Treating Hypoglycaemia: The 15g Rule, Paediatric Weight-Based Dosing, and Glucagon

Adult 15-gram rule, where it came from

The familiar shorthand most adults learn at diagnosis is the 15-gram rule. Take 15 grams of pure glucose, wait 15 minutes, retest, repeat if still below 4.0 mmol/L (72 mg/dL). The protocol that international and UK adult guidance anchors on (ISPAD 2024 Chapter 12 for the international anchor, NICE NG17 for the UK adult anchor) traces back to a 1984 outpatient dose-ranging study by Brodows and colleagues at Rochester (JAMA), which tested doses of 10, 15, 20 and 25 grams in adults with T1D. The 20-gram dose produced the largest glucose rise, the 10-gram dose was sometimes not enough, and the 15-gram dose with a 20-minute retest was the practical compromise. The clinical shorthand “15 and 15” rounded the retest down to 15 minutes for memorability, which is the version that has passed into structured education over the last forty years (DAFNE in the UK, BERTIE for adults, DYNAMIC for sport, GAME for adolescents).

The protocol works in adults because their body weight sits in a range where 15 grams of glucose, absorbed at the gut’s ceiling rate of around one gram per minute, is enough to lift glucose into the safe range within fifteen to twenty minutes most of the time. It is a population-average protocol, anchored on the population-average adult body. Where the protocol genuinely struggles is at the two ends: the small body that does not need 15 grams, and the larger body that arguably does. The paediatric trial evidence, which is more recent, addresses the small-body end of that.

Paediatric weight-based, the McTavish trial

The pivotal paediatric trial is McTavish and Wiltshire 2011 (Pediatric Diabetes), a randomised controlled trial in 39 children with T1D presenting with acute hypoglycaemia. Three doses were compared: 0.1, 0.2 and 0.3 grams of pure glucose per kilogram of body weight. The 0.3 g/kg dose recovered glucose to safe range within 15 minutes in over 80 percent of attempts; the lower doses underperformed. ISPAD Chapter 12 caught up with this in 2024, and now anchors the international paediatric protocol on 0.3 g/kg of pure glucose. For a 20 kg child that is 6 grams. For a 40 kg child it is 12 grams. The protocol scales with the body, not with a slogan.

The cap matters. A 70 kg adolescent on a 0.3 g/kg calculation would need 21 grams; the GNL platform caps the carb dose at 60 kg of body weight, so the maximum first dose surfaced anywhere on the GNL Hypo Treatment Explorer is 18 grams (60 kg multiplied by 0.3 g/kg). The reason for the cap is not that the body’s need stops at 60 kg; it is that the gut’s absorption ceiling sits around one gram of glucose per minute via the SGLT1 single-source pathway, and a 21-gram first dose would simply stack unabsorbed glucose in the stomach for the first three minutes, producing the rebound spike that makes everyone uneasy about hypo treatment to begin with. The cleaner answer for the larger body is a sensible first dose, a 15-minute retest, and a second dose if still low. This rule is shared with every other GNL carb-returning calculator on the platform, so a family who has used the Exercise Planner sees the same cap on the Hypo Explorer.

Why glucose first, and what about sucrose

Pure glucose is the cleanest treatment because the body does not have to convert it. It crosses the gut into the blood directly, raising glucose within ten minutes and finishing its job by twenty. Slama and colleagues (1990, Archives of Internal Medicine) ran the head-to-head: glucose tablets, glucose solution, glucose gel, fruit juice, sucrose, and starch hydrolysate, all tested side by side in adults with insulin reactions. Glucose tablets produced the fastest rise; starch was the slowest; the others sat in the middle. Husband and colleagues (2010, Pediatric Diabetes) added the paediatric piece, comparing 10 to 15 grams of glucose tablets, sucrose (Skittles) and fructose-led product (Fruit to Go) in 33 children with T1D. Glucose and sucrose recovered glucose at 15 minutes in 88 percent of attempts; the fructose-led product failed in 33 percent.

The lived consequence of this is twofold. First, glucose tablets are the most reliable rescue because they avoid the fructose half of sucrose entirely; they are portable, stable, measurable, and free of the variable that Husband showed failed one in three times in children at the 15-minute endpoint. Second, sucrose at a controlled 10 to 15 gram dose is not a slow-acting treatment in children; the popular wisdom that “sugar takes an hour” tends to describe a dose-and-stacking effect (larger servings, repeated treatments before the first has been absorbed, uncounted food alongside) rather than a property of 15 grams of pure sucrose. Fruit juice and fruit-led bars are genuinely slower because they are mostly fructose; they should not be the first-line rescue.

Glucagon, when oral does not work

Severe hypoglycaemia is defined as a hypo that needs help from someone else, regardless of the glucose number. The classic case is loss of consciousness or seizure. The less classic but commoner case is the conscious adult who cannot reliably swallow safely: the small child who is vomiting; the post-exercise teenager who has just collapsed and is not yet fully responsive. Oral glucose is unsafe in any of these settings; the airway is not protected and the family member at the kitchen table cannot be sure swallowing is intact. The glucagon route is the alternative.

The glucagon options have changed in the last seven years. The traditional reconstitution kit (mix the liquid with the powder, draw up, intramuscular injection) worked when used correctly, but the “used correctly” part was the failure mode. Rickels and colleagues (2016, Diabetes Care) ran the pivotal non-inferiority trial of intranasal glucagon (Baqsimi) in 75 adults with T1D, with hypoglycaemia induced by IV insulin to a nadir around 2.5 mmol/L (45 mg/dL). The intranasal 3 mg dose succeeded in 98.7 percent of cases versus 100 percent for the 1 mg intramuscular dose, with a clinically inconsequential 3-minute lag. The companion human-factors evidence (Yale 2017, cited in Rickels) showed that untrained bystanders succeeded with the nasal device in over 90 percent of simulated rescues, compared with around 20 percent for the reconstitution kit. The ready-to-use autoinjector format (Gvoke HypoPen) reaches similar success rates in untrained hands (Valentine 2019, Diabetes Technology and Therapeutics).

The lived consequence of the device shift is that the rescue plan in 2026 looks nothing like the one most adults with T1D were taught at diagnosis fifteen or twenty years ago. The household no longer needs to remember how to mix a vial under the pressure of a collapse; they need to know which pocket the nasal device or the autoinjector lives in, how to operate it once, and that the next call is 999. The diabetes team can supply a written household plan and the school nurse can be shown the same device. The diabetes care team is the route to the prescription and to the household training, in the UK and in most international settings.

Post-hypo carbohydrate, and the AID-era softer touch

The classic post-treatment advice to follow up the rescue dose with a long-acting carbohydrate (a slice of toast, a glass of milk, a snack) was useful in the era of long-acting insulins like NPH where the next insulin peak might still be coming. With modern basal insulins (degludec, glargine U300, the algorithm-modulated basal of an AID system) the picture is gentler. The international current position (ISPAD 2024 Chapter 12) is that follow-up carbohydrate is needed only when the next meal is more than an hour away, and not routinely after every hypo. Routine over-treatment is one of the most common drivers of post-hypo rebound highs and of the slow weight gain that some adults notice after years of treating frequent lows.

The AID-era softening is its own conversation. When a hybrid closed-loop system has already suspended insulin in anticipation of a low (the predictive low-glucose suspend mechanism Buckingham 2015 validated in paediatrics), the body is no longer being driven down by ongoing insulin. ISPAD 2024 Chapter 12 recommends a smaller treatment dose (5 to 10 grams) in this AID context, because a full 15-gram dose layered on top of a system suspension tends to overshoot. The Karges 2024 DPV registry (Lancet Diabetes Endocrinol) underwrote this overall pattern at scale: across 13,922 young people, hybrid closed-loop dropped hypoglycaemic coma by around a third (incidence rate ratio 0.68, 95 percent CI 0.48 to 0.97) without raising overall severe hypoglycaemia. The technology era has not eliminated treating lows; it has reduced both the frequency and the severity of the worst end of the curve.