For years we have dutifully told our patients to only get their lipid levels measured in the fasting state. We believed that a meal would artificially elevate their levels and lead us to make the wrong treatment decisions. For diabetes, we measure fasting and 2 hour postprandial blood sugars, but for cholesterol it’s fasting or come back another time. In fast, we have been trained to ignore a non-fasting result. Our lab technicians are so well indoctrinated; they dutifully turn away patients who have eaten.
Studies from 2008 and 2009 suggest that we are wrong.
- In a 2009 JAMA metanalysis of 68 studies involving 302,430 patients without vascular disease, hazard ratios for cardiovascular disease were at least as strong in patients who did not fast as those who did (JAMA 2009 Nov 11;302(18):1993).
- Two cohort studies from Denmark in 2008 involving 33,391 patients showed that lipid levels varied very little between fasting and non-fasting patients (Circulation 2008 Nov 11;118(20):2047):
- Total cholesterol: -0.2 mmol/L (-7.8 mg/dL) with p < 0.001
- LDL: -0.2 mmol/L (-7.8 mg/dL) with p < 0.001
- HDL: -0.1 mmol/L (-3.9 mg/dL) with p < 0.001
- Triglycerides: 0.3 mmol/L (26.7 mg/dL) with p < 0.001
After reading these studies, I have changed my habits. It is more important to get the test done, than worry about when my patients ate — and the evidence supports this.