Statistics from Altmetric.com
I read with great interest the paper by Talley and colleagues (Gut 2001;49:395–401) and the accompanying editorial by Tack and Peeters (Gut 2001;49:317–8). There are many important issues that are raised in the paper and editorial. I believe the paper provides an opportunity to identify areas where study design might be enhanced in future studies.
Firstly, the fact that gastric emptying was not measured at the end of the study leaves wide open the question of whether the prokinetic approach should be abandoned in the treatment of dyspeptic symptoms in diabetics. Thus it would be inappropriate to conclude from this study that prokinetics are not indicated. This point is also emphasised in the editorial by Tack and Peeters.
Secondly, the authors conclude that baseline gastric emptying does not influence the response to ABT-229. This conclusion is based on weak foundations as the method used to measure gastric emptying appears to provide data that are scarcely believable. Thus the t50% recorded in healthy subjects (130±50 (SD?) minutes) is remarkably outside the normal range reported using the gold standard scintigraphy (mean 110±4 (SEM) minutes, 10th percentile 70 minutes, 90th percentile 150 minutes in our laboratory). The methods section does not unequivocally state what mathematical analysis was used with the stable isotope breath test at the central laboratory used in the study. Improved mathematical analyses of gastric emptying using breath tests in the more recent literature provide a higher level of accuracy relative to scintigraphy.1–4 It is claimed that the method was validated in 19 diabetics in whom a significant correlation (r=0.73) was observed between scintigraphy and breath test data. Correlation does not equate to accuracy and, in the absence of a Bland-Altman or similar analysis, the gastric emptying data are suspect and cannot be used to classify patients to assess the relationship between symptoms and emptying, or to address the role of baseline gastric emptying as a covariate in the response to treatment. It is also unclear if the study was sufficiently powered to appraise an effect of delayed gastric emptying on response to therapy, given the fact that only 29% of the study cohort were classified as having delayed gastric emptying. A type II error cannot be excluded.
Thirdly, the theoretical point is made by Tack and Peeters regarding tachyphylaxis of this particular motilin agonist, previously demonstrated in the study of Verhagen and colleagues.5 However, other prokinetics, indeed other motilin agonists, may prove effective in the treatment of dyspepsia in diabetics with impaired gastric emptying.6,7
Fourthly, the observation that over time some of the symptoms continued to be aggravated in the active arm of the study suggests that the drug was still effective and worsened symptoms, rather than simply being ineffective in the patients evaluated.
Fifthly, the study illustrates the importance of thoroughly characterising the pharmacology of a novel agent before embarking on expensive potentially harmful therapeutic trials. Inhibition of accommodation by motilin agonists may indeed be responsible for aggravation of bloating and other symptoms over time. Fortunately, these effects are likely to be reversible and no permanent harm was reported.
However, it is still worth emphasising the general point—clinical pharmacology and pharmacodynamic studies have an important role to play in the drug development process. This is especially relevant in the context of “gastroparesis” or dyspepsia as there are non-invasive approaches to study gastric emptying, accommodation, and postprandial symptoms. These methods permit proper dose-response studies prior to exposing patients to potentially harmful agents or inappropriately selecting subgroups of patients for such large and expensive studies. Among patients with diabetes, neuropathy may alter both gastric emptying and gastric accommodation via different mechanisms (for example, extrinsic vagal v intrinsic nitrergic neuropathy). Thus selection of those with only impaired emptying (based on a reliable test) and normal accommodation might have provided a fairer opportunity to assess the efficacy of the drug.
Finally, as acknowledged by Talley et al, assessment of autonomic neuropathy requires a more formal assessment than the “opinion of the attending endocrinologist”. In fact, disturbances of the autonomic nervous system, evaluated with detailed tests, have been shown to significantly influence the symptom response to a prokinetic.8 Approaches that carefully characterise the drug before exposure of patients and selecting subgroups of patients after thorough understanding of the effects of the drug may save potentially effective medications from being abandoned. These patients need effective therapies. As one of many physicians who struggle to help relieve these patients' symptoms, we cannot afford to ...“throw out the baby with the bath water”. I trust that this appeal may encourage pharmaceutical companies to reconsider whether the medication or a derivative with improved pharmacokinetics should be given a “second chance”.
A number of the issues raised by Dr Camilleri are important and relevant although some of the points require clarification. We stand by our position that drugs which act solely as gastric prokinetics are unlikely to be beneficial in either diabetic gastropathy or functional dyspepsia. Our data (both in this study and eleswhere1) suggested that the motilin agonist tested actually worsened symptoms in both diabetics and non-diabetics with unexplained dyspepsia, regardless of baseline gastric emptying status. Other recent data suggest that motilin agonists impair fundic accommodation and this physiological disturbance may induce symptoms in a subset with dyspepsia.2 Cisapride relaxes the fundus and this may explain its therapeutic benefits in dyspepsia.3 Our observations have important implications for future drug development; we agree with Dr Camilleri that ideally the mechanisms of drug action need to be understood prior to planning clinical trials, although this is often completely impractical and could impair progress at times. It is also fair to point out that data on fundic accommodation have only become available relatively recently, and preceded the planning of the trials.
While we agree that there are limitations with C13-octanoic acid breath testing, we believe that the data are reasonably robust. Indeed, we applied a number of cut offs for delayed gastric emptying versus normal but were unable to identify any influence of baseline gastric emptying on the response of the motilin agonist tested.1
Dr Camilleri has emphasised the fact that gastric emptying was not measured at the end of the study. There has been a reluctance on the part of the pharmaceutical industry to re-measure gastric emptying in clinical trials because of the recognised lack of correlation of changes in gastric emptying with symptom improvement.4 Furthermore, there remains an absence of reliable standardised reference methods for gastric emptying that can be applied in multicentre trials. However, we agree that it is optimal in prokinetic trials to test gastric emptying at baseline and on drug, and this should be the “gold standard”.
The issue of tachyphylaxis is important. We conclude, based on the available evidence, that tachyphylaxis was unlikely but agree the issue needs to be carefully considered in all studies evaluating prokinetics. Indeed, in our studies, as Dr Camilleri points out, the drug was actually deleterious (this study and Talley and colleagues1). This strongly suggests that tachyphylaxis did not occur and did not explain the negative results with ABT-229.
We stand by the study design used although further improvements are feasible. Phase I data were available indicating that there were unlikely to be any significant serious effects of ABT-229 and therefore we dismiss the concern raised about potential harm; this was borne out in the phase II trials (present study and Talley and colleagues1). However, we agree that this may not apply to other novel pharmacological agents in development for diabetic gastropathy and functional dyspepsia. We conclude that the motilin agonist class is likely to be disappointing in unexplained dyspepsia unless agents in this class with quite different physiological effects are developed.