Authors (including presenting author) :
Chan PK(1), Fung WC(1), Cheung A(2), Chan VWK(2), Fu H(1), Cheung MH(1), Chiu KY(1)
Affiliation :
Department of Orthopaedics and Traumatology, (1) The University of Hong Kong (2) Queen Mary Hospital
Introduction :
Adequate sunlight exposure and dietary vitamin D levels (VitD) are essential for maintaining bone health. Despite sufficient sunshine in Hong Kong (HK) throughout the year, general population in HK was postulated to be at risk of Hypovitaminosis D(HypoD) because of sedentary lifestyle, Chinese diet and air pollution. Recent studies also showed that adequate level of VitD is also important to decrease perioperative complications in arthroplasty. There is increasing evidence supporting an association between HypoD and adverse postoperative surgical outcomes and worse clinical outcomes, that includes periprosthetic fracture and periprosthetic joint infection (PJI), in patients undergoing arthroplasty. Despite HypoD was easily modifiable, there is a paucity of local studies focusing on this area.
Objectives :
(1) To assess the incidence of HypoD, and its risk factors in patients undergoing lower limb arthroplasty (2) To perform multivariable logistic regression analysis and to develop a risk prediction model for preoperative screening of HypoD
Methodology :
Consecutive patients undergoing lower limb arthroplasty in our institution from Feb 2020 to June 2021 were included. Serum 25-hydroxyvitamin D (25-OHD) level was measured in preoperative assessment clinic and the VitD level was defined as (according to International Endocrine Society classification): Deficient (< 50 nmol/L), Insufficient (50-74 nmol/L) and Sufficient (≥75 nmol/L). Potential variables predicting 25-OHD were collected from standard questionnaires and clinical assessments. Statistical analyses were undertaken using SPSS 26.0 and R studio 1.4. Multivariable logistic regression model was built with cut-off at 50nmol/L (with HypoD vs without HypoD). Model predictive performance was measured by the area Under the receiver operating characteristic curve (AUC)
Result & Outcome :
The mean 25-OHD level in the included patients was 56.5±21.9 nmol/L (41.4% Deficiency (< 50 nmol/L), 41.4% Insufficient(50-74nmol/L), 17.3% Sufficient (≥75 nmol/L)). Among the potential predicting variables, age group, vitamin D supplementation, body mass index, albumin level and weekday outdoor time were the significant predictors (p< 0.05) included in the final multivariable logistic model. AUC of the final model was 0.74 which showed high accuracy in predicting HypoD. A multivariable logistic regression Model could identify patients with HypoD and screening and repletion of VitD could reduce PJI risk so as to reduce the subsequent cost in public hospital.
