Cost-Effectiveness of Antibiotic-Impregnated Bone Cement Used in Primary Total Hip Arthroplasty

Transcription

1 This is an enhance PDF from The Journal of Bone an Joint Surgery The PDF of the article you requeste follows this cover page. Cost-Effectiveness of Antibiotic-Impregnate Bone Cement Use in Primary Total Hip Arthroplasty Justin S. Cummins, Ivan M. Tomek, Stephen R. Kantor, Ove Furnes, Lars Birger Engesæter an Samuel R.G. Finlayson J Bone Joint Surg Am. 2009;91: oi: /jbjs.g This information is current as of March 10, 2009 Supplementary material Reprints an Permissions Publisher Information Commentary an Perspective, ata tables, aitional images, vieo clips an/or translate abstracts are available for this article. This information can be accesse at Click here to orer reprints or request permission to use material from this article, or locate the article citation on jbjs.org an click on the [Reprints an Permissions] link. The Journal of Bone an Joint Surgery 20 Pickering Street, Neeham, MA

2 634 COPYRIGHT Ó 2009 BY THE JOURNAL OF BONE AND JOINT SURGERY, INCORPORATED Cost-Effectiveness of Antibiotic-Impregnate Bone Cement Use in Primary Total Hip Arthroplasty By Justin S. Cummins, MD, MS, Ivan M. Tomek, MD, Stephen R. Kantor, MD, Ove Furnes, MD, PhD, Lars Birger Engesæter, MD, PhD, an Samuel R.G. Finlayson, MD, MPH Investigation performe at the Department of Orthopaeic Surgery, Dartmouth Hitchcock Meical Center, Lebanon, New Hampshire Backgroun: Antibiotic-impregnate bone cement is infrequently use in the Unite States for primary total hip arthroplasty because of concerns about cost, performance, an the possible evelopment of antibiotic resistance an because it has been approve only for use in revision arthroplasty after infection. The purpose of this stuy was to moel the use of antibiotic-impregnate bone cement in primary total hip arthroplasty for the treatment of osteoarthritis to etermine whether use of the cement is cost-effective when compare with the use of cement without antibiotics. Methos: To evaluate the cost-effectiveness of each strategy, we use a Markov ecision moel to tabulate costs an quality-ajuste life years (QALYs) accumulate by each patient. Rates of revision ue to infection an aseptic loosening were estimate from ata in the Norwegian Arthroplasty Register an were use to etermine the probability of unergoing a revision arthroplasty because of either infection or aseptic loosening. The primary outcome measure was either all revisions or revision ue to infection. Perioperative mortality rates, utilities, anisutilities were estimate from ata in the arthroplasty literature. Costs for primary arthroplasty were estimate from ata on in-hospital resource use in the literature. The aitional cost of using antibiotic-impregnate bone cement ($600) was then ae to the average cost of the initial proceure ($21,654). Results: When all revisions were consiere to be the primary outcome measure, the use of antibiotic-impregnate bone cement was foun to result in a ecrease in overall cost of $200 per patient. When revision ue to infection was consiere to be the primary outcome measure, the use of the cement was foun to have an incremental costeffectiveness ratio of $37,355 per QALY compare with cement without antibiotics; this cost-effectiveness compares favorably with that of accepte meical proceures. When only revision ue to infection was consiere, it was foun that the aitional cost of the antibiotic-impregnate bone cement woul nee to excee $650 or the average patient age woul nee to be greater than seventy-one years before its cost woul excee $50,000 per QALY gaine. Conclusions: When revision ue to either infection or aseptic loosening is consiere to be the primary outcome, the use of antibiotic-impregnate bone cement results in an overall cost ecrease. When only revision ue to infection is consiere, the moel is strongly influence by the cost of the cement an the average age of the patients. With few patients less than seventy years of age unergoing total hip arthroplasty with cement in the Unite States, the use of antibiotic-impregnate bone cement in primary total hip arthroplasty may be of limite value unless its cost is substantially reuce. Level of Evience: Economic anecision analysis, Level II. See Instructions to Authors for a complete escription of levels of evience. Deep infection following total hip arthroplasty is a evastating complication that can require costly revision surgery an reuce a patient s functional status. Several methos to reuce the incience of infection, incluing improve surgical technique, improve perioperative preparation, an use of prophylactic antibiotics, have been introuce since Disclosure: The authors i not receive any outsie funing or grants in support of their research for or preparation of this work. Neither they nor a member of their immeiate families receive payments or other benefits or a commitment or agreement to provie such benefits from a commercial entity. No commercial entity pai or irecte, or agree to pay or irect, any benefits to any research fun, founation, ivision, center, clinical practice, or other charitable or nonprofit organization with which the authors, or a member of their immeiate families, are affiliate or associate. A commentary is available with the electronic versions of this article, on our web site ( an on our quarterly CD-ROM/DVD (call our subscription epartment, at , to orer the CD-ROM or DVD). J Bone Joint Surg Am. 2009;91: oi: /jbjs.g.01029

3 635 the 1960s 1,2. These changes have been creite with reucing the infection rate associate with total hip arthroplasty from nearly 10% to between 0.5% an 2% 1,2. The use of antibiotic-impregnate bone cement has been avocate as one metho to further reuce the nee for revision surgery following primary total hip arthroplasty 3. Large registry atabases, such as the Norwegian Arthroplasty Register an the Sweish Arthroplasty Register, have shown a ecrease rate of revision surgery in patients who receive both perioperative intravenous antibiotics an antibiotic-impregnate bone cement at the time of primary total hip arthroplasty 4-7. Proponents of the use of antibiotic-impregnate bone cement in the Unite States point to these ata as evience that it shoul be use in all primary proceures involving use of cement 8. Opponents of the use of antibiotic-impregnate bone cement frequently cite its cost as the primary concern, especially given the alreay low rates of infection an revision 9. Other concerns inclue the possible evelopment of antibiotic resistance, allergic reactions, an possible compromise of the mechanical properties of the cement from the amixture of antibiotics Also, since antibiotic-impregnate bone cement has been approve by the Unite States Foo an Drug Aministration only for secon-stage reimplantation after revision ue to infection, use of antibiotic-impregnate bone cement in primary total hip arthroplasty represents an off-label usage in the Unite States. The purpose of this stuy was to employ a ecision analytic moel that accounts for competing risks, benefits, an costs to assess the cost-effectiveness of the use of antibioticimpregnate bone cement for primary total hip arthroplasty. Specifically, we sought to answer the question of whether the up-front cost of using antibiotic-impregnate bone cement in all patients unergoing total hip arthroplasty woul be justifie by the expecteecrease in future revisions. Materials an Methos Markov ecision analysis was use to moel the clinical scenario of eciing whether to use antibiotic-impregnate bone cement or stanar polymethylmethacrylate bone cement without antibiotics for primary total hip arthroplasty. Markov moels are useful when ecision problems involve competing risks over time, or when the timing of events is important, as is the case with revision surgery after primary total hip arthroplasty 13. Assumptions in the moel regaring revision rates, costs, utilities, mortality risks, an all other parameters were rawn from theliteratureanareescribeinetailbelow. Moel Design The esign of the Markov moel is illustrate schematically in Figure 1 an in etail in the Appenix. The ecision tree represents the potential clinical course of hypothetical patients in the scenario moele (i.e., the ecision to use antibioticimpregnate bone cement or conventional cement). The ecision tree inclues five main health states, an the arrows between the health states represent chance events that can occur over time. The five health states in this moel are: well after total hip arthroplasty with antibiotic-impregnate bone cement, well after total hip arthroplasty without antibioticimpregnate bone cement, well after aseptic revision, well after revision ue to infection, aneath. Over the course of the simulation, hypothetical patients transition from one health state to another on the basis of transition probabilities associate with each chance event moele. Patients who are well after total hip arthroplasty can transition to any of the other health states uring each time cycle of the moel (one year). Transition probabilities can change over time (e.g., the risk of eath from all causes increases with the patient s age). The moel continues to cycle until all hypothetical patients eventually reach the eath state. Each health state is assigne a utility. Utilities are efine as a measure of how a patient efines the value of a specific health state. Utility values are typically scale from 0 (eath) to 1 (perfect health) 13. The value of time in a given health state is measure in quality-ajuste life years (QALYs), which are calculate as the time in the health state multiplie by the utility assigne to the health state (years utility = QALY). Costs associate with each strategy are calculate on the basis of the occurrence of events in the moel, such as the cost of unergoing an aseptic revision when transitioning to the well after aseptic revision health state. The moel then tabulates the total utilities an costs accumulate by a hypothetical cohort of patients uring their simulate life span (before they transition to the eath state) for both the antibioticimpregnate-cement an stanar-cement strategies 14.Allcosts an utilities are iscounte at a stanar rate of 3% per year. The cost-effectiveness of using antibiotic-impregnate bone cement is expresse as the ratio of aitional costs attributable to this strategy to the aitional benefits (QALYs). The moel was esigne with use of ecision analysis software (TreeAge Pro 2005; TreeAge Software, Williamstown, Massachusetts). Moel Assumptions an Parameters In constructing the ecision moel, we use the following general assumptions: (1) all patients are unergoing primary total hip arthroplasty for the treatment of osteoarthritis, an a cemente femoral stem is use in all cases; (2) each patient can unergo only one revision uring his or her lifetime; (3) revision ue to a ocumente infection is performe as a twostage proceure, in which the prosthesis is remove, the patient is treate with intravenous antibiotics for six weeks, an a new prosthesis is then implante; an (4) the use of antibiotic-impregnate bone cement oes not affect the utility (value) of the ifferent health states moele. The parameters of the moel are iscusse below. Patient Population The age of the patients in the moel was set at sixty-eight years to coincie with the average age reporte in the Norwegian Arthroplasty Register from 1987 to Patients were assume to be of average health for their age an to be unergoing total hip arthroplasty because of egenerative arthritis an not a fracture.

4 636 Fig. 1 Clinical pathway of patients assume to have hip arthritis for which meical management has faile. Each patient receives a total hip arthroplasty with or without antibiotic-impregnate bone cement (ABC). When patients survive the operation, they are assume to stay well until they ie of other causes or nee a revision. When they have a revision an survive, they are assume to stay well with that revision until they ie from other causes. The moel continues until all patients reach the eath state. Revision Rates Revision rates were base on a recent stuy on the ifference in revision rates, as reporte to the Norwegian Arthroplasty Register, between patients who ha receive systemic antibiotics an antibiotic-impregnate bone cement at the time of primary total hip arthroplasty an patients who ha receive systemic antibiotics only 4,5. The results in that stuy were ivie into revisions ue to a ocumente infection (positive cultures) an aseptic revisions (negative cultures). The rate of revision was slightly higher in the first several years after the arthroplasty, but a linear rate of revision was assume for the moel. This approach provies a more conservative estimate of the cost-effectiveness of antibiotic-impregnate bone cement because revision costs are shifte further into the future an therefore iscounte. The ten-year revision rates were converte into yearly rates of both revisions ue to infection an those ue to aseptic loosening. For patients treate with antibiotic-impregnate bone cement, the rate of revision ue to infection was set at 0.04% per year an the rate of revision ue to aseptic loosening was set at 0.31% per year (Table I). The revision rates for the group treate with stanar bone cement were then etermine by multiplying the revision rates for the group treate with antibiotic-impregnate cement by the increase risk of revision in subgroups in which antibiotic-impregnate bone cement was not use. The relative risk of revision ue to infection was set at 1.8 an the relative risk of aseptic revision was set at 1.3 for the stanar-cement branch (Table I). The rates of infection an aseptic loosening use in the moel are on the low en of rates reporte in the literature 7,15, again proviing a conservative estimate of the costeffectiveness of antibiotic-impregnate bone cement. Mortality Rates The age-specific probability of eath from all causes was estimate from U.S. Life Tables from The risk of eath increase as the hypothetical patients cycle through the moel, an the probability of eath was set at 100% at age 101 to terminate the simulation. The risk of perioperative eath for patients treate with primary total hip arthroplasty was set at 0.23% on the basis of ata from the Norwegian Arthroplasty Register 5. Revisions were assume to be associate with the same mortality risk. Utilities Laupacis et al. use the time-trae-off technique to etermine utility scores for patients before an after primary total hip arthroplasty 16. These patients ha a mean age of sixty-four years, an the utility value average 0.80 two years after the total hip arthroplasty. Rorabeck et al. foun a similar result using the time-trae-off technique 17. These ata were use to assign a utility value of 0.80 for patients who ha unergone primary total hip arthroplasty in our moel (Table I). There are few stuies in the orthopaeic literature in which functional outcome was measure following aseptic revision of a total hip arthroplasty, an we foun no stuies in which a utility score was irectly etermine for patients who ha unergone that proceure. Hozack et al. use scores on the Short Form-36 (SF-36) to compare patients treate with primary total hip arthroplasty with those treate with aseptic revision total hip arthroplasty 18. The patients who unerwent revision ha lower scores in every category postoperatively, espite having ha similar preoperative scores. The utility value

5 637 TABLE I Markov Moel Variables Variable Value Reference Rate of revision ue to 0.4% over 10 years 5 infection with antibiotic cement Aseptic revision rate with 3.1% over 10 years 5 antibiotic cement Relative risk of revision 1.8 baseline rate 5 ue to infection with regular cement Relative risk of aseptic 1.3 baseline rate 5 revision with regular cement Probability of eath U.S life tables from all causes Probability of eath from 0.23% increase risk 5 total hip arthroplasty Cost of primary total $21, hip arthroplasty Increase cost of 1.6 baseline cost 21 aseptic revision Increase cost of revision 4.44 baseline cost 21 ue to infection Aitional cost of $600 antibiotic cement Utility of primary total hip arthroplasty Utility of aseptic revision 0.72 (10% ecrease) 18 Utility of revision ue 0.64 (20% ecrease) to infection Disutility of total hip 20.1 arthroplasty Disutility of aseptic (20% more) revision Disutility of revision ue to infection Discount rate per year 3% for cost an utilities for aseptic revision was conservatively estimate to be 10% lower than that for primary total hip arthroplasty ( % = 0.72). Comparative functional outcome ata following revision ue to infection were not ientifie in the literature, an a conservative estimate of a 20% ecrease in utility was assigne to the patients treate with that proceure ( % = 0.64). Disutilities Disutilities are a measure of the transient lower quality of life associate with unesirable events 13. They were use in this moel to represent the temporary health state of a patient in the perioperative perio, when patients have increase pain an ecrease mobility, as well as the potential for other complications (Table I). The isutility is assesse as a one-time toll within the moel, an the assigne amount is eucte from the patient s accumulate QALYs at the time that they unergo one of the proceures. The isutility for primary total hip arthroplasty was set at 20.1, which is the equivalent of eucting just over five weeks of perfect health. The isutility for aseptic revision was set at 20.12, a slight increase to account for the increase in complexity of a typical revision proceure. For revisions ue to infection, the assigneisutility was 20.20, which takes into consieration that the majority of patients unergo a two-stage proceure an thus spen a longer perio of time in an unesirable health state. Costs All cost estimates are in 2002 U.S. ollars. Cost estimates for primary total hip arthroplasty were obtaine from the orthopaeic literature as well as from the National Inpatient Survey (NIS) ata. These estimates accounte for the costs associate with the proceure an the acute hospitalization. Surgeons fees, costs for a rehabilitation stay, an lost wages ue to misse workays were not inclue. Estimates range from $12,846 to $31,000 for a primary total hip arthroplasty 19,20. The cost assigne to a primary total hip arthroplasty in our moel was $21,654, which was base on stuies, publishe in 2005, by Bozic et al., who estimate the hospital resources use for primary total hip arthroplasties, aseptic revisions, an revisions ue to infections 21,22. In the literature, estimates of the increase in cost for aseptic revisions, as compare with the cost for primary arthroplasty, have range from 20% to 60% 22, although in many reports it is unclear whether revisions ue to infection were inclue in the estimate. In our moel, the cost of an aseptic revision was estimate to be $34,866, as reporte by Bozic et al. 21. Cost estimates for revisions ue to infection were base on the assumption that the majority of patients were treate with a two-stage revision, with intravenous antibiotics aministere uring the perio between the removal of the prosthesis an the revision implant proceure. The cost assigne to revisions ue to infection in the moel was $96,166, on the basis of the ata reporte by Bozic et al. 21. The aitional cost of using antibiotic-impregnate bone cement was estimate by calculating the ifference between the cost of stanar polymethylmethacrylate bone cement an the cost of commercially available premixe gentamicin antibioticimpregnate bone cement. Gentamicin antibiotic-impregnate bone cement is the cement that has been stuie the most often, an its beneficial effects have been shown by ata in both the Sweish an the Norwegian Arthroplasty Register 5,6. The estimate cost of a 40-g packet of antibiotic-impregnate bone cement at our institution is approximately $365, an stanar bone cement costs approximately $65. It was estimate that two packets of cement are use on the average, resulting in an aitional cost of $600 per primary total hip arthroplasty. Analysis Cost-effectiveness analysis is a useful tool for evaluating meical interventions when one strategy is more costly but also more effective 14. If a treatment strategy is more costly an less ef-

6 638 fective, it is sai to be ominate by the alternative treatment strategy an shoul never be chosen 13. Similarly, if a strategy is less costly an more effective, it shoul always be chosen. The measure of a treatment s cost-effectiveness is expresse as an incremental cost-effectiveness ratio 13, which is calculate by iviing the ifference in cost between the two strategies by the ifference in effectiveness (i.e., the net cost ivie by net benefit). The unit of measure for effectiveness in this analysis is QALYs, resulting in a ratio expresse in ollars per QALY. Although no specific ollar value has been universally agree on as the threshol for cost-effectiveness, a meical intervention is generally consiere to be cost-effective if the incremental costeffectiveness ratio is $50,000 per QALY 13. The incremental costeffectiveness ratio was etermine for this moel by calculating the ifference between the costs accumulate by the patients treate with the antibiotic-impregnate bone cement an those accumulate by the patients treate with the stanar bone cement. This ollar amount was then ivie by the ifference in accumulate QALYs between the two strategies. A secon analysis was performe with use of the same methos but with the relative risk of aseptic revision set at one to evaluate the costeffectiveness of antibiotic-impregnate bone cement when only ocumente infections are consiere as outcomes. By analyzing revisions ue to ocumente infection separately, we coul compare the minimal expecte benefit from antibioticimpregnate bone cement with the maximal benefit that coul be expecte when all revisions were consiere in the moel. Sensitivity analysis was then performe on each of the parameters within the moel. Sensitivity analysis is use to evaluate how the outcome of the moel might change when cost, benefit, or risk assumptions are varie over a plausible range of values. For example, if the cost of a primary total hip arthroplasty is evaluate over the range reporte in the literature ($12,846 to $31,000), we can etermine whether uncertainty surrouning the true cost of total hip arthroplasty substantially weakens the conclusions rawn from the moel. Source of Funing No external funing source was use for this stuy. Results When all revisions (those ue to infection or aseptic loosening) were consiere to be the primary outcome measure, the use of antibiotic-impregnate bone cement for primary total hip arthroplasty was foun to be less costly an more effective (ominant), resulting in an overall cost ecrease of $200 per patient. When only revision ue to infection was consiere to be the primary outcome measure, the use of antibiotic-impregnate bone cement was foun to have an incremental cost-effectiveness ratio of $37,355 per QALY, which suggests that it is a cost-effective strategy if all of the moel parameters remain constant (Table II). Revision Rates Sensitivity analysis reveale that the relative risk of aseptic revision after the use of stanar bone cement woul nee to TABLE II Total Cost an QALYs Primary Outcome Measure/Cement Type Cost QALYs Cost per QALY Incremental Cost- Effectiveness Ratio All revisions Antibiotic cement $23, $2533 Dominant Regular cement $24, $2551 Revisions ue to infection Antibiotic cement $23, $2533 $37,355 Regular cement $23, $2509 be <1.2 before the use of antibiotic-impregnate bone cement was no longer the ominant strategy (Table III). Antibioticimpregnate bone cement remaine cost-effective (<$50,000 per QALY) until the relative risk of aseptic revision after the use of stanar bone cement was <1.0 i.e., antibioticimpregnate bone cement woul nee to be associate with an increase risk of the patient requiring aseptic revision before it woul no longer be a cost-effective strategy. When only revisions ue to infection were consiere, it was foun that the relative risk of revision ue to infection with the use of stanar bone cement ha to be <1.7 before the use of antibiotic-impregnate bone cement was no longer a cost-effective strategy. If the relative risk of infection after the use of stanar bone cement were >2.4, the use of antibiotic-impregnate bone cement woul become the ominant strategy (Table IV). Utilities Sensitivity analysis of the utility values assigne to the ifferent health states reveale that, when all revisions are consiere to be the primary outcome measure, antibiotic-impregnate bone cement remains the ominant strategy even when the utility of life after a single primary total hip arthroplasty is equal to the utility of life after a revision ue to aseptic loosening or infection (Table III). If only revision ue to infection is consiere as the primary outcome measure, the utility of a revision ue to infection woul nee to be >0.70 (close to the value of life after primary total hip arthroplasty) before the use of antibiotic-impregnate bone cement woul no longer be cost-effective (Table IV). Disutilities In the sensitivity analysis, the isutility values assigne to the proceures were foun to have essentially no effect on the moel over a broa range of values. This hel true when all revisions were consiere to be the primary outcome measure as well as when only revision ue to infection was consiere to be the primary outcome measure. Costs Sensitivity analysis of costs reveale that the moel is much more sensitive to cost parameters than it is to the other parameters

7 639 TABLE III Sensitivity Analysis with All Revisions as Primary Outcome Variable Threshol at Which Antibiotic Cement No Longer Dominant Strategy (Still Cost- Effective, Incremental Cost-Effectiveness Ratio <$50,000 per QALY) Threshol at Which Antibiotic Cement No Longer Cost-Effective (Incremental Cost-Effectiveness Ratio >$50,000 per QALY) Cost of antibiotic cement >$700 >$1500 Cost of total hip arthroplasty <$17,000 <$0 Cost of revision ue to infection <2.9 total hip arthroplasty cost <0 Cost of aseptic revision <1.1 total hip arthroplasty cost <0 Utility of aseptic revision >0.85 >0.9 Utility of revision ue to infection >1.0 >1.0 Age >73 yr >83 yr Relative risk of aseptic revision <1.2 <1.0 Relative risk of revision ue to infection <1.6 <0.8 analyze. When all revisions were consiere to be the primary outcome measure, it was foun that the cost of a primary total hip arthroplasty woul nee to be <$17,000 before the use of antibiotic-impregnate bone cement is no longer the ominant strategy. The use of antibiotic-impregnate bone cement remains cost-effective for all possible values of the cost of a primary total hip arthroplasty. The cost of an aseptic revision woul nee to be <1.1 times the cost of a primary total hip arthroplasty before the use of antibiotic-impregnate bone cement is no longer the ominant strategy. The aitional cost of antibiotic-impregnate bone cement woul nee to be >$700 before the use of antibioticimpregnate bone cement is no longer the ominant strategy, an the use of antibiotic-impregnate bone cement woul remain cost-effective (<$50,000 per QALY) until its aitional cost was >$1500 (Table III). When only revisions ue to infection were evaluate, it was foun that the cost of a primary total hip arthroplasty woul nee to be >$35,000 before the use of antibiotic-impregnate bone cement becomes the ominant strategy. The cost of a primary total hip arthroplasty woul nee to be <$17,000 before the use of antibiotic-impregnate bone cement is not costeffective. The cost of a revision ue to infection woul nee to be more than seven times the cost of a primary total hip arthroplasty before the use of antibiotic-impregnate bone cement becomes the ominant strategy. The aitional cost of using antibiotic-impregnate bone cement woul nee to be <$400 before the use of antibiotic-impregnate bone cement becomes the ominant strategy. The use of antibiotic-impregnate bone cement remains cost-effective until the aitional cost of antibioticimpregnate bone cement is >$650 (Table IV). Age The average age of the patients within the moel also influence the outcome. As the average age at surgery becomes oler, patients are less likely to live long enough to nee a revision, so the use of antibiotic-impregnate bone cement becomes less costeffective. When all revisions were consiere as the primary outcome, it was foun that the average age of patients woul nee to be greater than seventy-three years before the use of antibiotic-impregnate bone cement is no longer the ominant TABLE IV Sensitivity Analysis with Only Revisions Due to Infection as Primary Outcome Variable Threshol at Which Antibiotic Cement Becomes the Dominant Strategy (Less Costly, More Effective)* Threshol at Which Antibiotic Cement No Longer Cost-Effective (Incremental Cost-Effectiveness Ratio >$50,000 per QALY)* Cost of antibiotic cement <$400 >$650 Cost of total hip arthroplasty >$35,000 <$17,000 Cost of revision ue to infection >7.3 cost of primary total hip arthroplasty <3.5 Cost of aseptic revision NA NA Utility of aseptic revision NA NA Utility of revision ue to infection <0 >0.7 Age <46 yr >71 yr Relative risk of revision ue to infection >2.4 <1.7 *NA = not applicable.

8 640 strategy. The use of antibiotic-impregnate bone cement remains cost-effective until the average age of the patients is greater than eighty-three years (Table III). When only revisions ue to infection were consiere, it was foun that the average age of the patients nees to be less than forty-six years before the use of antibiotic-impregnate bone cement becomes the ominant strategy. The use of antibioticimpregnate bone cement remains cost-effective until the average age of patients excees seventy-one years (Table IV). Age an Cost Age an cost were the two parameters in the moel that ha the greatest influence on the outcome, with changes over a reasonable range of values. Two-way sensitivity analysis was performe on these parameters to emonstrate their interaction with each other. When all revisions were consiere to be the primary outcome measure, it was foun that the cost has to ecrease substantially as the average age of the population increases. For example, if the average age is eighty-five, the cost of antibioticimpregnate bone cement must be <$500 per case to remain a cost-saving strategy. The range for cost-effectiveness (<$50,000 per QALY) is much broaer, as emonstrate in the Appenix. When only revisions ue to infection were use as the primary outcome measure, the parameters were much tighter. With an average age of seventy years, the cost of antibioticimpregnate bone cement woul nee to be <$350 per case to provie cost-savings. If the average age increases to eighty-five years, the cost of antibiotic-impregnate bone cement woul nee to be <$200 to provie cost-savings an woul nee to be <$400 per case to remain cost-effective (see Appenix). Discussion This ecision moel emonstrate that the off-label use of antibiotic-impregnate bone cement is a strategy that is very epenent on the average age of the patients as well as the cost of the antibiotic-impregnate bone cement that is being use. This moel showe antibiotic-impregnate bone cement to be costeffective when the patient population is young (less than seventyone years ol) an the cost of the cement is low (<$650). With the cost of antibiotic-impregnate bone cement still being relatively high in the Unite States an most American patients uner the age of seventy being treate with an uncemente femoral stem, use of antibiotic-impregnate bone cement in primary total hip arthroplasty may be of limite value at this time. The moel is quite sensitive to changes in key parameters when only revision ue to ocumente infection is consiere to be the primary outcome measure. The cost of antibioticimpregnate bone cement in this setting only nees to excee $650 before it becomes cost-inefficient to use it for all primary total hip arthroplasties. Similarly, if the use of antibioticimpregnate bone cement oes not reuce the risk of eep infection by at least 70%, then it is no longer a cost-effective intervention. It is important to keep in min that the infection rates use in the moel were very low (0.7% over ten years for patients treate with stanar bone cement an 0.4% over ten years for those treate with antibiotic-impregnate bone cement) compare with some rates reporte in the literature 7. If infection rates are actually higher, the use of antibiotic-impregnate bone cement woul be a more cost-effective option over a wier range of assumptions regaring costs an outcomes. Also, the risk of infection was assume to be a linear function over time, as oppose to an exponential function with the majority of infections occurring within two years after the initial arthroplasty. The assumption of linearity results in a more conservative estimate of the cost-effectiveness of antibiotic-impregnate bone cement by isplacing the costs an utility eclines associate with revision into the future. The use of antibiotic-impregnate bone cement woul become more cost-effective if more infections occurre within two years, which is likely the case given that the antibiotics are generally fully elute within four weeks. Age also strongly influence the results of the moel. In the analysis in which all revisions were consiere to be the primary outcome measure, the use of antibiotic-impregnate bone cement was foun to cross the $50,000 per QALY threshol when the age of the patients unergoing the primary total hip arthroplasty was greater than eighty-three years. The average age was lower when only revision ue to infection was consiere as the primary outcome measure, with an average age of greater than seventy-one being the threshol. With the majority of total hip arthroplasty implants in the Unite States being uncemente, the average age for the use of a cemente prosthesis may be above these threshols. If the true average age for the use of cement is higher than the threshols note above, then perhaps the use of antibioticimpregnate bone cement is not justifie in that population. The results of this moel were more stable over a broaer range of parameter estimates when all revisions were consiere to be the primary outcome measure. Those analyses emonstrate that the cost of using antibiotic-impregnate bone cement woul nee to be >$1500 before it becomes costinefficient, which is well above the current cost estimates. A potential concern is that antibiotic-impregnate bone cement has been shown to ecrease the rate of aseptic revisions as well as revisions ue to infection 5. One potential explanation is that a low-grae infection that is not etectable by culture is the cause of some aseptic revisions 5. The reason for the ecrease in aseptic revisions is not entirely clear, which is why we also evaluate the moel with only revisions ue to infection as the outcome measure. Limitations of the moel are generally relate to the quality of the ata that are use to evaluate the efficacy of antibioticimpregnate bone cement. The estimations of the revision rates use in this moel were base on ata from registers, which coul have introuce bias into the ata. The Norwegian an Sweish Arthroplasty Registers have continue to show a substantial ifference in infection rates with the use of antibiotic-impregnate bone cement 5,6, but to ate no high-quality ranomize controlle trials have been performe to evaluate the efficacy of antibiotic-impregnate bone cement 23,24, probably because of the large number of patients an extene follow-up that woul be neee for such an investigation. Other concerns about the use of antibiotic-impregnate bone cement are the potentials for an allergic reaction as well

9 641 as for the evelopment of antibiotic-resistant organisms 10,12.To our knowlege, no allergic reactions have been ocumente to ate, but gentamicin, which rarely causes allergic reactions, has been the antibiotic primarily use in bone cement in Europe. The possibility of an allergic reaction may become greater if other antibiotics such as the cephalosporins are use. Although antibiotic resistance is a theoretical concern, there have not been reports of a greater percentage of resistant infections in Europe, where antibiotic-impregnate bone cement is use extensively. These concerns certainly warrant continue surveillance, but currently there is no evience that shouleter one from using antibiotic-impregnate bone cement for primary total hip arthroplasty on those grouns. In summary, the off-label use of antibiotic-impregnate bone cement for primary total hip arthroplasty with cement appears to be a cost-effective strategy if the patient population is young an the cost of the cement is relatively low. This may limit the usefulness of antibiotic-impregnate bone cement in primary total hip arthroplasty in the Unite States, given its current cost an the oler average age of patients being treate with cemente femoral stems. In our moel, we evaluate costs from a hospital resource-use perspective, which was chosen to etermine if the aitional up-front cost of antibiotic-impregnate bone cement was justifie by a future ecrease in costs for revisions. The results can potentially be use by surgeons an policy-makers to help ecie whether the use of antibioticimpregnate bone cement for primary total hip arthroplasty is justifie in view of the aitional costs, current infection rates, an average age of the population in which a cemente prosthesis is use. Appenix A figure showing the Markov moel an figures emonstrating the results of the sensitivity analyses of age versus cost are available with the electronic versions of this article, on our web site at jbjs.org (go to the article citation an click on Supplementary Material ) an on our quarterly CD/ DVD (call our subscription epartment, at , to orer the CD or DVD). n Justin S. Cummins, MD, MS Ivan M. Tomek, MD Stephen R. Kantor, MD Samuel R.G. Finlayson, MD, MPH Departments of Orthopaeic Surgery (J.S.C., I.M.T., an S.R.K.) an Surgery (S.R.G.F.), Dartmouth Hitchcock Meical Center, One Meical Center Drive, Lebanon, NH aress for J.S. Cummins: jcummins.m@gmail.com Ove Furnes, MD, PhD Lars Birger Engesæter, MD, PhD Department of Orthopaeic Surgery, Haukelan University Hospital, N-5021 Bergen, Norway References 1. Ligren L. Joint prosthetic infections: a success story. Acta Orthop Scan. 2001;72: Walenkamp G. Surveillance of surgical-site infections in orthopeics. Acta Orthop Scan. 2003;74: Bourne RB. Prophylactic use of antibiotic bone cement: an emerging stanar in the affirmative. J Arthroplasty. 2004;19(4 Suppl 1): Espehaug B, Engesaeter LB, Vollset SE, Havelin LI, Langelan N. Antibiotic prophylaxis in total hip arthroplasty. Review of 10,905 primary cemente total hip replacements reporte to the Norwegian Arthroplasty Register, 1987 to J Bone Joint Surg Br. 1997;79: Engesaeter LB, Lie SA, Espehaug B, Furnes O, Vollset SE, Havelin LI. Antibiotic prophylaxis in total hip arthroplasty: effects of antibiotic prophylaxis systemically an in bone cement on the revision rate of 22,170 primary hip replacements followe 0-14 years in the Norwegian Arthroplasty Register. Acta Orthop Scan. 2003;74: Malchau H, Herberts P, Ahnfelt L. Prognosis of total hip replacement in Sween. Followup of 92,675 operations performe Acta Orthop Scan. 1993;64: Ahnfelt L, Herberts P, Malchau H, Anersson GB. Prognosis of total hip replacement. A Sweish multicenter stuy of 4,664 revisions. Acta Orthop Scan Suppl. 1990;238: Bourne RB. Antibiotic bone cement approval: fuss-in at the Fes! Orthopeics. 2002;25: Hanssen AD. Prophylactic use of antibiotic bone cement: an emerging stanar in opposition. J Arthroplasty. 2004;19(4 Suppl 1): Joseph TN, Chen AL, Di Cesare PE. Use of antibiotic-impregnate cement in total joint arthroplasty. J Am Aca Orthop Surg. 2003;11: Arciola CR, Campoccia D, Montanaro L. Effects on antibiotic resistance of Staphylococcus epiermiis following ahesion to polymethylmethacrylate an to silicone surfaces. Biomaterials. 2002;23: Tunney MM, Ramage G, Patrick S, Nixon JR, Murphy PG, Gorman SP. Antimicrobial susceptibility of bacteria isolate from orthopeic implants following revision hip surgery. Antimicrob Agents Chemother. 1998;42: Hunink MGM, Glasziou P, Siegel JE, Weeks J, Pliskin J, Elstein AS, Weinstein M. Decision making in health an meicine: integrating evience an values. Cambrige: Cambrige University Press; Gol MR, Siegel JE, Russell LB, Weinstein MC, eitors. Cost-effectiveness in health an meicine. New York: Oxfor University Press; Persson U, Persson M, Malchau H. The economics of preventing revisions in total hip replacement. Acta Orthop Scan. 1999;70: Laupacis A, Bourne R, Rorabeck C, Feeny D, Wong C, Tugwell P, Leslie K, Bullas R. The effect of elective total hip replacement on health-relate quality of life. J Bone Joint Surg Am. 1993;75: Rorabeck CH, Bourne RB, Mulliken BD, Nayak N, Laupacis A, Tugwell P, Feeney D. Comparative results of cemente an cementless total hip arthroplasty. Clin Orthop Relat Res. 1996;325: Hozack WJ, Rothman RH, Albert TJ, Balerston RA, Eng K. Relationship of total hip arthroplasty outcomes to other orthopaeic proceures. Clin Orthop Relat Res. 1997;344: Antoniou J, Martineau PA, Filion KB, Haier S, Zukor DJ, Huk OL, Pilote L, Eisenberg MJ. In-hospital cost of total hip arthroplasty in Canaa an the Unite States. J Bone Joint Surg Am. 2004;86: Barrack RL. The evolving cost spectrum of revision hip arthroplasty. Orthopeics. 1999;22: Bozic KJ, Ries MD. The impact of infection after total hip arthroplasty on hospital an surgeon resource utilization. J Bone Joint Surg Am. 2005;87: Bozic KJ, Katz P, Cisternas M, Ono L, Ries MD, Showstack J. Hospital resource utilization for primary an revision total hip arthroplasty. J Bone Joint Surg Am. 2005;87: McQueen M, Littlejohn A, Hughes SP. A comparison of systemic cefuroxime an cefuroxime loae bone cement in the prevention of early infection after total joint replacement. Int Orthop. 1987;11: McQueen MM, Hughes SP, May P, Verity L. Cefuroxime in total joint arthroplasty. Intravenous or in bone cement. J Arthroplasty. 1990;5: