The CDC has not tracked attacks of any type by breed since 1998:
[P]ost 1998, the CDC stopped tracking which breeds were involved in fatal attacks because, according to a spokesperson, that information isn’t of discernible value. ["Breed-ban talk usually starts with pit bulls." THE NEWS TRIBUNE Saturday, November 13th, 2004.]
On another site the CDC states:
A CDC study on fatal dog bites lists the breeds involved in fatal attacks over 20 years (Breeds of dogs involved in fatal human attacks in the United States between 1979 and 1998). It does not identify specific breeds that are most likely to bite or kill [emphasis mine], and thus is not appropriate for policy-making decisions related to the topic. Each year, 4.7 million Americans are bitten by dogs. These bites result in approximately 16 fatalities; about 0.0002 percent of the total number of people bitten. These relatively few fatalities offer the only available information about breeds involved in dog bites. There is currently no accurate way to identify the number of dogs of a particular breed, and consequently no measure to determine which breeds are more likely to bite or kill."Breeds of dogs involved in fatal human attacks" is occasionally cited as proof that certain breeds bite more often than other breeds: a claim that is impossible to verify, and one that has no relation to the conclusions of this study.
Many practical alternatives to breed-specific policies exist and hold promise for preventing dog bites. For prevention ideas and model policies for control of dangerous dogs, please see the American Veterinary Medical Association (AVMA) Task Force on Canine Aggression and Human-Canine Interactions: A community approach to dog bite prevention.
Further discussion of this study can be found here.