 |
Audit of morbidity and mortality following neck of femur
fracture using the POSSUM scoring system
William Young, Richard Seigne, Shona Bright, Marysha
Gardner
 |
Abstract
Aims The aims of
this study were to compare the morbidity and mortality data for patients
undergoing surgical fixation of a fractured neck of femur (during a 6-month
period) to the predicted morbidity and mortality rates obtained from the POSSUM
(Physiological and Operative Severity Score for the enUmeration of Morbidity and
Mortality) scoring system, adapted for orthopaedic patients. The predictive
accuracy of the orthopaedic POSSUM system is evaluated for this population. The
1-year mortality for the males and females of the study group (mean ages) is
compared to the 1-year mortality of male and female New Zealanders of the same
age.
Methods
Physiological and operative data was collected from patient notes; patient
morbidity and mortality were obtained at 30 days and at 1 year postoperatively.
The data were analysed with the orthopaedic POSSUM scoring system.
Results 225
complete datasets were obtained. The mean age of the patients was 83 years; 75%
were female. The observed 30-day morbidity and mortality rates were 58% and 12%
respectively. The observed 1-year mortality was 38% for males (mean age 79
years) and 29% for females (mean age 84 years). New Zealand census data predicts
7% and 6.4% mortality respectively based on these mean ages.
Conclusions The
POSSUM system allocates patients into groups of varying risk. The observed data
shows higher numbers of complications, including death, in patients allocated
into higher risk groups. The 1-year mortality is much higher than that predicted
based on mean patient age from the New Zealand abridged life table.
|
|
|
Patients admitted to hospital following fractured neck of
femur have a high postoperative mortality rate of between 20% and 35%1 after 1
year. The rate in New Zealand has been published at 24%.2 Walker found that
patients from the Christchurch area had a 1-year mortality rate 30% higher than
24% (i.e. 32%). Walker’s study incorporated all patients with hip
fractures over age 65 years and included those
not receiving surgical fixation.2 We excluded conservatively managed
patients, but did not exclude patients on the basis of age or mechanism of
injury.
There is scant data for morbidity rates following surgery
for fractured neck of femur in the literature; most studies document mortality
only, or study a specific complication.
The POSSUM system was developed to allow comparative audit
of outcome between different populations and subgroups of patients undergoing
surgical procedures. Originally described in general surgical patients,3 it has
recently been adapted for use in orthopaedic patients.4–6 It has also been
applied to other surgical specialties,7 with varying success and
adaptations.
The scoring system is based on a set of physiological and
operative variables (Table 1).4
|
Variable
|
Physiological
score
|
|
1
|
2
|
3
|
4
|
|
Age (years)
|
<60
|
61-70
|
>71
|
|
|
Cardiac signs
|
Normal
|
On cardiac drugs or
steroid
|
Oedema or warfarin
|
Raised JVP
|
|
Chest radiograph
|
Normal
|
|
Borderline
cardiomegaly
|
Cardiomegaly
|
|
Respiratory signs
|
Normal
|
SOB exertion
|
SOB stairs
|
SOB rest
|
|
Chest radiograph
|
Normal
|
Mild COAD
|
Moderate COAD
|
Any other change
|
|
Systolic BP
(mmHg)
|
110 to 130
|
131 to 170
100 to
109
|
>171
90 to 99
|
<89
|
|
Pulse (/min)
|
50 to 80
|
81 to 100
40 to
49
|
101 to 120
|
>121
<39
|
|
Coma score
|
15
|
12 to 14
|
9 to 11
|
<8
|
|
Blood urea
(mmol/L)
|
<7.5
|
7.6 to 10
|
10.1 to 15
|
>15.1
|
|
Blood Na
(mmol/L)
|
>136
|
131 to 135
|
126 to 130
|
<125
|
|
Blood K
(mmol/L)
|
3.5 to 5
|
3.2 to 3.4
5.1 to
5.3
|
2.9 to 3.1
5.4 to
5.9
|
<2.8
>6
|
|
Hb (g/100ml)
|
13 to 16
|
11.5 to 12.9
16.1 to
17
|
10 to 11.4
17.1 to
18
|
<9.9
>18.1
|
|
White cell count (x10[12]/L)
|
4 to 10
|
10.1 to 20
3.1 to
3.9
|
>20.1
<3
|
|
|
ECG
|
Normal
|
|
AF (60 to 90)
|
Any other change
|
|
Variable
|
Operative
severity score
|
|
1
|
2
|
3
|
4
|
|
Magnitude
|
Minor
|
Intermediate
|
Major
|
Major+
|
|
Number of operative variables within 30 days
|
1
|
|
2
|
>2
|
|
Blood loss per operation (ml)
|
<100
|
101 to 500
|
501 to 999
|
>1000
|
|
Contamination
|
None
|
Incised wound (i.e.
stab)
|
Minor contamination or
necrotic tissue
|
Gross contamination or
necrotic tissue
|
|
Presence of malignancy
|
None
|
Primary only
|
Node metastases
|
Distant metastases
|
|
Timing of operation
|
Elective
|
|
Emergency
Resuscitation
possible
<48 hours
|
Emergency
Immediate
<6
hours
|
Two equations are employed to convert the individual's
scores into predictions of morbidity and mortality for the 30-day postoperative
period. For comparison, observed morbidity and mortality should be collected for
the first 30 days postoperatively; morbidity is defined by a standardised list
of complications.3 The predicted morbidity and mortality allows stratification
of patients into groups with similar risk probability.
The POSSUM system was designed to provide a 30-day
postoperative prediction which could then be compared to observed rates taking
into account case mix. The ratio of observed to predicted morbidity and
mortality allows comparisons between individual surgeons or surgical units and
for comparison of rates over time.
Methods
Data was initially collected prospectively over 6
months, but was incomplete—largely due to the number of variables
requiring recording. Therefore a retrospective search for patients having
femoral fractures in the orthopaedic procedure database was completed. The notes
of patients undergoing surgical fixation within the 6-month period were then
reviewed. Only patients undergoing surgery for fractures of the femoral neck
were included.
The final 6 month population undergoing surgery was 249
patients. Of these, data was incomplete or absent for 24 patients (in most cases
due to unavailability of patient notes), thus resulting in a population of 225
patients. Data was collected from the operating theatre data set; the number of
patients receiving conservative management in this period is not known. A
separate audit in the hospital (within 1 year of this data collection period)
revealed a 2% preoperative mortality rate in this group of patients.
The operative and physiological variables required for
the orthopaedic adaptation of POSSUM3,4 were collected from patient notes.
Physiological variables were collected as close to time of surgery as possible.3
The scores were totalled and the original POSSUM equations were used to produce
individual morbidity and mortality scores.
Each of the physiological and operative variables are
scored on an exponential scale with three or four bands, with a minimum score of
one point and a maximum of eight points, depending on the degree of
physiological abnormality or surgical insult. A score of 1 (i.e. normal result)
was recorded for a variable if investigations were not performed or the results
were unavailable.
The patients’ general practitioners and rest
homes were individually contacted to collect a 30-day outcome for each patient.
Complications described in the Orthopaedic POSSUM scoring system were collected.
These morbid events range from minor to severe, and comprise both medical and
specific orthopaedic complications. Examples include postoperative haemorrhage,
infections at any site, cardiac events (including infarction or heart failure),
thromboembolic phenomena, and specific problems such as wound breakdown or
prosthetic dislocation.4 This provides comparative data to test the predictive
accuracy of the POSSUM system.
Further analysis was modelled on that detailed in
Copeland’s initial study,3 with generation of standard receiver operation
characteristic (ROC) curves.
The New Zealand Health Information Service provided
1-year mortality data for our dataset.8
Results
Our population thus comprised 225 patients, with a mean age
of 83 years and a median age of 87 years. Of these, 75% were female with a mean
age of 84 years; the mean male age was 79 years.
Of the 225 patients, 158 (70%) suffered an adverse event,
131 (58%) patients had recorded complications, and there were 27 (12%) deaths
within the 30 days of surgery. The 1-year mortality was 71 patients (32%).
Using the POSSUM equations, the predicted morbidity and
mortality was calculated for each patient. These figures were allocated into
risk bands for analysis, and receiver operation characteristic (ROC) curves were
compiled from the sensitivity and specificity of the comparison between observed
and predicted outcome in each risk band. This was modelled on the analysis
described in Copeland’s original paper.3
The predicted life expectancy for our population was derived
from data in the New Zealand 2000–2 abridged life tables.8 The mean age of
the female study population is 84 years. This equates to a continued life
expectancy of 7 years and a 7.2% chance of dying within 1 year. For the male
study population, mean age of 79 years, and the life expectancy is very similar
at almost 8 years (with a 6.4% chance of dying during the following year). Our
observed 1-year mortality rate was 32% across the whole population; male
patients had a 38% mortality rate and female patients had a 29% mortality rate
(Tables 2 and 3; Figures 1 and 2)
|
Predicted
risk of event % (i.e. risk band)
|
Observed
rate % [total number of patients in group]
|
|
>30
26–30
21–25
16–20
11–15
6–10
0–5
|
21.6
[51]
25.0
[16]
13.0
[23]
0.0
[25]
13.2
[38]
4.3
[47]
8.0
[25]
|
|
Predicted
risk of event % (i.e. risk band)
|
Observed
rate % [total number of patients in group]
|
|
>79
70–9
60–69
50–59
40–49
30–39
|
76.1
[67]
51.6
[31]
67.7
[31]
62.5
[24]
50.0
[26]
47.8
[23]
|
Discussion
The
POSSUM scoring system provides a means for calculating the 30-day mortality and
morbidity for a population of patients. The system was developed through
collection of extensive patient data pre- and intraoperatively, observing the
morbidity and mortality, and then using logistical regression analysis to
develop the POSSUM equations.3 It was designed as a retrospective audit tool,
not as a predictor of morbidity or mortality, and was designed to be applied to
groups of patients rather than individuals. The requirement for intraoperative
data renders it unsuitable for preoperatively predicting postoperative outcome.
The POSSUM scoring system does allow comparison between
cohorts of patients. Individual predictions of outcome following surgery for
neck of femur fracture can be made based on ASA status,9–11 type of
fracture,12,13 confusion status,10 or site of residence prior to injury.14
From our audit, the POSSUM scoring system provides a useful
assessment of both the morbidity and mortality, with patients allocated to
arbitrary ‘risk bands’. There is a degree of over-prediction of both
morbidity and mortality with this dataset. This is not unique to our
study.6,7,15 The exponential methods of analysis used in the P-POSSUM
equations16 have produced more accurate predictions of morbidity and
mortality6,7,15 than the linear methods used in POSSUM.3
The weighting of POSSUM scoring variables may also be a
factor. Most of our patients scored maximum points based on their age (greater
than 71 years), points from electrocardiograph (abnormalities other than rate
controlled atrial fibrillation score maximally), and the operative magnitude
scores all of these patients as a ‘major’ procedure. It is also
apparent that the accuracy of prediction could be greater with greater numbers
of patients within each risk band. The caveat to this is that if our risk bands
are widened to increase numbers in a group then predictive accuracy is
reduced.
In this group of patients, the sensitivity and specificity
of the scoring system is less than that in the original study,3 and the
orthopaedic application of the POSSUM system.4 However, a more recent study
applying the POSSUM system to patients with fractured neck of femur6 shows
results very similar to the present study, with very similar ROC curves.
However, this group6 did not collect physiological data at time of surgery (as
originally described3 and as in the present study), but at time of admission to
hospital. Significant changes in some of the physiological markers may occur
between the time of admission and operation, which makes direct comparison to
similar studies difficult.
The issues arising from sensitivity and
specificity are exacerbated by the
relatively small numbers of patients followed in the present study. Other
studies involving POSSUM have used 1372,3 2326,4 and 1164[6] patients. Those
using smaller numbers of patients or higher risk groups have often adjusted the
scoring system in various ways in order to improve the correlation between
observed and predicted results (e.g. P-POSSUM, V-POSSUM,
RAAA-POSSUM).16–19
The small numbers involved means that the accuracy of the
present study is limited, but highlights the difficulty involved in applying any
scoring or prediction system to a relatively small patient group. Unfortunately
in New Zealand this may limit the applicability of this scoring system to large
or multicentre audits, or require the use of extended durations of study to
generate sufficient patient numbers.
Our population appears to be fairly typical of patients
suffering from fractures of the femoral neck, with a mean age of 83 years, and
75% female patients. When considering 1-year mortality in comparison with the
1-year mortality based on the mean ages by gender, a marked difference is seen.
The male patients’ 1-year mortality is predicted at 6.4%;8 but following
this surgery, the 1-year mortality is 38%, an almost six-fold increase.
Female patients fare slightly better, with a 1-year
mortality predicted at 7%8 but observed is still over four times greater at 29%.
These observed figures are similar to others from the literature.20–22
This study confirms that the sequelae from what often begins with a simple fall
persist far beyond a single hospital admission and a journey through the
operating theatres.
Neck of femur fracture is a life-changing event for elderly
patients. In many it becomes a life-ending event.
Author information:
William Young, Anaesthesia Registrar, Department of Anaesthesia,
Christchurch Hospital, Christchurch; Richard D Seigne, Specialist Anaesthetist,
Department of Anaesthesia, Christchurch Hospital, Christchurch; Shona Bright,
Emergency Registrar. Armadale Kelmscott Memorial Hospital, Armadale, WA,
Australia; Marysha Gardner, Emergency Registrar, Department of Emergency
Medicine, Christchurch Hospital, Christchurch
Correspondence: Dr
William Young, Department of Anaesthesia, Christchurch Hospital, Private Bag
4710, Christchurch, Fax: (03) 364 0289; email: wyoung@doctors.org.uk
- Goldacre
MJ, Roberts SE, Yeates D. Mortality after admission to hospital with fractured
neck of femur: database study. BMJ. 2002;325:868–9.
- Walker
N, Norton R, Vander Hoorn S, et al. Mortality after hip fracture: regional
variations in New Zealand. N Z Med J. 1999;112:269–71.
- Copeland
GP, Jones D, Walters M. POSSUM: a scoring system for surgical audit. Br J Surg.
1991;78:355–60.
- Mohamed
K, Copeland GP, Boot DA, et al. An assessment of the POSSUM system in
orthopaedic surgery. J Bone Joint Surg Br. 2002;84:735–9.
- Copeland
GP. The POSSUM system of surgical audit. Arch Surg. 2002;137:15–19.
- Ramanathan
TS, Moppett IK, Wenn R, Moran CG. POSSUM scoring for patients with fractured
neck of femur. Br J Anaesth. 2005;94:430–3.
- Neary
WD, Heather BP, Earnshaw JJ. The Physiological and Operative Severity Score for
the enUmeration of Mortality and morbidity (POSSUM). Br J Surg.
2003;90:157–65.
- Statistics
NZ. New Zealand Life Tables 2000-2. Available online. URL: http://www.stats.govt.nz/datasets/population/life-tables
Accessed May 2006.
- Davis
FM, Woolner DF, Frampton C, et al. Prospective, multi-centre trial of mortality
following general or spinal anaesthesia for hip fracture surgery in the elderly.
Br J Anaesth. 1987;59:1080–8.
- Clague
JE, Craddock E, Andrew G, et al. Predictors of outcome following hip fracture.
Admission time predicts length of stay and in-hospital mortality. Injury.
2002;33:1–6.
- Michel
JP, Klopfenstein C, Hoffmeyer P, et al. Hip fracture surgery: is the
pre-operative American Society of Anesthesiologists (ASA) score a predictor of
functional outcome? Aging Clin Exp Res. 2002;14:389–94.
- Fox
KM, Magaziner J, Hebel JR, et al. Intertrochanteric versus femoral neck hip
fractures: differential characteristics, treatment, and sequelae. J Gerontol A
Biol Sci Med Sci. 1999;54:M635–40.
- Su
H, Aharonoff GB, Hiebert R, et al. In-hospital mortality after femoral neck
fracture: do internal fixation and hemiarthroplasty differ? Am J Orthop.
2003;32:151–5.
- Bhandari
M, Koo H, Saunders L, et al. Predictors of in-hospital mortality following
operative management of hip fractures. Int J Surg Investig.
1999;1:319–26.
- Mohil
RS, Bhatnaga D, Bahadur L, et al. POSSUM and P-POSSUM for risk-adjusted audit of
patients undergoing emergency laparotomy. Br J Surg. 2004;91:500–3.
- Whiteley
MS, Prytherch DR, Higgins B, et al. An evaluation of the POSSUM surgical scoring
system. Br J Surg. 1996;83:812–5.
- Prytherch
DR, Whiteley MS, Higgins B, et al. POSSUM and Portsmouth POSSUM for predicting
mortality. Br J Surg. 1998;85:1217–20.
- Prytherch
DR, Ridler BM, Beard JD, Earnshaw JJ. A model for national outcome audit in
vascular surgery. Eur J Vasc Endovasc Surg. 2001;21:477–83.
- Prytherch
DR, Sutton GL, Boyle JR. Portsmouth POSSUM models for abdominal aortic aneurysm
surgery. Br J Surg. 2001;88:958–63.
- Davidson
TI, Bodey WN. Factors influencing survival following fractures of the upper end
of the femur. Injury. 1986;17:12–14.
- Schroder
HM, Erlandsen M. Age and sex as determinants of mortality after hip fracture:
3,895 patients followed for 2.5-18.5 years. J Orthop Trauma.
1993;7:525–31.
- Sutcliffe
AJ, Parker M. Mortality after spinal and general anaesthesia for surgical
fixation of hip fractures. Anaesthesia.
1994;49:237–40.
|
|
