Standfirst

The pathogenesis of exercise-related death among those with SCT is not understood. We hypothesise that potassium efflux from red cells and skeletal muscle causes hyperkalaemic death.

Sickle cell haemoglobin

Each of the haemoglobin (Hb) molecule's four subunits contains an oxygen-carrying haem group and a globin molecule. Two α-subunits are bound to two γ-subunits in fetal haemoglobin (HbF, α₂γ₂) or two β-subunits in adult haemoglobin (HbA, α₂β₂). The haemoglobin in which two α-subunits are associated with two mutant β-subunits (in which a hydrophobic valine replaces a hydrophilic glutamic acid) is referred to as sickle cell haemoglobin (HbS).

Hypoxia and falling pH cause HbS to polymerise, distorting red blood cell (RBC) shape (‘sickling’), whose extent depends upon the RBC concentration of HbS relative to HbA/HbF. This is lower in heterozygotes (HbAS with SCT) than homozygotes (HbSS with sickle cell disease (SCD)). Sickling-induced vaso-occlusion can lead to so-called ‘sequestration crises’, but can also damage organs including the bone, brain, lung and gut.

The mechanism of microvascular occlusion in SCD is not simply flow limitation through RBC deformation, however. Reticulocytes and permanently deformed dense cells adhere to the postcapillary venule endothelium via adhesion molecules including thrombospondin, vascular cell adhesion molecule-1 and cluster of differentiation 36. Leucocytes bind to sickled cells to form heterocellular aggregates, worsening obstruction through their inferior deformability, and through leucocyte-endothelial adhesion. Local inflammation amplifies endothelial activation and cellular adhesion. Vaso-occlusion worsens local hypoxia and thus HbS polymerisation, driving further occlusion (reviewed in reference 1).

Variation in SCD phenotype

While possession of HbS is determined monogenically, variation in other genes influences SCD phenotype2: leucocyte concentration affects the incidence of acute chest syndrome and cerebral infarction while higher HbF concentration ameliorates SCD phenotype, and both are under strong genetic influence.2 Nitric oxide (NO) bioavailability influences endothelial/sickle cell adhesion and variation in the endothelial nitric oxide synthase (eNOS) gene is associated with variation in the incidence of chest syndrome in SCD.3

Other candidate genes of influence might include those regulating RBC membrane cation transport, which regulate cell hydration state and consequently cellular HbS concentration and the propensity to polymerisation. Such cation transporters include the calcium-activated K⁺ (Gardos) channel, the potassium–chloride co-transporter and the deoxygenation-induced cation-selective channel (P_sickle), which promotes calcium influx and Gardos activation.4

SCT and sudden death

SCT is common (affecting <8% of the USA black population) and is associated with a modest increased risk of thromboembolism, haematuria and impaired urinary concentrating ability. However, the risk of exertional death is greatly elevated, with athletes now being screened for SCT in the USA.5 Among 2.1 million US military recruits, rates of otherwise unexplained death were 32.2/100 000 and 1.1/100 000 for recruits with and without SCT, respectively (RR 27.6). Risk has been confirmed in US military recruits6 and in recreational and professional athletes,7 and rises with age (an eightfold cline from late teens to late twenties) and intensity of exertion.8

In such cases, intense exercise immediately precedes a rapid decline and death within minutes or hours. Such rapid unanticipated collapse (together with ignorance of HbS status and of the clinical scenario) means that the mode of death is often poorly described. While ‘heat stroke’ and rhabdomyolysis are often associated, they might occur independently and neither seems universally present (perhaps one third of cases exhibit neither). Profound hyperkalaemia is, however, frequently identified,9 whether in isolation or in the context of rhabdomyolysis, and may prove resistant to interventions including haemodialysis.10

Putative model

Sustained exercise evokes four pro-sickling forces: local hypoxia, acidosis, raised temperature and RBC dehydration. Even in the absence of sickling-related microvascular occlusion, however, exercise-related changes in RBC rheology may impair tissue oxygen delivery directly and indirectly by driving local inflammatory responses and adhesion molecule expression.11

Whether through exercise-induced microvascular sickling or rheological impairment, SCT may suddenly impede skeletal muscle oxygen delivery and worsen local acidaemia, thus worsening flow limitation in a positive feedback loop and triggering skeletal muscle ‘metabolic failure’. The frequent identification of rhabdomyolysis is in keeping with this hypothesis: a fall in ATP production plays a fundamental role in the pathogenesis of exertional rhabdomyolysis through failure of the muscle sodium pump. In this context, associated potassium efflux could drive catastrophic hyperkalaemia.

Serum potassium might abruptly rise for other reasons too. RBC membrane deformation during HbS polymerisation induces potassium leak12 that worsens cellular dehydration, favouring further HbS polymerisation. The anion exchanger 1 (AE1), a major structural membrane component that also mediates chloride–bicarbonate exchange may play a crucial role in this. Normal RBCs demonstrate a (likely AE1-mediated) cation leak when undergoing shear stress,13 and AE1 is the proposed mediator of deoxygenation-induced cation-selective channel (P_sickle) cation flux.14 When compared with normal RBCs, sickle RBCs demonstrate an exaggerated potassium leak in response to shear stress, which is further enhanced at lower pHs.15 Furthermore, sickle RBC membrane components including AE1 are readily oxidised and even subtle peroxidation causes exaggerated cation leaks from the RBC membrane in response to mild shear stress. Furthermore, RBC deformation and oxidation interact synergistically to promote cation leak.16 Thus, deformation, hypoxia and low pH conditions that the RBC encounters in exercising muscle promote the efflux of potassium from RBCs.

The skeletal muscle vascular bed and thus the mass of affected RBCs and the scale of potassium efflux is large. Augmenting that derived from skeletal muscle metabolic failure and in the context of the described positive feedback loops of sickling and potassium efflux, catastrophic hyperkalaemia may result (figure 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

A proposed schematic of the pathophysiological processes which might culminate in exertional death in SCT subjects: a final common pathway of hyperkalaemia is suggested.

Death in SCD

If such a mechanism exists, ought it not to be readily identified in those with SCD?

In general, SCD-related death may result from acute end-organ damage from sickle crises or sepsis related to functional hyposplenism. Chronic end-organ damage (renal failure or pulmonary hypertension) contributes to SCD deaths, rarely, exclusively explaining a sudden fatal deterioration. Transfusion-related iron overload (causing cirrhosis and cardiomyopathy) also occurs.

However, SCD death is ‘frequently sudden and unexpected’,17 one detailed series documenting ‘sudden unexpected hospital or home death in an otherwise relatively healthy patient, with or without vaso-occlusive crisis’ as the cause in 23.4%.18 A further 22.6% died of ‘renal failure’ and 4.2% of ‘multiorgan failure’, the authors commenting that ‘most patients had more than one factor contributing to death’. The prevalence of identified hyperkalaemia or rhabdomyolysis (or a history of preceding exertion) in these cases is not known. It might thus be that hyperkalaemia related to muscle microvascular flow limitation does occur in SCD but goes unrecognised. Alternatively, such cases might not be prevalent in adults if a propensity to such pathology proved lethal earlier in life (‘survivor bias’).

A polygenic pathogenesis?

But why would only some individuals with SCT be affected by exertional sudden death? Environmental factors (exercise intensity, altitude, hydration status and the presence of intercurrent infection) may interact with genetic variation to influence risk.

A variety of candidate genes might be suggested. Gene variation influences cation leak via AE119 or the ammonium transporter rhesus-associated glycoprotein,20 while variation in the eNOS gene may also play a role (above), especially given the male preponderance of sudden death in SCT and the reduced bioavailability of NO in males with SCD.1 Further candidates include genes influencing HbA/HbF/HbS concentrations; nicotinamide adenine dinucleotide phosphate oxidase 1 (NOX-1) or NOX-4,which are expressed in vascular smooth muscle cells and generate vasoconstricting local superoxides; and glucose-6-phosphate dehydrogenase (G6PD), the rate-limiting step in the pentose phosphate pathway that generates nicotinamide adenine dinucleotide phosphate, eliminating vasoconstricting reactive oxygen species. G6PD gene variants would also influence RBC oxidation and thus potassium permeability, while G6PD deficiency decreases endothelial NO bioavailability.

Summary

In summary, SCT is associated with exertional sudden death. We propose a fundamental role for hyperkalaemia in mediating such deaths. We suggest that local hypoxia and acidosis worsen HbS polymerisation in a vicious (and viscous) circle. Resultant rheological impairment (possibly augmented by local sickling) leads to a sudden decline in skeletal muscle oxygen availability causing metabolic failure. Resultant potassium extrusion is augmented by that resulting directly from RBC deformation. Variation in genes influencing microvascular flow and potassium efflux may modulate the risk of exertional death across a seemingly homogeneous population.

We would advocate research in this area. Once initiated, catastrophic decline may prove hard to arrest, meaning that a preventive strategy might best be deployed. While this should certainly include modification of training patterns, it might also be that individuals at risk could be identified through the use of genetic markers in the future.