So which cardiac patients should get an ICD? The first recommendations are clear and well established. If a patient has ventricular fibrillation (VF) or ventricular tachycardia (VT) associated with cardiac arrest or sustained VT (i.e. >30 seconds) with syncope or haemodynamic compromise (excluding people with a treatable cause e.g. acute myocardial infarction) or asymptomatic sustained VT and an ejection fraction of <35% and they are not in NYHA class IV heart failure then they should have an ICD. We all agree with that. Putting it simply survivors of VF/VT cardiac arrest or near cardiac arrest and people with sustained VT and a severely impaired left ventricle should have an ICD. However this is a relatively small group of patients compared to the large group with heart failure who if they do have VT it is usually the non-sustained and asymptomatic type. So what does NICE say with regards to them?
Here the guidelines get more difficult. NICE say that an ICDs or CRT with defibrillator (CRT‑D) or CRT without defibrillator (CRT‑P) are recommended for people with heart failure who have an EF of <35% depending on their QRS duration on ECG. If the QRS is normal (ie <120msec) then they recommend an ICD if there is a "high risk" of sudden cardiac death. The rest of the table summaries the other recommendations based on the QRS duration and the NYHA class of heart failure.
Perhaps it time for some pragmatism. You want to implant an ICD in a patient at high risk of sudden cardiac death but you don't have a good way of identifying these patients. The outcome you are trying to prevent is very serious, occurs suddenly and unpredictably. If it happens you don't usually get a second chance to treat it. This means your strategy for prevention will need to be deployed with a very high sensitivity to treat and by definition you will need to accept a lower specificity. Thinking of it a different way if the treatment was a tablet you would put everyone on it because the thing you are trying to prevent is so serious and you only get one chance to do it. That must be the only sensible approach. For people with EF<35% the question is Why shouldn't they have an ICD?