Albumin to globulin ratio, a predictor or a misleader?

We appreciate the letter to the editor by Alkan et al. [1] in response to our recently published article [2]. Alkan et al. bring up some potentially valid points about our study. They suggest rheumatic diseases should have been excluded from the study population, as these conditions are ‘important causatives for chronic inflammation.’ They also suggest the inclusion of elderly subjects older than 60, comprising 24.2% of the study population, may be misleading to the results, as these subjects are ‘prone to malnutrition and also hypoalbuminemia.’ In order to address these suggestions, we have investigated the association of low albumin-to-globulin ratio (AGR) with cancer after having further excluded (in addition to the exclusion criteria originally implemented in our study) elderly subjects older than 60, as well as subjects with previous history of major rheumatic diseases, including rheumatic arthritis (ICD-10 M05), systemic lupus erythematosis (M32), ankylosing spondylitis (M45), systemic sclerosis (M34), dermatopolymyositis (M33), other connective tissue disease (M35–M36), as well as Crohn’s disease (K50), and ulcerative colitis (K51). The study population thus decreased from 26 974 to 20 695 subjects. The results are shown in Table 1. As expected, event numbers have significantly decreased (∼70%) but the overall trend is shown to be apparently intact, with adjusted hazard ratios remaining similar in general throughout all AGR groups. Overall, the association between low AGR and cancer is a legitimate observation, although its clinical application remains undetermined, deserving elucidation by future studies.