Endocrine Abstracts

Vitamin D deficiency is common in patients with chronic renal failure (CRF) based on routinely determined total-25-hydroxyvitaminD (t-25OHD) levels. The present knowledge is contradictory whether the direct measured free-25OHD (dm_f-25OHD) or the calculated free-25OHD (c_f-25OHD) is the best indicator of the vitamin D supply for CRF patients. Adding methods to our previously described first experiences on this subject, in this study we aimed to serve newer data to make this question clearer.

Method: 95 patients [39 men; 56 women; 69 ± 12 years: 30 on peritoneal dialysis (PD) 34 on haemodialysis (HD), both on 3000 IU D₃/day; 31 pre-dialysis (preD) on 1400 IU D₃/day] were investigated. Their sera were analysed for vitamin D-binding protein (DBP) (Dako), albumin (immunoturbidimetry, Integra), Ca, PTHi (Roche) dm_f-25OHD (ELISA, FutureDiagnostics), t-25OHD [using two methods: liquid chromatography/mass spectrometry (t-25OHD_MS) and chemiluminescence immunoassay (t-25OHD_CLIA; DiaSorin]. The c_f-25OHD levels were calculated from both t-25OHD (c_f-25OHD_MS; c_f-25OHD_CLIA). Differences between the methods were evaluated by Passing&Bablok regression and Bland & Altman analysis.

Results: Albumin levels were the lowest (PD:33 ± 4 vs HD:37 ± 4; preD:39 ± 4 g/l; P < 0,001) and DBP concentrations the highest (PD:351 ± 39 vs HD:307 ± 48 mg/l; P < 0,001) in PD group. The t-25OHD_CLIA levelswere lower compared to t-25OHD_MS t-(73.4 ± 22.4 vs 106.2 ± 30.5 nmol/l; P < 0,0001) in all groups, but the greatest bias was in PD patients (37%; test of linearity: P = 0.660). Free 25OHD levels obtained by calculation were overestimated compared to dm_f-25OHD concentrations in all three groups (all groups: dm_f-25OHD: 13.2 ± 4.5 vs c_f-25OHD_MS: 26.4 ± 11.4 and c_f-25OHD_CLIA: 20.0 ± 8.2 pmol/l; P < 0.001). The negative biases were significantly higher in case of c_f-25OHD_MS in all groups compared to c_f-25OHD_CLIA (preD: −53% vs −35,2%; PD: −67% vs −31%; HD: −45% vs −66%) but the test of linearity showed the best correlation between direct measured and c_f-25OHD_MS in PD (P = 0,999). All 25OHD fractions were significantly lower (P < 0,010) in PD compared to HD (t-25OHD_CLIA:73.4 vs 94.8 nmol/l; t-25OHD_MS:106 vs 122 nmol/l; dm-f-25OHD: 11.9 vs 14.9 pmol/l; c_f-25OHD_CLIA:24.2 vs 31.3 pmol/l) though the patients received the same dose of vitamin D3. Out of the five obtained 25OHD levels only dm_f-25OHD showed significant positive correlation (r = 0.39) with Ca and only in preD patients.

Conclusions: Our results highlight that determination of cut-off values that reflect the appropriate vitamin D supply should be defined according to the diseases and the methods as well. Calculated free levels overestimate the 25OHD supply. The dm_f-25OHD promises to be the most reliable marker in preD. All five methods certify that patients on PD need much higher doses of cholecalciferol.