08/31/2026
After reviewing blood work from roughly 2,000 patients at Iron Direct Primary Care, I’ve become much less enthusiastic about relying on tests that require a “perfect” fast.
In real life, patients drink coffee. They forget. They eat late. They assume black coffee doesn’t count. Then we’re left trying to interpret a fasting insulin or lipid panel without knowing how clean the fast actually was.
That creates a logistical problem: a test can be scientifically useful, but still be impractical if the pre-test conditions are unreliable.
For metabolic and cardiovascular screening, I’m increasingly favoring markers that are much less dependent on whether a patient fasted perfectly:
• ApoB — tells us the number of atherogenic lipoprotein particles and is generally far less affected by recent food intake than triglycerides or calculated LDL-C.
• Hemoglobin A1c — reflects average glucose exposure over roughly 2–3 months and does not require fasting.
Fasting insulin can still be useful when it is collected correctly, but caffeine, recent food intake, stress, exercise and other factors can influence the result. Likewise, triglycerides can change substantially after eating, which can affect interpretation of a conventional lipid panel.
My takeaway after years of trying to coordinate fasting labs in a real primary-care population:
The best laboratory test is not just the one with good physiology behind it. It also has to work reliably in real patients.
For many routine visits, that increasingly means I would rather know the ApoB and A1c than spend time debating whether someone was “really fasting.”
And if we specifically need a fasting insulin or fasting triglyceride measurement: 8–12 hours, water only. Even black coffee stays out.
