Osteoporos Int (2010) 21:18351839

DOI 10.1007/s00198-009-1154-0

ORIGINAL ARTICLE

Excess mortality after pelvic fractures in institutionalized

older people
K. Rapp & I. D. Cameron & S. Kurrle & J. Klenk &
A. Kleiner & S. Heinrich & H.-H. Knig & C. Becker

Received: 19 October 2009 / Accepted: 3 December 2009 / Published online: 8 January 2010
# International Osteoporosis Foundation and National Osteoporosis Foundation 2009

Abstract
Summary Mortality after pelvic fracture was calculated in
residents of nursing homes. Compared with a matched
comparison nonfracture group, excess mortality was found
during the first 2 months after pelvic fracture.
Introduction Low energy pelvic fractures are mainly
observed in people of advanced older age. The incidence
of these fractures has increased considerably during the last
decades. Information about excess mortality after pelvic
fractures in older people is not available.
Methods To calculate excess mortality, a retrospective
cohort study was conducted. Data from residents institutionalized in Bavarian nursing homes between 2001 and
K. Rapp : J. Klenk : C. Becker
Department of Clinical Gerontology, Robert-Bosch-Hospital,
Auerbachstr. 110,
70376 Stuttgart, Germany
K. Rapp (*) : J. Klenk : A. Kleiner
Institute of Epidemiology, Ulm University,
Helmholtzstr. 22,
89081 Ulm, Germany
e-mail: kilian.rapp@rbk.de
I. D. Cameron
Rehabilitation Studies Unit, Sydney Medical School,
University of Sydney,
Ryde, NSW, Australia
S. Kurrle
Division of Rehabilitation and Aged Care,
Hornsby Ku-ring-gai Hospital,
Hornsby, NSW, Australia
S. Heinrich : H.-H. Knig
Department of Psychiatry, Health Economics Research Unit,
University of Leipzig,
Liebigstrae 26,
04103 Leipzig, Germany

2006 were used. For each patient with a pelvic fracture (n=
1,154), five residents without pelvic fracture (n=5,770)
were matched by sex, age, date of admission to the nursing
home, and level of care (measure for the need of care).
Hazard regression models were applied.
Results An excess mortality was found during the first
months after pelvic fracture. In women, the increased
mortality risk was limited to the first (hazard rate ratio
(HR) 1.83, 95% confidence interval (CI) 1.422.37) and
second (HR 1.52, 95% CI 1.132.04) months after the
injury. In men, excess mortality was more pronounced (HR
2.95, 95% CI 1.575.54 for the first month) and appeared
to last longer than in women. The majority of deaths due to
pelvic fractures in the first 2 months after injury occurred
following discharge from the hospital to the nursing home.
Conclusion Pelvic fractures are associated with an increased
mortality. These results should encourage the development of
preventive measures to reduce this excess mortality.
Keywords Excess mortality . Nursing home . Pelvic
fractures

Introduction
Pelvic fractures can be the result of high energy trauma,
such as a motor traffic crash or low energy trauma, such as
a fall. While pelvic fractures due to high energy trauma are
observed typically in young or middle-aged men, pelvic
fractures due to low energy trauma occur predominantly in
old and very old women [1, 2]. Therefore, pelvic fractures
represent injuries which are very heterogeneous regarding
cause, age, and sex.
Very limited data about the epidemiology of pelvic
fractures are available. There is no question that the number

1836

of pelvic fractures due to low energy trauma in industrialized countries has increased considerably during the last
decades [3-6]. Most of the change can be explained by the
demographic transition to an older population. Studies in
the Finnish population have shown that not only the
absolute number but also the age-adjusted incidence of
pelvic fractures in elderly people has increased steadily [4,
5]. Therefore, pelvic fractures cause an increasing burden of
disease in the aging societies of industrialized countries.
Low energy pelvic fractures are usually stable fractures
and do not require surgical treatment. Their prognosis has
been assumed to be excellent. There are, however, a few
studies which report considerable 1-year mortality after
pelvic fractures which ranges from 12% to 33% [7-9]. The
baseline mortality in this age group, however, is high, and a
large part of the fracturemortality association may be
related to underlying comorbidities. Therefore, an appropriate comparison group is needed in order to analyze
fracture-related excess mortality.
Residents of nursing homes commonly have risk factors
for pelvic fractures including advanced age, female gender,
osteoporosis, and falls. Therefore, they are an important
population with reference to pelvic fractures.
We analyzed excess mortality in residents from nursing
homes after pelvic fractures. To the best of our knowledge,
it is the first study which used a comparison group and also
the first study which analyzed excess mortality in people
living in long-term care.

Materials and methods

The study dataset comprised 134,353 residents aged
65 years and older, insured through the Allgemeine
Ortskrankenkasse (AOK), and resident in a nursing home
in Bavaria between 1 January 2001 and 31 December 2006.
The AOK is Germany's largest statutory health insurance
company and covers about 50% of all residents living in
nursing homes in Bavaria, a federal state with 12.5 million
inhabitants in the south of Germany. During the observation
period, 1,048 pelvic fractures were registered in women and
106 in men.
To calculate excess mortality, a retrospective cohort
study using an exposed cohort (those who had experienced
pelvic fracture) and an unexposed cohort (those who had
not experienced pelvic fracture) was constructed. Since
personal characteristics may be associated both with
fracture risk and mortality, a matching procedure was used
in order to make both groups as comparable as possible.
Residents in the comparison group had to be alive and
institutionalized at the date of the matched patient's pelvic
fracture. They were matched by sex, age (in 5-year
intervals), level of care (see below), and date of admission

Osteoporos Int (2010) 21:18351839

to the nursing home. To each patient with a pelvic fracture

(n=1,154), five residents without pelvic fracture (n=5,770)
were matched. Thus, the final dataset comprised 6,924
residents.
Routine data collection systems of the health insurance
company were utilized to obtain data on sex, age, date of
admission to the home, level of care (see below), and if
appropriate, pelvic fractures, femoral fractures, and date of
death for each individual. Hospital discharge diagnoses
were used to identify pelvic fractures (International Classification of Diseases-10 (ICD-10): S32 except lumbar spine)
and femoral fractures (ICD-10: S72). The localization of
the pelvic fracture was specified in 730 of 1,154 cases. Of
these specified cases, 50.8%, 20.0%, 21.10%, 3.6%, 3.4%,
and 1.1% occurred in the pubic bone, ischium, acetabulum,
iliac bone, sacrum, and coccyx, respectively. Data were
only available for fractures which resulted in a hospital
admission.
In 1995, long-term care insurance which is compulsory
for all citizens was introduced in the German social
insurance system [10, 11]. Depending on the amount of
care required, recipients are categorized into one of three
levels after an assessment by a physician (level I, II, and III
requiring basic care such as washing, feeding, or dressing
for at least 0.75, 2, and 4 h per day, respectively). The level
of care is therefore a proxy measure for the extent of
functional impairment.
Survival time for patients and their matched residents
started at the date of the injury (admission date to the
hospital). Person-months were accumulated between date of
injury and date of death or end of the study (31 December
2006). Mortality rates were calculated by dividing the
number of deaths by the total number of person-months.
Since survival curves did not meet the proportional hazard
assumption, proportional hazard regression models and
calculations of absolute mortalities were performed stratified by six time intervals (0 to 1 month, >1 to 2 months, >2
to 3 months, >3 to 6 months, >6 months to 1 year, >1 year).
In order to compare mortality following pelvic fracture
with mortality after femoral fracture, mortality rates for
femoral fractures (8,107 in women, 1,295 in men) were
calculated in the same way as for pelvic fractures. For this
analysis, the same study dataset of 134,353 residents was
used. Analyses were performed separately for women and
men.

Results
The study cohort consisted of the group with pelvic
fractures (1,048 women, 106 men) and the comparison
group (5,240 women, 530 men). The median age was
88.0 years in women and 86.6 years in men. More than

Osteoporos Int (2010) 21:18351839

1837

Women

54% of the study population died during the follow-up

observation period. The survival curves demonstrate excess
mortality of residents with pelvic fractures particularly
during the first months after the injury (Fig. 1).
In women, the increased mortality risk was limited to the
first (hazard rate ratio (HR) 1.83, 95% confidence interval
(CI) 1.422.37) and second (HR 1.52, 95% CI 1.132.04)
months after the injury (Table 1). From the third month
onward, risk was not significantly increased, and survival
curves converged after about 15 months (Fig. 1).
In men, excess mortality after pelvic fracture was more
pronounced and lasted longer than in women. Mortality risk
was increased threefold and twofold during the first and
second month after the injury, respectively (Table 1). There
was also an increased relative risk in the period >6 months
to 1 year (HR 1.86 95% CI 1.033.38). This, however, was
more due to a low risk in the comparison group than to a
high risk in the male fracture group.
Mortality rates in the first and second months after pelvic
fracture fell between the mortality rates after femoral
fracture and the rates of the comparison nonfracture group
for the same time period (12.7, 6.8 femur versus 7.9, 6.3
pelvis versus 4.3, 4.1 nonfracture and 22.4, 10.1 femur
versus 15.5, 6.8 pelvis versus 5.2, 3.1 nonfracture per 100
person-months in women and men, respectively).
In women and men combined, deaths in the first
2 months after pelvic fracture occurred in 26% of subjects
while they were hospital inpatients with the remainder
occurring after discharge back to the nursing home.

1,0
Female controls
Women with pelvic fracture

Survival rate

0,8

0,6

0,4

0,2

0,0
0,0

0,5

1,0

1,5

2,0

Years
Men
1,0
Male controls
Men with pelvic fracture

Survival rate

0,8

0,6

0,4

0,2

0,0
0,0

0,5

1,0

1,5

2,0

Years

Fig. 1 Mortality in female and male residents of nursing homes with

pelvic fractures and in their comparison nonfracture groups (matched
by sex, age, level of care, and date of admission to the nursing home)

Discussion
We found an excess mortality during the first months after
pelvic fracture. This has not previously been detected in

Table 1 Mortality rates and risk of mortality in the group with pelvic fractures (n=1,154) and in the comparison nonfracture group (n=5,770) in
different time periods after the injury in female and male residents of nursing homes in Bavaria between 2001 and 2006
Women

Men

Mortality/100 person-months

0 to 1 month
>1 to 2 months
>2 to 3 months
>3 to 6 months
>6 months to 1 year
>1 year

Fracture cases

Comparison groupa

7.9
6.3
3.5
3.5
2.7
2.6

4.3
4.1
3.9
3.5
3.0
2.5

HR (95% CI)b

1.83
1.52
0.91
1.01
0.90
1.03

(1.422.37)
(1.132.04)
(0.621.34)
(0.791.28)
(0.731.12)
(0.901.18)

Matched for sex, age, level of care, and date of admission to the nursing home

Hazard rate ratio (95% confidence interval)

Mortality/100 person-months
Fracture cases

Comparison groupa

15.5
6.8
6.1
5.3
4.6
3.9

5.2
3.1
4.8
3.7
2.5
3.4

HR (95% CI)b

2.95
2.22
1.29
1.41
1.86
1.13

(1.575.54)
(0.865.71)
(0.493.41)
(0.732.73)
(1.033.38)
(0.721.75)

1838

community living older people or in institutionalized older

people. In hip fractures, a similar pattern with an increased
risk particularly during the first months after the injury has
been observed in different studies [12-17]. This is not
unexpected since nearly all hip fractures are treated by a
surgical procedure which is associated with an intra- and
postoperative mortality risk. Surgery after a low energy
pelvic fracture, however, is very unusual. Therefore, the
excess mortality after pelvic fracture must be attributed to
other causes.
Clinicians know that pelvic fractures are usually very
painful particularly when mobilizing or sitting, and strong
analgesic therapy is often required. Mobilization of these
patients can be difficult. From our data, only 3.5% of all
nursing home residents with a pelvic fracture received
treatment in a rehabilitation unit following the fracture. All
other patients were discharged directly back to their nursing
homes from hospital. In nursing homes, the continuation of
thromboprophylaxis is unusual, and limited physiotherapy
is available. Therefore, we speculate that the expected
complications of immobilization, particularly pulmonary
embolism or infections (pneumonia or urinary sepsis),
contributed to the observed excess mortality. This theory
has limited support from a study from a London hospital
which found in postmortem examinations of patients with
low energy pelvic fractures that five out of seven deaths
were caused by massive pulmonary embolism. The other
two deaths were due to pneumonia and congestive heart
failure, and all deaths occurred within a range of 925 days
after admission [7].
We have no definite information about the causes of the
observed excess mortality and can, therefore, give no
definite recommendations. However, the results are so
important that they should provoke a discussion about
treatments with potential to reduce excess mortality after
pelvic fractures. About three quarters of the deaths within
the first 2 months occurred after discharge back to the
nursing homes. Treatments could, therefore, include continuation of low dose anticoagulation for 2 months following the fracture, or transfer to a rehabilitation unit for
ongoing therapy. If this is not possible due to dementia or
other reasons, an intensive mobilization program could be
provided in the nursing home to which the resident is
discharged. Furthermore, close cooperation between the
resident's physician, nurses, and physiotherapist is necessary in order to optimize analgesic therapy and allow early
mobilization.
Excess mortality after pelvic fracture was higher in men
than in women. This is a phenomenon which was also
reported consistently after hip fracture [12, 14, 15, 17, 18].
The reasons are incompletely understood. Poorer health in
men at the time of the injury or a higher susceptibility to
specific causes of death like coronary heart disease or

Osteoporos Int (2010) 21:18351839

infections have been discussed but the evidence is

contradictory [18-22]. In our data, the distribution of the
level of care at the time of the injury was very similar
between women and men and cannot explain the sex
difference in the excess mortality. Thus, men seem to be
more vulnerable to the injury and to its complications than
women.
Major strengths of the study are its large number of
study participants with pelvic fractures and the accurate
ascertainment of the date of the pelvic fracture and the date
of death with appropriate matching of controls. A common
problem in studies in this population has been to
disentangle two of the components of excess mortality:
one due to the fracture itself and one due to morbidity
associated both with an increased risk of falls and fractures
and an elevated baseline risk of mortality [23]. In our
study, the matching procedure included a measure for
functional status (level of care), resulting in a more robust
similarity between residents with and without pelvic
fracture.
The study has also some limitations. Only fractures
which led to a hospital admission were included, and it is
not known what percentage of pelvic fractures is treated
without hospital admission. Our impression is that this is
not common but the consequence would be an underestimation of the fracture rate. There is, however, no reason
why the selection of fracture patients should have influenced the pattern of excess mortality. In addition, only the
date of death but not the cause of death is available. We
also have no information about the exact cause of the
fracture or its subsequent treatment. In our experience,
pelvic fractures in residents of nursing homes are nearly
always caused by falls, and are treated conservatively.
Thus, they seem to represent, therefore, a good model for
low energy pelvic fractures.
Our data are derived from institutionalized older people
and cannot be generalized to community-living older
people. While fracture-related problems, such as pain and
immobilization, are present also for community-living older
people, this population is likely to be less frail, and these
factors may not influence mortality rates after pelvic
fractures to the same extent.
In summary, we found an excess mortality in residents of
nursing homes following pelvic fractures particularly
during the first months after the injury. The burden and
the consequences of pelvic fractures in old and very old
people probably have been underestimated.
Acknowledgments We thank Regina Merk-Buml, Ralf Brum,
Johannes Laws-Hofmann, Gerhard Dahlhoff, Otto Gieseke, Michaela
Heil, Stefanie Drfler, and Markus Gindl from the Allgemeine
Ortskrankenkasse (AOK) and Ulrich Rissmann from the Robert-Bosch
Hospital, Stuttgart for the admission to the data and for the support of our
analyses.

Osteoporos Int (2010) 21:18351839

Conflicts of interest None.

Funding The analysis was supported by a grant of the Forschungskolleg Geriatrie of the Robert Bosch Foundation and by the
Bundesministerium fr Bildung und Forschung (Frderkennzeichen:
01EL0702, 01EL0717, 01EL0718).

References
1. Ragnarsson B, Jacobsson B (1992) Epidemiology of pelvic
fractures in a Swedish county. Acta Orthop Scand 63(3):297300
2. Balogh Z, King KL, Mackay P, McDougall D, Mackenzie S,
Evans JA, Lyons T, Deane SA (2007) The epidemiology of pelvic
ring fractures: a population-based study. J Trauma 63(5):1066
1073
3. Parkkari J, Kannus P, Niemi S, Pasanen M, Jarvinen M, Luthje P,
Vuori I (1996) Secular trends in osteoporotic pelvic fractures in
Finland: number and incidence of fractures in 19701991 and
prediction for the future. Calcif Tissue Int 59(2):7983
4. Kannus P, Palvanen M, Niemi S, Parkkari J, Jarvinen M (2000)
Epidemiology of osteoporotic pelvic fractures in elderly people in
Finland: sharp increase in 19701997 and alarming projections for
the new millennium. Osteoporos Int 11(5):443448
5. Kannus P, Palvanen M, Parkkari J, Niemi S, Jarvinen M (2005)
Osteoporotic pelvic fractures in elderly women. Osteoporos Int 16
(10):13041305
6. Boufous S, Finch C, Lord S, Close J (2005) The increasing
burden of pelvic fractures in older people, New South Wales,
Australia. Injury 36(11):13231329
7. Spencer JD, Lalanadham T (1985) The mortality of patients with
minor fractures of the pelvis. Injury 16(5):321323
8. Morris RO, Sonibare A, Green DJ, Masud T (2000) Closed pelvic
fractures: characteristics and outcomes in older patients admitted
to medical and geriatric wards. Postgrad Med J 76(900):646
650
9. Leung WY, Ban CM, Lam JJ, Ip FK, Ko PS (2001) Prognosis of
acute pelvic fractures in elderly patients: retrospective study. Hong
Kong Med J 7(2):139145
10. Becker C, Leistner K, Nikolaus T (1998) Introducing a statutory
insurance system for long-term care (Pflegeversicherung) in
Germany. In: Michel JP, Rubenstein LZ, Vellas BJ, Albarede JL
(eds) Geriatric programs and departments around the world. SerdSpringer, Paris-New York, pp 5564

1839
11. Bundesministerium fr Gesundheit long-term care insurance.
Homepage (serial online) Available at: http://www.bmg.bund.de/
cln_040/nn_617014/EN/Long-term-care-insurance/long-termcare-insurance-node,param=.html__nnn=true
12. Jacobsen SJ, Goldberg J, Miles TP, Brody JA, Stiers W, Rimm
AA (1992) Race and sex differences in mortality following
fracture of the hip. Am J Public Health 82(8):11471150
13. Cooper C, Atkinson EJ, Jacobsen SJ, OFallon WM, Melton LJ III
(1993) Population-based study of survival after osteoporotic
fractures. Am J Epidemiol 137(9):10011005
14. Schroder HM, Erlandsen M (1993) Age and sex as determinants
of mortality after hip fracture: 3,895 patients followed for 2.5
18.5 years. J Orthop Trauma 7(6):525531
15. Forsen L, Sogaard AJ, Meyer HE, Edna T, Kopjar B (1999)
Survival after hip fracture: short- and long-term excess mortality
according to age and gender. Osteoporos Int 10(1):7378
16. Hannan EL, Magaziner J, Wang JJ, Eastwood EA, Silberzweig
SB, Gilbert M, Morrison RS, McLaughlin MA, Orosz GM, Siu
AL (2001) Mortality and locomotion 6 months after hospitalization for hip fracture: risk factors and risk-adjusted hospital
outcomes. JAMA 285(21):27362742
17. Rapp K, Becker C, Lamb SE, Icks A, Klenk J (2008) Hip
fractures in institutionalized elderly people: incidence rates and
excess mortality. J Bone Miner Res 23(11):18251831
18. Fransen M, Woodward M, Norton R, Robinson E, Butler M, Campbell
AJ (2002) Excess mortality or institutionalization after hip fracture:
men are at greater risk than women. J Am Geriatr Soc 50(4):685690
19. Holt G, Smith R, Duncan K, Hutchison JD, Gregori A (2008)
Gender differences in epidemiology and outcome after hip
fracture: evidence from the Scottish Hip Fracture Audit. J Bone
Jt Surg Br 90(4):480483
20. Wehren LE, Hawkes WG, Orwig DL, Hebel JR, Zimmerman SI,
Magaziner J (2003) Gender differences in mortality after hip fracture:
the role of infection. J Bone Miner Res 18(12):22312237
21. Piirtola M, Vahlberg T, Lopponen M, Raiha I, Isoaho R, Kivela
SL (2008) Fractures as predictors of excess mortality in the aged-a
population-based study with a 12-year follow-up. Eur J Epidemiol
23(11):747755
22. Cameron ID, Chen JS, March LM, Simpson JM, Cumming RG, Seibel
MJ, Sambrook PN (2009) Hip fracture causes excess mortality due to
cardiovascular and infectious disease in institutionalized older people:
a prospective five-year study. J Bone Miner Res. Oct 19. [Epub ahead
of print]
23. Browner WS, Pressman AR, Nevitt MC, Cummings SR (1996)
Mortality following fractures in older women. The study of
osteoporotic fractures. Arch Intern Med 156(14):15211525

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.