More than 43,000 hip replacements are carried out in the UK National Health Service every year. It is one of the most common and effective surgical procedures performed in the field of orthopaedics. Single-stage bilateral operations, which have the advantages of reduced hospital stay, one anesthetic session, and the subsequently reduced costs, are becoming increasingly common. However, there are concerns regarding the degree of surgical trauma to the body, increased blood loss, and the impact these may have on existing comorbidities. Macaulay, Salvati, Sculco, and Pellicci, (2002) previously reported that single-stage bilateral operations should be contraindicated in patients with significant comorbidities such as heart disease and diabetes.
This study compared bilateral sequential and bilateral simultaneous hip replacements with a single hip system: collarless polished taper. The design principles of this system have been used for more than 30 years in the treatment of osteoarthritis, rheumatoid arthritis, and other disabling conditions of the hip. The acetabular components are a combination of cemented and uncemented. While this may have a bearing on operative time, and by extrapolation, an effect on the complication rate, it was not thought to be a likely major confounding factor within the context of this study.
Literature Review
Several previous studies have compared bilateral sequential and simultaneous groups, but no study has yet been published using only one cemented femoral component type. Eggli, Huckell, and Ganz (1996) concluded that there was no difference in complication rates, but that the simultaneous procedure resulted in a reduced length of hospital stay and its associated cost. This study included relatively large numbers: 128 hips in the simultaneous group and 382 hips in the sequential group. The limitations of this study were the use of different prosthesis and the short-term follow-up (mean follow-up of 1.5 years). Another study by Parvizi et al. (2006) reported reduced length of stay with the simultaneous procedure and, in addition, showed higher complication rates in the sequential group. This study only evaluated uncemented stems. A further study by Alfaro-Adrian, Bayona, Rech, and Murray (1999) advocated the simultaneous procedure but accepted that there was an associated higher blood transfusion rate. Again, this study included more than one stem type. Macaulay et al. (2002) accepted that in medically fit patients, postoperative complications are within acceptable limits in simultaneous bilateral operations, but stated that complication rates can be 1.3 times higher in this group in comparison to unilateral surgery. Lorenze, Huo, Zatorski, and Keggi (1998) reported no differences between bilateral sequential and simultaneous groups in perioperative complications but highlighted the reduced length of stay associated with the simultaneous procedure. One limitation of this study was that it included only 40 patients in each group. Berend et al. (2005) used similar outcome measures but only compared bilateral simultaneous to unilateral hip arthroplasty. Their study showed prosthetic survival to be similar in both groups, with no difference in Harris Hip Scores (HHS). However, an increased risk of thromboembolic complications in the simultaneous bilateral group implied that this procedure should perhaps be performed on patients with no significant comorbidities. Deep infection rates in bilateral sequential and simultaneous groups have been reported to be similar (Huotari, Lyytikainen, & Seitsalo, 2007).
Materials and Methods
Data were extracted from a multicenter hip database. Bilateral cases dated back to 1993, so operations from 1993 to 2005 were included. This database contains data on hip systems collected by independent practitioners and processed within our clinical audit services unit. Surgeons participate in this outcome study to monitor their own performance, the performance of the prostheses used, and patient satisfaction. All patients of the 32 participating surgeons from 10 centers throughout the United Kingdom admitted for elective hip arthroplasty surgery consented for electronic storage of their data and subsequent usage for dissemination. Each center had a clinical audit practitioner, who was trained by a single coordinator to collect the data according to a single protocol. This multiplicity of surgeons and centers avoids drawing conclusions on the basis of an isolated practice in a single "center of excellence." It therefore is relevant to everyday orthopaedic practice.
Preoperative baseline information and HHS, as described by Harris (1969), were collected by independent clinical audit practitioners. The same fully trained independent practitioner then reviewed patients at 1, 3, 5, and 7 years postoperatively. HHS, complications, revisions, and patient satisfaction were recorded at each review. Patient satisfaction was ascertained by asking patients whether they were happy with their surgery and if not why not. From the database, 594 patients were identified as having undergone bilateral hip replacement surgery. The simultaneous group comprised 68 patients (26 men and 42 women) with the sequential group having 526 (208 men and 318 women). The sequential group had more patients as it is presently much less common to perform bilateral simultaneous hip replacement surgery than bilateral sequential staged surgery in the United Kingdom.
Statistical Methods
Variables were explored for distribution and extreme values. Plausible normality of distribution was determined using the Shapiro-Wilks test. Outlying data points were identified and investigated.
Where appropriate, confidence intervals for quantitative variables in each group (sequential and simultaneous), and the difference between them, were established. Repeated-measures analysis of variance was applied to HHS at pre- and postoperative reviews. The within-subjects factor was Review and the between-subjects factor was Group. Variables with extreme values were investigated using the Mann-Whitney U test. Categorical variables were investigated using chi-square methods.
Results
The demographics of patients in the sequential and simultaneous groups were similar (see Table 1). The proportions of men and women in the sequential and simultaneous groups were comparable (39.5% male in the sequential group and 38.2% in the simultaneous group; p = .84). Patients in the sequential group were slightly older than those in the simultaneous group (mean age = 66.5 and 61.5 years, respectively; p < .001), but both groups lay firmly within the "standard" age range for the procedure.
The sequential group had a lower percentage of patients with Charnley classification C (bilateral disease with comorbidities); 76% compared with 93% in the simultaneous group. This may reflect the practice of some surgeons in offering a single operation to less-healthy patients, in the belief that this may lead to less cumulative morbidity than two operations, particularly in the controlled environment of the operating theatre. Clearly the simultaneous group did not include carefully selected healthier patients.
The mean BMI for each group was in the overweight range but comparable across groups (27.2 and 27.4 for sequential and simultaneous patients, respectively; p = .78). As can be expected, the preoperative HHS scores were in the poor range with marginally higher scores in the sequential group, however not significantly different. At 1 year postoperatively, HHS scores were improved significantly with slightly higher scores in the sequential group (82.2 [0.5] falling in the good range) compared to the simultaneous group (78.9 [1.8] falling in the fair range). This difference was significant at the p = .04 level. In looking at HHS improvement scores, the difference between the 1 year and the preoperative score was calculated. Improvements in HHS at 1 year were comparable (41.4 and 42.3, respectively; p = .66) across groups. At 3- and 5-year reviews, HHS was comparable in both groups with mean scores falling in the good range. The 7-year review for the simultaneous group had HHS of 79.5 in comparison with 83.2 for the sequential group, but the numbers were small (see Table 1).
There were several exceptional lengths of stay in each of the two groups that were attributable to complications. Including extremes, the median stay in the sequential and simultaneous groups were 9 and 13.5 days, respectively (p < .001). When cases with complications were excluded, median stays were 9 and 12 days, respectively (p < .001). Because of the long-term follow-up involved in this study, some operations dated back to 1993, thus explaining the long lengths of stay. The mean length of stay of patients operated on for a primary hip replacement in 1993 was 15.9 days in comparison with 8.9 days in 2005.
Satisfaction rates are displayed as a percentage of those who answered that they were satisfied with their surgery against those who were not satisfied. This question was asked at each review period. There were no significant between-group differences in satisfaction rates at any postoperative review (see Table 2). However, patients who underwent sequential hip replacement were less likely to have adverse events, as shown by the lower percentage of readmissions (see Table 2). With the exception of superficial infection, these differences were not statistically significant.
The incidence of superficial wound infection in the simultaneous group was 5.8% (4 cases of 68) in comparison with 1.5% (8 cases of 526) in the sequential group. This difference was statistically significant (p = .04). There were no cases of thromboembolic episodes in the simultaneous group although 8 cases (1.5%) occurred in the sequential group. The dislocation rate in the simultaneous group was 5.8% and 3.2% in the sequential group. Dislocation rates of 14.5% at 10-year follow-up have been reported using the Trent regional arthroplasty register (Allami et al., 2006).
Discussion
The results from our collarless polished taper data showed a reduced length of stay but an increased incidence of adverse events in the simultaneous bilateral group. This is consistent with the findings of other studies.
The incidence of superficial wound infection in the simultaneous group was 5.8% in comparison with 1.5% in the sequential group. Our finding differs from that of Huotari et al. (2007), who reported no between-group difference regarding infection.
An increase in the incidence of thromboembolic episodes in the simultaneous group has been suggested in the literature. Deep venous thrombosis (DVT) rates of 2.8% and 9.1% have been recorded after unilateral total hip replacement (Sudo et al., 2003; White, Romano, Zhou, Rodrigo, & Bargar, 1998). None was recorded in the simultaneous group in our study, although 1.5% occurred in the sequential group. Many surgeons cite anecdotal evidence of increased thromboembolic problems in patients with bilateral surgery as a reason not to perform it. The present study suggests that this view is not supported by the evidence.
The dislocation rate in the sequential group was lower than that in the simultaneous group (3.2% compared with 5.8%) and those rates reported previously, but much lower than those reported in the Trent registry after long-term review (Allami et al., 2006; Khatod, Barber, Paxton, Namba, & Fithian, 2006; Meek, Allan, Mcphillips, Kerr, & Howie, 2006).
The rate of revision appeared higher in the simultaneous group, because of a single case having a fall resulting in a periprosthetic femoral fracture. Reasons for revision in the sequential group were infection, groin pain due to leg lengthening, and femoral fracture after a fall.
Interestingly, the mean preoperative HHS for the simultaneous group was lower (see Table 1), implying a slightly more disabled patient group in terms of pain and function. This also is supported by the fact that the simultaneous group had 93% patients with Charnley classification C, indicating the presence of comorbidities in comparison with 76% in the sequential group, which had 15% of patients with Charnley classification A, with simultaneous group having none. Thus, simultaneous bilateral procedures were performed on patients with more comorbidities than the sequential group. This is contrary to what may be expected.
The strengths of this study are the use of a single protocol, independent assessors, independent analysis, and one type of femoral stem for all patients (see Figure 1), thus reducing bias from confounding factors. The limitations are the relatively small numbers in the simultaneous group and the participation of more than one surgeon, which may have resulted in the use of different approaches and surgical techniques. In addition, more in-depth information could have been collected on existing comorbidities. It would also have been of interest to ascertain any nursing care issues related to simultaneous bilateral operations. There was an increased dislocation rate in the simultaneous group, but only one occurred during the acute hospital stay.
Conclusion
Sequential bilateral surgery is performed more commonly than simultaneous bilateral surgery in the United Kingdom. While there are cost implications with increased length of stay for the sequential bilateral group, this may be offset by the lower incidence of adverse events.
Patient satisfaction rates were consistent and high in both simultaneous and sequential groups. HHS improvements were also similar and reflected a marked improvement in pain and function. These results do not show any major problem with either approach to the common clinical presentation of bilateral hip arthritis. Authoritative statements that either the sequential or the simultaneous procedure is better seem unjustified. The decision is likely to continue to be based upon surgeon and patient preference, rather than on dubious claims related to either mortality or cost, both of which are often cited within the surgical community.
ACKNOWLEDGMENTS
The authors have no professional or financial affiliations that may be perceived to have biased the presentation. The authors thank the contributing surgeons for access to their anonymized data.
REFERENCES