Accuracy of digital templating of uncemented total hip arthroplasty at a certified arthroplasty center: a retrospective comparative study

Preoperative planning is an essential and integral part of the total hip arthroplasty (THA) procedure [1]. Digital templating of THA is well established, facilitates the determination of the correct implant size and helps restore the patient-specific physiological biomechanical conditions such as leg length, offset, center of rotation as well as lateralization [1‐3]. Moreover, preoperative surgical planning improves postoperative range of motion and stability, shortens the operative time and reduces wear caused by malpositioning of the implant components [1‐5].

A variety of factors that might negatively influence the accuracy of digital templating are mentioned in the current literature [4, 6‐8]. Difficulties in determining the correct magnification factor for calibrating digital X-rays have been described, especially for obese patients [4, 6‐8]. In addition, better results affecting the accuracy of the predicted component size using digital preoperative planning software have been shown in combination with the planner’s experience [1‐4, 9‐11]. According to the literature, even implant design seems to have an effect on planning accuracy [11].

There is evidence to support the issues and the extent to which digital planning matches the actual intraoperatively selected THA size [1, 7, 12, 13], but to the best of our knowledge our study is the first to investigate the accuracy of preoperative THA planning at an EndoCert EPZmax center  [14, 15]. At such a certified arthroplasty department, endoprosthetic interventions and preoperative templating are performed according to the guidelines and specifications of EndoCert [16, 17]. EndoCert is an initiative of the German Society for Orthopaedics and Orthopaedic Surgery (DGOOC) and is the world's first system for the certification of medical facilities in the field of joint replacement [14]. The EndoCert certification system is intended to ensure high-quality patient care and high patient safety in endoprosthetic procedures [16]. The process quality, interdisciplinary and constantly evolving treatment paths as well as a high level of training for all professional groups involved in the treatment are of central importance in the certification process [14, 17]. In addition, certified surgeons have to perform a predetermined number of joint replacements per year [17]. Institutions that meet the requirements can be certified as an endoprosthetic center or endoprosthetic center for maximum care [15, 17].

We, therefore, aimed to (a) determine in how many cases preoperative templating accurately matched the implant size chosen intraoperatively and (b) analyse factors that might have influenced the accuracy of preoperative digital templating in patients who underwent THA in an EndoCert max center (experience, body mass index (BMI), sex, implant design and preoperative diagnosis).

The study protocol was approved by the local ethics committee of the Medical University (No. 1150/2019) and performed in line with the principles of the Declaration of Helsinki.

We retrospectively investigated and included all patients who underwent primary THA at our department between January 2017 and August 2019. A total of 786 patients, giving a total of 843 implanted THAs, were investigated. Inclusion criteria were defined solely as the implantation of a primary uncemented total hip implant in the above-mentioned and defined period of time. Exclusion criteria were prior surgical interventions in the hip joint, previous fractures with joint involvement, cemented THA, revision surgery and intraoperative complications such as periprosthetic fractures and malalignment of the femoral stem in postoperative anterior–posterior X-rays (defined as 5° < varus or valgus). A total of 578 patients met our inclusion criteria, resulting in 620 uncemented THAs.

Socio-demographic data and patient-related factors were collected to analyse a possible influence on the preoperative planning and the intraoperatively selected component size. These factors included body mass index (BMI), preoperative diagnosis, side of the operation, cut-to-suture time, positioning of the THA (cup: inclination and anteversion; stem: 5° < varus or valgus), surgical approach as well as the planner’s experience and intraoperative and postoperative complications. We included the uncemented implants most commonly used at our department:

Trident PSL cup in combination with the Accolade II stem (both Stryker Orthopaedics, Mahwah, NJ, USA) and the Pinnacle cup combined with the Corail stem (both DePuy Synthes, Warsaw, IN, USA). The Trident PSL cup is 1.8 mm wider than the stated size. This is meant to achieve an interference fit at the periphery of the implant [18, 19]. The Accolade II has a morphologic wedge and a size-specific medial curvature [20]. The Trident-Accolade II THA was found to be a common implant combination in Germany. According to the annual report of the German Joint Replacement Registry, 1875 Trident PSL cups and 3086 Accolade II stems were implanted in Germany in 2018 [21].

The Pinnacle cup is a spherical cup with a single radius [1, 22]. The Corail stem is designed to sit in the cancellous bone. It is hydroxyapatite-coated and has trapezoidal-like proximal cross-sections to provide rotational stability [1, 23]. According to the German Joint Replacement Registry in 2018 the Pinnacle cup was the most commonly used hip cup in Germany with a total of 17,878 implantations. The Corail stem was implanted 20046 times in Germany during the same period [21].

The decision for one or the other type of cup or stem was made by the surgeon himself, who preferred to use a particular cup or stem type, and did not depend on the patient. Thus, there is no selection bias for the implanted cup or stem component.

A total of 620 (left: 277; right: 343) cementless THAs in 578 patients (female: 298; male: 280) were investigated. Mean age at surgery was 66.1 (range: 16.7–90.3) years. In 578 cases the THA was performed on one side and in 42 cases on both sides simultaneously. Mean body mass index was 26.7 (range: 16.9–59.2) kg/m². According to the definition of the World Health Organization (WHO), 207 participants were under- or normal weight (BMI up to 24.9), 255 were overweight (BMI between 25 and 30) and 157 were obese (BMI from 30.1 upwards). The most common indication for THA was primary osteoarthritis in 533 out of 620 cases. In 56 cases necrosis of the femoral head, in 29 hip dysplasia and in 12 cases protrusion osteoarthritis of the hip was the indication for THA surgery. Mean cut-to-suture time was 67.7 (range 37–181) minutes. In all patients, a direct anterior approach was performed for the THA procedure [27, 28]. A Trident PSL cup and an Accolade II stem were implanted in 88% (n = 544), and a Pinnacle cup and Corail stem were used in 12% (n = 76) of the procedures. Details are given in Table 1 and Fig. 2.

The most important findings of the study were that templating was found to be accurate in 90% of the stems and in 85% of the cups. Regarding the investigated factors that potentially influenced templating accuracy, it was found that the type of cup implant had a significant influence (p = 0.016). Moreover, greater accuracy of stem templating was achieved in female patients (p = 0.004). For the rest of the investigated factors no such effect was determined.

When comparing our findings with those of previous research it appears that Holzer et al. analysed 632 preoperatively planned uncemented THAs, of which the cup and the stem were preoperatively determined to be within one size in 78% and in 87%, respectively [1]. In the study by Whiddon et al., planning accuracy was shown to be 78% for the acetabular and 90% for the femoral component (both within ± one implant size) [26]. Sershon et al. showed an accuracy of digital templating within two sizes of the final acetabular and femoral implants in 99.1% and in 97.1% of cases, respectively [25]. However, it may be questioned whether ± two implant sizes should still be taken as accurate. Wiese et al. showed a planning accuracy of 71% for the acetabular and 79% for the femoral component [29]. Roughly, the results of the current investigation match those of the studies mentioned above (Table 3).

Underlying studies, like the present work, are preoperative predictions based on two-dimensional electronic X-ray images. However, today there are numerous other options for preoperative planning in THA. Studies using CT-based three-dimensional planning show a significantly higher planning accuracy than do preoperative predictions based on two-dimensional electronic X-ray images. Sariali et al. and Osmani et al. demonstrated an accuracy of more than 95% when planning was based on three-dimensional CT images [30, 31]. The three-dimensional imaging based on CT images provides surgeons with more bone structures to assist in planning and increase accuracy [32]. Schiffner et al. were able to demonstrate the superiority of CT-based three-dimensional planning over 2D planning but emphasized that greater planning accuracy did not necessarily mean better clinical outcome [33]. Additionally, nowadays it is possible to plan automatically using CT imagination. Kagiyama et al. developed a system that is able to determine the most suitable implant by collecting data from an experienced surgeon [34]. Nevertheless, CT-based planning is still controversial. While Rübberdt et al. pointed out the greater radiation exposure in the area of the gonads, Henckel et al. argued that radiation exposure is negligible if CT scans are performed with special low-dose recordings [35, 36]. Furthermore, technical opportunities for preoperative three-dimensional imaging based on CT scans are not given in every hospital and sufficient precision can also be achieved with two-dimensional planning methods. Although the methods of CT-based three-dimensional and computer-assisted planning are manifold, the individual orthopaedist achieves the safest results with the method in which he was trained and has experience [37]. Thus, two-dimensional X-ray-based planning is still the most widely used in clinical practice. Though there is a manifold number of two-dimensional planning software, technically the majority is based on the same process. Preoperative radiographs are taken to a standard antero-posterior view of the pelvis and a calibration object of individual size is placed between the legs of the patients. The digital templating software automatically calibrates the image and template-overlays according to the known size of the marker. Most clinics have switched from planning with analogue solid templates to digital planning, and therefore the current literature clearly shows that digital two-dimensional planning is well established [4].

The presented study analysed the Trident PSL and the Pinnacle acetabular component systems and demonstrated that the Pinnacle cup achieved greater templating accuracy than did the Trident PSL cup (p = 0.016). When comparing the different femoral components, no significant difference was seen between the Accolade II and the Corail stem (p = 0.052). One possible explanation might be found in the “peripheral self-locking” system of the Trident PSL cup. The outside diameter of the cup is 1.8 mm wider than the actual reamed size. The surgeon has to take this into consideration when planning the cup size as well as intraoperatively when reaming the cup [18]. Furthermore, the differences in bone stock from patient to patient might have an influence, as surgical protocols mention that reaming should be performed with special attention to bone quality [18].

The presented study showed a tendency to predicting the acetabular component too large (30% planned too large, 19% too small), which coincides with the results reported by Wiese et al., where a tendency to estimate both components too large was proven [29].

Several studies have reported the experience of the orthopaedist who performed the planning to be a significant factor influencing the accuracy of preoperative planning. Carter et al. demonstrated a significant impact on the planning of both components, Holzer et al. only for the femoral component [1, 38]. Efe et al. and Strøm et al. showed no significant influence of the experience of the planning orthopaedic surgeon [39, 40], which is in line with the findings of the present study (p = 0.353 for the acetabular component; p = 0.169 for the femoral component). This might be due to the fact that the study population could possibly be too small since the planning surgeon could be identified in only 42% of the performed operations. However, the results show a trend, according to which specialized arthroplasty surgeons are more precise in their preoperative planning. With a larger study population, this might be a statistically significant influence.

No statistically significant impact of BMI was found on the accuracy of preoperative planning (p = 0.422 for the acetabular component; p = 0.216 for the femoral component). Similarly, Sershon et al. showed no significant effect of BMI on templating accuracy [25]. Whiddon et al. divided the patients into obese and non-obese patients (BMI ≥ / ≤ 30) and did not assess any differences in the accuracy of planning [25, 26]. In contrast, Holzer et al. showed a difference between normal- and overweight patients (BMI 18.5–24.9 for normal-weight patients or 25–29.9 for overweight patients) regarding accuracy [1].

In the presented study, significantly greater precision in preoperative planning was observed in female patients than in male patients, but only for the femoral component (p = 0.004) and not for the acetabular component (p = 0.602). Holzer et al. found no significant difference in the accuracy of preoperative planning between sexes [1].

Templating with accurate and reliable calibration markers is of the utmost importance in THA, as it has been shown that calibration errors using external calibration markers significantly influence component sizes [41]. In a recent study, Warschawski et al. compared the accuracy in the preoperative component selection of the double marker (King Mark) method, which may be more accurate than a single marker method, with the conventional metal ball method in the general population and in obese patients. However, the study found no difference between the King Mark method and the conventional metal ball method in the ability to accurately predict component sizes. In the subgroup of obese patients, the King Mark technique offered no advantage for accurately predicting component sizes [42].

Recently, Kase et al. presented a classification system to aid surgeons during their preoperative analysis, outlining the importance of considering femoral head translation during preoperative templating [43]. The authors describe a classification system to distinguish five types of architectural hip deformities, based on femoral head translation patterns, and advise surgeons to adapt their templating strategy accordingly. In a consecutive study, the authors evaluated whether mismatch between planned and real implant sizes compromises THA outcomes [44]. According to the given results, implanting a component of a different size than planned seemed not to compromise THA outcomes in terms of the Forgotten Joint Score (FJS) and Oxford Hip Score (OHS). Therefore, the authors advise that surgeons should respect their intraoperative findings when it comes to the ultimate implant size selection.

The following limitations are acknowledged. There was a lack of variability in the study population and therefore the subgroups were too small. Thus, the accuracy in planning for the different underlying diagnoses is not sufficiently comparable (osteoarthritis, dysplasia, femoral head necrosis etc.). Another limiting factor is the small amount of information available about the planning surgeon due to the retrospective character of the study. Of 620 included hip prostheses only 260 planning surgeons could be assessed. The study demonstrates a trend, according to which planning accuracy tends to increase with the experience of the planning surgeon, but no statistical significance was found (p = 0.353 for the cup; p = 0.169 for the stem). With a larger study population, there might be a relevant possibility to detect a significant difference.

From our findings, we conclude that preoperative 2D templating is accurate in 90% of the stems and 85% of the cups. In female patients, greater accuracy may be achieved. In addition to gender, the type of implant used may influence planning accuracy as well. Surgeon experience, BMI and preoperative diagnosis did not influence templating accuracy.