Ionised hypocalcaemia in emergency and acute medicine

Abstract

Acute hypocalcaemia can be life-threatening and must be diagnosed promptly. The gold-standard investigation is ionised calcium, which is measured on most blood gas analysers. Total calcium measurements are inaccurate in severe depletion even if ‘corrected’ or ‘adjusted’ for albumin. We present an illustrative case of a woman in her 30s with symptomatic hypocalcaemia and a very low ionised calcium on VBG analysis. Emergency calcium replacement was delayed due to a falsely reassuring corrected calcium result. Our discussion includes a systematic literature review on the use of ionised calcium in emergency and acute medical settings. We suggest cognitive biases that may explain clinical over-reliance on corrected calcium, and call for the inclusion of ionised calcium values in major treatment guidelines for acute hypocalcaemia.

Background

Acute hypocalcaemia can be life-threatening. It may cause seizures, hypotension and arrhythmia. In adults, hypocalcaemia is most commonly encountered in surgical settings (eg, post-parathyroidectomy) but is also seen in vitamin D deficiency, hypomagnesaemia, pancreatitis and alkalosis.¹

The classically taught symptoms of acute hypocalcaemia include numbness and paraesthesia around the mouth and at the extremities. Clinicians may elicit Trousseau’s and Chvostek’s signs. Regardless, many patients—even with severe calcium depletion—report no symptoms and examine normally.² The ECG may show QT prolongation but this finding is non-specific. It is for these reasons that biochemical parameters are relied on to diagnose hypocalcaemia.

There are two measurements of calcium available in practise: (1) total calcium and (2) ‘ionised’ calcium, with the latter considered gold standard.^{3 4} Ionised calcium is the proportion of serum calcium that is unbound to circulating proteins. It reflects calcium that is biologically active—for example, in cardiac action potentials. Total calcium differs from ionised calcium as it additionally measures inert, bound calcium. Formulae for ‘corrected’ or ‘adjusted’ calcium are used to ‘fine-tune’ total calcium values after measuring serum proteins. These formulae are unreliable in severe calcium derangement.^3–5

In the past, measuring ionised calcium was logistically challenging. Samples had to be stored on ice or rapidly transported to the laboratory to minimise cell metabolism. Total calcium was a more convenient test. Modern blood gas machines now provide a rapid assessment of ionised calcium.^3–5 However, this test is frequently under-valued in acute settings⁶ and is not included in major adult treatment guidelines (eg,¹). Overall, the usefulness of ionised calcium is appreciated more in the literature than it is in clinical practice.

We present a case of severe hypocalcaemia in which treatment was delayed by a misleading corrected calcium level and a failure to recognise ionised hypocalcaemia on venous blood gas (VBG) analysis.

Case presentation

A woman in her 30s presented to our urgent treatment centre with a 2-day history of vomiting and loose, non-bloody stools. She also reported a ‘tingling’ sensation in her lips with some intermittent cramping of the wrist. She had eaten rice and chicken in a new restaurant the evening before her symptoms began. She denied use of alcohol or recreational drugs.

Medical history included Ewing’s sarcoma treated 10 years prior with chemotherapy and surgery. She was on no regular medications.

On arrival to hospital, the patient was still vomiting and could not tolerate oral fluids. She had the following observations: temperature 36.4°C, heart rate 63 bpm, blood pressure 128/86 mm Hg, respiration rate 17 bpm and SpO2 100% on room air.

The patient was diverted to the emergency department and assessed under acute medicine. On examination, she had a soft and non-tender abdomen with normal bowel sounds. She had dry mucous membranes. Cardiorespiratory examination was unremarkable. She was breathing comfortably.

The documented impression was ‘gastroenteritis with dehydration and possible clinical symptoms of hypocalcaemia or other electrolyte imbalance’.

Investigations

A urine dip was positive for ketones and protein but negative for blood, nitrites and beta-hCG. A nose and throat swab was negative for SARS-CoV-2 RNA.

Blood tests were ordered, including a VBG. The results of the VBG are shown below, with normal ranges indicated in parentheses:

pH=7.67 (7.32–7.43)

Na+=134 (136–145)

K+=7 (3.5–5.1)

Cl−=96 (98–107)

Ca2+=0.98 (1.16–1.33)

Glucose=6 (3.6–5.3)

Lactate=2.6 (0.3–2)

BE=13.2 (−2 to +2)

HCO3=35.1 (22–29)

The raised lactate and potassium were noted, but the ionised hypocalcaemia was not. On the basis of the raised potassium, an ECG was taken. The trace (figure 1) was of poor quality but demonstrated normal sinus rhythm with a severely prolonged QT interval: 551 ms as corrected with the Bazett formula. An ECG taken 2 years prior had shown only a corrected QT interval of 366 ms.

Figure 1.

ECG on admission showing severely prolonged Q-T interval.

Antihyperkalaemic medications were prepared but prior to their administration, the full blood test results came back. These full blood results are shown below, with normal ranges indicated in parentheses:

*Total calcium=2.67 (2.2–2.6)

*Corrected calcium=2.52 (2.2–2.6)

Phosphate=0.89 (0.8–1.5)

C-reactive protetin=2 (0–5)

White cell count=11.3 (4–11)

Haemaglobin=158 (>115)

Platelets=390 (150–400)

Neutrophils=9.69 (1.8–7.5)

Lymphocytes=0.76 (1–4)

*Total protein=86 (60–80)

*Albumin=49 (34–48)

Mg2+=0.8 (0.7–1)

Na+=140 (133–146)

* K+=3.5 (3.5–5.3)

Urea=5.4 (2.5–7.8)

Creatinine=72 (45–84)

eGFR >90

These results were documented in the medical notes as follows: ‘normal bloods reviewed—electrolytes normal, pseudohyperkalaemia on VBG’.

Treatment

The patient was admitted to an acute medical ward, where she was given intravenous fluids and antiemetics. The next day, she was included in the consultant-led ward round. Hypocalcaemia was again considered on the basis of symptoms. Trousseau’s sign was elicited (video 1). Repeat blood tests were taken without a tourniquet. Uncuffed adjusted calcium was normal (2.39), as was her vitamin D level. Unexpectedly, parathyroid hormone was normal (2.9) at this time.

Video 1. Trosseau sign elicited on first day of hospital admission.

bcr-2022-251611supp001.mp4 ^{(668.9KB, mp4)}

DOI: 10.1136/bcr-2022-251611.video01

On-call biochemistry was consulted at this point. Advice was given to check the patient’s ionised calcium level using a blood gas analyser. The VBG taken in the emergency department was reviewed and the low ionised hypocalcaemia was noted. A repeat VBG showed calcium 1.11 with a pH of 7.40. Intravenous calcium gluconate was given immediately with ECG monitoring. A slow infusion of calcium was then administered over the next 10 hours. Oral calcium supplements were prescribed.

Outcome and follow-up

Over the next few days, the patient’s gastrointestinal symptoms resolved but her paraesthesia continued. Chvostek’s sign was elicited. She was seen by gastroenterology, who suggested that her symptoms were likely attributable to bacterial gastroenteritis, with bacillus cereus suggested as an organism.

The patient was seen by endocrinology on her sixth day of hospital admission. By this time her ionised calcium had improved to 1.19 and her symptoms had resolved. Her corrected calcium was 2.35. The blood gas machine used by the hospital (GEM Premier 500) provides results normalised to a pH of 7.4 and therefore it was suggested by endocrinology that the severity of the ionised hypocalcaemia may be attributable to the patient’s metabolic alkalosis on admission. She was discharged home on oral calcium supplements.

The patient was seen in the ambulatory care unit 2 weeks later. At this time, she was feeling well and was asymptomatic. Blood tests showed her corrected calcium to be 2.46 and her ionised calcium to be 1.24. She received no further follow-up.

Discussion

Hypocalcaemia is a potentially life-threatening electrolyte abnormality. Its gold-standard test—ionised calcium—is available on standard hospital blood gas analysers. However, ionised calcium is often under-valued by emergency and acute physicians. Total calcium measurements may provide false reassurance and delay treatment, even in symptomatic patients. The case that we have shared demonstrates a risk posed to patients by neglect of ionised calcium.

To contextualise our discussion of this case, we conducted a systematic review of MEDLINE and EMBASE databases (see online supplemental material 1). Our interest was in studies using ionised calcium in the treatment or prognosis of patients in the emergency department or on acute medical wards. In total, 2585 records were screened and eight relevant articles were identified.^6–13 Major trauma studies were excluded as comprehensive reviews have been published on ionised calcium in these settings.^{14 15}

On admission, the patient had an ionised calcium of 0.98 and was symptomatic. A key consideration in hypocalcaemia should be QT interval, of which hers was severely prolonged. We believe that this abnormality was attributable to hypocalcaemia: she was not on any regular medications, and electrolytes such as potassium and magnesium remained normal throughout admission. To what extent was she at risk of arrythmia? This question was investigated in a recent study by Thongprayoon et al.⁷ In 3353 emergency department patients with low ionised calcium (≤1.1) ventricular arrhythmia was observed in 164 (4.9%). This finding highlights the importance of requesting an urgent ECG when blood gas analysis indicates ionised hypocalcaemia. Indeed, ionised calcium has been shown to be superior to total calcium in predicting ECG changes.¹¹

More broadly, patients who present with this degree of ionised hypocalcaemia appear to be at high risk of mortality. Vroonhof et al⁹ analysed data from 1806 emergency department patients and found a 7-day mortality rate of 15.5% in those with an ionised calcium below 1.12. Thongprayoon et al⁷ demonstrated in-hospital and 1-year mortality rates of 8% and 24%, respectively, in addition to high rates of organ failure. Similar findings have been noted throughout the COVID-19 pandemic: ionised hypocalcaemia has been a common abnormality in swab-positive patients¹² with predictive value for hospital admission¹³ and potentially death.¹⁰

Despite abnormal blood gas results, in our case, calcium replacement was only provided on the second day of hospitalisation. This may not be an uncommon occurrence. Kalidindi et al⁶ found that only half of patients with ionised hypocalcaemia on their admission blood gas underwent treatment or repeat testing. In a subsample of patients described by Çap et al,¹¹ 19 had ionised hypocalcaemia and a prolonged QT interval. Yet only four of these patients received calcium replacement. Nine had a normal corrected calcium level on admission, which may—as in our case—have given false reassurance to clinicians.

What might be causing clinicians to under-use the blood gas calcium?⁶ We suggest two cognitive biases. The first is a straightforward authority bias: omission of ionised calcium from expert guidelines¹ may have led clinicians to regard the test as non-valued by experts. The second bias relates to a well-documented phenomenon in cognitive psychology called ‘pricing bias’. Essentially: humans under-value items that have a low price, or are very easy to obtain.¹⁶ This pricing bias may affect clinical decisions about blood tests. Total calcium may be over-valued as a ‘premium product’ because it requires sending a sample to the laboratory and waiting for a report. Ionised calcium, available within minutes at the point of care, may feel too ‘cheap’ to be a gold standard investigation. The same bias may affect the interpretation of other electrolytes measured by VBG analysis. A 2008 survey of 64 UK doctors found that just over half would wait for laboratory confirmation of abnormal blood gas potassium before initiating treatment.¹⁷

Although the outcome of our case was positive overall, we believe that it signals a need for a revision of guidelines on the treatment of acute hypocalcaemia. Updated guidelines should include an explanation to clinicians of the significance of ionised calcium on blood gas analysis and the unreliability of total calcium measurements. Ideally, they should also include treatment thresholds based on ionised calcium values^3–5 although our review has shown that there is currently limited evidence to support such recommendations. This would be a fruitful area for future research. We suggest an ionised calcium level of 1.15 as a starting point for investigation: below this level, a sharp increase in rates of ventricular arrythmia⁷ and mortality⁹ has been observed.

Learning points.

• The ionised calcium value on blood gas analysis is a gold-standard test for hypocalcaemia.

• Total calcium values, even if ‘adjusted’ or ‘corrected,’ can over-estimate serum calcium and provide false reassurance.

• Cognitive biases such as ‘pricing bias’ may cause clinicians to mistakenly over-value tests that are difficult to perform, and under-value tests that are easy to perform.

Ethics statements

Patient consent for publication

Consent obtained directly from patient(s).