Abstract
Purpose
The direct anterior approach (DAA) for total hip arthroplasty (THA) is claimed to be as effective but less invasive than the conventional posterior approach (PA). However, the higher risk of femoral fracture and soft tissue damage cannot be underestimated. The present authors believe that the difficult femoral exposure and the surgeon's knowledge of possible complications related to femoral preparation may result in a higher rate of undersized stems when compared to PA, even when a short femoral component is employed to minimise these risks.
Methods
A retrospective study was performed to compare the effective adherence of surgical sizing to preoperative planning in a series of 112 short stem THAs, of which 59 were implanted through PA and 53 through DAA without intraoperative imaging, by a single surgeon skilled in both techniques.
Results
the frequency of stems smaller than expected was 3-times higher with DAA than with PA (54.72% vs 16.95%, p<0.001), while the frequency of stems at least 2 sizes smaller than expected was more than 6-times higher with DAA than with PA (24.53% vs 3.39%, p = 0.001). Conversely, no differences of size discrepancy distribution were found for cups and prosthetic heads. Postoperative x-rays confirmed an inferior mean canal fill in the DAA group, but excluded a different incidence of frontal malalignment between the groups.
Conclusions
The technical difficulty of femoral preparation and the surgeon's knowledge of possible related complications might lead to implant undersized stems more frequently through DAA than through PA, especially if intraoperative imaging controls are not used.
Introduction
The indications for different surgical approaches for total hip arthroplasty (THA) remain controversial, and mostly depend on a surgeon's experience (1). Some investigators state that the direct anterior approach (DAA) provides faster recovery and less pain than the direct lateral approach (DLA) and the posterior approach (PA) (2, 3), even though the posited reduced muscular damage is still to be proved. Mjaaland et al (4) could not demonstrate that DAA determines a lower C-reactive protein elevation when compared with DLA, and unexpectedly found higher creatin kinase levels in the DAA group than in the DLA group. On the contrary, when the biochemistry of post-DAA muscular damage was compared with post-PA, the result was favorable, showing lower creatin kinase levels (5). Apart from these conflicting data concerning muscular injury, the anterior approach seems to be associated with a very low dislocation rate (6, 7), and this factor is often used by surgeons promoting DAA to claim its superiority against PA.
However, an increased rate of intraoperative complications has been reported with anterior approach techniques (8, 9). Intraoperative complications include trochanteric and calcar fractures, femoral perforations, bleeding, tensor fasciae latae tears, etc. These complications are mostly due to the more difficult exposure of the femur, both with and without the aid of a fracture table. For this reason, several surgeons using DAA have changed their implant choice from standard stems to short stems, whose reduced length lowers the chance that the broaches impinge against the iliac wing and the origin of the tensor fasciae latae, resulting in less chance of stem malalignment, femoral fractures and muscular damage (10, 11).
Despite an adequate surgical technique and the choice of a short stem, the authors propose that difficult femoral preparation and the surgeon's awareness of a significant rate of related complications might interfere with proper stem sizing. To test this hypothesis, a retrospective chart review was performed to assess the adherence of actual component sizing to preoperative planning in 2 groups of patients operated on with short-stem THA respectively through PA and DAA, both without intraoperative imaging.
Methods
A retrospective study was carried out reviewing the medical records and x-rays of all patients who underwent short stem THA. The study, being retrospective, guaranteeing participants anonymity and conforming to the principles of the Declaration of Helsinki as adopted in 1964 and revised in 2008, did not need specific ethical review board approval. Our indications to short-stem THA were limited to osteoarthritic hips in patients aged 75 years or less and with BMI ≤35. Inflammatory arthritis, severe osteoporosis, previous hip surgery or bony deformity needing femoral osteotomy were considered contraindications to a short stem.
Inclusion criteria for this study were conformity of the cases to the above set of indications/contraindications, usage of PA or DAA and implantation of the same short femoral implant (Fitmore femoral stem, Zimmer Inc, Warsaw, IN). Exclusion criteria were the unavailability of full medical records (included imaging and templating), usage of a different surgical approach (i.e. DLA) or stem design, previous contralateral THA (since the awareness of contralateral sizes and radiographic outcome might have influenced intraoperative decisions).
112 unilateral cases were selected according to these criteria: 59 hips underwent PA (group 1), while 53 hips underwent DAA (group 2). All the cases were operated by the same surgeon (FL), who had wide experience with both techniques. The indication to PA or DAA was determined through quasi-random allocation, following the order of inclusion on the waiting list and alternating the procedures performed with PA and DAA. The 2 groups were comparable as for age, body mass index (BMI) and sex distribution (Tab. I).
Demographic features of the 2 groups
All the patients had uncemented hip implants, which included an acetabular component with ceramic liner (Continuum cup, Zimmer Inc, Warsaw, IN), short femoral stem (Fitmore femoral stem, Zimmer Inc, Warsaw, IN) and ceramic head. All operations in group 1 were carried out under spinal anaesthesia. A standard PA was performed. All operations in group 2 were carried out under general anaesthesia to obtain the deepest muscle relaxation. A DAA was used, without fracture table aid (12).
Intraoperatively, neither fluoroscopic nor x-ray controls were used for estimation of cup inclination and anteversion or stem size and alignment, but all cases underwent accurate preoperative planning based on the AP pelvic x-ray.
In both groups of the study, preoperative planning was performed by 2 surgeons with long experience with this templating technique.
AP pelvic radiographs were taken in a standing position with both iliac spines at the same distance from the film. To estimate the length of the femoral neck, both femora were positioned at 15° to 20° of internal rotation, corresponding to the natural femoral anteversion. To control the magnification factor, the relation between the x-ray source, the hip and the film was fixed. Magnification was checked by use of a metal 32 mm head positioned as a calibration object at the level of the hip joint in the anteroposterior plane.
The templating was performed superposing the acetate implant drawings on the radiographs. Traditional landmarks (13, 14) were used to restore normal hip anatomy. Acetabular size, femoral size and head length were calculated and recorded for each planned case.
To evaluate the adherence to preoperative planning, we compared the expected acetabular and femoral size and the expected head length with implanted acetabular and femoral size and implanted head length in all patients of group 1 and group 2. For each component of each case the difference implanted size - expected size was calculated (size discrepancy, SDis).
Although in clinical practice the size prediction is usually considered successful if the implanted component is between +1 and -1 size from the expected one, for statistical purposes we defined nonagreement as any case of size discrepancy (SDis≠0). Thus the actual component was defined as undersized if at least 1 size smaller than expected (SDis ≤-1), and severely undersized if at least 2 sizes smaller (SDis ≤-2). Chi-square test (when frequencies were higher than 5) and Fisher's exact test (in the remaining cases) were performed to determine whether the rates of nonagreement, undersizing and severe undersizing in the 2 groups differed for a random effect or not. The level of statistical significance was set at p = 0.05.
If the null hypothesis was confuted (i.e. the rate of undersized stem is different between the 2 groups, and particularly is higher in the DAA group than in the PA group), to ascertain if the stems implanted through DAA are truly undersized or if their size was simply overpredicted, the canal fill index would have been calculated in postoperative x-rays as the ratio between stem width and canal width 2 cm below the tip of the lesser trochanter. Since there are no published reference values for the Fitmore stem, we would not have attempted to classify the cases into undersized and properly sized, but would have compared the mean canal fill between the two groups with a Student's t-test.
Lastly, the axial alignment would have been studied and the frequency of malaligned components (whose axis differ more than 3° from the proximal femoral axis) compared, in order to evaluate whether malalignment may contribute to stem undersizing.
Results
The distribution of size discrepancy for stem, cup and prosthetic head are shown in Figure 1 for PA (group 1) and in Figure 2 for DAA (group 2).
Size discrepancy distribution in group 1 (posterior approach THA).
Size discrepancy distribution in group 2 (direct anterior approach THA).
As for the nonagreement rate, no statistically significant differences between the 2 groups were observed for any component, although the stem size nonagreement rate was higher in group 2 than in group 1. If undersizing is analysed, the stem undersizing rate in group 2 turns out to be 3 times higher than in group 1 (54.72% vs 16.95%, p<0.001), while the stem severe undersizing rate is more than 6 times higher than in group 1 (24.53% vs 3.39%, p = 0.001) (Tab. II).
Rates of nonagreement, undersizing and severe undersizing compared between the 2 groups by Chi-square test or (*) Fisher's exact test
Noticeably, there is no significant difference of undersizing/severe undersizing rate of cup and head between the two groups, confirming that the difference of predictability is limited to the femoral component.
The size discrepancy distribution graphically represented in Figures 1 and 2 shows that the mode is 0 (i.e. perfect agreement) for both groups and for all the components (stem, head and cup), but the stem distribution in group 2 (DAA) is markedly asymmentrical, with undersized stems clearly prevailing over the oversized ones. This does not occur in any other series, whose distribution is symmetrical around the mode.
Given the above results, the postoperative x-rays were reviewed. 35 cases in each group had adequate anteroposterior postoperative imaging (neutral rotation of the limb, extension at least 2 cm below the end of the stem) and were used to estimate the canal fill. The head size was used for calibration, to determine the adequate level for the measuring stem and canal width. The canal fill index was compared between the two groups, and shown to be significantly lower in group 2 (82.63 ± 5.10% on average, ranging from 68% to 94%) than in group 1 (85.00 ± 5.47%, ranging from 77% to 94%) with p = 0.035, confirming that stems implanted through DAA filled the canal less tightly than stems implanted through PA.
As for stem alignment, only the frontal plane could be adequately assessed, since postoperative axial views were frequently compromised by malrotation, poor quality, and insufficient distal extension of the field. Just one stem turned out to be malaligned (varus), and was the most severely undersized (SDis = -4, canal fill index 68%). It belonged to group 2, although this is not statistically significant.
Discussion
The present study retrospectively compares the predictability of prosthetic component sizes with acetate templating between 2 groups of THAs, respectively implanted through PA and DAA by a surgeon skilled in both techniques. The main finding is the higher rate of undersized stems in the DAA group.
The prevalence of undersized (and severely undersized) stems in the DAA group is interpreted by the authors as an indirect sign of the greater difficulty of achieving proper sizing and positioning of the femoral component through DAA as compared to PA.
On the other side of the comparison, the symmetrical distribution of size discrepancy of prosthetic heads and acetabula around 0, with no different rates of undersized components between PA and DAA, confirms not only that such intrinsic technical difficulty is limited to the stem (as is reasonably expected in the light of the equally wide acetabular exposure guaranteed by both anterior and posterior approaches), but also that preoperative planning was performed with equal accuracy in both groups.
To evaluate size predictability, preoperative acetate templating was used and the predicted size was compared with the actual size. Despite the use of digital radiographs and planning software becoming more and more widespread (15), acetate templating is still reported to be the gold standard and has proven precision and accuracy (16, 17). The accuracy of hip templating is related to experience, and may reach an agreement rate between planned and implanted component size of 60% or more (14). For this reason we considered preoperative acetate planning performed by two expert surgeons a reproducible and accurate method of assessing the differences of predictability in these 2 groups of THAs.
The above conclusions were confirmed and refined by the assessment of postoperative x-rays, that showed that canal fill was truly higher in the PA group than in the DAA group, but that the frequency of frontally malaligned stems did not differ significantly. Thus the role of varus positioning should not be overemphasised, neither as a possible cause of undersizing, nor as a common pitfall of DAA. This result is consistent with the recent findings of Abe H and coworkers (18), who found a similar incidence of malaligned stems after DAA and PA in the coronal plane.
This study has some limits and drawbacks that should be considered. First, no inter- nor intraexaminer reliability test was performed. Second, no clinico-functional outcome was evaluated: we cannot conclude whether the undersized stems resulted in lower function, patient satisfaction or implant durability. Third, the postoperative x-ray assessment did not consider the axial view due to frequent technical inadequacy of the images, and this choice impeded the evaluation of the sagittal alignment; this is often found to be suboptimal in DAA if compared to PA (18) and could have contributed to poor sizing. Fourth, since postoperative CT scan is not routinely performed after THA, we could not assess the incidence of rotational malalignment and its potential contribution to sizing errors.
The direct anterior approach uses the intermuscular and internervous plane between the sartorius muscle and the rectus femoris medially and the tensor fasciae latae and glutei muscles laterally. While acetabular esposure is never an issue, femoral exposure can be demanding if the hip cannot be sufficiently extended. This technical difficulty can be increased in case of short vari necks, muscular and overweight patients. The use of short stems may facilitate the broach alignment into the femoral canal, but varus malpositioning, undersizing and risk of related fractures cannot be underemphasized. DAA demonstrated a reduction in operative time and fluoroscopy time only after initial 100 cases (19), therefore surgeons considering this approach should expect a substantial learning period. In consideration of the 30-year experience of the surgeon who performed all the THAs of this series, not using intraoperative fluoroscopy can only partially explain the relevant rates of undersized stems in the DAA group. As fluoroscopy was not employed in the PA group either, it could not impact on the differences between the 2 groups, but might have facilitated the occurrence of undersizing in both approaches, preventing the surgeon from correcting his possible errors.
Independently from a surgeon's experience, the awareness of a significant risk of intraoperative complications associated with femoral preparation (8, 9) and particularly the awareness of the technical difficulty of fixing an intraoperative femoral fracture through a DAA can make a surgeon excessively prudent. A surgeon's caution may influences the choice to be aggressive with sizing, avoiding the use of the largest stem possible (20).
In conclusion, our study demonstrates that the direct anterior approach may reduce the chance that the surgeon implant the planned femoral component, usually erring on the side of undersizing. Such a systematic mistake may be related to the technical difficulty of in obtaining an adequate broach alignment, although in our series frontal malalignment occurred exceptionally and cannot be considered a feasible explanation for undersizing, and to the caution related to the knowledge of possible femoral complications. For this reason we recommend intraoperative imaging to check stem size and alignment when a DAA is performed.
Further, prospective studies are needed to confirm these data and to evaluate if undersizing short stems through a direct anterior approach results in worse clinico-functional outcomes.
Footnotes
Financial support: None.
Conflict of interest: None.