Zonisamide-induced distal renal tubular acidosis and critical hypokalaemia

Background

Renal tubular acidoses (RTAs), first described in 1951, are relatively rare disorders which can be challenging to diagnose and manage.1 They can be either inherited, or acquired in association with a range of other diseases and medications. This case presents a stepwise approach to a medication-induced distal (type 1) RTA and explores its successful treatment.

Case presentation

A woman in her 20s presented to the emergency department with a 2-day history of rapidly progressive muscle weakness. Symptoms began in the proximal muscle groups of her lower limbs with a slowed gait and leg heaviness and extended to her upper limbs, leaving her first unable to stand, then lift her limbs off her bed unaided.

For the preceding month, she had increased fatigability, nausea and vomiting up to five times per day, with limited oral intake. She had a history of well-controlled asthma, irritable bowel syndrome (IBS) and epilepsy. Her seizure pattern conformed to the International League against Epilepsy generalised motor tonic–clonic type, with a diagnosis of an idiopathic generalised tonic–clonic seizure alone epilepsy syndrome.2 3 She had a previously normal MRI of the brain and electroencephalogram under the neurology services in an alternative institution.

Her medications included terbutaline and budesonide/formoterol inhalers, montelukast, lamotrigine 275 mg nocte, zonisamide 300 mg two times per day and a combined contraceptive pill. She had previously used alverine citrate for symptomatic relief of IBS.

She had a family history of migraine in two first-degree relatives and ulcerative colitis in one first-degree relative. She lived independently with no alcohol, tobacco or illicit drug usage.

Her pulse, blood pressure and temperature were all within normal limits. Her muscle strength was graded at 1/5 at her shoulder and hip flexors and extensors bilaterally, 3–4/5 at her elbow flexors and extensors, 1/5 at her knee flexors and 4/5 for her knee extensors bilaterally and 5/5 for her wrist and ankle flexors and extensors. Reflexes were brisk and symmetrical with normal tone and sensation and no cerebellar signs. The remainder of her clinical examination was unremarkable.

Her initial investigations, when summarised, showed hypokalaemia, a compensated hyperchloraemic metabolic acidosis (using Winter’s formula to predict expected respiratory compensation), hypophosphataemia and mild microcytic anaemia. Her serum potassium was 1.8 mmol/L (normal 3.5–5 mmol/L) and phosphate 0.63 mmol/L (normal 0.8–1.4 mmol/L). Her arterial pH was 7.15 (normal 7.35–7.45), with an arterial oxygen partial pressure of 9.4 kPa (normal 11–15 kPa), arterial carbon dioxide partial pressure of 3.2 kPa (normal 4.5–6 kPa), bicarbonate 10 mmol/L (normal 22–28 mmol/L), chloride 119 mmol/L (normal 95–105 mmol/L) and anion gap 10 mmol/L (normal <11 mmol/L). Her haemoglobin was 101 g/L (normal 115–165 g/L) and mean corpuscular volume 78.9 fL (normal 80–96 fL).

Her white cell count, sodium, calcium, magnesium, liver function tests, serum lactate and C reactive protein were within the normal ranges.

Her ECG showed sinus rhythm with T-wave inversion in leads V3–V6 and a prolonged corrected QT interval of 581 ms (normal 360–460 ms).

Investigations

An extensive range of serum and urinary biochemical investigations was conducted to tease out the cause and guide further treatment of her critical hypokalaemia.

Her serum osmolality was 292 mOsm/kg (normal 285–295 mOsm/kg). Her morning cortisol was 140 nmol/L (normal 166–507 nmol/L). Her vitamin D level was 19 nmol/L (normal >50 nmol/L). Her iron level was 6 µmol/L (normal 10–30 µmol/L), total iron-binding capacity 57.7 µmol/L (normal 50–80 µmol/L) and transferrin saturation 10% (normal 15%–50%).

Her thyroid function, B₁₂, folate, ferritin, connective tissue disease screen, tissue transglutaminase, serum ACE and serum electrophoresis were all within normal limits.

Her urinary biochemistry showed a urinary potassium of 18 mmol/L, urinary calcium 1.56 mmol/L, urinary creatinine 1.8 mmol/L (normal 0.6–1.1 mmol/L), urinary pH 5.5 (normal 6–7.5), urinary anion gap 10 mmol/L and urinary osmolality 253 mOsm/kg (normal 50–1400 mOsm/kg).

The significance and relevance of each finding are explored below in relation to each of the potential differential diagnoses.

Differential diagnosis

The delta ratio was first calculated to confirm this was a purely normal anion gap metabolic acidosis (NAGMA). The principle of the delta ratio is that when a strong acid is added to plasma, a rise in the anion gap should be matched by a fall in serum bicarbonate. In a true NAGMA, this does not occur: no new anion is added, so the change in anion gap numerator remains unchanged while the serum bicarbonate change denominator enlarges, leading to the overall delta ratio falling to <0.4.4 In this case, the delta ratio was 0.08 supporting a NAGMA diagnosis.

The causes of a NAGMA can broadly be divided into three categories as outlined below5:

• Adrenal insufficiency usually presents with hyponatraemia, hypoglycaemia and hyperkalaemia with a suppressed serum cortisol.6 This patient did not have any of these biochemical findings which excluded adrenal insufficiency from the differential list.

• Iatrogenic chloride supplementation, such as with hydrochloric acid infusions (used previously to treat metabolic alkalosis), hypertonic saline or large volume extracellular fluid expansion with normal saline (based on in vitro studies) can all cause a hyperchloraemic metabolic acidosis, though not a serum hypokalaemia.7–10 This cause can be discounted based on the patient’s history.

• Loss of bicarbonate, either through the gastrointestinal or renal tract, is the third major cause of a NAGMA.5 This patient did not have significant diarrhoea, nor a history of fistulae nor enterostomies suggesting gastrointestinal bicarbonate loss; however, a renal cause had to be confirmed.

The normal renal response to acid loading is to increase ammonium excretion.11 As many laboratories do not measure urinary ammonium levels directly, the urinary anion gap is an indirect measure of its excretion.11 A gastrointestinal cause of NAGMA would lead to a negative urinary anion gap, as would defective bicarbonate resorption in the proximal renal tubule from a proximal RTA, since distal ammonium production and net acid excretion are maintained in these settings; in contrast, her positive urinary anion gap suggests urinary acidification in the collecting duct is impaired due to a distal RTA.12 13

Finally, renal potassium wasting is confirmed in two ways, by measuring either the urinary potassium:creatinine ratio (UK/Cr) or the transtubular potassium concentration gradient. A spot urinary potassium can vary based on urine volume, whereas creatinine is excreted at an approximately steady rate. Accordingly, a UK/Cr corrects for individual urine volume variations. The UK/Cr in this case was 10 mmol/mmol, with a cut-off of >1.5–2.5 mmol/mmol reported as diagnostic of renal potassium wasting.14 15 The validity of the transtubular potassium concentration gradient to gauge renal potassium secretion generally has been questioned; it was only ever applicable if the urinary osmolality equals or exceeds the plasma osmolality, which did not apply in this case, so for both reasons was not performed.16 17

Accordingly, a failure of distal renal tract urinary acidification and renal potassium wasting was confirmed using this stepwise approach. The underlying pathophysiology is thought to be a failure of the usual H⁺-ATPase and H⁺/K⁺-ATPase activity of hydrogen ion excretion and potassium resorption in the distal tubule, as well as potentially increased permeability of the tubular cells in the distal nephron allowing passive potassium ion secretion.1 18 Further evidence supporting a distal RTA is provided by the persistently alkaline urinary pH, which should be <5.3 in the context of acidaemia if renal function was preserved.12 18 The elevated calcium:creatinine ratio at 0.87 mmol/mmol (normal range 0.07–0.41 mmol/mmol) reflects the hypercalciuria found in distal RTA in inverse proportion to plasma bicarbonate concentration for multifactorial reasons.19 20

The final element was to determine the underlying cause of her distal RTA, which can be either inherited or acquired. She had been well until approximately 1 month prior to presentation without the clinical features or family history associated with the inherited renal acidoses, which often present in infancy with failure to thrive, vomiting, polyuria and polydipsia, especially in autosomal recessive gene variants, escalating to growth retardation and sensorineural hearing loss with more pathogenic genetic variants.20–22 Other causes of inherited RTA include carbonic anhydrase II deficiency, a rare disorder with osteopetrosis, cerebral calcification and a variable RTA picture, and nephropathic cystinosis, an autosomal recessive inborn error of metabolism which can present with the renal Fanconi syndrome and proximal RTA (unlike the distal RTA seen here).23 24 This patient’s acute history of illness, with normal blood tests 9 months previously and without prior health issues, made an inherited acidosis unlikely.

Acquired distal RTA can be associated with a range of conditions, especially Sjögren’s syndrome, but also systemic lupus erythematosus, primary sclerosing cholangitis, sickle cell anaemia and Wilson’s disease; her normal connective tissue disease autoantibody screen and serum ceruloplasmin did not support any of these as underlying causes.18 She had structurally normal kidneys on ultrasound, with no evidence of proteinuria, and a normal serum protein electrophoresis, making amyloidosis unlikely. Furthermore, she had no exposure to heavy metals such as lead which can cause tubulointerstitial damage.25

Her vitamin D levels were noted to be low, which is a common finding in the general population in her country of origin.26 Low vitamin D levels have been associated with proximal RTA and the renal Fanconi syndrome, so, while she needed vitamin D replacement given her low serum levels, this could not have been the cause of her distal RTA.27 28

Lastly, distal RTA has been associated with a range of medications such as amphotericin B, lithium and high-dose ibuprofen.18 This patient had initially been prescribed zonisamide approximately 2 months prior to hospital admission. The dose had been titrated from 150 mg two times per day to 300 mg two times per day 1 month prior to admission, which precisely corresponded temporally with onset of nausea and vomiting. Metabolic acidosis from renal bicarbonate loss due to zonisamide therapy is well recognised, and is included as a special warning in its summary of product characteristics approved by the European Medicines Agency as part of its marketing authorisation, which further raised suspicion that zonisamide was a possible causative agent.29 Direct measurement of serum zonisamide drug levels is unhelpful. A direct relationship between serum drug concentration and clinical response has not been established, making routine therapeutic drug level monitoring of uncertain value; similarly, the development of metabolic acidosis is not related to either zonisamide dosage level or treatment duration.30–32 Even in cases of zonisamide overdose, measurement of serum drug levels is not mandated due to poor correlation with clinical features.33 34

Although the formoterol element of her combination inhaler can contribute to hypokalaemia, none of her other medications can cause a distal RTA.35

Treatment

The treatment approach to a drug-induced distal RTA is threefold. First, any causative medications are withdrawn. Second, potassium replacement is initiated to correct critical hypokalaemia, with either oral or intravenous potassium chloride.18 The total body potassium deficit is difficult to quantify, but it has been estimated conservatively that every 0.3 mmol/L decrease in serum potassium below the lower limit of normal represents a deficit of 100 mmol of total body potassium.36 Ongoing losses must also be included in estimating replacement rates. Lastly, alkali therapy is given to correct the metabolic acidosis. Sodium or potassium bicarbonate at the rate of 1–2 mmol/kg/day is usually sufficient to equal daily acid production, though potassium citrate may be preferred if there is a history of nephrolithiasis since large doses of sodium salts can increase urinary calcium production and stone formation.18 22 Alkali therapy is given after potassium replacement to prevent a bicarbonate-induced intracellular potassium shift worsening hypokalaemia.14 37

On the advice of the neurology service, her zonisamide was tapered rapidly by initially halving the dose to 150 mg two times per day, then weaning it by 50 mg every day over 7 days to stop. It was replaced by brivaracetam 50 mg two times per day, with temporary bridging cover with clobazam 5 mg two times per day, given the rapidity of the zonisamide taper.

Intravenous potassium replacement was commenced initially at a rate of 240 mmol/day. This was gradually switched after 48 hours to enteral potassium with bicarbonate replacement at a rate of 102 mmol of potassium and 47 mmol bicarbonate daily, with intravenous replacement limited to intermittent top-up boluses as required guided by regular venous potassium sampling. Other electrolytes such as magnesium and phosphate were kept within the normal ranges.

Intermittent intravenous potassium replacement was required until day 14 of her inpatient stay, at which time she was maintained on enteral replacement alone. She was also administered parenteral iron which corrected her iron-deficient anaemia.

Outcome and follow-up

Her total hospital length of stay was 27 days, of which 16 were in the intensive care unit pending stabilisation of her serum potassium levels. Her serum potassium, pH, bicarbonate and chloride levels during her critical care stay are shown in figure 1 below.

She was discharge on a tapering dose of oral potassium chloride and sodium bicarbonate, which were rationalised to monotherapy with potassium citrate alone with monthly serum potassium monitoring by her primary care physician and nephrologist.