Objectives: We calculated upper 95% reference limits for serum alpha-fetoprotein (AFP) and carcinoembryonic antigen (CEA) according to the CLSI/NCCLS C28-A3 guideline.
Material and methods: Serum samples from 1400 healthy male subjects were collected from the Fangchenggang Area Male Health and Examination Survey (FAMHES). Serum AFP and CEA values were measured by electrochemiluminescence immunoassay on COBAS 6000 system E601 (Elecsys module) immunoassay analyzers.
Results: Serum AFP and CEA values were not normally distributed but log normally distributed. The upper 95% reference limits of the reference intervals were ≤ 4.76 IU/mL (nonparametric) or ≤ 4.56 IU/mL (parametric) for AFP and ≤ 5.57 ng/mL (nonparametric) or ≤ 5.82 ng/mL (parametric) for CEA. The distribution of AFP levels was found to be consistent between the non-smoking and smoking (p=0.740) and non-drinking and drinking groups (p=0.698). The distribution of serum CEA levels was significantly higher in the smoking than the non-smoking group (p<10⁵), whereas there was no significant difference in this respect between the non-drinking and drinking groups (p=0.147). A significant increase with age was found both for serum AFP and CEA levels, and the age-dependent reference intervals were calculated.
Conclusions: The reference intervals for serum AFP and CEA show a slight deviation compared to previously reported reference levels. Distinct reference intervals of serum CEA must be established for smoking and non-smoking populations. In addition, age-dependent reference intervals should be implemented in clinical laboratories.