Regression of Electrocardiographic Left Ventricular Hypertrophy After Transcatheter Aortic Valve Implantation for Aortic Stenosis

Abstract

Background: The changes in electrocardiographic left ventricular hypertrophy (ECG-LVH) after transcatheter aortic valve implantation (TAVI) are not fully elucidated.

Methods and Results: The study group included 64 patients who underwent TAVI for aortic stenosis. Their 12-lead ECGs before and at 2 days and 1, 6 and 12 months after TAVI were analyzed, and ECG-LVH was evaluated using various definitions. Values and prevalence of each ECG-LVH parameter significantly decreased between 1 and 6 months after TAVI. Values of ECG-LVH parameters decreased especially in patients with ECG-LVH at baseline.

Conclusions: Regression of ECG-LVH was observed between 1 and 6 months after TAVI.

Left ventricular hypertrophy (LVH) is commonly observed in patients with aortic stenosis (AS) due to an increase in LV afterload.¹ Several studies have reported regression of LVH after transcatheter aortic valve implantation (TAVI), as evaluated by echocardiography,² cardiac magnetic resonance imaging (MRI),³ or cardiac computed tomography (CT).⁴ However, few studies have investigated the ECG changes in LVH after TAVI,⁵ and because previous reports have suggested that electrocardiographic LVH (ECG-LVH) is not always equivalent to echocardiographic LVH,^6,7 it is important to evaluate the ECG changes as well. We used various ECG criteria to investigate changes in ECG-LVH after TAVI.

Methods

Study Design and Patient Group

This was a retrospective, single-center, observational study conducted at Nagoya University Hospital, Aichi, Japan. A total of 156 consecutive patients with severe AS who underwent TAVI between April 2016 and December 2019 were included. Patients with an implanted permanent pacemaker (n=6), complete atrioventricular block (n=1), or bundle branch block (n=22) before TAVI were excluded. Patients with a newly implanted permanent pacemaker (n=8), with newly recorded bundle branch block at any time point (n=28), who died within 12 months after TAVI (n=10), or whose ECG records at any time point were not available (n=17) were also excluded. Finally, 64 patients were included in the analysis. This study complied with the Declaration of Helsinki and was approved by the institutional ethics committee.

ECG Analysis

We evaluated 12-lead ECGs obtained before TAVI, at 2 days, and 1, 6, and 12 months after TAVI. The ECG obtained before TAVI was defined as the baseline. Heart rate, PR, corrected QT intervals (QTc), QRS duration, and QRS axis were measured automatically. As an ECG assessment of LVH, the Sokolow-Lyon voltage criteria (sum of the amplitude of S in V1 and the amplitude of R in V5 or V6, with a cutoff ≥3.5 mV; SV1 + RV5 or RV6), Cornell voltage criteria (sum of the amplitude of R in aVL and the amplitude of S or QS complex in V3, with a cutoff >2.8 mV in men and >2.0 mV in women; RaVL + SV3), Cornell product criteria {[Cornell voltage (+0.8 mV in women)] × QRS duration, with a cut off of 244 mV × ms; [RaVL + SV3 (+0.8 mV in women)] × QRS}, and the Peguero-Lo Presti voltage criteria (sum of the amplitude of the deepest S wave in any lead and the S wave in lead V4; SD + SV4) were evaluated.^8,9 The PR segment was used as the baseline. The voltage of R or R’ and S or QS was measured both manually by the same cardiologist and automatically. When the voltages were different in the same lead, the tallest R or R’ and the deepest S or QS complex were measured.

The number of patients with the strain pattern and inverted T wave were counted. Strain pattern was defined as ≥1 mm convex ST-segment depression with asymmetrical T wave inversion.⁸ The T wave was considered to be inverted when it was deeper than 1 mm.¹⁰ A beat after premature beats was not used for measurement.

Echocardiographic Analysis

Echocardiographic data including the LV mass (LVM) index at baseline and 12 months were obtained. LVM was calculated by the Devereux formula: LVM (g) = 0.80 × {1.04 × [(septal thickness + internal diameter + posterior wall thickness)³ − (internal diameter)³]} + 0.6. LVM was indexed according to body surface area.

Statistical Analysis

Categorical variables are presented as numbers and percentages, and continuous variables are presented as mean±standard deviation or median (interquartile range). The Wilcoxon signed rank test was used to compare continuous ECG variables between baseline and 2 days and at 1, 6, and 12 months. To compare the prevalence of ECG-LVH, strain pattern, and inverted T wave between baseline and 2 days, 1, 6, and 12 months after TAVI, the chi-square test or Fisher’s exact test was used. Intraclass correlation coefficient (ICC) was evaluated to assess the inter- and intra-observer variabilities. An ICC value >0.90 was considered as excellent reliability. Differences were considered statistically significant at P<0.05. All statistical analyses were performed using SPSS version 27.0 (SPSS, Chicago, IL, USA).

Results

The baseline characteristics of the enrolled patients are shown in Table 1: one-third of the patients were male (35.4%), and the majority of patients had hypertension (72.3%).

Table 1. Baseline Characteristics

	n=64
Age, years	82.6±5.7
Male sex, n (%)	22 (34.4)
Height, cm	149±9
Body weight, kg	50.6±10.6
Body mass index	22.6±3.7
Hypertension, n (%)	47 (73.4)
Diabetes mellitus, n (%)	23 (35.9)
Dyslipidemia, n (%)	38 (59.4)
Smoking history, n (%)	20 (31.3)
eGFR <60 mL/min/1.73 m2, n (%)	44 (68.8)
Atrial fibrillation, n (%)	9 (14.1)
Obstructive pulmonary disorder, n (%)	16 (25.0)
Extracardiac arteriopathy, n (%)	11 (17.2)
Previous stroke/TIA, n (%)	9 (14.1)
Previous cardiovascular surgery, n (%)	5 (7.8)
Previous PCI, n (%)	10 (15.6)
Previous MI, n (%)	4 (6.3)
NYHA III/IV, n (%)	32 (50.0)
Echocardiography
Aortic valve area, cm2	0.59 (0.46–0.71)
Peak velocity, m/s	4.5±0.7
Peak pressure gradient, mmHg	79.5 (65.8–99.5)
Mean pressure gradient, mmHg	43.4 (36.4–59.5)
LVEF, %	67 (61.3–73)
Computed tomography
Annulus area, cm2	412±64
Annulus perimeter, mm	72.8±5.7
STS score	5.8 (3.8–7.28)
EuroSCOREII	4.0 (2.5–6.4)
Logistic Euro SCORE	12.7 (9.5–19.8)
Transcatheter heart valve
Type
SAPIEN XT/3, n (%)	43 (67.2)
Evolut R/Pro, n (%)	21 (32.8)
Size
20 mm, n (%)	2 (3.1)
23 mm, n (%)	31 (48.4)
26 mm, n (%)	23 (35.9)
29 mm, n (%)	8 (12.5)
Approach site
Transfemoral, n (%)	60 (93.8)
Non-transfemoral, n (%)	4 (6.3)
Procedural complication
Disabling/ nondisabling stroke, n (%)	2 (3.1)
Life-threatening/disabling bleeding, n (%)	1 (1.6)
Major vascular complication, n (%)	0 (0.0)

LVEF, left ventricular ejection fraction; MI, myocardial infarction; NYHA, New York heart Association; PCI, percutaneous coronary intervention; STS, Society of Thoracic Surgeons; TIA, transient ischemic attack.

The inter- and intra-observer variabilities for the analysis of the voltage of RaVL, RV5, SV1, SV3, SV4 and SD were well correlated [RaVL: inter-, 0.96 (P<0.001), intra-, 0.97 (P<0.001); RV5: inter-, 0.96 (P<0.001), intra-, 1.00 (P<0.001); SV1: inter-, 0.97 (P<0.001), intra-, 1.00 (P<0.001); SV3: inter-, 0.99 (P<0.001), intra-, 0.99 (P<0.001); SV4: inter-, 0.98 (P<0.001), intra-, 0.98 (P<0.001); SD: inter-, 0.97 (P<0.001), intra-, 0.97 (P<0.001)].

Figure 1 and Table 2 show the electrocardiographic and echocardiographic changes. QRS duration significantly decreased at 6 and 12 months after TAVI. The number of patients with strain pattern decreased significantly at 1, 6, and 12 months when compared with baseline, and the number of those with an inverted T wave also tended to decrease over time. The LVM index at 12 months after TAVI significantly decreased compared with baseline.

Figure 1.

Changes in the values and prevalence of ECG-LVH according to various criteria: (A) Sokolow-Lyon voltage criteria; (B) Cornell voltage criteria; (C) Cornell product criteria; (D) Peguero-Lo Presti voltage criteria; (E) QRS duration; (F) strain pattern; (G) inverted T wave. The values and prevalence at each time point were compared with those at baseline, and P values are described. In the box plot, the tops and bottoms of boxes show the 3rd and 1st quartile, respectively. Horizontal lines in the boxes represent median values. The tops and bottoms of bars show the maximum and minimum values without outliers. LVH, left ventricular hypertrophy.

Table 2. Electrocardiographic and Other Parameters

	Baseline	Day 2	1 month	6 months	12 months
ECG
Heart rate, beats/min	76 (67~81)	78 (69~88)	73 (68~80)	75 (67~83)	73 (66~78)
PR interval,# ms	165 (147~187)	167 (153~187)	164 (148~183)	163 (150~182)	168 (152~188)
QRS duration, ms	97 (91~107)	100 (94~107)	97 (92~102)	96 (90~104)•	95 (90~102)*
QTc, ms	435 (416~456)	435 (423~459)	434 (416~447)	431 (411~450)	431 (416~443)
QRS axis, °	23 (4~54)	27(0~58)	28 (−1~53)	28(−12~51)	25 (−3.3~50)•
SV1, mV	1.68 (1.10~2.27)	1.73 (1.18~2.50)	1.53 (1.02~1.82)*	1.26 (0.98~1.84)*	1.25 (0.85~1.64)*
RV5, mV	2.26 (1.76~2.89)	2.31 (1.80~2.76)	2.25 (1.65~2.85)	1.78 (1.33~2.30)*	1.71 (1.31~2.29)*
Sokolow-Lyon voltage (SV1 + RV5), mV	4.03 (3.27~4.98)	4.03 (3.33~4.88)	3.84 (2.89~4.69)•	3.19 (2.75~3.79)*	3.05 (2.47~3.71)*
RaVL, mV	0.67 (0.36~0.95)	0.69 (0.32~0.97)	0.62 (0.25~0.91)•	0.48 (0.26~0.80)*	0.53 (0.26~0.78)•
SV3, mV	1.90 (1.26~2.68)	2.04 (1.26~2.57)	1.56 (1.05~2.20)*	1.48 (0.97~2.08)*	1.34 (1.05~1.98)*
Cornell voltage (RaVL + SV3), mV	2.50 (1.99~3.51)	2.66 (1.91~3.59)	2.12 (1.46~2.87)*	1.98 (1.53~2.69)*	1.91 (1.48~2.59)*
Cornell product,ψ mV × ms	312 (240~412)	326 (224~431)	268 (201~347)*	245 (188~306)*	223 (192~306)*
SD,‡ mV	2.32 (1.82~3.21)	2.50 (1.94~3.30)	2.14 (1.60~2.65)*	1.92 (1.45~2.23)*	1.78 (1.49~2.39)*
SV4, mV	1.21 (0.59~1.74)	1.30 (0.58~1.82)	0.92 (0.45~1.50)*	0.83 (0.40~1.39)*	0.79 (0.42~1.43)*
Peguero-Lo Presti voltage (SD + SV4), mV	3.63 (2.63~4.70)	3.73 (2.71~5.30)	2.88 (2.27~4.18)*	2.63 (2.00~3.62)*	2.74 (1.91~3.64)*
Strain T wave, n (%)	13 (20.3)	18 (28.1)	5 (7.8)•	4 (6.3)•	2 (3.1)•
Inverted T wave, n (%)	33 (51.6)	42 (65.6)	29 (45.3)	23 (35.9)	23 (35.9)
Echocardiography
End-diastolic IVS, mm	11.4 (9.8~12.7)				10.2 (9.0~12.0)•
End-diastolic posterior wall thickness, mm	10.9 (9.5~12.1)				10.0 (9.0~11.9)•
LVM index,† g/m2	129 (102~162)				104 (85~126)*
Moderate to severe AR, n (%)	10 (15.6)				3 (4.7)•
BNP, pg/mL	136 (80~358)				65 (37~121)*

^#Not measured in patients with atrial fibrillation. ^ψCornell product = Cornell voltage (+0.8 in women) × QRS duration. ^‡SD = amplitude of the deepest S wave in any lead. ^†Calculated by Devereux formula: LVM (g) = 0.80 × {1.04 × [(septal thickness + internal diameter + posterior wall thickness)³ − (internal diameter)³]} + 0.6. LVM was indexed according to body surface area. •P<0.05 vs. baseline, *P<0.001 vs. baseline. AR, aortic regurgitation; BNP, B-type natriuretic peptide; IVS, interventricular septum thickness; LVM, left ventricular mass.

The Sokolow-Lyon voltage, Cornell voltage, Cornell product, and Peguero-Lo Presti voltage at 1, 6, and 12 months decreased significantly compared with baseline. The voltage of RV5 decreased significantly at 6 and 12 months. The voltage of SV1, RaVl, SV3, SV4 and SD decreased significantly at 1, 6 and 12 months. The rate of ECG-LVH for each parameter is shown in Figure 1. Regarding the criteria of Sokolow-Lyon voltage, Cornell voltage and Peguero-Lo Presti voltage, the prevalence of ECG-LVH decreased significantly at 6 and 12 months compared with baseline. For the Cornell product criterion, the rate of LVH decreased significantly at 1, 6, and 12 months compared with baseline.

Figure 2 shows the change in each ECG-LVH parameter when the patients were divided into those with and without ECG-LVH at baseline. ECG-LVH parameter values decreased significantly, especially in patients with ECG-LVH at baseline.

Figure 2.

Changes in the ECG-LVH parameters when patients were divided into two groups: with and without baseline ECG-LVH. (A) Sokolow-Lyon voltage criteria; (B) Cornell voltage criteria; (C) Cornell product criteria; (D) Peguero-Lo Presti voltage criteria. Tops and bottoms of boxes show the 3rd and 1st quartile, respectively. Horizontal lines in the boxes represent median values. Tops and bottoms of bars show the maximum and minimum values without outliers. The values and the prevalence at each time point were compared with baseline. •P<0.05; *P<0.001. LVH, left ventricular hypertrophy.

Discussion

We have presented the changes in ECG-LVH parameters and the prevalence of ECG-LVH according to various criteria. Compared with baseline, significant reductions in the values and prevalence of ECG-LVH were observed between 1 and 6 months after TAVI, especially in patients with ECG-LVH at baseline. In addition, the QRS duration significantly decreased at 6 months, and the ratio of patients with the strain pattern decreased at 1 month after TAVI.

Regression of ECG-LVH after surgical aortic valve replacement (SAVR) has been previously reported,^11,12 but the same phenomenon after TAVI has not been fully investigated. Recently, Tanaka et al showed regression of ECG-LVH after TAVI, as evaluated using the Sokolow-Lyon voltage.⁵ However, there are various criteria for ECG-LVH, and it is considered important to validate using the other criteria.

With regard to the time course of ECG-LVH, regression has been observed as per several criteria between 2 weeks and 6 months after SAVR.¹¹ After TAVI, regression of ECG-LVH evaluated by Sokolow-Lyon voltage was observed at 1 month.⁵ Our results, revealing the time course of various ECG parameters after TAVI, support those conclusions.

Similar to ECG-LVH, previous studies have shown regression of LVH after TAVI, as assessed quantitatively by echocardiography, MRI, or CT. Regression of LVH was observed by echocardiographic assessment at 1 month,² and regression of LVM was observed by MRI 4 days after TAVI.³ Therefore, regression of ECG-LVH is thought to indicate regression of anatomic LVH.

However, the relationship between ECG- or anatomic LVH and prognosis after TAVI is unclear.^5,13–17 Although the existence of echocardiographic LVH before TAVI was not associated with 1-year clinical outcomes,¹³ severe echocardiographic LVH before TAVI was reported to be related to poor outcomes at 5-year follow-up.¹⁴ Conversely, for ECG assessment, studies have reported that the presence of ECG-LVH at baseline was associated with better clinical outcome.^5,15 Regarding the change in LVH, regression of echocardiographic-, CT- and ECG-LVH after TAVI has been associated with favorable clinical outcomes.^4,5,16,17 The difference in the method of evaluating LVH might be the main reason for these conflicting outcomes, and anatomic LVH is thought to be not always equal to ECG-LVH. A relative voltage deficit caused by reduced intercellular coupling in hypertrophic hearts has been proposed.⁶ Furthermore, myocardial injury, fibrosis or amyloid deposition decreases the voltage of the QRS wave.^18–20 It is important to understand the difference between ECG- and anatomic LVH and to evaluate both when evaluating patients with AS before and after treatment. Further investigations are required to fully elucidate the relationship between LVH and TAVI.

Study Limitations

First, this was a retrospective, single-center observational study with a limited number of patients. Second, patients with bundle branch block or permanent pacemaker and those whose data were unavailable were excluded, which might have caused selection bias. Third, the relationship between LVH/LVH regression and clinical outcomes after TAVI was not evaluated in this study because of the limited number of patients and clinical events.

Conclusions

Regression of ECG-LVH after TAVI was demonstrated as per various criteria. The voltages and prevalence of ECG-LVH decreased significantly between 1 and 6 months after TAVI.

Funding

None.

Disclosures

H.I. received lecture fees from Astellas Pharma Inc., Astrazeneca Inc., Daiichi-Sankyo Pharma Inc., and MSD K. K. A.U. received lecture fees from Terumo, Japan Blood Products Organization. A.U. received an unrestricted research grant for the Department of Cardiac Surgery, Nagoya University Graduate School of Medicine from Edwards Lifesciences Corporation, Senko Medical Instrument Mfg., Co., Ltd., Medtronic, Terumo, and Lifeline. T.M. received lecture fees from Bayer Pharmaceutical Co., Ltd., Daiichi-Sankyo Co., Ltd., Dainippon Sumitomo Pharma Co., Ltd., Kowa Co., Ltd., MSD K. K., Mitsubishi Tanabe Pharma Co., Nippon Boehringer Ingelheim Co., Ltd., Novartis Pharma K. K., Pfizer Japan Inc., Sanofi-Aventis K. K., and Takeda Pharmaceutical Co., Ltd. T.M. received an unrestricted research grant for the Department of Cardiology, Nagoya University Graduate School of Medicine from Astellas Pharma Inc., Daiichi-Sankyo Co., Ltd., Dainippon Sumitomo Pharma Co., Ltd., Kowa Co., Ltd., MSD K. K., Mitsubishi Tanabe Pharma Co., Nippon Boehringer Ingelheim Co., Ltd., Novartis Pharma K. K., Otsuka Pharma Ltd., Pfizer Japan Inc., Sanofi-Aventis K. K., Takeda Pharmaceutical Co., Ltd., and Teijin Pharma Ltd.

H.I., A.U., and T.M. are members of Circulation Journal’s Editorial Team. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.

IRB Information

This study was approved by the Ethics Committee of Nagoya University Hospital (approval no. 2019-0179).

Acknowledgments

None.

Data Availability

The de-identified participant data will not be shared.

References

• 1.   Rader F, Sachdev E, Arsanjani R, Siegel RJ. Left ventricular hypertrophy in valvular aortic stenosis: Mechanisms and clinical implications. Am J Med 2015; 128: 344–352.
• 2.   Dahiya G, Kyvernitakis A, Joshi AA, Lasorda DM, Bailey SH, Raina A, et al. Impact of transcatheter aortic valve replacement on left ventricular hypertrophy, diastolic dysfunction and quality of life in patients with preserved left ventricular function. Int J Cardiovasc Imaging 2021; 37: 485–492.
• 3.   Dobson LE, Musa TA, Uddin A, Fairbairn TA, Swoboda PP, Erhayiem B, et al. Acute reverse remodelling after transcatheter aortic valve implantation: A link between myocardial fibrosis and left ventricular mass regression. Can J Cardiol 2016; 32: 1411–1418.
• 4.   Szilveszter B, Oren D, Molnar L, Apor A, Nagy AI, Molnar A, et al. Subclinical leaflet thrombosis is associated with impaired reverse remodelling after transcatheter aortic valve implantation. Eur Heart J Cardiovasc Imaging 2020; 21: 1144–1151.
• 5.   Tanaka T, Yahagi K, Asami M, Ninomiya K, Kikushima H, Okuno T, et al. Prognostic impact of electrocardiographic left ventricular hypertrophy following transcatheter aortic valve replacement. J Cardiol 2021; 77: 346–352.
• 6.   Bacharova L, Mateasik A, Krause R, Prinzen FW, Auricchio A, Potse M. The effect of reduced intercellular coupling on electrocardiographic signs of left ventricular hypertrophy. J Electrocardiol 2011; 44: 571–576.
• 7.   Bacharova L. The structural and electrical remodeling of myocardium in LVH and its impact on the QRS voltage. Anadolu Kardiyol Derg 2007; 7(Suppl 1): 37–42.
• 8.   Hancock EW, Deal BJ, Mirvis DM, Okin P, Kligfield P, Gettes LS, et al. AHA/ACCF/HRS recommendations for the standardization and interpretation of the electrocardiogram: Part V: Electrocardiogram changes associated with cardiac chamber hypertrophy: A Scientific Statement from the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society. Endorsed by the International Society for Computerized Electrocardiology. J Am Coll Cardiol 2009; 53: 992–1002.
• 9.   Peguero JG, Lo Presti S, Perez J, Issa O, Brenes JC, Tolentino A. Electrocardiographic criteria for the diagnosis of left ventricular hypertrophy. J Am Coll Cardiol 2017; 69: 1694–1703.
• 10.   Hanna EB, Glancy DL. ST-segment depression and T-wave inversion: Classification, differential diagnosis, and caveats. Cleve Clin J Med 2011; 78: 404–414.
• 11.   Kurisu S, Inoue I, Kawagoe T, Ishihara M, Shimatani Y, Nakama Y, et al. The decrease in QRS amplitude after aortic valve replacement in patients with aortic valve stenosis. J Electrocardiol 2009; 42: 410–413.
• 12.   Yamabe S, Dohi Y, Higashi A, Kinoshita H, Sada Y, Hidaka T, et al. Clinical value of regression of electrocardiographic left ventricular hypertrophy after aortic valve replacement. Heart Vessels 2016; 31: 1497–1503.
• 13.   Varshney AS, Manandhar P, Vemulapalli S, Kirtane AJ, Mathew V, Shah B, et al. Left ventricular hypertrophy does not affect 1-year clinical outcomes in patients undergoing transcatheter aortic valve replacement. JACC Cardiovasc Interv 2019; 12: 373–382.
• 14.   Gonzales H, Douglas PS, Pibarot P, Hahn RT, Khalique OK, Jaber WA, et al. Left Ventricular hypertrophy and clinical outcomes over 5 years after TAVR: An analysis of the PARTNER trials and registries. JACC Cardiovasc Interv 2020; 13: 1329–1339.
• 15.   Zhang R, Jou S, Cao Y, Keller K, Xiao E, de La Rosa A, et al. Absence of electrocardiographic left ventricular hypertrophy in patients undergoing transcatheter aortic valve replacement is associated with increased mortality. J Electrocardiol 2020; 63: 12–16.
• 16.   Lindman BR, Stewart WJ, Pibarot P, Hahn RT, Otto CM, Xu K, et al. Early regression of severe left ventricular hypertrophy after transcatheter aortic valve replacement is associated with decreased hospitalizations. JACC Cardiovasc Interv 2014; 7: 662–673.
• 17.   Chau KH, Douglas PS, Pibarot P, Hahn RT, Khalique OK, Jaber WA, et al. Regression of left ventricular mass after transcatheter aortic valve replacement: The PARTNER trials and registries. J Am Coll Cardiol 2020; 75: 2446–2458.
• 18.   Zema MJ, Collins M, Alonso DR, Kligfield P. Electrocardiographic poor R-wave progression: Correlation with postmortem findings. Chest 1981; 79: 195–200.
• 19.   Finocchiaro G, Merlo M, Sheikh N, De Angelis G, Papadakis M, Olivotto I, et al. The electrocardiogram in the diagnosis and management of patients with dilated cardiomyopathy. Eur J Heart Fail 2020; 22: 1097–1107.
• 20.   Ternacle J, Krapf L, Mohty D, Magne J, Nguyen A, Galat A, et al. Aortic stenosis and cardiac amyloidosis: JACC Review Topic of the Week. J Am Coll Cardiol 2019; 74: 2638–2651.