4.1 Introduction

The death rate from fracture of femur is apparently higher in West Midlands than in any other region. Among 65-84 year olds the age-standardised death rate from this cause is nearly three and a half times that of the lowest region (Yorkshire and Humberside) (Figure 4.1) and among those aged 85 and a similar difference is observed (Figure 4.2).

Figure 4.1: Regional comparison (Age 65-84) deaths, 2003-05

Regional comparison (Age 65-84) deaths, 2003-05

Figure 4.2: Regional comparison (Age 85+) deaths, 2003-05

Regional comparison (Age 85+) deaths, 2003-05

In the period 2003/05 fracture femur accounted for 148 deaths per year in the 65-84 age band in the West Midlands and 250 deaths per year in the 85 and over age band. Nearly all these fractures (90%) are fracture neck of femur (ICD10 S72.0), the next largest fraction being fracture of femur part unspecified (ICD10 S72.9) while the numbers of fractures of lower and mid shaft (ICD10 S72.3-S72.8) are negligible.

4.2 Anatomy and Clinical Presentation

The neck of femur may fracture in several places. The fracture may be intracapsular or extracapsular (trochanteric or sub trochanteric) (Figure 4.3). With intracapsular fracture the blood supply to the head of femur is impaired and there is a major risk of avascular necrosis and non union, while with extracapsular fractures the blood supply is more robust and with adequate fixation the fracture usually unites well ⁽¹⁾.

Figure 4.3: Anatomy of the hip joint

Anatomy of the hip joint

The fracture is most commonly the result of a fall or stumble. The fracture usually causes severe pain in the hip and inability to walk. Sometimes the person is unable to get up and has to lie where they have fallen until help comes.

4.3 Operative Management

In former days treatment was commonly by traction with prolonged bed rest and the death rate was very high. Nowadays treatment nearly always involves fixation of the fracture or prosthetic replacement of the femoral head and early mobilisation. The surgical procedures are Primary open reduction (W19, W20, W21) Closed reduction (W24, W25, W26) or prosthetic replacement (W37, W38, W39, W46, W47, W48).

4.4 General Epidemiology

Death rates from fracture neck of femur rise steeply with age and are slightly higher in females than males (Figure 4.4). Age standardised death rates from this cause (ICD10 S72.0 from 2001, ICD9 820 & 821 before) have risen slightly in the past 15 years (Figure 4.5).

Figure 4.4: Deaths by quinary age band, England and Wales 2005

Deaths by quinary age band, England and Wales 2005

Figure 4.5: Deaths trend (Age 65+), England and Wales 1991-2005

Deaths trend (Age 65+), England and Wales 1991-2005

Although one has to be careful of artefacts due to the coding change age specific rates (5 year age bands) for all ages above 74 appear to be rising. In the 1960s and 70s the hospital admission rate was noted to be rising more markedly  than could be accounted for by an ageing population ⁽²⁾ and this rise  continued through to the mid 90s ^{(3, 4)} but now appears to be stabilising ⁽⁵⁾. Mortality within one year of a fracture neck of femur is said to be between 20-35% ⁽⁶⁾ with 5-10% dying in the first month. In the presence of co-morbidities such as respiratory or heart disease mortality is higher ⁽⁷⁾.

Patients who have suffered one fracture neck of femur are at greater risk of suffering a second fracture ⁽⁸⁾. Factors increasing risk are osteoporosis, malnutrition, low body weight, factors that increase the risk of a fall and smoking ⁽⁹⁾. It has been suggested that vitamin D supplements would be protective but this remains to be proven ⁽¹⁰⁾. Hip protectors have been claimed to offer some protection against fracture but patient compliance is poor and more recent studies suggest they are not effective ⁽¹¹⁾.

So why are death rates for this fracture apparently higher in West Midlands than other regions? Regional differences are not accounted for by age differences since the figures are age standardised (but see comments on age standardisation later in this chapter).  Is the high mortality due to a high incidence of the fracture or a high fatality rate or some other cause?

4.5 A note on Age Standardisation

The average age of the age group 85 years and over varies considerably between regions. The percentage of those in this age band who are aged 90 years or over is lowest in the North East region (50.7%), slightly higher in the West Midlands region (52.2%) and highest in the South East region (58.7%). When event rates rise very steeply with age differences within age bands may become important. Since age standardisation has been done using quinary age bands with a top band of 85 years and more it will not have entirely removed the effect of age.

4.6 Hospital Admission Rates

Since nearly all fractures are admitted to hospital, first hospital admission rates are a fair indicator of incidence. As with mortality, hospital admission rates rise steeply with age and are higher in females than males (Figure 4.6).

Figure 4.6: England admissions by quinary age bands, 2004-05

England admissions by quinary age bands, 2004-05

Hospital admission rates in the West Midlands are slightly (but significantly) higher than the England average but the regional difference is much less marked than for mortality the difference in admission rate between highest and lowest region being only about 30% (Figure 4.7).  The rank order of regions for hospital admissions also differs considerably from that for mortality (Compare Figures 4.1 and 4.2 with 4.7). It thus seems unlikely that incidence of fracture neck of femur is unusually high in the West Midlands. A small number of people with fracture neck of femur die without being admitted to hospital but this is unlikely to affect the conclusion.

Figure 4.7: Regional comparison ages (Age 65+) admissions, 2004-05

Regional comparison ages (Age 65+) admissions, 2004-05

Admission rates have hardly changed in recent years (Figure 4.8). The apparent fall in 2001/02 is likely to be an artefact. Comparison of Figures 4.4 and 4.6 also reveals a much greater difference between males and females for admission rate than for mortality and also for most age sex bands the ratio of deaths to admissions is 0.1 or less, far less than the 0.2 - 0.35 that would be expected from the reported death rate after fracture neck of femur.

Figure 4.8: Trends in hospital admissions (Age 65+), 2000-01 to 2004-05

Trends in hospital admissions (Age 65+), 2000-01 to 2004-05

4.7 Outcomes and Quality of Care

If higher incidence of fracture neck of femur does not explain the higher death rate one has to look at outcomes of hospital care. Risk of death is increased in the months after suffering a fracture neck of femur. Comparisons of the indirectly standardised death rates for deaths within 30 days of emergency admission for fracture neck of femur show the West Midlands to have the second highest but not significantly different from that for England (Figure 4.9).

Figure 4.9: Regional comparison deaths within 30 days of admission, 2003-04

Regional comparison deaths within 30 days of admission, 2003-04

Figure 4.10: Regional comparison of timely surgery (within 48 hours), 2003-04

Regional comparison of timely surgery (within 48 hours), 2003-04

Speed of surgical intervention is probably associated with better outcome ⁽¹²⁾ The West Midlands has the second highest proportion of patients admitted with fracture neck of femur operated on within 48 hours and is significantly higher than the England average (Figure 4.10). The proportion in different regions operated on within first two days of admission ranges from 31% to 46%. Proportion of patients discharged to their usual residence is another indicator of quality of care for fracture neck of femur. For this indicator West Midlands has the lowest percentage discharged to usual residence with 41%, a figure significantly below the England average. However the range between regions (41% to 52%) is not large (Figure 4.11).

Figure 4.11: Regional comparison of discharged to usual residence, 2003-04

Regional comparison of discharged to usual residence, 2003-04

Figure 4.12: Regional comparison of operation type (Age 65+), 2005-06

Regional comparison of operation type (Age 65+), 2005-06

The operative management of patients by open reduction, closed reduction, prosthetic replacement or no operative procedure are shown in Figure 4.12. While the West Midlands does more primary open reduction than other regions (16%) the differences in treatment patterns are not striking. There is thus no reason to think that care of fractured neck of femur in the West Midlands has worse outcomes or is of poorer quality than in other regions.

4.8 Explaining the High Mortality

It appears that neither a higher incidence of fracture neck of femur nor a higher 30 day post-admission fatality rate can explain the high rate of deaths ascribed to fractured neck of femur in the West Midlands, nor the other variations in regional death rates for this condition. One is therefore left with the probability that the difference, large though it is, is due to differences in death certification practice. In 1993 Goldacre showed that in those who died within 4 weeks of admission for a fractured neck of femur the fracture was only mentioned on 25% of the death certificates ⁽¹³⁾ and in many of these it did not appear as underlying cause. Other authors also noted that recording of fracture neck of femur on death certificates was very unreliable (14, 15). It is possible that some certifying doctors are hesitant to mention fracture neck of femur on a death certificate not wishing to involve the relatives in a coroner’s enquiry ⁽¹⁶⁾. In order to pursue this investigation further it would be necessary to look at death rates at different time after fracture of neck of femur by linking hospital episode and death files. It is surprising that certification practice should differ so widely between regions but that appears to be the most likely explanation for the very high death rates of fracture neck of femur in the West Midlands.

The data specification for mortality from fracture neck of femur, which accompanies the Clinical and Health Outcomes Knowledge Base (NCHOD) from which most of the data in this chapter were taken warns “cause data for fracture should be used with great care because of the effects of artefactual local differences resulting from variations in certification procedures between coroners. Deaths caused by fracture femur are also under-recorded because there are a number of alternatives for classifying such deaths. For these reasons, variations between areas should be interpreted with caution”. This chapter demonstrates how true this is.

References:

1. Parker M, Johansen A. Hip fracture BMJ 2006; 333: 27-30
   
   
2. Lewis AF Fracture neck of femur: changing incidence. British Medical Journal 1981; 283, 1217-1219.
   
   
3. Evans JG, Seagroatt V, Goldacre MJ. Secular trends in proximal femoral fracture. Oxford record linkage study area and England 1968-86. Journal of Epidemiology and Community Health 1997; 51; 424-429.
   
   
4. Lyons AR Clinical outcomes and treatment of hip fractures. American Journal of Medicine 1997: 103 (2A), 51S-64S.
   
   
5. Balasegaram S, Majeed A, Fitz-Clarence H.  Trend in hospital admissions for fractures of the hip and femur in England 1989-1990 to 1997-98.  Journal of  Public Health Medicine 2001: 23, 11-17.
   
   
6. Goldacre MJ, Roberts SE, Yeates D. Mortality after admission to hospital with fractured neck of femur: database study. BMJ 2002: 325,868-869.
   
   
7. Roche JJW, Wenn RT, Sahota O, Moran CG. Effect of comorbidities and postoperative complications on mortality after hip fracture in elderly people: prospective observational cohort study.  BMJ 2005: 331, 1374-1376.
   
   
8. Parker MJ, Pryor GA, Anand JK, Lodwick R, Myles JW.  A comparison of presenting characteristics of patients with intracapsular and extrtacapsular proximal femoral fractures. Journal of Royal Society of Medicine 1992; 85, 152-155.
   
   
9. Law MR, Hackshaw AK.  A meta-analysis of smoking, bone mineral density and risk of hip fracture: recognition of a major effect.  BMJ 1997: 315, 841-84
   
   
10. Avenell A, Gillespie WJ, Gillespie JD, O’Connell DL. Vitamin D and Vitamin D analogues for preventing fractures associated with involutional and post menopausal osteoporosis.  Cochrane Database Systematic Review 2005(3) CD000227.
    
    
11. Parker MJ, Gillespie WJ, Gillespie LD. Effectiveness of hip protectors for preventing hip fractures in elderly people: systematic review. BMJ 2006: 206, 27-30.
    
    
12. Bottle A, Aylin P. Mortality associated with delay in operation after hip fracture: observational study. BMJ 2006: 332, 947-951.
    
    
13. Goldacre MJ.  Cause specific mortality: Understanding uncertain tips of the iceberg of disease. Journal of Epidemiology and Community Health 1993; 47: 491-496
    
    
14. Pemberton J.  Are hip fractures underestimated as a cause of death? The influence of coroners and pathologists on the death rate.  Community Medicine 1988: 10, 117-123.
    
    
15. Donaldson LJ, Parsons L, Cook AJ. Death certification in fractured neck of femur.  Public Health 1989: 103, 237-243.
    
    
16. Calder SJ, Anderson GH, Gregg PJ.  Certification of death in patients dying soon after proximal femoral fracture.  BMJ 1996: 312, 1515.