Is the QRS duration useful for determining the optimal timing of pulmonary valve replacement after tetralogy of Fallot repair?

In pulmonary valve replacement (PVR) after tetralogy of Fallot (TOF) repair, the right ventricular end-diastolic and end-systolic volume index (RVEDVI and RVESVI) of cardiac magnetic resonance imaging (cMRI) are often used as indicators of the RV volume. We examined the utility of QRS duration, cardiothoracic ratio (CTR), and plasma brain natriuretic peptide (BNP) as indicators of the appropriate timing of cMRI to assess the RV volume and function before PVR. We assessed the correlation of QRS duration, CTR, and BNP with RVEDVI and RVESVI on cMRI in 26 patients after TOF repair. Fifteen underwent PVR (age, 45.2 ± 11.4 years). Twelve underwent post-PVR cMRI. The RV volume change from before to after PVR was investigated. QRS duration, BNP, and CTR were positively correlated with RVEDVI and RVESVI after TOF repair. The post-PVR QRS duration was also positively correlated with post-PVR RVEDVI (p = 0.017) and RVESVI (p = 0.001). From before to after PVR, in 5 cases with QRS duration ≤ 160 ms, the QRS duration decreased from 110.4 ± 28.9 to 101.8 ± 30.5 ms (p = 0.063). Both RVEDVI and RVESVI decreased to the normal range in 4 of 5 cases. In contrast, in 7 cases with QRS duration > 160 ms, the QRS duration decreased from 183.0 ± 17.4 to 160.3 ± 23.8 ms (p = 0.013); however, RVESVI did not normalize in 6 of 7 cases. A prolonged QRS duration is a useful marker of RVEDVI and RVESVI enlargement after TOF repair. We recommend performing cMRI before the QRS duration reaches 160 ms due to normalization of the RV volume after PVR.


Introduction
An increasing number of adult patients experience pulmonary regurgitation (PR) after tetralogy of Fallot (TOF) repair, and the resultant chronic volume overload can lead to right ventricular (RV) dilation, biventricular dysfunction, supraventricular or ventricular arrhythmias, and sudden death [1][2][3][4]. PVR should be performed while RV enlargement and functional deterioration are reversible. Several studies suggest the optimal timing of PVR in terms of the threshold of the preoperative RV end-diastolic volume index (RVEDVI) or RV end-systolic volume index (RVESVI) by cardiac magnetic resonance imaging (cMRI) for RV normalization after PVR [5][6][7][8][9]. Therrien et al. recommended that PVR be performed before RVEDVI reached 170 ml/m 2 or RVESVI reached 85 ml/m 2 to normalize the RV volume after PVR [5]. Their report is also cited in the guidelines for the management of congenital heart disease in adults established by the Japanese Circulation Society [10]. We investigated whether three parameters, the resting electrocardiogram (ECG) QRS duration, cardiothoracic ratio (CTR), and plasma brain natriuretic peptide (BNP) level, could be used as indicators of the appropriate timing of cMRI for RV normalization to assess the RV volume and function before PVR.

Patients
The present study included 26 patients who received TOF repair and who underwent cMRI between August 2013 and November 2019 at 38.9 ± 14.0 years of age to consider PVR. Fifteen of the 26 patients underwent PVR between April 2016 and September 2020. The mean age at PVR was 43.1 ± 11.4 years, the mean interval after TOF repair was 37.6 ± 9.2 years, and the mean body weight at PVR was 71.5 ± 32.5 kg. The mean age at TOF repair was 5.0 ± 3.8 years. Twenty-one cases were reconstructed with a transannular monocusp patch and 5 cases received a pulmonary valve sparing procedure. Post-PVR cMRI was performed in 12 of 15 patients after PVR at a mean of 2.1 ± 1.3 years after PVR (Table 1). PVR indications for moderate or greater PR were symptoms and signs attributable to RV volume overload, presence of significant associated lesions such as tricuspid regurgitation and sustained tachyarrhythmias. For asymptomatic patients, we considered PVR when RVEDVI assessed by cMRI approached around 160 ml/m 2 .
The Institutional Review Board of Ehime University Hospital approved this retrospective study and waived individual informed consent. Data were obtained by review of medical records.

Methods
We assessed the correlation of QRS duration of resting 12-lead ECG, CTR on chest radiography, and plasma BNP level at the same time of cMRI with RVEDVI, RVESVI, and RVEF on cMRI in 26 patients who had received TOF repair. We also assessed the correlation of the post-PVR QRS duration, CTR, and plasma BNP level, at the same time of post-PVR cMRI, with post-PVR RVEDVI, RVESVI, and RVEF on cMRI in 12 patients who received cMRI after PVR. Changes of the QRS duration, RVEDVI, RVESVI and RVEF before and after PVR were also investigated. Normal range of RVEDVI and RVESVI defined as < 108 ml/m 2 and < 47 ml/m 2 , respectively [5].

QRS duration
Electrocardiography was analyzed automatically using the ECAPS12C ECG program (NIHONKODEN, Tokyo, Japan). Standard (speed, 25 mm/s and 1 mV/cm standardization) resting 12-lead ECG was used. The QRS duration was defined as the maximal QRS length in any lead from the first inflection to the final sharp vector changing to gentle inclination.

BNP
Venous blood was withdrawn and immediately sent for centrifugation. The plasma BNP level was determined by chemiluminescent immunoassay (CLIA) with BNP-JP・Abbott reagent (Abbott Japan, Matsudo, Japan) within 4 h after blood sampling under room temperature.

Cardiac MRI
All cMRI examinations were performed using a clinical 3 T MR scanner (MAGNETOM Skyra; Siemens Healthcare, Erlangen, Germany). True fast imaging with steady state precession (true FISP) was used for the retrospective ECGgated cine cMRI scans of all participants. Short-axis cine cMRI were obtained in a stack of eight contiguous slices spanning the entire ventricle from the base to the apex. The imaging parameters were as follows: repetition time, 36.6 ms; echo time, 1.4 ms; flip angle, 50°; section thickness, 6 mm; field-of-view, 312 × 384 mm; and voxel size, 0.9 × 0.9 × 6 mm. For quantitative measurements, the stack of eight contiguous short-axis slices of cine cMRI was assessed using a dedicated software package (SYNAPSE VINCENT; Fujifilm Corp., Ltd, Tokyo, Japan). The epicardial and endocardial contours were automatically traced on short-axis images. Contours rendered by the automated analysis were reviewed and manually corrected, as necessary.

PVR
PVR was performed through median repeat sternotomy using standard cardiopulmonary bypass. We performed PVR on beating, in cases without concomitant mitral valve plasty or a left atrial Maze procedure. The concomitant procedures are shown in Table 1

Statistical analysis
The data are given as absolute numbers and percentages, with continuous values presented as the mean ± standard deviation. The statistical analyses were performed using a commercially available statistical program. Differences between parameters before and after surgery were analyzed using a paired Student's t test. The association between two continuous variables was assessed by a linear regression analysis. As plasma BNP levels showed a skewed distribution, logarithmically transformed BNP values were used in correlation and regression analyses. p values of < 0.05 were considered statistically significant.

Parameters correlated with RV volume or RVEF after TOF repair
In all 26 cases, the QRS duration, log BNP, and CTR showed good positive correlations with RVEDVI and RVESVI after TOF repair (Fig. 1, Table 2). The QRS duration and log BNP also showed a good negative correlation with RVEF (r = 0.56, p = 0.003; r = 0.55, p = 0.003, respectively, Fig. 2, Table 2) after TOF repair.

Parameter changes before and after PVR related to QRS duration
In our data, the correlation between RVEDVI (X) and QRS duration (Y) was demonstrated by the following regression 1 3  Fig. 1A). Therrien et al. reported that the cutoff value of RVEDVI for RV normalization after PVR 170 ml/m 2 [5]. This cutoff value was applied to our regression equation (QRS duration: 156 ms).
We found that a standard QRS duration of > 160 ms after TOF repair could identify patients with RVEDVI > 170 ml/ m 2 with 88.9% sensitivity and 70.6% specificity, and  According to the QRS duration, we divided the patients into the W group (QRS duration > 160 ms) and the N group (QRS duration ≤ 160 ms).
Post-PVR cMRI was performed in 7 of 9 patients in the W group, and 5 of 6 patients in the N group. Parameters before and after PVR were compered in these 12 cases (N group, n = 5; W group, n = 7).
From before to after PVR, the QRS duration decreased from 110.4 ± 28.9 to 101.8 ± 30.5 ms in all 5 cases in the N group (p = 0.063), and the QRS duration decreased from 183.0 ± 17.4 to 160.3 ± 23.8 ms in all 7 cases but 1 in the W group (p = 0.013, Fig. 4A).
RVEF did not change significantly from before to after PVR in either the N group (from 43.2 ± 8.0% to 43.7 ± 8.5%, p = 0.93) or the W group (from 30.9 ± 9.6% to 34.9 ± 9.6%, p = 0.51, Fig. 4D). In the N group, the post-PVR RVEF was > 40% in 4 of 5 cases, whereas in the W group, it was < 40% in 6 of 7 cases.

QRS duration changes and RVEDVI changes after TOF repair
The QRS duration changed simultaneously with the RVEDVI in each case after TOF repair. Although there was no linear relation between QRS duration change and RVEDVI change after TOF repair, an increased QRS duration was accompanied by an increase of RVEDVI, and no change of QRS duration was accompanied by no change of RVEDVI in many cases (Figs. 5, 6). The QRS duration of TAP cases tended to be wider in comparison to pulmonary valve sparing cases in each generation after TOF repair. In Fig. 5, the dashed line indicates the pulmonary stenosis and regurgitation (PSR) cases with an RVOT pressure gradient of ≥ 25 mmHg; the QRS duration of PSR cases was larger than that of PR cases. However, because most PSR cases were older than the other cases of PR, the influence of RVOT stenosis was not obvious (Fig. 5).

Discussion
It is not practical to perform frequent cMRI to determine the optimal timing of PVR due to normalize the RV volume and prevent RV dysfunction in patients after TOF repair. We considered it ideal that parameters measured in the daily clinical setting could indicate the optimal timing of cMRI which was the gold standard for the evaluation of the RV volume and function.
In chronic PR, the RV systolic function is initially preserved, and many cases remain relatively free of symptoms. Once the compensatory mechanisms cannot be maintained, the RV mass-to-volume ratio decreases, then the RVESV increases, RVEF decreases, and patients finally fall into RV dysfunction. The RV dysfunction is also associated with RV wall stress due to RV afterload from RVOT obstruction, RV fibrosis, RVOT aneurysm at the site of transannular patch, an impaired RV diastolic function, and left ventricular (LV) dysfunction [2][3][4][11][12][13].
To increase the chances for the patient to reach a normal RV volume after the repair, Therrien et al. recommended PVR be undertaken before the RVEDVI reaches 170 ml/m 2 or the RVESVI reaches 85 ml/m 2 [5]. This report was cited in the guidelines of the Japanese Circulation Society [10], which provided the impetus for carrying out the present study. PVR is recommended to be performed under the condition that the RVEDVI is < 160-170 ml/m 2 and the RVESVI is < 80-90 ml/m 2 , as thresholds for the normalization of the RV volume after PVR in previous reports [5][6][7][8][9].  [2,7,[13][14][15]. Some studies reported that BNP and NT-pro BNP reflected the RVEDVI and RVEF of patients after TOF repair [16,17]. In present study, the plasma BNP level was also correlated with the RV volume in patients after TOF repair. However, the use of the plasma BNP level for the estimation of RV volume during outpatient follow-up was thought to be complex as it requires logarithmic transformation. Moreover, Eindhoven et al. reported that NT-pro BNP was correlated with the systolic RV function, but more strongly correlated with the systolic LV function and dimensions. They said that the likelihood that NT-pro BNP will be released from the right ventricle is generally lower in comparison to left ventricle [18].

Parameters correlated with post-PVR RV volume
In the present study, only the post-PVR QRS duration showed good correlations with both post-PVR RVEDVI and RVESVI. This indicates that the QRS duration is a good marker of the RV volume both before and after PVR in patients after TOF repair, as Doughan et al. reported [15].
In their report of 26 cases, Huysduynen et al. noted that the change in the QRS duration (ΔQRS duration) from before to after PVR is also correlated with the change in RVEDVI (ΔRVEDVI). They considered that RV dilatation may increase the QRS duration by increasing the distance that the electrical activation front has to move in the right ventricle, and they considered that the relationship between RVEDVI and the QRS duration also applies to the ΔRVEDVI and the ΔQRS duration before and after PVR [19].

Relationship between QRS duration and RV volume after PVR
The positive correlation between RVEDVI and the QRS duration in our data yielded a cutoff value of the QRS duration of 160 ms when RVEDVI 170 ml/m 2 was used as the threshold for RV normalization by Therrien In the present study, both RVEDVI and RVESVI normalized after PVR in most patients (RVEDVI, all 5 patients; RVESVI, 4 of 5 patients) in the N group (QRS duration ≤ 160 ms), although RVESVI did not normalize in 6 of 7 patients in the W group (QRS duration > 160 ms). Because RVESVI has been reported to have a better relationship with RVEF than RVEDVI [2,7,14], our results about the threshold of QRS duration (160 ms) for RV normalization are considered to have clinical significance. The post-PVR RVEF was > 40% in 4 of 5 patients in the N group; however, the post-PVR RVEF was < 40% in 6 of the 7 patients in the W group.
Geva et al. also used a cutoff value of QRS duration of ≥ 160 ms to predict persistent post-PVR RV dilatation and dysfunction, and suggested that a QRS duration of < 140 ms was associated with a normal RV size and function after PVR [9]. Cocomello et al. reported that patients with a QRS duration approaching 160 ms or those with RVEDVI ≥ 166 ml/m 2 or RVESVI ≥ 89 ml/m 2 should be considered for PVR to prevent a gradual increase of RV volume after PVR [20]. These 2 studies support the validity of our results regarding the cutoff value of QRS duration (160 ms) after TOF repair.
Scherptong et al. reported that a severely prolonged QRS duration (> 180 ms) was associated with an increased incidence of adverse outcomes, and PVR should be performed before severe QRS prolongation occurs [21]. Romeo et al. reported that the QRS duration could be used as a biomarker for RV volumetric and functional change, and as a risk factor for late adverse outcomes, and that if PVR can be performed early enough to prevent the progression of QRS prolongation, it may reduce the incidence of adverse events after PVR [22].

QRS duration changes and RVEDVI changes after TOF repair
Regarding the question of whether the QRS duration can be used as a preoperative assessment of the RV volume in outpatients after TOF repair, there was a tendency for the QRS duration to change simultaneously with the RV volume in each case after TOF repair, although this was only observed in a small number of cases. In other words, RVEDVI increased in patients with an increased QRS duration, especially in some PVR cases, while RVEDVI did not change in patients with a steady QRS duration, especially in cases without PVR (Figs. 5, 6). Over time, some of the patients showed a longer QRS duration of around 160 ms, PVR was indicated for some of these patients due to an increased RV volume. Such a change in QRS duration is essential for the continuous use of the QRS duration as an indicator of RV volume progression after TOF repair. To the best of our knowledge, limited studies have focused on the QRS duration as an indication for PVR for RV enlargement due to PR [8,18]. The QRS duration was particularly long in patients who underwent PVR more than 35 years after TOF repair; however, this may have been due to the historical background of large right ventriculotomy. In the no PVR cases, some patients did not show an increase in QRS duration over time; however, we consider that cMRI should be performed if there is a trend toward an increase in QRS duration or if the QRS duration exceeds 140 ms, based on the above-mentioned report by Geva et al. [9,23].
Considering additional factors of QRS prolongation, right bundle branch block (RBBB) due to RV infundibular resection or RVOT incision in TOF repair could influence the QRS prolongation. In this study, the QRS duration of TAP cases tended to be wider than that of pulmonary valve sparing cases in each generation after TOF repair (Fig. 5). In our study, 14 of the 15 PVR cases (the one exception had a narrow QRS duration of 92 ms) showed RBBB before PVR, and all 11 no PVR cases showed RBBB at the last follow-up examination. RBBB due to infundibular resection will progress with RV dilatation from PR after TOF repair.
The results regarding the association between pulmonary stenosis and the QRS duration were not significant due to the small number of PSR cases and longer time after TOF repair in PSR cases in comparison PR cases.
The QRS duration was a useful marker of the RV volume and RVEF; however, this study lacked sufficient power to determine whether the QRS duration can be used directly as a threshold to predict RV normalization after PVR. After TOF repair, we recommend that cMRI to evaluate the RV volume for the optimal timing of PVR be performed before the QRS duration reaches 160 ms.

Limitations
The present study was associated with some limitations, including its retrospective and monocentric design. The relatively small number of patients in this study resulted in a loss of statistical power. Not all patients could perform cMRI more than once after TOF repair, and a cMRI volume study could not be performed in all PVR patients. Our study requires validation by a further study of a larger population.

Conclusions
A prolonged QRS duration is a useful marker of the enlargement of RVEDVI and RVESVI. A QRS duration of 160 ms is a sensitive cutoff value for RV volume normalization after PVR. In order to evaluate the RV volume to determine the optimal timing of PVR, we recommend that cMRI be performed before the QRS duration reaches 160 ms.