Keywords
proximal humeral fracture - elderly - nonoperative treatment - reverse shoulder arthroplasty
Introduction
Proximal humeral fractures (PHFs) are common, accounting for 10% of all skeletal fractures.
Their incidence increases exponentially after age 40.[1]
[2]
[3]
[4] We expect the PHF number to rise three-fold in the next decade.[5]
Although most of these fractures do not present displacement and their conservative
management is successful, several surgical techniques treat higher-complexity fractures.[6] The most appropriate treatment must consider the age of the patient, bone quality,
fragment number, and displacement.[7]
[8]
Three- and four-part fractures in elderly patients remain controversial. Reverse arthroplasty
is the most proper surgical option for these patients. Since there are reports of
specific surgical complications in significant percentages 9, conservative management
may be a feasible alternative in older patients with lower functional demand and potentially
compromised bone quality.[10]
[11]
[12]
This study compared the functional outcomes from two cohorts of elderly patients with
PHFs with three or four fragments per the Neer classification. Subjects underwent
a shoulder arthroplasty or conservative management. The minimum follow-up period was
12 months.
Material and Methods
We have designed an ambispective (with a historical cohort), single-center, non-randomized
study with two cohorts of elderly patients with PHFs treated at our center from October
2013 to June 2019.
The inclusion criteria were the following:
Sample Size Calculation and Recruitment
We have calculated a sample size of 15 patients for each cohort and an alpha error
of 5%, estimating an analytic power of 96% to detect mean differences of 14 points
or more on the ASES score with a standard deviation of 10.[13]
Patients undergoing reverse shoulder arthroplasty represented a historical cohort
of 15 patients recruited from October 2013 to April 2016. All of them were over 75
years old and had a PHF with three or four fragments and a minimum follow-up period
of 12 months. These patients' recruitment occurred for a prospective study to compare
the outcomes of the reverse arthroplasty of PHFs to another prospective cohort undergoing
shoulder hemiarthroplasty. The previously mentioned study is registered in Clinical
Trials under the name FRALUX34, NCT 03339570, with a currently pending publication.
From November 2017 to June 2019, we prospectively included 19 patients in the conservative
treatment cohort. Three subjects were excluded from the study due to death or lack
of adherence to the rehabilitation treatment. Therefore, the conservative treatment
cohort consisted of 16 patients with a minimum follow-up period of 12 months.
We informed all patients about the study, and they gave their written informed consent
for inclusion. The Ethics and Clinical Research Committee from our hospital approved
this study, which was subsequently registered in Clinical Trials under the name SCIARPA
3/4 NCT03339570.
Surgical Treatment
All patients undergoing reverse arthroplasty were operated on in the first week after
trauma (days 2-7) by two senior surgeons (FLM and JCAG). The surgical approach was
the standard deltopectoral approach, with tenodesis of the long head of the biceps
to the pectoralis major and coracoacromial ligament section. The implant was the Delta
Xtend reverse shoulder prosthesis (DePuy, Warsaw, IN, USA) with a cemented stem. We
re-anchored the greater and lesser tuberosities to the prosthetic stem using horizontal
cerclage alone with a high-strength suture (Etibond #5). We did not perform vertical
cerclage in any subject. The metaglene fixation used two polyaxial screws in a superior
and inferior position. The retroversion of the humeral stem was calculated between
10 and 20°. We did not perform any tendon transfer or lateralization techniques of
the rotation center (BIO-RSA).
Rehabilitative Treatment
All patients in both cohorts similarly received specific rehabilitation treatment.
Immobilization with a sling occurred during the first three weeks ([Table 1]). Elbow, wrist, and shoulder pendulum exercises were allowed from the first day
as tolerated. After 15 days, the patients began passive/assisted mobility exercises
in flexion and abduction. In the fourth week, active mobilization exercises started,
including rotations and progressive sling removal.
Table 1
|
Variable
|
Conservative management
(n = 16)
|
Reversed arthroplasty (n = 15)
|
p-value
|
|
Mean age, years (range)
|
82.7 (77-91)
|
79.4 (75-86)
|
0.0029
|
|
Gender, n (%)
|
|
|
0.55
|
|
Female
|
13 (81%)
|
11 (73%)
|
|
|
Male
|
3 (19%)
|
4 (27%)
|
|
|
Dominant hand, n (%)
|
|
|
0.64
|
|
Dominant
|
10 (63%)
|
11 (73%)
|
|
|
Non-dominant
|
6 (37%)
|
4 (27%)
|
|
|
Fracture type, n (%)
|
|
|
0.174
|
|
Three fragments
|
13 (81%)
|
10 (67%)
|
|
|
Four fragments
|
3 (19%)
|
5 (33%)
|
|
|
American Society of
Anesthesiologists (ASA), n (%)
|
|
|
0.362
|
|
ASA II
|
7 (44%)
|
8 (53%)
|
|
ASA III
|
7 (44%)
|
6 (40%)
|
|
ASA IV
|
2 (12%)
|
1 (7%)
|
|
Last follow-up, n (%)
|
16 (100%)
|
15 (100%)
|
|
Clinical and Radiological Evaluation
Follow-up visits for all patients from both cohorts occur 1 week, 3 weeks, 3 months,
6 months, and 12 months after the fracture.
The 12-month follow-up visit assessed mobility. The same specialist evaluated all
subjects, recording flexion with the arm close to the body and external rotation in
both abduction and adduction in degrees. Internal rotation assessed the thumb' position
to anatomical references (trochanter, gluteus, sacroiliac joint, T12 vertebra, or
interscapular level).
The patients answered functional assessment questionnaires at 3, 6, and 12 months.
These questionnaires included ASES,[14] Constant-Murley,[15] and DASH.[16] We also used the VAS scale for a single pain assessment in the final 12-month follow-up.[17]
The radiological study consisted of plain anteroposterior and axial shoulder radiographs
at each follow-up. In the conservative management cohort, we evaluated the time to
fracture consolidation, tuberosity positioning, and the presence of cephalic necrosis
or nonunion. In the reverse arthroplasty cohort, we assessed tuberosity positioning,
complications (including neurovascular injury, postsurgical infection, and implant
instability or loosening), peri-implant osteolysis, and scapular notching.[18]
We also recorded complications such as neurovascular injury, surgical wound infection,
implant instability, or the need for reintervention for any reason.
Statistical Analysis
The statistical analysis used the SPSS 22.0 program for Mac (SPSS Inc., Chicago, Illinois,
USA). To study quantitative variables with dichotomous qualitative features, we used
the Student's t-test if the distribution was normal or the Mann-Whitney U test if
it was non-normal. The determination of relationships between qualitative variables
employed the Chi-square test. Differences with a p-value less than 0.05 were statistically
significant.
Results
Demographics
There were no statistically significant differences regarding gender, fracture type
(three or four fragments), American Society of Anesthesiologists (ASA) classification,
or dominant side. However, the mean age was slightly higher for the conservative treatment
group (82.7 years) than for the surgical treatment (79.4 years) (p = 0.0029) ([Table 1]).
Functional Outcomes
[Table 2] shows the results on the functional assessment scales.
Table 2
|
Variable
|
Conservative management
(n = 16)
|
Reversed arthroplasty (n = 15)
|
p-value
|
|
Constant-Murley score, mean +/- SD
|
51.9 +/- 12.4
|
75.1 +/- 10.3
|
0.001
|
|
ASES score, mean +/- SD
|
72.4 +/- 11.4
|
74.1 +/- 11.9
|
0.552
|
|
DASH score, mean +/- SD
|
23.8 +/- 14.8
|
29.4 +/- 7.3
|
0.208
|
|
VAS, mean +/- SD
|
1.2 +/- 2
|
1.3 +/- 1.5
|
0.423
|
|
Flexion, mean (°)
External rotation in abduction (°)
External rotation in adduction (°)
|
99.3 +/- 20.1
23.4 +/- 15.5
21.6 +/- 13.8
|
128.9 +/- 17
35.7 +/- 13.9
29.3 +/- 7.6
|
0.001
0.032
0.066
|
|
Internal rotation
|
|
|
0.189
|
|
Trochanter
|
2 (12%)
|
1 (7%)
|
|
|
Gluteus
|
6 (38%)
|
5 (33%)
|
|
Sacroiliac joint
|
6 (38%)
|
6 (40%)
|
|
T12 vertebra
|
1 (6%)
|
2 (13%)
|
|
Interscapular
|
1 (6%)
|
1 (7%)
|
The Constant-Murley score at the 12-month follow-up was significantly higher in the
reverse arthroplasty cohort (p = 0.001). At 12 months, the mean score in the surgical
cohort was 75.1 +/- 10.3 compared to a mean score of 51.9 +/-12.4 in the conservative
management cohort ([Fig. 1])
Fig. 1 Constant-Murley score.
We did not find statistically significant differences in mean ASES, DASH, and VAS
scores between the two cohorts at 12 months of follow-up. The final mean ASES score
was 74.1 +/- 11.9 in the surgical cohort versus 72.4 +/-11.4 in the conservative management
cohort (p = 0.552). The DASH score was 29.4 +/- 7.3 in the surgical group and 23.8
+/- 14.8 in the conservative management group (p = 0.208). The final VAS score was
1.3 +/- 1.5 in the surgical group and 1.2 +/-2 in the non-surgical group (p = 0.423).
The range of mobility at 12 months presented statistically significant differences
for flexion and external rotation in abduction in favor of the surgical cohort. The
final mean flexion in the surgical group was 128.9 +/- 17 degrees versus 99.3 +/-
20.1 in the conservative management group (p = 0.001). External rotation in abduction
was 35.7 +/-13.9 in the surgical group and 23.4 +/- 15.5 in the conservative management
group (p = 0.032). The mean values at 12 months for external rotation in adduction
and internal rotation were better in the surgical group but with no statistical significance.
Both cohorts showed a significantly progressive improvement in the average Constant-Murley,
ASES, and DASH scores at 3, 6, and 12 months ([Figs. 1], [2] and [3]).
Fig. 2 American Shoulder and Elbow Surgeons (ASES) score.
Fig. 3 Disabilities of the Arm, Shoulder and Hand score (DASH) score.
Radiological Outcomes and Complications at 12 Months
All patients managed conservatively present fracture radiological consolidation in
the 12-month radiographs. However, most images were consistent with malunion or tuberosity
malpositioning ([Table 3]). Moreover, there were three cases of avascular necrosis.
Table 3
|
Variable
|
Conservative management
(n = 16)
|
Reversed arthroplasty (n = 15)
|
|
Pseudoarthrosis, n (%)
|
0 (0%)
|
−
|
|
Malunion, n (%)
|
10 (62%)
|
−
|
|
Avascular necrosis, n (%)
|
3 (19%)
|
−
|
|
Normal positioned tuberosities, n (%)
Scapular notching, n (%)
Loosening, n (%)
|
-
-
-
|
7 (43%)
5 (33%)
0 (0%)
|
Fifty- three percent of the patients undergoing reverse arthroplasty presented correct
tuberosity positioning. In the remaining cases, the tuberosities were absent or malpositioned.
Patients with normal positioned tuberosities had better results on the functional
scales at 12 months, although with no statistical significance ([Table 4]). In the evaluation 12 months after the fracture, there were five cases of scapular
notching (33%), including four type 1 and one type 2 notching,[18] but causing no higher pain or functional repercussion.
Table 4
|
Variable
|
Correct positioning (n = 8)
|
Incorrect positioning (n = 7)
|
p-value
|
|
Constant-Murley score, mean
|
73.8
|
74.4
|
0.0897
|
|
ASES score, mean
|
74.6
|
71.5
|
0.625
|
|
DASH score, mean
|
30.8
|
27.2
|
0.359
|
The surgical group had no complications (infection, instability, prosthetic loosening,
or surgical revision required for any reason). One patient presented lower brachial
plexus involvement in the immediate postoperative period, resolving without specific
treatment in the first 3 months.
Discussion
The literature comparing outcomes from reverse arthroplasty and conservative management
of PHFs 3 or 4 per the Neer classification in elderly patients remains scarce.
Roberson et al.[19] conducted a retrospective, non-randomized study with a 2-year follow-up in two cohorts
of 19 and 20 patients undergoing conservative treatment or reverse arthroplasty. These
authors find no differences in mobility or functional scores. Although their study
population was significantly younger compared to ours (average age, 71), their results
were consistent with our study regarding degrees of mobility (mean flexion, 120 degrees)
and mean ASES score (72 points). Fifteen percent of their surgical patients required
a reintervention, with no differences between early and delayed surgery. These authors
propose to offer non-surgical treatment for this group of patients.
Similarly, Chivot et al.[20] conducted a retrospective, multicenter study with two cohorts of 28 and 32 patients
with an average age closer to our population (77 and 79 years) and a minimum follow-up
period of 2 years. Like us, they found statistically significant differences favoring
the reverse arthroplasty group for the Constant-Murley score. However, the improvement
was small and perhaps of little clinical relevance (56.5 points for the surgical group
vs. 50.5 for the non-surgical group). As in our study, there were no differences in
the DASH or VAS scores, concluding that only patients with high functional demands
should undergo reverse arthroplasty.
As far as we know, the most complete study on this subject is the recently published
work by Lópiz et al.[21] This prospective, randomized trial compared 30 patients treated non-surgically with
29 patients undergoing reverse arthroplasty. All patients were at least 80 years old.
The only statistically significant difference occurred in the VAS scale for pain,
favoring the surgical group at the end of the 12-month follow-up (1.6 vs. 0.9). They
found no differences on the Constant-Murley score even though the surgical group had
a favorable score of six points, potentially representing a clinically relevant improvement.
In addition, they found no differences in the DASH score. The study concluded that
the high comorbidity rate in both cohorts could mask the potential benefits of reverse
arthroplasty for patients with higher functional demands and better health status.
The most striking finding from our study was the 20-point difference in the Constant-Murley
score favoring the surgical cohort, well above the minimum limit for a clinically
relevant outcome.[22] This difference is largely explained by the better outcomes in terms of mobility
in the surgical group. Moreover, the ASES and DASH functional scores, which do not
evaluate the range of mobility, had no statistically significant differences. Pain
perception also had no differences between the cohorts, consistent with other publications.
We also agree with other studies in not obtaining significant results when comparing
mobility in surgical patients with or without correct tuberosity positioning.[23]
[24]
[25]
Since our study is non-randomized and has a historical cohort, a selection bias could
occur and explain the differences in the range of mobility between groups, assigning
patients with lower functional demand to the non-surgical cohort.
It remains to be defined which patient is “elderly” since different studies use this
designation for subjects at least 75 years old.[26]
[27] Elderly patients often form a heterogeneous group of subjects with critical differences
regarding comorbidities, functional demand, and life expectancy. Again, this definition
may lead to selection bias and wrong conclusions. A recent study[28] comparing mortality in the geriatric population (over 75) with PHF in three or four
fragments undergoing reverse arthroplasty or conservative treatment found no differences
in the mortality rate 1 year after the fracture (8.1% for the surgical group and 10.8%
for the conservative treatment group). Likewise, a report claimed that the mortality
associated with these fractures does not depend so much on age but on previous functional
capacity and social factors.[29]
Our study has several limitations, including its non-randomized design, the use of
a retrospective cohort, its small sample size, and sufficient but limited 12-month
follow-up period. Its qualities include homogeneity in monitoring and data collection
and the standardization of the surgical procedure.
Conclusions
The surgical treatment of the elderly patient with PHF in three and four fragments
resulted in better outcomes in some movements (anterior flexion and external rotation
in abduction) and the Constant-Murley score. However, we did not find significant
differences in pain (VAS) or functional requirements for daily living activities,
reflected in ASES and DASH scores.
Delayed surgery also seems a valid option; some studies do not find significant differences
in functional outcomes or complications between patients who undergo an initial reverse
arthroplasty compared to those operating after attempting conservative treatment.[19]
[30]
For this reason, we propose offering an initial conservative treatment to most elderly
patients with EPH fractures in three or four fragments. The characteristics of patients
with high demands and good functional reserve who may undergo an initial surgical
treatment remain to be defined.
