By Michael Erman and Bhanvi Satija

(Reuters) -Bristol Myers Squibb posted much better-than-expected second-quarter results on Thursday on strong sales of its best-selling brands, but failed to convince investors that newer products could drive future growth and shares fell nearly 5%.

Revenue in the quarter, which analysts had expected to fall due to the loss of patent protection on some of its products, including blood cancer therapy Revlimid, rose 1% to $12.3 billion. Analysts, on average, were looking for revenue of $11.4 billion, according to LSEG data.

The quarterly sales beat was driven by legacy products such as the blood thinner Eliquis, which is expected to face competition from cheaper generics next year, and demand for Revlimid, which has held up better than expected, said Raymond James analyst Sean McCutcheon.

However, "investors don't see the beat as an indicator of future growth prospects," McCutcheon said.

Sales of Eliquis, which Bristol Myers shares with Pfizer, rose 8% to $3.7 billion in the quarter, while cancer immunotherapy Opdivo brought in $2.6 billion, up 7% from a year ago. Analysts, on average, had forecast sales of $3.3 billion and $2.3 billion, respectively.

Revlimid sales fell 38% to $838 million, but still topped analyst estimates by about $300 million.

The U.S. drugmaker said it earned $2.9 billion, or $1.46 per share, down from $4.2 billion, or $2.07, a year earlier. Analysts had expected earnings of $1.07 a share.

Bristol has been contending with a steep revenue decline from Revlimid, which raked in nearly $13 billion in 2021 but $5.8 billion last year due to generic competition. Some of its other cancer drugs such as Pomalyst, Sprycel and Abraxane are contending with the same issue.

Still, Revlimid has performed better than initially feared, and the company now expects 2025 sales of around $3 billion, Chief Commercialization Officer Adam Lenkowsky said. The company had previously said it expected 2025 Revlimid sales to be closer to $2.5 billion.

Investors are closely watching Bristol's newer products, including cell therapy Breyanzi and schizophrenia treatment Cobenfy, to gauge whether they can drive the next phase of growth as sales of older blockbuster treatments decline.

McCutcheon said company comments suggest data from a trial testing Cobenfy for Alzheimer's disease may come a bit later than previously anticipated - by the end of this year rather than early fourth quarter.

The company said it was conducting reviews of trial sites, which could delay reporting of results, but Bristol is still targeting by the end of this year.