Journal of Orthopaedic Trauma Vol No. 2, pp. 103-174 ©°199 Raves Press, Lid., New York Thomas A. DeCoster, David B. Heetderks, Daniel J. Downey, James S. Ferries, and Department of Orthopaedics and Rehabilitation, University of New Mexico School of Medicine, and *Sandia Optimizing Bone Screw Pullout Force *Wendell Jones National Laboratories, Albuquerque, New Mexico Summary: Stability of internal fixation by plate and screw depends on the interface of bone and screw threads. Bone-screw thread design was tested in a synthetic bone model to determine maximum bone-serew pullout force. The ‘model was confirmed to provide reproducible results in a clinically relevant range, Consistent material properties were reflected in the relatively small standard deviations in pullout force, which were less than 10% of mean values. As expected, major diameter was an important determinant of pullout force in a roughly linear manner. Pitch was important with a finer thread giving greater purchase. Minor diameter and the ratio of major to minor diameter had a small but significant effect on pullout force. This study was significant for establish- ing a good bone-screw testing model and suggesting important thread param- eters in selecting or manufacturing a bone screw to optimize its holding power. Key Words Internal fixation by plate and screw or by bone screws alone is an important mode of fracture treat- ment. Failure of this treatment is most common on the bone side of the screw-bone interface. This fail- ure may be manifest in a variety of clinical sequelae including nonunion, malunion, and plate failure (loss of reduction). The impact of bone-screw thread design on failure of the screw-bone interface has been studied previously, but optimal design param- eters have not been identified (1-3,7,9,13,18). Pull- out force is directly proportional to the torque re- quired to strip the threads, a common cause of plant failure, and bas been used as a measure of holding power in previous reports (1,7,8,15). Iden- tification of optimal design parameters would allow for improved screw design, which could reduce fix- ation failure and improve the prognosis for patients with fractures. “Address correspondence and reprint requests to Dr. Thomas A. DeCoster at Department of Orthopaedics and Rehabilitation, University of New Mexico School of Medicine, Albuquerque, New Mexico 87131, U.S.A. 169 rracture fixation—Bone screw—Thread design In prior studies, variability in the quality of bone models used has contributed to wide variation in results (2,3,6,8,13,17). Cadaveric and animal bone density, both from site to site in the same bone and from specimen to specimen, varies widely; this variation is well-recognized as a problem in biome- chanical testing (1,5,11,16). Previous testing has been plagued by this variability with standard devi- ations in the 30-50% range for pullout force 8.9.17). In the presence of this variability, most studies have been unable to demonstrate a signifi- cant relationship between bone-screw thread pa- rameters and holding power. This difficulty has led to the search for a material that closely simulates human bone yet has consistent mechanical proper- ties, Many materials have been evaluated and used in other testing protocols, including a phenol form- aldehyde fabric laminate (Formica) (2), assorted woods (1,5), and most recently, a rigid polyure- thane foam (5,15,16). With one exception, these materials have proven to be poor models, giving either results with too much variability, as does real bone, or results much different than the range seen170 T. A, DeCOSTER ET AL. for real bone. The last material (RF-100 foam, Daro Products, Butler, WI) seems to be the most appro- priate bone substitute in that it has consistent me- chanical properties that are close to those of human bone and it can be cast in configurations and den- sities resembling many different human bone con- ditions (5,17). Prior studies done by screw manufacturers have not been well-publicized and have tested commer- cial screws without optimizing the thread parame- ters. The availability of a promising synthetic bone material, in combination with the need for a defin- itive, independent assessment of bone-screw design parameters, provided impetus for the present study. MATERIALS AND METHODS The screws tested in this study included custom manufactured screws of 316-L stainless steel and ‘commercial 3.5 mm and 4.5 mm screws (Richards Manufacturing Co., Memphis, TN) (Fig. 1). Syn- thetic bone material was manufactured from ure- thane foam and cast in 5 cm diameter cylindrical molds using a ratio of 10 parts resin to 9 parts cat- alyst. The synthetic foam diameter was chosen to maintain a constant pattern of failure by keeping the screw diameter less than 5% of the circumference of the synthetic bone. Casting was performed at a uni- form temperature of 25°C. Synthetic bone blocks were mounted in a jig and drilled with bits 0.1-mm larger than the minor diam- eter of the screw to be tested to maximize pullout force (4,12). All holes were drilled and tapped 10- mm deeper than the actual depth of insertion of the screws to avoid interference with the screw tip. AE ter tapping, each screw was inserted manually to the desired depth. The synthetic bone specimen and screw were mounted in the Instron machine as il- lustrated in Fig. 2. Displacement was applied through the longitudinal axis of the screw at a rate of 1 m/s until failure occurred. A small preload was applied prior to pullout to stabilize the Instron- foam-screw system. A load cell was used in series with the system to produce a calibrated load readout via a signal con- ditioner in volts to a chart recorder. This recorded pullout force in pounds (load) versus displacement in millimeters (deformation) of Instron crosshead and screw for each screw tested. Loading produced a sharp rise in the load deformation curve, and fail- ure was manifest by a rapid decline in the load de- J Orthop Trauma, Vol 4, No, 2, 1990 CUSTOM SCREWS Sorew Type #1 eeenesrsameneson (6184.0), , Sorew Type #200 Ga (212.5105) . Soraw Type #9 El ———wrsroptnn (iai2.s) Serew Type #4 sana (618 5) F COMMERCIAL SCREWS Screw Type #6 (3.52.0/1.2) Sorew Type # 8 (4.513.2/1.75) FIG. 1. Custom and commercial screw types (major diame- teriminor diameteripitch, formation curve (Fig. 3) as well as fracture failure of the foam at the synthetic bone-screw interface. Force data were also reported in Newtons, using a conversion factor of 4.448 Newtons per pound of force. To obtain foam density with pullout characteris- tics similar to those of human bone, we performed a series (10 times per screw) of pullout tests using common orthopaedic bone screws (3.5-mm and 4.5- mm commercial screws) on a variety of foam den- sities. Synthetic bone samples were manufactured with volumes of 27, 32, 38, 42, and 48 cc of urethane in a bone mold with 85-cc total volume. Screws were inserted into the synthetic bone models to a depth of 16 mm. This depth was chosen to allow us to compare our results further with those obtained from similar pullout testing done on cadaveric hu- man bones (17). The model was verified by a series of tests varying the depth of insertion of screws into the foam model. Bone-screw thread-parameter importance wasOPTIMIZING BONE FIG. 2. Instron-foam-screw system. tested by use of custom manufactured serews with specific parameters made by a local machine shop and by use of commercial 3.5 mm and 4.5 mm screws. The parameters tested were major diame- ter, minor diameter, major/minor diameter ratio, and thread pitch (Fig. 4). Custom screws were de- signed to enable us to compare each of these design parameters individually while holding all other pa- rameters constant. Major diameter varied from 3 to \CREW PULLOUT FORCE 171 FIG. 3. Typical load deformation curve. 6 mm. To measure the effect of minor diameter on pullout force, we compared two custom screws with different minor diameters of 4 mm and 5 mm. The major/minor diameter ratio values ranged from. 1.2 to 1.5 among the custom screws tested. Two screws identical except for different pitches, one at 2.5 mm and the other at 1.0 mm, allowed us to look at the effect of a change of pitch has on pullout force while all other parameters are held constant. An- other comparison was obtained by designing one screw to have major and minor diameters exactly one half of those of another custom screw. It was determined that four types of properly de- signed and manufactured screws, in addition to the commercial screws, would allow us to make these comparisons. The custom screws were designed with the following parameters (major diameter! minor diameterithread pitch): screw type 1: 6/5/1.0; serew type 2: 3/2.510.5; screw type 3: 6/4/2.5; screw type 4: 6/5/2.5; screw type 5: 3.5/1.9/1.2; and screw type 6: 4.5/3.1/1.75. Screw 4 was identical to screw 3 except for a different minor diameter, and it was identical to screw 1 except for a different pitch. Screws 5 and 6 were the commercial screws. RESULTS The first phase of testing of synthetic bone den- sity revealed the expected (14) linear relationship between pullout force and foam density (Fig. 5). The mean pullout value of 946 N reported by Trader et al. (17) for bone screws tested to a depth of 16 ‘mm in a human cadaveric radius model was repre- sentative of the literature (2,10,12,18) and corre- sponded to a volume of 35-cc urethane. This 35-cc urethane/85-cc total volume was used in subsequent testing. The range of values for a given test condi- tion was much less for the synthetic bone (average J Onthop Trauma, Vol.4, No.2, 1990172 T. A, DeCOSTER ET AL. PITCH — ~ MINOR DIAMETER -- MAJOR DIAMETER FIG 4, Parameter of bone-srew threads pullout value SD, 9.4%) than for real bone (pullout value SD, up to 54.3%) (8,9,13,18). The synthetic model also showed fracture characteristics at pull- out failure similar to those observed in human bone with a “button” type fracture pattern (1,7,16). The model was validated by demonstrating a linear re- lationship of depth of insertion and pullout force. The third phase of testing determined the effect of changes in specific thread design parameters on PULLOUT STRENGTH (Ibs) FOAM DENSITY (cc) FIG. 5. Pullout force versus foam density. J Onthop Trawna, Vol. 4, No.2, 1990 pullout force. Figure 6 gives the mean load to failure along with the standard deviations for each of the screws tested. As major diameter was increased, within a range of 3-6 mm, the mean pullout force increased in a roughly linear fashion from 105.4 Ibs [469 N] to 305.8 Ibs [1,360 N]. When the minor di- ameter was changed from 4 to 5 mm (screws 3 to 4, respectively) the mean pullout force decreased from 277.8 Ibs (1,236 N] to 247.8 Ibs [1,102 N]. The in- crease in minor diameter from 4 to 5 mm translated toa decrease in the major/minor diameter ratio from 1.5 to 1.2. This decrease in the major/minor diam- eter ratio caused a decrease in pullout force from 2778 Ibs [1,236 N] to 247.8 Ibs [1,102 N]. A de- crease in pitch (finer threads) from 2.5 to 1.0 mm (screws 4 to 1, respectively) caused a significant increase in mean pullout force from 247.8 Ibs [1,102 1] to 305.8 Ibs [1,360 N]. Decreasing design param- eters (major diameter and minor diameter) by one half (screws 1 versus 2) caused the mean pullout force to decrease by over 50% (305.8 Ibs [1,360 N] to 105.4 Ibs [469 N]). Figure 7 gives pullout force values for all thread-parameter comparisons. DISCUSSION Pullout testing performed on screw 5 and 6 showed that pullout force increased in a linear fash- ion as synthetic bone density was increased (Fig. 5). Overall, the consistency of the data obtained through using the synthetic bone model in these pullout tests is much better than that reported from similar studies using bone or other models. The syn- thetic foam has two important characteristics mak- ing it an excellent bone model: it gives accurate, reproducible results, and its composition can be al- tered, allowing comparison to a variety of human bone conditions (17). Testing currently underway in our laboratory suggests that foam has similar me- chanical characteristics to bone including pore size, modulus of elasticity, and surface hardness. This foam has no capacity to respond biologically over time. Results using this model can be used to design in vivo experiments efficiently. The results from the screw testing comparing var- ious thread parameters demonstrate that the syn- thetic bone model is sensitive to changes in thread design (Fig. 7). This holds true even among screws of equal major diameter, a fact other testing sys- tems were unable to demonstrate. Data from our pullout testing reveal a significant relationship be-OPTIMIZING BONE SCREW PULLOUT FORCE 173 w c 400 3 1600 2 —_ Eso a 5 0 FG. 6. Moan joad tofailurratioforeach «EB wwe sereow ype: (major dlameteriminer diam- 200 00 stort 3 9g oo Zz 100 ZF ‘oo 2 i x0 eA 0 ° hn clos ley ce eine, asain ‘tween thread design parameters and pullout force. As expected, major diameter followed a direct lin- ear increase in pullout force (p = 0.001). Pitch is also an important determinant, with lower pitch fine thread) yielding greater pullout force (p 0.001). Minor diameter was seen to have a smaller effect on pullout force, with increasing minor diam- eter leading to decreased pullout force (p = 0.014). Increasing the major/minor diameter ratio, while holding all other parameters constant, resulted in a small but measurable increase in pullout force (p = 0.014). ‘The data indicate that there now exists a syn- thetic bone model that has been previously shown to have properties similar to bone (5) and that has now been shown to yield consistent data, with sen- sitivity to changes in thread design configuration. ‘The data also suggest that pitch may be more clin- ically important than major diameter. The pullout force of a fine-pitch (1.2-mm) screw with a 3.5-mm major diameter was often greater than a 4.5- mm-major-diameter screw with a coarser (1.75-mm) pitch. Although pullout force varied more widely with major diameter than pitch, the major diameter of a clinically usable screw is limited by the size of 400 MEAN LOAD TO FAILURE. On tne tem TES the bone, whereas pitch is not (3,8). The results also show that other design parameters (minor diameter, major/minor diameter ratio) have a small but signif- icant effect on pullout force. ‘This study is significant for establishing a good bone-screw testing model and suggesting important thread parameters in selecting or manufacturing a bone screw for internal fixation. Some emphasis should be placed on the number of threads per length of screw (a measure of pitch) as it had sig- nificant impact on pullout force. Areas of further study suggested by the results include optimization of important thread-parameter values for a given bone condition, testing other modes of failure, and validation in living bone models. Acknowledgment: This project was supported in part by NIH BRSG S07 RR-05583-22, awarded by the Biomedical Research Support Grant Program, Division of Research Resources, National Institutes of Health, and by U.S. Department of Energy Contract Number DE-ACO4 ‘76DPOO789. It was presented in part at the Orthopaedic n Annual Mecting, October 1988, Dal- las, TX. The mechanical testing was performed with the assistance of Frank Vigil at Sandia National Laborato- ries. FIG. 7. Pullout force for thread parameter comparisons. J Orthop Trauma, Vol.4, No. 2, 1990174 T. A. DeCOSTER ET AL. REFERENCES, Wagenknecht M, Donkerwolcke M, Burny F: In Vitro Testing of Orthopedic Implants: The Choice of a ‘Model. Presented at the 12th International Conference on Hoffmann External Fixation, Garmisch-Partenkirchen! ‘Murnau, Bavaria, West Germany, October 1986, p. 88 2. Ansell RH, Scales JT: A study of some factors which affect the strength of screws and their insertion and holding power in bone. J Biomech 1:279-302, 1968 3. Baker BE, Spadaro JA, Baker B, et al: Bone screw optim: zation in the Modified Bristow Procedure” for anterior dis- location of the shoulder. In: Proceedings of the 34th Annual ‘Meeting. Orthopedic Research Society, Atlanta, 1988, p 447 4. Heim U, Pfeiffer KM: Small Fragment Set Manual 2nd ed), Heidelburg, Springer-Verlag, 1982, pp 8-15 5. Hein TJ, Perissinotto A, Hotchkiss R, Chao EYS: Analysis of Bone Model Material for External Fracture Fixation Ex- periments. Proceedings of the 24th Annual Rocky Mountain Bioengineering Symposium and 24th Intemational ISA Bio- medical Sciences Instrument Symposium, Fargo, ND, April 27-28, 1987 6. Kennedy JG, Carter DR: Long bone torsion: Effects of het- cerogeneity, anisotropy and geometric irregularity. J Bio- ‘mech Eng 107:183-188, 1985 17. Koranyi E, Bowman CE, Knecht CD, Janssen M: Holding. power of orthopedic screws in bone. Clin Orthop 72:28 286, 1970, 8. Krag MH, Beynnon BD, Pope MH, etal: An internal fixator for posterior application to short segments of the thoracic, J Orthop Trauma, Vol. 4, No.2, 1990 10. 1 2 1. 4. lumbar or lumbosacral spine. Design and testing. Clin Or- thop 203:75-98, 1986 Lyon WF, Cochran JR, Smith L: Actual holding power of various screws in bone. Ann Surg 114:376-384, 1941 ‘Nunamaker DM, Perren SM: Force measurements in screw fixation. J Biomech 9:669-675, 1976 Pettine KA, Kelly PJ, Chao EYS, Huiskes R: Histologic and biomechanical analysis of unilateral external fixator pinbone interface. In: Proceedings of the 32nd annual meeting, Or- thopedic Research Society, New Orleans, 1986, p. 472 ‘Schatzker J, Sanderson R, Mumaghan JP: The holding power of orthopedic screws in vivo. Clin Orthop 108:115— 126, 1975 Schatzker J, Horne JG, Sumner-Smith G: The effect of ‘movement on the holding power of screws in bone. Clin, Orthop 111:257-262, 1975 Seligson D, Pope D: Pullout Strength of 2mm Mini Half Pins. Presented at the 12th International Conference on, Hoffmann External Fixation, Garmisch-Partenkirchen/ ‘Murnau, Bavaria, West Germany, October 1986, p. 83 Shaw JA: A biomechanical comparison of scaphoid screws. J Hand Surg 124:347-353, 1987 ;. Wagenknecht M, Andrianne Y, Burny F, Donkerwolcke M: Study of the mechanical characteristics of external fixation pin anchorage. Preliminary results. Orthopedics 7:629-632, 1984) Trader JE, Johnson RP, Kalbfleisch JH: Bone mineral con- tent, surface hardness and mechanical fixation in the human radius. J Bone Joint Surg 61A:1217-1220, 1979 Uhthoi? HK: Mechanical factors influencing the holding power of screws in compact bone. J Bone Joint Surg SSB: 633-639, 1973