What is the best way to measure vitamin B12?
Total B12 alone misses too many cases; ideally combine it with MMA, and consider active B12 (holotranscobalamin) as a supplement. There is no perfect test, so always discuss your results together with your symptoms with your doctor.
Serum B12 alone misses roughly half of people with a deficiency. Someone who scores normal on the standard test can still have a functional B12 deficiency. This makes the standard measurement unreliable as a stand-alone test.
In everyday clinical practice, a combination of total B12 and methylmalonic acid (MMA) is the most useful approach. MMA is a metabolic by-product that rises early in B12 deficiency, even when B12 in the blood still appears normal. Note that MMA is less reliable in older people and in those with reduced kidney function. There is no perfect test.
Holotranscobalamin, also known as 'active B12', is the fraction of B12 that cells can actually use. In research involving healthy older adults, low active B12 was more strongly associated with measurable neurological abnormalities (slower nerve conduction, more abnormalities on brain MRI) than total B12. This makes active B12 an interesting addition, but this measurement also has its limitations.
An unexpectedly high B12 level also deserves attention. It can sometimes be a signal of underlying disease, including cancer. The method in which proteins are precipitated with polyethylene glycol to explain high values is not reliable and is not recommended. For high values, further blood testing of the B12 transport proteins may be considered.
If you are already taking high-dose B12 supplements or receiving injections, blood tests are of little value as a monitoring tool. Adjusting injections based on blood values is not recommended. Whether the treatment is working should be judged by improvement in symptoms, not by numbers in the blood.
All claims are based on associational research; no RCTs are available that directly compare diagnostic strategies. The recommendations reflect expert and review consensus, not randomised trials.