SARS-CoV-2 and hypokalaemia: evidence and implications

Introduction

The global pandemic secondary to the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is leading to unprecedented global morbidity and mortality. Anecdotal data and clinical observation from patients with the World Health Organization named disease coronavirus disease 2019 (COVID-19) has highlighted a bewildering array of presentations and pathologies. These include respiratory complications (acute hypoxic respiratory failure, severe pneumonia, acute respiratory distress syndrome (ARDS) and pulmonary fibrosis)¹, cardiovascular complications (myocardial injury, myocarditis, cardiomyopathy, cardiac tamponade, acute myocardial infarction, heart failure, dysrhythmias and venous thromboembolic events)², neurological complications (anosmia, myopathy, acute stroke, Guillain-Barre syndrome and encephalopathy)³, acute liver and/or pancreatic injury⁴, cytokine storm syndrome, septic shock, DIC, diarrhoea, Kawasaki-like disease⁵and renal complications (acute tubular injury, rhabdomyolysis, proteinuria, secondary focal segmental glomerulosclerosis and possible renin-angiotensin-aldosterone system activation)⁶.

SARS-CoV-2 may affect the kidneys directly and indirectly by both renal and renin-angiotensin-aldosterone system (RAS) involvement. What is uncertain are the exact pathophysiological mechanisms that play a role in this. Early reports have suggested that hypokalaemia may be seen frequently in patients infected with the SARS-CoV-2 virus. Here we aim to review the evidence and the potential clinical implications of this well recognised but potentially serious electrolyte disturbance in patients with SARS-CoV-2 infection and touch on renal features of the infection discovered so far.

Firstly, it is important to highlight how transmission of the virus occurs in humans. The SARS-CoV-2 spike protein enters host cells via the angiotensin-converting enzyme 2 (ACE2) receptor on the surface of pulmonary type 2 alveolar cells^7,8. ACE2 is crucial to counter-regulate RAS and shares approximately 60% homology with ACE. RAS pathway components are co-expressed in most organs and tissues in the body and communicate via both paracrine and autocrine signalling. ACE2 converts angiotensin-II into angiotensin-(1-7), which acts on the Mas receptor expressed on a variety of tissue cell lineages relevant to cardiovascular disease in addition to type 2 alveolar epithelial cells. It lowers blood pressure through vasodilation and promotion of renal sodium and water excretion and it also attenuates inflammation by producing nitric oxide. Meanwhile, ACE converts angiotensin-I to angiotensin-II which has directly opposing effects to signalling mediated by ACE2. Angiotensin-II acts at the type 1 angiotensin receptor (AT1R) to increase blood pressure by the induction of vasoconstriction and renal sodium and water reabsorption. This also creates oxidative stress which promotes inflammation and fibrosis. The balance between these two opposing pathways determines potential tissue injury, predominantly in the heart and kidneys^7,8. Not only has the ACE2 receptor found to be used for host cell entry by the SARS-CoV virus, but the affinity for SARS-CoV-2 is 10-20-fold higher⁹. During host cell entry, proteolytic cleavage of ACE2 by transmembrane serine protease 2 (TMPRSS2) occurs which is thought to suppress ACE2 expression. This in theory would lead to a reduction in angiotensin-(1-7) generation and angiotensin-II levels would increase resulting in oxidative-stress mediated tissue damage and hypertension^7,8. This mechanism resulting in lung parenchymal injury has already been demonstrated in mice injected with SARS-CoV-2 virus¹⁰.

Involvement of the RAS pathway has therefore been speculated as a mechanism to lead to more severe COVID-19 disease and patients with severe disease were more likely to have hypertension, chronic kidney disease (CKD), diabetes mellitus and cardiovascular disease than those with milder disease. This is, however not confirmed and is preliminary data that may be the subject of bias with disregard of age as a potential confounding factor⁸.

Hypokalaemia is known to cause muscle weakness, paraesthesia, thirst, muscle cramps and weakness, but hypokalaemia is an important complication of any disease due to it being a potentially life-threatening condition¹¹. In fact, the incidence of ventricular fibrillation has been shown to be fivefold higher in patients with hypokalaemia compared to those with hyperkalaemia¹². Furthermore, inadequate management of hypokalaemia has been identified in 24% of hospitalised patients¹³, reiterating that we need to be even more vigilant with the management of this life-threatening condition during a pandemic where resources are even more scarce.

Not all patients with hyperaldosteronism develop hypokalaemia

The theory of RAS activation causing hypokalaemia is popular, however it is recognised that hyperaldosteronism does not always cause hypokalaemia. This is explained by the ‘renal potassium switch’ mechanism (Figure 1), which involves the paradoxical actions of aldosterone on potassium; for example, aldosterone increases sodium reabsorption in states of low circulating volume without significantly altering potassium balance, but it also increases serum potassium excretion in high potassium states without affecting sodium balance. This paradoxical action all depends on the sodium delivery to the aldosterone-sensitive distal nephron (collecting duct) which is generated by the sodium-chloride co-symporter (NCC) in the early distal convoluted tubule. Activation of this ‘switch’ is by phosphorylation of NCC by the WNK/SPAK/OSR1 pathway^17,18 and is done by aldosterone in response to low or high serum potassium i.e. NCC is maximally phosphorylated when serum potassium is <2.5 mmol/l and not at all phosphorylated when serum potassium is >5.2 mmol/l¹⁹. In states of decreased circulating volume, ENaC activity is increased due to the effects of excess aldosterone (which generates a sufficient intracellular electrochemical gradient to cause active transport of potassium from blood via Na⁺K⁺-ATPase which would be passively excreted into the tubular lumen by ROMK and BK channels in the Principal cell). Whether this generates an effect on potassium or not depends on the integrated activity of the entire distal nephron. In a state of low circulating volume, NCC in the early DCT is ‘switched on’ which results in increased reabsorption of sodium and chloride in the early DCT and, along with sodium reabsorption in the PCT, a net increase of sodium reabsorption from the kidney occurs. As a result, sodium delivery to the aldosterone sensitive distal nephron (collecting duct) is decreased which offsets the effects of ENaC and potassium excretion is therefore unchanged. This results in a state of hyperaldosteronism without hypokalaemia.

Figure 1. NCC Phosphorylation downstream and changes in urinary potassium (renal potassium switch).

In states of hyperaldosteronism and volume depletion the renal potassium switch is ‘active’. Angiotensin-II (ang-II) stimulates Kir4.1/5.1 to hyperpolarise the basolateral membrane potential and lowers intracellular chloride. This action activates WNK1 (With-no-lysine-kinase-1) and WNK4 (With-no-lysine-kinase-4) by phosphorylation which subsequently phosphorylates SPS1-related proline / alanine-rich kinase (SPAK) and oxidative stress-responsive kinase (OSR1) which, again, phosphorylates the sodium chloride co-symporter (NCC). NCC then avidly reabsorbs sodium (and chloride) back into the body in the early part of the distal convoluted tubule (DCT1). This results in reduced tubular sodium delivery to the collecting duct. Subsequently there is less sodium for the Epithelial sodium Channel (ENaC) to reabsorb and its action is offset. ENaC usually reabsorbs sodium and, via an electrochemical gradient made by sodium-potassium-ATPase, potassium is normally excreted into the tubular lumen by ROMK (Renal Outer Medulla Potassium channel) and BK (Big Potassium/Maxi-K channel) and eliminated from the body. When ENaC is less active, as a consequence of less tubular sodium delivery, less potassium is excreted. This results in a state of hyperaldosteronism with a normal serum potassium. In states of hyperaldosteronism and normal cardiovascular volume (in response to high serum potassium) the switch is ‘inactive’. NCC is not phosphorylated which results in more distal sodium delivery to the collecting duct and subsequent ENaC activation resulting in excretion of potassium into the tubular lumen and a state of hypokalaemia^17,18. Arrows represent the direction of movement of electrolytes. Red arrows reflect a reduction in movement and green arrows represent an increase in movement. CLC-Kb, chloride voltage-gated channel Kb; Kir4.1/5.1, inwardly rectifying potassium channel 4.1; AT1R, type 1 angiotensin receptor.

Discussion

The data from China and evidence from the SARS-CoV case series provide good clinical justification to monitor potassium levels in patients infected with SARS-CoV-2. We do need to see this clinical finding replicated in multi-centres to be able to make definitive conclusions about this clinical sequela. Although hyperkalaemia can occur for many reasons in patients with SARS-CoV-2 infection, the incidence of ventricular fibrillation has been shown to be five-fold higher in patients with hypokalaemia compared to those with hyperkalaemia¹².

Based on evidence of significant urine potassium wasting which can be prolonged, we suggest checking the transtubular potassium gradient (TTKG) in patients with SARS-CoV-2 infection. This would help identify those at risk of severe or prolonged hypokalaemia and prompt pre-emptive cautious electrolyte monitoring and replacement which should help reduce the risk of hypokalaemic complications which can be fatal. The TTKG is easily measured and only requires measurements of serum and urine osmolality and potassium (urine potassium/serum potassium)/(urine osmolality/ serum osmolality). The estimation is relatively accurate providing the urine is not dilute and that urine sodium concentration is >25 mmol/l so that distal sodium delivery is not limiting²⁸.

Measuring components of the renin-angiotensin-aldosterone pathway would help us to identify the involvement of the SARS-CoV-2 virus; however, measuring these components in the acute setting has its limitations. The RAS system is heavily influenced by a multitude of factors. Many of the patients with SARS-CoV-2 infection are critically unwell and often present to hospital in states of low cardiovascular volume. It has been recommended that patients with ARDS also be kept in a relative state of volume depletion as to not worsen pulmonary interstitial oedema and thus oxygenation. Both of these factors would typically result in RAS activation and could be very misleading when determining virus-related RAS pathway involvement. Some studies have shown that low urine pH (causing acidic urine) can be a predictor of RAS activation²⁹ so future studies determining RAS system involvement of SARS-CoV-2 infection should, as well as TTKG, also consider urine pH as an early marker of RAS system activation.

Potassium levels <3.0 mmol/l can be arrhythmogenic and specifically can cause QTc interval prolongation, torsade de pointes, ventricular fibrillation and sudden cardiac death³⁰. This is particularly relevant for a couple of reasons. First, many of the drugs currently undergoing clinical trials in SARS-CoV-2 patients prolong the QTc interval (hydroxychloroquine³¹, azithromycin³¹, favipiravir, lopinavir/ritonavir and fingolimod: NCT04280588) and some drugs also risk hypokalaemia which may worsen pre-existing electrolyte disturbance (e.g. methylprednisolone: NCT04273321, NCT04263402; thalidomide: NCT04273529, NCT04273581; and bevacizumab: NCT04275414). Second, cardiac involvement in SARS-CoV-2 is high (44.4% of infected patients admitted to ICU experienced an arrhythmia). Induction of arrhythmias can be due to multi-factorial aetiologies of cardiac injury in SARS-CoV-2 patients such as hypoxia-mediated, direct tissue damage, cytokine-storm syndrome, worsening coronary perfusion and the direct effects of medications³². Both of these factors really emphasise the importance of maintaining normokalaemia in these patients to reduce morbidity and mortality.

In addition to the arrhythmogenic effects of both the SARS-CoV-2 cardiac sequelae and various clinical trial drugs, many patients are being given diuretics to improve the oxygenation in ARDS, which also risks hypokalaemic complications. If diuretics are to be used, it would be wise to consider potassium-sparing agents in hypokalaemic patients to reduce the cardiac complications of worsening hypokalaemia that can occur with those that are not potassium sparing.

It would be logical that with clear histopathological data suggesting tubular injury, hypokalaemia is to be expected in patients infected with SARS-CoV-2³³. Tubular injury seems to be multifactorial in these patients; ATN from SIRS/Cytokine Storm Syndrome, peritubular congestion, virion invasion, drug toxicity and pigment-cast nephropathy from rhabdomyolysis and therefore difficult to avoid.

The data on outcomes in cardiac arrest for SARS-CoV-2 patients is unfortunately poor. Chinese data suggests that in a series of 136 patients who underwent resuscitation efforts, return of spontaneous circulation was only achieved in 18 (13.2%) and only four patients were alive at 30 days³⁴. The strong association of hypokalaemia with arrhythmia and sudden cardiac death reiterates the importance of vigilance for detection and treatment of this electrolyte disturbance during and after admission given such poor outcomes when spontaneous circulation is lost.

To truly know the extent of involvement of the RAS pathway as a potential cause of hypokalaemia in SARS-CoV-2 patients, further studies require measurement of all components of the RAS pathway. Until we know this, we cannot make definite conclusions about the mechanisms of hypokalaemia in these patients outside of the histopathological evidence of tubulopathy published, of which multiple aetiological factors are likely.

Data availability

No data are associated with this article.