Cholesterol, MESA and EBCT

Thanks to Dave Lull for a copy of the latest MESA paper. I guess you could just say EBCT looks like a good predictor of heart attack risk, irrespective of age or ethnicity and leave it at that. It looks to be true.

I was just browsing through table one of the results and a few strange things came out. First is that it doesn't include LDL cholesterol at baseline! Most people with any shred of intelligence will realise that calculated LDL cholesterol is garbage, but it's not typical of the average cardiologist to get that far.... These guys must be good!

The second gem was that HDL was lower in the coronary event group, as happens in Framingham. But the difference between the means was 51.0 minus 47.3, I make that 3.7mg/dl. The standard deviation was 15 ish in both groups. If they had a Normal Distribution, which they should have, the range to include most participants is about three standard deviations around the mean. The overlap is phenomenal between heart attack and non heart attack groups. On a personal basis having an HDL of 20 could easily put you in either group, as could an HDL of 60. I guess cardiologists just get used to this.

Taking lipid lowering medication appears to be as bad as being a current of former smoker. Must be a message there but I can't work it out...

Anyone tempted to crash diet might be disappointed by the BMI in the coronary event group being .2 units lower than the spared group. Good job this has neither biological or statistical significance, or participants might have to be encouraged to gain a few pounds. Or get shorter!

I know all of these things need to be factored in to correct for know interactions as it's likely that all of the diabetics were on statins while being encouraged to eat healthy starches and shoot up with insulin, but the raw data still give me a chuckle.

So Detrano et al provide more evidence that coronary calcium, the physical evidence of on going disease, is a better marker of a physical on going disease process than a surrogate of any sort. They also seem less wedded to conventional risk factors. But I guess that's why they're doing the study...

Peter