Ahalf century ago Roberts and Wearn'published their classical study on human hearts, which demonstrated that 1) capillaries multiply dur-ing postnatal growth to compensate for the increase in muscle fiber size, 2) muscle-to-capillary ratio attains a value of approximately 1: 1at maturity, and 3) this ratio persists throughout life. They also found that hypertrophied hearts of adults consistently had lower capillary densities than hearts that were not hypertrophied. Since that time, the prevailing view has been that cardiac hypertrophy in humans is accompanied by little or no capillary growth. However, the article by Rakusan and colleagues2 that appears in this issue of Circulation is the first to systematically document capillary parameters with regard to postnatalgrowth and pressure-overload hypertrophy in humans. There are severalimportant new findings in this paper that provideinsights into coronary angiogenesis.

See p 38 One important finding in this study is that the data concerning vascular growth in humanhearts are gener-ally consistent with those obtained from experimental animals. The data on human hearts2 indicate that although capillary growth does not keep pace with the increase in mass of the heart during the transition from infancy to childhood to adulthood, capillary angiogen-esis is not arrested. Thus, a fourfold increase in heart mass from infancy to childhood is associated with only a 27% decrement in capillary density. Similarly, capillary density in rats declines only 23% between 4 and 10 weeks (weaning and puberty, respectively), whereas left ventricular (LV) weight increases 2.6-fold. 3 Moreover, the relation betweencapillary domains and heart mass during postnatal growth is quite similar in the two species. 24 More recently, Smolich and colleaguess have documented a rather marked capillary angiogenesis in sheep during postnatal growth. Their data show a decrease in myocyte-to-capillary ratio from birth to adulthood, a period characterized by nearly a fourfold increase in myocyte cross-sectional area. Data from the human hearts analyzed by Rakusan et a12 indicate that the timing of LV pressure overload