Dailey I am here again to help you with your delightful biological schemings! Jumping off from Separatrix' fine bat diagram I have figured it out.
Your question: "What point would elbow spikes serve?" is not relevant to dragons, because the spikes are not at the elbow. Allow me to explain.
First the bat.
Smaug has an additional finger in his wing proximal to the spike on his "elbow"!
This solves the problem: the "elbow" on dragons is not the juncture of the humerus and radius/ulna, but rather a wrist-equivalent. The humerus, radius and ulna are condensed into short, powerful, proximal structures as happened in the whale and icthyosaur. This fact leads to the conclusion that the dragon wing is in fact derived from a flipper, and dragons are evolved from aquatic ancestors.
The role played by the radius and ulna in bats and birds is played by elongated carpal bones in dragons - a long "wrist". The evolutionary equivalent of fingers spread from this site to form distal webs similar to those in the bat as well as proximal webs as seen in Smaug and related dragons. Additional digits not used in the webs persist as "spikes" or nearly formed hand-like structures (like Smaug has).
That gives Smaug more than 5 fingers, you may protest. But that is OK. Dragons are polydactylous. Polydactylous cats get by just fine.
ADDENDUM @KSmarts comment made me think that it worth adding an image showing how a dewclaw could be very medial, as is the case for a dog. In the dragon that dewclaw oriented phalange would not be reduced to just one bone and would extend out to be the medial wing finger.
from https://www.joshuanava.biz/human-figure/paws-and-hind-legs.html with my text box added