The cost-effectiveness of antibiotic-loaded bone cement versus plain bone cement following total and partial knee and hip arthroplasty

Abstract

Background

Postoperative infection is one of the most prevalent complications following total joint arthroplasty (TJA). As such procedures become more prevalent, it is imperative that we develop new prophylactic methods to prevent the need for revision procedures. In recent years, surgeons have opted to use antibiotic-loaded bone cement (ALBC) rather than plain bone cement (PBC) in primary hip and knee replacements due to its theoretical potential of lowering infection rates. However, the cost-effectiveness of this intervention remains in question.

Questions/Purposes: To determine the rate of infection and cost-effectiveness of antibiotic-loaded bone cement as compared to plain bone cement in hip and knee arthroplasty.

Patients and methods

We reviewed 4116 primary hip and knee arthroplasty cases performed between 2016 and 2018 at Morristown Medical Center in New Jersey. Data regarding demographics, complications, and any readmissions due to deep infection were collected retrospectively. During that time period there were a total of 4016 knee cases (423 ALBC, 3593 PBC) and 123 hip cases (63 ALBC, 60 PBC). The average cost for one bag of antibiotic-loaded bone cement and plain bone cement for hip and knee arthroplasty was $336.42 and $72.14, respectively. A statistical analysis was performed using Fisher's exact test; the National Healthcare Safety Network (NHSN) surgical site infection guidelines were used to distinguish between superficial and deep infections.

Results

Ten patients were readmitted due to deep infection, all of whom had undergone total knee arthroplasty. Of those cases, plain bone cement was used for the index procedure in seven instances and antibiotic-loaded cement was used in three. This resulted in an infection rate of 0.19% and 0.62%, respectively, p = 0.103. There was no statistically significant difference in infection rates between the two groups. A total of 778 bags of ALBC were used in 423 knee surgeries, and 98 bags of ALBC were used in 63 hip cases. The total cost for ALBC in TKA and THA procedures was $261,734.76 (778*336.42) and $32,969.16 (98*336.42), respectively. If PBC had been used during all index procedures, it would have resulted in a total savings of $231,509.28.

Conclusions

Antibiotic-loaded cement did not significantly reduce the rate of infection for either knee or hip arthroplasty. Thus, the routine use of antibiotic-loaded cement in primary hip and knee arthroplasty may be an unnecessary financial burden to the healthcare system. A larger sample size and a randomized controlled trial would help confirm our findings and would provide further information on the cost-effectiveness of ALBC cement versus PBC.

Significance/Clinical Relevance: In this review of cases performed from 2016 to 2018 there was no statistically significant difference between the rate of infection and the need for revision surgeries for patients treated with ALBC versus PBC. As hospital systems continue to transition towards a bundled payment model, it becomes imperative for providers to reduce any unnecessary costs in order to increase quality and efficiency. We estimate that our hospital system could save nearly $120,000/year by using plain bone cement instead of antibiotic-loaded cement.

1. Introduction

Periprosthetic joint infection (PJI) following hip and knee arthroplasty is a serious complication. In order to decrease the incidence of infection following these procedures, new prophylactic interventions and procedural modifications have been implemented. Antibiotic-loaded bone cement (ALBC) was first introduced and used in lieu of plain bone cement (PBC) by Buchholz and Englebrecht in 1970.^1,2 In a 1981 follow-up study, they reported a 77% success rate among 583 patients undergoing total hip arthroplasty (THA). Success was defined as no infection, no loosening of hardware, and useful function of the joint.³ Subsequent studies have investigated the efficacy of this intervention and have reported similar findings.²

Although the use of ALBC in revision total joint surgery is well supported by the literature, the regular use of this type of intervention in routine primary total joint arthroplasty (TJA) remains controversial. Several potential adverse complications have emerged, including the development of antibiotic-resistant strains of bacteria, antibiotic toxicity, a negative effect on the mechanical properties of the cement, and the economic implications of the added cost. Studies have reported that at high doses of antibiotics (>4.5 g of antibiotic powder per 40 g of cement), mechanical complications such as hardware loosening can occur much more frequently.^2,4, 5, 6 While the primary goal of ALBC is to deliver high, localized concentrations of the antibiotic, the risk of systemic exposure and toxicity remain concerning. In the past decade there have been a number of literature reports involving patients who received ALBC during a total knee or hip arthroplasty and subsequently experienced acute renal failure (ARF).7, 8, 9, 10, 11

The economic implications of this intervention also continue to be scrutinized. The price for one bag of ALBC can range from $200 to $500, while PBC costs under $100.^4,12,13 Jiranek et al. reported an estimated increase of $117 million dollars to the annual healthcare costs if ALBC is used in only half of the primary joint arthroplasty cases.⁴ In order to justify the incorporation of ALBC into routine primary joint arthroplasty, the cost savings must be greater than if PBC was used alone. This is contingent upon preventing subsequent revision surgeries due to deep infection. The goals of this study were to examine the cost-effectiveness and the rate of deep infection using ALBC versus PBC for patients treated at our facility for either a total knee arthroplasty (TKA), partial knee arthroplasty (UKA) or THA. Specifically, we aimed to look at whether or not the additional cost of ALBC would be justified by the expected reduction in deep infections.

2. Methods

A retrospective review was conducted on all cemented total and partial hip and knee arthroplasty cases performed between January 1, 2016 to December 31, 2018 at a large, suburban, regional medical center. Patient demographics, comorbidities, and any readmissions due to deep infection were extracted from patient charts. Data were also collected on patients who were readmitted to outside facilities for deep infections after their index procedure was done at our institution during the study period. For each procedure, either Stryker (Mahwah, NJ) or Zimmer-Biomet (Parsippany, NJ) bone cement was used. Choice of cement was based on the surgeon's discretion. One surgeon in particular routinely utilized ALBC for all primary joint replacements.

Inclusion criteria for the study were as follows: patients treated at our facility with a cemented primary total or partial hip or knee arthroplasty. Exclusion criteria included: revision arthroplasty patients and uncemented arthroplasty cases. A statistical analysis was performed using Fisher's exact test; the National Healthcare Safety Network (NHSN) surgical site infection guidelines were used to distinguish between superficial and deep infections. Demographics such as gender, age, body mass index (BMI), American Society of Anesthesiologists Score (ASA), and LOS (length of stay) were included in our analysis. Medical comorbidities such as diabetes, chronic obstructive pulmonary disease (COPD), congestive heart failure (CHF), renal disease, and obesity were also identified for each cohort (Table 1).

Table 1.

Patient comorbidities.

	Antibiotic-Loaded Cement Knee	Antibiotic-Loaded Cement Hip	Plain Bone Cement Knee	Plain Bone Cement Hip	Totals
Diabetes	107	12	436	3	558
COPD	16	2	93	4	115
Renal Disease	16	2	78	1	97
CHF	8	4	48	2	62
Obesity	183	10	913	10	1116

3. Results

During the study period there were a total of 4016 cemented knee cases: 423 ALBC (408 TKA, 15 UKA) and 3593 PBC (3561 TKA, 32 UKA). There were also 123 cemented THA cases: 63 ALBC and 60 PBC (Table 2). There were no cases of cemented hip hemiarthroplasty during the study period. The average age was 67.92 (range, 36 to 94) with 39.94% of patients being male and 60.06% being female. Average BMI, ASA, and LOS were 31.62 (range, 15 to 61.1), 2.32 (range, 1 to 4), and 2.28 days (range, 0 to 21), respectively. The large discrepancy in case volume between hip and knee surgeries can be attributed to the general U.S. trend of orthopedic surgeons using uncemented implants during routine THA.

Table 2.

Total cases.

Cases	Antibiotic-Loaded Cement	Plain Bone Cement	Totals
Total Knee Arthroplasty	344	2677	3021
Bilateral Total Knee Arthroplasty	40	665	705
Unicondylar Knee Arthroplasty	39	251	290
Total Hip Arthroplasty	12	34	46
Hip Resurfacing	24	10	34
Bipolar Hip Hemiarthroplasty	27	16	43
All Cases	486	3653	4139

Ten patients were readmitted due to deep infection, all of whom had undergone TKA. Of the cases reported, PBC was used during the index procedure in 7 patients (0.19%) and ALBC was used in 3 patients (0.62%). At our institution, the average cost for one bag of PBC is $72.14; the average cost for one bag of ABLC is $336.42. The discrepancy in price is due to the pricing differences between the two different brands of PBC and ALBC available at our institution. A total of 778 bags of ALBC were used in 423 knee surgeries (TKA, UKA, bilateral TKA) and 98 bags of ALBC were used in 63 THA cases. The total cost for ALBC in TKA and THA procedures was $261,734.76 (778*336.42) and $32,969.16 (98*336.42), respectively. If PBC had been used during all index procedures, it would have resulted in a total savings of $231,509.28. There were no infections in our UKA, bilateral TKA, and THA groups. There were no statistically significant differences between the various comorbidities with respect to infection rates between the PBC and ALBC cohorts (Table 3).

Table 3.

Results.

	Antibiotic-Loaded Cement	Plain Bone Cement	P-Values
All Cases	3/486	7/3653	0.103
All Knee Cases	3/423	7/3593	0.079
All Hip Cases	0/63	0/60	>0.999
Total Knee Arthroplasty	3/344	7/2677	0.081
Bilateral Total Knee Arthroplasty	0/40	0/665	>0.999
Unicondylar Knee Arthroplasty	0/39	0/251	>0.999
Total Hip Arthroplasty	0/12	0/34	>0.999
Hip Resurfacing	0/24	0/10	>0.999
Bipolar Hip Hemiarthroplasty	0/27	0/16	>0.999
All Diabetic Patients	1/119	0/439	0.213
All COPD Patients	0/18	0/97	>0.999
All CHF Patients	0/12	1/50	>0.999
All Renal Disease Patients	0/18	0/79	>0.999
All Obesity Cases	3/193	6/923	0.192

4. Discussion

The most important finding of this study is that there was no statistically significant difference between the rate of deep infection and the need for revision surgery in patients undergoing either primary knee or hip arthroplasty who were treated with ALBC versus PBC. The effectiveness of ALBC in primary hip and knee joint replacement remains a subject of controversy in the literature. As hospital systems transition towards a bundled payment model, the answer to this debate becomes even more essential, and may offer a significant avenue towards cost savings while at the same time maintaining the quality of care rendered.

With regard to ALBC in primary THA, a majority of the support for its use comes from early European registry data studies, where cemented THA is much more common than in the United States.^14,15 More recently, while there are no adequately powered randomized control trials in the literature addressing ALBC for primary THA, systematic reviews continue to support the use of ALBC to reduce the risk of deep periprosthetic infection in THA.16, 17, 18 However, the results of this paper directly contradict these findings, and did not demonstrate any statistically significant reduction in deep infection rates in patients undergoing primary THA. This is important, as most of the studies specifically looking at ALBC in primary THA are almost a decade old and rely on national registry data that have the potential limitations of selection bias due to incomplete data capture.

The authors of this paper also did not find a statistically significant difference in the rate of deep infection and subsequent revision surgery between ALBC and PBC in primary TKA, bilateral primary TKA or UKA. In the early 2000s, several authors, including Chiu et al. and Eveillard et al., were the first to suggest that ALBC could reduce PJI rates in primary TKA.^19,20 However, these studies had several limitations, including borderline statistical significance as well as the failure to address confounding comorbidities such as diabetes. The present study demonstrated adequate control for confounding comorbidities (CHF, COPD, renal disease, diabetes) and variables (age, gender, BMI) while still demonstrating a lack of efficacy of ALBC. More recent studies, including a randomized control trial by Hinarejos et al. and other larger case series, corroborate the present study's findings.^17,21, 22, 23 They also failed to demonstrate a decrease in infection rates when using ALBC versus PBC. More recent registry studies, systematic reviews, and meta-analyses also have arrived at similar conclusions.^13,16,24

Several studies have commented on the idea of risk-stratified usage of ALBC, especially in patients with diabetes and rheumatoid arthritis.^19,25 However, a more recent study by Qadir et al. found no advantage of ALBC over PBC in high-risk patients, such as those with inflammatory arthropathies, diabetes, and immunosuppression.²⁶ This was supported by our research, as there was no statistically significant difference between infection rates in the higher risk patients defined by the medical comorbidities analyzed in our study.

The routine use of ALBC in TKA and THA leads to a significant increase in cost as compared to PBC. Much of this is due to the relatively increased cost of ALBC in the United States specifically. At our facility there was a nearly $300 difference in cost between ALBC and PBC, leading to a potential loss in savings of $231,509.28 over two years. Studies have continued to fail to demonstrate its cost-effectiveness, specifically in the United States. Although certain models involving cheaper hand-mixed cement, younger patients, a higher overall infection rate, and the inclusion of revisions for aseptic loosening have theoretically been able to demonstrate the cost-effectiveness of ALBC, the same authors concluded that it would not be prudent to extrapolate this data to the current United States demographics and ALBC prices.^12,27 Arguably, any single episode of PJI carries a catastrophic increase in perioperative claims costs. However, the 1% absolute risk reduction of PJI by ALBC that would be needed to justify its cost, as described in the current literature, is already much higher than our institution's baseline PJI rate.^23,28 Without any strong cost-effectiveness evidence indicating the role of ALBC in reducing PJI in primary TKA, UKA and THA, the routine use of ALBC may not be justified.

There were several limitations to our study. The lack of patient randomization to either the PBC or ALBC cohorts and the retrospective nature of the study inherently make it difficult to account for all confounding variables. Furthermore, this study included patients from multiple surgeons, each with different surgical techniques and postoperative protocols. A more standardized perioperative protocol could help decrease the confounding factors. However subgroup statistical analysis was performed for certain demographics and comorbidities such as diabetes, COPD, CHF, etc. which showed no significant difference in infection rates. No patients were lost to follow up within the first 90 days. Patients who may have presented to outside hospitals with their deep infections were still kept track of by our department of infection control. Lastly, while our study included over 4000 patients, the authors realize it may not be sufficiently powered to detect the slight difference in PJI rates. In addition, the TJA infection rate of 0.24% at our institution is eight times less than the national average of 2%. While this rate may not be reproducible at all hospital settings, what can be reproduced is the cost-saving benefits of using PBC. Despite these limitations, the authors feel that these results can be broadly applied to all settings where hospitals have surgeons with varying training, protocols, preferences, and techniques. Further studies with larger patient volumes, randomization, and standardization of surgical and postsurgical protocols are needed to further validate the conclusions of this study.

5. Conclusion

The use of ALBC in primary TKA varies by country. The percentage of surgeons who routinely use it in primary TKA is >90% in some countries, such as the United Kingdom, Norway, and Sweden, compared with approximately 10% of surgeons in other countries such as the United States.^3,7,14,15 The present study demonstrates that there was no statistically significant difference in infection rates or in the need for revision surgery due to deep infection for patients with primary TKA, UKA, and THA who were treated with ALBC as opposed to PBC. The use of ALBC may add an unnecessary hospital expense in this setting. We estimate that our hospital system could have saved nearly $120,000 per year if PBC was used in lieu of ALBC for each index procedure. Given the rising necessity of hospitals to allocate funds sparingly under bundled care models, it would be valuable to have randomized controlled trials to confirm our findings and further evaluate the cost-effectiveness of ALBC versus PBC.

Author contributions

TH, JS, JP, and JW helped design the study and write the manuscript. CM, DG, EP, and TC II all assisted with data curation.

All surgeries and data collection for this study were performed at Morristown Medical Center.

Declaration of competing interest

“Each author certifies that he or she has no commercial associations (eg, consultancies, stock ownership, equity interest, patent/licensing arrangements, etc) that might pose a conflict of interest in connection with the submitted article.”