National Health and Nutrition Examination Survey

Analytical Note for 25-Hydroxyvitamin D Data Analysis using NHANES III (1988-1994), NHANES 2001-2006, and NHANES 2007-2010 (October 2015)

Data Advisory

The purpose of this analytical note is to inform researchers that serum 25-hydroxyvitamin D (25(OH)D) data from NHANES III (1988-1994) and NHANES 2001-2006 have been converted by using regression to equivalent 25(OH)D measurements from a standardized liquid chromatography-tandem mass spectrometry (LC-MS/MS) method. The LC-MS/MS method was used in the analysis of the newly released NHANES 2007-2010 25(OH)D data. This standardization was done so that researchers could use 25(OH)D data that are equivalent to 25(OH)D measurements assayed with a LC-MS/MS method that is traceable to the NIST reference materials. The LC-MS/MS-equivalent data for NHANES III (1988-1994) and NHANES 2001-2006 are needed to compare with 25(OH)D data from NHANES 2007-2010, which were obtained using the LC-MS/MS method. This analytical note replaces the previous analytical note entitled “Revised Analytical Note for NHANES 2000-2006 and NHANES III (1988-1994) 25-Hydroxyvitamin D Analysis (Revised November 2010)”. It is highly recommended that researchers use the LC-MS/MS-equivalent data for all analyses, but especially to analyze secular trends for 25(OH)D involving 1988-1994, 2001-2006 and 2007-2010 data.

Background

Measurements of serum concentrations of 25(OH)D were performed as part of the nutrition biomarker component of NHANES III (1988-1994) and in NHANES 2001-2006.1 Serum 25(OH)D concentrations were measured at the National Center for Environmental Health, CDC, Atlanta, GA using the DiaSorin RIA kit (Stillwater MN). The DiaSorin assay kit that was used to measure 25(OH)D in NHANES III was reformulated by the manufacturer in 1998 (after NHANES III analyses were completed) by introducing: 1) an antibody that provided improved binding; and 2) an altered washing solution to reduce non-specific binding. To assess the magnitude of changes associated with the reformulated assay, which might impact any observed trends in serum 25(OH)D in the population, the CDC laboratory reanalyzed a subset of banked serum samples from NHANES III using the reformulated version of the RIA. The NHANES III results, as measured with the reformulated assay, were regressed on the NHANES III values obtained with the original assay for these reanalyzed specimens. The average difference between the reformulated and original RIA was -12%. An equation was generated after first accounting for the assay drift occurring during analysis of NHANES 2003-2004 using the reformulated DiaSorin assay (see below). This harmonization equation allowed an approximation of NHANES III results to the level of the reformulated assay that was used for NHANES 2001-2006.

In addition to the DiaSorin assay reformulation, the CDC laboratory observed drifts in the serum 25(OH)D assay performance (as reflected in QC pool shifts in the mean, up or down, by up to 10%) over the period of NHANES 2003-2006. The variation in 25(OH)D values appeared to be due to method variation that probably resulted from reagent and calibration lot-to-lot variation. Several approaches were attempted for harmonizing the 2003-2006 25(OH)D. A model based on RIA quality control pool data was selected because the results should be independent of any empirical trend in the sample participant data. As a result, the NHANES 2003-2004 25(OH)D data were generally adjusted to lower values and the 2005-2006 data were generally adjusted to higher values. These RIA-harmonized data, using QC materials to make the adjustment, were publicly released in November 2010. Archived RIA-harmonized data files and documentation are available under “other data” on the NHANES web site (/nchs/nhanes/vitamind/RiaMethod.aspx) so researchers who previously used these data can still access them. However, CDC strongly recommends using the LC-MS/MS-equivalent data for all analyses, unless there is a compelling reason to use the RIA-harmonized data for 1988-1994 and 2001-2006. Use of the LC-MS/MS-equivalent data is strongly recommended for correct interpretation of trends in 25(OH)D data from NHANES III and 2001-2006 when compared with NHANES 2007-2010.

Concurrent with the recognition of quality control issues for the RIA data from NHANES, an increasing number of publications in the vitamin D field began to describe excessive method bias and imprecision in the methods to measure 25(OH)D available at that time. A major effort to develop a reference method, including reference standards, based on LC-MS/MS methodology, was undertaken by NIST. As a result, the 25(OH)D measurement has now been standardized using a LC-MS/MS method, which allows laboratories and surveys to compare 25(OH)D measurements. The CDC decided to develop a LC-MS/MS method traceable to the NIST-reference materials for NHANES, and used this method starting with NHANES 2007-2008 to measure 25(OH)D3, 25(OH)D2, and the C3 epimer of 25(OH)D3. For the CDC LC-MS/MS method, total 25(OH)D (in SI units of nmol/L) was defined as the sum of 25(OH)D3 and 25(OH)D2 and excluded the C3 epimer of 25(OH)D3. However, due to rounding, the sum of 25(OH)D3 and 25(OH)D2 will not necessarily be equal to the 25(OH)D. It is not appropriate to define 25(OH)D as the sum of 25(OH)D3 and 25(OH)D2 in conventional units (ng/mL) since it is not correct to add different analytes using mass concentrations (such as ng/mL) that do not reflect the relative amounts of the analytes. Molar concentrations (used in SI units of nmol/L) have to be used for adding together different measured quantities. The CDC recommends using the total 25(OH)D in SI units (nmol/L) measured directly by LC-MS/MS (variable LBXVIDMS, for NHANES 2007-2010) or predicted LC-MS/MS equivalent (variable LBDVIDMS for NHANES III and NHANES 2001-2006) and converting this quantity to conventional units (ng/mL), if needed (see below).

The CDC LC-MS/MS method has better analytical specificity and sensitivity compared to immunoassay methods, and fixed analytical goals for imprecision (≤10%) and bias (≤5%). Also, standardization of NHANES 25(OH)D data allows comparison of standardized 25-hydroxyvitamin D data to other national surveys, which is a major objective of the Vitamin D Standardization Program (https://ods.od.nih.gov/Research/vdsp.aspx).

CDC 25(OH)D assay, bridging study, and regression equations

The CDC LC-MS/MS method was used for measurement of 25(OH)D for NHANES 2007–2010. The processing of the NHANES 2007-2010 25(OH)D samples was delayed until the CDC developed and verified the LC-MS/MS method. In addition, sera from 1,448 participants with available 25(OH)D RIA data from NHANES 1988–1994 and 2001-2006 were reanalyzed using LC-MS/MS in order to perform a bridging (crossover) study to develop regression equations to predict a NHANES participant's LC-MS/MS-equivalent concentration from their previously measured un-harmonized (original) RIA value. A variety of regression models were evaluated. The CDC decided to provide regression equations for publicly available 25(OH)D data (NHANES 1988-1994, 2001-2002, 2003-2004, and 2005-2006) rather than using regressions based on single-year time periods that corresponded to laboratory assay dates, because that would require restricted access datasets available only in the NCHS Research Data Center.

The final regression equations were selected based on each model's predictive ability to convert RIA to LC-MS/MS-equivalents for NHANES 1988-1994, 2001-2002, 2003-2004, and 2005-2006. There were five separate regression equations (25(OH)D in nmol/L units):

NHANES III (1988-1994): a piecewise linear regression model was used

1988-1994: if RIAoriginal ≤102.2074 then LC-MS/MSequivalent = 1.57548 + 0.8429*RIAoriginal

1988-1994: if RIAoriginal >102.2074 then LC-MS/MSequivalent = 59.2296 + 0.2788*RIAoriginal

NHANES 2001–2002, 2003–2004, and 2005–2006: ordinary least square regression equations were used

2001-2002: LC-MS/MSequivalent = 6.43435 + 0.95212*RIAoriginal

2003-2004: LC-MS/MSequivalent = 1.72786 + 0.98284*RIAoriginal

2005-2006: LC-MS/MSequivalent = 8.36753 + 0.97012*RIAoriginal

Conversion of 25-hydroxy vitamin D SI units to conventional units

For NHANES 2007-2010, the CDC has released total 25(OH)D, 25(OH)D3, 25(OH)D2, and the C3 epimer of 25(OH)D3 in SI units (nmol/L). For NHANES 1988-2006, the CDC has released total 25(OH)D in SI units (nmol/L). To convert SI units (nmol/L) to conventional units (ng/mL), the following are suggested conversion factors:

Total 25(OH)D, 25(OH)D3 and C3 epimer: 1 nmol/L = 0.40066 ng/mL (1 ng/mL = 2.4959 nmol/L)

25(OH)D2: 1 nmol/L = 0.41266 ng/mL (1 ng/mL = 2.4233 nmol/L)

Conversion of SI units to conventional units should be done after any adjustment of the data such as using the regression equations to convert RIAoriginal to LC-MS/MSequivalent of 25(OH)D.

Comparison of unadjusted RIA (original), harmonized RIA, and LC-MS/MS-adjusted 25(OH)D

The following table shows weighted arithmetic means for unadjusted RIA (original), QC-harmonized RIA, and reference material-standardized LC-MS/MS serum 25(OH)D concentrations (nmol/L), by survey, for persons 12 years and older, in order to illustrate the effects of the adjustments:

Survey Sample Size RIA, original RIA, harmonized LC-MS/MS, standardized*
1988-199418,85174.465.362.3
2001-20026,81658.658.662.2
2003-20046,55362.159.562.7
2005-20066,48054.258.361.0

*Concentrations are standardized to NIST reference materials

Footnote

1Three surveys were undertaken between 2001 and 2006, namely, 2001-2002, 2003-2004, and 2005-2006. Serum 25(OH)D RIA data were also collected in the year 2000. These data are only available through the NCHS Research Data Center because of the potential risk for disclosure of confidential information for any single-year data release from any two-year continuous NHANES data cycle. The CDC decided that the NHANES 2000 dataset would not be adjusted to the LC-MS/MS method because this dataset is not publicly available, and it also requires special sample weights and special primary sampling units.

References

Yetley EA, Pfeiffer CM, Schleicher RL, et al. NHANES monitoring of serum 25-hydroxyvitamin D: A roundtable summary. J. Nutr. 2010;140:2030S-2045S.

Phinney KW. Development of a standard reference material for vitamin D in serum. Am J Clin Nutr 2008 August 1;88(2):511S-512S.

Tai SS-C, Bedner M, Phinney KW. Development of a candidate reference measurement procedure for the determination of 25-hydroxyvitamin D3 and 25-hydroxyvitamin D2 in human serum using isotope-dilution liquid chromatography-tandem mass spectrometry. Anal Chem 2010:82(5):1942–1948.

Looker AC, Pfeiffer CM, Lacher DA, Schleicher RL, Picciano MF, Yetley EA. Serum 25-hydroxyvitamin D status of the US population: 1988-1994 versus 2000-2004. Am J Clin Nutr 2008;88:1519-1527.

Schleicher RL, Encisco SE, Chaudhary-Webb M, Paliakov E, McCoy LF, Pfeiffer CM. Isotope dilution ultra performance liquid chromatography-tandem mass spectrometry method for simultaneous measurement of 25-hydroxyvitamin D2, 25-hydroxyvitamin D3 and 3-epi-25-hydroxyvitamin D3 in human serum. Clin. Chim. Acta 2011;412(17-18):1594-1599.

Sempos CT, Vesper HW, Phinney KW, Thienpont LM, Coates PM. Vitamin D status as an international issue: National surveys and the problem of standardization. Scandinavian J Clin. Lab Investigation 2012;72 (Suppl 243): 32-40.

Young DS. Implementation of SI Units for clinical laboratory data. Ann Int Med 1987; 106: 114-129.

Click to continue to:
VID_D.htm