BACKGROUND: Left axis deviation (LAD) discovered in children via electrocardiogram (ECG) is uncommon but can be associated with heart disease (HD). The optimal diagnostic approach in a seemingly healthy child with LAD is unclear. We sought to better stratify which patients with LAD but without previously known HD may warrant additional workup.
METHODS: A retrospective chart review was performed to identify patients ≥1 to <18 years of age with LAD (QRS frontal plane axis 0 to −90) on an ECG between January 2002 and December 2014. Patients with known HD before their initial ECG were excluded.
RESULTS: Overall, 296 patients were identified (n = 181 [61%] male; mean age: 10.8 ± 4.6 years; mean QRS axis: −24 ± 22°). An echocardiogram was performed in 158 (53%) patients, with 24 (15%) having HD. Compared with those with an echocardiogram but without HD (n = 134), patients with HD had a more negative mean QRS axis (−42 vs −27°; P = .002) and were more likely to have a QRS axis ≤−42° (58% vs 26%; P = .003), ECG chamber enlargement or hypertrophy (38% vs 5%; P < .0001), and abnormal cardiac physical examination findings (75% vs 8%; P < .0001).
CONCLUSIONS: LAD discovered in isolation in the asymptomatic pediatric patient may not necessitate further cardiovascular investigation. Clinicians should consider obtaining an echocardiogram in patients with LAD and ECG cardiac chamber enlargement or hypertrophy, a QRS axis ≤−42°, and/or the presence of abnormal cardiac physical examination findings.
- ASD —
- atrial septal defect
- CHD —
- congenital heart disease
- DCM —
- dilated cardiomyopathy
- ECG —
- HCM —
- hypertrophic cardiomyopathy
- HD —
- heart disease
- LAD —
- left axis deviation
- MVP —
- mitral valve prolapse
- NPV —
- negative predictive value
- OR —
- odds ratio
- PPV —
- positive predictive value
- RCM —
- restrictive cardiomyopathy
- RR —
- relative risk
- TTE —
- transthoracic echocardiogram
- VSD —
- ventricular septal defect
What’s Known on This Subject:
Left axis deviation (LAD) is associated with particular forms of cardiac pathology in children, with many cases identified during infancy. The prevalence of heart disease (HD) in asymptomatic children with LAD is unknown, and the additional workup required is unclear.
What This Study Adds:
We identify risk factors that make the presence of HD more likely in children with LAD but without previously known HD. These risk factors aid physicians in determining which patients may require further workup if LAD is detected incidentally.
Physicians caring for children may obtain an electrocardiogram (ECG) for any number of reasons during the life of a child. When an abnormal QRS axis is noted, this may implicate underlying congenital heart disease (CHD). Age-based pediatric normative values for the mean frontal plane QRS axis have been established previously.1 Left axis deviation (LAD) occurs when the summation of ventricular electrical forces results in a mean frontal plane QRS vector that is more negative than normative values, arriving via a counterclockwise loop in the vectorcardiogram. LAD is a relatively uncommon finding in the pediatric population but one that has been associated with specific forms of cardiac pathology, namely the following: tricuspid atresia,2 atrioventricular septal defects,3 and Wolff-Parkinson-White syndrome.4,5 LAD in an infant is highly suspicious for CHD, and an echocardiogram is generally warranted, but there is a paucity of literature in which the implications of LAD in the asymptomatic child outside of infancy have been outlined. This lack of data could result in over- or underuse of further testing to exclude heart disease (HD). Furthermore, the prevalence of previously unrecognized HD in this patient population is uncertain, making it difficult to know if further workup is warranted. In this study, we aimed to better stratify which children with LAD but without previously known HD may warrant further cardiac workup.
A retrospective review of Mayo Clinic’s electronic medical record was performed to identify all patients ≥1 to <18 years of age who had ECG evidence of LAD from January 2002 to December 2014. Patients with known HD (including channelopathies) before their initial ECG were excluded from this study, as were patients with left bundle branch block, paced ventricular rhythm, LAD occurring in the context of wide complex tachycardia, and LAD occurring after orthotopic heart transplant. In patients with more than 1 ECG revealing LAD, the initial ECG revealing LAD was used for analysis.
Standard 12- or 15-lead ECGs were obtained at the discretion of the ordering provider caring for each individual patient. Reported mean fontal plane QRS values were calculated automatically by the MUSE Cardiology Information System (GE Healthcare, Chicago, IL). ECGs were read by a pediatric cardiologist using age-based normative QRS axis values reported by Davignon et al.1 LAD was defined in our study as a QRS axis ≤0 and ≥−90°. Transthoracic echocardiograms (TTEs) were obtained at the discretion of the provider caring for each individual patient because screening echocardiograms are not performed universally at our institution for children with newly diagnosed LAD. TTEs were performed at Mayo Clinic during the study period and were conducted on commercially available cardiac ultrasound systems by employing an established CHD protocol.
We subsequently examined demographic variables, the mean frontal plane QRS axis, ECG evidence of cardiac enlargement or chamber hypertrophy, cardiac physical examination findings, echocardiographic findings (when available), patient follow-up duration, and outcomes. In a subset of patients in whom HD was discovered, we also reviewed the type of HD present and the indication for the ECG. This protocol was approved by the Mayo Clinic Institutional Review Board.
Continuous variables were expressed as a mean ± SD and ordinal variables as total number and percentage. Student’s t test or 2-tailed Fisher’s exact test were performed to compare continuous and ordinal values between groups, respectively, as appropriate. A P value ≤ .05 was deemed statistically significant. All data analyses were performed by using JMP (SAS Institute, Inc, Cary, NC).
Overall, 296 patients (mean age: 10.8 ± 4.6 years; range: 1–17 years; 181 male [61%]) met inclusion criteria for our study (Table 1). The mean frontal plane QRS axis for this group was −24 ± 22° with a mean follow-up duration of 3.9 ± 3.7 years. An echocardiogram was performed in 158 (53%) patients (Table 1). Compared with the 138 patients without an echocardiogram, these patients were significantly younger in age (10.2 vs 11.4 years; P = .02), had a more negative QRS axis (−29.0 vs −18.7°; P < .0001), and had a longer follow-up period (4.6 vs 3.2 years; P = .001). Because we were unable to verify the presence or absence of HD in those without a TTE, our subset analysis was focused on patients who received a TTE.
Of the 158 patients who were LAD-positive who had a TTE, 24 (15%) patients were determined to have some form of HD (Table 2). Detected forms of HD included secundum atrial septal defect (ASD) (n = 4), dilated cardiomyopathy (DCM) (n = 4), hypertrophic cardiomyopathy (HCM) (n = 4), partial atrioventricular canal defect (n = 4), and 1 instance of each of the following: bicuspid aortic valve, congenitally corrected transposition of the great arteries, left ventricular noncompaction cardiomyopathy, mitral valve stenosis, mitral valve prolapse (MVP), restrictive cardiomyopathy (RCM), membranous ventricular septal defect (VSD), and flail tricuspid valve leaflet (traumatic). Indications for the original ECG as documented by the patient’s care provider included murmur (n = 10), arrhythmia and/or tachycardia (n = 3), screening for familial HCM (n = 3), dyspnea (n = 2), and 1 instance of each of the following: chest pain, renal failure, scleroderma, stroke, syncope, and trauma. Abnormal cardiac physical examination findings were documented in 18 (75%) patients with HD. Specific heart defects that were detected on TTE in patients without a concomitantly documented physical examination abnormality included bicuspid aortic valve, secundum ASD, DCM, HCM, MVP, and RCM (1 patient each).
Compared with those with a TTE but no HD (n = 134), patients with HD had a significantly more negative mean QRS axis (−41.9 vs −26.7°; P = .002) (Fig 1). Although patients with HD were more likely to have a mean QRS axis ≤−30° (67% vs 39%; P = .01), a receiver operating characteristic curve suggested that a mean QRS axis ≤−42° may provide enhanced patient discrimination. Patients with HD were significantly more likely to have a mean QRS axis ≤−42° (58% vs 26%; P = .003) (sensitivity: 58%; specificity: 74%; positive predictive value [PPV]: 29%; negative predictive value [NPV]: 91%), evidence of cardiac chamber enlargement or hypertrophy on an ECG (38% vs 5%; P < .0001) (sensitivity: 38%; specificity: 95%; PPV: 56%; NPV: 89%), and abnormal cardiac examination findings as documented by the physician caring for them (75% vs 8%; P < .0001) (sensitivity: 75%; specificity: 92%; PPV: 62%; NPV: 95%) (Fig 2). Patients who had none of these 3 variables present (axis ≤−42°, ECG enlargement or hypertrophy, or abnormal cardiac examination) were unlikely to have HD (8% HD versus 65% without; P < .0001) (Table 3). The NPV in this case was 98%, with a sensitivity of 92%. There was no statistical difference between groups of patients with just 1 positive variable (38% HD versus 31% without; P = .6). Patients with 2 positive variables (29% HD versus 3% without; P = .0002) were significantly more likely to have HD (specificity: 97%; PPV: 64%; relative risk [RR]: 9.8; odds ratio [OR]: 13.4), as were patients with all 3 positive variables (25% HD versus 1% without; P < .0001) (specificity: 99%; PPV: 88%; RR: 33.5; OR: 44.3).
The mean follow-up duration for patients with HD was significantly longer than in those without HD (6.5 vs 4.3 years; P = .006). However, 77% (n = 103) of the patients without HD were followed for ≥1 year, and none of these patients subsequently developed HD or died of a cardiac cause. Likewise, none of the patients (n = 138) without a TTE subsequently developed HD or died of a cardiac cause in 3.2 ± 3.5 years of follow-up. Two (9%) patients within the HD patient subset died of cardiac causes during follow-up. One patient with RCM died of sudden cardiac death at the age of 16 after receiving an orthotopic heart transplant. The second patient had HCM and died of a cardiac dysrhythmia at 21 years of age.
The incidence of LAD in the pediatric population has not been well established, but previous estimates range from 1% to 2%.5–7 Patients with LAD and clinically significant HD are often identified during infancy. LAD discovered incidentally in the otherwise healthy child outside of infancy is less common, and the prevalence of HD in this patient population is unknown. Pediatricians and cardiologists alike must determine on an individual basis if the clinical scenario is suggestive of a child with previously unrecognized HD or simply a variant of normal. Objective data discriminating these 2 groups are lacking. Therefore, we evaluated a cohort of children without previously recognized HD who had LAD identified on an ECG at our institution. Ultimately, we determined that LAD found in isolation may be a variant of normal, but in certain patients, specific risk factors may suggest unidentified HD.
Across the spectrum of ages, LAD may result from multiple etiologies, including the following: tricuspid atresia,2 atrioventricular septal defects,3 Wolff-Parkinson-White syndrome,4,5 VSDs,5,8,9 disruption of the left anterior bundle branch,5,10 and natural aging.11 LAD may also be a normal variant in a small subset of children who seemingly experience a benign clinical course on the basis of short-term follow-up data.6,12 Out of 158 patients with LAD and a TTE in our study, 24 (15%) patients had some form of HD. Four of the identified patients had partial atrioventricular canal defects, in which LAD is expected secondary to displacement of the A-V node posteriorly and His bundle inferiorly.3 Although LAD has been described in patients with both primum and secundum ASDs13 and congenitally corrected transposition of the great arteries,14 it is not classically associated with the other defects that were appreciated in our study. LAD in these patients may reflect a normal, coincidental variant given that anatomically the conduction system should not be disrupted or displaced from the usual position.
Three objective clinical features were identified as being discriminators of patients with HD compared with those without: degree of LAD (≤−42°), presence of ECG chamber enlargement or hypertrophy, and abnormal cardiac physical examination findings. The latter 2 variables were the most sensitive discriminators of HD when taken on their own. However, the presence of any 1 variable found in isolation, statistically, did not predict the existence of HD in patients with LAD (PPV: 18%; RR: 1.2). Patients with a combination of any 2 or all 3 positive variables were 10 to 33 times more likely to have HD, and additional cardiovascular workup would be warranted based on these data. Conversely, patients with none of the 3 aforementioned positive variables were extremely unlikely to have HD, and additional cardiovascular workup may not be of benefit.
Two patients with HD had none of the 3 positive variables. One of the patients was being screened for palpitations and was found to have DCM, whereas the other was found to have asymptomatic HCM after being screened for familial HCM. The single patient without HD but who tested positive for all 3 variables was a 3-year-old with an innocent murmur who met right ventricular hypertrophy criteria on the basis of S-wave depth in lead V6, with a QRS axis of −53°. Nevertheless, the most compelling conclusion from our data is that patients with no positive variables are unlikely to have HD (NPV: 98%).
A previous study by Ravi et al13 revealed that consultation with a pediatric cardiologist was the most cost-effective initial step in further evaluation of patients with LAD. Echocardiograms were found to only be cost-effective in patients with LAD in the presence of abnormal cardiovascular physical examination findings. In our study, the presence or absence of abnormal cardiovascular examination findings was based on documentation from any provider who saw the patient and ordered the ECG, many of whom were not pediatric cardiologists. The following 6 cardiovascular abnormalities were not identified or suspected on the basis of the documented clinical examination: bicuspid aortic valve, secundum ASD, DCM, HCM, MVP, and RCM. It’s possible that some of these lesions could have been detected by a provider more adept at detecting subtle nuances of the cardiac examination. However, despite these patients with HD having no detected physical examination abnormalities, the NPV of having none of the 3 identified variables was still high (98%). Therefore, with these data we suggest that it may not be absolutely necessary to have a pediatric cardiologist examine each patient with LAD to determine if an echocardiogram is justified. It should also be noted that LAD in and of itself has not been shown to predispose an individual to life-threatening arrhythmia or sudden cardiac death, so additional cardiovascular workup or referral is likely unnecessary in the absence of additional cardiovascular risk factors or other clinical concerns.
As previously stated, Ravi et al13 concluded that the next most cost-effective step in the evaluation of patients with recognized LAD would be referral to a pediatric cardiologist who may be more judicious in determining if an echocardiogram is required. They did not, however, take into account additional identifiable risk factors (as we did) that may place the patient at an increased risk for structural HD, thus increasing the pretest probability of the echocardiogram. With our data, we suggest that a child with LAD and any concomitant combination of ≥2 positive variables (degree of LAD [≤−42°], presence of ECG chamber enlargement or hypertrophy, and abnormal cardiac physical examination findings) would benefit from an echocardiogram to rule out the presence of structural HD, particularly in the context of abnormal physical examination findings. The decision to obtain the echocardiogram first or refer directly to a pediatric cardiologist and then get an echocardiogram will vary depending on multiple factors, including access to testing and/or care and the comfort level of the provider following up on the results. If a referral is made first, cardiologists should use their clinical judgment regarding the need for an echocardiogram in patients with LAD, although an echocardiogram should be strongly considered in patients with ≥2 positive variables.
Our study was limited by a small sample size, although to our knowledge, this study represents the largest of its kind in the medical literature. Other limitations included selection bias because all patients presented to our medical center under the assumption that some type of medical problem was present, although the exact indication for obtaining the ECG was not always known. Patients were also selected to have ECGs and echocardiograms on the basis of the suspicion of disease from the physicians caring for them. An ascertainment bias was also present because we did not prove or disprove the presence of HD in those patients with LAD and no echocardiogram (n = 138), so they were not included in our analysis of variables. Secondary to these aforementioned reasons, and in conjunction with the fact that we excluded all patients with known HD from our study, our data should not be interpreted as an estimate of LAD prevalence in children.
LAD discovered in isolation in the asymptomatic pediatric patient may not necessitate further cardiovascular investigation, particularly if no additional findings are present. Clinicians should consider obtaining an echocardiogram in patients with LAD who exhibit at least 2 of these 3 variables: (1) abnormal cardiac physical examination findings, (2) ECG cardiac chamber enlargement or hypertrophy, and/or (3) a QRS axis ≤−42°.
- Accepted December 5, 2017.
- Address correspondence to Philip L. Wackel, MD, Mayo Clinic, 200 First St SW, Gonda 6, Rochester, MN 55905. E-mail:
FINANCIAL DISCLOSURE: Other than those already listed under Potential Conflict of Interest, the other authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: Dr Ackerman is a consultant for Audentes Therapeutics, Boston Scientific, Gilead Sciences, Invitae, Medtronic, MyoKardia, and St. Jude Medical. Dr Ackerman and Mayo Clinic have a potential equity and/or royalty relationship with AliveCor, Blue Ox Health, and StemoniX. However, none of these entities have participated in this study in any way; the other authors have indicated they have no potential conflicts of interest to disclose.
- Feldt RH,
- DuShane JW,
- Titus JL
- Allen HD,
- Shaddy RE,
- Penny DJ,
- Feltes TF,
- Cetta F
- Wershing JM,
- Walker CH
- Farrú-Albohaire O,
- Arcil G,
- Hernández I
- Shaw NJ,
- Godman MJ,
- Hayes A,
- Sutherland GR
- Grant RP
- Copyright © 2018 by the American Academy of Pediatrics