Fragmented QRS (fQRS) results from localised disruption of conduction due to myocardial fibrosis and scar and is evident on 12-lead ECG. Inhomogeneous myocardial electrical activation consequently manifests as spikes within the QRS complex. fQRS has been shown to be associated with increased mortality, ventricular arrhythmias and major adverse coronary events across a spectrum of pathologies ranging from coronary artery disease, cardiomyopathies (dilated cardiomyopathy, hypertrophic cardiomyopathy and left ventricular non-compaction) to acute pulmonary embolism.1 fQRS has also been defined as a marker of arrhythmogenic right ventricular (RV) dysplasia, long QT and Brugada syndrome.

fQRS morphologies are not only confined to acquired and hereditary cardiac conditions, but are also present on the surface ECGs of patients with congenital heart disease—including tetralogy of Fallot (TOF) and Ebstein's anomaly. fQRS has been reported to occur predominantly in the right and mid-praecordial leads in patients with repaired TOF (rTOF), with fragmented intracardiac electrograms previously demonstrated through invasive endocardial mapping.2 Until now, two cross-sectional studies of adult patients with rTOF have elucidated associations between fQRS and the extent of RV dilatation, dysfunction and myocardial fibrosis burden.3 ,4 RV fibrosis as quantified by late gadolinium enhancement (LGE) cardiovascular MR showed a strong correlation between the extent of fQRS (measured by the number of ECG leads in which fQRS was detectable) and the RV LGE score.3 fQRS was associated with higher RV LGE scores, lower RV ejection fractions,3 larger RV volumes and the presence of RV outflow tract aneurysms.4 However, its clinical prognostic value in TOF has been undefined thus far.

This knowledge gap has now been tackled by means of a multicentre study presented by Bokma et al,5 which incorporates 794 adult patients with TOF from the prospective Dutch CONCOR registry. The presence and extent of fQRS on resting ECGs were analysed. When baseline right bundle branch block was present (as in most patients with TOF), fQRS was defined by the presence of ≥3 R wave/notches in the R/S complex in ≥2 contiguous leads. When QRS interval <120 ms, fQRS was present when an additional R wave (R’) or notch was seen in the nadir of the S wave. fQRS extent was graded as none, moderate (2–4) or severe (5–12 leads with fQRS). All-cause mortality was the primary endpoint, while clinical ventricular arrhythmias were examined as the secondary endpoint.

Of the study cohort (median age 27 years, 55% men), fQRS was present in 48% of adult patients with TOF, with severe fQRS (≥5 leads) in 16%. Patients with fQRS had longer QRS durations (155±24 ms) than patients without fQRS (127±26 ms, p<0.01). Over a median 10.7-year follow-up period, 46 (6%) patients died and a further 35 (4%) patients had ventricular arrhythmias. The severity of QRS fragmentation independently predicted all-cause mortality and ventricular arrhythmias. Ten-year survival was 98% in the absence of fQRS, in contrast to 81% in patients with severe fQRS. The sensitivity of any QRS fragmentation (≥2 leads) was 87% versus severe fQRS of 46%, both being superior to the QRS duration sensitivity of 28% in predicting mortality. When multivariable Cox regression analysis was carried out, the extent of fQRS remained independently predictive of mortality (hazard ratio 2.24/higher class, p<0.001), whereas QRS duration did not (p=0.05). The prognostic value of fQRS was sustained regardless of QRS morphology (with and without bundle branch block).

A risk model was subsequently derived by the authors encompassing the multivariate predictors of mortality in the study: fQRS severity (0–2 points), patient age by decade (≥1 point for each decade of age from 30 to 40 years onwards), history of a previous shunt (1 point) and the presence of pacemaker (1 point). A twofold mortality risk was conferred with a risk score of ≥6 points.

This study provides novel prospective evidence on prognostication regarding fQRS in patients with TOF. The data suggest that fQRS is common; its presence and severity may be predictive of death and ventricular tachycardia, and thus, it should be considered for periodic assessment of patients with TOF. Furthermore, the authors proposed a prognostication model, incorporating simple, readily available clinical parameters with the severity of fQRS, towards identifying high-risk patients who may have otherwise been missed.6

Myocardial fibrosis-induced RV dysfunction and scar-related propensity to ventricular arrhythmias may be plausible mechanisms for increased mortality in patients with TOF and fQRS. In addition, mechanoelectrical interaction has been recognised in patients with rTOF for some time.7 ,8 Progressive RV dilatation occurs during adult life in patients with rTOF, usually secondary to obligatory pulmonary regurgitation following surgical repair in childhood. This, in turn, contributes towards QRS prolongation through RV-positive remodelling and longer myocardial conduction times. fQRS severity may therefore be a relevant factor to consider when optimising the timing of pulmonary valve replacement in patients with rTOF.

Significant QRS prolongation (≥180 ms) is widely regarded as a risk factor for sudden cardiac death in TOF.8 The heightened sensitivity of fQRS compared with QRS duration in this study hints at fQRS as an earlier indicator of mortality risk than QRS prolongation in isolation. As the authors allude to, this may imply that patients with shorter, fQRS complexes have higher mortality risk than patients with longer, non-fQRS complexes. This premise merits further investigation, including time-related changes in fQRS and their relevance to outcome. Such an interval change in QRS duration, for example, was more sensitive and specific in identifying patients at risk of sudden cardiac death and malignant arrhythmia compared with absolute QRS duration >180 ms.8 Furthermore, the superiority of fQRS over QRS duration, reported herewith, may also reflect, at least in part, the marked shift in surgical techniques in repairing TOF (from extensive ventriculotomies and frequent use of transannular patches to a right atrial approach and preservation of pulmonary valve function in recent decades). This, in turn, may have led to a shorter QRS duration both at postoperative baseline and with subsequent prolongation at follow-up. Data relating to the surgical approach undertaken within the current study patients were not presented, but may have been informative.

In conclusion, Bokma et al have demonstrated that the presence and severity of fQRS on surface ECGs provide prognostic insights through an inexpensive and reproducible, routinely applied modality in clinical practice. Identification of at-risk patients with rTOF and fQRS may better guide the choice and rationalisation of onward investigations. The predictive value of fQRS may be further enhanced by combining this electrical marker with cardiac functional and structural data from imaging such as echocardiography and MR in future work.