Comparing amyloid deposits in Alzheimers to those in diabetes and senile cardiomyopathy

Attention conservation notice: A narrow and probably biased comparison of three diseases characterized by the deposition of amyloid in tissues, to clarify points I was confused about while first learning about them. My confusion is not a property of the world.

One of the tidbits I will take away from the first two years of medical school is that Alzheimers is not the only disease with amyloid deposits. Far from it! The two other amyloidoses that most caught my attention, likely because they are the most common, were type 2 diabetes mellitus and systemic senile amyloidosis. The latter is sometimes, and in my view more aptly, called senile amyloid cardiomyopathy, because it almost exclusively affects myocytes of the heart.

1) Type 2 diabetes: Amyloid deposits in pancreatic islets can be seen in type 2, but not type 1, diabetes. They can also be found in non-diabetic elderly people, at somewhat lower frequencies. Here is a table with a summary of many studies making this comparison in autopsy studies:

Zhao et al, 2008; Higher Islet Amyloid Load in Men Than in Women With Type 2 Diabetes Mellitus [1]

If you add up the major sources from above, you get a positive likelihood ratio of about 4-5 for amyloid to be found in the islets of diabetic vs non-diabetic people.

However, according to some reports, they are usually only found extensively in those with diabetes [2].

Finally, whether islet amyloid deposits can contribute to the beta-cell failure that can be seen in later stages of type 2 diabetes seems to be hotly debated, with some papers (eg, [2]) arguing no, and other, perhaps more recent, papers (eg, [3]) arguing yes. The classic story we learned in med school (and, eg, can be found in First Aid for the USMLE Step 1) was that the amyloid does not contribute to pathology in type 2 diabetes, but classic stories are by definition lagging indicators of actual consensus.

2) Senile amyloid cardiomyopathy: Here the story appears to be simpler. Basically, wild-type transthyretin can deposit in the walls of the atria and ventricles, and this is a common and mostly asymptomatic finding at autopsy in ~ 25% of people > 80 years old. It only becomes symptomatic when there is so much amyloid deposition that ventricular wall function becomes impaired, leading to heart failure [4]. So, here, at least, plaque lesion load correlates with and is the likely cause of impaired function.

3) Alzheimers: If you have amyloid plaques at autopsy, what is the probability that you also have Alzheimers disease-related dementia? That question turned out to be surprisingly hard for me to answer, but this study offers an estimate:

Monsell et al., Comparison of symptomatic and asymptomatic persons with Alzheimer disease neuropathology

Monsell et al., Comparison of symptomatic and asymptomatic persons with Alzheimer disease neuropathology [5]

Importantly, these authors did not include the 339 patients without any amyloid plaques. However, if we assume that none of those patients had symptomatic AD, that all of the symptomatic patients are symptomatic due to AD, and if we threshold cutoff for “presence of plaques” at “frequent”, then we can compute a positive likelihood ratio estimate of ~ 10 (57% / 5.5%).*

Whether the asymptomatic patients with amyloid plaques have preclinical AD and would therefore inevitably progress to full-blown AD in the absence of other problems is a hotly debated topic [eg, see 6].

Importantly, this likelihood ratio estimate may be a slight overestimate. Consider the 339 patients without amyloid plaques, and imagine that some of them had AD-like dementia. The standard explanation for this would be that they probably had an atypical presentation of a different type of dementia, such as vascular or normal pressure hydrocephalus, instead. But if the amyloid hypothesis is truly wrong, is it possible that, like type 2 diabetes, some of these patients actually did have dementia as a result of whatever cause root cause leads to AD, just without the potential epiphenomenon of amyloid deposition? By definition, no, but of course definitions do not always carve reality appropriately at its joints.

*: Tau tangles would have a higher likelihood ratio and are the basis for Braak staging, but that is not the subject of this post.

[1]: Zhao, et al. 2008 Higher Islet Amyloid Load in Men Than in Women With Type 2 Diabetes Mellitus

[2]: Cornwell GG, et al. 1980 Senile amyloidosis: a protean manifestation of the aging process

[3]: Costes S, et al. 2013 β-Cell failure in type 2 diabetes: a case of asking too much of too few? 

[4]: Falka RH, et al. 2010 Amyloid Heart Disease

[5] Monsell et al. 2013 Comparison of symptomatic and asymptomatic persons with Alzheimer disease neuropathology

[6]: Balasubramanian et al. 2012 Alzheimer disease pathology and longitudinal cognitive performance in the oldest-old with no dementia