Because prior to the development of nuclear reactors, you could not easily produce radioactive poisons in quantity, and the first reactors were used to make fuel for nuclear weapons, a much better use of their output from a military standpoint.

To put it into perspective: the early unit of radioactivity, the Curie, is the amount of activity in 1 gram of radium. In 1921, the United States honored Marie Curie by pooling the cash necessary to get her a single gram of radium for her to do research with — it was worth $100,000 at the time (maybe $10 million USD today). They weren't "creating" that gram — they were extracting it from uranium ores. With perfect efficiency you can get about .14 grams of radium out of 1 ton of uranium ore — so they had to process some 7 tons of ore to get that radium (and probably more than that given inefficiencies). Not easy, not cheap.

You can also make radioactive materials with a particle accelerator. Prior to World War II, accelerators were still fairly rare and expensive. (And I haven't done the math, but I imagine that making military-significant quantities of radioactive isotopes is probably a time-consuming endeavor.) I guess you could imagine a military group running such things round-the-clock to make radioisotopic poison stockpiles... but it didn't happen.

The invention of nuclear reactors changed that dramatically. A 100 KW reactor can produce hundreds of thousands of Curies worth of fission products per day. The both peaceful and military possibilities of this were recognized immediately (you can use radioisotopes for peaceful things as well, but there is also the dirty bomb issue). In fact, there were fears that the Germans might use such weapons during D-Day, as it would be much easier than making an atomic bomb (it turns out the Germans didn't get their reactors working).

So nuclear reactors, weapons, and dirty bombs all pretty much emerged due to the same technologies and at the same time.

Separately, it is worth noting that radioactive poisons are pretty difficult to deal with practicably, and are of uncertain and dubious effectiveness in military situations. If your goal is to poison people, prior to but even after reactor development, there are many other ways to do that much more easily (like chemical weapons). They have only had very limited appeal to states; their major appeal is to terrorists, less because of their acute damage, more because of their long-term terror function.