Every letter we have printed

As a matter of fairness: the supplier that started reporting content after you contacted them deserves to be named. You name people when they behave badly.

Timing consistency is undervalued. Same laboratory, same time of day, same fasting state removes a large part of the noise between draws for no cost whatever, and it makes a modest true change visible that would otherwise be buried.

The discipline your article recommends is the right one: decide what you would do differently for each result before ordering it. Anything that fails that test is generating a number that will be worried about and not acted on.

As a biomedical scientist I would add one point to your reference-interval section: many laboratories do not derive their own intervals at all. They adopt the manufacturer interval for the platform, which was established in a population that may have nothing to do with the one being tested. The interval on the report can be a document about a different country.

Where the same event can be coded under two dictionary terms, incidence is split between them and each looks smaller. Coding conventions are a technical matter with a large effect on the numbers people quote.

Self-reported incidence collected in online communities is not comparable with trial incidence in any direction, and it is put side by side with it regularly. Different populations, different definitions, different collection, one chart.

The material this publication covers is not approved for human use in any jurisdiction, which changes what a safety discussion can even mean. A pharmacovigilance system attaches to an approved product and there is no analogous system here at all.

Your list of twelve method values omits the equilibration time between injections. On a shallow gradient with an aqueous-rich starting composition, an inadequate re-equilibration will shift early-eluting impurities into the solvent front and quietly remove them from the denominator. It is the cheapest way to improve a figure by accident.

We changed to core-shell particles two years ago for exactly the reasons you set out, and I would add one that you do not mention. The pressure headroom means a column survives a badly filtered sample instead of blocking, which in a service laboratory is worth more than the extra plate count.

Two-dimensional liquid chromatography now exists in a form a service laboratory can afford, and it collapses the whole argument: the second separation happens on the fraction rather than on the sample. It is still a specialist purchase, but the price is falling and the case for it in this market is stronger than in most.

Capillary electrophoresis is the forgotten answer here. It separates on charge-to-size in free solution, it needs a few nanolitres, and for anybody who already has the instrument it is an afternoon. The barrier is not cost or difficulty; it is that hardly anyone in this trade has been trained on it.

The trials are conducted with pharmaceutical-grade material under supervision. Whatever they establish, they establish about that material in that setting, and the extension of their findings to a research-supply context is an assumption rather than a result.

The curves everybody reproduces are drawn from the last available observation for participants who left early, which is a specific analytical choice with a known direction of bias. Alternative analyses are usually in the appendix and rarely differ dramatically, but readers should know one was made.

Reasons given in public forums are shaped by the audience. A community that values persistence will hear fewer accounts of stopping, and a community that is sceptical will hear more. Neither is a sample of anything.

Reasons collected by a supplier, a clinician and a researcher would produce three different distributions from the same population, because people say different things to different audiences. Every dataset in this area is an artefact of who was asking.

Supplements taken before a draw interfere with several assays directly, which is a chemical interference rather than a physiological effect. It is well documented and almost never asked about.

Tube type and fill volume matter. An underfilled tube changes the ratio of sample to additive, and the result is affected in a direction that depends on the analyte. The laboratory notices; the person receiving the report usually does not.

Turnaround is the practical obstacle. A compendial sterility test takes a fortnight, which for material with a short commercial life is a genuine cost rather than an excuse. Rapid methods exist and validating one is not trivial.

A note on scale. A supplier filling thousands of units has a fixed-cost argument for better process control that a small operation does not, and the market’s structure therefore affects what is even feasible. That is worth a piece in its own right.

Naming an absence without implying a hazard is a difficult editorial line and your department walks it more carefully than most. Saying that something is unmeasured is a statement about measurement, not about risk.

Composition of regained tissue is barely studied and is the question I would most like answered. Whether what returns resembles what was lost has practical consequences and there is very little data on it.

The regain trajectory is presented as a mean curve and the variation around it is very wide. Some participants hold most of the change and some return close to baseline, and the average describes neither of them well.

Cost is a maintenance variable and it is treated as though it were outside the clinical question. For most people the sustainable dose is the affordable one, and a plan that ignores that will be abandoned rather than followed.

Where people give more than one reason, most analyses take the first mentioned, which is an ordering effect rather than a weighting. Asking respondents to rank would cost nothing and would change several published distributions.

Nobody has published what happens to intake over the course of these trials, though several must have collected it. A plateau explained by intake returning toward baseline is a different phenomenon from one explained by anything pharmacological.

The word plateau implies a stable state and what the data usually shows is a decelerating one. Those are different shapes with different implications for what happens next, and the distinction is lost in every summary I read.

The claim that nobody has randomised escalation intervals is too strong. There are protocol amendments in several programmes that effectively created slower-titration cohorts, and some of those have been analysed post hoc.

The tolerability curve and the efficacy curve have different shapes, and the interesting region is where they diverge. Almost all discussion of ceilings addresses one curve at a time.

The composition question sits underneath the plateau question and is almost never asked with it. A flat mass curve can conceal a continuing change in composition in either direction, and the scale on the floor cannot distinguish them.

Disposal is the part of this subject with actual legal content in most jurisdictions, and it is the part least discussed. The rules are public, they differ by locality, and they are not difficult to find.

Gauge numbering runs backwards and it catches people out constantly. A higher number is a thinner needle. A reader who assumes the number tracks the diameter will order the opposite of what they wanted, and the packaging does nothing to help.

A quibble on your dead-space figure. You give two to seven microlitres for insulin-type needles, which is right for fixed-needle syringes, but the detachable pen needle plus syringe hub combinations sold in this market are considerably worse and you should say so.

Amylin analogues sit adjacent to this discussion and are usually left out of it. The mechanism is different, the combination data are early, and the reason they belong in a piece about multi-receptor design is that they show the field is not converging on one receptor family.

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.

A small thing: you write "class B GPCR" and then "secretin-like receptor" as though these were different classifications. They are the same family under two naming conventions, and the piece would be clearer if it said so.

Subcutaneous absorption is not instantaneous and the rate depends on the depot. That is a pharmacokinetic reason for site rotation which has nothing to do with tissue irritation, and it is a more interesting reason than the one usually given.

Your account of electrospray is clear but it understates how much the additive matters. A mobile phase carrying trifluoroacetic acid suppresses ionisation substantially, which is why a laboratory running an excellent chromatographic method sometimes has to change it entirely before the same sample will give a usable spectrum. The two techniques want different solvents.

Ion suppression is the point I would put in bold. A species that ionises poorly in the presence of an excess of something that ionises well can be present at several per cent and invisible in the spectrum. Absence of a peak is not absence of a compound, and the report cannot tell you which you have.

A substudy is powered for the parent trial’s primary endpoint and not for its own, which is why the intervals around composition figures are so wide. Quoting the point estimate without the interval misrepresents the design as much as the result.

The substudies are the most expensive part of these programmes per participant, which is why they are small, and their size is then treated as a weakness of the finding rather than a consequence of the method.