People with 45 chromosomes (or to generalize to other species: individuals with mismatched chromosome counts as a result of evolutionarily recent chromosome fusion or splitting) do not generally have a high chance of being sterile. They may have a high chance of some kind of reduced fertility, though even that is not clear. There's a major detection bias here since almost no one ever gets karyotyped unless they believe they have fertility problems in the first place. This inflates estimates of how often these chromosome mismatches cause fertility problems.

The answer to the question of why any mutation that has any negative effect on fertility would spread is, as others have said, random success. In evolutionary terms, this is called genetic drift. Genetic drift is sometimes thought of as affecting neutral variation that doesn't have either positive or negative effects. However, it's been well understood from the beginning of genetic drift research that what really matters is the "strength" of genetic drift versus the "strength" of natural selection. Many things can make genetic drift stronger, like a small population or an expanding population or pops where some individuals reproduce more than others, etc. If enough of these drift exagerating factors are found in a population (as they often have to humans), then even variation with pretty strong negative effects can still spread. Beneficial mutations can also be lost in the same way.

>If a member of a species was born with an extra chromosome, or two chromosomes fused, their offspring have a high change of being sterile. How could the increase of decrease of a chromosome become wide spread in a species if that happens?

I think I understood the question and answered it. 1. The sterility of an offspring with an additional / fused chromosome isn't that high as shown by examples, it can be neutral; and 2. neutral mutations can become fixed.

The argument is quite similar to mutations in general. There is the general opinion that mutations are bad and overwhelmingly deleterious. They aren't. Most are neutral; the result is most people have mutations, often quite a few. Those mutations can become fixed simply because there are so many of them and they are not selected out. There are certainly bad mutations, which cause developmental or functional problems. They are sometimes really bad and really obvious, and people remember those. Sometimes they are good and increase selection.

Similarly, sometimes chromosomes fuse or split, and it doesn't make a difference. Sure, sometimes, in fact more often than not, they are bad and get selected out. But sometimes they are neutral, and sometimes the different number gets fixed. This is not unexpected.