An Unusual Presentation of Proton-Pump Inhibitor Induced Hypomagnesemia in a 93-Year-Old Nursing Home Resident

A 93-year-old Caucasian woman had been a resident in a nursing home (NH) for several years. She was followed for several comorbidities including dementia of the Alzheimer type, hypertension, osteoarthritis, non-ST segment elevated myocardial infarction, and a history of gastritis in 2017 with secondary iron deficiency anemia. 

The patient required one assist for transfers and experienced occasional urinary incontinence. She was able to feed herself and enjoyed reading the paper and conversing with her family over the phone. She was generally oriented and exhibited no behavioral or psychological symptoms related to dementia. Her medications prior to this case included daily doses of metoprolol succinate ER 25 mg, aspirin 81 mg, lisinopril 20 mg, omeprazole 20 mg, as well as a probiotic and laxatives as needed for constipation.

On routine visits, the patient voiced few, if any, complaints other than a desire to return home to live with her husband. She had lost approximately 10 pounds during the first few months of the COVID-19 pandemic, which she regained slowly over the summer. Her appetite was consistently described as very good, and she never experienced dysphagia, abdominal pain, diarrhea, nausea, or emesis. Over the past 12 months, she had been hospitalized once for acute bronchitis. 

A routine 60-day visit was conducted at which the patient expressed no concerns. Her vitals remained stable with a blood pressure of 147/66, pulse of 62, oxygen saturation of 96%, and weight of 171 lb (BMI 30.3). Her exam was remarkable for clear lungs with a regular heart rhythm without gallop or murmur and trace pedal edema. Her conjunctiva were without pallor. Labs done 3 weeks prior revealed a sodium of 136 mEq/l, potassium of 4.5 mEq/l, chloride of 102 mEq/l, and CO₂ of 25 mEq/l. Her BUN was 17 mg/dl, creatinine of 0.8 mg/dl, and calcium of 9.0 mg/dl. 

One week later, the patient was seen for a complaint of “I can’t use my right hand.” She noted numbness involving all fingers on the right hand, both anteriorly and posteriorly, and attributed her symptoms to a recent blood draw which was taken from the right antecubital vein. Nurses noted that when she was distracted, she appeared to have normal motor function of the right hand. On examination, there was no evidence of trauma. She had difficulty gripping with the right hand but maintained normal wrist flexion and extension as well as finger abduction and adduction. There were no demonstrable sensory or vascular deficits of either the right upper extremity or hand. While the etiology of the patient’s complaint was unclear, a referral to occupational therapy was made and analgesics offered for any associated discomfort.

Over the next week, the patient continued to report difficulty using her right hand but had no other complaints. She continued to eat and drink well and had no change in bowel habits, diarrhea, nausea, or emesis. At 1 week from her initial presentation, her examination was remarkable for what appeared to be spasm of all 5 fingers on the right. Her grip was significantly diminished. Labs were obtained that day, which were notable for the following: sodium of 141 mEq/l, potassium of 2.3 mEq/l, chloride of 106 mEq/l, CO₂ of 24.9 mEq/l, BUN of 18 mg/dl, creatinine of 0.9 mg/dl, glucose of 78 mg/dl, calcium of 6.7 mg/dl, and albumin of 3.2 (corrected calcium 7.3 with normal range of 8.2-10). 

Her liver panel was within normal range with an Hgb of 12.6 g/dl, WBC of 8.0 K/uL, and platelets at 169,000 K/uL. Radiographs of the right wrist and hand revealed osteopenia and degenerative changes; mild soft tissue swelling was also noted in the right hand. An electrocardiogram was notable for sinus rhythm, left bundle branch block, and possible inferior infarct, age indeterminate.

The tentative diagnosis, at this time, was of tetany involving the right hand secondary to hypocalcemia. In view of the above findings, a serum magnesium level was obtained and reported back at less than 0.6 mg/dl (range 1.6-2.8). The patient’s phosphorus level was 2.6 mg/dl (range 2.4-4.7), PTH at 45 pg/ml (range 12-88), and Vit D, 25OH at 115.6 ng/ml (range 30-80). Omeprazole was discontinued and aggressive oral repletion was immediately started with calcium carbonate 600 mg QID, potassium chloride 40 mEq TID, and magnesium oxide 400 mg BID.  

Within two days, her potassium had increased to 3.1 and her calcium to 6.9. Over the next 5 days, the tetany gradually resolved, and by one week, the serum calcium and potassium normalized (potassium 5.0; calcium 9.9). At 9 days, the patient continued to have some difficulty bending her right middle finger but was noting improvement with continued occupational therapy. The serum magnesium had increased to 1.3 after 9 days and 2.1 after 12 days (range 1.6-2.6). Calcium and magnesium were reduced to once a day, and the potassium chloride was discontinued. 

One month later, nurses reported that the patient had 2 episodes of coffee ground emesis. Physical exam was unremarkable, and she was tolerating her diet and denied abdominal discomfort. Given that the patient was hemodynamically stable, an H2 blocker was prescribed (Cimetidine), aspirin was discontinued, and a stat complete blood count (CBC) ordered. The next afternoon, the patient had 2 “black” stools but remained stable with a blood pressure of 136/68 and pulse of 64. An abdominal exam revealed no tenderness, masses, or organomegaly. Serial CBCs revealed a drop of hemoglobin from 11.0 g/dl the day of the episodes to 9.0 g/dl the next day. Blood urea nitrogen (BUN) increased to 62 mg/dl from a baseline of 18 with no change in creatinine at 0.8 mg/dl. 

Although the patient and her family preferred not to be hospitalized, melena persisted and her blood pressure dropped, necessitating transfer to the hospital. On admission, her hemoglobin had dropped to 6.6, requiring transfusion of 2 units packed red blood cells. Following stabilization, the family refused any further workup, including endoscopy, and the patient returned to the nursing facility on pantoprazole 40 mg daily. Magnesium oxide was prescribed at 400 mg twice daily to hopefully overcome the adverse effect of the PPI on magnesium absorption. With continued magnesium supplementation, no further abnormalities in serum magnesium, potassium, or calcium were documented nor was there recurrence of the presenting symptoms.

We believe that the 93-year-old NH resident described in this case study represents the oldest patient with PPI-induced hypomagnesemia. The oldest patient previously described in published reports was 90 years old with no reports found in NH residents.¹ Given that upwards of 27% of NH residents may be prescribed a PPI, the complications experienced by this patient are particularly noteworthy.² 

 PPI-induced hypomagnesemia typically occurs after years of use, often facilitated by urinary losses secondary to diuretics.³ Both loop diuretics and thiazides disrupt the normal mechanism of cation exchange in the ascending limb of the loop of Henle and distal convoluted tubule, respectively. Danzinger et al⁴ found that patients taking both a PPI and diuretic had an odds ratio of 1.54 for developing hypomagnesemia compared to an odds ratio of 0.92 for patients only taking a PPI.  

In contrast to diuretic-induced renal loss of magnesium, PPI-induced hypomagnesemia has also been attributed to changes in GI absorption.⁵ Magnesium is absorbed both actively and passively by the gastrointestinal tract. Two mediators in the active transport of magnesium include transient receptor potential melastatin 6 and 7 (TRPM6 and TRPM7). TRPM6 is especially important when luminal concentrations of magnesium are low, increasing the absorption of magnesium.⁶ It has been hypothesized that changes in intestinal pH could affect the TRPM ion channels and affect magnesium absorption.⁷

However, while H2 receptor antagonists also affect intestinal pH, they have not been associated with hypomagnesemia, which is suggestive of a pH-independent mechanism.⁴ In support of this hypothesis, Serfaty-Lacrosniere et al⁸ found that moderate hypochlorhydria seen in short-term omeprazole use did not increase the risk of developing mineral malabsorption.

Mackay et al³ speculated that PPI medications could affect the functioning of tight junctions—the passive mechanism of magnesium absorption—which accounts for 90% of magnesium absorption. The paracellular mechanism for magnesium absorption occurs via simple diffusion down the concentration gradient between enterocytes, and the rate of absorption is dependent on the transepithelial voltage gradient. The tight junction proteins implicated in magnesium absorption are claudin-16 and claudin-19.⁹

However, Cundy and Dissanayake¹⁰ suggested that because PPI-induced hypomagnesemia can be corrected with high doses of magnesium supplements, passive magnesium transport remains intact while the active transport is disturbed by the PPI. More research in this area is needed in order to definitively state the pathophysiology responsible for hypomagnesemia secondary to the use of PPIs.

Hypomagnesemia causes hypocalcemia because of magnesium’s role in parathyroid hormone (PTH) secretion. Initially, low serum magnesium levels produce an increase in PTH secretion, similar to the effect of low serum calcium. However, if magnesium levels are too low, there is a paradoxical block of the secretion of PTH.¹¹ Hypokalemia can be attributed to an increased secretion of potassium into the urine through the renal outer medullary potassium (ROMK) channel in the kidney. Magnesium is an inhibitor of the potassium transport through the ROMK channel and must be corrected before the hypokalemia will resolve.¹² 

Hypomagnesemia can cause a variety of manifestations among patients, associated with the low serum levels of magnesium, calcium, potassium, and phosphate. Published cases include patients experiencing hypocalcemic seizures, loss of consciousness, and cardiac arrhythmias, including U waves, prolonged QT interval, and ST depression.¹³ Our patient presented with tetany and muscle weakness in her right hand, a presentation similar to the patient described by Subbiah et al¹⁴ 4 weeks after being prescribed a PPI.

Although up to half of all PPIs are administered without appropriate indications, our patient’s recurrent GI bleed shortly after discontinuation of the PPI speaks to the potential adverse consequences of even well-intentioned deprescribing.¹⁵ She was once again prescribed a PPI upon her return from the hospital, and magnesium supplementation was continued to hopefully overcome its effects on absorption.  

Our case highlights the need to monitor serum magnesium levels for NH residents taking PPIs, even those who are not concomitantly taking diuretics. Further, our case scenario demonstrates how cognitive bias can often delay diagnosis. The examiner did not initially weigh all possible options given the patient’s insistence that her symptoms were the result of a recent blood draw and that the classic signs of tetany were not initially present. This case is a reminder that atypical presentations are indeed the norm in the NH and that medication side effects are often the etiology behind changes in condition

Authors: Sarah E. Crawford¹ • Paul Katz, MD¹

Affiliations:
¹Florida State University

Disclosures:
The authors report no relevant financial relationships.

Correspondence:
Paul Katz
Florida State University
1115 W Call St., Ste 4300
Tallahassee, FL
​​​​​​​Email: paul.katz@med.fsu.edu