That just sounds like pancreatitis, not salmonella and not something that would be caused by contaminated kibble.
When an animal (or human for that matter) becomes ill (nausea/vomiting is the biggest culprit), it can elevate certain liver enzymes which when tested point to a pancreatic episode. A lot of times a patient will be diagnosed with pancreatitis (acute) or chronic if the liver gets too overloaded and a harder recovery seems eminent (long term illness which would need to be managed).
Salmonella poisoning could easily be overlooked if not specifically tested for and the patient given a diagnosis of pancreatitis since the symptoms are very similar and if testing would show elevated liver enzymes which would fit a pancreatitis diagnosis. Profuse vomiting alone can often times cause the ALT and amylase to elevate but stopping the vomiting and re hydration bring the numbers down. The liver is capable of "healing" fairly quickly (lowering the elevated numbers within days) but often go unchecked that soon after a diagnosis of pancreatitis is made (in canines). This would be helpful in making an accurate diagnosis more often!
I guess what I am trying to say is.....
A dog with salmonella poisoning might very well be sick enough to alter their liver enzymes so that a pancreatitis diagnosis is made. If it were not a terribly serious case of salmonella poisoning (affected their heart, they became anemic, etc...) treatment for the pancreatitis (removing food for 24-48 hours among other things/meds) might very well be helpful. Typically a change in diet is warranted which might involve removing the salmonella source without knowing it.
I hope some of this makes sense. My DH has suffered from pancreatitis for about 20 years and every time he has an attack that lands him in the hospital they always want to test for salmonella (and then some) even though we tell them he has pancreatitis!