A case series of severe symptomatic peripartum hyponatraemia

Abstract

This single centre case series describes the presentation and management of six cases of peripartum hyponatraemia in women who were otherwise deemed low-risk at delivery. It highlights presenting symptoms such as fatigue, confusion and seizures as well as the effects on the neonate. It also focuses on areas of interest such as fluid intake, hormonal effects of ADH and oxytocin and the association with birthing pools for future research.

Introduction

Individual cases of symptomatic peripartum hyponatraemia have been reported since the 1960s,^1–3 but there remains limited understanding and a paucity of literature regarding this rare yet life-threatening and treatable disorder affecting mothers and babies. Reports in the UK have increased resulting in the publication of guidelines on the condition in Northern Ireland, ⁴ and a United Kingdom Obstetric Surveillance System (UKOSS) study. ⁵ This single centre series allows the review of case similarities and highlights possible risk factors for peripartum hyponatraemia that might provide avenues for further investigation. Nottingham University Hospitals NHS Trust is a tertiary referral centre with approximately 10,000 births per year in both consultant-led and midwifery-led birthing units. We performed a case review of severe symptomatic peripartum hyponatraemia between 2017 and 2019.

Case 1

A healthy 37-year-old woman in her first pregnancy was admitted in spontaneous labour at 39 weeks’ gestation. Urinalysis showed ketonuria (2+) with the midwife noting low volume and ‘concentrated’ urine output. She was encouraged to “hydrate to decrease ketones”. Intravenous (IV) fluid was started due to ketonuria. Cardiovascular observations were normal throughout. Two hours after delivery she became acutely confused and suffered a short tonic-clonic seizure. She was given IV magnesium sulphate 4 g for presumed eclampsia. After seizure resolution, she remained agitated and required general anaesthesia to facilitate brain computerised tomography (CT). Arterial blood gas analysis identified a markedly low serum sodium concentration, [Na+] 114 mmol/l. Brain CT was normal. Hypotonic hyponatraemia was indicated by serum osmolality 251 mosmol/kg and urine osmolality 265 mosmol/kg. She was admitted to critical care and correction of hyponatraemia was achieved after spontaneous diuresis and 12 h of fluid restriction. The baby was also hyponatraemic (serum [Na+] 120 mmol/l) and required fluid restriction and oral sodium chloride (NaCl). Retrospective calculation of fluid intake prior to the seizure estimated 3.5 l of oral fluid and 2 l of IV compound sodium lactate (CSL) over 12 h.

Case 2

A 29-year-old, woman in her first pregnancy spontaneously laboured at term. She spent 3 h in a birthing pool during her 12 h first stage. Urinalysis demonstrated (3+) ketonuria. Nausea and vomiting were treated with cyclizine. Oxytocin infusion was started at full dilation to augment the contraction frequency. The midwife noted that the woman seemed “spaced out” but that she had agreed to the management plan. She had a kiwi delivery due to cardiotocograph (CTG) concerns and staff noted she looked “exhausted”. There was a post-partum haemorrhage (PPH) of 2700 ml due to atony and perineal trauma. This was treated with intramuscular (IM) oxytocin 5 IU, Syntometrine® 5 IU/500 mg and oxytocin 40 IU in 500 ml of 0.9% sodium chloride at 125 ml/h. IM carboprost 250 mcg was given with a further dose given after 15 min. 500 ml of IV CSL was commenced. Though vital signs were normal after this, she was described as “delirious and clammy”. Maternal venous blood gas analysis revealed [Na+] 118 mmol/l. This was corrected over 3 h after a spontaneous diuresis and infusion of 1 L 0.9% NaCl on the labour suite. She was also transfused three units of packed red cells. Her neonate had a serum [Na+] 113 mmol/l and was treated with oral fluid restriction. Her partner estimated retrospectively that she had drunk 2 l of isotonic sports drinks and 4 l of water during her labour. She remarked that she drank as she felt hot and thirsty and not because of external encouragement. She was left with no memory of the birth and immediate post-partum period.

Case 3

A healthy 29-year-old woman in her first pregnancy was admitted at 39 weeks’ gestation in established labour. She entered the birthing pool after 5 h, at 7 cm cervical dilatation, and after a further 30 min both baby and placenta were delivered underwater. The woman had a PPH in the water where the mother felt faint and was syncopal on leaving the pool. The woman regained consciousness rapidly, but her husband noted that from this time onwards she was unable to talk in full sentences and had “vacant” episodes. Vital signs were unremarkable, and she received 500 ml of CSL and IM oxytocin 10 IU followed by an infusion of 40 IU oxytocin in 500 ml 0.9% NaCl at 125 ml/h. Total blood loss was estimated at 2.1 l. Four hours later [Na+] was 120 mmol/l. She was now confused and slightly tachypnoeic at 21 breaths per minute. A diagnosis of dilutional hyponatraemia was made and staff noted that she had been drinking regularly. She was treated with fluid restriction, 1 L 0.9% NaCl and furosemide 40 mg. She was then transferred to critical care and given 2.7% NaCl at 30 ml/h. Sodium corrected to 131 mmol/l within an hour and the infusion was stopped. After a further 12 h [Na+] was 137 mmol/l and all symptoms had resolved.

Case 4

A 34-year-old woman in her first pregnancy was admitted at 39 weeks’ gestation in spontaneous labour. She spent the first hour of her labour in the birthing pool and was encouraged to have isotonic drinks on account of ketonuria. A prolonged second stage and maternal exhaustion led to a forceps delivery. Electrolytes measured from umbilical cord blood at delivery showed [Na+] 110 mmol/l but this was not noted. She became more agitated and confused. A perineal tear was identified, and she gave consent for operative repair. In theatre it was clear that she could not comply with the checklist to confirm her details. A maternal venous blood gas was taken, almost 2 h following delivery, and revealed [Na+] 110 mmol/l. 150 ml 2.7% NaCl was given via a peripheral vein, and she was transferred to critical care. Surgery was postponed until the following day when [Na+] 136 mmol/l. The neonate was found to have serum [Na+] 116 mmol/l and was treated with fluid restriction and oral NaCl. In retrospect, it was discovered that the woman had drunk approximately 9 l of isotonic sports drinks during her 12 h labour.

Case 5

A healthy 34-year-old woman in her first pregnancy had an induction of labour with a prostaglandin pessary at 40 + 3 weeks’ gestation, as her fetus was large for gestational age. Artificial rupture of membranes was performed, and oxytocin augmentation was started due to delay in progress – 10 IU was prescribed in 500 ml of 0.9% NaCl and started at 3 ml/h (0.06 IU/hr). An epidural was sited for analgesia. Her cervix became fully dilated 17 h after the initiation of oxytocin, which was then running at a rate of 60 ml/h (1.2 IU/h). After a further 2 h, a forceps delivery was performed due to a pathological CTG. There was a subsequent PPH of 950 ml. This was treated with IM Syntometrine® 5 IU/500 mg, infusion of oxytocin 40 IU in 0.9% NaCl over 4 h and tranexamic acid 1 g. Electrolytes measured from umbilical cord blood at delivery showed [Na+] 119 mmol/l, although not noted at the time. One hour after delivery the woman seemed to be acting strangely which was initially thought to be an after effect of cyclizine. A venous blood sample showed [Na+] 122 mmol/l. Fluid restriction was started but the woman remained drowsy, confused and slurring her words. She was referred to critical care and given 150 ml 2.7% NaCl. The oxytocin infusion was stopped. Hypotonic hyponatraemia was indicated by serum osmolality 264 mosmol/Kg, urine osmolality 209 mosmol/Kg and urine sodium 10 mmol/l. Symptoms resolved after spontaneous diuresis and the woman was discharged back to obstetric care after 17 h when [Na+] 139 mmol/l.

Case 6

A well, 31-year-old woman in her first pregnancy presented in spontaneous labour at term. She made satisfactory progress until full cervical dilatation. Soon after this she vomited and started behaving abnormally. A forceps delivery was performed after a prolonged second stage of labour followed by IM Syntometrine® 5 IU/500 mg. Fifteen minutes after delivery the woman was unable to talk and obey commands. A venous blood sample showed [Na+] 115 mmol/l. A urinary catheter was inserted which facilitated a large spontaneous diuresis and fluid restriction was initiated. She was later started on an infusion of 1 l 0.9% NaCl at 40 ml/h. Four hours after delivery, when [Na+] 131 mmol/l, symptoms resolved. She was left with no memory of the immediate postpartum period. On retrospective questioning with her partner, it was noted that she had drunk a total of 7.75 l of fluid over 18 h of which 4.5 l had been drunk over the 9 h from admission to delivery.

Discussion

Symptomatic hyponatraemia occurs when there is a marked fall in plasma sodium concentration, particularly over a short time period. This change results in a fall in plasma osmolality leading to a net influx of water into cells, classically producing cerebral oedema. Acute hyponatraemia can cause acidosis by the reduction in strong ion difference and lead to a compensatory raised respiratory rate. ⁶ These six cases of severe symptomatic hyponatraemia all presented soon after delivery. The presenting features were agitation, confusion and drowsiness, with one case resulting in a maternal seizure. Some women experienced nausea earlier in labour which, although non-specific, may have been related to hyponatraemia. Subtle symptoms such as confusion, agitation and fatigue were assumed to be the side effects of drugs and exhaustion of labour, whilst seizures were misdiagnosed as eclampsia, leading to diagnostic delays.

Whilst the definitive cause of peripartum hyponatraemia is not known, current understanding is that oxytocin and anti-diuretic hormone (ADH) released during labour causes fluid retention which can be exacerbated by other factors. Potential additional risk factors from this case series include high fluid intake, exogenous oxytocin infusion and the use of birthing pools which are explored below.

Oral fluid intake

The excessive intake of hypotonic fluids has clear links with other types of hyponatraemia, particularly in the context of a reduced capacity for water excretion. ⁷ A prospective, observational, cohort study showed that the incidence of hyponatraemia immediately post-delivery is increased in relation to greater fluid intake in labour; an incidence of 26% was reported in women who had more than 2.5 l fluid intake, either orally or intravenously, compared to an incidence of 1% in women with a total intake of less than 1 l. ⁸ It also showed that babies’ cord sodium concentration was lower when women had a higher fluid intake in labour.

Oral intake in labour has increased over the last 30 years. Mendelson described the mortality and morbidity associated with aspiration of gastric contents during emergency general anaesthesia for caesarean section in 1946. This led to a widespread practise of restricting food and drink in labour. A relaxation in guidance for fasting during labour has stemmed from multiple studies showing no benefit from restricting food and fluid intake during low risk labour. ⁹ In higher risk labours, while solid intake may be restricted, women are often encouraged to have sugary drinks to maintain calorific consumption and hydration. This means that women in labour are rarely restricted in their oral fluid intake. Isotonic sports drinks have become a mainstay for labouring women with NICE recommending that they may be more beneficial than water. ¹⁰ Although these sports drinks are marketed as isotonic, they have a significantly lower concentration of sodium than blood. Hyponatraemia can occur by drinking either water or isotonic drinks to excess. ⁷

Outside of labour and pregnancy, hydration is widely encouraged, with its health benefits promoted by the media without a clear evidence base. ¹¹ This has created a phenomenon of cultural polydipsia. Society has widely adopted this, reflected in increased sales of drinks bottles and bottled water. ¹² Women who have already adopted a pattern of increased water intake, extend this behaviour into pregnancy and labour when physiological and hormonal influences increase the risk of fluid retention. Three of the women discussed in this review drank in excess of 6 l of fluid during labour and two commented on retrospective review with their birth partners that they drank because of their own impulse without external encouragement to do so.

Management of ketonuria

Four out of the six cases in this review had documented ketonuria in labour. Two women were encouraged to increase their oral fluid intake to treat ketonuria with one having additional IV fluid. One woman drank 9 l of isotonic drinks during her 11 h labour, whilst being encouraged to eat and drink to treat ketonuria. Treating ketonuria with oral and IV fluid in women who are fluid replete, has been a factor in previously published cases of peripartum hyponatraemia. ⁴ In these circumstances excessive oral fluid has been encouraged and IV fluids have been repeatedly administered, in futile attempts to treat worsening ketonuria. The presence of ketones in urine is common in labour and there are links with poor labour outcomes, although there is no robust evidence that it is a causal factor. ¹³ Other evidence suggests that ketonuria is part of the normal physiology of pregnancy. ¹⁴ Urinalysis dipstick tests for ketones vary in their sensitivity and specificity. ¹⁵ Ketogenesis is a result of the body‘s response to starvation, where the lack of the ability to use glucose as substrate for metabolism causes the breakdown of fatty acids. The treatment for ketosis is carbohydrates such as sugar, rather than fluid. The ongoing practise of using fluids to reduce the presence of ketonuria suggests that there is a requirement for further education in clinical practise. Future research to determine causal links between ketosis and pregnancy outcomes is also required.

Intravenous fluid

Intravenous fluids have been linked to early case reports of hyponatraemia, particularly when hypotonic solutions were used to dilute oxytocin for labour augmentation. ¹⁶ The use of 5% dextrose infusions in labour has reduced, with oxytocin commonly diluted in 0.9% sodium chloride in the UK. 5% dextrose is still infused alongside variable rate insulin infusions in women with diabetes, which should prompt close monitoring for hyponatraemia, especially around labour.

The NICE guidance on Intrapartum Care is clear that IV fluid should be considered for an abnormal CTG in the context of maternal hypotension. ¹⁰ In practise IV fluid is often used to treat an abnormal CTG in the absence of hypovolaemia. Hypotension itself is not always due to hypovolaemia. Mean arterial blood pressure is a product of cardiac output and total peripheral resistance. Cardiac output may be reduced by hypovolaemia, but also through decreased cardiac contractility, which can occur with volume overload. Reduced total peripheral resistance may be caused by the effect of epidural analgesia on vascular tone. A Cochrane review on the effect of prophylactic fluid preloading for low dose regional analgesia showed no difference in incidence of hypotension when fluid is administered before epidural analgesia. ¹⁷ Despite this fact, IV fluid is often routinely given with low dose epidural analgesia, without consideration of the woman‘s volume status (i.e. hypovolaemic, euvolaemic or hypervolaemic/overloaded) and without volumetric pumps to fix the rate of infusion. Local survey on fluid administration in 2019 showed that women with epidurals were more likely to receive over 2 l IV fluids when compared to other women with duration of labour over 6 h. One woman in the survey received more than 5 l IV fluid during a 7 h labour with epidural analgesia. This survey also showed that no maternity units in our region used volumetric pumps to administer fluid. ¹⁸ This led to a change in practise where IV fluid is now not routinely infused with epidural analgesia.

Fluid balance in labour is often poorly recorded because women may drink and void without informing midwives or being able to measure volumes. Overall, the aim in low risk settings has been to lessen such intervention. Local audit showed that 56% of women with labours of 6 h or more had inaccurate fluid balance documentation with only 49% having fluid balance charts. Staff involved in the cases described here, including midwives providing direct care, were unaware that women had drunk to excess until direct questioning regarding fluid intake occurred retrospectively. Four out of these six cases required retrospective fluid balance determination via direct questioning of the women and their birthing partners.

Birthing pools

Four of these six cases of peripartum hyponatraemia involve women spending part of their labour in birthing pools. This should be interpreted with caution. It is difficult to know if the pool itself is a causal factor, or merely an association. ¹⁹ The warm temperature of a birthing pool may make women sweat, encouraging them to drink. It also may lead healthcare staff to encourage further oral fluids, with guidelines on pool birth stressing the importance of drinking ‘plenty of fluids’. ²⁰ A pool environment confers particular difficulty in measuring urine output and overall fluid balance. More research is needed to assess whether birthing pools may either increase hormone secretion leading to water retention or whether free water can itself be absorbed during labour and delivery in a birthing pool. Overall caution should be exercised in women labouring for long periods in a birthing pool.

ADH and oxytocin

ADH and oxytocin can both control the resorption of water by the distal renal tubules. The release of ADH may itself increase around labour, particularly in response to nausea, pain and stress,^21,22 and after haemorrhage. ²³ In this series two women had a major obstetric haemorrhage over 2 l where ADH release may have contributed to fluid retention. Only two of these six cases received intrapartum, synthetic oxytocin which may have had an anti-diuretic effect – particularly where it was infused for 17 h. Three of the women received oxytocin post-delivery but were already demonstrating abnormal symptoms likely due to hyponatraemia prior to this. In these cases the anti-diuretic effect of oxytocin may have contributed to the severity of symptoms and slowed resolution. Whilst both hormonal mediators may contribute to fluid retention, further work is needed to establish the mechanisms of this potential interplay.

Management

Whilst fluid restriction and spontaneous post-partum diuresis aided resolution in these cases, three of these women received hypertonic sodium chloride (2.7% NaCl) as an emergency treatment for severe hyponatraemia. Our local guidelines advise the administration of 150 ml 2.7% sodium chloride over 20 min only in symptomatic cases where serum [Na+] less than 125 mmol/l. ²⁴ Whilst the risk of precipitating osmotic demyelination with hypertonic saline is low in acute hyponatraemia, the woman should be referred to critical care and hypertonic sodium chloride should be given under close supervision. The aim is to improve the degree of hyponatraemia to avoid severe complications but not to fully correct serum [Na+]. Spontaneous correction can occur quite rapidly post-partum if there is no ongoing oxytocin infusion. The Society for Endocrinology published general clinical guidelines on the management of severe hyponatraemia in 2016 though this does not feature the acute presentation of peripartum hyponatraemia. ²⁵

Importantly, these cases involved women in their first pregnancies who were deemed low risk and thus initially had less monitoring and intervention. Such women may have a longer duration of labour and lower body mass index. Reviews of exercise-associated hyponatraemia have shown links with smaller body mass index and longer duration of exercise.^7,26 Five of the six cases resulted in instrumental deliveries. Hyponatraemia has been associated with instrumental delivery and caesarean section for failure to progress. ⁸ A bench top pilot study has shown that hyponatraemia reversibly affects human myometrial contractility suggesting hyponatraemia can cause uterine dysfunction and lead to poor delivery outcomes. ²⁷ This link needs further evaluation.

In summary we have presented six cases of severe symptomatic hyponatraemia, highlighting potential risk factors for further evaluation. We also reflect on the physical and psychological impact on women who are left with little memory of the birth of their child. It is hoped that the upcoming UKOSS study into peripartum hyponatraemia will develop our understanding and drive research into causal factors and management for mothers and babies.