Every letter we have printed

Weight-bearing changes as mass changes, and the mechanical loading on the skeleton changes with it. Any bone finding in a mass-loss study is confounded by that mechanically, quite apart from anything pharmacological.

The precision of the instrument, the magnitude of the reported change and the duration of the study should always be given together. Any two of the three without the third leaves the reader unable to judge whether anything was measured at all.

I am sixty-eight, I have lost nineteen kilograms over fourteen months, and my consultant has twice told me my lean mass is fine on the basis of a handheld bioimpedance device in the clinic corridor. Having read your piece on what that device measures, I am no longer sure what I have been reassured about.

The studied interval was chosen partly for adherence, and adherence in a trial is measured and supported in ways it is not outside one. A schedule that works in a study population may not survive contact with an ordinary week.

Splitting an interval changes the peak-to-trough ratio and not the average exposure, which is a distinction with real consequences and no supporting outcome data in this class.

The higher-dose studies that do exist are informative and are frequently misquoted, because the incremental benefit at the top of the range is smaller than the excitement suggests. Printing the incremental figures rather than the totals would settle a good deal of argument.

You keep insisting on peptide content rather than purity when discussing dose certainty. I have looked at a dozen certificates from four different testing services and content is reported on perhaps a third of them. What are readers supposed to do with an absence?

Absolute and relative statements point in opposite directions here and both are true. Lean mass falls in absolute terms and the proportion of the body that is lean can rise. Choosing which to headline is an editorial decision and it should be made visibly.

A participant who is scanned is not a random participant. Substudy enrolment depends on site capability and on willingness, and the resulting sample can differ from the trial population in ways nobody records. It is a small caveat and it belongs in the methods paragraph.

Bone density in these studies is reported over horizons far shorter than the timescale on which bone remodels. A twelve-month change is a signal about an early phase of a slow process, and it is read as a conclusion about the process.

An activity component in a trial is also an attention component, and the two cannot be separated in an unblinded design. Some part of any benefit attributed to training is attributable to contact with the study team.

The signature block is not decoration. A named analyst with a role, over a determination date, is a person who can be asked a question. A scanned signature with no name under it is an image.

Your anatomy diagram should mark which fields are assertions by the manufacturer and which are results from an instrument. On most certificates they sit in the same table in the same typeface, and a reader has no way to tell a measurement from a claim.

A mechanistic argument can be tested only where an outcome study exists to test it against, and in this class several widely repeated mechanistic claims have never met one. Listing which have and which have not would be a useful reference entry.

The limits of what receptor pharmacology can explain are worth stating explicitly in every piece of this kind. A mechanism can be entirely correct and still fail to predict an outcome, and this literature has several such cases.

The published potency ratios for multi-target compounds vary by more than an order of magnitude between molecules described by the same term. The term names an architecture, not a profile.

Desensitisation kinetics differ between ligands at the same receptor, which is a well-established property of this family and almost absent from the popular account. A compound that resists desensitisation behaves differently over weeks regardless of its acute potency.

I teach a seminar that uses two of your pieces as set reading. Students arrive expecting advocacy and are visibly unsettled by the refusal to supply any.

The testing market for this subject is thin because the demand is thin, and the demand is thin because the certificate does not have a field for it. Formats create markets. Add a sterility line to the standard template and somebody will start selling the test within a year.

Research-use framing lets the whole question be avoided legitimately. Material sold for laboratory use has no requirement of this kind attached to it, and the framing is accurate as well as convenient.

Filtration removes organisms and does not remove endotoxin, which passes through a sterilising filter without difficulty. Anybody relying on a filtration step for both properties has been misinformed about the chemistry.

What I would like next is a template. A single page a buyer can send that asks for bioburden, endotoxin, closure integrity and the monitoring trend in the supplier’s own terms. Half the difficulty in this subject is that nobody knows how to phrase the request.

A ratio derived from a substudy of eighty people is being applied to a population of millions in general discussion, and the arithmetic is presented with a confidence the original authors were careful to avoid. The paper is fine; the citation chain is where the damage happens.

What I would most like to see is a piece on what has not been measured. Duration beyond the trial periods, composition after regain, and the effect of a second loss cycle are all open, and the absence of data is itself worth reporting.

As a DXA technologist of twenty-two years I would add one thing to your precision section: the largest source of error in practice is not the machine, it is positioning. A patient scanned with their arms two centimetres further from their trunk will report different regional values. We are trained to a protocol and the protocol is not always followed.

Where a subject is absent from every certificate, buyers infer that it does not matter. Silence is read as reassurance, which is the most consequential property of an absence and the least intended.

I asked four suppliers about this last year and all four answered plainly that it is not part of the standard panel and can be quoted separately. The silence is on the storefront rather than in the correspondence.

Your endotoxin table gives vial 12 at 112 EU per vial and then declines to say whether that is dangerous. I understand why. It is still frustrating to read a figure of that size next to the sentence "arithmetic, not a safety assessment".

You write that recombinant factor C is insensitive to the glucan branch of the cascade. It would be worth adding why anybody cares: cellulose filter media and certain paper wrappings shed glucans, and a laboratory that has chased a false positive through three repeat assays will never willingly go back to a reagent that responds to them.

Signalling downstream of this receptor branches, and a compound can favour one branch. Where a paper reports a single potency it has reported one branch and left the reader to assume the others follow.

Receptor reserve means that a large fraction of receptors can be unoccupied while the maximal response is still obtained, which undercuts every argument that reasons directly from occupancy to effect.

Your table lists orforglipron with a half-life of 29 to 49 hours. That is a wide range to report as a single figure. What accounts for it?

Your accumulation table gives 2.0 for a seven-day half-life at weekly dosing. I make it 2.0 as well, but I would point out that this assumes complete absorption of each dose, which for subcutaneous peptides is a generous assumption.

I have read three separate articles this month describing cagrilintide as a GLP-1 agonist and one describing tirzepatide as "semaglutide with an extra bit". Thank you for the glossary. Please run it again.

Research-use language belongs on the certificate rather than only on the invoice. A document that describes a material without stating what it is not approved for is a document that will be read as broader than it is.

A note on page numbering. Certificates that are not marked page one of two are trivially separable, and the second page is where the method usually lives. I have twice received a first page alone and had no way to know a second existed.

The composite test table you print as a model is unrealistic. No research supplier is going to run headspace GC and ion chromatography on every lot at these price points, and publishing an aspirational document as a benchmark just makes real certificates look worse than they are.

The batch number on the vial matching the batch number on the certificate is the single check I ask new staff to perform, and it takes four seconds. It is not sufficient and it is astonishingly informative, because the documents that fail it fail it obviously.

On the recycling of a genuine document across later lots: this is the failure mode I meet most often in the wider market, and it is the hardest to characterise, because nothing on the page is false. The determination happened. It simply happened to a different batch than the one in the parcel.

Biased signalling is the part of this literature that has moved most in the last few years and is barely present in general coverage. A receptor with more than one downstream pathway can be engaged differently by two ligands with identical binding.