The current study indicates that plasma HO-1 concentration is elevated in individuals with IGR in comparison with healthy controls, and it is significantly associated with odds of IGR. The association between plasma HO-1 concentrations and IGR retains rather consistent under adjustment for possible confounding factors including age, sex, BMI, smoking, alcohol drinking, hypertension, family history of diabetes, lipid profiles and CRP, which makes plasma HO-1 more convincing as an emerging independent marker for IGR.
IGR, also referred to pre-diabetes, has been recognized as a metabolic intermediate state between a “normal” state and T2DM 
. It should be noted that the IGR state has both progressive and reversible properties. IGR people shows approximately 10-fold higher risk of T2DM 
, and on the other hand, the onset of T2DM in individuals with IGR can be delayed or prevented through lifestyle modifications 
. Therefore, it is crucial to elucidate the underlying mechanisms of IGR and to develop biomarkers to understand the role of corresponding mechanisms in human.
Under high glucose exposure, HO-1 gene expression and enzyme activities in the islets are elevated remarkably in parallel with hyperglycemia-induced intracellular peroxide levels 
, prior to the elevation of classical antioxidant enzymes (e.g., superoxide dismutase, catalase and glutathione peroxidase) 
. In addition, HO-1 mediates the anti-inflammatory effect of interleukin-10 through a p38 mitogen-activated protein kinase-dependent pathway 
Recently, it has been found that HO-1 is present and detectable in serum or plasma samples and serves as a systemic stress marker 
. Elevated circulating HO-1 levels are subsequently found in several oxidative stress-related illness conditions, such as in chronic silicosis 
, T2DM 
, acute myocardial infarction 
, coronary microvascular dysfunction 
, Parkinson's disease 
, and critically ill patients 
. Similar to our previous report 
, we found that plasma HO-1 concentrations were significantly correlated with fasting plasma glucose, 2 hour OGTT glucose, HOMA-beta and HOMA-IR in this population. Taken together with our previous study 
, the study shows an association that could suggest that HO-1 is responsive to high blood glucose, even if under moderate hyperglycemia in the form of IGR. The current study also suggests that plasma HO-1 concentration can significantly improve the discriminative value for IGR. If the utility of HO-1 be confirmed in prospective cohort studies, it might be incorporated into established panel of biomarkers 
to help further improve the predictive value for IGR and T2DM.
There are several limitations to this study. First, the current case-control study design could not allow examining the causal relationship between plasma HO-1 and IGR, which remains to be confirmed in further prospective cohort studies. We also cannot directly evaluate the predictive value for IGR, thus we only evaluate the discriminative value instead. Second, we did not have information on diet and physical activity among the participants, while it is possible that individuals with IGR have a different dietary and exercise habits compared to controls. Whether such difference confounds our results remains unknown. Third, since it has been demonstrated that changes in 2-h postload glucose develops years prior to worsening in fasting glucose 
, IFG and IGT might represent not only different types of glucose abnormality, but also different stages along the same line of abnormality. However, we did not perform a separate analysis for IFG group and IGT group, because there were only 21 IFG subjects (8%) in our IGR cases group. Whether there is a significant difference in association between HO-1-IFG and HO-1-IGT merits further investigation. Fourth, diabetic comorbidities may occur in a small portion of patients with prediabetes. Although we excluded those with known cardiovascular disease, it was difficult to eliminate potential bias because elevated circulating HO-1 levels may be a result of diabetic comorbidities. Fifth, all participants in this study were of Chinese Han ethnicity, which minimizes the confounding effects by ethnic background. Whether these results can be generalized to other populations need to be studied further.
In conclusion, elevated plasma HO-1 concentration is significantly associated with increased ORs for IGR. However, its clinical utility should be validated in further studies, especially in prospective cohort studies.