No Difference in 30-day Mortality Between Patients Undergoing Bilateral Simultaneous Total Knee Arthroplasty With Technology Assistance Compared to Conventional Instrumentation

Abstract

Background: Bilateral simultaneous total knee arthroplasty (BSTKA) has decreased in frequency due to concerns about higher rates of early mortality and complications than unilateral or staged surgeries. Purpose: We sought to evaluate whether technology assistance (encompassing robotics and computer assistance) decreases early mortality following BSTKA. Methods: We conducted a retrospective cohort study using a national all-payer claims database. Patients who underwent BSTKA from October 2015 to December 2020 were identified. Univariate and multivariable analyses were conducted to compare outcomes in patients who underwent BSTKA with technology assistance compared to conventional instrumentation. The primary outcome was 30-day postoperative mortality. Secondary outcomes were respiratory failure and fat embolism. A post-hoc analysis was performed to evaluate length of stay, readmission, and other medical complications. Results: A total of 14,870 patients who underwent BSTKA were included in this study. Of these, 860 patients underwent technology-assisted BSTKA, and 14,010 patients underwent BSTKA without technology assistance. After a multivariable analysis, patients who underwent technology-assisted BSTKA had equivalent odds of 30-day mortality compared to those who underwent BSTKA without technology assistance. Technology assistance was not protective against the development of acute respiratory failure or fat embolism. Conclusion: This retrospective cohort study found no differences in the rates of 30-day mortality, respiratory failure, or fat embolism after technology-assisted BSTKA compared to conventional BSTKA. On the post-hoc analysis, technology use was associated with a decreased length of stay, lower readmission risk, and decreased rates of deep vein thrombosis, pulmonary embolism, and blood transfusion.

Introduction

For patients with end-stage arthritis of the knee, total knee arthroplasty (TKA) has been increasingly used to provide pain relief and improve function. Predictive models suggest that it will become even more common in the next 10 years [29]. Knee osteoarthritis is often either bilateral at presentation or it progresses in the contralateral extremity. In fact, 1 in 4 patients who undergo a single TKA choose to undergo contralateral TKA within 5 to 8 years [18].

Patients who have exhausted nonoperative treatments for bilateral knee arthritis must decide with their surgeons whether bilateral simultaneous TKA (BSTKA) or staged TKA is most appropriate. Advantages cited in favor of BSTKA over staged TKA include decreased time off work, lower cost, and greater convenience [10,12]. However, some studies have found that BSTKA increases the risk of perioperative mortality and complications [1,20,31]. Only offering BSTKA to younger patients with fewer comorbidities may reduce adverse perioperative outcomes [6].

The use of computer and robotic arm assistance in TKA has recently increased. As of 2018, nearly 8% of TKA surgeries used technology assistance, and about 5% of these used a robotic system [2]. The effects of these technologies on patient outcomes are currently under investigation [2,9]. One possible benefit of computer assistance and robotics in BSTKA is that these systems generally do not rely on instrumentation of the femoral canal, which may contribute to increased blood loss and risk of fat embolism [28]. A recent retrospective study of the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) of patients who underwent BSTKA found a substantial improvement in 30-day mortality (0.062% vs 0.57%; adjusted odds ratio [OR] 0.26; 95% confidence interval [CI] 0.09–0.74) and 90-day mortality (0.08% vs 0.22%; adjusted OR 0.26; 95% CI 0.10–0.67) among patients who underwent technology-assisted BSTKA [17]. However, this finding has not been validated in other large database studies, and the etiology for this mortality benefit has not been elucidated.

Using a national insurance database, we sought to evaluate whether the use of technology assistance decreases the rate of early mortality in patients undergoing BSTKA. Secondary outcomes were the rates of respiratory failure and fat embolism with technology-assisted BSTKA compared with conventional BSTKA. We hypothesized that technology-assisted BSTKA would be associated with reduced early mortality by decreasing the rate of respiratory failure and fat embolism associated with BSTKA.

Methods

A retrospective cohort study was conducted using data from October 2015 to December 2020 from the PearlDiver patient records database (www.pearldiverinc.com). Specifically, the Mariner data set was used; it includes payer claims data including commercial insurance, Medicare, and Medicaid in the United States. The data set follows patients over time based on distinct identifiers, minimizing loss to follow-up in the system. Patients who underwent BSTKA were identified using International Classification of Diseases 10 (ICD-10) procedure codes. In this analysis, ICD-9 procedure codes were not included because they cannot control for laterality and, thus, could not distinguish patients who received BSTKA. Patients who underwent BSTKA with computer or robotic-arm assistance were identified using the Current Procedural Terminology or ICD codes on the same date as BSTKA. Patients who underwent TKA with computer or robotic-arm assistance were collectively referred to as having undergone TKA with technology assistance to encompass both categories. Patients who underwent BSTKA without computer or robotic-arm assistance codes documented were assumed to have undergone the procedure with conventional instrumentation. The specific billing codes used to identify TKA and computer or robotic-arm assistance are provided in Supplement 1. Because PearlDiver releases only deidentified information, this study was exempt from institutional review board approval.

Patients were excluded if they were younger than 18 years at the time of TKA, had a traumatic arthropathy, a prior malignancy, or an unknown laterality of TKA. Patients were included if they underwent BSTKA for bilateral osteoarthritis and had active follow-up of 30 days from the index procedure. Patients without 30 days of follow-up, as determined using the patient manifest, were assumed to be lost to follow-up and were excluded.

Over the study period, a total of 14,870 patients underwent BSTKA and were eligible for inclusion. Of these patients, 860 patients underwent technology-assisted BSTKA, and 14,010 patients underwent BSTKA without technology assistance.

Demographic characteristics collected included age, gender, and the Elixhauser comorbidities for each cohort [5]. The primary outcome for this study was 30-day mortality. Secondary outcomes included incidence of respiratory failure and fat emboli within 30 days postoperatively. To evaluate other previously identified associations between technology-assisted TKA and postoperative complications, a post-hoc analysis was performed to assess possible differences in length of stay, readmission, blood transfusion, pneumonia, deep vein thrombosis (DVT), pulmonary embolism (PE), delirium, and cerebrovascular accident (Supplement 1) [14,23]. Lastly, the timing of the 30-day mortality was divided into 10-day intervals.

Statistical Analysis

Univariate analysis using the R software (R Foundation for Statistical Computing) provided by PearlDiver was used to analyze any differences in patient demographics, comorbidities, complications, and outcomes. This was conducted using chi-squared tests for categorical variables and Student’s t tests for continuous variables where appropriate. To mitigate confounding variables and covariates, all demographic variables and comorbidities with P < .2 on univariate analysis were included as independent variables for each multivariable analysis. All outcomes and complications with P < .2 were included as dependent variables for separate multivariable analyses. Using the PearlDiver software, logistic regression was conducted for multivariable analysis of the indicated variables; P < .05 was used as the level of significance. For confidentiality purposes, PearlDiver does not permit the reporting of count data less than 11. Any outcome with a patient number less than 11 was reported as <11, but the appropriate percentage was reported.

Results

Those who underwent technology-assisted BSTKA were significantly older (P < .001) and more likely to have hypertension (P = .001) and chronic kidney disease (P = .004) than those who underwent conventional BSTKA. However, patients who underwent technology-assisted BSTKA had lower rates of valvular heart disease (P < .001), peripheral vascular disease (P = .007), hypothyroidism (P = .021), depression (P = .038), and obesity (0.006). Other comorbidities did not differ significantly across cohorts (P > .05; Table 1).

Table 1.

Demographics and comorbidities of patients undergoing BSTKA.

	Total (n)	Technology assisted	Technology assisted (%/SD)	Conventional (n)	Conventional (%)	P value
Overall	14.870	860	-	14,020	-	-
Mean age (years)	-	64.76	-	63.48	-	<.001
Gender	-	-	-	-	-	.104
Male	6682	410	52.36%	6272	45.99%
Female	8188	450	57.47%	7738	56.74%
CHF	800	45	5.75%	755	5.54%	.905
Arrhythmias	3072	165	21.07%	2907	21.32%	.291
Valvular disease	1656	65	8.30%	1591	11.67%	<.001
Pulmonary circulatory disorders	537	28	3.58%	509	3.73%	.630
PVD	1420	59	7.54%	1361	9.98%	.007
HTN	1867	271	34.61%	1596	11.70%	.001
Paralysis	-	<11	0.26%	120	0.88%	.076
Other neurological disorders	423	17	2.17%	406	2.98%	.141
CPD	2481	133	16.99%	2348	17.22%	.347
Diabetes mellitus	2576	135	17.24%	2441	17.90%	.211
Hypothyroidism	2433	116	14.81%	2317	16.99%	.021
CKD	929	74	9.45%	855	6.27%	.004
Liver disease	1306	62	7.92%	1244	9.12%	.106
PUD	225	15	1.92%	210	1.54%	.669
Rheumatoid arthritis and CVD	1420	78	9.96%	1342	9.84%	.665
Coagulopathy	1390	66	8.43%	1324	9.71%	.094
Fluid and electrolyte disorders	2568	132	16.86%	2436	17.86%	.137
Blood loss anemia	586	31	3.96%	555	4.07%	.666
Deficiency anemia	1350	73	9.32%	1277	9.36%	.576
Alcohol abuse	-	<11	0.38%	61	0.45%	.914
Drug abuse	504	20	2.55%	484	3.55%	.093
Psychoses	-	<11	0.77%	136	1.00%	.536
Depression	2959	147	18.77%	2812	20.62%	.038
Smoking	1313	72	9.20%	1241	9.10%	.67
Obesity	4216	208	26.56%	4008	29.39%	.006

BSTKA bilateral simultaneous total knee arthroplasty, CHF congestive heart failure, PVD peripheral vascular disease, HTN hypertension, CPD cardiopulmonary disease, CKD chronic kidney disease, PUD peptic ulcer disease, CVD collagen vascular disease, <11 PearlDiver does not permit the reporting of count data with less than 11 patients. Bold values indicate statistically significant P-values < .05.

On univariate analysis, technology-assisted BSTKA was associated with equivalent likelihood of 30-day mortality (0.70% vs 0.51%; P = .48), respiratory failure (1.05% vs 0.54%; P = .052), and fat embolism (0.00% vs 0.04%; P = .579) compared with conventional BSTKA. The greatest number of deaths occurred in the 10 days following surgery. Patients who underwent technology-assisted BSTKA had a decreased length of stay (2.97 vs 3.38 days; P < .001) and lower odds of readmission (5.81% vs 9.54%; P < .001). Regarding other complication rates, technology-assisted BSTKA was associated with lower rates of DVT (1.40% vs 2.47%; P = .046), PE (0.47% vs 1.26%; P = .038), and blood transfusion (11.98% vs 17.64%; P < .001). Other complication rates were similar between groups (Table 2).

Table 2.

Univariate analysis of outcomes and complications following BSTKA with and without robotic assistance.

	Total (n)	Technology assisted (n)	Technology assisted (%)	Conventional (n)	Conventional (%)	P value
LOS	N/A	2.97	-	3.38	-	<.001
Readmission	1386	50	5.81%	1336	9.54%	<.001
Respiratory failure	a	<11	1.05%	75	0.54%	.052
Pneumonia	a	<11	0.81%	138	0.99%	.62
DVT	358	12	1.40%	346	2.47%	.046
PE	a	<11	0.47%	177	1.26%	.038
Blood Transfusion	2,575	103	11.98%	2472	17.64%	<.001
Delirium	a	<11	0.70%	134	0.96%	.446
CVA	a	<11	0.47%	60	0.43%	.873
Fat emboli	a	0	0.00%	<11	0.04%	.579
30-Day death	a	<11	0.70%	72	0.51%	.479
10 Days (acute)	a	<11	0.47%	38	0.27%	.298
20 Days (middle)	a	<11	0.12%	22	0.16%	.768
30 Days (delayed)	a	<11	0.12%	12	0.09%	.768

BSTKA bilateral simultaneous total knee arthroplasty, CVA cerebrovascular accident, DVT deep vein thrombosis, PE pulmonary embolism, <11 PearlDiver does not permit the reporting of count data with less than 11 patients.

^aDisclosing the exact number would cause disclosure of count data with less than 11 patients.

On multivariable analysis, patients who underwent technology-assisted BSTKA had comparable odds of 30-day mortality (OR 1.82; 95% CI 0.63–4.23; P = .207) and respiratory failure (OR 1.77; 95% CI 0.82–3.38; P = .108) to those who underwent surgery with conventional instrumentation. Technology-assisted BSTKA was also associated with lower odds of readmission (OR 0.45; 95% CI 0.32–0.61; P < .001), shorter length of stay (OR 0.64; 95% CI 0.54–0.75; P < .001), decreased use of blood transfusions (OR 0.44; 95% CI 0.31–0.62; P < .001), and lower odds of DVT (OR 0.41; 95% CI 0.19–0.75; P = .008) and PE (OR 0.34; 95% CI 0.11–0.81; P = .008; Table 3).

Table 3.

Multivariable analysis of complications following BSTKA with and without technology assistance.

Technology assisted vs conventional BSTKA
	Odds ratio	95% CI	P value
Mortality	1.825	0.628-4.226	.207
Readmission	0.446	0.318-0.609	<.001
LOS	0.640	0.545-0.752	<.001
DVT	0.408	0.194-0.747	.008
PE	0.344	0.106-0.814	.008
Respiratory Failure	1.774	0.822-3.381	.108
Blood transfusion	0.445	0.309-0.618	<.001

BSTKA bilateral simultaneous total knee arthroplasty, CI confidence interval, DVT deep vein thrombosis, PE pulmonary embolism, LOS length of stay.

Discussion

Patients with bilateral knee osteoarthritis who opt for surgical management may be interested in BSTKA because of its increased convenience, high patient satisfaction, lower costs, and decreased time off work [10,12,25]. However, the concern that BSTKA places patients at increased risk of postoperative complications and mortality has led to decreased use of this procedure [3,8,19,26,32]. A recent study using the AOANJRR found that technology-assisted BSTKA was associated with a significant reduction in early mortality compared to BSTKA with conventional instrumentation; this difference was possibly related to reduced fat emboli [17]. Using a national database with data from multiple payers, our study did not find an association between the use of technology assistance and reduced mortality after BSTKA. Furthermore, no association was observed between the use of technology assistance and postoperative respiratory failure and fat embolism.

There are limitations to this study. First, as a retrospective analysis of an administrative claims database, the data are limited by the accuracy of coding. For example, it is possible that in some cases, technology assistance was used but not coded. This database allows for the observation of postoperative but not intraoperative complications, which could cause the incidence of certain complications to be underestimated. While a multivariable analysis was performed using comorbidities commonly cited to contribute to outcomes after arthroplasty procedures, it is possible that other unaccounted-for risk factors could confound our findings. Also, because this analysis only used 30-day complications data, longitudinal differences in outcomes beyond 30 days could not be assessed. This database cannot isolate institutional or surgeon-level data to compare outcomes between conventional and technology-assisted TKA in the absence of surgeon or institution-dependent factors. A randomized multicenter study would be necessary to resolve this limitation. Finally, while commonly cited codes for technology-assisted TKA were used to identify these patients, the specific instrumentation used in each case was not available, and it is possible that a surgeon could have used intramedullary instrumentation as well.

This study also has several important strengths. It drew from a large sample of patients, including those with a range of different insurance types. Additionally, the PearlDiver database allowed patients to continue to be followed up beyond discharge and included presentations at hospitals other than that where the original surgery was performed. This helped to optimize follow-up and improved accuracy of the data. This study also allowed for the evaluation of more variables than the prior Australian registry study evaluating the association of technology assistance with mortality after BSTKA, which included only age, sex, procedure year, American Society of Anesthesiologists classification, and body mass index in the logistic regression [17].

Several large database and national registry studies have documented 30-day mortality rates under 1% following BSTKA [3,17,30,32]. Using the PearlDiver Mariner data set, we found a comparable mortality rate of 0.52% in this patient cohort. The specific causes of mortality after BSTKA are poorly characterized. In the early postoperative period after arthroplasty in general, ischemic heart disease is the most common cause of death, and deaths associated with digestive system–related diseases also increased [11]. The cause of death was not available in the data set used for our study. Factors previously associated with mortality after knee arthroplasty include primary knee surgery, advanced age, preexisting cardiopulmonary disease, use of a cemented prosthesis, and bilateral surgery under the same anesthesia [24]. After a multivariable analysis, technology assistance was not found to be protective against mortality after BSTKA in this study. Notably, the prior Australian registry study that found decreased mortality with technology-assisted TKA used fewer covariates in the regression analysis [17].

The association between instrumentation of the femoral canal and adverse events following TKA has been another area of interest in the literature. Compared to conventional instrumentation, computer-assisted TKA without an intramedullary femoral alignment jig decreases embolic loads as evaluated with transesophageal echocardiography (TEE) or a pulmonary artery catheter [4,21]. However, the clinical significance of fat emboli after TKA has been questioned [22]. Decreased embolic loads were the proposed mechanism for decreased mortality in the previously referenced Australian study although fat emboli were not documented [17]. Notably, the rate of documented fat embolism was <.05% in our study population. This database study cannot evaluate whether TEE or pulmonary artery catheters were used to monitor for fat emboli although most likely they were not as this is not standard of care. In the absence of direct monitoring intraoperatively, the clinical diagnosis of fat embolism can be difficult as the presentation may be variable and include the triad of pulmonary distress, neurological symptoms, and petechial rash [27]. However, in the literature, about half of all patients who develop fat embolism develop respiratory failure, and we, therefore, thought that respiratory failure may also be used as an indirect proxy of fat embolism [16]. Given similar rates of respiratory failure between technology-assisted and conventional BSTKA in our study, we do not see definitive evidence that technology assistance prevented fat embolism.

A post-hoc analysis to identify differences in other complications found that technology-assisted BSTKA was associated with lower rates of readmission and a shorter length of stay. A possible explanation for this is that technology-assisted procedures may be done in hospitals with higher arthroplasty volume and more streamlined postoperative care pathways, as such institutions may be more willing to make the required investments to purchase TKA assistance technologies. However, these factors were not directly assessed in the present study. Additionally, robot-assisted TKA has been associated with decreased soft tissue damage, which may facilitate early postoperative recovery [7,13]. Prior data also showed that robot-assisted TKA was associated with improved early postoperative function and reduced time to discharge, the latter of which corresponds with our study results [14]. Additionally, rates of DVT, PE, and blood transfusions were lower among patients who underwent BSTKA with technology assistance. The etiology of these findings is not clear although prior work has identified these findings for unilateral TKA as well [15,23].

In conclusion, when compared with BSTKA using conventional instrumentation, technology-assisted BSTKA was not found to be associated with lower likelihood of 30-day mortality, fat embolism, or respiratory failure. On post-hoc analysis, technology usage was associated with decreased length of stay, lower readmission risk, and decreased rates of DVT, PE, and blood transfusion. Future randomized controlled trials could further our understanding of the differences in outcomes based on the use of technology in patients undergoing BSTKA.