From the datasheet the maximum value for gain register is 0xFFFF (2^16-1). According to the equation on page 41
Code = D * (M + 1) / 2^16
So for 0xFFFF the code value is D (2^16-1+1)/2^16
So any number other than 0xFFFF will make the adjusted value smaller. In other words we only can trim the output value smaller but not larger with gain register. Is it correct?
Thanks for help!