Effect of intravenous infusion of glyceryl trinitrate on gastric and small intestinal motor function in healthy humans


Dr J. L. Madsen, Department of Clinical Physiology and Nuclear Medicine, 239, Hvidovre Hospital, Kettegård Allé 30, DK-2650 Hvidovre, Denmark.
E-mail: jan.lysgaard.madsen@hh.hosp.dk


Background  Glyceryl trinitrate is a donor of nitric oxide that relaxes smooth muscle cells of the gastrointestinal tract. Little is known about the effect of glyceryl trinitrate on gastric emptying and no data exist on the possible effect of glyceryl trinitrate on small intestinal transit.

Aim  To examine the effect of intravenous infusion of glyceryl trinitrate on gastric and small intestinal motor function after a meal in healthy humans.

Methods  Nine healthy volunteers participated in a placebo-controlled, double-blind, crossover study. Each volunteer was examined during intravenous infusion of glyceryl trinitrate 1 μg/kg × min or saline. A gamma camera technique was used to measure gastric emptying and small intestinal transit after a 1600-kJ mixed liquid and solid meal. Furthermore, duodenal motility was assessed by manometry.

Results  Glyceryl trinitrate did not change gastric mean emptying time, gastric half emptying time, gastric retention at 15 min or small intestinal mean transit time. Glyceryl trinitrate did not influence the frequency of duodenal contractions, the amplitude of duodenal contractions or the duodenal motility index.

Conclusions  Intravenous infusion of glyceryl trinitrate 1 μg/kg × min does not induce major changes in gastric or small intestinal motor function after a 1600-kJ meal in healthy volunteers.


After intravenous infusion, glyceryl trinitrate (GTN) is converted to nitric oxide (NO). NO has been established as an important non-adrenergic, non-cholinergic neurotransmitter in the gastrointestinal tract, inducing smooth muscle cell relaxation.1–3 NO activates the enzyme guanylate cyclase and stimulates the synthesis of cyclic guanosine monophosphate. Cyclic guanosine monophosphate stimulates calcium-dependent potassium-channels in smooth muscle cells, leading to their hyperpolarization and relaxation. Observations in animals suggest that NO may be an important modulator of gastroduodenal motility.4–6 There is less information about the role of NO mechanisms in the regulation of gastrointestinal motor function in humans, and the reports are not concordant. Previous studies with the NO donor isosorbide dinitrate did not show any change in the gastric emptying rate.7, 8 In other studies, GTN was observed to slow gastric emptying.9, 10 Most likely, the lack of consistency is due to variations in the type, the dose and the way of administration of the NO donors. It is known from animal studies that NO mechanisms are involved in the coordination of the peristaltic reflex,11–13 and previous studies indicate that NO modulates the human duodenal motility before and after a meal.10, 14 However, the influence of intravenous infusion of a NO donor on the propulsive activity of the entire human small intestine after a meal has not been assessed before. We found it relevant, therefore, to evaluate the effect of intravenous infusion of GTN on gastric and small intestinal motor activity after a mixed liquid and solid meal in healthy humans.

Materials and methods

Nine healthy male volunteers (23–31 years) participated in a placebo-controlled, double-blind, crossover study. Written informed consent was obtained beforehand. None of the participants had previous abdominal surgery other than appendectomy or received any medication. The study was performed in accordance with the Declaration of Helsinki and approved by the regional medical ethics committee.

Study design

The volunteers underwent paired studies on separate days, at least 7 days apart. After an overnight fast, at about 08:00 am a canula was positioned in an antecubital vein for intravenous infusion of GTN or saline. The manometry catheter was introduced as described below, and the duodenal pressure recordings began. In randomized and double-blind order, each subject was then given an intravenous infusion of either GTN in 0.9% saline at a rate of 1 μg/kg × min or the same volume of 0.9% saline. The infusion was continued throughout the study period. After 30 min, the volunteers ingested a radiolabelled meal. Subsequently, the scintigraphic assessments were started. Blood pressure and heart rate were measured at 15-min intervals. The subjects remained supine throughout the study, and they were asked to report any unusual sensations. During the study, the subjects drank 200 mL of water/h.


A fine (OD 2.8 mm) flexible polyurethane catheter that incorporated three piezoresistive pressure sensors located 3, 13 and 23 cm from the tip (Unisensor AG, Atticon, Switzerland) was used to record small intestinal motility. Under fluoroscopic control, the catheter was passed transnasally into the duodenum so that the tip of the catheter was located close to the ligament of Treitz. The catheter was secured by adhesive tape to the subject's cheek. Outputs from the pressure transducers were recorded with a computerized polygraph (UPS-2020, Medical Measurement Systems, Enschede, the Netherlands) and stored on a computer for later analysis. A software package (UPS-2020 gastro 7.1; Medical Measurement Systems) was used to process the pressure recordings. Thus, the mean frequency of duodenal contractions, the mean amplitude of duodenal contractions and the duodenal motility index were calculated over periods of 1 h at the distal recording site. The motility index was found from following equation: ln [(Sum of contraction amplitudes × Number of contractions) + 1].15 Furthermore, simple visual analyses of the motility patterns were also performed.


Scintigraphy was performed as previously reported with few modifications.16, 17 In 10 min, the subjects ingested a 1600-kJ mixed liquid and solid meal (200 mL water, 80 g bread, 120 g omelette) containing radiolabelled markers. About 4 MBq of 111In-diethylenetriamine penta-acetic acid (DTPA) was added to the water as a liquid-phase marker, and 20 MBq of [99mTc]-stannous colloid was added to the omelette as a solid-phase marker. If the examinations were not already finalized, further non-radiolabelled standard meals were consumed after 3 and 6 h. Scintigraphy was performed with a dual-head gamma camera equipped with medium energy, general purpose collimators (Millennium VG, General Electric Medical Systems, Milwaukee, WI, USA). To measure gastric emptying and small intestinal transit rates, static images of 2 min were obtained in anterior and posterior projections at 15-min intervals until no radioactivity could be detected in the small intestine. Gastric antral motility was assessed from dynamic images of 1 s each that were acquired for 5 min in the anterior projection. The dynamic imaging was repeated at 15-min intervals for 1 h.

For gastric emptying and small intestinal transit assessment, regions of interest for integration of radioactivity were drawn manually on each static image. The counts were corrected for physical decay, downscatter and attenuation. As the gastric input of the radiolabelled meal was almost instant, gastric time–activity curves were obtained directly. Construction of adequate small intestinal time–activity curves involved a deconvolution procedure that made it possible to compute the expected time–activity curves if the input of radiolabelled meal components from the stomach was instantaneous.16 For each experiment in each subject, the time–activity curves were condensed into gastric mean emptying times or small intestinal mean transit times. These variables, which express the mean time it takes for all the radiolabelled marker to leave the stomach or the small intestine, take into account the whole time–activity curve rather than focusing separately on only a part of the curve or specific time points.17 However, gastric half emptying time as found by interpolation of the raw data and gastric retention at 15 min were also applied.

The method used to process the dynamic images has been described elsewhere.18 A standardized region of interest was localized over the distal antrum close to the pylorus. Then, the position of the geometric centre of radioactivity was determined on each 1-s image. On the assumption that the horizontal oscillations of the geometric centre of radioactivity in the region of interest reflected the contractile activity of the antrum, fast Fourier-transform analysis was used to find the dominant frequency of these contractions.


Values are presented as mean and standard deviation. Comparisons were performed by paired, two-tailed Student's t-test. Differences were considered significant when P < 0.05.


Results of the gastric emptying measurements are shown in the Table 1 and Figure 1. It appears that GTN did not influence gastric mean emptying time, gastric half emptying time or gastric retention at 15 min of liquid or solid marker. Results of the small intestinal transit determinations are shown in the Table 1 and Figure 2. GTN had no effect on the small intestinal mean transit time of liquid or solid marker. The postprandial frequency of antral contractions, which appears from the Table 1, was not changed by GTN at any time point. Because of the fast small intestinal transit in some experiments, sufficient data for a proper statistical analysis of duodenal motility recordings were only available for a 4-h period. The results are shown in the Table 1, and it appears that intravenous infusion of GTN did not change the mean frequency of duodenal contractions, the mean amplitude of duodenal contractions or the duodenal motility index for any 1-h period. The simple visual analyses of the pressure recordings did not reveal specific or changed motility patterns during infusion of GTN. Infusion of GTN caused a small, but significant decrease in systolic blood pressure (122 ± 5 vs. 126 ± 8 mmHg, P = 0.047), whereas heart rate was not changed (62 ± 9 vs. 62 ± 6 beats/min, P =1.000). Six subjects reported mild or moderate headache and one subject had a sensation of heat during the infusion of GTN. One subject reported dizziness during the infusion of saline. None of the volunteers had other inconveniences during the experiments.

Table 1.  Effects of glyceryl trinitrate on gastrointestinal motility variables in nine healthy humans
 PlaceboGlyceryl trinitrateP-value
  1. Values are given as mean ± s.d.

Gastric mean emptying time (h)
 Liquids0.66 ± 0.200.78 ± 0.290.263
 Solids1.2 ± 0.221.5 ± 0.650.321
Gastric half emptying time (h)
 Liquids0.36 ± 0.110.61 ± 0.410.119
 Solids1.1 ± 0.141.4 ± 0.780.298
Gastric retention at 15 min (%)
 Liquids57 ± 1069 ± 170.094
 Solids88 ± 889 ± 100.770
Small intestinal mean transit time (h)
 Liquids4.1 ± 1.33.8 ± 1.40.374
 Solids3.7 ± 1.33.5 ± 1.40.347
Frequency of antral contractions (per min)
 0.00 h after test meal3.0 ± 0.273.1 ± 0.300.548
 0.25 h after test meal3.0 ± 0.263.0 ± 0.411.000
 0.50 h after test meal3.0 ± 0.272.9 ± 0.330.594
 0.75 h after test meal3.1 ± 0.253.1 ± 0.200.799
 1.00 h after test meal2.9 ± 0.173.1 ± 0.250.141
Mean frequency of duodenal contractions (per min)
 0–1 h after test meal3.3 ± 1.03.1 ± 1.10.636
 1–2 h after test meal3.5 ± 1.33.0 ± 1.00.057
 2–3 h after test meal2.8 ± 1.22.6 ± 0.900.201
 3–4 h after test meal2.9 ± 0.972.7 ± 0.530.642
Mean amplitude of duodenal contractions (mmHg)
 0–1 h after test meal31.7 ± 13.732.3 ± 14.50.939
 1–2 h after test meal28.7 ± 10.131.7 ± 13.70.631
 2–3 h after test meal33.8 ± 11.531.1 ± 6.60.560
 3–4 h after test meal36.1 ± 10.132.3 ± 5.70.421
Duodenal motility index
 0–1 h after test meal11.9 ± 1.011.7 ± 0.90.707
 1–2 h after test meal11.9 ± 0.911.7 ± 0.90.355
 2–3 h after test meal11.6 ± 1.011.4 ± 0.80.292
 3–4 h after test meal11.7 ± 0.911.6 ± 0.40.648
Figure 1.

Gastric emptying of radiolabelled liquid and solid marker after intravenous infusion of glyceryl trinitrate 1 μg/kg × min or placebo.

Figure 2.

Small intestinal transit of radiolabelled liquid and solid marker after intravenous infusion of glyceryl trinitrate 1 μg/kg × min or placebo.


It is generally accepted that the process of gastric emptying of complex meals involves a delicate interplay between the muscle tone of the proximal stomach and the coordinated antropyloric motor patterns providing resistant and grinding forces as well as pulsatile aboral flow of the meal components.19, 20 Drugs that change the tone of gastric and/or pyloric smooth muscles by influencing the production or degradation of NO, therefore, may at the same time inhibit and promote the process of gastric emptying. Thus, relaxation of gastric fundus probably delays gastric emptying, whereas a reduced tone of the pylorus may accelerate the emptying process. Furthermore, it is reasonable to assume that pharmacological interference with NO mechanisms may have variable effects on gastric emptying of liquid and solid meal components as the proximal stomach have a major role in the emptying of liquids, whereas the distal stomach seems to play a more pronounced role in the emptying of solid foods. Consequently, it is not surprising that previous studies have shown considerable variations in the effects NO donors on gastric motor activity in healthy humans. Previously, a study with sublingual administration of GTN showed a decreased gastric emptying rate of a liquid meal,9 whereas other studies did not reveal a changed gastric emptying of semisolid or mixed liquid and solid meals after oral administrations of the NO donor isosorbide dinitrate.7, 8

To the best of our knowledge, only one previous study addressed the influence of intravenous infusion of GTN on gastric emptying.10 Thus, Sun et al.10 found that intravenous infusion of GTN 5 μg/min slowed the gastric emptying of 300 mL of a 25% glucose solution. Their dose of GTN was lower than the dose used in our study and probably, therefore, they did not observe any haemodynamic changes during the infusion. In our study, the chosen dose of GTN, which corresponded to the dose normally given to patients with severe angina pectoris, gave rise to a reduced systolic blood pressure, but the gastric emptying of the 1600 kJ mixed liquid and solid meal was not changed. The reason for this discrepancy is not obvious, but it is conceivable that a more pronounced reduction in the pylorus-mediated resistance to flow had counterbalanced the concurrent effect of relaxation of the proximal stomach in our study. This explanation may indirectly be supported by the recent findings in dogs and pigs that inhibition of the NO synthase decreases gastric empting rate due to a change in the postprandial pyloric motility from relaxation to contraction dominant pattern.6, 21 However, it should be stressed that the number of subjects in our study was rather low, so it cannot be precluded that the observations were influenced by type 2 error. Moreover, it could be argued that our chosen dose of GTN was too low. On the other hand, our observations indicate that unacceptable influence on the systemic haemodynamics or unbearable discomfort might have occurred in at least some of the volunteers if we had chosen a higher dose of GTN.

The rhythm of the gastric electrical slow waves, which is generated by interstitial cells of Cajal, coordinates gastric peristalsis.22, 23 Cajal cells, distributed in specific locations within the enteric tunica muscularis, serve as electrical pacemakers and mediators of neuromuscular transmission.24 These cells have close relationships with neurones of the myenteric plexus and are specifically responsive to NO neurotransmission.25 Furthermore, the NO donor, sodium nitroprusside, was recently shown to increase the pacemaker frequency of guinea-pig gastric antrum.26 In assumption that the postprandial frequency of antral contractions reflects the rhythm of gastric slow waves, the present study indicated that the rhythm in normal humans is not changed by pharmacological activation of the nitrergic neurotransmission. This finding is in accordance with a previous electrogastrographic investigation showing that inhibition of the endogenous NO synthesis did not change the basal electrical rhythm of the stomach in healthy volunteers.27 On the other hand, it should be emphasized that the scintigraphic technique is quite a crude and not validated method to assess antral motility after a meal. Although it seems capable of measuring the frequency of antral contractions, it provides no information about the amplitude. Therefore, weakened or strengthened antral contractions of unchanged frequency would not have been detected in our study.

The impact of the nitrergic system on small intestinal motility has been the topic of extensive research based on animal models.11–13 Thus, NO mechanisms are involved in the regulation of coordinated activity of the circular and longitudinal smooth muscles known as the peristaltic reflex that brings about the postprandial propulsion of intestinal contents. The relaxation component of the reflex has been shown to be mediated by inhibitory motor neurones of the enteric nervous system that coexpress NO synthase and vasoactive intestinal polypeptide.13 In humans, the importance of the nitrergic system for normal as well as abnormal small intestinal motor function is far from clarified. In healthy fasted volunteers, NO has been reported to modulate the migrating motor complex.14 Furthermore, few studies indicate that NO plays a role in the development of chronic intestinal pseudo-obstruction and post-operative ileus.28, 29 However, the selective contribution of increased or decreased NO synthesis to the pathogenic basis of these conditions remains unestablished. We are not aware of any previous study that aimed to evaluate the effect of a NO donor on the propulsive activity of the entire small intestine in healthy humans. According to our findings, a dose of the NO donor GTN that gave rise to a slight lowering of the systolic blood pressure and subjective discomfort in most of the participating volunteers did not change the small intestinal transit rate, and we do not believe that this observation was due to an inferior measuring technique. By using a deconvolution principle, it was possible to exclude any influence of the gastric emptying process on the variable calculated to characterize the small intestinal transit rate.16 Moreover, the calculated small intestinal mean transit time took into account the whole bulk of the radiolabelled test meal and not only the head of the meal components as is the case for the caecal appearance time, but also a traditional and often applied variable. Besides, the manometric finding of an unchanged postprandial duodenal pressure activity during GTN infusion seems to lend support to our scintigraphic observations. More likely, therefore, the absence of any detectable effects of GTN on the propulsive capacity of the small intestine in our study reflected the complexity of small intestinal motor control. Integrated by the enteric nervous system, the control system may provide the opportunity of introducing compensatory reactions when one of the components is modified.

In conclusion, intravenous infusion of GTN 1 μg/kg × min does not seem to induce major changes in gastric or small intestinal motor function after a 1600-kJ mixed liquid and solid meal in healthy volunteers.


This study was supported by grants from The Augustinus Foundation and The Aase and Ejnar Danielsen Foundation.