I personally am in the camp of those who think that Lucretius did his dead level best to transmit Epicurus faithfully. Therefore whenever Lucretius says something clearly then I would wager that it reflects what Epicurus taught. But f someone were to ask me for an example of a line in Lucretius that has the most potential for deviating from what Epicurus himself taught, this - especially the part I underlined - would be the line that I would put under the microscope:
Quote from Lucretius 2:216But if they were not used to swerve, all things would fall downwards through the deep void like drops of rain, nor could collision come to be, nor a blow brought to pass for the first-beginnings: so nature would never have brought aught to being.
But before I were to reach the conclusion that this deviates from Epicurus' own views, I would first want to examine the Latin original to see whether there is any chance of a subtlety of translation or other way to reconcile it. I presume that this isn't a scholium addition since it's poetry, but I presume that is not likely due to it being in poem form.)
Certainly like everything else swerves would go back infinitely in time, so there would have been no "first" swerve. As Sedley says it doesn't seem likely that Epicurus would have looked to the swerve as the reason why atoms come together in bodies, but maybe there's something even subtler going on that would have led him to think that it did. I don't gather that the letter to Herodotus is clear as to what it was that causes atoms to "stick" together after their initial collisions? That doesn't seem likely either, but maybe he was developing a theory that whatever causes the swerve has other emergent effects as well.
Are there other alternatives? Maybe Epicurus simply developed the swerve theory later in life and the Herodotus letter was earlier, and never revised, as Sedley seems to think? This calls for creative thinking.