Hardware Removal: Indications and Expectations Abstract Although hardware removal is commonly done, it should not be considered a routine procedure. The decision to remove hardware has significant economic implications, including the costs of the procedure as well as possible work time lost for postoperative recovery. The clinical indications for implant removal are not well established. There are few definitive data to guide whether implant removal is appropriate. Implant removal may be challenging and lead to complications, such as neurovascular injury, refracture, or recurrence of deformity. When implants are removed for pain relief alone, the results are unpredictable and depend on both the implant type and its anatomic location. Current literature does not support the routine removal of implants to protect against allergy, carcinogenesis, or metal detection. Surgeons and patients should be aware of appropriate indications and have realistic expectations of the risks and benefits of implant removal. H ardware removal is frequently undertaken for symptoms at- tributed to the presence of hardware. In addition, concerns about system- ic and local effects of retained im- plants have led many patients to re- quest elective hardware removal. Although many orthopaedic sur- geons view the procedure as a rou- tine part of care, it is sometimes more challenging and prone to com- plications than the initial surgery. Although there is little debate that hardware should be removed in the setting of implant failure, infec- tion, nonunion, and soft-tissue com- promise, there is little consensus on routine hardware removal in the set- ting of healed fracture. Neither is there consensus on whether im- plants represent a risk for the patient whose vocation or avocation re- quires impact loading at that site. Furthermore, it is not clear how long patients should be protected from significant loads after hardware re- moval. Important considerations in de- termining whether to remove hard- ware include the potential for com- plications and the economic impact. To make the best decision regarding implant removal, the orthopaedic surgeon must be familiar with the potential risk of refracture or neural injury, pain caused by implants, metal sensitivity, carcinogenesis, and the possibility of implant detec- tion by security devices reported in the orthopaedic literature. Frequency and Cost Although there are not extensive data outlining occurrence of hard- ware removal, most sources identify Matthew L. Busam, MD* Robert J. Esther, MD, MSc* William T. Obremskey, MD, MPH *Dual first authorship Dr. Busam is Resident, Department of Orthopaedics and Rehabilitation, Vanderbilt University, Nashville, TN. Dr. Esther is Resident, Department of Orthopaedics, University of North Carolina, Chapel Hill, NC. Dr. Obremskey is Assistant Professor, Department of Orthopaedics and Rehabilitation, Division of Orthopaedic Trauma, Vanderbilt University. None of the following authors or the departments with which they are affiliated has received anything of value from or owns stock in a commercial company or institution related directly or indirectly to the subject of this article: Dr. Busam, Dr. Esther, and Dr. Obremskey. Reprint requests: Dr. Obremskey, Division of Orthopaedic Trauma, Vanderbilt University, Medical Center East, South Tower, Suite 4200, Nashville, TN 37232-8774. J Am Acad Orthop Surg 2006;14:113- 120 Copyright 2006 by the American Academy of Orthopaedic Surgeons. Volume 14, Number 2, February 2006 113 it as a common procedure, account- ing for approximately 5% of all or- thopaedic procedures done in the United States. 1 In a Finnish study, nearly all implants inserted for frac- ture fixation (81%) were removed af- ter fracture healing. 2 Removal of the implant accounted for 29% of elec- tive procedures and 15% of total or- thopaedic procedures performed at that institution during a 7-year peri- od, compared with a removal rate of 6% of all orthopaedic cases in Fin- land for the duration of that study. Despite the significant number of hardware removals performed, there is little published information re- garding the full cost of the proce- dure. In addition to the direct costs (ie, physician and hospital fees), indi- rect costs include patient lost work and productivity. These costs have not been quantified, and only a few studies of implant removal docu- ment patient time away from work. One study of removal of lower ex- tremity intramedullary nails found that patients required a mean of 11 days of sick leave. 3 Given the finite resources available for medical care, research is needed on the economic costs of elective implant removal. Additionally, there is a need for re- search into practice variations re- garding hardware removal in the United States. Peri-implant Fracture and Refracture Internal fracture fixation with either intramedullary or extramedullary implants creates a biologic environ- ment that leads to adaptive changes in bone, with the principal desired effect of fracture healing. Direct frac- ture healing does not produce frac- ture callus; the new osseous chan- nels form across the fracture site in the environment of rigid internal fix- ation, which is most commonly achieved with compression plating. Indirect fracture healing with callus formation occurs in the setting of less rigid fixation, such as intramed- ullary or external fixation. During the initial months of heal- ing after plate fixation, some bone mass loss is observed at the bone- plate interface. Some authors at- tribute this to stress shielding, in which the plates shield the bones from normal, functional stresses leading to bone loss. 4 Perren et al 5 at- tribute this osteopenia to the disrup- tion of blood supply caused by con- tact between the plate and the bone. They showed that osteopenia was temporary, produced even by flexible plastic plates, and occurred less of- ten and for a shorter time when the vascular supply to the bone was less disturbed. This work led to the de- velopment of low-contact plates and locked plates, which cause less peri- osteal and cortical vascular disrup- tion. Multiple reports on implant re- moval demonstrate lower rates of re- fracture when implants are retained longer, possibly further supporting the idea that osteoporosis is a self- limited, vascular phenomenon. Beaupré and Csongradi 6 retrospec- tively reviewed seven studies to ex- amine the refracture rate in 401 pa- tients from whom 459 forearm plates were removed after successful union. They reported higher rates of refracture with the use of large- fragment dynamic compression plates (DCPs) (21%), compared with one-third tubular plates (0%), small- fragment DCPs (5.6%), and semitu- bular plates (6.6%). 6 Removing a plate before complete fracture con- solidation increased the rate of re- fracture. A second area of concern is the stress riser at the cortical defect after screw removal. In a study of drilled dog femurs, Brooks et al 7 reported a mean 55% reduction in energy- absorbing capacity in the presence of a single 2.8- or 3.6-mm drill hole. In a cadaveric study, a 22% reduction in compressive load to failure oc- curred in calcanei after drilling with a 6.0-mm pin, the size commonly used when placing ankle-spanning external fixators for tibial pilon frac- tures. 8 However, Burstein et al 9 re- ported that radiographic evidence of a screw hole remained after the hole began to fill in with new bone. New woven bone eliminated the stress- concentrating effect of the hole within 4 weeks in a canine model, even though the hole was still radio- graphically present. Using single photon absorptiometry, Rosson et al 10 found that bone mass in young adult men returned to close to nor- mal 18 weeks after screw removal, leading them to recommend avoid- ance of contact activity for 4 months after screw removal. Although refracture after plate re- moval cannot be completely pre- vented, the available data lead to several conclusions that can be used to minimize the risk. (1) Achieving complete union and remodeling be- fore implant removal decreases the risk of refracture. (2) Avoiding un- necessary disruption of the vascular supply to the bone decreases os- teopenia. Furthermore, allowing suf- ficient time for the vascular supply to recover may correct the initial os- teopenia. (3) Screw holes may re- main as stress risers for as long as 4 months. Refracture is rarely reported after removal of an intramedullary im- plant. Wolinsky et al 11 reported on 551 fractures managed with reamed intramedullary femur fixation. They removed 131 nails and reported no refractures. In a study of femoral fractures in patients treated with static interlocked stainless steel nails, Brumback et al 12 compared 111 fractures managed with retained implants with 103 from which the implant was removed. No fractures occurred about the nail or locking screws in the first group, and only one patient refractured at the origi- nal fracture site in the second group. The authors concluded that stress shielding from intramedullary nail fixation was not clinically evident once the fracture had united. In addi- tion to radiographic evidence of cir- Hardware Removal 114 Journal of the American Academy of Orthopaedic Surgeons cumferential bridging external cal- lus, they recommend retaining the implant for at least 12 months post- operatively. 12 Although union occurs before 12 months, the additional time allows bone remodeling for hy- pertrophy and strength before hard- ware removal. Patients and physicians are often concerned about the risk of fracture in proximity to a retained implant. Because implants may biomechani- cally function as stress risers, theo- retically they may predispose a pa- tient to peri-implant fracture (Figure 1). However, few data exist indicat- ing an increased overall fracture risk caused by a retained implant. McKee et al 13 reported three cases of fracture occurring at the tip of a locked hu- meral nail, all as a result of low- energy trauma. These injuries were attributed to the type of bone in which the implants were inserted: humeral nails end in diaphyseal bone, whereas femoral and tibial im- plants end in metaphyseal bone. An analogous femoral implant is the in- tramedullary hip screw, such as the original Gamma nail (Stryker, Kalamazoo, MI), a device reported to have a risk of diaphyseal peri- implant fracture as high as 3.1%. 14 Periprosthetic fracture rates about the hip and knee have been reported to be as high as 2.3% and 1.2%, re- spectively. 15 Patients are often concerned about the consequences of a new fracture near a retained implant, but a retained implant may be beneficial if a second fracture occurs. Figure 2, A, demonstrates a distal tibia frac- ture caused by a motorcycle accident in a patient with a retained unlocked intramedullary nail. The tibial in- tramedullary nail was reduced back into the distal metaphysis and re- locked without having to place a new one (Figure 2, B). There is no consensus concerning the necessary amount of protection, weight-bearing limits, or activity modification after implant removal. The available data seem to support Figure 1 Oblique-lateral radiograph demonstrating a peri-implant ulnar fracture caused by a retained implant. The implant served as a stress riser. Figure 2 A, Lateral radiograph in a patient with a prior tibial fracture that was managed with an intramedullary nail. A repeat injury caused the nail to break out of the anterior cortex of the distal tibia. B, The retained nail simplified treatment by allowing reimplantation and relocking of the nail in the distal tibia without the need to replace the intramedullary device. Matthew L. Busam, MD, et al Volume 14, Number 2, February 2006 115 limiting impact and torsional loading for up to 4 months. 10 The timing of resuming contact activity, whether occupational or recreational, is a common question of patients and their families. Brumback et al, 12 ac- knowledging lack of data to support their recommendation, allowed pa- tients to participate in sports activ- ity with an intramedullary nail in place for the first athletic season af- ter fracture healing, provided the in- terlocking screws had been removed. They recommended nail removal upon completion of the first season of competition. Evans and Evans 16 re- ported no clinical problems in 13 of 15 professional rugby players (87%) who returned to participation with a variety of implants in situ. However, one player reported a new, peri- implant fracture after having open re- duction and internal fixation (ORIF) with plating for a both-bone forearm fracture. A second patient was symp- tomatic in the area of a tension-band fixation for a patellar fracture. The authors recommended allowing early return to competitive sports with re- tained implants because the minimal risk is offset by competitive and fi- nancial rewards. 16 The current orthopaedic litera- ture regarding fracture risk from re- tained implants does not support ei- ther universal retention or removal of hardware. There appears to be no significant risk of peri-implant frac- ture when hardware is left in place, even when the patient resumes con- tact activity. The local bone seems to adequately remodel to correct any deficit within 2 to 4 months after hardware removal. The decision to remove or retain hardware cannot be clearly decided solely on the basis of refracture risk; therefore, other fac- tors ought to be considered. Painful Hardware Persistent pain after radiographic ev- idence of fracture union commonly leads to implant removal. Rates of implant removal vary based on ana- tomic location and implant selec- tion. In one study of 55 patients un- dergoing tension band wiring of olecranon fractures, 61% required revision surgery for painful hard- ware. 17 In a retrospective review of surgically treated patellar fractures, 9 of 87 patients underwent removal of symptomatic hardware. 18 It is important to consider whether the patient may reliably expect pain relief after hardware removal. Brown et al 19 examined functional outcomes after inter nal fixation of ankle frac- tures and found lower pain scores and lower scores on the Medical Out- comes Study 36-Item Short Form for patients with pain overlying the lat- eral hardware. Of the 39 patients re- porting pain, 22 underwent removal of hardware, but only 11 (50%) of those had improved lateral ankle pain. These data contrast with that of Jacobsen et al, 20 who reported im- provement after hardware removal in 75% of patients who had previously undergone ORIF of the ankle. Pain relief following femoral in- tramedullary nail removal is simi- larly unpredictable. In their retro- spective review of 80 patients with femoral fractures, Dodenhoff et al 21 noted that 11 of 17 who underwent implant removal experienced pain re- lief. With tibial implants, knee pain is a common indicator for nail re- moval. Keating et al 22 showed a 45% rate of complete relief of knee pain after tibial nail removal; 35% of pa- tients experienced partial relief and 20%, no relief. In a retrospective re- view of 169 patients, Court-Brown et al 23 noted complete pain relief in 27% and marked relief in 69% after nail removal. However, 3.2% re- ported worsening pain after hardware removal. In another study, 17% of pa- tients noted an increase in knee pain after tibial nail removal. 3 Because the extent of pain relief varies after hard- ware removal, the surgeon must ex- ercise caution in attributing persis- tent pain to retained implants. No patient should be guaranteed com- plete pain relief. Fixation Across Joints Preventing implant failure is a com- mon indication for removal. The cy- clic loading associated with fixation across joints often leads to fatigue failure of metallic implants. Because of this concern, hardware is often re- moved from the distal tibiofibular syndesmosis after ankle injury fixa- tion as well as from the midfoot af- ter fixation of Lisfranc joint injuries. Removal versus retention of ankle syndesmosis transfixion screws re- mains controversial. There are no prospective, randomized studies comparing the results of retention versus removal of syndesmosis screws. Some authors routinely re- move the implant before unrest- ricted weight bearing, 24 but DeSouza et al 25 reported no complications from screw retention and removed screws only from those patients who were symptomatic on palpation or who requested removal. Kuo et al 26 followed 48 patients who underwent ORIF for Lisfranc joint injuries for an average of 52 months (range, 13 to 144). Twenty-eight patients required hardware removal secondary to pain, but the remainder demonstrated no clinical problems with the retained hardware. Another concern is the immobil- ity created by fixation across pelvic joints. Displaced fractures of the an- terior and posterior pelvic ring rou- tinely require fixation spanning the symphysis pubis and the sacroiliac joints. Displaced pelvic fractures in female patients have been associated with negative effect on genitouri- nary and reproductive function. 27 To date, no studies have been able to de- termine the ability of a female pa- tient to have a vaginal delivery after undergoing pelvic fracture fixation. However, obstetricians are generally unwilling to have their patients at- tempt vaginal delivery in the setting of symphyseal or sacroiliac fixation. This concern may be an indication for hardware removal in young fe- male patients. Hardware Removal 116 Journal of the American Academy of Orthopaedic Surgeons Metal Allergy Implants with nickel or chromium composition cause allergic respons- es in a small segment of the popula- tion. A review of approximately 50 studies shows the prevalence of met- al sensitivity in the general popula- tion to be 10% to 15%. 28 In fracture surgery, the incidence of sensitivity to any of the three ions in stainless steel (ie, chromium, nickel, cobalt) seems to be low (0.2%, 1.3%, and 1.8%, respectively). 29 Because of concerns about hypersensitivity to any of these ions, some authors have proposed using titanium implants in patients known to be allergic to the components of stainless steel. A patient who has metal sensitiv- ity or a nickel allergy may report nonspecific deep generalized pain over the area of injury and implant. It is very difficult to differentiate this nonspecific pain from either pain caused by the local injury or mechanical pain related to the im- plant. An example of clinical infor- mation that may suggest a metal sensitivity is the presence of symp- toms in a fair-skinned, red-haired woman with a history of earlobe ir- ritation caused by earrings that are not 14-carat gold or caused by cos- tume jewelry. The patient also may be sensitive to medications and have multiple allergies. Patients with sensitivity or allergy will ex- press significant relief almost im- mediately after hardware removal. It is not yet known whether metal sensitivity plays a notable role in implant failure in fracture surgery, or whether it is merely an unusual complication for a limited number of patients. Additionally, it is not known whether there is a cause- and-effect relationship between metal sensitivity and implant loos- ening. Currently, there is no evi- dence of an increased risk of implant failure in patients with positive skin patch testing sensitivity. 30 Carcinogenicity Because younger patients may re- quire insertion of metal implants, the carcinogenic risk of these im- plants must be assessed. The associ- ation between metallic implants and tumors has been established in ex- perimental animals. 31 In the absence of chronic infection, the pathogene- sis of metal-induced carcinogenesis may fall into two general categories: (1) metal-ion binding to DNA and (2) alteration of DNA and protein syn- thesis. Because binding is reversible, other effects are likely to be involved in carcinogenesis. Evidence points to reactive oxygen species created dur- ing corrosion and their effects on DNA and proteins as the likely sec- ond culprit in metal-induced car- cinogenesis. 32 Although basic sci- ence and animal studies may point to a correlation between metallic implants and cancer, one must be careful not to ascribe carcinogenesis to retained implants. There are fewer than 30 human cases of implant-associated tumors in the literature. The limitations of such case reports is that the denom- inator is not known, making it im- possible to quantify risk. Moreover, it is extremely difficult to differenti- ate correlation from causation when trying to establish a relationship be- tween implants and tumors. Gener- ally, sarcomas related to implants tend to be high-grade and occur many years after initial placement of the device. 33 There is no consensus, however, that implants pose a signif- icant risk for local tumor develop- ment. The overall risk, if any, ap- pears to be very low. The great majority of data related to cancer risk and metallic implants is found in the total joint literature. Gillespie et al 34 reported a 70% in- crease in hematopoietic cancers over the general population in their retro- spective review of 1,358 total joint patients over a 10-year period. Those results have not been duplicated in other studies, however. In the largest study to date, Signorello et al 35 con- ducted a nationwide cohort study in Sweden to examine cancer incidence in 116,727 patients who underwent total hip replacement from 1965 through 1994. Overall, they found no increased risk of cancer compared with the general population, but they did note slight increases in prostate cancer and melanoma as well as a reduction in stomach can- cer. Long-term follow-up (>15 years) showed an increase in multiple my- eloma and a statistically insignifi- cant increase in bladder cancer. The authors found no increase in bone or connective tissue cancer in either sex in any follow-up period. 35 The risk of carcinogenicity associated with metallic implants appears to be very small and does not warrant the routine removal of hardware. Metal Detection In this era of heightened security at venues ranging from airports and sporting events to hospital emergen- cy departments and high schools, pa- tients frequently inquire about the possibility that an implant will set off a metal detector. In 1992, Pearson and Matthews 36 tested a variety of arthroplasty and fracture implants. They postulated that only those im- plants with sufficiently high iron content would be detected and that because modern implants have lit- tle, if any, iron, detection is unlike- ly. In 1994, Beaupre 37 corrected that earlier assertion, explaining that 316L stainless steel is actually 60% iron. Detection depends on an ob- ject’s permeability (ability to tempo- rarily disrupt a magnetic field) and conductivity. Because modern pro- cessing techniques limit permeabil- ity and conductivity, the potential for detection is very low. The incidence of implant detec- tion during security screening may be low, but many orthopaedic sur- geons provide their patients with wallet cards containing a short statement providing documenta- Matthew L. Busam, MD, et al Volume 14, Number 2, February 2006 117 tion of a metallic implant as well as a telephone number that appropri- ate authorities may use to further confirm the presence of implanted metal. Our experience with a joint arthroplasty and airport travel is that the screeners do not pay atten- tion to an implant card. Given the low likelihood of detection by secu- rity measures, removing metallic implants to avoid travel concerns is not warranted at this time. Pediatric Patients The general practice at many institu- tions is to offer removal of implants to pediatric patients. The reasons cited for removing pediatric implants include difficulty in removing im- plants later because of exuberant cal- lus overlying the implant, stress shielding, risk of corrosion, metal al- lergy, and potential carcinogenesis. Concern about degenerative pro- cesses and the consequences of re- tained hardware when addressing later fractures also has driven the routine removal of implants in chil- dren. The same concerns may be ex- pressed in adults, but adults have fewer expected years of risk for com- plications. No data are available con- cerning the frequency of a retained implant’s posing a technical problem in the patient undergoing surgery for a second fracture or for joint degen- eration in that extremity. Flexible intramedullary rods used for treating pediatric fractures are routinely removed after bony union. There are no data in the literature re- garding whether these implants should be removed or what the con- sequences are if they are left in place. In a recent review of flexible nailing of pediatric femoral frac- tures, hardware removal was not un- dertaken routinely. 38 Removal of flexible intramedullary nails in chil- dren is frequently as difficult as or more difficult than implantation and requires larger incisions (Figure 3). Of the two major complications in the study by Luhmann et al, 38 one was a septic knee following implant removal. Removal of implants used for treating a slipped capital femoral epiphysis (SCFE) is also routinely done, but not without risk of complications. According to Swiontkowski, 39 a major complica- tion is blood loss and surgical time exceeding that of the original proce- dure. He noted such difficulty in 11 of 18 cases of SCFE hardware remov- al (61%). In another series of implant removal in patients with SCFE, four of seven patients (57%) undergoing implant removal had complications, such as breakage of the retained im- plant or intraoperative fracture. 40 Kahle 41 reported an overall compli- cation rate of 13% in pediatric hard- ware removal but a 42% rate in SCFE hardware removal. Based on these numbers, some surgeons ques- tion the practice of routine hardware removal in children. Kahle 41 stated that “there is very little clinical or experimental evidence to support a policy of routinely removing asymp- tomatic internal fixation devices.” There are no clear data in the lit- erature regarding routine removal of pediatric implants. Chapman states, in the orthopaedic textbook that he edited, “In children we advise rou- tine removal of implants.” 42 Howev- er, Green and Swiontkowski 43 do not recommend (and even discourage) routine removal of implants except in the pelvis and proximal femur, where retained hardware could be problematic during secondary recon- structive procedures. As with any elective procedure, parents need to be aware of the risks and benefits of hardware removal in the pediatric population. Surgical Complications Any surgical procedure carries inher- ent risks, including wound compli- cations, iatrogenic injury, and anes- thetic complications. In their report on implant removal in 86 patients, Richards et al 44 noted a 3% compli- cation rate, including one refracture, one radial nerve injury, and one he- matoma. Sanderson et al 45 reported an overall 20% complication rate in their series of 188 patients. The most common complication was in- fection, followed by nerve injury. They recommend senior surgeon su- pervision of forearm hardware re- moval; unsupervised junior surgeons produced three permanent nerve in- juries. 45 Langkamer and Ackroyd 46 reported on 55 patients who had forearm plate removal. They noted a 40% complication rate, including 4 infections, 5 poor scars, 17 nerve problems, 1 delay in wound healing, Figure 3 Anteroposterior view of a retained flexible intramedullary nail after management of a pediatric femur fracture. Hardware Removal 118 Journal of the American Academy of Orthopaedic Surgeons and 2 refractures. They recommend- ed leaving asymptomatic hardware in place and not delegating the pro- cedure to inexperienced surgeons. They reported complication rates of 13%, 60%, and 100% in cases per- formed by experienced surgeons, moderately experienced surgeons, and inexperienced surgeons, respec- tively. Takakuwa et al 47 reported on four intraoperative fractures of the tibia during elective removal of a slotted intramedullary tibial nail. Given this risk, the surgeon should consid- er intraoperative fluoroscopy to con- firm that no new fracture has oc- curred. Fur thermore, informing the patient about the possible risks of nail removal remains paramount. Summary Hardware removal, although a com- mon operation, should not be under- taken lightly and should not be a routine procedure. Although it is clearly indicated in some instances, the habitual removal of implants is not supported by the literature and exposes the patient to unnecessary costs and complications. Even in pa- tients reporting implant-related pain, removal of that implant does not guarantee relief and may be asso- ciated with further complications, including infection, refracture, nerve damage, and worsening pain. Addi- tionally, patients may request or sur- geons may recommend removal on unproved grounds, such as protec- tion from neoplasm or reduction of stress shielding. No data suggest that implant removal accomplishes these objectives or that retained im- plants increase the risk of neoplasm or cause stress shielding. As with any surgical procedure, it is impor- tant to understand the expected ben- efits from the procedure as well as to know the inherent risks. More re- search is needed regarding the tim- ing and expected benefits of remov- ing implants as well as the direct and indirect costs of the procedure. References Evidence-based Medicine: There are no level I or level II evidence-based studies in the articles referenced. Citation numbers printed in bold type indicate references published within the past 5 years. 1. Rutkow IM: Orthopaedic operations in the United States, 1979 through 1983. J Bone Joint Surg Am 1986;68: 716-719. 2. Bostman O, Pihlajamaki H: Routine implant removal after fracture sur- gery: A potentially reducible consum- er of hospital resources in trauma units. J Trauma 1996;41:846-849. 3. Boerger TO, Patel G, Murphy JP: Is routine removal of intramedullary nails justified? Injury 1999;30:79-81. 4. Harkness JW, Ramsey WC, Harkness JW: Principles of fractures and dislo- cations, in Rockwood CA, Green DP, Bucholz RW, Heckman JD (eds): Frac- tures in Adults, ed 4. Philadelphia, PA: Lippincott-Raven, 1996, vol 1, pp 3-120. 5. Perren SM, Cordey J, Rahn BA, Gauti- er E, Schneider E: Early temporary po- rosis of bone induced by internal fix- ation implants:A reactionto necrosis, not to stress protection? Clin Orthop Relat Res 1988;232:139-151. 6. Beaupré GS, Csongradi JJ: Refracture risk after plate removal in the forearm. J Orthop Trauma 1996;10:87-92. 7. Brooks DB, Burstein AH, Frankel VH: The biomechanics of torsional frac- tures: The stress concentration effect of a drill hole. J Bone Joint Surg Am 1970;52:507-514. 8. Juliano PJ, Yu JR, Schneider DJ, Jacobs CR: Evaluation of fracture predilec- tion in the calcaneus after external fixator pin removal. J Orthop Trauma 1997;11:430-434. 9. Burstein AH, Currey J, Frankel VH, Heiple KG, Lunseth P, Vessely JC: Bone strength: The effect of screw holes. J Bone Joint Surg Am 1972;54: 1143-1156. 10. Rosson J, Murphy W, Tonge C, Shear- er J: Healing of residualscrew holes af- ter plate removal. Injury 1991;22: 383-384. 11. Wolinsky PR, McCarty E, Shyr Y, Johnson K: Reamed intramedullary nailing of the femur: 551 cases. J Trauma 1999;46:392-399. 12. Brumback RJ, Ellison TS, Poka A, Bathon GH, Burgess AR: Intramedul- lary nailing of femoral shaft fractures: III. Long-term effects of static inter- locking fixation. J Bone Joint Surg Am 1992;74:106-112. 13. McKee MD, Pedlow FX, Cheney PJ, Schemitsch EH: Fractures below the end of locking humeral nails: A report of three cases. J Orthop Trauma 1996;10:500-513. 14. Docquier PL, Manche E, Autrique JC, Geulette B: Complications associated with gamma nailing: A review of 439 cases. Acta Orthop Belg 2002;68: 251-257. 15. Mabrey JD, Wirth MA: Periprosthetic fractures, in Rockwood CA, Green DP, Bucholz RW, Heckman JD (eds): Rockwood and Green’s Fractures in Adults, ed 4. Philadelphia, PA: Lippincott-Raven, 1996, vol 1, pp 539- 603. 16. Evans NA, Evans RO: Playing with metal: Fracture implants and contact sport. Br J Sports Med 1997;31:319- 321. 17. Romero JM, Miran A, Jensen CH: Complications and re-operation rate after tension-band wiring of olecra- non fractures. J Orthop Sci 2000;5: 318-320. 18. Smith ST, Cramer KE, Karges DE, Watson JT, Moed BR: Early complica- tions in the operative treatment of pa- tella fractures. J Orthop Trauma 1997;11:183-187. 19. Brown OL, Dirschl DR, Obremskey WT: Incidence of hardware-related pain and its effect on functional outcomes after open reduction and internal fixation of ankle frac- tures. J Orthop Trauma 2001;15:271- 274. 20. Jacobsen S, de Lichtenberg MH, Jen- sen CM, Torholm C: Removal of in- ternal fixation—the effect on pa- tients’ complaints: A study of66 cases of removal of internal fixation after malleolar fractures. Foot Ankle 1994; 15:170-171. 21. Dodenhoff RM, Dainton JN, Hutch- ins PM:Proximal thighpain afterfem- oral nailing: Causes and treatment. J Bone Joint Surg Br 1997;79:738-741. 22. Keating JF, Orfaly R, O’Brien PJ: Knee pain after tibial nailing. J Orthop Trauma 1997;11:10-13. 23. Court-Brown CM, Gustilo T, Shaw AD: Knee pain after intramedullary tibial nailing: Its incidence, etiology, and outcome. J Orthop Trauma 1997; 11:103-105. 24. Meyer TL Jr, Kumler KW: A.S.I.F. technique and ankle fractures. Clin Orthop 1980;150:211-216. 25. de Souza LJ,Gustilo RB, Meyer TJ: Re- sults of operative treatment of dis- placed external rotation-abduction Matthew L. Busam, MD, et al Volume 14, Number 2, February 2006 119 fractures of the ankle. J Bone Joint Surg Am 1985;67:1066-1073. 26. Kuo RS, Tejwani NC, DiGiovanni CW, et al: Outcome after open reduc- tion and internal fixation of Lisfranc joint injuries. J Bone Joint Surg Am 2000;82:1609-1618. 27. Copeland CE, Bosse MJ, McCarthy ML, et al: Effect of trauma and pelvic fracture on female genitourinary, sex- ual, and reproductive function. J Orthop Trauma 1997;11:73-81. 28. Jacobs JJ, Goodman SB, Sumner DR, Hallab NJ: Biologic response to ortho- paedic implants, in Buckwalter JA, Einhorn TA, Simon SR (eds): Ortho- paedic Basic Science: Biology and Biomechanics of the Musculoskeletal System, ed 2. Rosemont,IL: American Academy of Orthopaedic Surgeons, 2000, pp 401-426. 29. Swiontkowski MF, Agel J, Schwap- pach J, McNair P, Welch M: Cutane- ous metal sensitivity in patients with orthopaedic injuries. J Orthop Trauma 2001;15:86-89. 30. Hallab N, Merritt K, Jacobs JJ: Metal sensitivity in patients with ortho- paedic implants. J Bone Joint Surg Am 2001;83:428-436. 31. Memoli VA, Urban RM, Alroy J, Gal- ante JO: Malignant neoplasms associ- ated with orthopedic implant materi- als in rats. J Orthop Res 1986;4:346- 355. 32. Wang S, Shi X: Molecular mechanisms of metal toxicity and carcinogenesis. Mol Cell Biochem 2001;222:3-9. 33. Keel SB, Jaffe KA, Petur Nielsen G, Rosenberg AE: Orthopaedic implant- related sarcoma: A study of twelve cases. Mod Pathol 2001;14:969-977. 34. Gillespie WJ, Frampton CM, Hender- son RJ, Ryan PM: The incidence of can- cer following total hip replacement. J Bone Joint Surg Br 1988;70:539-542. 35. Signorello LB, Ye W, Fryzek JP, et al: Nationwide study of cancer risk among hip replacement patients in Sweden. J Natl Cancer Inst 2001;93: 1405-1410. 36. Pearson WG, Matthews LS: Airport detection of modern orthopaedic im- plant metals. Clin Orthop 1992;280: 261-262. 37. Beaupre GS: Letter: Airport detection of modern orthopedic implantmetals. Clin Orthop 1994;303:291-292. 38. Luhmann SJ, Schootman M, Schoe- necker PL, Dobbs MB, Gordon JE: Complications of titanium elastic nails for pediatric femoral shaft frac- tures. J Pediatr Orthop 2003;23:443- 447. 39. Swiontkowski MF: Slipped capital femoral epiphysis: Complications related to internal fixation. Orthopedics 1983;6:705-712. 40. Jago RD, Hindley CJ: The removal of metalwork in children. Injury 1998; 29:439-441. 41. Kahle WK: The case against routine metal removal. J Pediatr Orthop 1994;14:229-237. 42. Chapman MW: Principles of internal and external fixation, in Chapman MW (ed): Chapman’s Orthopaedic Surgery, ed 3. Philadelphia, PA: Lip- pincott Williams and Wilkins, 2001, vol 1, pp 375-376. 43. Green NE, Swiontkowski MF: Skele- tal Trauma in Children, ed 3. Phila- delphia, PA: Saunders, 2003, vol 3, pp 388-398. 44. Richards RH, Palmer JD, Clarke NM: Observations on removal of metal im- plants. Injury 1992;23:25-28. 45. Sanderson PL, Ryan W, Turner PG: Complications of metalwork remov- al. Injury 1992;23:29-30. 46. Langkamer VG, Ackroyd CE: Remov- al of forearm plates: A review of the complications. J Bone Joint Surg Br 1990;72:601-604. 47. Takakuwa M, Funakoshi M, Ishizaki K, Aono T, Hamaguchi H: Fracture on removal of the ACE tibial nail. J Bone Joint Surg Br 1997;79: 444-445. Hardware Removal 120 Journal of the American Academy of Orthopaedic Surgeons