The article was a press release, which is basically the same as an advertisement. It is not the same thing as the study itself which we can't evaluate not having seen it. Legitimate scientific studies are **published** in respected "peer-review" journals such that other scientists can replicate the study and hopefully get the same results. Sometimes they can't duplicate the results and when that happens, the original study become highly suspect.
Well constructed studies, rigourously performed and published in a reputable journal have genuine value, but that value is still limited until the results can be replicated.
The study (as quoted) makes no claim about improving the lifespan of dogs, it simply claims to reduce DNA damage as a result of poor nutrition, but that by itself is not proof of having an effect on longevity. A long term study might find that dogs don't live longer eating this food, and if true, then all the study has shown is a reduction in DNA damage, a useless result if it doesn't improve lifespan.
As an example, a major new heart medication for humans that *should* have increased the lifespan of diseased humans was pulled when it was shown to have the opposite of the intended effect, people were dying sooner, not later.
Just because a drug or treatment "should logically work" does *NOT* mean that it will work.
The press release also contains claims not supported by the study itself: to wit, a supposedly objective person (New Zealand Kennel Club President Ray Greer) when he said "It will also give more years of valuable companionship to the many pet owners in this countryâ€ (his opinion). Again while the study as quoted *implied* that dogs will live longer (by using the technique of saying, "because A is true and B is true that therfore C must be true") the study didn't actually say that dogs would live longer, it just said that DNA damage was less (and therefore dogs *should* live longer) but as I've said above, there can be a big difference between "should" and "does".