This is an unusually thoughtful synthesis, because it treats “lithium” not as a psychiatric drug headline, but as a trace element with plausible systems biology relevance to brain aging.
What I found most compelling is the reframing: endogenous lithium appears to be present in the brain, and in the Harvard/Rush datasets it’s reduced already at the MCI stage, while blood levels don’t necessarily track brain bioavailability. Mechanistically, the idea that amyloid can sequester lithium (turning plaques into a functional “sink”) provides a concrete bridge between the amyloid story and the long-standing puzzle of why pathology and symptoms don’t map 1:1. 
The downstream biology you highlight also fits what we see across neurodegeneration: lithium deficiency in models is associated with glial activation, synaptic and myelin-related injury signatures, and transcriptomic shifts overlapping with Alzheimer’s, suggesting a broad “network fragility” effect rather than a single-pathway tweak. 
Two clinician-facing nuances feel important:
1. This is not a green light for self-supplementation. Lithium has a narrow safety margin at pharmacologic doses, with renal/thyroid risks, drug interactions, and special concerns in older adults, so translating “endogenous depletion” into “replacement” needs properly designed human trials and careful formulation/dosing work. 
2. The biomarker implication is intriguing (brain-specific depletion, early change), but we’ll need clarity on what to measure (serum vs other compartments), how it behaves longitudinally, and how it performs against established risk markers. 
Overall: this is exactly the kind of hypothesis that can sharpen the field; less ideology (“amyloid vs not”) and more physiology (“what becomes unavailable, when, and to which cell types?”). If the lithium-bioavailability story holds up, it could open a prevention lane that’s both mechanistically grounded and clinically testable.
Excellent article. Curious, have you looked into whether dietary lithium from drinking water sources shows any meaningful correlation with cognitive outcomes in APOE4 carriers specifically?
Thank you, Doc! I have looked into it and couldn’t find anything stratified to carriers. I think we are in the early days of having a grasp on ApoE4 testing in asymptomatic carriers and all the lithium in water studies I found are at the population-level so this is why we don’t have good stratification by genotype. And I wonder how many people in regions with above average lithium levels in the water are now filtering their water. I would bet a lot more are filtering the lithium out than in years past…
Nutritional Lithium orotate has been used for decades without adverse side effects. It is the dose that makes the poison. In contrast to prescription lithium, nutritional lithium has a very wide therapeutic range sine you are taking only 2-7% of therapeutic doses..
I still wouldn’t care to call RX lithium a poison, but it is an incredible drug and there are precautions prescribers should take to monitor it’s therapeutic window. I’m personally glad to live in a time where we have this mechanistic insight into lithium “deficiency” and that there is an OTC option available to those of us interested!
I do prescribe RX lithium in my psychiatric practice so feel the need to defend it. For the right patient’s it can be lifesaving! But they need to commit to blood monitoring and the risk of thyroid issues….
This is an unusually thoughtful synthesis, because it treats “lithium” not as a psychiatric drug headline, but as a trace element with plausible systems biology relevance to brain aging.
What I found most compelling is the reframing: endogenous lithium appears to be present in the brain, and in the Harvard/Rush datasets it’s reduced already at the MCI stage, while blood levels don’t necessarily track brain bioavailability. Mechanistically, the idea that amyloid can sequester lithium (turning plaques into a functional “sink”) provides a concrete bridge between the amyloid story and the long-standing puzzle of why pathology and symptoms don’t map 1:1. 
The downstream biology you highlight also fits what we see across neurodegeneration: lithium deficiency in models is associated with glial activation, synaptic and myelin-related injury signatures, and transcriptomic shifts overlapping with Alzheimer’s, suggesting a broad “network fragility” effect rather than a single-pathway tweak. 
Two clinician-facing nuances feel important:
1. This is not a green light for self-supplementation. Lithium has a narrow safety margin at pharmacologic doses, with renal/thyroid risks, drug interactions, and special concerns in older adults, so translating “endogenous depletion” into “replacement” needs properly designed human trials and careful formulation/dosing work. 
2. The biomarker implication is intriguing (brain-specific depletion, early change), but we’ll need clarity on what to measure (serum vs other compartments), how it behaves longitudinally, and how it performs against established risk markers. 
Overall: this is exactly the kind of hypothesis that can sharpen the field; less ideology (“amyloid vs not”) and more physiology (“what becomes unavailable, when, and to which cell types?”). If the lithium-bioavailability story holds up, it could open a prevention lane that’s both mechanistically grounded and clinically testable.
Excellent article. Curious, have you looked into whether dietary lithium from drinking water sources shows any meaningful correlation with cognitive outcomes in APOE4 carriers specifically?
Thank you, Doc! I have looked into it and couldn’t find anything stratified to carriers. I think we are in the early days of having a grasp on ApoE4 testing in asymptomatic carriers and all the lithium in water studies I found are at the population-level so this is why we don’t have good stratification by genotype. And I wonder how many people in regions with above average lithium levels in the water are now filtering their water. I would bet a lot more are filtering the lithium out than in years past…
Nutritional Lithium orotate has been used for decades without adverse side effects. It is the dose that makes the poison. In contrast to prescription lithium, nutritional lithium has a very wide therapeutic range sine you are taking only 2-7% of therapeutic doses..
I still wouldn’t care to call RX lithium a poison, but it is an incredible drug and there are precautions prescribers should take to monitor it’s therapeutic window. I’m personally glad to live in a time where we have this mechanistic insight into lithium “deficiency” and that there is an OTC option available to those of us interested!
I do prescribe RX lithium in my psychiatric practice so feel the need to defend it. For the right patient’s it can be lifesaving! But they need to commit to blood monitoring and the risk of thyroid issues….