

The second important finding in Winbo’s current study is that even before birth, the LQTS phenotype, including both FHR and the risk for adverse cardiac events depends on mutation status and functionality. The prenatal and postnatal rhythm phenotype of individuals heterozygous for a single KCNQ1 R518X and Y111C mutation was mild and low risk: The average FHR was 134 ± 8 bpm and 2/93 fetuses had non-sustained FHR >200 bpm (presumed ventricular tachycardia). On the other hand, the average 3^rd trimester FHRs of double mutations carriers was 111±6 bpm, 1/13 fetuses had presumed ventricular tachycardia and 12 adverse cardiac events including appropriate ICD discharge, TdP and sudden death were reported after birth. These are important new findings in light of previous publications suggesting the fetal/neonatal KCNQ1 phenotype is low risk for serious arrhythmias^2,17,18.

Winbo and her colleagues raise important questions about the role of FHR in detection of LQTS¹⁶. Can FHR be used to risk stratify the LQTS fetus, or is it just a marker for LQTS? The answer is, both. A FHR ≤ 110 bpm correlates with a higher risk for perinatal life-threatening cardiac events, and LQTS mutations not commonly seen in adulthood, such as SCN5A R1623Q, uncharacterized mutations, dominant negative single KCNQ1 mutations and double KCNQ1 mutations^{1, 2,16–18}. As a marker for LQTS, a mean FHR of ≤ 133 bpm in the 3^rd trimester has a high specificity (>97%) but low sensitivity (<50%)¹⁶. It is not known if FHR at a different point in gestation might increase the sensitivity of FHR for LQTS detection.

How can we use this data in clinical practice? Can we extrapolate the 3^rd trimester LQTS FHR results to the low risk population? Do we ask our obstetrical and pediatric colleagues to refer newborns with a mean FHR ≤133 bpm during the 3^rd trimester for a postnatal ECG, knowing the sensitivity of that FHR in LQTS detection is < 50%¹⁶? Or do we evaluate only those newborns with a mean 3^rd trimester FHR ≤ 110 bpm?

Additional data are clearly needed. A reasonable approach would be to build on Winbo’s important findings¹⁶ and prospectively evaluate not only 3^rd trimester FHR in families with a KCNQ1 mutations but to develop a FHR/GA profile in families with KCNH2, SCN5A and other KCNQ1 mutations. Once defined, the FHR/GA profile for the common LQTS mutations could alert obstetrical care providers to the possibility of LQTS and the need for postnatal ECG testing. Incorporating FHR/GA data in neonatal LQTS ascertainment may improve the cost benefit ratio of neonatal ECG screening in order to identify those at risk for cardiac arrest^{19, 20}.

In summary, Winbo and her colleagues have made an important contribution to our understanding of the 3^rd trimester FHR and rhythm phenotype of specific KCNQ1 mutations¹⁶. Until technological advances allow for accurate and readily available fetal ECG recordings, the suspicion of neonatal LQTS rests on critical evaluation of the FHR, and recognizing what is an abnormal FHR. The failure to suspect LQTS before birth misses a critical window of opportunity for anticipatory care and treatment for LQTS subjects of all ages.