Glucagon‐like peptide‐1 receptor agonists and the appropriate measurement of gastric emptying

The Roman Emperor, Marcus Aurelius (AD121-180) stated: ‘We need to learn to measure what we value, not value what we can easily measure’. Information relating to the effects of glucagon-like peptide-1 (GLP-1) receptor agonists (RAs) on gastrointestinal symptoms and gastric emptying in type 2 diabetes (T2DM) continues to be compromised by the widespread use of suboptimal methodology; in the former, self-report of symptoms, which is known to be unreliable, rather than validated questionnaires, as employed in functional gastrointestinal disorders, and in the latter, the plasma absorption kinetics of oral paracetamol (acetaminophen) as a surrogate marker of gastric emptying, where the recently published study by Urva et al. is an example. Regulatory bodies should assume some responsibility for this situation, which has contributed to confusion, rather than clarity. Measurement of gastric emptying requires a precise technique. Gastric emptying exhibits a substantial interindividual variation in health and there are major differences in rates and patterns of emptying of solids, nutrient liquids and non-nutrient liquids. The interindividual variation in gastric emptying is greater in T2DM; 30% of those with longstanding, suboptimally controlled, T2DM seen in tertiary centres have gastroparesis, while in well-controlled T2DM, emptying is usually normal or modestly accelerated. Acute hyperglycaemia slows emptying, while insulin-induced hypoglycaemia accelerates it. Delayed emptying of solids is usually regarded as the dominant abnormality in gastroparesis. In general, the relationship of symptoms with the rate of emptying is relatively weak. Scintigraphy is the ‘gold standard’ technique to quantify gastric emptying and can be used to measure emptying of solids, nutrientand non-nutrient liquids, as well as intragastric meal distribution. Additional information is obtained with a dual isotope technique when emptying of solid and liquid meal components is measured concurrently. Limitations in the use of scintigraphy relate to its cost, requirement for expensive equipment and radiation exposure. Stable isotope breath tests, with the rate of gastric emptying calculated using an appropriate mathematical model, are simpler, less expensive and not associated with a radiation burden, providing an acceptable alternative, particularly for large-scale clinical trials and studies in animals, including mice. The paracetamol absorption test, while simple and inexpensive, has major limitations the implicit assumptions that paracetamol is always absorbed rapidly in the duodenum and that emptying of the paracetamol from the stomach is concordant with that of a meal, are incorrect. Paracetamol is usually included in a low-nutrient liquid, but this is inconsistent, and the paracetamol technique cannot be used to assess solid gastric emptying. While the plasma paracetamol area under the curve (AUC) 0-1 h may provide an assessment of liquid emptying, the absence of a difference in AUC after that time (e.g. 1-4 h) does not establish that an effect on emptying is only sustained for 1 h, simply because the paracetamol will eventually be absorbed over 4 h, even though it may be absorbed more slowly during the first 0-1 h. Gastric emptying is a major determinant of the postprandial glycaemic response and a therapeutic target in T2DM. The relationship of the postprandial glycaemic excursion with the entry of carbohydrate into the small intestine appears non-linear. Accordingly, minor changes in gastric emptying may have a major effect on the glycaemic response. Studies using scintigraphy and breath tests have established that ‘short-acting’ GLP-1RAs (i.e. exenatide twice daily, lixisenatide) lower postprandial glucose predominantly by slowing gastric emptying, an effect that is greater when baseline emptying is relatively faster and dependent on the magnitude of slowing. The concept that ‘long-acting’ GLP-1RAs have little or no effect on gastric emptying with sustained administration was supported by studies showing, in health, that the slowing of emptying by intravenous GLP-1 is diminished over 8.5 h and 24 h, that is, there is ‘tachyphylaxis’. However, in both studies, after continuous administration, gastric emptying was still slowed substantially by GLP-1. Recent studies, using scintigraphy, have established that the ‘long-acting’ GLP-1RAs, exenatide once-weekly and liraglutide, do slow emptying with chronic administration, albeit probably less than ‘shortacting’ GLP-1RAs. For example, in health, exenatide once-weekly slows emptying of solids and liquids after 8 weeks’ administration; moreover, the reduction in postprandial glucose is related to the magnitude of this slowing. An effect of exenatide once-weekly on gastric emptying was ‘missed’ using the paracetamol test. The acceleration of gastric emptying by acute hypoglycaemia may represent an important counter-regulatory mechanism. Frandsen et al. have suggested that liraglutide does not delay gastric emptying during hypoglycaemia in type 1 diabetes, but their conclusion was based on the paracetamol absorption technique. A study reported in abstract form, in which gastric emptying was measured by scintigraphy, suggests that dulaglutide has a sustained effect to slow gastric emptying in T2DM. The effects of subcutaneous and oral semaglutide on gastric emptying, regretfully, have only been assessed using the paracetamol test. In obese subjects given semaglutide (1 mg/week subcutaneous for 12 weeks), plasma paracetamol levels were 27% less Received: 23 June 2020 Revised: 20 July 2020 Accepted: 31 July 2020