Orthopaedics & Traumatology: Surgery & Research (2012) 98, 715—719

Available online at

www.sciencedirect.com

REVIEW ARTICLE

Epidemiology of proximal humerus fractures

managed in a trauma center
A. Roux a,∗, L. Decroocq a, S. El Batti a, N. Bonnevialle b, G. Moineau c,
C. Trojani d, P. Boileau d, F. de Peretti a

a
Department of Trauma and Orthopaedic Surgery, Saint-Roch Hospital, Nice Teaching Hospital Center, 15, rue Pierre-Devoluy,
06000 Nice, France
b
Department of Orthopaedic and Trauma Surgery, Purpan Hospital, Toulouse Teaching Hospital Center, place du Dr-Baylac,
31059 Toulouse cedex 9, France
c
Pasteur Lanroze Orthopaedic Surgical Center, 18, rue Amiral-Desfossés, 29200 Brest, France
d
Department of Orthopaedic Surgery, l’Archet Hospital, Nice Teaching Hospital Center, 151, route Saint-Antoine-de-Ginestière,
06200 Nice, France

Accepted: 21 May 2012

KEYWORDS Summary Proximal humerus fractures (PHF) are osteoporotic fractures that affect women
Epidemiology; over 70 years of age. Like fractures of the femoral neck they have become a public health
Fracture; concern. As the population ages there is an increase in the number of people in poor general
Proximal humerus; condition with an increased risk of falls on fragile bones. The incidence of these fractures
Prevalence; has increased by 15% per year. All patients managed for PHF in our center in the past year
Osteoporosis were included in this prospective study (prospective cohort study; level 2). Three hundred and
twenty-five patients were included with 329 fractures. There was a ratio of two women to one
man. At the final follow-up 50 patients had died (15%) and 25 patients were lost to follow-up.
The mean age was 70 years old. There were two types of risk factors. The first was fragile bones,
and the second was patient specific risk of falls. The severity of the fracture increased with the
age of the population. In the study by Charles S. Neer in 1970, 85% of PHF were not or were
only slightly displaced, while this category percentage was only 42% in our study. Hospitalization
was necessary in 43% of the cases in our study. Surgical management was necessary in 21%. This
lack of relationship between the percentage of displaced fractures (58%) and the percentage
of surgically treated fractures is a sign of the difficulties of managing this population, which is
usually in poor general condition.
© 2012 Elsevier Masson SAS. All rights reserved.

∗ Corresponding author.
E-mail address: [email protected] (A. Roux).

1877-0568/$ – see front matter © 2012 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.otsr.2012.05.013
716 A. Roux et al.

Introduction

Proximal humerus fractures (PHF) are the seventh most fre-

quent fracture in adults, and the third in patients over 65
following wrist and femoral neck fractures. They represent
5.7% of diagnosed fractures. This is mainly an osteoporotic
fracture and its prevalence increases as one moves north in
Europe. There is a linear increase in the incidence of this
entity after the age of 40. Like fractures of the femoral
neck, PHF have become a public health issue (French Law
Figure 1 Number of fractures according to age group.
No. 2004-806, August 2004, on public health policy). The
ageing population means there is an increase in the number
of people poor general condition with a greater risk of falling patients were seen for postoperative follow-up in an out-
on weak bones. This regular increase in the prevalence of patient consultation at 3 weeks, 6 weeks and 3 months.
osteoporotic fractures results in higher medical costs (hos- We contacted all the patients by telephone at 10 months
pitalization, treatment, convalescence. . .) and can result in of follow-up. They responded to a medical questionnaire
loss of autonomy. The management of these fractures will concerning: risk factors (osteoporosis, history of falling, low
be a real challenge for future healthcare policies in the level of physical activity, hip fracture in the mother, trou-
upcoming years. Kannus et al. [1,2] studied PHF between ble walking, pain in the lower limbs, trouble seeing, trouble
1970 and 1998 in patients over the age of 60 admitted to hearing, alcoholism, tobacco use), comorbidities (diabetes,
hospitals in Finland. The number of patients went from 208 epilepsy, depression, dementia, Parkinson’s disease, oth-
fractures in 1970 to 1105 in 1998 or increased by 15% per ers), the causes of the fracture, medical history in the
year. In 28 years and if the ageing of the population is taken fractured shoulder, associated traumas, the patient’s notion
into account, this fracture has increased by 166% in women of his/her own general condition on a three point scale
and 250% in men. These fractures will probably become (good, average, poor) and the MOS SF12 quality of life score
more difficult to treat due to delayed union, an increase [7]. All of these data were noted on an excel table. This was
in the number of complications and in the rate of pseu- a prospective cohort study (level 2).
darthrosis. Palaven et al. [3] estimated that the number of
shoulder fractures would increase three fold in the next 30
years. Court-Brown and Caesar [4] talk about a revolution in
Results
the management of fractures because trauma centers were
created in the developed countries between 1970—1980 to Three hundred and twenty-five patients were included with
manage trauma secondary to high-energy traumas, which 329 fractures. The population included 224 women (69%)
mainly affect young men. At present, the prevalence of that including three with bilateral fractures and 101 men (31%)
type of trauma is decreasing while traumas on osteoporotic one with a bilateral fracture, for a ratio of two women/one
bone are increasing. The latter are managed differently man. PHF was associated with another fracture in 34 cases.
because the short and long-term aims are different. Lind PHF represented 0.4% of adult emergency room consulta-
T et al. [5] noted that in 730 fractures, 29% of the patients tions at the CHU Nice (70,000 consultations per year). On
needed to be hospitalized; 75% of these were over 60 and a national level, there are 15,500,000 visits to the emer-
only 21% underwent surgery, which represents 583 hospital gency room per year and an estimated 65,000 isolated PHF
days per year for a Danish city of 250,000 inhabitants. The per year in France. At the final follow-up, 50 patients had
aims of the present study were to define the epidemiology of died (15%) and 25 patients were lost to follow-up. Initial
this population presenting a PHF and evaluate the severity data included the entire study population, while at the final
of the fracture and its therapeutic management in relation follow-up questionnaires were sent to 250 patients with 253
to different subgroups in this population. fractures (185 women and 65 men). The mean age was 70
(16—97). The fractured shoulder was on the right side 156
times and on the left side 173 times. The fracture affects
Materials and methods the dominant side in 48% of cases. Analysis of the distribu-
tion of fractures throughout the year showed that most of
All the patients managed in the emergency unit of Nice these fractures occurred during the ‘‘cold’’ season with 60%
University Hospital (CHU Nice) between November 2009 between October and March. Figs. 1 and 2 show that there
and November 2010 for PHF were included in this prospec- is a peak in the prevalence of fractures in patients in their
tive study. The initial evaluation included double oblique eighties and this is only found in women. The causes of frac-
AP view X-rays of the shoulder and a Lamy view while ture are summarized in Table 1. In men, 55% of the fractures
MRI was requested for more complex fractures. These dif- were due to a simple fall and 45% to a high-energy kinetic
ferent imaging techniques were visualized digitally. Initial trauma. In women, the cause was a simple fall in 82% of
management of each patient was noted in the medical file the cases. The cause of fracture was a high energy trauma
(number of days of hospitalization, period of the trauma, in young patients and low energy trauma in older patients,
surgical procedures, type of immobilization). Each fracture which can be found as of the age of 60 because in 56% of
was classified twice by three senior surgeons at 3-month cases it was due to a fall from standing height and this per-
intervals based on the Neer classification [6]. The most centage regularly increased until it reached 100% at the age
frequent response was taken for each fracture. All of the of 100. The risk factors and comorbidities are summarized
Epidemiology of proximal humerus fractures managed in a trauma center 717

Figure 2 Age group in relation to gender.

Table 1 Number of patients according to the cause of the

fracture. Figure 3 Distribution of fractures according to the Neer clas-
sification.
Cause of fracture Total Men Women

Falling from standing height 189 37 152 score is 50 for each). There was a history of shoulder injury
Road accident 34 18 16 in 10 patients (4%), (three fractures, two dislocations, five
Violent fall 17 7 10 rotator cuff tears, one arthritis and two anti-inflammatory
Aggression 9 3 6 injections). Fig. 3 shows the distribution of the different
Suicide attempt 1 0 1 fractures according to the Neer classification. Forty-two
percent of the fractures were NEER type 1; these are con-
sidered to be slightly or not displaced. Fifty-nine percent
respectively in Tables 2 and 3. The patients felt that they of the fractures were displaced. The type of fracture var-
were in good general condition in 154 cases (62%), average ied depending upon the age group. Under the age of 40,
in 66 cases and poor in 30 cases. The objective measure- 41% were slightly or not displaced (Neer type 1), the rest
ment of the general condition of this population based on were two part fractures (Neer type 3, 4 or 6). Following this,
the MOS SF12 quality of life scale showed a mean 41 for the fracture displacement worsened. In the 41—70 year old age
physical score and 53 for the psychological score (the normal group, the percentage of Neer type 1 fractures was similar
(43%) but three or four part fractures are found and repre-
sent 16% of all fractures. In the 71—100 year old age group,
Table 2 Risk factors for fracture. the percentage of type 1 Neer fractures was only 33% and
the percentage of three or four part fractures was 19%; this
Risk factors for fracture Number of patients %
is also the only age group with six patients with articular
Osteoporosis 85 34 fractures of the humeral head. Fractures were managed by
History of falling 82 33 conservative treatment in 79% of the cases (256 patients)
Limited physical activity 75 30 and surgery in 21% (69 patients). Surgical treatment included
History of femoral neck 31 12 internal fixation in 57 cases (53 intramedullary nails and
fracture in mother four ORIF with screws in the greater/lesser tuberosity) and
Trouble walking 62 25 12 shoulder arthroplasties (four anatomical replacements
Pain in the lower limbs 47 19 and eight reverse total shoulder replacements). Hospital-
Trouble seeing 66 26 ization was necessary in 43% of the patients (139 patients).
Trouble hearing 41 16 Therapeutic management varied depending upon the type
Alcoholism 37 15 of fracture. Surgery was performed in three Neer type 1
Tobacco use 71 28 fractures. Two presented with a diaphyseal fracture of the
humerus associated with the PHF, while the third presented
with a bilateral fracture, and it was decided to stabilize both
Table 3 Comorbidities. shoulder fractures for the patient’s comfort. Neer type 3
fractures (surgical neck fracture) were treated surgically in
Comorbidities Number of patients % 39% of cases, Neer type 8 fractures (surgical neck fractures
Diabetes 26 10 and the greater tuberosity) in 41% of the cases and Neer type
Epilepsy 7 3 12 (four part fractures) in 45% of cases. The other types of
Depression 53 21 fractures were rarely operated on (Table 4). Thirty-seven
Dementia 26 10 percent of the patients (125) who presented with displaced
Parkinson’s disease 4 1.6 PHF were not operated on. Conservative treatment was indi-
Psychosis 8 3 cated because of the patient’s poor general condition and
Cardiopathy 21 8 the risk of anesthesia. This population was a mean 78 years
HTA 10 4 old, older than the overall population, and in poorer general
History of stroke 8 3 condition (22% had died at 10 months, 14% had dementia,
20% had comorbidities, 26% had difficulty walking). At the
718 A. Roux et al.

the general French population. In our study 21% of patients

Table 4 Percentage of patients who underwent surgery for
were depressive. Lind T et al. [5] found an increase in these
PHF in relation to the Neer classification of fracture.
fractures in the winter, which we also observed usually in the
Neer Number of shoulders Surgery % middle of the day, but cannot confirm statistically. Although
alcohol or tobacco consumption was not considered to be a
I 136 3 2 risk factor by Chu et al. [12] they were present in 15 and 28%
III 102 40 39 of our cases respectively. The patient’s general condition
IV 3 0 0 is an equally important risk factor because an autonomous
VI 33 4 12 person in good condition will be less apt to fall and his/her
VII 1 1 100 bones will be less fragile. In a 5-year prospective study (1027
VIII 27 11 45 PHF) Court-Brown et al. [8] found that this population was
IX 1 0 0 in poorer general condition than the population presenting
XII 20 9 45 with wrist fractures, but in better general condition than the
XV 6 2 33 population presenting with hip fractures. Only 10% were in
an institution when the fracture occurred. Measures taken
to limit these risk factors and the lifestyle in certain areas
final follow-up our questionnaire was filled out by 88 of these
of the world can reduce the prevalence of PHF. In a study
patients (27 had died, eight were lost to follow-up, two had
performed between 1970 and 1995 and published in 2008
severe dementia). From a subjective point of view 52% (46)
in a population that was 80 or older, Kannus et al. [15],
of these patients considered themselves to be in good gen-
showed that the number of fractures increased from 88 to
eral condition 36% (32) in average condition and 11% (10) in
304/per 100,000 women. However the number of fractures
poor condition. Objectively the mean SF 12 physical score
has stabilized since 1995 suggesting that this population is in
was 39 and the mean SF 12 psychological score was 53.
better health, that preventive measures have been taken,
The results of these two scores were similar to those in the
and that management of osteoporotic bone has improved.
overall study population.
Hagino et al. [10] found that the traditional Japanese way of
life decreased the risk of falling, and therefore reduced the
Discussion prevalence of these fractures. The prospective study EPIDOS
[16] in osteoporotic fractures of the shoulder showed that
Like the studies by Kannus et al. [1] and Court-Brown et al. a patient with osteoporosis with falling risk factors, has a
[4,8], this epidemiological study shows that most PHF are greater risk of fracture than a patient without osteoporosis
osteoporotic fractures in women over the age of 70. The or falling risk factors. On the other hand, s/he has the same
risk of fracture begins to increase linearly in women in their risks as a patient without osteoporosis with high falling risks.
fifties [9,10]. The prevalence of PHF increases as the pop- The severity of the fracture increases as the population
ulation ages. There are two main types of risk factors for ages. In his article in 1970, Charles S. Neer [17] reported
osteoporotic fractures, in particular for PHF. The first risk type 1 fractures (not or slightly displaced - less than one
is fragile bones and the second is the risk of falling. The centimeter of displacement and an angle of less than 45◦ )
more fragile the bones are the more severe the fracture in 85% of the population. On the other hand in a prospec-
is [8]. Nguyen et al. [11] found that osteoporotic verte- tive 5-year study (1027 FESH) in 2001, Court-Brown et al.
bral compression fractures with loss of height, which is [8] found Neer type 1 fractures in 49%, two part fractures in
one of the most typical signs of existing osteoporosis, were 37%, three part in 9% and four part in 4%. We identified Neer
a predictive factor for these fractures. Ethnicity is also a type I fractures in 42% of the patients in our study in 2010
risk factor of fragile bones [12,13]. The prospective EPOS in 329 fractures. The percentage of non-displaced fractures
[9] study showed that in osteoporotic fractures, PHF were has decreased by 50% in 40 years. This increase in the num-
more frequent in women in northern Europe with an inci- ber of displaced fractures suggests that there is theoretically
dence of 5.2/1000/year, while in the rest of Europe it was a greater need for surgical management and a risk of poorer
1.3—1.9/1000/year. In men the incidence was comparable functional outcome. However, compared to the results by
in the different European regions with figures between 0.4 Charles S. Neer [17] in which approximately 15% of FEUH
and 1.2/1000/year. One or several risk factors of falling were were considered to be displaced, and thus require surgical
found in more than half of our population: 33% had a history management, the percentage of surgically treated fractures
of falling and 30% had low level of physical activity which in our study (21%) was similar to that in the Neer study. This
could indicate worsening of the patient’s general condition lack of relationship between the percentage of displaced
which was identified as a risk factor for PHF by Kelsey et al. fractures (58%) and the percentage of fractures that were
[14]. Difficulty seeing was identified in 26% of patients. Dif- surgically treated (21%) is a sign of the difficulties of man-
ficulty walking and pain in the lower limbs was identified aging this population. The technical difficulties associated
respectively in 25 and 19% of the cases. Cardiovascular dis- with fixation material that is not always adapted to mediocre
eases were found in 16% of cases, but their frequency in bone quality and the general condition of patients who are
the general population makes it difficult to confirm that usually extremely elderly with multiple comorbidities, are
this increases the risk of PHF, although in the elderly this the main reasons for this low percentage of surgeries. At 10
is probably the cause of attacks, which result in falls. Dis- months of follow-up the level of satisfaction in this specific
eases, which increase the risk of PHF, are diabetes, epilepsy, population is similar to that in the overall study population,
depression and dementia. In our study, 10.4% of patients but no firm conclusions can be drawn due after such a short
were diabetics (1 and 2 combined) while this figure is 3.8% in amount of time.
Epidemiology of proximal humerus fractures managed in a trauma center 719

Conclusion [6] Neer 2nd CS. Displaced proximal humeral fractures.

I. Classification and evaluation. J Bone Joint Surg Am
1970;52(6):1077—89.
PHF is now usually an osteoporotic fracture in women over
[7] Ware JE, Kosinski M, Keller SD. A 12-Item Short-Form Health
the age of 70. It is frequent and its prevalence is increasing. Survey: construction of scales and preliminary tests of reliabil-
The ageing population is the cause, resulting in a popula- ity and validity. Med Care 1996;34(3):220—33.
tion that is in poor general condition with an increased risk [8] Court-Brown CM, Garg A, McQueen MM. The epidemiol-
of falling on increasingly fragile bone. PHF are displaced in ogy of proximal humeral fractures. Acta Orthop Scand
50% of the cases, requiring hospitalization in at least one 2001;72(4):365—71.
out of two cases while surgery was performed in one out [9] Kristiansen B, Barfod G, Bredesen J, Erin-Madsen J, Grum B,
of five. Measures must be taken in this growing population Horsnaes MW, et al. Epidemiology of proximal humeral frac-
to prevent the risk factors of PHF because management of tures. Acta Orthop Scand 1987;58(1):75—7.
[10] Hagino H, Yamamoto K, Ohshiro H, Nakamura T, Kishimoto H,
these fractures may become another source of dependency
Nose T. Changing incidence of hip, distal radius, and prox-
in the elderly population.
imal humerus fractures in Tottori Prefecture, Japan. Bone
1999;24(3):265—70.
Disclosure of interest [11] Nguyen TV, Center JR, Sambrook PN, Eisman JA. Risk fac-
tors for proximal humerus, forearm, and wrist fractures in
The authors declare that they have no conflicts of interest elderly men and women: the Dubbo Osteoporosis Epidemi-
ology Study. Sydney, Australia: Bone and Mineral Research
concerning this article.
Program, Garvan Institute of Medical Research, St. Vincent’s
Hospital.
References [12] Chu SP, Kelsey JL, Keegan TH, Sternfeld B, Prill M, Quesenberry
CP, et al. Risk factors for proximal humerus fracture. Am J
[1] Kannus P, Palvanen M, Niemi S, Parkkari J, Järvinen M, Vuori Epidemiol 2004;160(4):360—7.
I. Osteoporotic fractures of the proximal humerus in elderly [13] Chapurlat R. Épidémiologie des fractures de l’extrémité
Finnish persons: sharp increase in 1970—1998 and alarm- supérieure de l’humérus: l’épaule est une autre cible de
ing projections for the new millennium. Acta Orthop Scand l’ostéoporose. Lett Rhumatol 2007;333:15—7.
2000;71(5):465—70. [14] Kelsey JL, Browner WS, Seeley DG, Nevitt MC, Cummings SR.
[2] Kannus P, Palvanen M, Niemi S, Parkkari J, Järvinen M, Risk factors for fractures of the distal forearm and proximal
Vuori I. Increasing number and incidence of osteoporotic humerus. The Study of Osteoporotic Fractures Research Group.
fractures of the proximal humerus in elderly people. BMJ Am J Epidemiol 1992;135(5):477—89.
1996;313(7064):1051—2. [15] Kannus P, Palvanen M, Niemi S, Sievänen H, Parkkari J. Rate of
[3] Palvanen M, Kannus P, Niemi S, Parkkari J. Update in the epi- proximal humeral fractures in older Finnish women between
demiology of proximal humeral fractures. Clin Orthop Relat Res 1970 and 2007. Bone 2009;44(4):656—9.
2006;442:87—92. [16] Lee SH, Dargent-Molina P, Bréart G, EPIDOS Group. Risk factors
[4] Court-Brown CM, Caesar B. Epidemiology of adult fractures: a for fractures of the proximal humerus: results from the EPI-
review. Injury 2006;37(8):691—7. DOS prospective study. Épidemiologie de l’Ostéoporose Study.
[5] Lind T, Kroner K, Jensen J. The epidemiology of frac- J Bone Miner Res 2002;17(5):817—25.
tures of the proximal humerus. Arch Orthop Trauma Surg [17] Neer II CS. Displaced proximal humeral fractures: part 1. Clas-
1989;108(5):285—7. sification and evaluation. JBJS Am 1970;52:1077—89.