International and national predictions from the late 1990s warned of alarming increases in hip fracture incidence due to an ageing population globally. Our study aimed to describe contemporary, population-based longitudinal trends in outcomes and epidemiology of hip fracture patients in a tertiary referral trauma centre.

A retrospective review was performed of all patients aged 65 years and over with a diagnosis of fractured neck of femur (AO classification 31 group A and B) admitted to the John Hunter Hospital, Newcastle, New South Wales between 1st January 2002 and 30th December 2009. Datawas collated and cross referenced from several databases (Prospective Long Bone Fracture Database, Operating Theatre Database and the Hospital Coding Unit). Mortality data was obtained via linkage with the Cardiac and Stroke Outcomes Unit, Planning and Performance, Division of Population Health. Main outcome measures were 30-day mortality, in-hospital mortality, length of stay.

The JHH admitted (427 ± 20/year, range: 391–455) patients with hip fractures over the 9 year study period. The number of admissions per year increased over the study period (p = 0.002), with no change in the age-standardised incidence (p = 0.1). The average age (83.5 ± 0.2) and average percentage female (73.7%) did not change. There was an overall trend to decreased 30-day mortality from 12.4% in 2002 to 7% in 2009 (p = 0.05). The factors that were associated with increased mortality were age (p < 0.0001), male gender (p = 0.0004), time to operating theatre (p = 0.0428) and length of stay (p < 0.0001).

In accordance with national and international projections on increased incidence of geriatric hip fractures, the incidence of fractured neck of femur in our institution increased from 2002–2009, reflecting our ageing population. 30-day mortality improved and longer length of stay corresponded with increased 30-day mortality.

Geriatric hip fracture patients have a 14-fold higher 30-day mortality than their age matched peers. Up to 50% of these patients receive blood transfusion perioperatively. Both restrictive and liberal transfusion policies are controversial in this population. Aim: The longitudinal description of transfusion practice in geriatric hip fracture patients in a major trauma centre.

An 8-year (2002–2009) retrospective study was performed on patients over the age of 65 undergoing hip fracture fixation. Yearly transfusion rate; the influence of transfusion on 30-day, 90-day and 1-year mortality and length of stay (LOS) was investigated. On admission haemoglobin (Hb), pre-transfusion Hb and post-transfusion Hb and their effect on transfusion requirement and mortality was also reviewed. The yearly changes in on-admission and pre-transfusion Hb were also examined. The influence of comorbidities, timing, procedure performed and operation duration on transfusion requirement and mortality was also studied. From the 3412 patients, 35% (1195) received transfusion during their hospital stay. There was no change in age, gender and co-morbidities during the study. Thirty-day mortality improved from 12.4% in 2002 to 7% in 2009. The transfusion rate showed a gradual decrease from the highest of 48.3% (2003) to 22.9% (2009) (Pearson correlation - R2 = −0.707, p=0.05). There was no change during the study period in on-admission and pre-transfusion Hb. The mortality for non-transfused and transfused patients was [9.6% vs. 10.3 % (30-day)], [17.2% vs. 18.4%(90-day)] and [27% vs. 30.5%(1-year), p=0.031]. LOS was 11±9 for non-transfused patients and 13±10 (p<0.001) for transfused patients. Patients with more comorbidities experienced a higher transfusion rate, (0 – 31%, 1 – 38%, 2 – 46%, 3 – 57%), (Pearson Chi-squared, p<0.001). The need for transfusion by different procedures in decreasing order was 47.6% intramedullary device, 44.0% DHS, 25.2% cemented hemiarthroplasty, 23.6% Austin-Moore, and 5.5% cannulated screws. The length of the operation increases the chance of transfusion (<1hrs, – 33%, 1–2hrs – 35%, 2–3hrs – 41%, >3 hours – 65%), (Pearson Chi-squared, p=0.010). Preoperative waiting time had no influence on transfusion frequency (<24hrs – 36%, 24–48hrs – 34%, 48–96hrs – 36%, >96hrs – 33%), (Pearson Chi-squared, p=0.823).

The percentage of transfused geriatric hip fracture patients halved during the eight-year period without changes in demographics and co-morbidities. Perioperative transfusion of hip fracture patients is associated with higher 1-year mortality and increased LOS. A more restrictive transfusion practice has been safe and may be a factor in the improved 30-day mortality.