There is a lack of consensus regarding the most appropriate specimen type for analysis of many biochemistry analytes. The aim of this study was to compare renal and lipid analyte profiles and phenytoin values in plain serum (S), serum gel (G) and plasma (lithium heparin, P) tubes and to investigate the stability of these analytes after prolonged contact with cells or gel at room temperature (RT, 20 degrees C) and as aliquoted and stored at 4 degrees C. Primary specimens were centrifuged once, maintained at RT and analysed within 2 h (T(0)) and after 24 h (T(24)) and 48 h (T(48)). For assessment of stability at 4 degrees C, two cell-free aliquots were separated from each of the primary tubes and stored at 4 degrees C and then analysed at T(24) and T(48). Differences in analyte concentrations between tubes at T(0) and following storage (at T(24) and T(48)) were evaluated for both statistical and clinical significance. At T(0) all analytes, except potassium, demonstrated equivalence between serum, gel and plasma tubes. Potassium and creatinine were more stable in gel tubes than in serum/plasma tubes. In contrast, phentytoin was stable in plain serum and plasma up to T(48) at RT, but showed a progressive and clinically significant decrease in concentration in gel tubes at T(24) and T(48) at RT. All analytes except CO(2) were stable up to T(48) when aliquoted and stored at 4 degrees C. We concluded that the serum gel tube has advantages over plain serum and plasma tubes for measurement of the analytes investigated in this study, with the exception of phenytoin. In practice, the gel tubes demonstrate enhanced analyte stability and reduce the need to aliquot specimens, with greater protection against possible contamination. Further investigation would be required to evaluate a broader range of analytes.