Severe Hypomagnesemia and Hypocalcemia Linked to Semaglutide in Type 2 Diabetes: A Case Report

Introduction

Hypomagnesemia complicating diabetes is common, affecting between 10% and 38% of people with type 2 diabetes [1–3]. It is usually asymptomatic at serum concentrations down to <0.5 mmol/L [4], but lower levels can be associated with non-specific symptoms, including nausea, vomiting, loss of appetite, and fatigue [5], which can progress to more severe neuromuscular (weakness, cramps, tremor, tetany, seizures, and coma) and cardiologic (arrythmias and myocardial ischemia) effects [6,7]. Hypomagnesemia is caused by inadequate dietary intake and/or increased gastrointestinal/renal losses [4].

Diabetes-specific predisposing factors include poor glycemic control and treatment with metformin, sulfonylureas, and dipeptidyl peptidase IV inhibitors [1,3]. Of the newer blood glucose-lowering agents, sodium-glucose cotransporter-2 inhibitors (SGLT2i) appear protective [8]. There is evidence that glucagon-like peptide-1 receptor agonists (GLP-1RA) have a neutral effect on serum magnesium homoeostasis [9], but there have been 3 case reports of severe hypomagnesemia in people treated with these agents [10–12]. All of these cases were characterized by continued and marked vomiting and/or diarrhea, suggesting that GLP-1RA-associated hypomagnesemia likely reflected inadequate intake and/or excessive gastrointestinal losses.

In the present case of a woman with type 2 diabetes, severe hypomagnesemia developed in the presence of a number of factors known to affect serum magnesium concentrations, but with an atypical presentation in the aftermath of GLP-1RA therapy that was discontinued due to adverse gastrointestinal effects.

Case Report

A 73-year-old woman with a 15-year history of type 2 diabetes was referred by her primary care physician for specialist management because of suboptimal glycemic control on multiple oral blood glucose-lowering agents. Her glycated hemoglobin at that time was 8.7% (72 mmol/mol) despite treatment with maximal doses of metformin (2 g daily), gliclazide (120 mg of the modified release formulation daily), sitagliptin (100 mg daily), dapagliflozin (10 mg daily), and acarbose (100 mg twice daily). She confirmed full adherence to this complex treatment regimen. She was also taking irbesartan for hypertension, rosuvastatin for dyslipidemia, esomeprazole for reflux, and atenolol and dabigatran for chronic atrial fibrillation. She was a non-smoker but consumed 10 to 20 units of alcohol on most days of the week. She did not have a history of cardiovascular disease, had normal renal function but persistent microalbuminuria, had no retinopathy but was being treated for glaucoma, and did not have peripheral sensory neuropathy. On examination, her body weight was 67 kg (body mass index 26.2 kg/m2), her blood pressure was 130/89 mmHg, and her pulse rate was 81 beats/min and irregularly irregular.

She was reluctant to consider a strong recommendation for insulin therapy, and so subcutaneous semaglutide 0.25 mg weekly was started as an alternative. Sitagliptin was stopped, and she also ceased dapagliflozin because she felt it caused dyspnea. Her other medications for diabetes remained unchanged, and she continued to take irbesartan and esomeprazole for hypertension and gastroesophageal reflux, respectively. Although she experienced persistent nausea on the starting dose, she increased the semaglutide to 0.5 mg weekly after 4 doses but began to vomit and so discontinued therapy.

Because she had tolerable adverse effects on the initial 0.25 mg weekly dose, had lost 7 kg in body weight since initiation of GLP-1RA therapy, and her home blood glucose test results had improved on this treatment, she agreed to restart this dose 2 weeks after stopping the higher dose. However, she redeveloped marked upper gastrointestinal symptoms and discontinued after 2 further 0.25 mg weekly doses. She improved symptomatically over the next week. Two weeks after her last semaglutide dose, she collapsed at home while making a cup of tea. She did not have premonitory symptoms and sustained a head injury and minor burns to her face and forearms from the hot beverage. Her daughter found her unresponsive soon afterward and called an ambulance. She had regained consciousness when she was assessed in the Emergency Department but was confused. On admission, there were no localizing neurologic signs. A computed tomography scan showed no acute intracranial abnormality. An electrocardiogram showed atrial fibrillation but no other abnormality. As part of biochemical screening, she was found to have a serum magnesium concentration below the laboratory assay detection limit and a low total corrected serum calcium concentration (see Table 1).

She was started on intravenous electrolyte replacement, and her serum magnesium and calcium normalized during the next 8 h. She was discharged after 2 days of hospitalization, on oral magnesium and calcium supplements and her pre-admission oral blood glucose-lowering treatments. She remained off semaglutide.

At review 3 weeks after discharge, she was feeling well. She spontaneously reported that her longstanding muscle cramps had resolved since her inpatient treatment. Her serum magnesium and calcium concentrations were normal on continued supplements of both electrolytes. Because of its relatively strong association with hypomagnesemia, compared with other blood glucose-lowering therapies [1,13], and her desire to simplify her treatment regimen, metformin was stopped, but she was kept on gliclazide, sitagliptin, and acarbose. The results of her home blood glucose tests, done at least twice daily, remained consistently ≥9 mmol/L throughout. She subsequently agreed to start insulin (pre-dinner degludec/aspart 70/30) and has since achieved improved glycemic control (home blood glucose concentrations <8 mmol/L before breakfast), without hypoglycemia.

Discussion

The present patient had several potential causes for her acute presentation with severe hypomagnesemia. She was being treated with metformin, a sulfonylurea and a dipeptidyl peptidase IV inhibitor [1,3], while she was no longer experiencing the potentially protective effects of a SGLT2i [8]. Chronic proton pump inhibitor therapy [14] and alcohol consumption [15], both present in this case, have been associated with hypomagnesemia. Given the link between hypomagnesemia and cramps [5], which resolved with restoration of normal serum magnesium concentrations in our patient, we hypothesize that she had unrecognized longstanding hypomagnesemia associated with poor glycemic control, oral blood glucose-lowering therapies, chronic proton pump inhibitor therapy, and regular alcohol use. Although her semaglutide-associated gastrointestinal symptoms had largely resolved by the time she was hospitalized, it is likely that they contributed to a low serum magnesium concentration, which then became critical in the presence of the other continuing contributing factors at the time of her episode of collapse.

In comparison with previous reports of GLP-1RA-assocated severe hypomagnesemia [10–12], the present case exhibited several differences. First, and most importantly, the present patient no longer had significant gastrointestinal symptoms when she collapsed, while the previously published cases presented in the context of continued marked vomiting and/or diarrhea. Nevertheless, the temporal relationship between semaglutide use and the present patient’s acute presentation strongly suggest that it was contributory. Second, our patient had a larger number of potentially causal factors other than GLP-1RA use than did past cases (4 non-GLP-1RA medications associated with hypomagnesemia plus regular alcohol use). Under our hypothesis, this background multifactorial contribution combined with intermittent GLP-1RA use, and its associated gastrointestinal effects on magnesium homoeostasis, would put such a patient at risk of clinically significant hypomagnesemia that might have a delayed manifestation in the aftermath of use of a medication in the GLP-1RA class. Lastly, although there were no suggestive features of an epileptic seizure on examination of the present patient, such as tongue biting and urinary incontinence, this was the presenting feature in 2 of the previous cases [10,12] and may have been the cause of her sudden state of altered consciousness.

Hypomagnesemia can induce secondary hypocalcemia, as in the present patient, through attenuation of parathyroid hormone secretion and its action on target organs [16,17]. Indeed, serum calcium levels can be restored by magnesium supplementation alone [18], even if both electrolytes were replaced in the present patient. It is possible that her hypocalcemia, albeit relatively mild in comparison to the hypomagnesemia, contributed to her acute presentation with collapse.

This case has clinical implications. When a patient with type 2 diabetes, especially in the context of poor glycemic control and/or treatment with medications such as metformin that are associated with hypomagnesemia, presents with an acute neurological event such as in this case, serum magnesium and calcium concentrations should be measured as part of initial laboratory screening. In the broader context of usual care, practitioners treating people with type 2 diabetes should be aware of the potential for the medications they prescribe, whether for blood glucose-lowering [1,3] or other indications [4], to contribute to low serum magnesium concentrations. Although initial symptoms of clinically significant hypomagnesemia are usually non-specific, neuromuscular and cardiological symptoms, such as cramps, as in the present case, and palpitations, can indicate its presence. We recommend periodic monitoring of serum magnesium levels with or without serum calcium when patients with type 2 diabetes are taking several medications that can provoke hypomagnesemia, and especially when GLP-1RA therapy causes prominent/sustained gastrointestinal symptoms. This applies to the present patient in whom this was not done. The presence of cramps could have prompted a serum magnesium measurement and prevention of her acute presentation if supplementation had been given to correct a low concentration. Long-term magnesium supplementation may both improve associated symptoms and, given available indirect evidence, ameliorate the risk of chronic cardiovascular complications of type 2 diabetes [1].

Conclusions

The present case illustrates the potential for severe hypomagnesemia when people with type 2 diabetes are taking medications known to reduce serum magnesium concentrations. While not directly influencing magnesium homeostasis, recent semaglutide use likely contributed, through gastrointestinal effects, to the degree of serum magnesium lowering, which continued through other medications after semaglutide was stopped. Periodic monitoring of serum magnesium concentrations in at-risk individuals with type 2 diabetes is recommended since, as in the present case, the clinical presentation of severe hypomagnesemia can be sudden and without indicative warning symptoms.