No Evidence of Superiority in Reducing Outliers of Component Alignment for Patient‐specific Instrumentation for Total Knee Arthroplasty: a Systematic Review

Abstract

Patient‐specific instrumentation (PSI) technology has been developed to improve alignment when implanting total knee arthroplasty (TKA) and is a new focus in the orthopaedic community. Current controversial data concerning PSI are discussed. A systematic review to compare PSI with conventional instrumentation and assess the radiographic outcomes was performed. Electronic databases (including PubMed, Medline, Embase, the Cochrane Library and the Science Citation Index database) and conference proceedings from 1950 to 2014 in the English language were searched. Data, including relevant patient characteristics, sample size, radiographic method, PSI system manufacturer and outliers of implant positioning and alignment on radiography were independently extracted from all eligible studies by two of the authors. A total of 2739 TKAs were included (1410 performed with PSI and 1329 with conventional instrumentation). There were more TCA outliers (malalignment >3°) and tibial slope outliers (malalignment >3°) in the PSI group than in the conventional group. The other radiographic outcomes assessed, including coronal, sagittal or rotational alignment outliers did not differ between the two groups. With regard to radiographic outcomes, our findings indicate that PSI technology is not superior in reducing outliers of component alignment.

Introduction

Total knee arthroplasty (TKA) is a surgical intervention that provides pain relief, enhances mobility, and improves quality of life of patients with end‐stage knee arthritis. With increases in population size and longevity, the incidence of TKA is predicted to increase in the future1. With TKA, component alignment and position are important predictors of clinical outcome and longevity2, 3. The accuracy of conventional instrumentation has been questioned by many authors; the rate of implant malposition can be as high as 20%–40%4. Since the introduction of these systems in the 1970s, they have been modified repeatedly but without radical changes or substitutes. Technological advances aimed at improving limb alignment and component position include computer‐assisted surgery and patient‐specific instrumentation (PSI)5, 6, 7. Computer‐assisted surgery is rarely performed because of its drawbacks, which include difficulty in identifying landmarks accurately intraoperatively, increased set‐up and operative time, increased perioperative cost and a substantial learning curve5, 6.

Patient‐specific instrumentation is a new concept in which computer‐assisted preoperative planning and rapid prototyping technology provides patient‐specific instruments that can replace conventional instrumentation systems. Preoperative CT or MRI scans are acquired and imported to a special software system. Planning of sizing, alignment, and bone cutting and verification of optimal implantation and positioning is performed on a computer and two virtual templates designed and transformed into physical guides preoperatively. Information built into the guides can be transferred to the patient's knee: surgeons can use these guides or conventional cutting blocks. TKA can then be performed without using intra‐or extramedullary guides7.

Patient‐specific instrumentation can theoretically improve the accuracy of component position and alignment; however, the clinical results are currently controversial8, 9, 10, 11. We therefore performed a systematic review to compare the radiographic outcomes between PSI and conventional instrumentation.

Materials and Methods

A systematic search of the PubMed, Medline, Embase, Cochrane Library and Science Citation Index (SCI) databases was performed. The search terms were as follows: (“patient‐specific” or “patient specific” or custom*) AND (instrument* or template* or cutting or guide* or block*) AND (“total knee replacement” or “Total knee arthroplasty” or “Total knee arthroplasties”). The search was performed on 1 June 2014. Inclusion criteria were as follows: i) clinical research studies comparing PSI technology to conventional technology in TKA; ii) analysis of component alignment; iii) primary TKA; and iv) patients aged over 18 years. The titles of all studies were identified by the initial search and assessed by two authors. Review articles, technique descriptions, and editorials were excluded, as were clearly non‐relevant studies that failed to meet the inclusion criteria. Abstracts and full reports of the remaining studies were then independently assessed by two authors according to agreed inclusion and exclusion criteria. Disagreements were resolved by discussion until consensus was reached. When the two reviewers were unable to reach consensus, a third reviewer was asked for a final opinion, resulting in a group consensus. A flow diagram shows the search process (Fig. 1).

Figure 1.

Study flow diagram. A summary of the search process and study identification. 24 studies were included in the final analysis.

Eventually, 24 studies were eligible for this study (Table 1). Two authors independently extracted relevant data from all eligible studies, including patient characteristics, sample size, PSI system manufacturer, radiographic method and outcomes in terms of radiographically‐identified outliers. Radiographic measurements for coronal alignment were as follows: hip‐knee‐ankle angle (HKA), femorotibial angle (FTA), femoral coronal angle (FCA, defined as the medial angle between the femoral component and the mechanical axis of the femur) and tibia coronal angle (TCA, measured as the angle between the tibial component and the mechanical axis of the tibia). Radiographic measurements for sagittal alignment were the lateral femoral component angle (LFC angle, defined as the anterior angle between the femoral component and the anterior cortex of the femur), LTC angle (lateral tibia component angle, defined as the anterior angle between the tibial component and the posterior cortex of the tibia), femoral flexion angle and tibial slope angle. Radiographic measurements for rotational alignment were femoral rotational angle and tibial rotational angle. Two types of outliers with two different cutoff values (malalignment >3° or >2°) were recorded. Two authors independently identified the radiographic measurements of each eligible study and recorded the outliers.

Table 1.

Details of eligible studies

Year	Journal	First author	Total	PSI	Con	PSI system	Study design	Measurement method
2014	Int Orthop (SICOT)	Kotela9	112	52	60	Signature; Biomet	RCT	X‐ray
2013	Clin Orthop Relat Res	Roh10	100	50	50	Signature; Biomet	RCT	X‐ray; CT
2014	J Bone Joint Surg Am	Woolson11	48	22	26	TruMatch; DePuy	RCT	CT
2014	Knee Surg Sports Traumatol Arthrosc	Silva12	45	23	22	Signature; Biomet	RCT	CT
2014	Knee	Chotanaphuti13	80	40	40	TruMatch; DePuy	RCT	X‐ray; CT
2013	Knee Surg Sports Traumatol Arthrosc	Parratte14	40	20	20	Unknown	RCT	X‐ray; CT
2013	Knee Surg Sports Traumatol Arthrosc	Boonen15	180	90	90	Unknown	RCT	X‐ray
2013	J Arthroplasty	Hamilton16	52	26	26	TruMatch; DePuy	RCT	X‐ray
2013	Knee Surg Sports Traumatol Arthrosc	Ng17	78	51	27	PSI system; Zimmer	RCT	CT
2013	Bone Joint J	Chareancholvanich18	80	40	40	PSI system; Zimmer	RCT	X‐ray
2012	Orthopedics	Dossett19	82	41	41	Unknown	RCT	CT
2014	Clin Orthop Relat Res	Victor20	125	61	64	Signature Biomet; TruMatch DePuy; Visionaire Smith & Nephew; PSI Zimmer	RCT	X‐ray; CT
2012	J Arthroplasty	Noble21	29	15	14	Visionaire; Smith & Nephew	Quasi‐RCT	X‐ray
2013	J Arthroplasty	Vundelinckx22	62	31	31	Visionaire; Smith & Nephew	Quasi‐RCT	X‐ray
2012	Clin Orthop Relat Res	Ng23	160	105	55	Signature; Biomet	Retrospective cohort	X‐ray
2014	J Arthroplasty	Marimuthu24	300	115	185	Visionaire; Smith & Nephew	Retrospective cohort	CT
2014	Orthop Traumatol Surg Res	Moubarak25	68	57	11	Visionaire; Smith & Nephew	Prospective cohort	X‐ray
2014	J Arthroplasty	Stronach26	120	58	62	Signature; Biomet	Retrospective cohort	X‐ray
2014	Int Orthop	Daniilidis27	340	170	170	Visionaire; Smith & Nephew	Retrospective cohort	X‐ray
2014	Knee	Heyse28	94	46	48	Visionaire; Smith & Nephew	Retrospective cohort	MRI
2012	J Bone Joint Surg Br	Barrack29	200	100	100	Unknown	Prospective cohort	CT
2012	Acta Orthop	Boonen30	80	40	40	Signature; Biomet	Retrospective Cohort	X‐ray
2012	Clin Orthop Relat Res	Nunley31	150	100	50	OtisMed; Styker	Retrospective Cohort	X‐ray; CT
2012	Clin Orthop Relat Res	Nunley32	114	57	57	Signature; Biomet	Retrospective Cohort	CT

Con, conventional instrumentation; PSI, patient‐specific instrumentation; Quasi‐RCT, quasi‐random controlled trial; RCT, randomized controlled trial.

A meta‐analysis was performed with RevMan software (“Review Manager,” version 5.1, Nordic Cochrane Centre, Copenhagen, Denmark). A random effects meta‐analysis was generated using these dichotomous data to compare PSI with conventional instrumentation with regard to outliers of coronal alignment, sagittal alignment and rotational alignment. A P value of 0.05 was considered statistically significant.

Results

The final 24 studies included 2739 total knee arthroplasties (1410 performed with PSI and 1329 with conventional instrumentation). The details of these 24 studies are shown in Table 1. Twelve studies were randomized trials9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, two were quasi‐randomized trials21, 22 and ten were non‐randomized comparative trials23, 24, 25, 26, 27, 28, 29, 30, 31, 32. Five different PSI systems were utilized in the studies: Biomet Signature (Warsaw, IN, USA), Zimmer Patient Specific Instruments (Minneapolis, MN, USA), Smith & Nephew Visionaire (Memphis, TN, USA), DePuy TruMatch (Raynham, MA, USA) and Stryker OtisMed (Kalamazoo, MI, USA).

Coronal Alignment

Sixteen studies reported HKA outliers of malalignment >3°, the odds ratio for HKA outliers of malalignment >3° in the PSI group compared with the conventional group was 1.04 and the odds ratio for HKA outliers of malalignment >2° in the PSI group was 0.70. Four studies reported FTA outliers of malalignment >3°; the odds ratio for HKA outliers of malalignment >3° in the PSI group compared with the conventional group was 1.08. Both the incidence of FCA outliers of malalignment >3° and malalignment >2° were not statistically significant; the odds ratios were 0.72 (P = 0.16) and 0.69 (P = 0.23) respectively. The incidence of TCA outliers of malalignment >2° was also not significantly different (odds ratio: 1.01, P = 0.97); however, the odds ratio for outliers of malalignment >3° was 2.22 (P = 0.001); there were thus significantly fewer TCA outliers of malalignment >3° in the conventional group (Figs 2, 3, 4, 5).

Figure 2.

Forest plot for HKA outliers (malalignment >3°).

Figure 3.

Forest plot for FTA outliers (malalignment >3°).

Figure 4.

Forest plot for FCA outliers (malalignment >3°).

Figure 5.

Forest plot for TCA outliers (malalignment >3°).

Sagittal Alignment

All femoral sagittal variables, including LFC outliers (malalignment >3°) and femoral flexion angle outliers (malalignment >3° or 2°) did not differ significantly between the two groups. LTC outliers of malalignment >3° were equivalent between the two groups (odds ratio, 1.46; P = 0.42). Thus, decreased tibial slope angle (outliers of malalignment >3°) occurred significantly less often with conventional instrumentation (odds ratio, 1.97; P = 0.001); however, outliers of malalignment >2° was not significantly different between the two groups (Figs 6, 7, 8, 9).

Figure 6.

Forest plot for LFC outliers (malalignment >3°).

Figure 7.

Forest plot for femoral flexion outliers (malalignment >3°).

Figure 8.

Forest plot for LTC outliers (malalignment >3°).

Figure 9.

Forest plot for tibial slope outliers (malalignment >3°).

Rotational Alignment

The incidence of femoral rotation outliers of malalignment >3° did not differ significantly between the two groups (odds ratio, 0.75; P = 0.38). The heterogeneity of the other two comparisons, namely femoral rotation outliers of malalignment >2° and tibial rotation outliers of malalignment >3° was unacceptable, so differences could not be accurately assessed (Fig. 10).

Figure 10.

Forest plot for femoral rotation outliers (malalignment >3°).

Discussion

This analysis showed that more TCA outliers of malalignment >3° and tibial slope outliers of malalignment >3°) in the PSI than in the conventional group. The other outliers of radiographic outcomes (HKA, FTA, FCA, LFC, LTC, femoral flexion angle and rotational angle) showed no differences between the two groups. These findings results are similar to those of previous studies33, 34, 35. Thus, there is still no evidence for routine use of PSI in TKA36.

These disappointing findings can be attributed to the following factors. First, PSI is a new technique for most surgeons, whereas conventional technology has been developed for several decades and surgeons are therefore more familiar with it. The different degree of familiarity may have influenced the outcomes. Second, the level of experience of orthopaedic surgeons influences the accuracy of operation; accordingly, the differences in radiographic outcome between the PSI group and conventional technology group were not statistically significant37. Third, the clinical use of PSI is at an exploratory stage and there are many problems in practice26, 38, 39. For example, some studies reported making intraoperative modifications. Stronach et al. made 161 changes in 66 cases in regard to thickness, angles and sizes (an average of 2.4 changes per knee)26. Issa et al. reported 29 intraoperative changes in 89 primary TKAs using PSI technology38. The above factors may account for the discrepancy between the theoretical advantages and the reality.

We also noted that some studies reported using PSI in some technically demanding cases in which the normal anatomical landmarks were hard to find and it was therefore difficult to achieve correct alignment. Thienpont et al. used TKA with the assistance of PSI to treat patients with knee arthritis and extra‐articular deformities from malunion or with retained femoral hardware40. Postoperatively, limb alignment was restored with a mean hip–knee–ankle angle of 179.3° ± 1.3° (P < 0.05). Maximum outliers were 177° to 181°. They concluded that mechanical alignment can easily be obtained with PSI technology by intra‐articular correction of deformities under 20°. Kerens et al. used PSI for revision of medial uni‐condylar knee arthroplasty to TKA41. Seven of 10 femoral prostheses were within the desired AP and sagittal angles of ±3°, all tibial components were within the desired AP angle of ±3° and 7 of 10 were within the desired sagittal angle. The hip–knee–ankle angle was within 0° ± 3° in 8 of 10 cases. The PSI technique provides a new and less demanding approach to preoperative planning and execution of the plan during surgery in some special situations.

This study has several limitations. One important limitation of this study is the intrinsic drawbacks of meta‐analysis. Although we assessed heterogeneity and bias in the randomized controlled trials, quasi‐ randomized controlled trials and retrospective cohort studies, the final conclusions should be interpreted with caution. Further studies with high evidence levels are necessary. An additional limitation is that the PSI systems of different manufacturers are based on MRI or CT, which provides a potential bias. There were too few data to perform subgroup analysis. We also did not analyze detailed data for each angle because postoperative radiographic methods differed between the studies and the measurements were not uniform. The variable of outliers is consistent and reliable; however, much potential information may be missing.

In conclusion, this study does not confirm the claimed benefit that PSI technology provides better accuracy of alignment than conventional technology.