Mapping a rotation that you will actually follow

One anatomical instruction is worth stating on its own, because it is universal, it appears in every set of published recommendations, and it is ignored more often than any other. The area immediately around the umbilicus — conventionally a radius of about two inches — is avoided because the tissue there is tethered to underlying structures and the subcutaneous layer is inconsistent in thickness. Beyond that boundary, the anterior abdominal wall is the largest, flattest and most forgiving surface available to most people, and it is where the great majority of injections in this drug class are given.

Accidental intramuscular delivery

Intramuscular delivery of a subcutaneous preparation accelerates and destabilises absorption. The insulin literature established this cleanly: intramuscular administration produces faster onset and markedly greater between-occasion variability than subcutaneous administration of the same preparation.¹

For a weekly acylated agonist the consequences of one such injection are less acute than for a mealtime insulin, because the depot is designed to release over days and albumin binding dominates the kinetics. It is nonetheless an unintended change in the input function, and where it happens repeatedly — a long needle used consistently in a lean thigh — it becomes a persistent alteration in exposure that no dose adjustment will explain.

The signals are not reliable. A deeper ache during and after injection, more bleeding, and a sensation of the injection being harder to push are all suggestive and none are diagnostic. This is why the answer is structural rather than perceptual: a 4 mm needle removes the possibility, and no amount of attentiveness makes a 12.7 mm needle in a lean thigh safe from it.

Where to inject

The four conventional subcutaneous sites are the anterior abdominal wall, the anterior and lateral thigh, the posterolateral upper arm, and the upper outer buttock. For most people the abdomen is the largest, most accessible and most forgiving, and it is where the majority of injections in this class are given.

Two exclusions apply. The area within roughly two inches of the umbilicus is avoided, because the tissue is tethered and the subcutaneous layer inconsistent. Any area of scarring, striae, active inflammation, bruising or existing lipohypertrophy is avoided, because absorption from altered tissue is altered.

Practical constraints matter as much as anatomy. The posterolateral arm is difficult to reach and to see on oneself, which makes it a poor routine site for self-injection even though it is perfectly good tissue. The buttock is similarly awkward without assistance. That leaves the abdomen and thighs as the realistic rotation surface for most people, which is ample: the abdomen alone offers a very large number of non-overlapping sites at a spacing of a couple of centimetres.

Rotation and lipohypertrophy

Lipohypertrophy is thickened, rubbery subcutaneous tissue produced by repeated injection into the same small area. Surveys of insulin-injecting populations have found it in a substantial proportion of patients — figures around a third or higher are commonly reported — and identified inadequate rotation and needle reuse as the principal risk factors.²

Its consequence is not cosmetic. Injection into lipohypertrophic tissue produces blunted and considerably more variable absorption; controlled work in insulin users found impaired and less predictable action following injection into such tissue compared with normal tissue.³ For any injected depot preparation the implication is the same: an unpredictable fraction of an intended dose.

The self-reinforcing feature is that lipohypertrophic tissue is less sensitive, so injecting into it is more comfortable, so people prefer it. Any rotation scheme therefore has to be followed against a mild incentive not to.

The schemes that survive contact with real life are simple: move at least a needle-length — practically, a couple of centimetres — from the previous site every time, and shift region on a fixed calendar cue rather than by memory. Palpating the sites occasionally for thickening is worth more than any grid diagram, because it detects the problem the scheme exists to prevent.

Why site matters less in this class than in insulin

Insulin absorption differs by site, which is why insulin regimens specify them. For the long-acting acylated incretin agonists, the labelling treats abdomen, thigh and upper arm as interchangeable, and the clinical pharmacokinetic literature reflects site comparisons conducted during development.⁴

The mechanistic reason is straightforward. These molecules are engineered to bind albumin reversibly and to release slowly from a subcutaneous depot; that release, and not regional blood flow, is the rate-limiting step. Where the depot sits therefore matters much less than it does for a preparation whose absorption is perfusion-limited.

This is a genuine practical simplification and it should be said clearly, because rotation advice imported wholesale from insulin practice can leave people believing that changing region will change their exposure. Rotation in this class is about protecting tissue from repeated trauma. It is not a dose-control measure, and a person who injects the thigh one week and the abdomen the next has not altered their treatment.

The exception is tissue that has already changed. Once lipohypertrophy is established, absorption from that area is unpredictable regardless of molecule, and the interchangeability above no longer applies.

How the Journal reports technique

Two bodies of evidence underlie this file. Questions of tissue, depth, needle length and rotation come from the insulin injection-technique literature, which is large, well conducted and directly transferable because it concerns anatomy rather than any particular molecule. Questions of absorption by site, in-use stability and exposure come from the incretin literature, which is smaller and where we say so. Where we describe practice rather than evidence, the text states it.

We give arithmetic in full rather than in tables of pre-computed unit counts, deliberately. A pre-computed table is correct only for the concentration it was computed for, and the recurring error in this market is precisely the reuse of a correct number under changed conditions. A reader who can perform the four-line calculation is protected against a class of error that no table can prevent.

Nothing in this file is medical advice. The Journal does not recommend doses, products, diluents or suppliers, and cannot assess an individual. Several compounds discussed are sold for research use only, are not approved for human use in any jurisdiction, and are not manufactured or released to any human sterility, content or endotoxin standard. Injection technique is properly taught in person by a clinician or nurse, and this file is not a substitute for that.

This file sits between two others in the department. Titration decides what dose is intended; tolerability decides whether it can be sustained; technique decides whether the intended dose is the one delivered. All three have to be right, and the third is the only one that can be got right in full by a careful person with a calculator.