In the present study, we found that in ACS patients, serum K level is significantly decreased during ischemic attack compared to the stable phase in individual subjects. We examined multiple clinical factors affecting the degree of the potassium dip (as indicated by ΔK), and found that the plasma glucose level during ischemic attack was the sole factor which was positively correlated with ΔK. In contrast, HbA1c level was negatively correlated with ΔK.
There was no link between the medication profiles and ΔK, such as the use of ACE-inhibitors, ARBs, MR-inhibitors and diuretics, after adjusting for any potential confounders of ΔK, even in patients who had been on these medications before admission (Table 3). Moreover, the renal function, as indicated by eGFR did not affect the potassium dip, suggesting that the potassium kinetics during ischemic attack are not simply regulated by renal elimination.
Although a couple of potential mechanisms for this K decrease have been proposed [1–4, 14], the precise mechanisms remain to be elucidated. Meanwhile, we have previously reported that K level decreases with the severity of heart failure if renal function is preserved .
It is possible that insulin stimulates an intracellular K shift into the cardiac and skeletal muscles via Na+/K+ATPase activation, leading to the decrease in serum K level [16–19]. In the present study, plasma glucose level during attack was positively correlated with HbA1c level, thus, at least theoretically, HbA1c level would also be positively correlated with ΔK. However, we found that HbA1c was not correlated with ΔK in the simple regression analysis and was actually negatively correlated with ΔK after adjusting for any potential confounders. Although this negative correlation is relatively weak, one can still say that an elevated glucose level during attack is tightly associated with an enhanced ΔK, regardless of the severity of diabetic condition. These findings suggest that insulin resistance may have had a role in attenuating the potassium dip, and that there are other serum K lowering systems that may overwhelm the effects of insulin resistance, as discussed below. Further investigations are required to fully demonstrate that the tight correlation between glucose and ΔK overwhelms insulin resistance using other parameters, such as homeostasis model assessment ratio (HOMA-R), which was not available in the current study.
The systemic sympathetic nerve system can be activated by ischemic stress, and elevated catecholamines stimulate Na+/K+ATPase primarily via β2-adrenergic receptor [1–4, 20, 21]. In fact, in the present study, non-selective β-blocker use on admission, but not β1 selective β-blocker use, reduced ΔK in the one-way ANOVA (Figure 4), although the multiple regression analyses showed only a tendency toward a negative correlation between non-selective β-blocker use and ΔK that did not achieve statistical significance (Table 3). These data indicate that catecholaminergic effects via β2-adrenergic receptor would be partially involved, but that they cannot explain the entire extent of the potassium dip.
Sodium-proton exchanger (NHE) also stimulates Na+/K+ATPase . Although insulin is one of the NHE activators [22, 23], diabetic condition (namely, hyperinsulinemia) as indicated by increased HbA1c, rather reduces ΔK. It is possible that other factors, such as intracellular acidification and some neurohumoral regulators, including the renin-angiotensin-aldosterone system are directly involved in the activation of NHE under the presence of insulin resistance [24, 25].
One can infer that the intracellular components, including potassium, leak out when cardiomyocytes are damaged by ischemic attacks, thus leading to an increase in serum K level during attack (namely, ΔK reduction) just like cardiac enzymes, such as CK(−MB). However, the present study demonstrated that a higher peak CK level was observed in patients with a larger ΔK and a lower K (K < 4.1) on admission (Table 4). Moreover, the subanalysis with MI subjects (n = 188) in a multiple regression analysis demonstrated that ΔK was not significantly correlated with peak CK level (if anything, ΔK was actually found to be positively correlated with peak CK level in a simple regression analysis, data not shown, n = 188, P = 0.002), thus suggesting that ΔK reflects the severity of ischemic stress rather than the extent of cellular injury.
In accord with previous reports demonstrating that a decrease in serum K level during the acute phase of ACS increases the risk of cardiovascular events [1, 3–7], we found in the present study that more severe ischemia was observed in patients with a lower K level on admission (Table 5). Moreover, the present study revealed that a lower serum K level on admission was associated with a larger ΔK (Figure 2b, Tables 2 and 3). Considering that the patients who presented with lower K level during ischemic attack did not necessarily continue to exhibit relatively low K concentration during stable phase, these data indicate that lower K subjects are more susceptible to larger potassium dip, thus suggesting that serum K level on admission per se reflects disease severity.
The main limitation of this study was that we did not measure the hormonal changes (i.e. serum concentrations of catecholamine, insulin, aldosterone, ACTH, cortisol etc.) or the urinary potassium concentration during attacks. Therefore, the mechanisms described above still remain speculative. Moreover, it would be very interesting to determine the K and glucose concentrations in the coronary sinus, so that the local potassium kinetics in the ischemic heart, where glucose becomes an important preferential substrate for metabolism [26, 27], could be evaluated.