Contents lists available at ScienceDirect Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech Lecture Achievable accuracy of hip screw holding power estimation by insertion torque measurement Paolo Erani, Massimiliano Baleani ⁎ Laboratorio di Tecnologia Medica, Istituto Ortopedico Rizzoli, Italy ARTICLE INFO Keywords: Proximal femoral fractures Hip screw Insertion torque Pullout strength Holding power prediction ABSTRACT Background: To ensure stability of proximal femoral fractures, the hip screw must firmly engage into the femoral head. Some studies suggested that screw holding power into trabecular bone could be evaluated, in- traoperatively, through measurement of screw insertion torque. However, those studies used synthetic bone, instead of trabecular bone, as host material or they did not evaluate accuracy of predictions. We determined prediction accuracy, also assessing the impact of screw design and host material. Methods: We measured, under highly-repeatable experimental conditions, disregarding clinical procedure complexities, insertion torque and pullout strength of four screw designs, both in 120 synthetic and 80 trabecular bone specimens of variable density. For both host materials, we calculated the root-mean-square error and the mean-absolute-percentage error of predictions based on the best fitting model of torque-pullout data, in both single-screw and merged dataset. Findings: Predictions based on screw-specific regression models were the most accurate. Host material impacts on prediction accuracy: the replacement of synthetic with trabecular bone decreased both root-mean-square errors, from 0.54 ÷ 0.76 kN to 0.21 ÷ 0.40 kN, and mean-absolute-percentage errors, from 14 ÷ 21% to 10 ÷ 12%. However, holding power predicted on low insertion torque remained inaccurate, with errors up to 40% for torques below 1 Nm. Interpretation: In poor-quality trabecular bone, tissue inhomogeneities likely affect pullout strength and insertion torque to different extents, limiting the predictive power of the latter. This bias decreases when the screw engages good-quality bone. Under this condition, predictions become more accurate although this result must be confirmed by close in-vitro simulation of the clinical procedure. 1. Introduction Surgical treatment is the most common approach to treat in- tracapsular and trochanteric fractures of the femur (Adam, 2014; Kaplan et al., 2008; Miyamoto et al., 2008). The type of surgery de- pends on fracture type, fracture pattern and patient conditions (Adam, 2014; Kaplan et al., 2008; Miyamoto et al., 2008; Schipper et al., 2004). Regardless of the implant used for the reduction of the fracture (in- tramedullary or extramedullary fixation), the options to fix the prox- imal part of the femur include the use of blades or screws (Schipper et al., 2004; Stern et al., 2011). Cut-out is one of the most important clinical complications (Baumgaertner et al., 1995; Bojan et al., 2010). The risk of cut-out depends on fracture type (De Bruijn et al., 2012; Geller et al., 2010; Zirngibl et al., 2013) and bone mineral density (Bonnaire et al., 2005; Konstantinidis et al., 2013). However, when a hip screw is used, ac- curate alignment of the device in the femoral head and appropriate distance of the screw tip from the head apex also reduce the risk of screw cut-out (Andruszkow et al., 2012; De Bruijn et al., 2012; Geller et al., 2010). Despite specific tools supplied by the manufacturer with the im- plantable device, accurate driving of self-tapping screws through the femoral head remains a challenge, especially in unstable fractures (Jin et al., 2014). Indeed, the surgeon must maintain the correct position of the proximal fragment during screw driving in order to achieve a good fracture reduction (Hsueh et al., 2010) to optimise stability (Gundle et al., 1995) and avoid the risk of anterior displacement (Mohan et al., 2000). Although anti-rotation wires or pins may be inserted across the fracture before screw driving (Adam, 2014; Saper and Tornetta, 2016), low insertion torque of screw is desirable (Chaturvedi et al., 2015; Georgiannos et al., 2015; O'Malley et al., 2012). It makes the insertion of a self-tapping screw easier, decreasing the need for pre-tapping the hole in good-quality trabecular bone as well as the risk of overcoming the anti-rotation capacity of additional devices or wires used as https://doi.org/10.1016/j.clinbiomech.2018.01.010 Received 1 June 2016; Accepted 16 January 2018 ⁎ Corresponding author at: Laboratorio di Tecnologia Medica, Istituto Ortopedico Rizzoli, Via di Barbiano, 1/10, 40136 Bologna, Italy. E-mail address: baleani@tecno.ior.it (M. Baleani). Clinical Biomechanics 52 (2018) 57–65 0268-0033/ © 2018 Elsevier Ltd. All rights reserved. T