Over-the-counter antacids linked to severe hypokalaemia in the context of threatened preterm labour

Background

While hypokalaemia is a common diagnosis in inpatients with an incidence of approximately 20%, it is only found in 1%–3% of outpatients.1 The incidence in pregnancy is accepted to be even lower as this is a cohort of younger women with fewer comorbidities and less medication use.

It is important to investigate and treat the underlying cause of hypokalaemia appropriately. Hypokalaemia is often asymptomatic but symptoms include weakness, palpitations, parasthesia, muscle cramps and abdominal symptoms. The degree of hypokalaemia can be classified as mild (3.0–3.5 mEq/L), moderate (2.5–3.0 mEq/L) or severe (<2.5 mEq/L). The differential diagnosis for hypokalaemia can be categorised into abnormal losses, inadequate intake, transcellular shifts or an erroneous result.2

Hypokalaemia is a common side effect of many classes of medications, for example diuretics, glucocorticoids and laxatives.3 On reviewing the literature around hypokalaemia, we have been unable to find a documented case where severe hypokalaemia was caused by the ingestion of over-the-counter antacids (calcium and magnesium carbonate).

Case presentation

A 26-year-old woman presented initially at 29 weeks’ gestation with abdominal pain. She was gravida 6, para 3 with three previous normal vaginal deliveries (all at term) 8, 3 and 2 years ago with birth weights of 3.4 kg, 3.8 kg and 3.2 kg, respectively. Her medical history included bulimia as a teenager. She was not under the care of the mental health team. In this pregnancy she denied symptoms of an active eating disorder and had gained an appropriate amount of weight. She was on sertraline (50 mg) for anxiety and depression but felt her mood was stable. No abnormalities were found on her dating and anomaly scans. Her body mass index was 18.7 kg/m² at booking. There were no social concerns. At 21 weeks’ gestation she had been treated for a urinary tract infection but had otherwise been under midwifery-led care as per the local low-risk multiparous pathway. She was group B Streptococcus positive on a genital swab in this pregnancy.

At 29 weeks’ gestation she presented with suprapubic pain. On speculum examination her cervix was normal and there was felt to be no evidence of preterm labour. She was diagnosed with pelvic girdle dysfunction and a referral was made to physiotherapy. She presented to the labour suite again at 32 weeks’ gestation. She was tightening once every 10 minutes and was very uncomfortable. The registrar was not able to assess her immediately but she was treated with nifedipine tocolysis and intramuscular steroids to expedite fetal lung maturation for suspected preterm labour. She was later assessed by the registrar and found to have a long and closed cervix, and PartoSure was negative. Cardiotocography was normal apart from the uterine tightenings. After further assessment by a senior obstetrician, preterm labour was felt to be unlikely and she was discharged.

She presented for a third time at 34 weeks and 5 days’ gestation with further abdominal pain and concerns about spontaneous rupture of membranes. She had continued to have intermittent tightenings in the intervals between reviews. At this time she had two to three palpable tightenings in 10 minutes and she reported a pink-coloured vaginal loss but was otherwise well with good fetal movements. On speculum examination there was no evidence of liquor, and vaginal examination found that her cervix was 2 cm long and 2 cm thick. Cardiotocography was normal. It was felt that threatened preterm labour was likely from her clinical symptoms and she was admitted. Her tightenings continued to be regular so she was re-examined 2 hours later and there was found to be no change in her cervix.

As there had been several admissions, a full set of routine blood tests were sent at this time (full blood count, urea and electrolytes, liver function tests and C reactive protein). A serum potassium level of 2.4 mmol/L was found. The other blood results were unremarkable.

Investigations

Initial blood results:

• Plasma potassium 2.4 mmol/L (3.5–5.3 mmol/L).

• Plasma creatinine 51 µmol/L (45–84 µmol/L), urea 2.7 mmol/L (2.5–7.8 mmol/L).

Repeat blood results:

• Plasma potassium 2.3 mmol/L (3.5–5.3 mmol/L).

Further investigations given the severe hypokalaemia:

• Plasma magnesium 0.61 mmol/L (0.7–1.0 mmol/L).

• Venous blood gas pH 7.552, bicarbonate 36.7 mEq/L and base excess (BE) 13.1.

• ECG sinus rhythm.

Differential diagnosis

The underlying cause of hypokalaemia was initially thought to be an erroneous laboratory result as the patient reported no additional symptoms. As the result was confirmed with a second laboratory sample, it was questioned whether the patient had had a relapse of her bulimia symptoms, which she was not wanting to admit, or poor oral intake in the context of an eating disorder.

She reported feeling well in herself with no bowel symptoms or vomitting and had a good appetite. A drug history was taken at this time and she reported only sertraline 50 mg once daily and two doses of naproxen, which she was advised were not suitable in pregnancy. She reconfirmed that she had not been symptomatic of bulimia for 2 years. Her blood pressure was 112/65 mmHg and her abdomen was soft and non-tender.

We considered the following list of differential diagnosis:

• Pseudohypokalaemia (laboratory error or delayed sample).

• Sertraline.

• Inadequate oral intake in the context of an eating disorder.

• Induced vomitting in the context of bulimia.

• Over-the-counter medications, for example laxatives.

• Pre-eclampsia.

None of these explained the finding of hypokalaemia in this case. The patient had been on sertraline for 2 years with good compliance, no recent dose changes and was taking the lowest possible dose. She had previously had blood tests taken while on sertraline with no electrolyte abnormalities found. Another attempt at exploring the cause with the patient revealed an unexpected addition to the drug history, over-the-counter calcium and magnesium carbonate (Rennie) tablets at a dose of 72 tablets daily throughout the majority of her pregnancy. She reported that she suffered badly from reflux and when the antacids had not initially helped, she had continued to take them in the hope of symptomatic relief. She denied using the medications for any other reason and realised that the dose she was using was excessive but felt that they were safe during pregnancy.

Final diagnosis:

• Severe hypokalaemia secondary to an overdose of over-the-counter antacids.

Treatment

Initially the causative agents (antacids) were stopped and the patient was informed regarding the importance of managing reflux using alternative medications, in this case we used omeprazole. Then to normalise the serum magnesium levels, 1 L of 20 mmol magnesium was given followed by 2 L of 0.9% normal saline with 40 mmol potassium chloride. Hypomagnesaemia is present in more than 50% of cases of severe hypokalaemia and is thought to contribute to the development of hypokalaemia by promoting potassium loss via impairment of the sodium–potassium ATPase pump.3 It is therefore critical to correct a magnesium deficiency initially, otherwise replacement of potassium to normal levels will be challenging.3 Oral replacement of phosphate was also given (two tablets of SandoPhos two times per day). After 48 hours’ admission her electrolyte levels were normalised and the patient continued to feel well. After discussion with the endocrinology registrar, discharge was felt to be safe on oral replacement (2 tablets of Sando-K 2 times per day) and the patient was switched to ranitidine.

Outcome and follow-up

Potassium levels took 2 days to stabilise as an inpatient, and normalisation of her pH and base excess followed. Her tightenings decreased in frequency. She was subsequently discharged on oral Sando-K (2 tablets 2 times per day) with a plan to check her potassium level 3 days later.

However, the patient presented 3 days later at 35 weeks and 4 days’ gestation with increasingly painful uterine tightenings. On vaginal examination she was found to be 5 cm dilated and preterm labour was diagnosed. Cardiotocography confirmed fetal well-being, and intrapartum antibiotics were given in view of her earlier group B Streptococcus positive vaginal swab. Less than an hour later, a live male infant was delivered by spontaneous vaginal delivery in good condition. Syntometrine was given and the placenta and membranes were delivered by controlled cord traction. Weighed blood loss was 200 mL. Neonatal cord gases results were venous pH 7.396 (BE −1.8) and arterial pH 7.209 (BE −2.7). Maternal blood tests demonstrated a potassium level of 3.4 mmol/L, phosphate 0.9 mmol/L and magnesium 0.72 mmol/L. Due to neonatal weight loss, the patient remained in hospital for 5 days post-delivery but had a routine discharge with both mother and baby doing well.