Patient Reported Outcomes After Conversion vs. Primary Total Hip Arthroplasty: A Propensity Matched Analysis

Abstract

Background

Conversion total hip arthroplasty (cTHA) is increasingly utilized as a salvage procedure for complications associated with fracture fixation around the hip and acetabulum and for failed hip preservation surgery. While primary THA (pTHA) has a high success rate, little is known about outcomes following conversion THA. The purpose of this study is to evaluate patient reported outcomes (PROs) and complication rates following conversion THA compared to primary THA.

Methods

Patients that underwent cTHA or pTHA from 2015-2020 at a large tertiary referral academic center were retrospectively identified. THA patients were propensity matched in a 1:1 fashion by age, body mass index (BMI), and sex. Pain scores and PROMIS physical function (PF), pain interference (PI), and depression (DA) scores were compared at preoperative and final postoperative follow up timepoints using independent t-tests. Differences in complication and reoperation rates between cohorts were assessed using chi square analysis.

Results

A total of 118 THAs (59 cTHA, 59 pTHA) were included in this analysis with an average follow up of 21.3 months. cTHAs were most commonly performed following hip fracture fixation (50.8%). The conversion cohort had significantly longer lengths of stay (3.6 days vs 1.9 days, p<0.01) and greater use of revision-type implants (39.0% vs 0.0%, p<0.01) compared to pTHA. There was no significant difference in complication rates (cTHA = 15.3%, pTHA = 8.5%; p=0.26), with intraoperative fracture being the most common for both. Primary and conversion THA groups also experienced similar reoperation rates (cTHA = 5.1%, pTHA = 6.8%; p=0.70). No significant differences in PROs at final follow up were identified between groups.

Conclusion

Patients undergoing cTHA required increased utilization of revision hip implants and had longer lengths of stay, but had comparable complication and reoperation rates, and ultimately demonstrated similar improvements in PROMIS scores compared to a matched cohort of pTHA patients.

Level of Evidence: III

Introduction

Conversion total hip arthroplasty (cTHA) is increasingly utilized as a salvage procedure after failed hip preservation surgery or for complications associated with fracture fixation around the hip and acetabulum. Most commonly, cTHA addresses failures that occur after prior treatment of fractures around the hip including those treated with open reduction and internal fixation, cephalomedullary fixation, and hemiarthroplasty. Despite best efforts, complications such as post-traumatic arthritis, femoral head osteonecrosis, nonunion, infection, implant failure and loss of fixation do occur.^1-3 In the setting of prior surgery with progressive disabling post-traumatic arthritis, loss of fixation, implant cutout, or pain, cTHA can be a good option for pain relief and improved function. As the incidence of hip fractures increases, a greater number of patients are likely to undergo treatment for surgically related complications, leading to an increased utilization of cTHA as a salvage option.^4,5

Despite being part of the same diagnosis related group (DRG), multiple studies have demonstrated that cTHA is more complicated than primary THA (pTHA) in many ways. Altered anatomy and previously placed implants add to the complexity of the procedure. Studies that have evaluated cTHA have found the procedure to be associated with increased resource utilization, cost, and perioperative complications.^6-14 Unfortunately, few studies have assessed improvement in patient reported outcomes (PROs) after cTHA.^15-17

As healthcare delivery in the United States shifts towards alternative payment models and focuses more on value-based, outcomes-driven care, PROs will become increasingly important in measuring and maximizing value in healthcare.¹⁸ The National Institutes of Health (NIH) Patient-Reported Outcomes Measurement Information System (PROMIS) is one such PRO which has been utilized in arthroplasty to evaluate pTHA.^19,20 The PROMIS Computer Adaptive Test (CAT) is capable of increasing the efficiency and accuracy of short-form versions of the questionnaire utilizing Item Response Theory to ask a more appropriate subsequent set of questions. While multiple domains of questions exist, Physical Function (PF), Pain Interference (PI) and Depression Assessment (DA) are commonly utilized. The purpose of this study is to compare the 1) procedure characteristics, 2) complication rates, and 3) pre-operative and post-operative PROMIS CAT PF, PI, and DA scores of patients undergoing cTHA compared to pTHA. Our hypothesis is that patients undergoing cTHA will have similar PROMIS scores as compared to the pTHA cohort but will have increased perioperative complications, length of stay and discharge to skilled nursing facilities (SNF).

Methods

Patient Population

Following institutional review board approval, patients that underwent a primary or conversion THA at a large tertiary referral academic center from 2015 to 2020 were retrospectively identified through billing records using current procedure terminology (CPT) codes 27130 and 27132. Arthroplasty performed for primary treatment of fracture, tumor, revision of prior arthroplasty, or in the setting of active infection were excluded. Patients were excluded if they had underlying dementia and/or were unable to appropriately complete the questionnaires. Patients were also excluded if they had incomplete preoperative PRO survey data, or if they did not have a minimum of 6 months of postoperative follow up.

Data Collection

Demographic and baseline characteristics of the patient population were recorded including age, sex, body mass index (BMI), smoking status, Charlson Comorbidity Index (CCI), and specific comorbidities such as hypertension and diabetes mellitus. Details related to cTHA were also collected including time to conversion, indication for conversion, and need for revision-type or specialty implants. Perioperative information was recorded including surgeon type (arthroplasty or trauma fellowship-trained), approach utilized, cemented or non-cemented implants, length of hospital stay (LOS), and discharge destination.

PRO questionnaire data was collected from preoperative and all postoperative follow up visits including PROMIS PF, PI, and DA domains as well as visual analog scale (VAS) pain scores. Any complications or reoperations during the perioperative or postoperative period were recorded (up to one year). Complications included intraoperative fracture, periprosthetic fracture, surgical site infection, dislocation, symptomatic heterotopic ossification (HO) requiring surgical excision, surgical site hematoma requiring surgical evacuation, and trochanteric nonunion. Surgical site infections were considered superficial if treated with antibiotics alone, while deep infections referred to those requiring surgical intervention. Reoperations included component revision and/or open reduction and internal fixation for periprosthetic fracture, irrigation and debridement and component exchange, hematoma evacuation, and HO excision.

Statistical Analysis

Statistical analysis was performed using Stata/SE® (StataCorp LP, College Station, TX). Conversion and primary THA patients were matched using one-to-one propensity score matching based upon age, sex, and BMI.²¹ These parameters were chosen as they are similar to previous investigations in arthroplasty literature.^11,14,22,23 Patient baseline and perioperative characteristics were compared between conversion and primary cohorts using chi square analysis and independent t-tests for categorical and continuous variables, respectively. Complication and reoperation rates were assessed between groups using chi square analysis. PROs at final follow up and improvement in PROs from baseline were compared between conversion and primary cohorts using independent t-tests. Statistical significance was set at p<0.05.

Results

A total of 151 cTHA were identified. Patients were excluded if they did not have pre-op PROMIS scores (29 patients), if they had incomplete PROMIS data, could not complete PROMIS questionnaires, or had <6 months follow-up (58 patients), or underwent staged procedures for infection or revision arthroplasty (5 patients), leaving 59 cTHAs that met inclusion criteria. A total of 118 THAs (59 conversions, 59 primaries) in 115 patients were included in this analysis after successful 1:1 propensity-score matching. The average length of postoperative follow up was 21.3 ± 14.2 months. Table 1 details the characteristics of this patient population. There were no differences identified between primary and conversion cohorts with regard to age, sex, BMI, smoking status, or comorbidity burden (p>0.05 for all).

Table 1.

Baseline Characteristics

	Primary (N=59)	Conversion (N=59)	†p-value
Age (Mean ± SD, years)	61.4 ± 13.5	60.9 ± 15.5	0.875
Sex (n)			0.257
Female	66.1% (39)	55.9% (33)
Male	33.9% (20)	44.1% (26)
Body Mass Index (n)			1.000
Non-obese (< 30 kg/m2)	62.7% (37)	62.7% (37)
Obese (≥ 30 kg/m2)	37.3% (22)	37.3% (22)
Smoking Status (n)			0.068
Non-smoker	94.9% (56)	84.8% (50)
Smoker	5.1% (3)	15.2% (9)
CCI (Mean ± SD)	2.3 ± 1.7	2.4 ± 1.9	0.724
Hypertension (n)			0.460
No	49.2% (29)	42.4% (25)
Yes	50.8% (30)	57.6% (34)
Diabetes Mellitus(n)			0.729
No	93.2% (55)	91.5% (54)
Yes	6.8% (4)	8.5% (5)

SD = Standard deviation; CCI = Charlson Comorbidity Index †p-values calculated using independent t-tests and chi square analysis for continuous and categorical variables, respectively

Details regarding the cTHAs are reported in Table 2. On average, cTHA were performed 7.9 ± 11.4 years after the index procedure. Conversion THAs were most commonly performed following hip fracture fixation (50.8%). Revision-type implants were utilized in 39.0% of the cTHAs, most common being a revision-type (diaphyseal engaging) femoral component (28.8%), while no pTHAs utilized revision-type implants (p<0.01).

Table 2.

Conversion Details

	Conversion (N=59)
Time to Conversion (Mean ± SD, years)	7.9 ± 11.4
Reason for Conversion (n)
Femoral neck fracture fixation	20.3% (12)
IT/ST fracture fixation	30.5% (18)
Acetabular fracture fixation	18.6% (11)
Hip resurfacing/hemiarthroplasty	13.6% (8)
Other (SCFE, DDH, PAO)	17.0% (10)
Revision Implants (n)	39.0% (23)
Revision femoral component	28.8% (17)
Revision femoral component + dual mobility	5.1% (3)
Revision acetabular component + dual mobility	1.7% (1)
Revision femoral + acetabular components	1.7% (1)
Dual mobility	1.7% (1)

SD = Standard deviation; SCFE = Slipped capital femoral epiphysis; DDH = Developmental dysplasia of the hip; PAO = Periacetabular osteotomy

Table 3 details the operative characteristics for the conversion and primary cohorts. Primary THAs were most likely to be performed by an arthroplasty fellowship-trained surgeon (74.6%). Conversion THAs were most often performed by trauma fellowship-trained surgeons (47.5%), followed closely by arthroplasty fellowship-trained surgeons (39.0%). The approach utilized was significantly different between groups, with conversions more frequently performed through a posterior approach compared to primaries (61.0% vs 39.0%, p=0.02). Cement fixation was infrequently utilized in both cohorts (pTHA = 3.4%, cTHA = 8.5%; p=0.24). More cTHA patients were discharged to a skilled nursing facility compared to pTHA patients (20.7% vs 8.6%, p=0.07), however, this did not reach statistical significance. Patients undergoing cTHA had significantly longer postoperative lengths of stay (3.6 days vs 1.9 days, p<0.01) compared to pTHA. Of note, one patient in the conversion cohort had a 70-day hospitalization due to the development of acute respiratory failure during the postoperative period requiring a prolonged stay in the intensive care unit. This outlier was excluded from the length of stay analysis to avoid skewing the data.

Table 3.

Perioperative Characteristics*

	Primary (N=59)	Conversion (N=59)	†p-value
Surgeon (n)			<0.001
Arthroplasty fellowship-trained	74.6% (44)	39.0% (23)
Trauma fellowship-trained	23.7% (14)	47.5% (28)
Other	1.7% (1)	13.6% (8)
Approach (n)			0.017
Anterior/Lateral	61.0% (36)	39.0% (23)
Posterior	39.0% (23)	61.0% (36)
Cemented (n)			0.242
No	96.6% (57)	91.5% (54)
Yes	3.4% (2)	8.5% (5)
Length of Hospital Stay (Mean ± SD, days) ǂ	1.9 ± 1.2	3.6 ± 3.0	<0.001
Discharge Destination (n)			0.066
Home	91.4% (53)	79.3% (46)
Skilled Nursing Facility	8.6% (5)	20.7% (12)

SD = Standard deviation

*Boldface indicates statistical significance

†p-values calculated using independent t-tests and chi square analysis for continuous and categorical variables, respectively

ǂ One outlier with 70-day hospital stay in Conversion group excluded from this analysis

Complication rates and patient reported outcomes are described in Table 4. There was no significant difference in complication rates between groups (cTHA = 15.3%, pTHA = 8.5%, p=0.26). The most common complications among both groups were intraoperative fracture (cTHA = 4, pTHA = 2) and periprosthetic fracture (cTHA = 2, pTHA = 2). Additional complications observed in the conversion group included surgical site infection (n=1), dislocation (n=1) and trochanteric nonunion (n=1). Additional complications in the primary group included symptomatic HO (n=1) and hematoma requiring surgical evacuation with subsequent superficial surgical site infection (n=1). Primary and conversion THA groups also experienced similar reoperation rates (cTHA = 5.1%, pTHA = 6.8%; p=0.70), with the most common being component revision/fixation for periprosthetic fracture (cTHA = 2, pTHA = 2). Other reasons for reoperation included irrigation and debridement of a surgical site hematoma (n=1) and HO excision (n=1) in the primary group, and irrigation and debridement with component exchange for infection (n=1) in the conversion group. PROMIS physical function, PROMIS pain interference, and VAS pain scores were similar between conversion and primary cohorts at final follow up (p>0.05 for all). Conversion THA trended toward higher PROMIS depression scores (50.7 vs 46.9, p=0.052) compared to pTHA at final follow up, though this did not reach statistical significance. However, improvement in PROMIS depression from preoperative values was similar between groups (cTHA Δ = -3.2, pTHA Δ = -4.3; p=0.51).

Table 4.

Patient Outcomes

	Primary (N=59)	Conversion (N=59)	†p-value
Final Follow Up (Mean ± SD, months)	22.3 ± 14.4	20.1 ± 14.0	0.397
Surgical Complications (n)	8.5% (5)	15.3% (9)	0.255
Reoperations (n)	6.8% (4)	5.1% (3)	0.697
PROMIS Physical Function (Mean ± SD)
Preoperative	36.5 ± 5.5	34.6 ± 6.3	0.087
Final Δ	6.2 ± 9.0	6.6 ± 8.6	0.788
Final follow up	42.7 ± 9.3	41.3 ± 9.8	0.417
PROMIS Pain Interference (Mean ± SD)
Preoperative	64.0 ± 6.0	64.6 ± 6.5	0.590
Final Δ	-8.2 ± 10.8	-7.6 ± 10.4	0.784
Final follow up	55.7 ± 9.6	57.3 ± 10.6	0.388
PROMIS Depression (Mean ± SD)
Preoperative	51.3 ± 10.8	53.8 ± 10.0	0.194
Final Δ	-4.3 ± 9.0	-3.2 ± 8.4	0.505
Final follow up	46.9 ± 9.0	50.7 ± 11.4	0.052
VAS Pain Score (Mean ± SD)
Preoperative	5.8 ± 3.1	5.6 ± 3.0	0.756
Final Δ	-3.0 ± 3.7	-3.0 ± 3.6	1.000
Final follow up	2.8 ± 3.1	2.6 ± 3.1	0.760

SD = Standard deviation; PROMIS = Patient reported outcomes measurement information system; VAS = Visual analog scale †p-values calculated using independent t-tests (continuous variables) and chi square analysis (categorical variables) Δ = Final follow up score - Preoperative score

Discussion

Conversion THA is being increasingly utilized as a salvage surgery after failed fracture fixation, progressive arthritis after previous hip surgery, or prior hip salvage procedures. Patient reported outcomes after pTHA have been extensively evaluated and the success of THA at relieving pain and improving quality of life is evident.^19,24-26 Despite being part of the same diagnosis related group, cTHA is a much more complex procedure due to multiple factors including altered anatomy and pre-existing implants requiring larger and more extensive exposures. Studies have demonstrated increased resource utilization for patients undergoing cTHA secondary to increased length of stay and readmission rates, need for blood transfusions, disposition to subsequent inpatient care, length of surgical procedure and increased cost of implants. Perioperative complications including prosthetic joint infection, hip dislocation, and need for revision surgery have also been found to be higher in this population.^6-14

In a large propensity matched database study evaluating National Surgical Quality Improvement Project (NSQIP) data files, cTHA patients demonstrated greater rates of complications (7.5% vs. 4.5%, odds ratio 1.68, CI 1.39-2.02), longer LOS, and statistically significant differences in non-home bound discharge.¹³ A subsequent review of the American College of Surgeons NSQIP database found that patients undergoing cTHA are more similar to patients undergoing revision arthroplasty (rTHA) with similar demographic, clinical, and perioperative characteristics, underscoring the complex nature of the patient population.¹² Ryan et al. reviewed 163 cTHA performed at a tertiary medical center. As compared to pTHA, cTHA was associated with about 19% greater cost, increased operative times, LOS, blood loss, and perioperative complications.⁷ Similarly, Schwarzkopf et al. compared 119 cTHA to 251 pTHA and found that patients undergoing cTHA had increased LOS (3.8 days vs 2.8 days), longer surgical time, and greater likelihood of utilizing revision-type implants.⁶ This is in line with the current study, as our cTHA cohort similarly demonstrated increased LOS and higher likelihood of utilizing revision-type implants. Although patients undergoing cTHA trended toward a higher rate of discharge to a skilled nursing facility, statistical significance was not reached. In spite of the increased risks, patients continue to elect to undergo cTHA to ease their pain and improve their function. However, little is known about patient reported outcomes after cTHA.

To our knowledge, this is the first study evaluating PROMIS scores in a propensity matched cohort comparing cTHA to pTHA. Our results demonstrate that in a one-to-one, propensity matched study of patients undergoing pTHA versus cTHA, there are no significant differences in PROs, including VAS, PROMIS CAT PF, PI, and DA, at preoperative and final postoperative time-points. At final follow-up, cTHA patients were functioning as well as pTHA patients and exhibiting similar improvements in PROs. Previous studies have demonstrated improved WOMAC and Harris Hip Scores after cTHA from failed bipolar hemiarthroplasty as well as conversion from failed CMN, despite relatively high short-term complication rates.^9,17,27 Additionally, Tadevich et al. evaluated a matched cohort of cTHA and pTHA patients and found that UCLA activity level scores between the two groups were similar at one year follow-up.¹⁵ These reports support the findings of the present study, suggesting that cTHA leads to acceptable patient outcomes.

Increased scrutiny is being placed on patient outcomes, complications, and cost of episode of care.⁷ In an effort to elucidate the perceived value in conversion arthroplasty, this study demonstrates that despite the increased short-term risk profile and increased resource utilization associated with this procedure, patient outcomes may ultimately be similar to that of primary arthroplasty, which has a previously established record of high success. In the present study, cTHA patients did demonstrate increased resource utilization, with increased length of stay and greater use of revision-type implants. This is also in-line with previous studies. This may be useful for pre-operative patient counseling and provides justification for the increased resource utilization, risk profile, and appropriate reimbursement.

Limitations of this study include the retrospective nature of chart review as well as exclusion of patients who may have been lost to follow up or were without sufficient PRO data. Selection bias may exist against those who did not complete PROMIS questionnaires either due to disinterest, time constraints, or due to tablet availability in the setting of COVID-19. In addition, this study is somewhat limited by a small sample size. However, this procedure is relatively uncommon and there is very limited information on the subject in current literature. This study was also performed at a large, tertiary-care academic medical center in New York, so these results may not be generalizable across all patient populations. Despite these shortcomings, the current study was able to evaluate long term PROs for a large group of patients undergoing a relatively rare procedure in a 1:1 propensity matched fashion.

Conclusion

Patients undergoing cTHA required increased utilization of revision hip implants, had longer lengths of stay and comparable complication rates, but ultimately demonstrated similar improvements in PROMIS scores compared to a matched cohort of pTHA. Further research should focus on multicenter studies examining this relatively rare but quickly growing complex patient population.