Potassium supplementation and the prevention of atrial fibrillation after cardiac surgery (TIGHT-K) trial

The Prevention of Dysrhythmias on the Cardiac Intensive Care Unit - Does Maintenance of High-normal Serum Potassium Levels Matter (TIGHT-K) trial is a multicentre, randomized, open-label, non-inferiority study investigating whether relaxed vs. tight potassium control, with supplementation only when the serum levels drop <3.6 mmol/L vs. already with a decrease <4.5 mmol/L, respectively, leads to a similar incidence of post-operative atrial fibrillation (AF), atrial flutter, or tachyarrhythmia, in patients undergoing coronary artery bypass grafting (CABG).1 The trial recruited patients undergoing isolated CABG surgery, without history of atrial arrhythmias or use of rhythm-controlling drugs. Patients undergoing cardiac surgery are at high risk for perioperative morbidity and mortality, with electrolyte disturbances a potential contributer.2 Atrial fibrillation is the most frequent post-operative adverse event, occurring in roughly one out of three patients.3 Many strategies including beta-blocker therapy, rate-controlling calcium blockers, intravenous magnesium, amiodarone, sotalol, bi-atrial pacing, colchicine, corticosteroids, or sodium–glucose co-transporter-2 inhibitors are used or tested for post-operative AF prevention, yet most strategies have relatively weak evidence in support.4 Potassium is the most important intracellular cation, and low potassium levels are associated with worse outcomes including a higher chance of atrial arrhythmias in observational studies, with the optimal serum potassium level associated with the lowest risk ∼4.5 mmol/L.5 Therefore, the TIGHT-K trial had a solid hypothesis with a clinically relevant question. The trial was registered at clinicaltrials.gov (NCT04053816). Randomized (allocation 1:1 to potassium supplementation only when serum levels <3.6 mmol/L vs. already with serum levels <4.5 mmol/L) Open-label with blinded outcome assessment Parallel arm (two arms) Multicentre: 23 study locations in the UK and Germany Non-inferiority design with 10% non-inferiority margin Follow-up time: 5 days Total number of trial participants: 1690 (843 randomized to relaxed potassium control; 844 randomized to tight potassium control, 3 randomization errors that were excluded) Seventeen patients excluded before the randomized intervention, leaving 1667 (98.6%) in the intention-to-treat analysis Mean age: 65 years Men/women: 85%/15% White/Asian/other race: 87%/10%/3% Body mass index: 29.1 kg/m2 EuroSCORE II: 1.5 ± 1.3 Beta-blocker use 78% Potassium concentration coming off bypass: 5.0 ± 0.7 Patient ≥18 years of age scheduled to have isolated coronary artery bypass graft Sinus rhythm Previous history of AF, atrial flutter, and/or atrial tachyarrhythmia Pre-operative high-degree atrioventricular block, defined as Mobitz type 2 second-degree atrioventricular block or complete heart block Pre-operative serum potassium concentration >5.5 mmol/L Current or previous use of medication for the purposes of cardiac rhythm management Dialysis-dependent end-stage renal failure Concurrent patient involvement in another clinical trial assessing cardiac rhythm post-operative interventions Unable to give informed consent Episode of AF (or atrial flutter or tachyarrhythmia) lasting ≥30 s that is both clinically detected and electrocardiographically confirmed (on either a 12-lead electrocardiogram, telemetry, or Holter monitoring) within the first 5 days after surgery. The incidence of any episode of AF (or atrial flutter or tachyarrhythmia) lasting ≥30 s, either clinically detected or through ambulatory heart rhythm monitoring within the first 5 days after surgery The incidence of any episode of AF (or atrial flutter or tachyarrhythmia) lasting ≥30 s, detected through ambulatory heart rhythm monitoring only within the first 5 days after surgery Non-atrial arrhythmias including ventricular tachycardia and ventricular fibrillation detected through ambulatory heart rhythm monitoring only within the first 5 days after surgery In-patient mortality Critical care length of stay Hospital length of stay Costs related to potassium supplementation Quality of care assessed with the EQ-5D-5L questionnaire after 6 months (results not presented or published currently) There was a clear difference in the number of potassium supplements administered between both arms, with more than half of the patients in the relaxed arm receiving no supplementation, while the tight arm received a median of seven doses of supplemental potassium. Despite this marked difference, patients in the tight arm had a serum potassium level that was on average only 0.2–0.3 mmol/L higher than in the relaxed arm. From Day 3, potassium levels dropped slightly below the 4.5 mmol/L target in the tight arms, while in both study arms, the average potassium levels remained well-above 4 mmol/L on every of the first 5 days after surgery. In the intention-to-treat analysis, post-operative atrial arrhythmias occurred in 27.8% of patients in the relaxed group vs. 26.2% of patients in the tight group, which met the criteria for non-inferiority. Results were confirmed in a per-protocol analysis. The incidences of atrial arrhythmias only detected by the ambulatory heart rhythm monitoring (27.8% vs. 33.1%) or by clinical detection (33.1% vs. 33.0%), and the incidence of non-atrial arrhythmias including ventricular tachycardia and ventricular fibrillation (19.1% vs. 21.1%) were very similar in both groups, respectively. There was no difference in mortality up to 5 days with 4 deaths (0.5%) in each study arm. Furthermore, the median length of critical care stay at 2 days and hospital admission at 6 days was identical in both groups. Because of a higher use of potassium supplements in the tight group, the costs for these supplements were increased four-fold in the tight group. The TIGHT-K trial was a well-designed and well-executed trial with a clear hypothesis and a clinically relevant question. The trial included the intended population, which was a typical CABG population, although there might have been an underrepresentation of women (only 15%). Patients were at low risk with a EuroSCORE II of only 1.5. The investigators did not include patients with concomitant valve surgery to ensure a homogenous population with results that could be clearly interpreted. Whether these results translate to higher risk populations therefore remains unsure. There was a high use of beta-blockers in 78% of patients before cardiac surgery. The use of beta-blockers does reduce the incidence of post-operative atrial tachyarrhythmia, which might have influenced the study results.4 In fact, the main paper of the TIGHT-K trial reports a subgroup analysis, yielding an interaction P-value of 0.06 for the baseline use of beta-blockers.1 Although this interaction is strictly speaking not statistically significant, the study was likely underpowered to assess this. While the point estimate of −0.4% indicated absolutely no difference of tight potassium supplementation in patients on beta-blockers, there was a 9.4% risk difference in favour of tight control in patients not treated with a beta-blocker. As a consequence of the study design, most patients included had normal to high potassium levels immediately after surgery with a mean serum potassium level of 5.0 mmol/L. Therefore, potassium levels dropped <4 mmol/L in only a small minority of patients, even in the relaxed group, with a very small overall difference between the intervention and control groups (i.e. 0.2–0.3 mmol/L difference in serum potassium levels). This might have biased the trial towards a neutral result and confirmation of non-inferiority. A better approach might have been to randomize only patients with baseline levels <4 mmol/L to ensure a more meaningful difference between the study arms. With the results of the TIGHT-K trial, we can now conclude that aggressive potassium repletion is not routinely indicated for low-risk patients undergoing cardiac surgery. Serum potassium levels down to 4 mmol/L do not seem to be associated with worse outcomes in terms of arrhythmic events, especially not when concomitant beta-blocker therapy is applied. Future studies may focus on whether low potassium levels <4 mmol/L are still associated with increased risk and might be enriched with patients who have not an indication for beta-blocker treatment. None declared. No new data were generated or analysed in support of this research.