Affiliate disclosure
Book titles on this page link to Amazon. As an Amazon Associate, DataField.Dev earns from qualifying purchases — at no additional cost to you.
Chapter 19 — Further Reading
⚠️ Read the recovered trials themselves
This is the one thing on this page I'd insist on, because both sides describe them and almost nobody reads them.
Ramsden CE et al., Sydney Diet Heart Study reanalysis — BMJ, 2013. Ramsden CE et al., Minnesota Coronary Experiment recovered data — BMJ, 2016.
Both open access. ⚠️ Read the methods and the limitations sections, not the abstracts. The authors are notably careful about what their findings do and don't establish — considerably more careful than most people citing them in either direction. This is what §19.6's 🟡 is made of.
Then read the correspondence and editorials responding to them, which the BMJ publishes alongside. ⚠️ You will see qualified people disagreeing about the same data in real time, which is worth more than any summary of the disagreement.
On what actually changed
Search "adipose tissue linoleic acid content secular trend United States." ⚠️ This is §19.3, the strongest fact on the sceptical side, and it is a real literature. Look for the work compiling adipose biopsy composition across decades.
Then search "linoleic acid intake trends soybean oil food supply twentieth century" for the supply- side data.
⚠️ Read these before reading anyone's argument about them. The measurement is more interesting than the interpretations, and it will inoculate you against both the dismissal ("there's no evidence anything changed") and the over-claim.
On the mechanism, and its testing
Search "dietary linoleic acid inflammatory markers systematic review controlled trials." ⚠️ This is §19.5's first prediction. Note the trial sizes and durations, and note that the outcomes are markers rather than disease — the limitations are real and I'd rather you saw them yourself.
Search "linoleic acid intake tissue arachidonic acid conversion humans." ⚠️ This is the decisive one — the broken link. Look for work examining whether increasing LA raises AA in plasma phospholipids and tissue.
On desaturase regulation: search "FADS1 FADS2 polymorphism fatty acid desaturase activity." ⚠️ Genetic variation in this pathway is substantial, differs markedly between populations, and is almost never mentioned in the popular debate — which is §19.12's empty "in whom" slot, and it's a genuinely interesting research area.
On substitution and cardiovascular outcomes
Cochrane reviews on "reduction in saturated fat intake for cardiovascular disease" — free
plain-language summaries at cochranelibrary.com. ⚠️ Read the certainty ratings. "Moderate
certainty" is doing specific work and the review says exactly what it means.
Then read a critique of the same evidence base. Search for Zoë Harcombe's or Nina Teicholz's published critiques of the saturated fat guidelines. ⚠️ I disagree with parts of both and I think you should read them anyway — they are precise about the methodological weaknesses, and Chapter 9 §9.6's handling of the Keys cherry-picking charge is the model for how to read a critic who is right about the method and wrong about the conclusion.
On oxidation and frying
Search "total polar compounds frying oil degradation regulation." ⚠️ Then find your own jurisdiction's limit and enforcement arrangements. Case Study 2's central point is that this chemistry is settled enough to be a number on an inspection form, and seeing your local number makes that concrete.
On the compounds: search "4-hydroxynonenal frying oil dietary exposure" and "aldehydes repeatedly heated cooking oil." ⚠️ Note how consistently the studied condition is repeated heating — and note where the human outcome evidence stops.
On oxidative stability versus smoke point: search "oxidative stability index cooking oils". ⚠️ This will correct the popular argument in both directions, which is unusual and useful.
On the processing objections
On hexane: search FDA and EFSA assessments of "hexane extraction solvent residue edible oils" for the residue levels and limits. ⚠️ The occupational and environmental case for hexane is separate, legitimate, and worth reading — it just isn't about your cooking oil.
On trans generation in refining: search "trans fatty acid formation deodorization refining vegetable oil." ⚠️ This is §19.10's honest concession and the sceptics raising it are raising something real. Compare the levels to the partially hydrogenated fats Chapter 9 covered — the difference in magnitude is the point.
Both sides, deliberately
For the sceptical case: search for Chris Knobbe's presentations on vegetable oils and chronic disease, and Tucker Goodrich's writing. ⚠️ Read them for the §19.3 material, which they present better than the mainstream does, and watch carefully for where they move from Claim 2 to Claim 4 (§19.2) without new evidence. That transition is the whole thing.
For the mainstream case: search Stephan Guyenet's writing on linoleic acid, and Bill Harris's work on the omega-3 index. ⚠️ Harris's index is the practical answer §19.5 points to — measure the omega-3 end rather than arguing about the ratio.
And read at least one exchange between them, if you can find one. ⚠️ Watching two informed people disagree about the same papers is the single most useful thing available on this topic and it is not replicable by reading either alone.
Practical
Your national food safety agency on domestic frying oil reuse. Short, boring, and it's §19.8's practical content.
TPC test strips — search "cooking oil test strips total polar compounds." ⚠️ A few pounds a pack. If you fry at home more than occasionally, or if you're just curious what your local takeaway is doing, they exist and they're not professional equipment.
On oily fish: your national guidance on portions per week and species advice in pregnancy (mercury limits apply to some species). ⚠️ Two portions a week is 🟢 and cheap — tinned sardines, mackerel, salmon.
A note on what to be careful with
⚠️ This chapter has more low-confidence verdicts than any other in the book, and I'd rather you knew where.
§19.13's confidence table is the honest summary. ⚠️ The part I'd most expect to revise is §19.6 — how to weigh the recovered trials — and the part I'd bet on holding is §19.11's confounding.
Two failure modes, and this chapter can produce either:
1. The sceptic's version: you read §19.3, find it striking, and conclude the case is made. ⚠️ But §19.5 tested the mechanism that connects §19.3 to disease, and the prediction failed. An unprecedented exposure with no demonstrated pathway is a research question, not a finding.
2. The dismisser's version: you read §19.5, conclude the whole thing is nonsense, and stop. ⚠️ But §19.3 is real, §19.8 is real, §19.10's trans concession is real, and §19.11 means the observational data can't exonerate seed oils any more than it can indict them.
The position this chapter actually lands on is: ⚠️ an unprecedented dietary change, a coherent mechanism whose central prediction failed direct testing, a genuine oxidation concern that applies to a different situation than the one people act on, and a confounder so complete that the question may not be answerable with the evidence we're going to get.
That is genuinely unsatisfying. ⚠️ It is also, as far as I can tell, correct — and Chapter 17 §17.14 warned you that ⚗️ and 🟡 are the right answer more often than anyone finds comfortable.
Chapter 22 is where most of the energy in this argument should have been going all along.