May-adjusted 25OHD level in serial sampling study
| Time point | n | 25OHD (nmol/L) | Difference (nmol/L) | P value* | P value† |
| Preoperative | 92 | 48.3 (43.3–53.4) | Reference | Reference | Reference |
| Intraoperative | 30 | 38.2 (31.8–44.7) | −12.8 | 0.0004 | 7.8e-5 |
| 1–2 | 62 | 29.0 (25.1–32.9) | −17.3 | 3.6e-9 | 5.5e-7 |
| 3–5 | 58 | 29.6 (25.2–34.1) | −17.2 | 4.4e-9 | 6.5e-5 |
| 6–9 | 45 | 35.0 (28.7–41.2) | −12.7 | 3.0e-6 | 0.002 |
| 30–120 | 58 | 40.0 (34.4–45.5) | −6.4 | 0.002 | 0.04 |
| >162 | 40 | 64.9 (53.5–76.2) | 18.3 | 3.1e-5 | 2.0e-8 |
Difference in 25OHD at each time point accounts for missing samples within paired comparisons. Preoperative samples were taken median 7.5 days before surgery (range 49 to 0 day). >162 day samples taken median 225 days postoperatively (range 163–315 days).
*P values given represent paired Wilcoxon sign-rank test results with preoperative level as reference. C-reactive protein (CRP) was assayed in 70 patients preoperatively. Of these, 30 then had repeat CRP assayed in intraoperative samples while 40 patients had CRP assayed at various postoperative time points.
†P values given represent results from linear mixed model including the fixed covariates age, gender, body mass index (BMI) and CRP, and patient ID included as a random effect.
25OHD, 25-hydroxyvitamin D.