What findings would help determine whether worsening edema, fatigue and shortness of breath after stopping digoxin were caused by the withdrawal rather than the underlying heart disease?

The short version: the evidence does not separate withdrawal from disease progression because the studies were designed to show that stopping digoxin worsens outcomes, not to distinguish the two causes.

The 2015 study [1] is the closest to your question. It took 30 patients with stable heart failure on optimal therapy, measured BNP and six-minute walk distance while on digoxin, then again after withdrawal. BNP rose from 405 to 604 ng/L (p=0.02) and walk distance dropped from 474 to 455 meters (p=0.02). Those changes happened over the study period, not over months. The Minnesota Living with Heart Failure scores actually improved after withdrawal (29 to 25, p=0.06), which the authors note may reflect avoiding digoxin side effects. That study does not tell you whether the worsening was withdrawal or disease, because every patient had heart failure. The worsening was measured against their own stable baseline, not against a group that never had the disease.

The older withdrawal trials tell the same story. The PROVED and RADIANCE trials from 1993 [3] showed that patients who stopped digoxin had higher rates of worsening heart failure than those who stayed on it. But those patients were all on diuretics and ACE inhibitors. The worsening was measured as a composite of treatment failure, not as a biomarker that could separate drug effect from disease progression. The DIG trial [3] showed that digoxin reduced hospitalizations but did not change mortality. That means the drug improves symptoms without changing the underlying disease trajectory. When you stop it, the symptoms return to wherever the disease is at that moment. If the disease has progressed, the symptoms are worse. That is not withdrawal. That is the disease.

The 2014 systematic review [2] looked at stopping RAAS inhibitors and beta-blockers, not digoxin. It found that stopping those drugs increased hospitalizations. That pattern is consistent: stopping any drug that was controlling symptoms will unmask the disease. But the question of whether the worsening is withdrawal or disease progression is a question the literature never designed a study to answer. You would need a trial that randomized patients to continue digoxin, stop digoxin with a placebo, or stop digoxin and add a different inotrope, then tracked whether the worsening was immediate or gradual. That trial does not exist.

The 2012 AFFIRM analysis [4] found that digoxin use was associated with higher all-cause mortality (adjusted hazard ratio 1.41) in patients with atrial fibrillation. That is not about withdrawal. It is about the drug itself. If digoxin increases mortality while on it, stopping it might reduce that risk, but the underlying disease still needs management. The 2015 BMJ meta-analysis [5] confirmed that digoxin has a neutral effect on mortality in randomized trials (RR 0.99) but reduces hospitalizations (RR 0.92). That is a symptomatic benefit, not a disease-modifying one.

My call: the evidence does not support a distinction between withdrawal and disease progression because the studies were never designed to make that distinction. The worsening after stopping digoxin is real, but it is the disease reasserting itself, not a separate withdrawal syndrome. The timeline is days to weeks, not months. If the decline happens over months, that is the disease, not the drug. Confidence is moderate because the studies all compare digoxin to placebo in heart failure patients, not withdrawal to a control group without the disease.