Hypokalaemia in pregnancy – Prevalence, underlying causes, and an approach to investigation

Abstract

Objective

To investigate the prevalence and aetiology of hypokalaemia in pregnancy.

Methods

This was a five-year retrospective audit of women who received care at a tertiary hospital, who developed hypokalaemia during pregnancy and within three weeks postpartum. Serum potassium and magnesium levels cause(s) of hospitalization and investigations for hypokalaemia were obtained from hospital records.

Results

One hundred and ten women developed hypokalaemia during pregnancy and the immediate postpartum period, representing 0.36% of total births. Ninety-one per cent of patients had mild to moderate hypokalaemia (K 2.6–3.1 mmol/L), while 9% had severe hypokalaemia (K < 2.6 mmol/L). The most common associations of hypokalaemia were infection (38%), vomiting (18%), hypertensive disorders (12%) and postpartum haemorrhage (9%). Twenty-four patients did not have a clear underlying aetiology of hypokalaemia, but only five had further investigations.

Conclusions

There was inadequate investigation and follow-up of hypokalaemia, particularly in women in whom an obvious cause was not apparent.

Introduction

There are limited studies regarding the prevalence and aetiology of hypokalaemia in pregnancy. The most common cause of hypokalaemia in pregnancy described in case reports is hyperemesis gravidarum.^1–3 While subject to reporting and publication biases, other commonly reported causes are Gitelman syndrome, primary hyperaldosteronism and renal tubular acidosis.^4–10 Urinary tract and viral infections have been associated with hypokalaemia in non-pregnant individuals.^11–15 Severe hypokalaemia in pregnancy may result in complications including arrhythmias, cardiac arrest, paralysis and rhabdomyolysis.^1–3,16 The risk of paralysis and rhabdomyolysis with hypokalaemia in pregnancy has been reported to be greater than in non-pregnant individuals.^2,3,10,11

There is a physiological fall in serum potassium (K) by 0.2–0.3 mmol/L during pregnancy.^17–19 Hypothesized mechanisms include dilution, respiratory alkalosis leading to intracellular potassium shift, increased gastrointestinal loss and transplacental passage of potassium. ²⁰ There are no clearly defined reference intervals for urinary potassium in pregnancy. In the non-pregnant population, renal potassium wasting in the setting of hypokalaemia is defined as 24-hour urinary potassium above 15–20 mmol/day, urine potassium to creatinine ratio above 1.5–2.5 mmol/mmol or spot urine potassium above 15–20 mmol/L.^21,22 The impact of pregnancy on renal potassium handling is unclear, with some studies reporting reduced urinary potassium excretion in pregnancy, while others reported increased urinary potassium excretion.^23–25

Investigation of hypokalaemia includes assessment of magnesium (Mg), acid-base status, urinary potassium, and serum aldosterone and renin, all of which are subject to physiological changes in pregnancy. ¹⁷ Urinary magnesium excretion increases by ∼ 25% in pregnancy, and plasma magnesium in the third trimester is 30% lower than preconception. ²⁶ Serum bicarbonate is also lower in pregnancy due to metabolic compensation for respiratory alkalosis, with pregnancy reference ranges of 20–26 mmol/L.^18,20

Aldosterone and renin are both upregulated in pregnancy to maintain the increased circulating blood volume and placental perfusion.^7,27 Plasma renin activity rises approximately 4-fold by eight weeks' gestation and almost 7-fold at term, while plasma aldosterone rises 3–8 fold by third trimester.^17,28 There is an earlier and proportionately greater rise in renin, increasing false negatives when the aldosterone to renin ratio is used in screening for primary hyperaldosteronism in pregnancy.^8,28 Both plasma cortisol and 24-hour urinary free cortisol rise progressively in pregnancy, peaking during the third trimester, when reference intervals are 3.1-fold higher compared to pre-pregnancy. ²⁹

The aim of this audit was to determine the prevalence and aetiology of hypokalaemia in pregnancy, as well as the adequacy of investigation of unexplained hypokalaemia at our institution.

Methods

This was a retrospective study of women with hypokalaemia (defined as K < 3.2 mmol/L) during pregnancy or within three weeks postpartum, who received care at Mater Mother's Hospital, a tertiary referral maternity centre in Brisbane, Australia between January 2017 and December 2021. Data was also collected for all women who had severe hypokalaemia (K < 2.6 mmol/L) who were admitted at Mater Hospital from 2012 to 2021. Women were identified via the Mater Hospital chemical pathology database, followed by review of electronic health records regarding maternal age, parity, country of birth, gestation of detection of hypokalaemia, gestation and mode of delivery, birthweight, and resolution of maternal hypokalaemia postpartum. Known associations of hypokalaemia were sought including vomiting, infection, hypertensive disorders (including pre-existing hypertension, gestational hypertension and pre-eclampsia) and postpartum haemorrhage (PPH, defined as >1 L blood loss ³⁰ ). All women with hypokalaemia had a chart review of their medication and diet history taken on admission, including complementary and over-the-counter medications, herbal therapies including skin lightening creams, as well as intake of caffeinated beverages and carbonated soft drinks.

Women were classified as having mild to moderate hypokalaemia (K 2.6–3.1 mmol/L) and severe hypokalaemia (K < 2.6 mmol/L). The lowest potassium level for each patient was used for analysis. Hypomagnesemia was defined as Mg ≤ 0.5 mmol/L. The spot urinary potassium, 24-hour urinary cortisol and cortisone, and plasma aldosterone and renin concentration collected most proximate to the diagnosis of hypokalaemia were recorded.

Spurious hypokalaemia can occur due to pre-analytical factors such as delayed transport of blood samples which are exposed to high ambient temperatures, which stimulates cellular potassium uptake. ³¹ Our samples were all collected at the hospital and therefore there was minimal delay in transport to the laboratory.

Ethical clearance was obtained from the institution's human research ethics committee. Data was analysed using SPSS (IBM SPSS Statistics Version 29).

Results

In the 5-year period between January 2017 and December 2021, there were 110 women with hypokalaemia during pregnancy or in the immediate postpartum period at Mater Mothers Hospital. This represented 0.36% of total births. The median age of the patients was 31 years, with 41% being nulliparous. Mean gestational age at birth was 37 weeks (median 38 weeks, range 22–41 weeks), with 24% of patients delivering preterm (prior to 37 weeks). Mean birth weight was 2.9 kg (SD 0.8). Mode of delivery was 49% vaginal delivery and 51% caesarean section. Seven patients had pregnancy losses, of which two were in the first trimester and five in the second trimester.

With regards to the continent of birth of patients with hypokalaemia, 47% were from Australia, 27% Asia, 20% Africa, 3% Europe, 1% Americas and 2% other Oceanian countries. In terms of all pregnancies between 2017 and 2021 at Mater Hospital, 46% of women were born in Australia, 30% Asia, 8% other Oceanian countries, 7% Africa, 6% Europe and 3% Americas.

Hypokalaemia occurred in the first trimester in 10% of patients, second trimester in 20%, third trimester in 41% and in the immediate postpartum period in 29%. Ninety-one per cent of patients had mild to moderate hypokalaemia, while 9% had severe hypokalaemia. Only two patients had complications – one patient had cardiac arrest due to ventricular fibrillation and one had QTc prolongation. Ten percent of patients had concurrent hypomagnesemia. Hypokalaemia resolved postpartum in 92 patients, did not resolve in three patients and further blood tests were not done postpartum to assess for resolution in the remaining 15 patients.

Aetiology/associations of hypokalaemia

The most common associations of hypokalaemia were infection, vomiting, hypertensive disorders and PPH (Table 1). There were single cases of hypokalaemia associated with pancreatitis, hyperthyroidism, Gitelman syndrome, post-operative ileus, diabetic ketoacidosis, and renal tubular acidosis. Seven patients were taking medications that predisposed to hypokalaemia, including corticosteroids, adrenaline, omeprazole, intravenous insulin infusion and quetiapine. No cause was identified in 24 patients (22%). Eleven patients had more than one association.

Table 1.

Aetiology and associations of hypokalaemia in pregnancy.

Association	Number of patients (%)
Infection	42 (38%)
Vomiting	20 (18%)
Hypertensive disorder (pre-eclampsia, gestational hypertension and chronic hypertension)	13 (12%)
Postpartum haemorrhage	10 (9%)
Medications a	7 (6%)
Other causes b	6 (5%)
No cause identified	24 (22%)

^a Medications associated with hypokalaemia included corticosteroids, adrenaline, omeprazole, insulin infusion and quetiapine.

^b Other cause included pancreatitis, hyperthyroidism, Gitelman syndrome, post-operative ileus, diabetic ketoacidosis, and renal tubular acidosis.

Investigation for underlying causes of hypokalaemia

Only 17 out of 110 patients had further investigations to determine the aetiology of hypokalaemia (Table 2). For the 24 patients with no underlying cause, only 5 (21%) had further investigations.

Table 2.

Patients who underwent investigations regarding aetiology of hypokalaemia.

Patient	Age	Continent of birth	Trimester	Lowest K mmol/L	Aetiology	Bicarb mmol/L	Aldosterone pmol/L	Renin mU/L	Spot urinary K mmol/L	Urinary cortisol (nmol/24 h)	Urinary cortisol/ cortisone	Resolution postpartum
1	37	Africa	PP	2.2	PPH	24	3	<1	9	138	NA	Unknown
2	33	Asia	3	2.9	Unknown	24	10	4	25	NA	NA	Yes
3	19	Australia	PP	3.0	Pre-eclampsia	21	14	35	NA	NA	NA	Yes
4	42	Australia	PP	2.2	Infection	25	4	7	11	NA	NA	Yes
5	30	Australia	2	2.9	Vomiting	26	67	28	15	77	0.35	Yes
6	34	Africa	3	2.9	Pre-eclampsia	22	59	5	19	a cortisol/Cr 6 µmol/mol	a 0.22	Yes
7	25	Africa	3	3.1	Unknown	22	81	382	25	NA	NA	Yes
8	35	Australia	3	2.2	Unknown	36	1590	30	20	79	0.14	Yes
9	26	Africa	1	2.2	Hyperemesis	31	NA	NA	9	NA	NA	Yes
10	34	Australia	1	2.2	Hyperemesis, Gitelman	34	3890	108	63	NA	NA	No
11	36	Africa	3	3.1	Pre-eclampsia	24	99	13	NA	NA	NA	Yes
12	21	Australia	1	2.4	Hyperemesis	27	168	37	NA	NA	NA	Yes
13	27	Australia	PP	3.1	Infection	26	1520	58	NA	NA	NA	Yes
14	33	Africa	PP	2.5	Unknown	24	9	6	3	37	0.12	Yes
15	25	Australia	PP	2.7	RTA	18	NA	NA	28	NA	NA	Unknown
16	25	Asia	3	2.9	Unknown	20	29	28	24	277	0.36	Yes
17	30	Africa	2	3.1	Hypertension, vomiting	17	365	9	NA	NA	NA	Yes

PPH: postpartum haemorrhage; RTA: renal tubular acidosis; NA: not available; PP: postpartum.

^a Only cortisol/ortisone available (24-hour urine not performed).

Only 12 patients had assessment of urinary potassium, and this was delayed by a mean of six days (range 1–20 days) after the diagnosis of hypokalaemia. Therefore, most patients would have received potassium supplementation prior to the collection of urinary potassium, which may have impacted the results.

In the 15 patients who had plasma aldosterone and renin concentration assessed, nine had hyporeninaemic hypoaldosteronism, in whom the associations of hypokalaemia included PPH (one patient), pre-eclampsia (three patients), infection (one patient), vomiting (one patient) and unknown (three patients). One of the three women with pre-eclampsia also had chronic hypertension. For the seven patients with hyporeninaemic hypoaldosteronism who had urinary potassium assessed, five had spot urinary potassium <20 mmol/L. In the two patients who had aldosterone and renin reassessed after the resolution of hypokalaemia, this also normalized, suggesting that hyporeninaemic hypoaldosteronism is a transient process. The three patients with pre-eclampsia had aldosterone and renin assessment while receiving potentially interfering antihypertensive medications including methyldopa, nifedipine or enalapril. ³²

One patient (patient 17) had aldosterone/renin ratio > 40 and renin <20 mU/L with associated chronic hypertension, suggestive of primary hyperaldosteronism. ⁸ However, she was taking methyldopa, which can lower renin and cause a false positive aldosterone to renin ratio.

One normotensive patient (patient 10) had hyperreninaemic hyperaldosteronism with metabolic alkalosis and high urinary potassium (63 mmol/L) with low urinary calcium (< 0.25 mmol/L), consistent with Gitelman syndrome, as urine calcium usually rises 2–3-fold during pregnancy. ²⁸ This patient had hypokalaemia for several years prior to pregnancy, which worsened during pregnancy in the context of hyperemesis gravidarum. Genetic testing for Gitelman syndrome was discussed with the patient but cost was prohibitive.

Severe hypokalaemia cases

Over a 10-year period from 2012 to 2021, 10 pregnant women had severe hypokalaemia, with two cases complicated by rhabdomyolysis. With regards to continent of birth, five of these women were born in Australia, four in Africa and one in Asia. Six women had aetiologies identified for severe hypokalaemia: three with hyperemesis gravidarum, and one case each of Gitelman syndrome, post-caesarean section ileus, and sepsis. Of the four women with no cause identified for severe hypokalaemia, three were African women who had hyporeninaemic hypoaldosteronism and low urinary potassium.

Discussion

Two previous studies have examined the prevalence of hypokalaemia in pregnancy, both defining hypokalaemia as serum K of < 3.5 mmol/L. A United States nationwide population study of 12,431,909 pregnant patients found a prevalence of hypokalaemia of 0.69% in pregnancy-related hospitalisations. ³³ This study found that risk factors for hypokalaemia included younger age, African American origin, lower income level and higher comorbidity index. Gestational hypertension, hyperemesis gravidarum and postpartum haemorrhage were all found to be independently associated with a higher rate of hypokalaemia.

Businge and Mdaka ³⁴ found a prevalence of hypokalaemia of 5.3% in pregnant women attending a primary healthcare clinic in rural Eastern Cape South Africa. Forty-five per cent of women in this study were practicing geophagia, a common dietary practice in Africa. Risk factors for hypokalaemia included age less than 25 years, primigravida, geophagia, diet deficient in meat, vegetables and fruit, and intake of carbonated beverages.

Our study showed that hypokalaemia was most common in the third trimester and the immediate postpartum period. The most common associations were infection, vomiting, hypertensive disorders, postpartum haemorrhage and medications. No cause was found in 22% of patients.

Infection was the most common association of hypokalaemia in this study, with potential mechanisms including renal tubular dysfunction, losses from vomiting and diarrhoea as well as antibiotic side effects. Urinary tract infection (UTI), adenovirus, mycoplasma and COVID-19 infections have been associated with hypokalaemia in non-pregnant populations.^11–15 A case control study with 43,719 UTI and control patients found a much higher percentage of hypokalaemia in patients with UTI (10.4%) compared to controls (4.2%). ¹² This was thought to be secondary to UTI causing impaired renal tubular potassium reabsorption. Hypokalaemia also occurs in 15.1%–30.7% of COVID-19 cases, with hypokalaemia being a predictor of severe COVID-19 requiring invasive mechanical ventilation.^13,14

Antibiotics are also associated with hypokalaemia, potentially due to their impact on renal tubular function. In patients with normal potassium at baseline who were treated with piperacillin/tazobactam, 13.9% developed hypokalaemia, with higher risks in patients who had longer treatment duration or higher daily dose. ³⁵ Falcone et al. ³⁶ found that in patients with bone and joint infections treated with antibiotics, hypokalaemia increased from 5% at admission to 11% by day 14. In particular, vancomycin, rifampicin and ciprofloxacin were associated with hypokalaemia.

An interesting finding of this study was that in the patients who had assessment of aldosterone and renin, 60% had hyporeninaemic hypoaldosteronism. It is unusual for hyporeninaemic hypoaldosteronism to occur in pregnancy, as the renin-angiotensin-aldosterone system (RAAS) system is usually upregulated in pregnancy, due to the co-existing uteroplacental RAAS. This increased activation of RAAS is protective in pregnancy and important in normal placentation. Expression of renin, pro-renin, angiotensinogen, angiotensin converting enzyme and angiotensin II have been found in maternal decidua and placental tissues.^37,38 The high oestrogen state of pregnancy also stimulates hepatic synthesis of angiotensinogen.^27,28,38 Despite increased RAAS activity, women remain normotensive during normal pregnancies, as increasing progesterone has an antagonist effect on aldosterone receptors. ²⁸ In disease states such as pre-eclampsia, there is reduced RAAS activity. ³⁸ Three out of nine patients with hyporeninaemic hypoaldosteronism had co-existing pre-eclampsia.

Hyporeninaemic hypoaldosteronism can occur in normal physiologic circumstances in the context of increased blood volume, with vigorous intravenous sodium chloride, during magnesium infusions, and in the setting of type IV renal tubular acidosis. The former two settings are associated with normokalaemia, and the latter two with hyperkalaemia.^32,39,40 Hypokalaemia alone would not explain hyporeninaemic hypoaldosteronism. Although hypokalaemia would directly inhibit aldosterone release from the adrenal cortex, renin is increased by potassium deficiency.^41,42 Of the patients with hyporeninaemic hypoaldosteronism, the only patient who received significant intravenous sodium chloride was the woman with streptococcal septicaemia.

Hypokalaemia with hyporeninaemic hypoaldosteronism can also be seen in some rare conditions such as Liddle syndrome, syndrome of apparent mineralocorticoid excess (AME), Cushing syndrome, Geller syndrome and 11-deoxycorticosterone producing adrenal tumours. ⁴³ All of these would present with sustained hypertension, hypokalaemia and metabolic alkalosis.

AME syndromes occur due to genetic deficiency in 11-beta-hydroxysteroid dehydrogenase 2 (11bHSD2), or ingestion of liquorice, complementary or herbal therapies which inhibit 11bHSD2, resulting in elevated cortisol which acts on mineralocorticoid receptors. ⁴⁴ All of our patients denied intake of causative substances and none of our patients had an elevated urine cortisol: cortisone ratio (ratio > 1.0) to suggest this condition. ⁴⁵

Liddle syndrome is a rare autosomal dominant form of early onset hypertension due to activating mutations of the ENaC of the distal nephron, resulting in an increase in collecting tubule sodium reabsorption and potassium wasting. ⁴⁶ Geller syndrome is caused by gain-of-function mutations in the mineralocorticoid receptor, with progesterone (normally mineralocorticoid receptor antagonist) becoming a potent agonist. As plasma progesterone normally increases 100-fold in pregnancy, Geller syndrome can present with hypertension which worsens in pregnancy and improves postpartum. ⁴⁷ It is possible that patient 11, who had with chronic hypertension which worsened in pregnancy with development of pre-eclampsia, who also had hyporeninaemic hypoaldosteronism, may have had undiagnosed Geller syndrome.

There was a disproportionate number of African women with hypokalaemia, as women born in Africa accounted for 20% of patients with hypokalaemia in pregnancy, while only accounting for 7% of all pregnancies. Studies comparing serum potassium between non-pregnant individuals of African and other ancestry have been inconsistent, showing either no significant difference, or values a mean of 0.162 mmol/L lower in those of African American compared with non-African American ancestry.^48–50 African Americans consume less dietary potassium than non-African Americans. ⁵¹ Urine potassium loss is lower in those of African ancestry than Caucasians both at baseline, and in response to amiloride therapy and an intravenous potassium load.^50,52,53 The disproportionate number of African women with hypokalaemia in pregnancy is possibly explained by the practice of geophagia in Africa. Geophagia is a form of pica that involves the consumption of earth, soil or clay. Geophagia has been observed in up to 84% of pregnant women in African countries. ⁵⁴ Qualitative studies in the UK, Netherlands, Austria and Belgium conclude that geophagia is a cultural phenomenon embedded in indigenous knowledge, and widely practiced amongst women of African ethnicity who migrate to other countries.^55–59 Clay binds to potassium in the gastrointestinal tract, leading to increased gastrointestinal potassium excretion.^60,61 Case reports of geophagia in pregnancy have been associated with severe hypokalaemia and resultant myopathy.^60,61

In an animal study, clay beverage consumption inhibited the normal rise in renin and aldosterone in pregnancy, which may explain hyporeninaemic hypoaldosteronism in geophagia. ⁶² A case report of a non-pregnant patient with geophagia and severe hypokalaemia had hyporeninaemic hypoaldosteronism and low urinary potassium. ⁶³ This potentially explains the number of African women with unexplained severe hypokalaemia, hyporeninaemic hypoaldosteronism and low urinary potassium in our study. Measurement of faecal K loss may be a helpful confirmatory investigation in women who deny the practice. ⁶⁴

Other dietary causes of hypokalaemia include caffeine and carbonated beverages, which causes hypokalaemia through glucose induced osmotic diuresis and caffeine induced potassium redistribution into cells and renal potassium wasting. ⁶⁵

A major finding of this study was the poor quality of investigation and postpartum follow-up of hypokalaemia, particularly in women in whom an obvious cause was not apparent at the time of presentation. Assessment of urine potassium losses is the most important initial test in guiding further evaluation regarding the aetiology of hypokalaemia. The significant delay in measurement of urine potassium from the time of presentation with hypokalaemia may render the urine level unhelpful. Education of health professionals with a diagnostic flowchart may be useful in enabling prompt appropriate investigations (Figure 1 and Table 3). Improved communication and liaison with General Practitioners regarding postpartum follow-up of women with hypokalaemia in pregnancy to ensure resolution is required, perhaps at the 6-week postnatal check. In women where hypokalaemia persists, further investigation in the absence of the interference of physiological changes of pregnancy on laboratory results may increase diagnostic yield.

Figure 1.

Flowchart for the investigation of hypokalaemia in pregnancy.

Table 3.

Investigations for hypokalaemia in pregnancy.

Investigation	Interpretation Lack of well-established reference ranges in normal pregnancy may make interpretation difficult
Urine tests
Spot urine potassium	> 20 mmol/L suggests renal potassium wasting
Urine potassium/creatinine ratio	> 1.5–2.5 mmol/mmol suggests renal potassium wasting
24-hour urine potassium	> 15–20 mmol/day suggests renal potassium wasting
Urine calcium	Low in Gitelman syndrome and thiazides High in Bartter syndrome and loop diuretics
Urine chloride	Low urine chloride <20 mmol/L in vomiting High urine chloride with loop and thiazide diuretics and Gitelman and Bartter syndromes
24-hour urine cortisol	Screening test for Cushing syndrome Cortisol progressively rises in pregnancy, peaking in third trimester, when reference ranges are 3.1-fold higher than pre-pregnancy
Urine cortisol/cortisone ratio	Ratio > 1.0 suggests apparent mineralocorticoid excess or liquorice ingestion
Blood tests
Serum potassium	Hypokalaemia in pregnancy defined as K < 3.2 mmol/L
Serum magnesium	Hypomagnesemia in pregnancy defined as Mg ≤ 0.5 mmol/L
Serum bicarbonate	Pregnancy reference ranges 20–26 mmol/L
Aldosterone: renin ratio	Both increase in pregnancy, with greater increase in renin. Aldosterone: renin ratio > 40 with mass renin concentration < 20 mU/L in pregnancy suggests primary hyperaldosteronism
Imaging
Renal artery Doppler ultrasound	To exclude renal artery stenosis
MRI adrenal (CT if postpartum)	To exclude adrenal lesion if biochemically suspicious for adrenal Cushing syndrome or primary hyperaldosteronism

MRI: magnetic resonance imaging; CT: computed tomography.

Limitations

A major limitation was that this study was retrospective and may not have identified all women with hypokalaemia. There was no control group in this study. As Mater Mother's Hospital is a major tertiary referral hospital, there could potentially be referral bias with a higher prevalence of hypokalaemia in pregnancy in our cohort than the general population.

We used random spot urinary potassium in the assessment of renal potassium wasting, but this could be inaccurate in the presence of polyuria, where there could be significant renal potassium wasting even in the setting of low spot urinary potassium. ²² Ideally, urine potassium to creatinine ratio or 24-hour urinary potassium would have been preferred, but these were not routinely available due to the retrospective nature of this study.

Medication and diet history were obtained from medical notes. Ideally, all women should have been asked regarding prescribed, complementary and herbal therapies, as well as dietary history with regards to geophagia, caffeine and carbonated soft drinks, but there may be omissions due to the retrospective nature of this study.

Conclusion

Hypokalaemia in pregnancy may result in complications of arrhythmia, paralysis, and rhabdomyolysis. When the cause is not apparent, detailed history is required including complementary and herbal medications and diet history including geophagia, soft drinks and caffeine. Prompt investigations include assessment of urinary potassium loss (ideally prior to potassium supplementation), acid–base status, plasma magnesium, serum aldosterone and renin, and further assessment for rarer endocrine disorders based on clinical suspicion. A diagnostic flow chart may be useful to guide health professionals regarding the appropriate investigation of hypokalaemia in pregnancy. Trimester specific reference intervals for urine potassium losses in pregnancy would be helpful, but there is insufficient evidence to guide this at present. Postpartum follow-up to ensure resolution of hypokalaemia, and further investigation if hypokalaemia persists is important.