Background: Early detection of delayed union or non-union of a fracture is not possible only by clinico-radiological assessment. The changes in the bone turnover markers (BTMs) during fracture healing may lead to early detection of delayed or non-union. Material and methods: Serial measurement of serum Bone Specific Alkaline Phosphatse (BAP) was undertaken after intramedullary fixation of thirty cases (24 males and 6 females) of isolated closed diaphyseal fractures of tibia in healthy young adults (18-45 years age) with 24 weeks follow-up. The fracture was assessed clinically every four weeks and radiologically at 4, 8, 12, 18 and 24 weeks and RUST scoring was done. Serum BAP levels were measured at 0, 4, 8 and 12 weeks using commercial ELISA method. Results: 27 (90%) patients had bridging callus in 3 or 4 cortices (radiological union) and 3 patients (10%) with delayed union had callus in only two or less cortices at 24 weeks. In patients with normal union peak BAP levels were seen at 8 weeks. BAP levels at 4, 8 and 12 weeks were significantly higher in patients achieving normal union as compared to patients with delayed union (p= 0003, <0.001 and 0.041 respectively). In delayed union cases though baseline BAP was comparable to that in normal union cases, but it declined afterwards and remained persistently low. RUST score at 12 weeks correlated significantly with BAP level at 8 weeks (r=0.441, p=0.015). RUST score at 18 weeks correlated significantly with BAP at 8 weeks (r=0.402, p=0.028) and BAP at 12 weeks (r=0.391, p=0.033). RUST score at 24 weeks correlated significantly with BAP at 8 weeks (r=0.406, p=0.026). Conclusions: Peak BAP levels were seen at 8 weeks and correlated well with RUST scores at 12, 18 and 24 weeks. Thus a rising BAP level up to 8 weeks may predict normal radiological healing in coming weeks. While a failure of BAP to rise and persistent low BAP levels than baseline up to 8 weeks may be predictive of delayed union, and a larger study is recommended for better analysis.