Methods

The data reported in this paper come from NHANES III and NHANES 2001-2006, which are cross-sectional surveys representative of the US civilian noninstitutionalized population carried out by the National Center for Health Statistics of the Centers for Disease Control and Prevention from 1988 to 1994 and twice yearly from 2001 to 2006, respectively. A stratified, multistage sampling design was used to recruit participants from household clusters, with oversampling of non-Hispanic blacks and Mexican Americans. Participants were initially interviewed at home, before visiting mobile centers, where they underwent extensive physical examinations. Full details have been published of all survey methods, including sampling, interviews at home, examinations at mobile centers, laboratory measurements of blood samples, ethical approval, and informed consent.

Participants were initially interviewed at home, where information was collected on a wide range of variables, including age, gender, race or ethnicity (self-assigned as non-Hispanic white, non-Hispanic black, Mexican American, or other). Participants were asked about current tobacco smoking, if they were currently taking prescribed medicines for high blood pressure, and the number of times a range of common physical activities were undertaken in leisure time during the previous month. Metabolic equivalents were assigned for each physical activity, and participants aged ≥60 years were classified as doing moderate or vigorous activities if the metabolic equivalent for any activity was ≥3.0 or ≥6.0, respectively, while those aged 20 to 59 years were similarly classified if the metabolic equivalent for any activity was ≥3.5 or ≥7.0, respectively.

Participants were defined as having cardiovascular disease if they reported that doctors had ever told them they had had heart attacks or strokes or if they had pain or discomfort in their chest consistent with angina defined by the Rose criteria (asked only of subjects aged ≥40 years in the 2001 to 2006 surveys) and as having diabetes if they had ever been told by doctors that they had diabetes or were currently taking diabetes medications.

Physical measurements were carried out on most participants at mobile examination centers. Participants were dressed in underpants and disposable light clothing and slippers while being weighed on electronic scales in pounds, which were converted to kilograms. Standing height was measured with a fixed stadiometer to the nearest millimeter. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. Blood pressure was measured at the mobile examination centers by physicians with mercury sphygmomanometers using a standard protocol. Up to 3 blood pressure measurements were collected from each participant while in the sitting position and averaged after excluding the first measurement if >1 measurement was collected. Systolic blood pressure was defined as the point at which the first Korotkoff sound was heard, and diastolic blood pressure was the level of mercury 2 mm below the point at which the last sound was heard. Radial pulse was measured by physicians for 15 seconds and multiplied by 4 from 1988 to 1994 and for 30 seconds and multiplied by 2 from 2001 to 2006, to convert to beats per minute. Pulse rate (beats/min) and systolic blood pressure (mm Hg) were multiplied to calculate the RPP as an indirect measure of myocardial work.

Blood samples collected during the examination were centrifuged, aliquoted, frozen to −70°C on site, and shipped on dry ice to central laboratories, where they were stored at −70°C until analysis. Serum 25(OH)D was measured by a radioimmunoassay kit after extraction with acetonitrile (DiaSorin, Inc., Stillwater, Minnesota) by the National Center for Environmental Health of the Centers for Disease Control and Prevention (Atlanta, Georgia). Serum 25(OH)D concentrations ranged from 2.0 to 90.1 ng/mL after excluding 1 subject with value of 160.3 ng/ml.

A total of 16,573 adults from 1988 to 1994 and 14,542 from 2001 to 2006 aged ≥20 years attended mobile examination centers. Data in this report are restricted to 27,153 non-Hispanic white, non-Hispanic black, and Mexican-American adults aged ≥20 years who attended the mobile examination centers (n = 14,671 from 1988 to 1994 and n = 12,482 from 2001 to 2006), after excluding those of other races or ethnicities (n = 662 from 1988 to 1994 and n = 565 from 2001 to 2006), those who had no serum 25(OH)D measurements (n = 762 from 1988 to 1994 and n = 856 from 2001 to 2006) or very high serum 25(OH)D (n = 1 from 1988 to 1994), and those who did not have systolic blood pressure or pulse rate measurements (n = 477 from 1988 to 1994 and n = 639 from 2001 to 2006).

Statistical analyses were carried out using SUDAAN version 10.0.0 (RTI International, Research Triangle Park, North Carolina) to correct standard errors for any design effect arising from clustered sampling. Sampling weights were not used in analyses to weight up to the US population, because data from 4 NHANES random samples were combined. PROC CROSSTAB was used to compare distributions of categorical variables, and PROC REGRESS was used to calculate adjusted means and mean differences between subgroups.