An Assessment of Automated Vitamin D Measurement Methods Including a Point-of-Care Testing Method, I-CHROMATM Using the Randox International Quality Assurance Scheme (RIQAS)

25Hydroxy vitamin D (25-OH), more commonly referred to, as Vitamin D is made up of a group of anti-rachitic fat-soluble 9,10-secosteroids. It is synthesized in the skin when the body is exposed to UV sunlight, and vitamin D deficiencies have been linked to several pathologies including osteoporosis and rickets, cardiovascular disease, cancer, depression and recently, schizophrenia. There is an increasingly high prevalence of people suffering from Vitamin D deficiency worldwide [1-2]. There has therefore been an increased demand for reliable automated Vitamin D assays. Vitamin D is usually measured using immunoassays, HPLC and liquid chromatography tandem mass spectroscopy (LCMS/MS) [3-5]. Various manufacturers including Diasorin, Abbott, Siemens, IDS SYS and Roche have more recently come into play as automated assay providers for the analysis of Vitamin D. There have been many studies carried out that have compared these assays to the traditional assays, and it seems that they are still under scrutiny for their accuracy. It has been reported that some of the assays such as the Siemens Advia Centaur tend to over-estimate Vitamin D [57], whilst the Abbott Architect and the Roche Elecsys methods tend to under-estimate [5,7].


VITAMIN D Concentration Estimation
The assay was performed following the manufacturer's instructions: a.
Place the test cartridge into the i-chamber slot, which has been set at 35oC.

b.
Transfer of 50µL of releasing buffer into a sample mixing tube and add 50µL of sample to sample mixing tube containing releasing buffer and mix well by pipetting 10 times. c.
Insert the sample mixing tube into the inserting block in the i-chamber and leave the tube in the inserting block at 35oC for 5 mins. d.
Add 100µL of detection buffer to the sample mixing tube containing the releasing buffer and sample mixture. Mix well and leave in the inserting block again at 35oC for 15 mins.

e.
Draw 75 µL of incubated material into the sample well on the test cartridge and push into i-Chamber and leave for 8 mins. f.
Immediately insert the test cartridge into the cartridge holder of the i-CHROMA™ reader and press the 'Select' button and read the result on the display screen.
The Randox international quality assessment scheme used by 40,000 laboratories worldwide. Helps laboratories to meet their quality requirements and provide best patient care. The ALTM was calculated by taking the average of all Vitamin D samples 1-36 obtained from using all vitamin D methods. The R2 value was calculated by taking the average of the vitamin D methods compared with the ALTM. The Bias was calculated by taking the average of all the differences between samples using the vitamin D method vs the ALTM.

Results
From the RIQAS information provided, the average number of laboratories using Abbott Architect, Roche Total Vitamin D, Roche Elecys, Siemens Centaur XP/XPT/Classic, Roche COBAS® 4000/ e411, Ortho Vitros, Diasorin Liaison, IDS SYS ELISA, Roche Modular and Beckman 600/800 vitamin D methods that submitted returns for the distribution samples were 35, 25, 10, 13, 18, 5, 14, 2, 3, 3 respectively indicating that the Abbott Architect and the Roche Vitamin D Total are the commonest methods used with the IDS SYS being the least common method used ( Figure 1). The mean Vitamin D results from all ten registered Vitamin D methods was calculated: Abbott Architect (n=24), Roche Vitamin D Total (n=24), Siemens Centaur XP/XPT/Classic (n=24), Roche Elecsys (n=24), Diasorin Liasion (n=24), Roche COBAS® 4000/e411 (n=24), Ortho Vitros 3600/5600/ECi (n=23), IDS SYS (n= 12), Roche Modular (n=21) and Beckman 600/800 (n=17)( Table 2).  The mean estimates of the following vitamin D methods were lower than the ALTM: Abbott Architect, Roche Cobas vitamin D Total, Roche Elecys, Roche Cobas 4000/e411, while the mean estimates of the following methods were higher than the ALTM: Siemens Centaur XP/XPT/Classic, Ortho Vitros, Roche Modular and Beckman 600/800.The mean estimates of the following methods were comparable to the ALTM: Diasorin Liaison and IDS SYS (   Although, the Roche Vitamin D Total, Siemens Centaur XP/XPT/ Classic, Roche Elecys, Roche Cobas 4000/e411, Roche Modular Vitamin D methods achieved a good correlation with the ALTM of the RIQAS apart from Ortho Vitros methods. They did not meet the criteria of a very good correlation (r2> 0.9) and bias (+/-15% difference). The laboratory methods with the best combination of a very good correlation (r2> 0.9) and bias (+/-15% difference) with the ALTM in the RIQAS were the Abbott Architect, Diasorin Liaison, IDS SYS and Beckman 600/800 vitamin D methods. The ALTM compared well with the new point of care test Boditech i-CHROMA™ Vitamin D method: (r2=0.7) (Figure 3) with a reasonable bias of -9.97nmol/L (Figure 4).

Discussion
Based on the number of laboratory vitamin D methods that were submitted to the RIQAS, the Abbott Architect and Roche Vitamin D Total methods constitute the most common methods, 28% and 20%. The IDS SYS, Roche Modular and Beckman 600/800 were the least common methods making up 2% each of the total methods. Similar correlations were also observed in three other comparative studies of automated vitamin D methods compared with LC-MS/MS methods: Farrell et al. [6,7] Abbott Architect, DiaSorin Liaison, IDS SYS, Roche E170, Siemens Centaur [8]; Janssen et al. [9] DiaSorin RIA, Diasorin Liaison, Roche Cobas Total Vitamin D, IDS ISYS, Siemens Advia Centaur, Abbott Architect i1000 and i2000 methods [9] and Freeman et al. [10] Siemens Advia Centaur, Diasorin Liaison, Roche Elecys and Abbott Architect methods [10]. In this study using the RIQAS submissions compared with the ALTM, the methods with least biases (<10%) were the Beckman 600/800 method 1.08nmol/L (2.63%), the IDS SYS method -2.18nmol/L (-5.14%), DiaSorin Liaison -2.60nmol/L (-5.78%) and the Abbott Architect 4.32nmol/L (9.59%). In the Farrell et al. [6,7] study, the DiaSorin Liaison method also had the smallest bias (-0.5nmol/L) compared to the LC-MS/MS method [8]. The Janssen et al study showed that the DiaSorin RIA method had the smallest bias (0.3nmol/L) compared to the LC-MS/MS method [9].
In addition, the Freeman et al. [10] study, the Siemens Advia Centaur method had the smallest bias (-6.8nmol/L) [10]. The methods that showed the largest biases were the Ortho Vitros3600/5600/ECi method -23.67nmol/L (-53.49%), Siemens Centaur XP/XPT/Classic -21.62nmol/L (-48.05%), Roche COBAS® 4000/e411 14.94nmol/L (33.19%), Roche Elecsys 14.64nmol/L (32.54%) and Roche Modular 12.65nmol/L (28.96%). In the comparative studies with LC-MS/MS, the methods with the largest bias were the Abbott method bias (11.4nmol/L), the Siemens Advia Centaur method bias (-10.7nmol/L) and Abbott Architect method bias (-18nmol/L) [8][9][10]. The observations of the RIQAS assessment were not too far from the data observed in the comparative studies, although some of the comparative studies did not cover as many methods. The DiaSorin Liaison vitamin D method was the most consistent method with the smallest bias in the RIQAS evaluation and the other comparative studies compared to LC-MS/MS method or ALTM.
The Siemens XP/XPT/Classic method had the largest bias in the RIQAS evaluation compared to the ALTM but the Abbott Architect method appears to be the method with the largest bias in the comparative studies to the LC-MS/MS. In the introduction, we described that assays such as the Siemens Advia Centaur tend to over-estimate Vitamin D [5][6][7], whilst the Abbott Architect and the Roche Elecsys methods tend to under-estimate [5,7]. The Diasorin Liaison has been described to both over-and underestimate Vitamin D [5]. In this RIQAS evaluation, we confirmed these observations as the Siemens Centaur XP/XPT/Classic method, reported a mean value higher than the ALTM, while Abbott Architect and the Roche (Vitamin Total, Elecsys, Cobas 4000/e411 and Modular) methods all reported means lower than the ALTM. The DiaSorin Liaison method mean was similar to the ALTM.
In the RIQAS evaluation, the laboratory methods with the best combination of a very good correlation (r2>0.9) and bias (+/-15% bias difference) with the ALTM were the Beckman 600/800, Abbott Architect, Dia Sorin Liaison and IDS SYS vitamin D methods. In a study, of 170 randomly selected patient samples and Vitamin D was measured using a similar number of laboratory methods, Farrell et al. [6,7] concluded that the DiaSorin Liaison, the IDS SYS and the Siemens Centaur methods were the 3 methods that met the minimum performance goal for the measurement of vitamin D [8]. In a recent review article, the DiaSorin Liaison and Abbott Architect methods were described to have performed well based on a 2009 proficiency testing survey, these methods failed to perform as well in the 2012 proficiency survey, where the IDS SYS method performed well [11].

Conclusion
In conclusion, using the data from the RIQAS evaluation, which was compared to the ALTM, and the most extensive Vitamin D method comparison study which was compared to LC-MS/MS [8], there appears to be some concordance that the best performing Vitamin D methods are the Dia Sorin Liaison and the IDS SYS. The other methods that follow are the Abbott Architect, Siemens Centaur and the Beckman 600/800. Point of Care Tests (POCT) are becoming quite common and in this evaluation, a new POCT method, Boditech i-CHROMA™ Vitamin D method provides us with some insight into how rapidly this field is growing. The Boditech i-CHROMA™ Vitamin D method compared well with the ALTM (r2=0.7) but not as good as the other methods apart from the Ortho Vitros 3600/5600/ECi method. The Boditech i-CHROMA™ Vitamin D method also had reasonable bias of -12.84nmol/L (-29.58%). This method therefore still requires further investigation in the clinical setting.