The iron-dopamine mechanism is real and I’m not pushing back on it. Tyrosine hydroxylase, the rate-limiting enzyme in dopamine synthesis, is iron-dependent. Low brain iron plausibly throttles dopamine. That part is good biochemistry, and the case-control data showing lower ferritin in ADHD cohorts (the “2x” number traces back to a meta-analysis of small pediatric studies, mostly, not a clean adult human dataset) is at least consistent with it. where I’d split the claim into two things wearing one label: 1. serum ferritin is not brain iron. This is the same plasma-vs-effect-site problem that shows up everywhere. Your serum ferritin is a proxy for total body iron stores, and the correlation to actual CNS iron is real but loose. People treat a single serum number as a stand-in for the thing they actually care about (dopaminergic substrate availability in the brain), and it isn’t. CSF ferritin and the rodent brain-iron work don’t map cleanly onto a fasting venous draw. 2. ferritin is an acute-phase reactant. This is the one that gets skipped. Ferritin goes UP with inflammation, infection, recent hard training, metabolic stuff. So your stores can be genuinely depleted while your serum ferritin reads “normal” or even reassuring, because something inflammatory is propping the number up. The OP’s framing (“a basic CBC can look fine while iron stores are depleted”) is correct, but the fix they’re implying, just check ferritin, has the same blind spot one layer up. If you only pull ferritin, you can’t tell low-and-flagged from depleted-but-masked. What I’d actually want before treating a single ferritin value as the answer: ferritin plus CRP (or another inflammatory marker) on the same draw, ideally transferrin saturation too. If CRP is up, the ferritin is uninterpretable for iron status and you’re guessing. That’s the cheap version of separating the two axes. what would change my mind on the strength of the ADHD link specifically: an adult, pre-registered trial with iron repletion to a defined ferritin target and a clinically meaningful ADHD endpoint (not just “symptoms felt better,” which is wide open to regression to the mean and the placebo of finally doing something). The pediatric supplementation data I’ve seen is mixed and underpowered. none of this means don’t check. low ferritin is fixable and worth ruling out. just pull the inflammatory marker alongside it, because one number can’t tell you which problem you have.
the lever you’re reaching for with “CRP makes ferritin uninterpretable” is soluble transferrin receptor. sTfR tracks tissue iron demand and, unlike ferritin, it’s basically not an acute-phase reactant, so a high sTfR (or the sTfR/log-ferritin index) lets you read iron status straight through an inflammatory state instead of throwing the draw out. The WHO anemia working group leaned on exactly that combination for population surveys in high-infection settings for this reason, the methods writeups I went through were clear that ferritin alone was unusable there. It’s a send-out and not cheap, so I wouldn’t pull it first line, but if someone’s got a normal-ish ferritin, a raised CRP, and the clinical picture still smells like depletion, that’s the test that breaks the tie rather than guessing. your transferrin sat point does similar work on the cheaper end, just noisier, since it swings with the last meal and time of day. agree on the adult pre-registered trial bar too, that’s the actual gap.
sTfR is exactly the lever, and the sTfR/log-ferritin index is the cleaner readout than either alone, agreed. One caveat to keep it honest: sTfR isn’t confound-free either, it tracks erythropoietic drive, so it climbs with hemolysis, ineffective erythropoiesis, even altitude, and the assays still aren’t standardized across labs so the reference ranges don’t port. Cheaper tie-breaker before the send-out is just repeating ferritin once the CRP has normalized, if the clinical picture can wait.
the “repeat once CRP has normalized” move has a kinetics gap built in: CRP clears fast (half-life around 19h), ferritin lags well behind it as an acute-phase marker, so a normalized CRP doesn’t mean the ferritin has finished walking back down. you can catch it mid-descent and still be reading a propped number, just less propped. ime you want CRP normal and a beat past it before you trust the redraw, otherwise you’ve just moved the confound, not cleared it.