The prognosis of patients following myocardial infarction is adversely
affected by the finding of late potentials at the time of hospital di
scharge. Loss of late potentials has been previously reported during s
erial testing during the first year after infarction, but it is not kn
own whether such patients remain at risk of arrhythmic events. This st
udy prospectively followed 243 patients after myocardial infarction. L
ate potentials were observed in 92 patients (group I) at the time of h
ospital discharge. Of these patients, 23 no longer had late potentials
at 6-week follow-up and 8 had had an arrhythmic event (sudden death o
r ventricular tachycardia). In patients with loss of late potentials,
overall QRS duration had decreased from 109 +/- 11 msec at discharge t
o 104 +/- 11 msec (P < 0.01), terminal QRS voltage rose from 15 +/- 4
muV to 31 +/- 9 muV (P = 0.001), and late potential duration fell from
42 +/- 6 msec to 28 +/- 6 msec (P = 0.001) at the 6-week study. Predi
ctors of loss of late potentials were: initial duration of the QRS dur
ation (P < 0.001) and terminal voltage (P < 0.005); non-Q wave infarct
ion (P < 0.001); and being a male (P < 0.05). After the 6-week assessm
ent, 11 additional arrhythmic events occurred during median follow-up
of 31 months. The risk of arrhythmic events was similar in patients wi
th loss of late potentials and those who retained late potentials in g
roup I (9% vs 11%, P = NS) but significantly greater than patients wit
h no late potentials at discharge (group II, 2%). Of those patients wi
th events beyond 6 weeks, a normal signal-averaged ECG (either lost la
te potentials or group II) was observed in 6/11 (55%) patients on at l
east one occasion prior to the occurrence of the event. Hence, a signi
ficant number of arrhythmic events occurring greater-than-or-equal-to
6 weeks after myocardial infarction occur in patients with a normal si
gnal-averaged ECG even when late potentials are initially present. ''L
oss'' of late potentials does not necessarily confer an improved progn
osis in terms of risk of arrhythmic events.