Cardiac magnetic resonance (CMR) height-based allometric scaling in Southeast Asians; implications in the diagnosis and prognosis of left ventricular hypertrophy (LVH): the REMODEL study

Abstract Background Left ventricular hypertrophy (LVH) is a strong predictor of adverse cardiovascular outcomes (1). Although normalizing LV mass (LVM) to height exponents has been shown to reduce variability from body size, specific recommendations for height exponents are lacking due to the scarcity of normal cohorts to define appropriate height exponents (2). Cardiac magnetic resonance (CMR) is increasingly used in assessing LVM, and the existing body of research is based primarily on the MESA cohort (3). While the MESA cohort is ethnically diverse, the intrinsic effects of ethnicity may not fully be discerned if certain ethnic groups are under-represented. Purpose We aimed to demonstrate the diagnostic and prognostic implications of establishing height exponents specific to sex in a Southeast Asian cohort. Methods Non-overweight/obese Southeast Asian healthy volunteers (n=416) were used to establish appropriate height exponents and normal reference ranges of LVM. The impact of these height exponents was examined in a separate cohort of Southeast Asians with hypertension (n=878). All individuals underwent standardized cardiovascular magnetic resonance imaging (CMR). Primary outcome was a composite of acute coronary syndrome, heart failure hospitalization, stroke, and cardiovascular mortality. Cox regression analysis of various LVH measures were adjusted for 24-hour systolic blood pressure, diabetes, and hyperlipidemia. Results The height exponents for healthy females and males were 1.57 and 2.33, respectively. In patients with hypertension, sex-specific height exponents significantly increased the detection of LVH compared with body surface area (BSA) indexing (47% [n=409/878] versus 27% [n= 236/878], p<0.001). Most individuals re-classified to LVH with sex-specific height exponents were overweight (33.1%, n=58/175) or obese (64.6%, n=113/175). There were 37 primary events over 60 [37-73] months of follow-up. Regardless of indexing method, LVH was independently associated with increased adverse events (sex-specific height exponent hazard ratio (HR): 2.8 [1.25-6.29], P=0.013). The prediction ability of LVH diagnosis by sex-specific height exponents was similar to BSA indexing (HR: 5.43 [2.49-11.8], P<0.001). Conclusions Reference ranges specific to the ethnicity, sex and imaging modality are necessary to establish appropriate height exponents. Although utilizing sex-specific height exponents to index LVM resulted in significantly more LVH reclassification when compared with BSA indexing, this did not translate to a notable improvement in event prediction. Further research into alternative methods of LVM indexing is warranted.Abstract SummaryPredictive Ability of LVH Measures