Assumption of no continuous between two stochastic losses

Observing the cascade step function raises a question. Calculating the point of the next interaction, there must be an assumption or a function i don't see.

In general the next interaction point will be calculated out of the mean free pathlength using an exponential distribution. However the pathlength is a function of current energy, which changes through continuous energy losses. If the energy cut is not negligible small, the particle loses significantable energy traveling to the next interaction point but approximated with initial energy the whole continuous steplength.

InverseGrammageType const total_inv_lambda = fProcessSequence.GetTotalInverseInteractionLength(vParticle);
corsika::random::ExponentialDistribution expDist(1 / total_inv_lambda);
GrammageType const next_interact = expDist(fRNG);

Does CORSIKA assume, that continuous energy losses don't make a measurable effect calculating the interactionpoint or do i only not find the method which takes this into account?