Every letter we have printed

A fair hit, and we have amended the paragraph to carry the same caveat in both places. The distinction we should have drawn is that the substudy figures are at least attached to a stated population and a stated instrument, which the textbook rule is not. Neither is a constant.

Because grip strength in an observational cohort of people who chose to take a drug, and who differ from those who did not in age, motivation and comorbidity, cannot separate the drug effect from the selection. We report those cohorts and we do not treat them as answering the question.

Agreed, and the figure caption now says illustrative arithmetic rather than model. The point survives the simplification, which is that a small amount of selection produces a large apparent effect, but we should not have dressed a demonstration as an estimate.

That is the correct reading, and the unmeasurable part is the honest part. We would add only that vendor-published reports are not worthless — a vendor willing to commission testing at all is behaving better than one that will not — they are simply weaker in a specific way.

Probably true in part, and it is the strongest argument against our proposal. Our answer is that gaming requires effort and leaves traces, which the present arrangement does not, and that a partially gamed record is more informative than no record. We would not claim more than that.

A fair hit. The tag under which this coverage sits predates the distinction we now draw, and we have added the distinction to the second paragraph and to the table. The department name will follow at the next reorganisation of the site.

A fair objection and we have adjusted the wording. Our intention was to locate the absence with the people who could have funded the trial rather than with the clinicians managing without it, and on rereading the original paragraph did not achieve that.

Agreed, and the standards desk is compiling exactly that. It will appear as a standing reference page rather than inside an article, so that it can be kept current.

Twenty-four months is a common default and “industry standard” is an accurate description of the practice rather than a justification of the claim. The distinction we would press is between an expiry date, which asserts shelf life, and a retest date, which asserts only a review interval. The second is defensible without stability data. The first is not.

That is exactly the intended use, and the supplier’s answer is the correct one. The remaining question is why the relationship between the two codes is not printed on the document, since it takes a line.

A useful qualification and the text has been softened. The observation remains worth making; the inference we drew from it was stronger than the practice supports.

A fair hit. We can describe the pharmacology accurately and still be writing at a distance from how the decision lands, and the paragraph you object to does both. The reframing is offered as a corrective to a stigma, not as a claim that the stigma is imaginary.

Correct on both counts, and the table now carries the steady-state assumption explicitly. Your second point is the more useful one: interrupting mid-escalation clears faster than interrupting from a settled dose, and the practical reading of the table should be adjusted accordingly.

Well put, and we had not written it down. The cost of discarding partially used material is a real input into dosing behaviour in a market where much of the spend is out of pocket, and it deserves treatment in The Ledger rather than a sentence here.

It is not arbitrary — the two maxima come from separate dose-selection programmes with different primary endpoints, and 2.0 mg was established against 1.0 mg in a dedicated glycaemic comparison. We have added a clause. The underlying point, that indication shapes the ladder as much as the molecule does, is worth more space than we gave it.

We can say where it comes from, which is the in-use period established for licensed pen presentations of specific formulations in specific containers. Whether it transfers to a different peptide reconstituted in a different diluent in a different vial is not something the stability literature permits anyone to assert. Declining to guess is not agnosticism; it is the difference between a study and a convention.

Well observed, and now in the priming section and the sidebar. The wet skin is exactly the useful feedback signal — unlike most of the errors in this file, this one announces itself, and the announcement is being misread.

Because rotation protects tissue rather than controlling absorption. Repeated injection into one small area produces lipohypertrophy, and absorption from lipohypertrophic tissue is blunted and erratic for any injected depot. Rotation prevents the condition that would make site matter. The advice is consistent; we should have made the causal order clearer.

Collection services are not interested in what was in the syringe. A pharmacy or local authority sharps point will take a rigid sealed container without inquiry, and the barrier you describe is built entirely of anticipated judgement. We would rather say that plainly than add to the lecturing.

A real gap and we will address it properly rather than in a reply. The short version is that it is a pharmacokinetic question — how much excess exposure, over what half-life — and a clinical one about symptom burden, and neither is answerable in the abstract. It also belongs in the titration file, which currently discusses omission and not excess.

The restraint is about where the fault lies. Customs authorities are performing a lawful function and owe nobody a thermal record. The claim of end-to-end control is the thing we criticise, and we do criticise it, in the article and again in the closing. What we will not do is convert an unavoidable feature of international freight into an allegation against the shipper who could not see it either.

A fair criticism of the reporting. Parcel 7 was submitted for purity determination on arrival and returned a figure within a percentage point of the supplier’s stated value, which is consistent with the solid-state argument and proves very little on its own, since we had no pre-shipment measurement on that vial. The design fault is ours: a shipment study without a paired baseline sample cannot answer the question we most wanted answered, and the next round will.

That is a fair reallocation of the criticism and we accept it. The word is chosen by whoever commissions the report, and laboratories are answering the question they were paid to answer. Our complaint is with the practice, not with the analysts, and the article should have located it more precisely.

Agreed, and it may be the best suggestion we have received on this subject. It is now a seventh line in the version of the list we send to suppliers, with the note that internal calibration should be stated where it was used.

Nor do we, and the file should have been clearer. The four-week interval is a floor below which the previous rung is still accumulating, not a threshold below which escalation is unsafe. Plenty of people tolerate faster ascent. Our objection is to the inverse inference — that because you did, everybody should — and to the absence of a trial that would let anyone say which is which in advance.

This is the hardest editorial question this department faces and we do not think you are wrong. Our position is that a practice this widespread is better described accurately, with its evidentiary status stated, than left to circulate in fragments. We accept that the distinction does work that a reader may not do.

The second. The published interquartile ranges around those means are wide, and outcomes like yours are well within them. The trials were not designed to explain why some people hold weight after cessation and others do not, and nothing measured at randomisation predicts it usefully. You are not an anomaly; you are part of a distribution nobody quotes.

Well put, and better than our own phrasing. We have adopted the point in the text with attribution to a reader.

Correct, and the distinction matters. We have added a clause: no arm overshot at a group level, which is not the same as no participant overshooting. The trials do not report individual overshoot rates and we have not found them published anywhere.

On the basis of the synthesis and analytical literature, where epimer formation during solid-phase assembly and during storage at extremes of pH is well characterised. What is missing is not evidence that it occurs but evidence about how much of it is present in any particular commercial vial, which is a different absence and the one we should have named.

A fair criticism of the reporting. Parcel 7 was submitted for purity determination on arrival and returned a figure within a percentage point of the supplier’s stated value, which is consistent with the solid-state argument and proves very little on its own, since we had no pre-shipment measurement on that vial. The design fault is ours: a shipment study without a paired baseline sample cannot answer the question we most wanted answered, and the next round will.

Retest, with a stated interval and a note that no formal stability study supports it. That is an honest description of a chemical supplier’s position and it is standard practice in the wider chemical trade. Printing expiry implies a study exists, which is the specific inference we object to.

The folk model that suffering indexes efficacy is widespread and the published analyses do not support it. It is also actively harmful when it drives escalation, which is why we gave it a line in the closing section rather than burying it in the tables.

The mechanism transfers; the incidence figures transfer only to the extent that your actual exposure matches the trial exposure, which is unknown unless the content has been measured. That is not evasion. It is the reason we argue for peptide content as a standard reported field rather than purity alone.

It is, and it is a compromise with how readers search. The tag exists because that is the word people use; the glossary exists because it is the wrong one. We would rather be findable and then precise than precise and unread.