Reassessing Vitamin D Deficiency: Data Driven Indirect Reference Intervals for 25-Hydroxyvitamin D in Pakistan via refineR Algorithm.

Abstract

Objectives: To establish population-specific reference intervals (RIs) for serum 25-hydroxyvitamin D [25(OH) D] in the Pakistani population using an indirect big data approach, and to evaluate the prevalence of vitamin D deficiency across age and sex strata.

Methods: This retrospective study analyzed serum 25(OH) D levels from 201,625 individuals aged ≥18 years who underwent testing at a tertiary care center over five years. Data were extracted from electronic medical records. Serum 25(OH)D was measured via chemiluminescent immunoassay (DiaSorin Liaison), with internal and external quality controls. The refineR algorithm with modBoxCox transformation was used to derive RIs. Statistical analysis was conducted using R software.

Results: The median serum 25(OH)D level was 17.2 ng/mL, with 25.7% of individuals classified as deficient (≤20 ng/mL), and 33% severely deficient (≤12 ng/mL). Vitamin D deficiency was more pronounced in men and younger adults. Reference intervals derived using modBoxCox transformation were 10.3-51.5 ng/mL for females and 6.1-50.6 ng/mL for males. Notably, females demonstrated higher vitamin D levels across all age groups compared to males.

Conclusion: This study presents large-scale, data-driven estimated reference intervals for 25(OH)D in Pakistan using an indirect approach. The findings highlight a high burden of vitamin D deficiency, particularly among men and younger adults, and reinforce the need for localized reference values to improve diagnostic accuracy. These insights support the development of targeted screening and supplementation strategies, and the reevaluation of clinical thresholds based on population specific data.