Most likely one of your variable is int32. As a result, all the other variables are implicitly converted to int32. Convert it to double before concatenating it.
concat = [dblvar, dblvar, double(int32var)];
rant: Most languages have made the sensible decision that implicit conversions are widening. That is, an implicit conversion will only happen if there is no loss of data (e.g. int8 -> int32, float->double or int->double). Matlab, on the other hand has narrowing conversions. An int32 will get demoted to an int8 with all the loss of data that ensues. Same with double->float or as in your case double->int. Completely absurd!
