Previous reports have suggested that high-dose L-arginine could be used in diabetic patients as a prophylactic blocker for the initial glycation reaction of proteins by methylglyoxal (MG), a reactive dicarbonyl compound of glucose metabolism. Here, we present several lines of evidence to substantiate that this prophylactic intervention may be inappropriate and should be used with care. First, we demonstrated that when various concentrations of L-arginine (2.0–8.0 mM) were added to a fixed concentration of MG (1.56 µM) in a buffered lucigenin solution, dose-dependent generation of superoxide anion (O–2)-mediated ultraweak chemiluminescence (uwCL) occurs. The suppression of uwCL generation by exogenously added superoxide dismutase further substantiated that the interaction between MG and L-arginine generated O–2. This phenomenon can also be demonstrated in a serum-based system. Furthermore, when a fixed concentration of L-arginine (8.0 mM) was added exogenously to a group of sera obtained from either diabetic patients (n = 10) or their matched nondiabetic controls (n = 10), a marked discrepancy in the generation of O–2-mediated uwCL could be demonstrated (12,534 ± 3,147 vs. 950 ± 350 counts; p < 0.001). Taken together, this evidence demonstrates that the appropriateness of using high-dose L-arginine for prophylactic measures in diabetic patients may be questioned, because the inhibition of the glycation reaction between MG and proteins by high-dose L-arginine unexpectedly produces plethoric O–2 as a by-product, which may subsequently aggravate the preexisting oxidative stress status of diabetic patients.