Elsevier

The Lancet

Volume 382, Issue 9898, 28 September–4 October 2013, Pages 1097-1104
The Lancet

Articles
90-day mortality after 409 096 total hip replacements for osteoarthritis, from the National Joint Registry for England and Wales: a retrospective analysis

https://doi.org/10.1016/S0140-6736(13)61749-3Get rights and content

Summary

Background

Death within 90 days after total hip replacement is rare but might be avoidable dependent on patient and treatment factors. We assessed whether a secular decrease in death caused by hip replacement has occurred in England and Wales and whether modifiable perioperative factors exist that could reduce deaths.

Methods

We took data about hip replacements done in England and Wales between April, 2003, and December, 2011, from the National Joint Registry for England and Wales. Patient identifiers were used to link these data to the national mortality database and the Hospital Episode Statistics database to obtain details of death, sociodemographics, and comorbidity. We assessed mortality within 90 days of operation by Kaplan-Meier analysis and assessed the role of patient and treatment factors by Cox proportional hazards model.

Findings

409 096 primary hip replacements were done to treat osteoarthritis. 1743 patients died within 90 days of surgery during 8 years, with a substantial secular decrease in mortality, from 0·56% in 2003 to 0·29% in 2011, even after adjustment for age, sex, and comorbidity. Several modifiable clinical factors were associated with decreased mortality according to an adjusted model: posterior surgical approach (hazard ratio [HR] 0·82, 95% CI 0·73–0·92; p=0·001), mechanical thromboprophylaxis (0·85, 0·74–0·99; p=0·036), chemical thromboprophylaxis with heparin with or without aspirin (0·79, 0·66–0·93; p=0·005), and spinal versus general anaesthetic (0·85, 0·74–0·97; p=0·019). Type of prosthesis was unrelated to mortality. Being overweight was associated with lower mortality (0·76, 0·62–0·92; p=0·006).

Interpretation

Postoperative mortality after hip joint replacement has fallen substantially. Widespread adoption of four simple clinical management strategies (posterior surgical approach, mechanical and chemical prophylaxis, and spinal anaesthesia) could, if causally related, reduce mortality further.

Funding

National Joint Registry for England and Wales.

Introduction

Although death as a result of total hip joint replacement is rare, the risk still needs to be quantified and minimised through reduction of risk factors. The UK National Institute for Health and Care Excellence has recommended several measures to reduce postoperative mortality,1 including use of mechanical and chemical thromboprophylaxis. However, obtaining good evidence about the effectiveness of preventive measures is difficult because mortality is low.

The National Joint Registry for England and Wales, established in April, 2003, records all total hip replacements done in England and Wales, the Office for National Statistics records all deaths in England and Wales, and the Hospital Episode Statistics records all inpatient recipients of NHS-funded care in England. Combination of these datasets provides an opportunity to assess many aspects of care in total hip replacement and their association with mortality. Additionally, the effect of patient factors on mortality can be studied.

We analysed postoperative mortality after hip replacement for osteoarthritis with the aims of: estimating the risk of mortality after hip replacement, assessing whether mortality after hip replacement has decreased with time, and identifying which treatment factors are associated with reduced mortality after accounting for patient factors.

Section snippets

Data sources

In this retrospective observational study we analysed data from the National Joint Registry. Details from the National Joint Registry of patients who had had primary total hip replacement were passed to the NHS Personal Demographics Service, which provided dates of death from the Office for National Statistics if the NHS number was traceable. National Joint Registry data were also linked to inpatient and day case records from Hospital Episode Statistics. We produce part 3 of the annual report

Results

Over the 8 years of follow-up available for the whole cohort, the hazard rate increased with time from operation (data not shown), with steeper slopes for men and older age groups, as expected for normal ageing, especially in view of the ages of patients at primary operation (mean 68·4, SD 10·6 years). 1743 patients died in the first 90 days after surgery. The hazard rate within the first 90 days (figure 1) suggests a short-term peak risk of death in the perioperative period that then subsided.

Discussion

We have shown a large fall in early mortality after primary hip joint replacement for osteoarthritis between 2003, and 2011, which was unexpected. If deaths in the earlier period were systematically under-reported, this trend could be an artifact. However, we believe that this scenario is unlikely because most deaths occurred after registration; thus, any under-reporting should have been random and not related to future mortality. Additionally, we report a steady year-on-year fall, rather than

References (24)

  • AJ Pugely et al.

    Differences in short-term complications between spinal and general anesthesia for primary total knee arthroplasty

    J Bone Joint Surg Am

    (2013)
  • Cardiff and Vale NHS weight loss plan for operations

  • Cited by (236)

    View all citing articles on Scopus
    View full text