Glucagon is absorbed from the rectum but does not hasten recovery from hypoglycaemia in patients with type 1 diabetes

Abstract

AIMS

A failure to secrete glucagon during hypoglycaemia is near universal in patients with type 1 diabetes 5 years after disease onset and may contribute to delayed counter-regulation during hypoglycaemia. Rectal glucagon delivery may assist glucose recovery following insulin-induced hypoglycaemia in such patients and has not been previously studied.

METHODS

Six male patients (age 21–38 years) with type 1 diabetes (median duration 10 years) without microvascular complications, were studied supine after an overnight fast on two separate occasions at least 14 days apart. After omission of their usual morning insulin and 45 min rest, hypoglycaemia was induced by an intravenous insulin infusion which was terminated when capillary glucose concentration reached 2.5 mmol l⁻¹. Subjects were randomized to insert a rectal suppository containing 100 mg indomethacin alone (placebo) or 100 mg indomethacin plus 1 mg glucagon at the hypoglycaemic reaction. Serial measurements were made for 120 min.

RESULTS

In the two groups, mean (SD) plasma glucose concentrations fell to a similar nadir of 1.8 (0.7) mmol l⁻¹ (placebo) and 2.1 (1.2) mmol l⁻¹ (glucagon). Peak plasma glucagon following hypoglycaemia was higher in the glucagon group; 176 (32) ng l⁻¹vs. 99 (22) ng l⁻¹ after placebo (P = 0.006). However, the glucose recovery rate over 120 min after hypoglycaemia did not differ significantly.

CONCLUSIONS

Our results provide evidence for the absorption of glucagon from the rectum. They also indicate that 1 mg does not constitute a useful mode of therapy to hasten recovery from hypoglycaemia in patients with type 1 diabetes.

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT

• Patients with type 1 diabetes experience recurrent hypoglycaemia and have abnormal glucose counter regulatory responses with a failure to secrete glucagon. It is unknown if rectal glucagon is absorbed and what effect this may have on counter regulation from hypoglycaemia.

WHAT THIS STUDY ADDS

• A rectal suppository of glucagon results in a rise in plasma glucagon with metabolic effects in normal subjects. Similarly rectal glucagon results in a rise in plasma glucagon in patients with type 1 diabetes, but 1 mg does not improve recovery rates from experimental hypoglycaemia when compared with placebo.

• Larger doses of glucagon per rectum may provide pharmacological circulating concentrations with resulting therapeutic benefit during recovery from hypoglycaemia and deserves further study.

Introduction

Recurrent hypoglycaemia is a common adverse effect of insulin therapy in patients with type 1 diabetes and can lead to hypoglycaemic unawareness in a substantial number of patients [1]. Parenteral glucagon administration is a well established method of treating hypoglycaemic coma in the emergency department. However, in the domestic setting injections may be unreliable as training is required to perform this procedure quickly and effectively [2]. To circumvent potential delays in restoring blood glucose concentration to normal, alternative methods of glucagon administration have been reported. The effects of the intranasal route are transient and as a voluntary sniff is required [3, 4], this route of administration is not useful during profound hypoglycaemia. There are no reports detailing the absorption of glucagon from the rectum or in using this mode of delivery to explore if this results in improved recovery from hypoglycaemia in patients with type 1 diabetes.

In a pilot rectal absorption study, we investigated the effects of rectal administration of placebo or 0.1, 0.5 or 1 mg of glucagon in the presence of an absorption-promoter (indomethacin) on fasting plasma glucose, glucagon and insulin concentrations (glucose only in placebo group) in healthy individuals. We then investigated the clinical usefulness of this method of glucagon administration on recovery from induced hypoglycaemia in patients with type 1 diabetes.

Methods

Rectal absorption study – normal volunteers

Ten healthy male volunteers aged 18–47 years (BMI range 22–24 kg m⁻²) were studied on four separate occasions in random order using a double-blind protocol after an overnight fast and following defaecation. An in-dwelling intravenous plastic cannula was inserted into the left antecubital fossa for venesection and baseline blood samples were taken for plasma glucose, insulin and glucagon concentrations. The volunteers then self inserted a suppository containing 100 mg of indomethacin (as an absorption promoter) alone (placebo) or with a variable dose of glucagon (0.1 mg, 0.5 mg and 1 mg). The suppositories used in this study were manufactured in the Department of Pharmacy at Bristol Royal Infirmary using porcine powered glucagon (supplied by Novo Nordisk) in witepsol H15 base. This base melts at body temperature allowing spreading of the suppository up the rectum to enhance absorption. Plasma glucose, serum insulin and glucagon were measured at 30 min intervals for 120 min.

Hypoglycaemia study

Six male patients with type 1 diabetes aged 21–38 years were examined on two separate occasions in random order after an overnight fast having omitted their morning dose of insulin. Patients were maintained on human regular insulin (Actrapid, Novo Nordisk) before each meal and intermediate acting insulin (Protophane, Novo Nordisk) at bedtime. Type 1 diabetes was diagnosed clinically. All were of normal body weight (BMI range 21.2–24.5 kg m⁻²; median 22.4 kg m⁻²) with a median duration of diabetes of 10 years (range 6–20 years). All required insulin therapy from diagnosis. None had any evidence of significant diabetic microvascular complications on clinical examination and all reported normal hypoglycaemia awareness. All avoided any significant hypoglycaemic reactions in the 7 days prior to each study. Mean HbA_1c concentration over the prior three visits to the diabetes clinic in the group was 7.8 ± 1.2% (normal range 4–6%).

All patients attended on two occasions separated by at least 14 days and were studied in the supine position. Two indwelling venous cannulae were inserted into the left arm, one at the wrist in a retrograde direction for arterialized blood sampling (the hand was inserted into a thermostatically controlled hot air chamber at 55°C) and the other in the antecubital fossa for administration of fluid and insulin. An automated sphygmomanometer was attached to the right arm and blood pressure and pulse rate were recorded every 3 min during the study. At the end of a 30–45 min rest period during which time pulse rate and blood pressure were allowed to stabilize, baseline blood samples were taken for measurement of glucose, insulin, glucagon, catecholamines, growth hormone, cortisol and intermediary metabolites (lactate, acetoacetate, glycerol and free fatty acids).

Our standard unit protocol for inducing hypoglycaemia was used [5]. An infusion of insulin was commenced (Human Actrapid, Novo Nordisk 2.5 mU kg⁻¹ h⁻¹) until capillary glucose concentration as measured at the bedside fell to 2.5 mmol l⁻¹. The infusion was then stopped. At the hypoglycaemic reaction (R) characterized by a sudden rise in heart rate, the clock was reset to zero. At this point, each subject was given an identical suppository containing 100 mg indomethacin alone (placebo group) or 100 mg indomethacin plus 1 mg of glucagon in a combined preparation (glucagon group) in a witepsol base in a randomized double-blind design. These suppositories were inserted with a water-based lubricant. Serial measurements of plasma glucose, serum insulin, glucagon, catecholamines, growth hormone, cortisol and intermediary metabolites were repeated at intervals for the next 120 min. At the end of the study, each subject was fed.

Plasma glucose concentration was measured by a glucose oxidase method, C-terminal pancreatic glucagon concentration via radioimmunoassay (intra-assay coefficient of variation (CV) of 5.5% at 150 ng l⁻¹, interassay CV of 12.2% at 150 ng l⁻¹ and a sensitivity of 5 ng l⁻¹) [6], insulin by radioimmunoassay (intra-assay CV of 7.8%; interassay CV of 11.5% at 10 mU l⁻¹), and catecholamines by HPLC. This study was approved by the local ethics committee and all subjects gave their written informed consent for participation.

Data were analyzed using a mixed models approach to repeated measures. Since it was unlikely that carryover effects could influence the model, simple main effects (i.e. group and time) and their interaction were modelled. Significant main or interaction effects were further investigated using the method of Tukey to preserve the overall pair wise error rate. All analyses were performed using the MIXED procedure of SAS (SAS Institute Inc, v9.1). Data were normally distributed. All tests were two-tailed and P < 0.05 was considered statistically significant. Data are presented as mean ± SD and in the figures as mean ± 95% CI. A one-way random effect analysis of variance model was used to assess intra-individual and interindividual variability. The distribution of errors was assumed to be independent and intra- and interindividual variability was assumed to be random. The nested procedure of SAS (SAS Institute Inc, v 9.1) was used. The reliability coefficients were uniformly high, −0.89 for insulin, 0.86 for glucose and 0.90 for glucagon.

Results

Rectal absorption study in normal subjects (no hypoglycaemia, Figure 1) n = 10

Figure 1.

The effects of a rectal suppository containing 100 mg indomethacin alone (placebo x) or 100 mg indomethacin and 0.1 mg (open circles), 0.5 mg (open triangles) or 1 mg (open squares) glucagon on mean (95% CI) fasting plasma glucose (A), glucagon (B), and insulin (C) in 10 healthy volunteers. Plasma glucose rose in a dose dependent manner in response to increasing doses of glucagon. Plasma glucagon showed a significant rise from baseline after the suppository containing 1 mg glucagon. Insulin concentration rose after glucagon insertion but there were no significant dose dependent effects

Suppositories containing indomethacin alone had no effect on plasma glucose concentration.

Following 1 mg glucagon suppositories, the plasma concentration of glucagon increased from baseline concentrations and peaked at 30 min (P_dose*_time < 0.001). There was a statistically significant (P_dose*_time < 0.001) increase in glucagon concentration over time with the biggest change being found in the highest dose group. Insulin concentration tended to rise over the next 2 h (P*_time < 0.0001). However the magnitude of the rise could not be differentiated between doses of glucagon administered (P_dose*_time = 0.24). Increases in glucose concentration were significant with all three doses of glucagon and peak increases occurred 60 min after administration (i.e. later than peak glucagon increases). The overall response of plasma glucose concentration to rectal glucagon was also related to dose (P_dose*_time < 0.001). Two of the 10 volunteers described mild nausea and colicky abdominal pain following the 1 mg glucagon suppository.

Hypoglycaemia study (Figures 2 and 3) n = 6

Figure 2.

Mean (95% CI) plasma glucose (A) and insulin (B) concentrations following the administration of an intravenous infusion of insulin in six fasting male patients with type 1 diabetes studied on two separate occasions. Patients received a suppository of 100 mg indomethacin alone (placebo ×, solid lines) or 100 mg indomethacin and 1 mg glucagon (open circles, dotted lines) at the hypoglycaemic reaction (↓). The nadir of glucose and peak insulin concentration at the hypoglycaemic reaction did not differ on the two study days, and the recovery from hypoglycaemia was no different after placebo or glucagon

Figure 3.

Mean (95% CI) plasma glucagon (A), adrenaline (B) and noradrenaline (C) concentrations following the administration of an intravenous infusion of insulin in six fasting male patients with type 1 diabetes studied on two separate occasions. Patients received a suppository of 100 mg indomethacin alone or 100 mg indomethacin and 1 mg glucagon at the hypoglycaemic reaction (↓). Glucagon concentrations rose significantly after the glucagon suppository was inserted and remained elevated for 90 min but did not change after placebo. Peak catecholamine responses to hypoglycaemia were no different on the two study days

In two patients (one after placebo and one after glucagon) the capillary glucose concentration as measured at the bedside was still under 2.5 mmol l⁻¹ at 120 min after the hypoglycaemic reaction and intravenous glucose was administered. The time taken to reach R in the two groups (range 60–120 min, median 92 min) and rate of recovery from R on successive visits in the same individual was not significantly different (data not shown).

Plasma glucose concentration (Figure 2A) fell to a similar nadir on both study days from 12.6 (2.9) to 1.8 (0.73) mmol l⁻¹ in the placebo group and from 13.2 (3.2) to 2.1 (1.2) mmol l⁻¹ in the glucagon group. The rate of recovery of glucose from the nadir was similar with or without active glucagon in the suppository. In the first hour it was 0.63 (0.27) mmol l⁻¹ h⁻¹ (placebo) and 0.63 (0.44) mmol l⁻¹ h⁻¹ following glucagon. In the second hour after R it was 1.47 (0.37) mmol l⁻¹ h⁻¹ (placebo) and 1.68 (1.03) mmol l⁻¹ h⁻¹ after glucagon (P_dose*_time = 0.99).

The units of insulin administered to induce hypoglycaemia did not differ significantly on either study day (data not shown) and baseline and peak insulin concentrations were similar (Figure 2B) (P_dose*_time = 98).

Plasma glucagon concentration (Figure 3A) did not change in response to hypoglycaemia in the patients after placebo (baseline 111.7 (32), peak 111.7 (35) ng l⁻¹ at R + 60 min) but following glucagon administration there was a significant increase from the baseline of 105 (27) ng l⁻¹ within 15 min which peaked at 175.8 (31) ng l⁻¹ after 30 min and remained above baseline for 90 min (P = 0.0043).

There was no significant difference in the concentration of catecholamines at baseline or in the response to hypoglycaemia on each of the study days (Figure 3B,C adrenaline (P_dose*_time = 0.36), noradrenaline (P_dose*_time = 0.87)). Similarly there was no difference in baseline concentrations of growth hormone, cortisol or intermediary metabolites nor in their response to hypoglycaemia between the glucagon and placebo arms (Table 1-all P > 0.54). Cortisol, growth hormone and lactate increased significantly following hypoglycaemia in both groups (P = 0.02, P = 0.0005 and P = 0.037, respectively) (Table 1). Pulse rate and systolic blood pressure rose at R in both groups and the magnitude of the changes was not different between study days (data not shown).

Table 1.

Mean ± SD concentrations of growth hormone, cortisol and intermediary metabolites in six patients with type 1 diabetes taken at baseline, and peak/nadir response after induced hypoglycaemia with the addition of rectal glucagon (1 mg) or placebo at the hypoglycaemic reaction. There were no significant differences in the concentration of all parameters at any time point between groups, nor in the timing of the peak/nadir response to hypoglycaemia

		Basal	Peak/Nadir
Growth hormone (mU l−1)	Placebo	7.2 ± 3.6	59.3 ± 6.1
	Glucagon	5.8 ± 3	60.5 ± 3.1
Cortisol (nmol l−1)	Placebo	528.8 ± 63.4	744.2 ± 48.5
	Glucagon	486.5 ± 65	690 ± 18.1
Lactate (nmol l−1)	Placebo	1.38 ± 0.07	2.18 ± 0.31
	Glucagon	1.37 ± 0.12	2.12 ± 0.12
Acetoacetate (nmol l−1)	Placebo	0.89 ± 0.28	0.26 ± 0.07
	Glucagon	1.27 ± 0.38	0.18 ± 0.05
Glycerol (μmol l−1)	Placebo	10.4 ± 2	14.6 ± 2.8
	Glucagon	10.8 ± 2.1	18.4 ± 2.9
Free fatty acids (nmol l−1)	Placebo	0.63 ± 0.17	0.47 ± 0.14
	Glucagon	0.68 ± 0.24	0.32 ± 0.09

Glucagon reached a peak 30 min after administration and decayed to a half-life at 60 min (95% CI 55, 65), post suppository. Insulin reached a nadir 20 min (95% CI 15, 25) after glucagon administration. Both one and two phase elimination models were fitted to glucagon and to insulin. Congruence between the models suggests a single elimination model and also mitigates against nonabsorption.

Discussion

The principal counter regulatory hormones to hypoglycaemia are glucagon and adrenaline. Counter regulatory failure may either represent a reduced capacity to secrete the hormone or an alteration in the glycaemic threshold at which hormone secretion is initiated. Both types of abnormality are observed in patients with type 1 diabetes and deficient responses to hypoglycaemia have been described for all the principal counter regulatory hormones [7]. The impaired glucagon response contributes to the delay in recovery from hypoglycaemia and has been confirmed by other studies [8–10]. The underlying mechanism of its impaired secretion is uncertain but it is a specific, intrinsic defect of glucoreceptors in the pancreatic α-cells [11]. It does not appear to be related to a reduction in pancreatic islet cell mass [12] or to autonomic dysfunction [13–15] but may be influenced centrally [16] or by relative intra-islet hyperinsulinaemia [17, 18]. The present study confirms that the glucagon response to hypoglycaemia in patients with type 1 diabetes with duration > 5 years is absent.

The rectal route is an effective method of delivering hormones such as testosterone [19] and hydrocortisone [20] as well as anticonvulsants, analgesics, sedatives, anti-inflammatory drugs, anti-emetics and antibiotics, and is a readily learned technique [21–25]. The absorption of drugs from the rectum is significantly enhanced by the presence of an absorption promoter [23, 24] and nonsteroidal anti-inflammatory drugs have been most widely used in this respect. Volumes of 10–25 ml can be retained in the rectum without difficulty and the relatively constant environment of the rectum favours the absorption process [25].

The partial avoidance of hepatic first pass metabolism is an important aspect of rectal drug administration for drugs with peripheral action. The upper part of the rectum drains into the superior mesenteric vein which drains into the liver via the portal vein. The middle and inferior rectal veins in contrast drain the lower part of the rectum and venous blood is returned to the systemic circulation via the inferior vena cava and thus avoids hepatic first pass metabolism [22]. However, the witepsol-based suppository consists of saturated triglycerides and tends to spread to 5–7 cm within the rectum and therefore it is likely that a significant amount of glucagon absorbed is delivered directly to the liver via the portal vein which is an advantage in terms of its main site of action being in the liver parenchyma where it binds to hepatocyte receptors, exerts its metabolic effects and undergoes degradation (although some is excreted unchanged into the bile). In the basal state, up to 65% of glucagon delivered to the liver is extracted in a single transhepatic passage [26] and thus circulating plasma concentrations will be much lower than portal vein concentrations after rectal delivery.

The normal volunteer non-hypoglycaemia study showed that glucagon absorption occurs from the rectum in the presence of an absorption promoter and that significant increases in plasma glucose concentration can result in normal man without causing significant side-effects. Although the rectal administration of glucagon in the present series of patients during hypoglycaemia induced a significant increase in plasma glucagon concentration, this increase had no significant effect on glucose recovery from hypoglycaemia when compared with placebo. It is noteworthy in this regard that the dose of glucagon given parenterally to correct hypoglycaemia is a grossly supraphysiological one. Peripheral circulating concentrations of glucagon after IM injection of 1 mg glucagon reach a peak of 1800 ± 600 ng l⁻¹[27]. Peak glucagon concentrations achieved in the present study after rectal suppository were around 200 ng l⁻¹ and are similar to those achieved spontaneously during previous studies of hypoglycaemia [5, 28]. Another possibility is that the suppository was administered too late to affect glucose recovery. The patients with diabetes were hyperinsulinaemic at the time of hypoglycaemia (and when the suppository was used) and for the first 30–60 min thereafter due to the insulin infusion, and this may have counteracted positive effects of glucagon on gluconeogenesis. The peak insulin concentration observed was not different from previous studies using this protocol in normal individuals in our unit [5] but the duration of the hyperinsulinaemia in the patient with diabetes, was longer than normal subjects probably reflecting the longer duration (and hence dose) of insulin infusion required to achieve a glucose <2.5 mmol l⁻¹ in these patients. Within 15 min of the hypoglycaemic reaction, circulating insulin concentrations had fallen significantly to 100 mU l⁻¹ and by 30–60 min to 30–50 mU l⁻¹. There is little known about circulating concentrations of insulin during hypoglycaemia after subcutaneous insulin injection in diabetic patients. Bolli et al.[13] reported peak insulin concentrations of 30–40 mU l⁻¹ 1–3 h after 0.15 U kg⁻¹ subcutaneous injection of regular insulin in a group of 20 patients with type 1 diabetes which resulted in hypoglycaemia in 17 of the patients for many hours. Plasma insulin remained at this level for many hours despite the withdrawal of long acting insulin for 48 h before the experiments. In practice, therefore, a suppository containing 1 mg of glucagon is unlikely to be of practical benefit in the treatment of hypoglycaemia in patients with type 1 diabetes. Additional studies of higher dose suppositories are needed. However, two normal subjects experienced nausea and abdominal cramps after 1 mg glucagon and thus higher doses may not be well tolerated.

A potential weakness of the study is the relatively small number of patients with type 1 diabetes studied. However the near identical glucose recovery rates shown within each patient between the two studies despite very large differences in glucagon concentrations, suggests that if any effect were to be shown in a larger study, it is unlikely to be of clinical significance. This is further supported by the data showing very similar responses of insulin, catecholamines and intermediary metabolites to hypoglycaemia therefore excluding confounding influences on glucose recovery.

In summary the present study provides evidence that glucagon in suppository form with an absorption promoter is absorbed through the rectal mucosa with resulting metabolic effects in normal man. However, our observations with induced hypoglycaemia in patients with type 1 diabetes mellitus indicate 1 mg by suppository is not a useful therapeutic method to aid recovery from hypoglycaemia. However further studies with higher doses of labelled glucagon suppositories that result in pharmacological circulating concentrations of glucagon are needed to address the pharmacodynamics/kinetics of this route of delivery as well as its clinical utility to hasten recovery from hypoglycaemia in patients with type 1 diabetes.

GB and DP received funding from Bristol Diabetes Research fund.

Greg Gamble, Statistician, Department of Medicine, University of Auckland performed the statistical analysis.