A total of 135 patients with a fresh tibial shaft fracture and with no other significant injuries underwent primary conservative treatment. Data on their smoking habits were obtained from hospital records and by questionnaire. Although the smokers had better prospects for healing of the fracture at the outset than non-smokers (lower mean age and less fractures caused by high-energy injuries), the smokers were found to have a significantly longer mean time to clinical union and a higher incidence of delayed union. According to a crude calculation, smokers had a 4.1-fold risk of tibial shaft fracture caused by low-energy injury, compared with non-smokers. An accelerated failure time model showed that the more comminuted or open the fracture, the higher the number of cigarettes smoked and the older the patient, the longer was the time to clinical union of the tibial shaft fracture. Female sex appeared to be a further risk factor for delayed healing. A logit model indicated that comminution of the fracture, smoking and female sex were associated with delayed union and non-union. If a patient has a markedly raised probability of delayed union of tibial shaft fracture because of many risk factors as reported in the previous literature or in this study, operative treatment should be considered as the primary alternative instead of conservative treatment. Stopping smoking during healing of tibial shaft fracture could also promote the union of the fracture.