Diminishing returns, quantified
Where the curve flattens, what flattens with it, and what does not.
Explainers
Page 6 of 12 of this archive, newest first.
Page 6
The impurity hiding underneath the parent peak
Where two methods disagree, the conservative convention is to report the lower figure. It is not universal, and whether a laboratory follows it belongs on the report.
Units, millilitres, milligrams: three scales on one small object
We work through the arithmetic in full, because it is short, and because the errors it prevents are order-of-magnitude errors.
Gallstones during rapid weight loss: drug, or weight loss?
A small number of serious gastrointestinal events occur in people taking these drugs. Distinguishing them from the expected effect profile is the most consequential…
What the trial protocols permitted, and what the labels omitted
The practice is near-universal, clinically sensible, and supported by observational data rather than randomised comparison. We say which is which.
The incretin receptors, ranked by how much we actually know about them
Selectivity, potency and efficacy are three different measurements. The trade routinely reports none of them.
Why purity is the cheapest number to flatter
The Journal has read several hundred certificates from twenty companies. The number is almost always there; the method behind it almost never is.
Site selection when the tissue has already changed
The evidence base here is the insulin injection-technique literature, which is large and transfers well on tissue questions.
The specification sheet says 500,000. The report says conforms.
Calibration drift is real, unremarkable, and the reason serious laboratories run internal standards.
Vagal afferents, the area postrema, and the anatomy of nausea
A tour of the tissues where the receptor is expressed, and what happens in each.
Mapping a rotation that you will actually follow
Repeated injection into the same small area produces lipohypertrophy — thickened, rubbery subcutaneous tissue with blunted and erratic absorption. It is common, it is…
The arithmetic behind every identity confirmation, worked in full
A peptide has a monoisotopic mass and an average mass, they differ by several daltons at this molecular size, and a certificate that does not say which it quotes cannot be…
Three receptors, one peptide: the design problem behind triple agonism
Head-to-head data exists for some of these comparisons and not for others. This piece says which.
When a symptom is information and when it is noise
A dose held long enough stops being a pause and becomes a maintenance decision. That transition is rarely made explicitly.
Do you need the top of the ladder?
The dose-response curve in this class flattens near its top. That has direct consequences for whether the final rung is worth climbing.
The incidence tables, read line by line
Trial adverse-event tables count episodes reported to a study nurse. They are the best data we have and they systematically under-record the mundane.
Six things incretin pharmacology cannot currently predict
Receptor pharmacology is strong on average effects and almost silent on individual variation. That gap is where most reader questions live.
How a correct calculation becomes a wrong dose
A pharmacist catches most of these in licensed practice. In this market there is no pharmacist, so the checks have to be structural.
Delayed emptying explains some of this, and not all of it
What scintigraphy and breath-test studies established about emptying rate, and what they did not.
The restart nobody plans for
We work through the residual-exposure table so the decision can be made from numbers rather than from feel.