I’m of two minds when it comes to blood testing. For myself, I’m not a huge fan of obsessive, frequent testing and optimization. I have a good idea about how to optimize my health through the actions I take and the foods I eat, and by monitoring how I’m feeling, looking, and performing in response. It’s worked well for me. Whenever I do get a checkup or have blood drawn, my numbers are great.
But many people are the opposite. They like to quantify what’s happening under the hood. That’s great, and often necessary. The problem is that there are big problems with many of the most common blood tests.
A glaring problem for almost every blood test are the reference ranges used. What’s wrong with those?
Reference ranges reflect what’s common, not normal. A reference range for a blood test refers to the values possessed by 95% of the normal population.
Reference ranges for blood tests are based on the people who get lab tests. Who gets lab tests? People who go to the doctor, often because there’s something wrong with them. They may not reflect normal ranges for truly healthy people.
Reference ranges are extremely broad, which may give a false sense of security. Having blood sugar on the high end of “normal” isn’t healthy. It actually presages type 2 diabetes.
Another problem inherent to almost every blood test you’ll take is that the result represents a snapshot in time, a brief glimpse at a situation in constant flux. Your cholesterol was elevated today at 12 noon. What does that say about your levels tomorrow at six in the evening? Next week? Nothing. To account for natural fluctuations, get tested at regular intervals and observe the trends.
Let’s get more specific.
“Oh, gee, Tom, steak and eggs for lunch again? You ever get your cholesterol checked?” We hear this all the time, the chorus of pleas that we please go get a “cholesterol test” before we keel over. Say you decide to humor the skeptics. You go get a cholesterol test. What should you watch for?
Most of the time, LDL is calculated, not directly measured. If you have low triglycerides, as is common on low-carb diets, your calculated LDL will be higher than the reality .
LDL-C refers to the passengers in the cars, not the number of cars on the highway. Most basic tests don’t measure LDL particle number. All evidence points to the number of LDL particles being far more predictive  of heart disease risk than the more common LDL-C. More LDL particles means LDL is hanging around in the blood, increasing the chance they’ll become oxidized and atherogenic.
The guidelines aren’t supported in the literature. Many studies have shown a disconnect between supposedly dangerous cholesterol levels and actual heart attacks. In a 2009 study , 75% of people hospitalized for a heart attack had “healthy” cholesterol numbers. A 2016 review  found the lowest levels of LDL were actually associated with higher mortality in the elderly—you know, some of most vulnerable among us.
What can you do?
Request an advanced lipid test. Tests like VAP, NMR, ApoB, and LDL particle number testing all provide deeper insights into the state of your blood lipids.
If you’re stuck with the basic test, take a look at ratios. Total cholesterol/HDL ratio is a good indicator of how long LDL is hanging around in the blood and remains the best standard assessment of heart disease risk. Another good one is triglyceride/HDL ratio, which is a strong surrogate marker for insulin resistance. In both cases, lower is better. An ideal T/HDL ratio is 1:1; 2:1 is about as high as you want to go. An ideal TC/HDL ratio is 3.5:1 or lower.
There are three primary types of blood sugar tests: fasting, postprandial, and hemoglobin A1c (HbA1c). Postprandial blood sugar measures your blood sugar response to eating at various intervals after meal. Fasting blood sugar measures your blood sugar levels at rest, when no food is coming in. HbA1c measures your average blood glucose across the previous three months. They’re all important, but the tests all have issues.
“Normal” might not be normal. According to the American Diabetes Association, a fasting blood sugar (FBG) under 100 is completely normal. It’s safe. It’s fine. Don’t worry, just keep eating your regular diet, and did you get a chance to try the donuts in the waiting room? They only start to worry at 110-125 (pre-diabetic) and above 125 (diabetic).
This may be unwise. Healthy people subjected to continuous glucose monitoring have much lower average blood glucose—89 mg/dl . A 2008 study  found that people with a FBG of 95-99 were 2.33 times more likely to develop diabetes in the future than people on the low-normal end of the scale.
How about HbA1c? A “normal” HbA1c is anything under 5.7. And 6.0 is diabetic. That’s what the reference ranges, which mostly focuses on diabetes. What does the research say? In this study , under 5 was best for heart disease. In this study , anything over 4.6 was associated with an increased risk of heart disease.
That 5.7 HbA1c isn’t looking so great.
Healthy FBG depends on your BMI. At higher FBG levels, higher BMIs are protective. You read that right. A recent study  showed that optimal fasting blood glucose for mortality gradually increased with bodyweight. Low-normal BMIs had the lowest mortality at normal FBG (under 100), moderately overweight BMIs had the lowest mortality at somewhat impaired FBG (100-125), and the highest BMIs had the lowest mortality at diabetic FBG levels (over 125).
The oral glucose tolerance test is unrealistic. The standard way to test postprandial blood sugar is the oral glucose tolerance test: 75 grams of pure glucose in liquid form. Unless you’re downing jumbo Slurpees, you’re not consuming that much pure glucose in a single sitting, so the results may not be relevant.
HbA1c depends on a static red blood cell lifespan. A1c seeks to establish the average level of blood sugar  circulating through your body over the red blood cell’s life cycle, rather than track blood sugar numbers that rapidly fluctuate through the day, week, and month. If we know how long a red blood cell lives, we have an accurate measurement of chronic blood sugar levels. The clinical consensus assumes the lifespan is three months. Is it?
Ironically, people with healthy blood sugar levels might have inflated HbA1c levels. One study  found that folks with normal blood sugar had red blood cells that lived up to 146 days, and RBCs in folks with high blood sugar had life cycles as low as 81 days. For every 1% rise in blood sugar, red blood cell lifespan fell by 6.9 days. In those with better blood sugar control, RBCs lived longer and thus had more time to accumulate sugar and give a bad HbA1c reading. In people with poorer blood sugar control, red blood cells live shorter lives and have less time to accumulate sugar, potentially giving them “better” HbA1c numbers.
Anemia can inflate HbA1c. Anemia depresses the production of red blood cells. If you have fewer red blood cells in circulation, the ones you do have accumulate more sugar since there are fewer cells “competing” for it.
If you’re very low-carb, postprandial blood glucose will be elevated. This is because very low-carb, high-fat diets produce physiological insulin resistance  to preserve what little glucose you have for the tissues that depend on it, like certain parts of the brain. The more resistant you are to insulin, the higher your blood glucose.
What can you do?
If you need to pass a test, 150 grams of carbs a day will do it. Eat that way for three days to a week before your OGGT.
Try meals, not sugar solutions. Mixed meals of fat, protein, and carbs are better candidates for testing one’s real-world glucose response.
Take all three measurements into account. If your HbA1c is higher than you’d like but you ace all the postprandial tests, you’re probably fine.
Follow trends. Even if your red blood cells are centenarians, you can track the trend of HbA1c with multiple readings.
The most common liver enzymes you’ll test are alanine transaminase (ALT) and aspartate transaminase (AST). Another one is gamma-glutamyl transpeptidase, or GGT. When the liver is damaged or inflamed, liver enzymes generally go up. They’re usually pretty straightforward—more straightforward than the cholesterol and blood sugar tests—but not completely.
Weight loss can change them. Weight loss transiently increases ALT and AST in women and decreases them in men . If you’re in the process of losing weight, you can safely ignore small improvements or increases in liver enzymes.
Normal isn’t healthy. The normal range for GGT is 8 to 65 U/L, yet men under 70 years of age with GGT levels over 38—right smack in the middle of “normal”—have an increased risk of heart failure . Another study  found that even low-normal levels of GGT were associated with an increased incidence of atrial fibrillation, a harbinger of more serious cardiovascular conditions.
What can you do?
Just be aware. Don’t fear weight loss for transient changes to your ALT/AST, and don’t rest on your laurels if your GGT looks “normal.”
Testing your cholesterol, blood sugar, and liver enzymes is helpful, informative, and often necessary—but only if you’re aware of the limitations and you know how to account for context.
Thanks for reading, everyone. Take care!