Left ventricular hypertrophy on electrocardiogram: prognostic implications from a 10-year cohort study of older subjects: a report from the Bronx Longitudinal Aging Study

Abstract

Objective: The objective of this study was to report on the prevalence, incidence and prognosis of left ventricular hypertrophy (LVH) on the electrocardiogram (ECG) in a cohort of ambulatory older men and women.

Design: A prospective, longitudinal study of 10 years duration with ECGs obtained at baseline and on an annual basis.

Setting and patients: A community-based cohort study consisting of 459 subjects (aged 75-85, mean age 79 years).

Measurements: Baseline and follow up ECGs were interpreted using the Minnesota Code. Prevalence and incidence of LVH and ECG were determined as well as regression of ECG LVH. Clinical event rates measured were incidence of total mortality, myocardial infarction (MI, fatal and non-fatal), cardiovascular mortality, cardiovascular disease (fatal and non-fatal), stroke (fatal and non-fatal), all-cause dementia, and multi-infarct dementia. Differences in event rates between groups (those subjects with and without LVH) were compared as tests between proportions. A Cox Proportional Hazards Regression Analysis was performed to compare the relative independent predictive values of different competing factors, including age, gender, serum cholesterol, digitalis use, body mass, index, Blessed Dementia Scale, cigarette smoking, LVH at baseline, LVH ar baseline (persisting), new LVH, new LVH (persisting), new LVH (regressed), previous MI by history of ECG, hypertension by history, and cardiomegaly by X-ray (cardiothoracic ratio > or = 50%).

Results: At baseline, 9.2% of subjects (n = 42) had LVH on ECG and a mortality rate of 11.7/100 persons years versus 4.9/100 persons years for subjects without baseline LVH (P < .0001), and MI rate of 7.5/100 persons years with LVH versus 2.6/100 persons years without LVH (P < .0001), and a cardiovascular mortality rate of 7.2/100 persons years without LVH versus 2.7/100 person years without LVH. Subjects who developed new LVH on ECG (n = 39) had a mortality rate of 14.4/100 person-years compared with 4.4/100 person-years for those without LVH (P < .0001), a cardiovascular mortality rate of 11.1/100 person years versus 2.0/100 person years without LVH (P < .0001), and an MI rate of 6.1/100 person years versus 2.0/100 person years without LVH (P < .01). Subjects in whom the ECG LVH pattern disappeared over time had fewer cardiovascular mortal and morbid events than those with persistent LVH. According to the regression analyses, persistent LVH from baseline was an independent predictor of MI, overall cardiovascular disease, and total mortality. Newly developing LVH with subsequent regression was an independent predictor of overall cardiovascular disease and death.

Conclusions: An increased prevalence and incidence of LVH on ECG, irrespective cause, is associated with a poor prognosis in very old men and women. Regression of ECG LVH in older people, irrespective of cause, may confer improvement in risk for cardiovascular disease.