Keyword
brachial plexus birth injury - hand function - long-term outcomes - surgery
Introduction
Most birth-related injuries to the brachial plexus involve spinal nerves C5 and C6,
which merge into the upper trunk of the plexus and are commonly known as Erb's palsy.[1]
[2] The upper trunk is responsible for the innervation of the shoulder muscles. Thus,
management and evaluation of the shoulder are often of primary concern in these patients.
The shoulder deficit is a marker of the degree of injury.[3]
[4] Spontaneous recovery with minimal impairment is seen in most patients with a C5–C6
lesion, ranging from approximately 50% to over 95% of patients.[5]
[6] The remaining patients (10–30%) suffer permanent impairment of shoulder function,
regardless of treatment.[7]
[8]
[9]
[10] Surgical intervention is generally accepted for patients with an upper/middle (C5/C6/C7)
injury and global palsies[5]; however, there has been more uncertainty surrounding early nerve reconstruction
for upper brachial plexus birth injuries (BPBI).
More recently, attention has been drawn to limitations of hand function in patients
with upper BPBI, contrary to the accepted dogma that the hand remains unaffected.[2]
In this study, we report long-term outcomes in children with C5–C6 pattern injuries
who underwent early nerve reconstruction.
Methods
A retrospective chart review of patients undergoing primary nerve reconstruction performed
by the senior author for Erb's palsy between 2000 and 2005 was conducted. Of the 131
patients who received early surgical intervention, 55 were identified as having an
upper BPBI. From this cohort of 55 patients with upper BPBI requiring reconstruction,
32 (58.2%) were available for long-term follow-up. These patients underwent clinical
evaluation, including calculation of shoulder function by Mallet grading system and
Miami shoulder score as documented in [Table 1] of this article. For reference we have also included the Miami shoulder score classification
system in [Fig. 1]. A 9-hole peg test (9-HPT) was performed on each patient, which was used to determine
hand function, and which had been previously validated in 2012[2] ([Table 2]).
Table 1
Miami shoulder classification: total grade for active abduction/forward flexion and
external rotation
|
Average active abduction/forward flexion (degrees)
|
Active external rotation (degrees)
|
Total score[a]
|
Grade
|
|
0
|
Fixed deformity
|
0
|
0 (no function)
|
|
<45
|
Full passive <10
|
1–2
|
1 (poor)
|
|
<90
|
<30
|
3–4
|
2 (fair)
|
|
90–120
|
30–45
|
5–6
|
3 (satisfactory)
|
|
120–150
|
45–90
|
7–8
|
4 (good)
|
|
>150
|
90
|
9–10
|
5 (excellent)
|
a Maximum shoulder score =10; decrease score by 1 point for a contracture of >20 degrees.
Table 2
Aggregate results of the nine-hole peg test
|
9-HPT involved hand(s) (mean of attempts)
|
9-HPT uninvolved hand(s) (mean of attempts)
|
Difference in performance(s) (involved − uninvolved)[a]
|
Percent difference (percent of uninvolved)[b]
|
Normative mean(s) (involved hand)[b]
|
Normative mean(s) (uninvolved hand)[b]
|
Normative Differences involved − uninvolved)
|
Normative percent difference (percent of uninvolved)[a]
|
|
Mean
|
25.2
|
22.4
|
2.9
|
12.8
|
17.7
|
17.0
|
0.68
|
4.2
|
|
SD
|
5.5
|
4.4
|
3.1
|
13.6
|
0.87
|
0.96
|
0.96
|
5.66
|
|
Range
|
19 to 48
|
18–42
|
−4 to 11
|
−14 to 43
|
17–19
|
16–18
|
−1 to 2
|
−7 to 14
|
Abbreviations: 9-HPT, 9-hole peg test; SD, standard deviation.
a Percent data were rounded to the nearest total percentage.
b Normative data were extracted from Poole et al and are sex, age, and hand dominance
matched. Discrepancies between the involved and the uninvolved extremities are recorded
as positive when the involved hand was worse (longer time to complete the test) and
negative when it was better (shorter time to complete the test).
Fig. 1 Miami classification. Shoulder evaluation: C5, C6 repair long-term follow-up (n = 32).
Basic analysis, including mean and standard deviation (SD), was performed on multiple
measures, as reported below ([Table 2]). Comparisons were made to the available normative population mean times matched
for age, sex, and hand dominance. Statistical analysis was performed using the Student's
t-test.
Results
Of the 55 patients with C5–C6 injuries that underwent primary nerve reconstruction,
32 patients were available for long-term follow-up. We detail the demographic makeup
and surgical history of all patients seen in follow-up. There were 13 males and 19
females. The mean age at surgery was 10 months (SD = 4.5 months). We do not generally
operate in our program on pure C5/C6/upper trunk lesions until 8 months of age except
in unusual circumstances. The mean age at the final follow-up was 15 years (SD = 2
years 2 months; [Supplementary Tables S1] and [S2], available in online version only). [Supplementary Tables S3] and [S4] (available in online version only) include the information available to the authors
about the patients not seen in long-term follow-up.
All patients underwent primary nerve surgery. The patients who presented with a shoulder
contracture or loss of full passive external rotation of the shoulder when examined
under anesthesia had a simultaneous surgical correction. Seven patients reported that
the involved extremity was their dominant hand (two left-handed and five right-handed
patients). The remainder reported that the involved hand was nondominant. In [Table 1], cumulative shoulder function was evaluated by Miami score as “good” or “excellent”
in 23 individuals and 4 individuals received a “poor” analysis of shoulder function.
The mean Miami score was 4.1 (SD = 1.3; [Fig. 1]).
9-HPT performance times are reported in [Table 2]. Twenty-three of the 32 patients took longer to complete the task using the involved
hand than the uninvolved hand. Eighteen of these patients are right-hand dominant,
and five are left-hand dominant. The mean time difference between the involved and
the uninvolved extremities was 2.9 seconds, compared with the expected difference
of 0.68 seconds (based on age-, sex-, and hand dominance-matched normative data).[11] This difference was statistically significant (p = 0.00035). On average, individuals took 12.8% longer to complete the test with the
involved hand; this was significantly different from the expected difference of 4.2%
(p = 0.001).
Two of the seven patients with hand dominance on the affected side demonstrated a
markedly poorer score for the involved hand. The mean time difference for these five
individuals was 2.7 seconds (SD = 5.3), compared with 2.9 seconds (SD = 2.3) for the
remainder of the cohort (not significant). Interestingly, three patients in this cohort
displayed no time difference between involved and uninvolved hands by 9-HPT.
Discussion
We report on both long-term shoulder range of motion and fine motor hand function
in children treated for upper brachial plexus birth palsy. We used the 9-HPT as a
validated outcome tool for the hand. The majority of the individuals in our cohort
had detectable hand function deficits. The persistent hand deficits seen in our study
are contrary to the accepted “normal” hand function in individuals with a C5–C6 lesion.
In 2012, our group published similar findings with a much smaller cohort with a shorter
follow-up time.[2]
In our current study, the majority of children reported their noninjured hand to be
dominant. If the function of the two hands is equal, hand dominance is likely determined
by a mechanism unrelated to the brachial plexus injury. The reported prevalence of
left-handedness in the general population is 8 to 10%; in our cohort, it is 31.2%.
This suggests that hand dysfunction can affect hand dominance, as it is likely that
some of the children selectively became left-handed to compensate for a persistent
functional deficit in their right hand.
One possible limitation of this study is that the 9-HPT requires active shoulder function
to position the hand to execute the test. Residual shoulder deficit might affect the
results. However, most individuals had “excellent” shoulder function on the Miami
shoulder scale ([Table 1]). No correlation was observed between the shoulder scores and the 9-HPT score.
Our data were derived from two attempts by each child individually. Each normative
data point was obtained by testing 21 to 43 children, with each child first performing
a practice followed by a timed test.[11] Therefore, the normative data are an average of many children performing the test.
These findings define the expected “norm” for the population of a given sex, age,
and hand dominance. We intended to compare our participants involved and noninvolved
hands to what would be expected in an unaffected age-matched population. However,
the normative data did not provide specific hand-to-hand comparisons for individual
children, only an aggregate for the tested group's left- and right-hand times. It
seems unlikely that typically developing children would exhibit striking differences
in performance between the dominant and nondominant hands, so we believe that our
comparison is still valid.
Conclusion
Most patients in our cohort demonstrated long-term deficits in hand function with
good or excellent shoulder function, challenging the generally accepted notion that
C5–C6 pattern birth injuries do not influence hand function. Injuries that involve
the upper brachial plexus exclusively may include subtle unrecognized injuries to
the lower brachial plexus as well. Early referral and implementation of multidisciplinary
strategies allow these children the best chance of functional recovery from upper
brachial plexus injury. Further study of hand function is warranted in these individuals
to redirect early treatment strategies.
