On the stability of immunoreactive glucagon in plasma samples