Omeprazole and refractory hypomagnesaemiaBMJ 2008; 337 doi: https://doi.org/10.1136/bmj.39505.738981.BE (Published 10 July 2008) Cite this as: BMJ 2008;337:a425
- N Shabajee, medical student1,
- E J Lamb, consultant clinical scientist1,
- I Sturgess, consultant geriatrician2,
- R W Sumathipala, consultant geriatrician2
- 1Department of Clinical Biochemistry, East Kent Hospitals NHS Trust, Kent and Canterbury Hospital, Canterbury CT1 3NG
- 2Department of Health Care of the Older Person, East Kent Hospitals NHS Trust
- Correspondence to: E J Lamb
- Accepted 25 October 2007
Hypomagnesaemia is common in hospital patients and is often accompanied by other electrolyte abnormalities, such as hypocalcaemia, hypokalaemia, and hypophosphataemia, that may remain refractory to treatment until the underlying magnesium deficiency is corrected.1 We present two patients with refractory chronic hypokalaemia and hypocalcaemia secondary to hypomagnesaemia that resolved after withdrawal of the proton pump inhibitor omeprazole.
A 78 year old woman was admitted to hospital after an exacerbation of chronic obstructive pulmonary disease accompanied by diarrhoea and vomiting. Her medical history included breast cancer, ischaemic heart disease, myocardial infarction, osteoporosis, and spinal stenosis. A several year history of paraesthesia, numbness, and weakness in her limbs had been attributed to spinal stenosis, but surgery had been ruled out because anaesthesia was risky. Seven years earlier she had been investigated for postprandial pain, early satiety, nausea, and weight loss. Non-erosive duodenitis, diverticular disease, and a hiatus hernia had been diagnosed. She had been prescribed omeprazole (40 mg/day) and her symptoms improved slightly. In addition, she was receiving spironolactone, bumetanide, furosemide, gabapentin, co-codamol, hyoscine butylbromide, glyceryl trinitrate, losartan, aspirin, atorvastatin, ipratropium bromide, and salmeterol.
She was hypokalaemic on admission (table 1⇓), and this failed to respond to withdrawal of diuretics and intravenous and oral potassium replacement. On day 4 she developed hallucinations and became agitated: muscular excitability was noted but Chvostek’s sign was absent. She remained markedly hypokalaemic and was also hypocalcaemic and hypomagnesaemic: retrospective analysis of samples from the day of her admission showed that magnesium and calcium concentrationshad been low then.She was also hypophosphataemic. Her symptoms resolved after treatment with intravenous magnesium sulphate, calcium gluconate, and continued potassium. Magnesium concentration was normal while she received intravenous replacement, but when it was stopped magnesium fell again. She was discharged after 10 days and was taking oral magnesium glycerophosphate and a phosphate supplement; her diuretics were withheld even though she had some ankle oedema.
At outpatient follow-up, serum magnesium and calcium concentrations remained low, and her history was reviewed for potential causes of magnesium loss. She had had no further diarrhoea and denied using laxatives or alcohol. Measurement of urinary calcium and magnesium concentrations suggested appropriate renal conservation (table 1⇑). In light of a report of two cases of hypomagnesaemic hypoparathyroidism associated with omeprazole,2 we discontinued omeprazole and prescribed the H2 receptor antagonist ranitidine. Electrolyte status improved dramatically and was maintained even after her magnesium supplements were stopped (table 1⇑). The patient remains well and reported an improved appetite after omeprazole was withdrawn.
An 81 year old man who lived independently was admitted to hospital after presenting to his general practitioner with dizziness and nausea that had been ongoing for three months but had worsened in the previous two days. He had vomitedafter meals but had had no diarrhoea; he had also had urinary incontinence for two days and reported polyuria over the previous month.
His medical history included hypertension, ischaemic heart disease, benign prostatic hyperplasia, and diabetes controlled by diet. He was taking omeprazole (40 mg/day), isosorbide mononitrate, atenolol, atorvastatin, lisinopril, quinine, amlodipine, olmesartan, and aspirin. He had no history of alcohol misuse.
He had muscle cramps and paraesthesia, with pins and needles in his hands; Trousseau’s sign was elicited at venepuncture. He was very unsteady, falling to both sides, and had an irregular heartbeat. Hypokalaemia, hypocalcaemia, and hypomagnesaemia were seen, and his symptoms were attributed to these. Parathyroid hormone concentration was subsequently reported as within the reference range but inappropriate for his degree of hypocalcaemia (table 2⇓). His incontinence and raised C reactive protein were attributed to a urinary tract infection, which was treated with trimethoprim.
With oral potassium and intravenous calcium gluconate and magnesium glycerophosphate replacement his calcium and potassium concentrations gradually normalised and his dizziness and paraesthesia improved, but he developed hypomagnesaemia whenever the supplements were stopped. While in hospital he developed bradycardia (pulse 48 beats/min) and hypotension (blood pressure 90/55 mm Hg), and it was suggested that this might account for his dizzy episodes. His antihypertensive drug was stopped. An electrocardiogram showed atrial flutter and long (4 second) pauses. As his hypomagnesaemia was thought to be contributing to the abnormal electrocardiogram, we considered that the hypomagnesaemia should be corrected before fitting a pacemaker.
Drug review raised the possibility that omeprazole was causing his electrolyte disturbances, and it was stopped. Within a few days the patient “felt great.” Normal electrolytes were maintained without supplementation. He was discharged with outpatient follow-up.
Both cases of refractory hypomagnesaemia associated with concurrent electrolyte abnormalities resolved after omeprazole was withdrawn. In the first case we could document renal conservation of magnesium and calcium while the patient was receiving omeprazole. In the second case we documented an inappropriately normal parathyroid hormone response to hypocalcaemia, classically associated with hypomagnesaemia.3 These data seem to support a causative role of omeprazole in the development of hypomagnesaemia, as described earlier by Epstein et al.2
The symptoms of magnesium deficiency relate to the central role of magnesium in ATP metabolism and neuromuscular transmission, but it is generally difficult to ascribe symptoms solely to magnesium deficiency because other electrolyte abnormalities are present. However, refractory cardiac dysrrhythmias such as those seen in case 2 have been ascribed to magnesium deficiency.1 The polyuria in case 2 was possibly related to the patient’s chronic hypokalaemia, which causes downregulation of aquaporin-2 water channels and hence resistance to antidiuretic hormone.4 Both patients had been receiving omeprazole for a long time. Vomiting and diarrhoea, exacerbated by use of a loop diuretic in case 1, may have unmasked an underlying magnesium deficit.
Absorption of dietary magnesium occurs in the ileum and colon via carrier mediated transport and simple diffusion.1 The kidneys are highly efficient organs in magnesium conservation: about 3% of filtered magnesium is lost in the urine, mostof which is reabsorbed in the thick ascending limb of the loop of Henle.5 The renal conservation seen in case 1 and also by Epstein et al2 suggests that the action of omeprazole is unlikely to be at the renal level. Low gastric pH is thought to be important for the absorption of minerals. Metal ions bind to ligand binding sites on dietary fibre and may be displaced by hydrogen ions, facilitating absorbtion. Waves of acidity entering the small intestine from the stomach may help to keep mineral salts in solution until they can be absorbed.6 Omeprazole induced hypochlorhydria could therefore theoretically cause mineral deficiency, although there is no evidence that omeprazole use, at least in the short term, inhibits magnesium absorption.7
Although the mechanism remains unclear, there seems little doubt that omeprazole may cause hypomagnesaemia. Given the frequency with which this drug is prescribed it would seem likely that use of the drug in the absence of other illness predisposing to electrolyte imbalance does not invariably cause hypomagnesaemia, but these cases are unlikely to be isolated. Doctors should consider omeprazole as a possible causative agent when investigating hypomagnesaemia.
Cite this as: BMJ 2008;337:a425
We thank the staff of the clinical biochemistry department.
Competing interests: None declared.
Provenance and peer review: Not commissioned; externally peer reviewed.