Good lag screw holding power in trabecular bone of the femoral head is a requisite to achieve stability in the management of proximal femoral fractures. It has been demonstrated that insertion torque and pullout strength of lag screw are linearly correlated. Therefore, insertion torque measurement could be a method to estimate the achieved screw purchase. Manual perception is not reliable [1], but the use of an instrumented screwdriver would make the procedure feasible. The aim of this study was to assess the accuracy achievable using the insertion torque as predictor of lag screw purchase. Four different screw designs (two cannulated and two solid-core screws) were investigated in this study. Each screw was inserted into a block of trabecular bone tissue following a standardised procedure designed to maximise the experimental repeatability. The blocks of trabecular tissue were extracted from human as well as bovine femora to increase the range of bone mineral density. The prediction accuracy was evaluated by plotting pullout strength versus insertion torque, performing a linear regression analysis and calculating the difference (as percentage) between predicted and measured values. Insertion torque showed a strong linear correlation (coefficient of determination R. 2. : 0.95–0.99) with the pullout strength of lag screw. However the prediction error in pullout strength estimation was greater than 40% for small values of insertion torque, decreasing down to 15% when the lag screw was driven into good quality bone tissue. Measuring insertion torque can supply quantitative information about the achieved lag screw purchase. Since screw design and insertion procedure have been shown to affect both the insertion torque and the pullout strength [2], the prediction model must be screw-specific and determined, closely simulating the clinical procedure defined by the screw manufacturer. However, the surgeon must be aware that, even under highly repeatable experimental conditions, the prediction error was found to be high when small insertion torque was measured, i.e. when the screw was driven in low quality bone tissue. Therefore, insertion torque is not reliable in evaluating lag screw purchase in the management of proximal femur fracture of osteoporotic patients

Summary. A rotational limit for screw insertion may improve screw purchase and plate compression by reducing stripping, as compared to a torque based limit. Introduction. Over-tightening screws results in inadvertent stripping of 20% of cortical bone screws. The current method of “two-fingers tight” to insert screws relies on the surgeon receiving torque feedback. Torque, however, can be affected by screw pitch, bone density and bone-thread friction. An alternative method of tightening screws is the “turn-of-the-nut” model, commonly used in engineering applications. In the “turn-of-the-nut” method, nuts used to fasten a joint are rotated a specific amount in order to achieve a pre-specified bolt tension. When applied to orthopaedics, bone assumes the role of the nut and the screw is the bolt. The screw is turned a set angular rotation that is independent of torque feedback. Potentially the “turn-of-the-nut” method provides an easier way of screw insertion that might lessen inadvertent screw stripping. The purpose of the current study was to use the “turn-of-the-nut” method to determine the angular rotation that results in peak plate compression and peak screw pullout force. Methods. Three pairs of human humeri in each of three groups (osteopenic, osteoporotic, and normal) underwent plate compression and pullout protocols. For plate compression, 3.5-mm screws were tightened into strain gauge instrumented plate until screw stripping occurred. Insertion torque, plate compression, and screw rotation were measured. For pullout, 3.5-mm screws were inserted until the head contacted the plate, additionally rotated (90, 180, 270, or 360 degrees), and then pulled out. A generalised linear and latent mixed model was used to check for significant associations (P < 0.05). Results. Mean (95% CI) peak plate compression occurred at 286 degrees (range, 261 – 311 degrees) beyond screw seating. Plate compression significantly increased at 90 to 135 degrees but not after 180 degrees. At 270 degrees, 39% of the screws had already reached their peak ability to compress. Peak screw torque lagged behind peak plate compression by 31 ± 50 degrees, and in seeking peak screw torque, a loss of 104 ± 115 N in plate compression resulted. Screw pullout force was greatest at 90 degrees, but it was not significantly different from that of the other angle groups. Conclusions. Screw rotation at 180 degrees provides plate compression and pullout strength statistically similar to those at greater rotations but without the loss of purchase associated with greater rotations