Are the First or the Second Hips of Staged Bilateral THAs More Similar to Unilateral Procedures? A Study from the Swedish Hip Arthroplasty Register

Abstract

Background: Bilateral THAs performed in the same patient should not be considered independent observations, neither biologically nor statistically. As a result, when surgical results are reviewed, it is common to analyze only the first of the two hips, assuming that the first, and not the second hip of a staged bilateral THA, better resembles unilateral THAs. This assumption has not been empirically justified.Question/purposes (1) In patients with staged bilateral THA, is the first or second hip more similar to a unilateral THA in terms of age at surgery, presence of any preoperative Charlson comorbidity, and risk of postoperative reoperation? (2) Should the date of a first or second hip surgery of a staged bilateral THA be used as a starting point for patient survival to better resemble patients with unilateral THA?

Methods: We identified 68,357 THAs due to osteoarthritis in 63,613 patients from the Swedish Hip Arthroplasty Register (SHAR) in 1999-2015. Of those THAs, 14,780 concerned the first hip of a staged bilateral procedure performed between 1999 and 2004; 28,542 were unilaterals from 2004 to 2008, and 25,035 concerned the second hip of a staged bilateral procedure performed 2008 to 2015. We excluded patients who underwent one-stage bilateral THAs. We used different inclusion periods to distinguish unilateral procedures from the first and second hips from staged bilateral procedures because sufficiently long set-up and follow-up periods were needed before and after each period to identify possible contralateral THAs. This introduced potential period confounding, meaning that possible group differences might not be distinguished from unrelated outcome differences over time. We investigated if such time trends existed. It did not for age and reoperation rates, but it did for comorbidity and patient survival. Our primary study endpoint was whether patients with unilateral THAs were more similar to patients with a first hip of a staged bilateral THA, or to patients with their second hip operated. We used Student's t-test to compare mean age at surgery. The proportion of patients with at least one presurgery Charlson comorbidity were compared by 95% bootstrap confidence intervals, after subtracting the yearly time-trend to avoid period confounding. Postoperative risks of reoperation were compared by log-rank tests of Kaplan-Meier curves and by comparing 5-year reoperation rates by pair-wise 95% CIs. Our secondary study endpoint was to compare patient survival for patients with a unilateral THA, a first hip of a staged bilateral THA, or a second hip of a staged bilateral THA. We evaluated this by relative 5-year survival, comparing patients of each group with the general Swedish population of the same age, sex, and year of birth. This way, possible survival differences would be less likely explained by period confounding.

Results: Patients undergoing unilateral THA were older than those undergoing a first hip of a staged bilateral THA (70 ± 10 versus 66 ± 9 years, mean difference of 4; p < .001), but they were not different from patients undergoing the second hip of a staged bilateral THA (70 ± 9 years, mean difference of 0; p = 0.74). The period-adjusted proportion of patients with unilateral THA and presurgery comorbidity (Charlson index > 0) was 20% (95% CI: 19.8-20.7). This was no different from patients with a second hip from a staged bilateral THA (20%; 19.7-20.6), but higher compared to patients with a first hip of a staged bilateral THA (15%; 14.5-15.4). For reoperation rates, the log-rank tests showed no difference between unilateral THAs and the second hips of staged bilateral THAs ((Equation is included in full-text article.)). Such difference was found for unilaterals compared with the first hips of staged bilateral THAs ((Equation is included in full-text article.)). The Kaplan-Meier estimate of reoperation rates at 5 years after surgery were also no different for the unilateral THAs compared with the second hips of staged bilateral THAs (3% [95% CI 2.8 to 3.2] for both groups). It was lower (2% [95% CI 1.8 to 2.3]) for a first hip of a staged bilateral THA. For the secondary outcome, the relative 5-year survival differed for all groups. It was 105% (95% CI 104.9 to 105.9) for patients with unilateral THA, 107% (95% CI 106.3 to 107.4) for patients with a second hip from a staged bilateral THA and 109% (95% CI 108.8 to 109.5) for patients with a first hip of a staged bilateral THA. Patients with only a first hip of a planned staged bilateral THA who did not survive long enough to undergo their second THA were classified as unilaterals. The rank-order of survival curves are therefore by design ("immortal time bias"). We conclude, however, that survival for patients with unilateral THA more closely resembles the survival of patients with a second hip of a staged bilateral THA, compared with the first.

Conclusions: Our findings, which are based on observational register data, challenge the common practice in epidemiologic studies of analyzing only the first hip of a staged bilateral THA. We recommend analyzing the second THA in a patient who has undergone staged bilateral THA rather than the first because the second procedure better resembles unilateral THA.

Level of evidence: Level III, therapeutic study.

Fig. 1

This is a schematic representation of all THAs. The black and gray circles indicate observed and unobserved insertions. The blended black and gray circle indicates that the procedures contributed to categorization into study groups but were not themselves included. Primary THA before 1992 was used in a similar way (marked by triangle). Such cases were only recorded retrospectively in SHAR if the hip was reoperated after 1979.

Fig. 2

This graph shows the proportion of unilateral THAs, as well as the first and second hips of staged bilateral THAs performed from 1999 to 2015 (the period with comorbidity data available). Each group of THAs had stable proportions during different periods, leading to different inclusion periods for each group. The solid lines, over stable periods, indicate procedures included in the study.

Fig. 3

This flowchart shows the inclusion criteria and number of hips and patients. OA = osteoarthritis.

Fig. 4

This graph shows the age distribution among the three groups. Patients with unilateral THA had almost the same age distribution as patients with a second hip of a staged bilateral THA. Patients with staged bilateral THAs were younger by the time of their first insertion. (Absolute values on the Y-axis are not meaningful for distribution curves).

Fig. 5

This Kaplan-Meier graph shows the time to the first (if any) reoperation for each group, with 95% CIs.

Fig. 6

This curve shows the relative survival for each group with 95% CIs. Patients from each group were compared with the general Swedish population of the same age, sex and year of birth. The average relative survival for all patients is indicated by the dotted line. These curves should be interpreted carefully, since some patients are included with both their first and second hip of a staged bilateral THA (immortal time bias).