Does the frequency of respiratory tract infections help to identify humoral immunodeficiencies in a primary health-care cohort?