No Clinically Important Differences in Thigh Pain or Bone Loss Between Short Stems and Conventional-length Stems in THA: A Randomized Clinical Trial

Abstract

Background

Short-length stems were developed to reduce bone loss of the proximal femur and potentially decrease the incidence of thigh pain after cementless THA. However, it remains unknown whether short stems indeed reduce bone loss or the frequency of thigh pain.

Questions/purposes

Is there a difference between short- and standard-length stems in terms of: (1) the frequency or severity of thigh pain, (2) modified Harris hip scores, (3) implant loosening, or (4) bone mineral density as measured by dual-energy x-ray absorptiometry?

Methods

Between March 2013 and January 2014, three surgeons performed 205 primary THAs. To be eligible, patients needed to be at least 20 years of age, have not undergone previous history of hip surgery, and have no metabolic bone disease. A total of 100 patients were randomized to receive THA either with a short stem (n = 56) or with a standard-length stem (n = 44). Both stems were proximally coated, tapered, cementless stems. Compared with standard stems, short stems typically were 30- to 35-mm shorter. A total of 73% (41 of 56) and 77% (34 of 44) of those groups, respectively, were accounted for at a minimum of 5 years and were analyzed. The presence of thigh pain during activity was evaluated using a 10-point VAS, and the modified Harris hip score was calculated by research assistants who were blinded to the treatment groups. Plain radiographs were taken at 6 weeks, 6 months, and 12 months postoperatively, and every 1 year thereafter; loosening was defined as subsidence > 3 mm or a position change > 3° on serial radiographs. Radiological assessment was performed by two researchers who did not participate in the surgery and follow-up evaluations. Bone mineral density of the proximal femur was measured using dual-energy x-ray absorptiometry at 4 days, 1 year, 2 years, and 5 years postoperatively. The primary endpoint of our study was the incidence of thigh pain during 5-year follow-up. Our study was powered at 80% to detect a 10% difference in the proportion of patients reporting thigh pain at the level of 0.05.

Results

With the numbers available, we found no difference between the groups in the proportion of patients with thigh pain; 16% (9 of 56) of patients in the short-stem group and 14% (6 of 44) of patients in the standard-stem group experienced thigh pain during the follow-up period (p = 0.79). In all patients, the pain was mild or moderate (VAS score of 4 or 6 points). Among the 15 available patients who reported thigh pain, there was no difference between the implant groups in mean severity of thigh pain (4.3 ± 0.8 versus 4.2 ± 0.7; p = 0.78). There were no between-group differences in the short versus standard-length stem groups in terms of mean modified Harris hip score by 5 years after surgery (89 ± 13 versus 95 ± 7 points; p = 0.06). No implant was loose and no hip underwent revision in either group. Patients in the short-stem group showed a slightly smaller decrease in bone mineral density in Gruen Zones 2, 3, and 5 than those in the standard-stem group did; the magnitude of the difference seems unlikely to be clinically important.

Conclusion

We found no clinically important differences (and few differences overall) between short and standard-length THA stems 5 years after surgery in a randomized trial. Consequently, we recommend that clinicians use standard-length stems in general practice because standard-length stems have a much longer published track record in other studies, and short stems can expose patients to the uncertainty associated with novelty, without any apparent offsetting benefit.

Level of Evidence

Level I, therapeutic study.

Introduction

Conventional cementless femoral stems have shown excellent clinical performance with durable survival and have especially been used in young patients with a long-life expectancy. However, one problem with conventional stems is thigh pain [12]. Theoretically, transmission of stem load to the proximal femoral metaphysis might reduce stress shielding and the risk of thigh pain [20, 26]. In addition, stress shielding around conventional cementless stems remains a matter of concern at long-term follow-up because it may increase the potential risk of loosening and periprosthetic femoral fracture [3, 13, 26]. Short stems, which are 30- to 35-mm shorter than conventional length stems, have been developed under the assumption that these new designs might move the stem-to-femur load more proximally, and thus might reduce stress shielding and the development of thigh pain [20]. A variety of short metaphyseal-fitting stems have been developed [20].

However, whether these short stems really reduce stress shielding or the frequency of postoperative thigh pain is unknown. Some studies have presented similar clinical outcome scores and survivorship rates of short stem compared with standard stems [13, 14, 15]. In terms of thigh pain, previous studies reported the various incidence ranging from 1% to 24% [1, 13, 14], and no definitive study of which we are aware has been able to conclude whether the frequency of this painful complication is lower in patients treated with short stems.

We, therefore, performed a randomized clinical trial, in which we asked: Is there a difference between short- and standard-length stems in terms of: (1) the frequency and severity of thigh pain, (2) modified Harris hip scores, (3) implant loosening, or (4) bone mineral density as measured by dual-energy x-ray absorptiometry?

Patients and Methods

Study Design

This randomized clinical trial was conducted at two tertiary referral hospitals, and its design and protocol were approved by the institutional review board of each hospital. The study was registered in the Clinical Trials.gov protocol registration system (trial no. NCT02087436). Patients were informed about the study, and those who participated provided consent for this study before undergoing THA.

Patients

Patients who underwent primary THA at the two institutions were eligible for inclusion in the study. The inclusion criteria were age older than 20 years, no history of hip surgery, and the patient’s willingness to participate in the study. Patients who could not be operated on with the predetermined prostheses because of a severe defect or hypoplasia of the acetabulum or proximal femur were excluded. Patients who had metabolic bone disease such as Paget’s disease of bone or renal osteodystrophy also were excluded.

Enrolled patients were randomly assigned to undergo THA with either a short-length stem (short-stem group) or a standard-length stem (standard-stem group). A simple randomization was performed by means of an interactive central telephone system. Both the short stem and standard stem were templated preoperatively for each patient. The surgical team was notified which stem to use as they performed the surgical approach. The patients and investigators (research assistants), who asked about thigh pain and modified Harris hip score elements, were blinded during the evaluation. A total of 205 patients who underwent THA between March 2013 and January 2014 were screened for eligibility of this study. Thirty-three patients who had a history of previous surgery, three patients with renal osteodystrophy, and two patients who had hypoplastic femurs with a narrow (≤ 8 mm) medullary canal were excluded. Sixty-seven patients declined to participate in this study. The remaining 100 patients who met our inclusion criteria were enrolled in this study (Fig. 1). Eight patients underwent staged bilateral THA at an interval of 1 week between procedures. For these eight patients, only the first operated-on hip was enrolled and evaluated in this study.

Fig. 1.

This flowchart shows follow-up in both groups.

Fifty-six hips were assigned to the short-stem group and received the TaperLoc® Microplasty stem (Zimmer Biomet, Warsaw, IN, USA). Forty-four hips were assigned to the standard-stem group and received the standard TaperLoc stem (Zimmer Biomet) (Fig. 2). There were no differences in baseline demographics between the two groups at the time of THA, except for age. The short-stem group was slightly older than standard-stem group (50 ± 11 versus 45 ± 12; p = 0.04) (Table 1).

Fig. 2.

A-B This image shows (A) the TaperLoc® Microplasty stem (Zimmer Biomet, Warsaw, IN, USA) and (B) the standard TaperLoc stem, which are made of titanium alloy and have identical geometry in the proximal portion. The Microplasty stem is 35 mm shorter than the standard stem.

Table 1.

Comparison of basic demographics between the short-stem and standard-stem groups

Parameter	Short-stem group (n = 56)	Standard-stem group (n = 44)	p value
Surgeon			0.73
A, % (n)	34 (19)	27 (12)
B, % (n)	30 (17)	36 (16)
C, % (n)	36 (20)	36 (16)
% of women (n)	43 (24)	36 (16)	0.51
Age in years, mean ± SD	50 ± 11	45 ± 12	0.04
BMI in kg/m2, mean ± SD	25 ± 4	24 ± 4	0.73
Primary diagnosis			0.26
Femoral head osteonecrosis, % (n)	68 (38)	73 (32)
Primary osteoarthritis, % (n)	27 (15)	16 (7)
Secondary osteoarthritis, % (n)	5 (3)	7 (3)
Femoral neck fracture, % (n)	0 (0)	5 (2)
Preoperative modified Harris hip score, mean ± SD	55 ± 14	58 ± 16	0.26
Operation time in minutes, mean ± SD	104 ± 33	111 ± 39	0.33

A total of 73% (41) of the short-stem group and 77% (34) of the standard-stem groups were followed-up for at a minimum of 5 years, were analyzed (Fig. 1).

Stem Designs

Both the TaperLoc Microplasty (Zimmer Biomet) and the standard TaperLoc (Zimmer Biomet) stem designs are made of titanium alloy and have identical geometry in the proximal portion. The proximal portions are porous-coated, and the average pore size is 247 μm. The TaperLoc Microplasty stem (Zimmer Biomet) is 35 mm shorter than the standard stem.

Each stem has 17 sizes; the TaperLoc Microplasty stem (Zimmer Biomet) length ranges from 93 mm to 132 mm and the standard stem ranges from 128 mm to 167 mm. The distal tip of the TaperLoc Microplasty stem (Zimmer Biomet) is rounded in the lateral aspect to avoid varus positioning of the stem (Fig. 2).

A single cementless acetabular component design (ABT; Zimmer Biomet) and Delta ceramic-on-ceramic bearings (Biolox Delta, CeramTec, Plochingen, Germany) were used in both groups. The diameter of the ceramic bearing was 32 mm in 48 mm to 52 mm metal shells and 36 mm in those ≥ 54 mm.

Operative Technique

All operations were performed by three high-volume (more than 200 hip procedures per year) hip surgeons (YKL, YCH, KHK) using the Kocher-Langenbeck posterolateral approach [29]. The acetabular and femoral components were inserted in a press-fit manner.

During the study period, we used the combined anteversion concept [4]. We prepared the femoral side first. Cancellous bone in the femoral canal was removed with a box chisel and a single starter reamer was inserted into the distal femoral canal to the level appropriate to the templated stem size. The femoral canal was prepared using stem-specific broaches. We used the smallest size broach first and then progressively increased the broach size until it tightly engaged the medial and lateral cortex of the proximal femur. After that, we measured stem version of the final broach. The target anteversion of the metal shell was calculated with a formula: cup anteversion = 37.3° - 0.7 x stem anteversion [30]. The target abduction of the metal shell was 43° [10]. The acetabular cup was positioned at the target using the method described by Ha et al. [10]. The posterior capsule and short external rotators were repaired using transosseous suture through three drill holes in the trochanteric crest to prevent dislocation [11].

Postoperative Management

Patients were encouraged to start walking with crutches the next day after the surgery. We recommended using crutches and restricting full weightbearing for 1 month.

Study Outcomes

The primary outcome measures were the proportion and severity of thigh pain in each group during the follow-up period, and the secondary outcome measures were the change in bone mineral density (BMD) of the proximal femur, improvement in the modified Harris hip score, radiologic changes, implant fixation, and survival at 1 year, 2 years, and 5 years after THA.

Evaluation for thigh pain and modified Harris hip score were performed by three research assistants (JOS, SK, EHC) who were blinded to study-group allocation.

Follow-up Evaluations and Assessment of Endpoints

Follow-up visits were scheduled at 6 weeks, 6 months, and 1 year postoperatively, and every 1 year thereafter. At each follow-up visit, the presence and intensity of thigh pain were evaluated, the modified Harris hip score was calculated [5], and plain radiographs (AP and translateral views) were taken. The BMD of the proximal femur was measured at 4 days, 1 year, 2 years, and 5 years postoperatively. Follow-up evaluations were done in 99 hips at 6 weeks, 98 hips at 6 months, 93 hips at 1 year, 89 hips at 2 years, and 75 hips at 5 years (Fig. 1). The proportions of missing in BMD measurement were similar between both groups; 5% (3 of 56) and 9% (4 of 44) at 1 year; 11% (6 of 56) and 11% (5 of 44) at 2 years, 27% (15 of 56) and 23% (10 of 44) at 5 years.

Patients were specifically asked with a questionnaire and face-to-face interview whether they had thigh pain during activities of daily living [12]. A diagnosis of thigh pain was made according to the definition of Barrack et al. [2]: pain on the anterior and/or lateral thigh below the inguinal area. When a patient had pain over the posterior thigh or gluteal region or pain that radiated to the lower leg, we reviewed radiographs of the lumbar and sacral spines. When we saw spine arthrosis, we thought that spinal stenosis or lumbosacral arthritis might be the etiology of pain. The intensity of thigh pain, if present, was measured on a 10-point VAS (0 = no pain; 10 = severe pain) at each follow-up visit [12].

Femoral stem fixation was classified according to the method of Kim et al. [16] and acetabular component fixation was classified according to the method of Latimer and Lachiewicz [18]. A stem was considered loose when there was a subsidence > 3 mm or position change > 3° on follow-up radiographs compared with the baseline radiograph [16]. These radiographic evaluations were done by two independent observers (SHW, JWP) who did not participate in surgery. The 6-week radiographs were used as the baseline for radiographic comparison.

The original Gruen system was designed for a cemented stem, and the system divided the stem into thirds: proximal, middle, and distal [8]. Patil et al. [25] modified the Gruen zones for cementless stems. In the modified system, the proximal half of the lateral porous coated area is Zone 1 and the distal half is Zone 2. Zones 6 and 7 represent the corresponding areas on the medial side. Zones 3, 4, and 5 represent the lateral smooth portion, the distal smooth portion, and the medial smooth portion of the stem, respectively.

We measured the BMD of the proximal femur on dual-energy x-ray absorptiometry (Lunar densitometer; GE Healthcare, Madison, WI, USA) in the metal-removal hip-scanning mode. Some patients declined the BMD measurement because of a concern about the radiation hazard of dual-energy x-ray absorptiometry. The femoral scan was started approximately 2.5 cm distal to the tip of the femoral stem and continued proximally to 4 cm above the tip of the greater trochanter. The scan had a field width of 15 cm and oriented the longitudinal axis of the stem shaft in the vertical position. We measured the BMD at the seven zones of modified Gruen et al. [8] (Fig. 3). The BMD measured on the first dual-energy x-ray absorptiometry image, which was taken 4 days postoperatively, served as the baseline value for comparison. A follow-up dual-energy x-ray absorptiometry scan was taken in 83 hips (47 short-stem hips and 36 standard-stem hips) at 1 year, 88 hips (49 short-stem hips and 39 standard-stem hips) at 2 years, and 47 hips (26 short-stem hips and 21 standard-stem hips) at 5 years.

Fig. 3.

A-B Dual-energy x-ray absorptiometry images of the hip show BMD, which was measured according to the seven zones of Gruen et al. [8], in hips with (A) the TaperLoc® Microplasty stem (Zimmer Biomet) and (B) the standard TaperLoc stem. A color image accompanies the online version of this article.

Sample Size Calculation

The number of patients required for the study was estimated with a sample size calculation. The calculation was based on detecting a difference in the incidence of postoperative thigh pain between two study groups: the short-stem group and standard-stem group, assuming an overall α error (two-sided) of 5% with a statistical power of 80% (β error = 0.20). We assumed that the incidence of thigh pain in the standard-stem group would be 15% compared with 5% in the short-stem group. With these assumptions, 40 patients would be needed in each of the two groups [13]. To compensate for future dropouts and loss to follow-up, we determined that a sample size of 50 patients would be needed in each group.

Statistical Analysis

Continuous variables were compared between the two groups using a t-test, and categorical variables were compared using a chi-square test. All statistical analyses were performed with the Statistical Package for the Social Sciences (version 16.0; SPSS Inc, Chicago, IL, USA). Differences were considered significant at p < 0.05.

Results

Frequency and Severity of Thigh Pain so

With the numbers available, we found no difference between the groups in the proportion of patients with thigh pain; 16% (9 of 56) of patients in the short-stem group and 14% (6 of 44) of patients in the standard-stem group experienced thigh pain during the follow-up period (p = 0.79). In all patients, the pain was mild or moderate (4 to 6 points on the VAS). There was no difference between the implant groups in the mean severity of thigh pain (4.3 ± 0.8 versus 4.2 ± 0.7; p = 0.78). The pain did not limit activity in any patient and no patients used medication to control pain. The pain developed at a median (range) of 8.6 months (2 to 24) postoperatively in the short-stem group and at a median of 9.5 months (1.5 to 35) postoperatively in the standard-stem group. The pain disappeared at a mean of 20 ± 14 months in eight patients with short-stem hips and at a mean of 19 ± 17 months in five patients with standard-stem hips (p = 0.85). Two patients (one with a short stem and one with a standard-length stem) had persistent thigh pain until the last follow-up examination.

Modified Harris Hip Score

There were no differences in the modified Harris hip scores between the two groups at each follow-up visit, except at 1 year (Table 2), at which time patients who received a standard-length stem had slightly higher Harris hip scores; however, the observed difference was unlikely to be clinically important. Five years after surgery, there were still no differences between the short and standard-length stem groups with the numbers available (89 ± 13 versus 95 ± 7 points, mean difference 6 points [95% CI 86.3 to 91.1 points versus 92.6 to 97.7]; p = 0.06).

Table 2.

Postoperative modified Harris hip score in the short-stem group and standard-stem group

Time of evaluation	Short-stem group	Standard-stem group	p value
6 months postoperatively	83 ± 13	83 ± 15	0.89
1 year postoperatively	89 ± 11	94 ± 9	0.04
2 years postoperatively	95 ± 7	97 ± 5	0.13
5 years postoperatively	89 ± 13	95 ± 7	0.06

Data are presented as the mean ± SD.

Stem Loosening

No implant became loose, and no hip underwent revision in either group.

Change in BMD

In general, there were few differences in BMD between the stem designs, and the differences we observed were small, intermittent, and clinically unimportant. In both groups, BMD decreased in all Gruen zones during the follow-up period (Fig. 4). The decrease in BMD from the baseline measurement was slightly less severe in the short-stem group than in the standard-stem group in Gruen Zone 2 at 1 year and 2 years and Gruen Zones 3 and 5 at 5 years (Table 3).

Fig. 4.

A-G These graphs show that the mean with SD bone mineral density decreased in (A) Gruen Zone 1, (B) Gruen Zone 2, (C) Gruen Zone 3, (D) Gruen Zone 4, (E) Gruen Zone 5, (F) Gruen Zone 6, and (G) Gruen Zone 7 during the follow-up period. A color image accompanies the online version of this article.

Table 3.

Comparisons of BMD changes at the seven Gruen zones between the short-stem and standard-stem groups

Gruen zone	Time lapse (years)	Short-stem group	Standard-stem group	p value
Zone 1	1	-7 ± 16	-13 ± 13	0.11
	2	-8 ± 19	-8 ± 23	0.47
	5	-19 ± 17	-27 ± 8	0.14
Zone 2	1	-2 ± 11	-6 ± 11	0.03
	2	-1 ± 14	-7 ± 10	0.04
	5	-19 ± 18	-25 ± 6	0.12
Zone 3	1	-3 ± 10	-1 ± 11	0.41
	2	-1 ± 12	-2 ± 8	0.80
	5	-12 ± 15	-19 ± 6	0.04
Zone 4	1	-4 ± 10	-4 ± 8	0.59
	2	-4 ± 11	-4 ± 6	0.29
	5	-12 ± 15	-15 ± 5	0.43
Zone 5	1	-5 ± 11	-4 ± 8	0.61
	2	-6 ± 10	-5 ± 6	0.40
	5	-7 ± 13	-15 ± 6	0.01
Zone 6	1	-3 ± 11	-4 ± 9	0.57
	2	-3 ± 14	-3 ± 8	0.94
	5	-13 ± 20	-20 ± 6	0.16
Zone 7	1	-2 ± 24	-1 ± 15	0.37
	2	-1 ± 25	-4 ± 14	0.79
	5	-24 ± 31	-31 ± 12	0.98

Data are presented as mean ± SD.

Discussion

Although numerous designs of cementless stems have been developed and show satisfactory long-term results and survival rates [9, 19], thigh pain and periprosthetic bone loss are unsolved problems related to these stems [9]. Short stems have been developed based on the assumption that transmission of stem load to the proximal femoral metaphysis might reduce stress shielding and possibly the risk of thigh pain. In this amply powered randomized trial, we found no substantial differences between short- and standard-length stem designs in terms of thigh pain, hip scores, loosening, or bone remodeling. Consequently, we recommend using the standard-length stem design, which has more long-term follow-up studies supporting its durability.

Limitations

There were some limitations to this study. First, loss to follow-up was about 25% at 5 years postoperatively. However, there was no difference in the proportion of patients who were lost to follow-up between the two groups, and so it is unlikely that this level of missingness would assert a differential effect between groups. Second, we performed a relatively large number of statistical comparisons without a correction for multiple inference tests. This should cause the reader to de-emphasize the few, small differences we observed in BMD over time. Third, we used only one stem design in each study group in this study. Currently, there are various short-stem designs with different lengths, shapes, tapering angles, and extent of porous coating. Our findings might not be generalizable to other short-stem designs, although there is little reason to believe they would not generalize; the stem designs we used employed geometries and surface coatings that are offered by several manufacturers and are in wide use. Fourth, we diagnosed thigh pain according the definition of Barrack et al. [2]. Although we carefully evaluated the nature of thigh pain to exclude the chance that it could have been related to spinal problems, it was difficult to discriminate true thigh pain from radiating pain in some patients. In these patients, spinal problems might have partially contributed to the development of thigh pain. Fifth, our study was done in South Korea, the mean body mass index of our patients was low (24.5 kg/m²), and the most common diagnosis for THA was osteonecrosis of the femoral head. Our results may not generalize to populations that differ in those respects, but we suspect that our findings probably do generalize, as those factors are unlikely to influence the endpoints we evaluated. Sixth, we could not determine to what degree the decrease of BMD may have been related to age-related bone loss of the proximal femur because we did not compare the BMD change in study patients with control patients who did not undergo THA. Finally, patients who received the short stem were older and had lower baseline BMD than patients who received the standard-length stem. Nevertheless, the decrement of BMD was less severe in the patients who received the short stem.

Frequency and Severity of Thigh Pain

We found no difference in the patient proportion or the severity of thigh pain between the two groups. Prior studies reported various frequencies of thigh pain after the use of numerous short-stem designs [1, 13, 14]. Although none of 60 patients had thigh pain in a study of Proxima stem (DePuy, Warsaw, IN, USA) [15], 24% of hips (58 of 238) had thigh pain after THA with use of TaperLoc Microplasty stem (Zimmer Biomet) in another study [1]. The different diagnostic criteria in each study, the different patient population, and the different study nature (prospective study with a specific questionnaire on thigh pain versus retrospective review) might explain the differences in observed thigh pain frequency. In our study, the primary outcome was the occurrence of thigh pain, which we evaluated using a questionnaire, face-to-face patient interview, and a physical examination to exclude radiating pain from spinal problems. This thorough assessment of thigh pain could explain the higher frequency of thigh pain here than has been observed in some other studies.

Modified Harris Hip Score

There were few between-group differences in hip scores; the slight difference observed 1 year after surgery favored the standard-length stem group, but it was small, and unlikely to be clinically important. This is generally consistent with studies of both short [7, 21, 23] and standard-length stems [22, 24], both of which generally have shown good hip scores at short- to intermediate-term follow-up. We believe that no-difference findings should generally favor the more established approach, which in this case, is the standard-length stem.

Stem Loosening

In this study, we observed no loosening in either stem group. This is consistent with considerable evidence at intermediate term both with short stems [7, 21, 23] and standard-length stems [22, 24]. Our study was almost certainly underpowered on this endpoint, and so we recommend no firm conclusions be drawn from our findings; we present them mainly so that they can be pooled into future systematic reviews.

Change in BMD

The differences in BMD between the groups, while slightly favoring the short stems, were generally small (on the order of a few percent, at most), and should be no basis for a stem choice.

Several studies have compared bone remodeling of the proximal femur between short-stem designs and conventional-length stems by measuring the change in BMD [6, 13, 14, 17, 28]. In those studies, short stems preserved the proximal femur better than the conventional-length stems. However, the proximal geometry of the short stem and standard-length stem were different in those studies, and the region of interest for the BMD measurement was different in each study. In the current study, we compared two stem designs that had identical geometry of the proximal porous-coated portion, and we evaluated BMD changes according to the modified Gruen zones [8, 25]. Schilcher et al. [27] also compared the BMD changes between TaperLoc Microplasty stem (Zimmer Biomet) and standard TaperLoc stem in a randomized clinical trial of 60 THAs. They did not find significant differences in periprosthetic bone loss between the two stem designs at 2 years [27]. Considering that bone loss caused by stress shielding might be an insidious phenomenon after cementless THA, the difference of bone loss of the proximal femur between the study of Schilcher et al. [27] and ours might be caused the difference of follow-up period.

Conclusion

In this randomized trial, we found no substantial differences between short- and standard-length stem designs, in terms of thigh pain, hip scores, loosening, or bone remodeling 5 years after THA. Consequently, we recommend using standard-length stems in practice unless there is some compelling reason to do otherwise (such as hardware or deformity) because standard-length stems have a much longer published track record in other studies. In addition, short stems can expose patients to the uncertainty associated with novelty, without any apparent offsetting benefit.