Drug interactions, renal impairment and hypoglycaemia in a patient with Type II diabetes

Introduction

In this report we describe a ‘Dangerous corner’ in the life of a patient with diabetes. As in J. B. Priestley's play, a trivial upset triggered a sequence of adverse interactions that would never have arisen had not the protagonists, in this case drugs, coexisted so precariously. We discuss the mechanisms of these drug interactions, how they might have been prevented, and how drug therapy was optimized in their wake.

Case history

A 64 year old retired Polish farm worker with Type II diabetes mellitus was brought to the hospital unconscious. According to his wife, during the previous week his appetite had dwindled and his stools had been intermittently loose. Earlier on the day of admission, he had become generally unwell, tremulous and perspiratory, with periods of confusion and dysarthria. On arrival, capillary blood glucose was estimated at 2.2 mmol l⁻¹ by fingerprick testing (Medisense). He regained consciousness rapidly after a 50 ml intravenous bolus of 50% glucose and a more detailed history was obtained. A laboratory report later confirmed that venous blood glucose at the time he presented was 1.1 mmol l⁻¹.

Diabetes had been diagnosed 4 years earlier by his General Practitioner and he was taking two oral hypoglycaemic agents, glibenclamide and metformin. There had been no previous admissions to hospital in connection with diabetes. For at least 2 years he had taken a nonsteroidal anti-inflammatory drug (NSAID), naproxen, to relieve pain arising from osteoarthritis of the spine and knees. In the past, a duodenal ulcer had been diagnosed by barium meal and he was taking ranitidine, an H₂-receptor antagonist, regularly. He had also consulted a urologist for symptoms of bladder outflow obstruction for which he was taking terazosin, an α-adrenoceptor antagonist. His medication at the time of admission is set out in Table 1.

Table 1.

The patient's prescription at the time of admission.

He had been treated for hypertension for 3 years with bendrofluazide, and 2 months earlier he had been referred to a nephrology clinic with elevated blood pressure and proteinuria. At the clinic, he was found to have early diabetic retinopathy, and investigations revealed a 24 h urinary protein excretion of 1.6 g, and raised plasma urea and creatinine concentrations [Table 2; column (a)], implying the presence of diabetic nephropathy; abdominal ultrasound scan showed no obstructive uropathy or other renal abnormality. He was advised to take ramipril, an angiotensin-converting enzyme (ACE) inhibitor. One month later, renal function was reassessed by his GP and found to be essentially unchanged [Table 2; column (b)]. There was no family history of relevance. He was an ex-smoker and admitted to only occasional use of alcohol—he had certainly had no alcohol shortly prior to his admission.

Table 2.

Serial changes in plasma electrolytes, urea and creatinine.

On examination he weighed 110 kg and was obese. The pulse was 90 beats min⁻¹ and regular, blood pressure 190/95 mmHg. Examination of the heart and chest was unremarkable, and oxygen saturation was 97% while breathing room air. The abdomen was soft, there was slight epigastric tenderness, and the bladder was palpable. Rectal examination revealed a moderately enlarged prostate. The central nervous system was normal apart from decreased vibration sense at the right ankle. Dipstick testing of his urine revealed: glucose—negative; protein—‘one plus’; blood—trace.

Plasma biochemistry is shown in Table 2 [column (c)]. The electrocardiogram showed sinus rhythm and incomplete left bundle branch block. A radiograph of the chest was within normal limits. Later investigations revealed 5.9% glycosylated haemoglobin (reference range 3.8–5.8%) and prostate-specific antigen <2.0 ng ml⁻¹ (reference range 0–4 ng ml⁻¹).

He was admitted to hospital. The oral hypoglycaemic drugs and potential nephrotoxins (ramipril and naproxen) were stopped. An intravenous infusion of saline was started and a urinary catheter inserted to monitor urine output. Capillary blood glucose was measured at regular intervals. Urine output was well maintained at over 100 ml h⁻¹, but brief hypoglycaemic episodes recurred for about 24 h after admission (Figure 1). Blood pressure remained slightly elevated at 160/95 mmHg. There were no further gastrointestinal symptoms during his admission, although he began to complain once more of arthralgia.

Figure 1.

Blood capillary glucose concentration (estimated by fingerprick testing) plotted against time after admission. Three boluses of glucose (one intravenous followed by two oral) were required in the first 24 h to restore and maintain physiological blood glucose.

Once blood glucose had risen to a consistent safe level, an alternative sulphonylurea, gliclazide, was introduced at a dose of 80 mg daily. Five days after admission, when renal function was stable [Table 2; column (d)], he was discharged from hospital.

Two months later, the patient was readmitted to a surgical team with intermittent abdominal pain and a further, more dramatic decline in renal function. He suffered a cardiac arrest whilst an inpatient and could not be resuscitated. An autopsy disclosed a recent myocardial infarction with severe atheromatous disease affecting the coronary and renal arteries.

Discussion

Hypoglycaemia following impairment of renal function alerted us to consider the elimination of the oral hypoglycaemic drugs he was taking. Metformin is excreted unchanged by the kidney and accumulates in renal failure. It acts by decreasing glucose absorption from the intestine, depressing hepatic gluconeogenesis and increasing peripheral glucose uptake. It has not been found to cause hypoglycaemia when taken alone [1, 2], but might exacerbate hypoglycaemia induced by sulphonylureas, especially when present at high concentration. It can cause diarrhoea, but this usually becomes apparent soon after the drug is introduced. Our patient had been taking metformin for 4 years without previous untoward effects.

Glibenclamide, like all sulphonylureas, is capable of inducing hypoglycaemia which can be fatal [3, 4]. It acts by increasing insulin release from pancreatic beta cells and sensitizing peripheral tissues to the effect of insulin. The risk of prolonged hypoglycaemia in patients taking a sulphonylurea depends partly on the relative potency of the drug concerned—glibenclamide is much more potent than other sulphonylureas [5]—and partly on the drug's route of elimination. Glibenclamide is cleared by hepatic metabolism but has two active metabolites, 4-trans-hydroxyglibenclamide and 3-cis-hydroxyglibenclamide, both renally excreted [6]; renal impairment thus increases the risk of hypoglycaemia in patients taking glibenclamide, especially when combined with metformin, and seems likely to have been the reason for our patient's presentation.

Whereas initial management involved an i.v. infusion of 0.9% saline (in view of the concern about his state of hydration), this ought to have been combined with an infusion of i.v. glucose. His blood glucose would almost certainly have risen faster had he received 5-10% glucose (the preferred crystalloid in a patient with sulphonylurea-induced hypoglycaemia) by infusion. It may sometimes be necessary to continue the glucose infusion for several days in such cases.

This patient's renal function was impaired when he first presented to the nephrology clinic and the coexistence of proteinuria suggests that he may have had diabetic nephropathy. The deterioration in renal function probably resulted from coprescription of an NSAID and an ACE inhibitor in the presence of volume depletion. In volume-replete individuals with normal renal function, these drugs can be used together relatively safely, but when renal perfusion is reduced, for example, by hypovolaemia or renal artery stenosis, they interfere with vital autoregulatory mechanisms protecting glomerular filtration [7]. In response to reduced renal blood flow, the afferent glomerular arteriole is dilated by prostaglandins formed within the glomerulus [8], whereas the efferent arteriole is constricted by angiotensin II, generated following release of renin from the juxtaglomerular apparatus [9]. Inhibition of either one of these compensatory mechanisms may not compromise glomerular filtration, but if prostaglandin synthesis is inhibited by an NSAID while the formation of angiotensin II is blocked by an ACE inhibitor, glomerular filtration can be seriously impaired. Our patient, we now know, had bilateral renal artery atheroma, as well as acute volume depletion due to diarrhoea; renal perfusion was therefore very sensitive to the concurrent inhibition of prostaglandin and angiotensin II production.

The importance of renal artery atheroma relative to volume depletion as prime facilitator of the interaction in this patient is difficult to quantify, even with hindsight. One would expect a man aged 64 years with Type II diabetes to have significant atheromatous arterial disease, but he displayed no clinical manifestations of peripheral vascular disease and the renal ultrasound scan showed no reduction or disparity in renal size. Moreover, renal function did not change during the first month after introduction of the ACE inhibitor—had there been significant bilateral renal artery stenosis one would have expected renal function to deteriorate sooner than this. The presence or absence of significant renal artery stenosis could have been established by renal isotope scanning, followed if necessary by contrast aortography or spinal CT angiography, had not all the clinical evidence been reassuring.

The interaction of ACE inhibitors with NSAIDs is a class effect insofar as it concerns reduced angiotensin II activity [7]. It is likely that angiotensin II receptor antagonists, such as losartan, would interact in the same way; animal models of renal artery stenosis support this assumption [10], although no human data are as yet available.

There have been several reports of ACE inhibitors precipitating hypoglycaemia in diabetic patients treated with sulphonylureas [11, [12]. The mechanism of this is unclear but it has been postulated that systemic vasodilatation may increase insulin delivery to the periphery or that ACE inhibitors directly retard hepatic gluconeogenesis. This interaction is unlikely to have been relevant in this patient.

After managing the acute event, we faced a dilemma about the continuing treatment of our patient's diabetes, hypertension and arthritis. Gliclazide, which is eliminated by hepatic metabolism, was started in place of glibenclamide, though one could argue that a low dose of insulin might have been even safer. It would clearly have been undesirable to reintroduce both naproxen and ramipril, yet good indications for each of these drugs remained: our patient's main daily complaint was of arthritic pain and stiffness, which had been well controlled by naproxen, and the evidence in favour of using an ACE inhibitor in a patient with diabetic nephropathy and hypertension was difficult to ignore [13]. In the immediate discharge period, both drugs were withheld, and paracetamol combined with a minor opioid was used for analgesia. To maintain blood pressure control, the dose of terazosin was increased to 4 mg daily and bendrofluazide reintroduced at half the original dosage. The final decision about whether to restart naproxen or ramipril would have depended upon his symptom control and renal function in the medium term. Sadly, these considerations were overtaken by a fatal coronary event, another ‘Dangerous corner’ that many diabetic patients with progressive nephropathy unfortunately fail to negotiate.

Learning points

Renal function deteriorates with increasing age in patients with diabetes, and can affect the elimination of drugs they are taking; it should be carefully monitored.

Glibenclamide has active metabolites that accumulate and can cause hypoglycaemia in patients with renal impairment.

Patients with sulphonylurea-induced hypoglycaemia should be given i.v. glucose by infusion to maintain blood glucose concentration following initial correction.

Concurrent use of an ACE inhibitor and an NSAID can lead to impaired renal function in patients with reduced renal perfusion. Those most at risk include patients with bilateral renal artery stenosis, or with hypovolaemia resulting from, for example, pathological fluid loss or overdiuresis.

Bilateral renal artery stenosis should be suspected in any patient with evidence of coronary, cerebral or peripheral vascular disease, or with a strong risk factor for atheromatous arterial disease, such as diabetes. Where suspicion is high, appropriate imaging should precede treatment with agents that either inhibit angiotensin II production or antagonize its effects. Once treatment is established renal function should be monitored regularly (at least every 6 months).