Subjects
As part of the VIDEO trial, 174 subjects were recruited at Southampton. Of these, 50 had serial CE MRIs including one baseline and at least another CE scan during the first 2 years of the study. Of these, 24 were assigned to vitamin D therapy and 26 to placebo. Table
1 shows baseline clinical data and, radiographic and MRI characteristics of these subjects. Of the 50 patients who contributed data, the treatment and placebo groups appeared well matched. Reasons for excluding the remaining 124 participants included: no MRI data, an absence of baseline MRI (particularly CE-MRI) or follow-up MRI.
Table 1
Subject characteristics at baseline by treatment group
Age (years) | 63.0 (5.8) | 63.6 (7.2) |
Females, n (%) | 17 (70.8) | 20 (76.9) |
Index knee, n (% Right) | 14 (58.3) | 15 (57.7) |
Height (m2) | 1.68 (0.09) | 1.68 (0.10) |
Weight (kg) | 79.2 (10.1) | 82.1 (17.0) |
Body mass index (BMI) (kg/m2) | 28.2 (4.0) | 29.2 (5.6) |
Serum vitamin D3 (μg/L) | 26.6 (8.5) | 25.3 (8.7) |
Vitamin D deficient (< 20 μg/L), n (%) | 6 (25) | 7 (27) |
Worst Kellgren–Lawrence (KL) gradea (medial/lateral) in index knee |
Grade 1, n (%) | 6 (25.0) | 2 (7.7) |
Grade 2, n (%) | 8 (33.3) | 14 (53.9) |
Grade 3, n (%) | 8 (33.3) | 8 (30.8) |
Grade 4, n (%) | 2 (8.3) | 2 (7.7) |
WOMACb pain score | 29.4 (15.6) | 34.3 (19.5) |
WOMAC stiffness score | 45.3 (22.9) | 50.4 (22.4) |
WOMAC function score | 33.0 (20.4) | 38.6 (21.2) |
WOMAC total score | 33.3 (18.8) | 38.7 (19.9) |
Synovial tissue volume (mm3) | 11,879.8 (7112.3) | 12,646.6 (5305.5) |
Total Subchondral BML volume (mm3) | 7159.1 (7739.2) | 5765.6 (8839.8) |
STV and subchondral BML volumes of the index knee were missing for a total of 6 patients (12.0%) at year 1 (placebo = 3, vitamin D = 3) and 19 patients (38.0%) at year 2 (placebo = 9, vitamin D = 10). Missingness was caused in part by the exclusion of one visit from 1 patient due to MRI parameters which were beyond the normal scope for the sequence type. The remaining missing observations were due to an absence of imaging data due to technical and/or logistic reasons. Missingness of MRI data did not vary by treatment arm. There were no differences in the baseline characteristics between those 50 subjects used in the analysis and the remaining Southampton study population (
N = 124) other than that the mean baseline serum vitamin D which was significantly higher in those who contributed data to the analysis reported here compared to those who did not; see Additional file
1: Table S1.
Effect of intervention on Vitamin D levels
Of the 50 patients who contributed data, serum vitamin D
3 levels increased significantly in the vitamin D group from a mean of 27.5 μg/L (SD 8.3) at baseline to 32.6 μg/L (SD 7.2) at 12 months follow-up; mean change 5.2 μg/L (SD 5.0) (
P < 0.001). In contrast, serum vitamin D
3 levels decreased significantly in the placebo group from a mean of 25.3 μg/L (SD 8.7) at baseline to 22.7 μg/L (SD 8.5) at 12 months follow-up; mean change was − 2.5 μg/L (SD 6.1) (
P = 0.05). Further, the number of subjects with vitamin D deficiency (< 20 μg/L) [
5,
17] fell from 25.0% (
N = 6) to 4.2% (
N = 1) in the vitamin D intervention group at 12 months follow-up though increased in the placebo group from 26.9% (
N = 7) to 34.6% (
N = 9).
Effect of intervention on STV and subchondral BML volume
After adjusting for baseline values of the outcome, age and gender there was no significant difference at year 1 in the mean change in total synovial tissue volume from baseline between the vitamin D and placebo groups (− 54.6 mm
3, 95% CI -1537.8 to 1428.5); see Table
2. Looking within groups, compared to baseline, there was a small, non-significant decrease in total STV in those on vitamin D (− 364.3 mm
3, 95% CI -1439.1 to 710.6) and also in the placebo group (− 309.6 mm
3, 95% CI -1330.6 to 711.4). Further, there was no significant difference in the mean change from baseline in total synovial tissue volume between groups at 2 years (93.9 mm
3, 95% CI -1605.0 to 1792.7); with both the vitamin D group (155.4 mm
3, 95% CI -1097.3 to 1408.0) and placebo group (61.5 mm
3, 95% CI -1085.6 to 1208.6) showing a small, non-significant increase in total STV compared to baseline. The visit-by-treatment interaction was non-significant.
Table 2
Treatment effect estimates for synovial tissue volume for vitamin D and placebo groups
Vitamin D Therapy | −364.3 | 155.4 |
(−1439.1 to 710.6) | (−1097.3 to 1408.0) |
Placebo | −309.6 | 61.5 |
(− 1330.6 to 711.4) | (−1085.6 to 1208.6) |
Adjusteda Mean Difference between groups | −54.6 | 93.9 |
(−1537.8 to 1428.5) | (−1605.0 to 1792.7) |
After adjusting for baseline values of total subchondral BML volume, age and gender, there was no significant difference at year 1 in the mean change in total subchondral BML volume from baseline between the vitamin D and placebo groups; 2175.7 mm
3 (95% CI -1263.8 to 5615.1); see Table
3. In the placebo arm, total subchondral BML volume decreased but not significantly from baseline by a mean of − 506.7 mm
3 (95% CI -2870.3 to 1856.9), while in the vitamin D arm, BML volume increased but not significantly by a mean of 1669.0 mm
3 (95% CI -827.9 to 4165.9). Further, there was no significant difference at year 2 in the mean change in total subchondral BML volume from baseline; − 313.5 mm
3 (95% CI -4244.7 to 3617.7). In those on vitamin D, there was a small, non-significant decrease in BML volume from baseline (− 506.9 mm
3, 95% CI -3395.6 to 2381.9) and also in the placebo group (− 193.4 mm
3, 95% CI -2845.7 to 2459.0). The visit-by-treatment interaction was non-significant.
Table 3
Treatment effect estimates for subchondral BML volume for vitamin D and placebo groups
Vitamin D Therapy | 1669.0 | −506.9 |
(−827.9 to 4165.9) | (−3395.6 to 2381.9) |
Placebo | −506.7 | − 193.4 |
(− 2870.3 to 1856.9) | (−2845.7 to 2459.0) |
Adjusteda Mean Difference between groups | 2175.7 | −313.5 |
(−1263.8 to 5615.1) | (−4244.7 to 3617.7) |
In sensitivity analysis, further adjustment for baseline body mass index (BMI) and baseline K&L grade did not affect the between group difference for either STV or subchondral BML volume at 2 years follow-up (see Additional file
2: Table S2). Using a one-sided test, there was no significant difference between vitamin D and placebo groups at 2 years follow-up in mean change from baseline for STV (93.9 mm
3, 95% one-sided CI 1190.1, one-sided
P = 0.47) or subchondral BML volume (− 313.5 mm
3, 95% one-sided CI 5062.1, one-sided
P = 0.9).