Cornell Voltage Criteria

Cornell LVH

Standard Voltage Analysis for LVH

Baseline Reference (Sex)

R wave in aVL

S wave in V3

Enhanced Sensitivity (Cornell Product)

The Cornell Criteria is highly reliable for specificity. However, if the sum is borderline, check for the "RaVL Criterion" (R in aVL > 11 mm) or signs of Left Atrial Enlargement.

Input EKG leads
to analyze LVH

Guidelines & Evidence

Verified

Last Review: 2026-07-17

When to Use

Primary Indications

ECG screening for left ventricular hypertrophy (LVH) – Non-invasive, widely available, low-cost initial test

Chronic hypertension monitoring – Assess target organ damage from long-standing or poorly controlled hypertension

Risk stratification in hypertensive patients – LVH by ECG predicts increased cardiovascular morbidity and mortality (2-4x higher risk of MI, stroke, heart failure, sudden cardiac death)

Sex-specific assessment – Superior sensitivity in women compared to Sokolow-Lyon (Cornell: 42% vs Sokolow-Lyon: 23% sensitivity at 95% specificity)

Screening for hypertrophic cardiomyopathy (HCM) – When combined with clinical suspicion and family history

Aortic stenosis severity assessment – LVH is an adaptive response to pressure overload; absence of LVH in severe AS suggests alternate diagnosis

Athlete vs pathology differentiation – Physiologic LVH in athletes (increased voltage with normal LV dimensions, no ST/T changes) vs pathologic LVH (with repolarization abnormalities)

Prognostic stratification post-MI – Regression of LVH with ACE inhibitors/ARBs predicts improved outcomes

Contraindications / Limitations

Left bundle branch block (LBBB) – Voltage criteria unreliable; rely on LBBB-specific LVH criteria (e.g., Sclarovsky-Birnbaum)

Ventricular paced rhythms – Artificial depolarization alters voltage distribution

Left anterior fascicular block (LAFB) – Alters frontal plane voltages; caution with interpretation

Poor ECG quality – Baseline wander, muscle artifact, improper lead placement invalidate measurements

Low voltage ECG – Obese patients, COPD, pericardial effusion, amyloidosis (Cornell will underdiagnose LVH)

Athletes with physiologic LVH – High voltage may be normal; requires clinical correlation (ST/T wave abnormalities suggest pathology)

NOT a substitute for echocardiography – ECG LVH criteria have modest sensitivity (30-50%) but high specificity (85-95%). Echo (LV mass index) is gold standard for LVH diagnosis.

Comparison of LVH ECG Criteria

Criterion	Formula	Threshold (LVH Present)	Sensitivity (Echo-proven LVH)	Specificity	Best For
Cornell Voltage (original, 1991)	RaVL + SV3	Men: >20 mm Women: >28 mm	40-50%	90-95%	Women (far better sensitivity than Sokolow-Lyon). General screening.
Cornell Product (1991)	(RaVL + SV3) × QRS duration (ms)	>2440 mm·ms	45-55%	90-95%	Added accuracy when QRS is prolonged (intraventricular conduction delay)
Sokolow-Lyon (1949)	SV1 + RV5 or RV6 (whichever larger)	≥35 mm	35-45%	85-90%	Men (historically derived from young men). Simpler to calculate.
Romhilt-Estes (1968)	Point score system - Voltage criteria: 3 points - ST/T changes: 3 points - Left atrial enlargement: 3 points - Left axis deviation: 2 points - QRS duration >90ms: 1 point - Intrinsicoid deflection >50ms: 1 point	≥5 points = definite LVH 4 points = probable LVH	30-40%	95-98%	Highest specificity for autopsy-proven LVH (point score system)
Peguero-Lo Presti (2017)	Deepest S wave in any lead (V1-V6) + SV4	Men: >28 mm Women: >23 mm	55-65%	85-90%	Obesity, COPD (where QRS voltages are generally lower)
Framingham (1990)	SV1 + RV5/V6 or RaVL + SV3 (sex-specific continuous score)	Sex-specific percentile-based (e.g., >95th percentile)	Variable (depends on percentile)	Variable	Research settings, not commonly used clinically

How it Works

Standard Cornell Voltage (Sex-Specific)

Cornell Voltage (mm) = RaVL + SV3 • RaVL = amplitude of the R wave in lead aVL (measured from baseline to peak of R wave) • SV3 = amplitude of the S wave in lead V3 (measured from baseline to nadir of S wave) THRESHOLDS: • Men: RaVL + SV3 > 20 mm (some guidelines use ≥28 mm? NO: Men >20 mm, Women >28 mm is standard) • Women: RaVL + SV3 > 28 mm (higher threshold due to generally lower QRS voltages in women) ALTERNATE (if QRS duration >120 ms): Cornell Product = (RaVL + SV3) × QRS duration (ms) > 2440 mm·ms • Accounts for prolongation of QRS in LVH due to slower conduction through hypertrophied myocardium

Measurement Technique (Standard ECG Calibration: 10 mm/mV)

Quick Reference: LVH Thresholds

Criterion	Men	Women	Notes
Cornell Voltage (RaVL + SV3)	>20 mm	>28 mm	Primary criterion; sex-specific thresholds improve accuracy in women (original Casale 1991 data)
Cornell Product (× QRS duration)	>2440 mm·ms	>2440 mm·ms	Single threshold for both sexes when QRS prolonged; best for LVH with intraventricular delay
Sokolow-Lyon (SV1 + RV5/6)	≥35 mm	≥35 mm (but less sensitive)	Not sex-specific; lower sensitivity in women
Peguero-Lo Presti (deepest S + SV4)	>28 mm	>23 mm	Alternative for low-voltage ECGs (obesity, COPD, pericardial effusion)

Clinical Pearls

Critical Pearl #1: Sex-Specific Thresholds Matter

The Cornell criteria were the first sex-specific ECG LVH criteria. Women have lower QRS voltages than men due to smaller body habitus and less LV mass. Using a single threshold for both sexes (as in Sokolow-Lyon) reduces sensitivity in women. Studies show: at 95% specificity, Cornell sensitivity in women is 42% vs Sokolow-Lyon 23%. ALWAYS use the sex-specific threshold: Men >20 mm, Women >28 mm.

Critical Pearl #2: When QRS is Wide, Use Cornell Product

LVH causes slower intraventricular conduction due to enlarged myocytes and increased myocardial fibrosis. This is reflected by QRS prolongation (>100-120 ms). The Cornell Product (voltage sum × QRS duration in ms) accounts for this and improves diagnostic accuracy. Threshold >2440 mm·ms for both sexes. If QRS duration is 160 ms, even modest voltage sum of 15 mm yields product 2400—close to threshold—suggesting LVH with conduction delay.

Critical Pearl #3: Physiologic vs Pathologic LVH

Athletes frequently meet voltage criteria for LVH (increased QRS voltages) without having true pathologic hypertrophy. Distinguishing features: Physiologic LVH in athletes: (1) Normal LV cavity size (not small), (2) No ST segment depression or T wave inversions (repolarization normal), (3) QRS duration normal or mildly prolonged but no left atrial enlargement, (4) Rapid regression with deconditioning. Pathologic LVH (hypertension, HCM, AS): (1) LV cavity small or normal, (2) ST segment depression and T wave inversions (strain pattern) in lateral leads (I, aVL, V5-V6), (3) Left atrial enlargement (P wave duration >120 ms in lead II, negative terminal P wave in V1 >1 mm deep), (4) May have deep Q waves (HCM). If athlete with isolated voltage criteria (no repolarization changes or left atrial enlargement), re-evaluate after 3-6 months of reduced training.

Limitations and Interpretation Pitfalls

False positives (meet ECG LVH criteria but no LVH on echo): (1) Young adults with thin chest walls, (2) Lean body habitus (low body fat, voltages appear higher), (3) Athletes with physiologic LVH (normal adaptation, not disease), (4) Electronic artifact (muscle tremor mimicking voltage).

False negatives (normal ECG but LVH on echo): (1) Obese patients (adipose tissue attenuates QRS voltages), (2) COPD/hyperinflated lungs (increased distance from heart to chest wall electrodes), (3) Pericardial effusion (fluid attenuates voltage), (4) Amyloidosis (infiltrative cardiomyopathy paradoxically produces low voltage despite increased LV mass), (5) Early/established LVH without voltage changes (some patients never develop voltage criteria despite significant hypertrophy).

Conduction delays confuse interpretation: LBBB, RBBB, LAFB, and ventricular pacing all alter normal voltage distribution. Do NOT apply standard Cornell criteria in LBBB (specific criteria exist but not widely validated).

Age-related voltage decline: QRS voltages decrease with age (>65 years) due to increased chest wall impedance (fat, muscle atrophy). Sensitivity of voltage criteria declines in elderly.

Poor R wave progression (PRWP) – When V3 has a small R wave and deep S wave, the SV3 measurement is large (increasing Cornell sum, potentially false positive). In PRWP, the transition zone is delayed (V5-V6). True LVH usually has preserved transition (V3-V4).

Cornell vs Sokolow-Lyon: Detailed Comparison

Aspect	Cornell Voltage	Sokolow-Lyon	Winner
Sex-specific thresholds	Yes (Men >20 mm, Women >28 mm)	No (single ≥35 mm for both sexes)	Cornell (far superior in women)
Sensitivity (Echo-LVH, 90% specificity)	~45% (men), ~42% (women)	~38% (men), ~23% (women)	Cornell (especially in women)
Specificity (avoiding false positives)	~92-95%	~85-90%	Cornell (better)
Test-retest reproducibility	Good (less variation than Sokolow-Lyon)	Fair (Sokolow-Lyon varies more between ECGs)	Cornell
Ease of calculation	Requires summing two leads (aVL and V3)	Requires summing two leads (V1 and V5/V6)	Tie (both simple)
Obesity impact	Reduced sensitivity (obesity attenuates all voltages)	Significantly reduced sensitivity	Cornell (less affected because V3 is anterior, not lateral)
Black patients (race-specific)	Not race-adjusted; may overcall LVH in Black men (higher voltages at baseline)	Known to have higher voltages in Black individuals; thresholds may need adjustment	Neither (both need race-specific normative data)
Correlation with echo LV mass index	r = 0.55-0.65 (moderate)	r = 0.45-0.60 (moderate)	Cornell (slightly better)

Next Steps

Step-by-Step Clinical Actions Based on Cornell Criteria

Prognostic Implications of Cornell-Positive LVH

Finding	Risk (vs normal ECG)	Management Implication	Evidence
Any ECG LVH (Cornell or Sokolow positive)	2-3x increased risk of CV death, MI, stroke, heart failure	Aggressive risk factor modification. BP target <130/80. Consider aspirin/statins if 10-year ASCVD risk >7.5%.	Framingham Heart Study, LIFE trial
LVH with strain pattern (ST/T changes)	3-4x increased risk (highest risk group)	Urgent BP optimization. Referral to cardiology. TTE within 2-4 weeks.	Devereux RB, 2000; Okin PM, LIFE 2006
Regression of Cornell voltage (decrease ≥5mm over 12 months)	~40% reduction in CV events compared to no regression	Continue current therapy (RAAS blockade effective for LVH regression). Monitor BP to goal.	LIFE trial (losartan vs atenolol)
Progression of Cornell voltage (increase ≥5mm over 12 months)	~2x increased risk of CV events despite treatment	Intensify BP treatment. Add second or third agent. Reassess for secondary hypertension.	LIFE trial
Cornell product >4000 mm·ms in hypertensive patient	~5x increased risk of heart failure hospitalization	TTE within 2 weeks, BNP monitoring, cardiology referral. Low threshold for admission if dyspnea.	Okin PM, JACC 2007

Special Populations: Modified Thresholds

Population	Cornell Threshold Modification	Rationale	Recommendation
Obesity (BMI ≥30 kg/m²)	Cornell voltage reduced (attenuation). Use Peguero-Lo Presti criteria (deepest S + SV4) >28 mm (men) or >23 mm (women)	Adipose tissue acts as electrical insulator; voltages decrease. Cornell may have false negatives.	Use Peguero-Lo Presti as adjunct (higher sensitivity in obesity) OR interpret Cornell only if markedly positive (>25 mm men, >35 mm women)
Black/African American patients	Higher baseline voltages; some suggest raising thresholds: Men >25 mm, Women >32 mm	Black patients have higher QRS voltages at same LV mass due to chest wall configuration (less adipose, more muscle).	No formal race-specific thresholds endorsed. Use clinical judgment: Cornell mildly positive + no strain pattern in asymptomatic Black patient may be normal variant.
Elderly (≥75 years)	Voltages decline with age; Cornell >17 mm (men) may be abnormal given age-related voltage drop.	Age-related increase in chest wall impedance (fat, muscle atrophy) reduces voltages. Fixed thresholds underdiagnose LVH in elderly.	No formal age-adjusted thresholds. If clinical suspicion high (severe HTN, HF symptoms, murmur), obtain TTE regardless of ECG.
Athletes (competitive)	Physiologic LVH common. Do NOT diagnose LVH based on voltage alone. Must have ST/T wave changes, left atrial enlargement, or pathologic Q waves to suspect HCM.	Athletes have increased LV cavity size (eccentric hypertrophy) unlike pathologic LVH (concentric). Cornell may be positive in 15-30% of endurance athletes (rowers, cyclists, marathon runners, soccer players).	Obtain TTE if: (1) Family history of HCM/SCD, (2) Murmur, palpitations, syncope, (3) ST/T wave inversion in inferior/lateral leads (>2 leads), (4) Q waves (pathologic). Otherwise, no further testing needed.
COPD (hyperinflated lungs)	Voltages reduced (diagnostic challenge). Use Peguero-Lo Presti criteria. Cornell rarely positive even with severe LVH.	Hyperinflated lungs increase distance from heart to chest wall electrodes; attenuation effect. Cornell sensitivity <20% in COPD with echo-proven LVH.	Do NOT rely on ECG to rule out LVH in COPD. Obtain TTE for diagnostic confirmation if symptoms (dyspnea) or risk factors (HTN, known CAD).

Origins & History

Historical Development

The Cornell voltage criteria were developed at Cornell University Medical College (now Weill Cornell Medicine) in New York City between 1988 and 1991. Dr. Richard B. Devereux, a cardiologist and echocardiography expert, led the Hypertension Center at Cornell. He noted that existing LVH criteria (Sokolow-Lyon, 1949; Romhilt-Estes, 1968) were derived predominantly in young men and performed poorly in women, elderly, and hypertensive patients. Dr. Paul N. Casale (Cardiology Fellow) and Dr. Peter Kligfield (ECG expert) collaborated with Devereux to develop improved, sex-specific ECG criteria with better sensitivity for echo-proven LVH. Using 183 patients (derivation cohort) and 92 patients (validation cohort), they tested dozens of lead combinations. The combination of RaVL (limb lead, affected by frontal plane hypertrophy) and SV3 (precordial lead, affected by horizontal plane hypertrophy) emerged as the best at balancing sensitivity and specificity. The sex-specific thresholds (men 20 mm, women 28 mm) reflected lower baseline voltages in women (smaller body habitus, less LV mass). The product with QRS duration (Cornell product) was concurrently proposed for patients with intraventricular conduction delay. The seminal 1991 Circulation paper has been cited over 2,500 times and transformed clinical practice, establishing Cornell as the preferred ECG LVH criterion, especially in women.

Key Contributors and Timeline

Year	Contributor(s)	Institution	Contribution
1949	Sokolow M, Lyon TP	University of California, San Francisco	Sokolow-Lyon criteria: SV1 + RV5/6 ≥35 mm. Dominant LVH criterion for 40+ years. Not sex-specific, lower sensitivity in women.
1968	Romhilt DW, Estes EH	Duke University	Romhilt-Estes point score system. Higher specificity (autopsy-validated) but complex. Still used for research.
1991	Casale PN, Devereux RB, Kligfield P, et al.	Cornell University Medical College (Weill Cornell Medicine), New York	Original description of Cornell voltage criteria with sex-specific thresholds. Higher sensitivity than Sokolow-Lyon (validated with echocardiographic LV mass index). Cornell product (× QRS duration) also described.
2001-2004	LIFE Trial Investigators (Okin PM, Devereux RB, Dahlöf B, et al.)	Multiple international sites	LIFE trial (n=9,193) validated Cornell criteria as predictor of CV events and demonstrated LVH regression with RAAS blockade improves outcomes. Cornell criteria entered mainstream clinical practice.
2007	Pewsner D, Jüni P, Egger M, et al.	University of Bern, Switzerland	Meta-analysis confirming Cornell sensitivity 48% vs 34% for Sokolow-Lyon, identical specificity (~88-90%). Cemented Cornell as preferred ECG LVH criterion.
2017	Peguero JG, Lo Presti S, et al.	Montefiore Medical Center, New York	Alternative Peguero-Lo Presti criteria (deepest S + SV4). Higher sensitivity than Cornell, especially in obesity and low voltage states.
2018	ACC/AHA and ESC/ESH Hypertension Guidelines	American College of Cardiology / European Society of Cardiology	Both guidelines endorse ECG LVH assessment (Cornell or Sokolow-Lyon). Class I recommendation for all hypertensive patients.

Limitations of Original Cornell Derivation Study

Derivation cohort predominantly white and male (86% men). Sex-specific thresholds validated but limited racial/ethnic diversity. Unclear generalizability to Black, Hispanic, Asian populations (Black individuals have higher voltages, Asians have lower voltages).

Echocardiographic LVH reference standard (not autopsy). Echocardiography overestimates LV mass compared to autopsy (systematic error ~10-15%). However, clinical practice uses echo, so appropriate comparator.

LVH defined by LV mass index >131 g/m² (men) and >100 g/m² (women) – These thresholds are now considered too high (current guidelines: men >115 g/m², women >95 g/m²). Modern echo thresholds would likely increase sensitivity (detect milder LVH).

Exclusion of patients with LBBB, prior MI, poor ECG quality – Limits applicability to real-world populations with conduction disease or prior infarction. No LBBB-specific LVH criteria provided.

Modest sensitivity (54%) even in derivation cohort – Means nearly half of echo-proven LVH patients have normal ECG. ECG cannot rule out LVH.

Last Comprehensive Review: 2026-07-17