In the US, due to free speech protections (i.e. the first amendment), it seems that the majority of lawsuits for negligent publication fail. It's worth noting that the author might be more liable than the publisher, since the publisher typically isn't expected to have the expertise to catch such errors. More cases and details can be found here.
The only case I've found where a publisher was held liable for negligent publication of a textbook is the following. Note that it's somewhat old, and standards for products liability have likely changed since.
D. Wallis, "Negligent publishing" Implications for University Publishers:
On September 4, 1980, a jury in the U.S. District Court in Springfield, Massachusetts awarded C. Carter $100,000 in her suit against Rand McNally Publishing Company. C. Carter, an eighth grade chemistry student had been severely burned during a chemistry class while attempting an experiment described in a textbook published by Rand McNally. This experiment called for the use of methyl alcohol, a highly inflammable substance. The plaintiff argued, and the jury agreed, that her injuries were caused, in part, by the publishing company's negligent failure to adequately warn students and teachers about the dangerous qualities of these experiments. Rand McNally did not appeal the verdict.
This case is of special interest to publishing companies because it is the first time a book had been considered a "product" and the first time products liability theory has been used to impose liability upon a publishing house for an injury caused by technical information contained within a textbook.
In this case it's not clear whether the disclaimer was present or not. However, there is nothing magical about a disclaimer that would stop lawsuits from occurring, nor do they automatically protect against liability:
The "disclaimer," although it will rarely be effective in shielding the publisher from liability if the court finds that the publisher breached its duty of care to the reader, is more likely to be effective by demonstrating that the publisher did not make any specific guarantees regarding the safety or reliability of the contents of the publication.