Introduction
Median nerve compression occurs mostly at sites of fibro-osseous or fibromuscular
tunnels. The most common site is at the carpal tunnel, causing carpal tunnel syndrome
(CTS). However, proximal compression of the median nerve such as pronator syndrome
and Lacertus syndrome should not be overlooked.
The etiology of CTS is mostly an idiopathic cause. Chronic trauma, rheumatoid arthritis,
menopause, pregnancy, hypothyroidism, diabetes mellitus, acromegaly, and amyloidosis
are also associated with the development of CTS.[1]
[2]
Median nerve compression due to a benign or malign tumor is rare. A lipofibromatous
hamartoma (LFH) is the most common histological type of tumor, but also lipomas, ganglion
cysts, and fibromas are rarely found in the trajectory of the median nerve. These
tumors can be incidental findings during diagnostics or routine carpal tunnel surgery.
Unfortunately, they can have an uncommon presentation of symptoms which makes it hard
to diagnose or distinguish from an idiopathic cause, with the possible consequence
of a prolonged duration of the diagnostic process or even multiple surgical procedures.
If missed or neglected, irreversible damage to the median nerve may occur.[3] This narrative review presents a list of uncommon tumors which can cause median
nerve compression and may help to avoid diagnostic delay and unnecessary surgery.
The objective of this review is to provide a comprehensive overview of tumors described
in recent publications that cause median nerve compression and to evaluate treatment
options.
Materials and Methods
Search Strategy
PubMed, Embase, and Web of Science were searched for studies which described unique
or rare cases of median nerve compression due to tumors. The first search was performed
on May 11, 2020, and the last update of the search was performed on May 25, 2020.
Due to the lack of a valid bias analysis for narrative reviews, specific criteria
for article selection and inclusion and exclusion were developed. The following criteria
have been used to include a study to this review: papers describing a case, cause
or etiology of median nerve compression, tumors, year of publication from 2000, age
of patients 18 years or older, and written in English.
Excluded studies involved cadavers, animals, and cases of other causes of median nerve
compression (including idiopathic CTS, pregnancy, edema, rheumatoid arthritis, gout,
diabetes mellitus, hypothyroidism, tenosynovitis, acromegaly, repetitive tasks, anatomic
anomalies, traumatic neuropathy, and infectious causes). A detailed electronic search
strategy is described in [Supplementary Appendix A] (available in the online version).
Data Extraction
Data of patient characteristics, side of compression, uni- or bilateral compression,
cause of compression, symptoms at presentation, duration of symptoms at presentation,
physical examination, diagnostics, treatments, location of tumor, results of treatment,
complications, follow-up, and recurrence were extracted and analyzed by two reviewers
with the use of EndNote X8.2 (Clarivate Analytics, Philadelphia, PA).
Disagreements were resolved by discussion or by the third reviewer. All outcomes were
processed in SPSS (IBM SPSS statistics version 26). If any data was missing studies
were excluded from that particular part of statistical analysis.
Data Analysis
All included studies were analyzed together for the general characteristics. In addition,
three categories were formed: neural, nonneural, and combined. The neural category
includes tumors consisting of neural elements. The nonneural category includes tumors
arising from the surrounding tissue. In case of the presence of multiple tumors from
both categories, the case was listed in the combined category. Furthermore, cases
of proximal median nerve compression were analyzed separately from compression at
carpal tunnel.
All papers were evaluated and analyzed in a systematic way to reduce bias. Studies
with a limited case description are still included to reduce reporting bias and to
include all available information.
Results
Study Selection
The initial search revealed 2,024 studies. After removing duplicate studies (n = 534), abstracts from 1,490 studies were evaluated and 1,396 studies were excluded
based on the title and abstract. Furthermore, 15 studies were excluded on the base
of full-text analysis leaving a total of 79 papers for full-text review. Also, there
were 15 additional studies found in the reference lists. [Supplementary Appendix B] (available in the online version) presents the study selection process.
Case Characteristics
Eventually, 94 studies with a description of 100 cases of median nerve compression
due to a tumor have been reviewed. There were 45 (46%) male patients and 54 (56%)
female patients with a mean age of 45.14 (ranging from 18 to 86) years. In 39 cases
the left side was affected, in 56 cases the right side. Only two bilateral cases were
described. In 8 cases (8%) the patient presented for the second time, after failing
of initial therapy (6 carpal tunnel releases and 2 nonsurgical treatments). [Table 1] describes the general characteristics of the reviewed cases.
Table 1
Characteristics of cases
|
Frequency (n)
|
Percent (%)
|
|
Gender[a]
|
Male
|
45
|
46
|
|
Female
|
54
|
56
|
|
Side[b]
|
Left
|
39
|
40
|
|
Right
|
56
|
58
|
|
Bilateral
|
2
|
2.1
|
|
Sensory symptoms
|
|
82
|
82
|
|
Motor symptoms
|
|
27
|
27
|
|
Pain
|
|
53
|
53
|
|
Swelling
|
|
40
|
40
|
|
Triggering
|
|
6
|
6
|
|
Preoperative imaging[c]
|
|
81
|
87
|
|
Treatment[d]
|
Carpal tunnel release
|
52
|
55
|
|
Complete resection mass
|
58
|
62
|
|
Partial resection mass
|
4
|
4.3
|
|
Only biopsy
|
7
|
7.4
|
|
Treatment outcome
|
Recovery
|
43
|
43
|
|
Improvement
|
28
|
28
|
|
Not described/unknown
|
29
|
29
|
|
Recurrence
|
Yes
|
3
|
3
|
|
No
|
47
|
47
|
|
Not described/unknown
|
50
|
50
|
a For 1 case missing data.
b For 3 cases missing data.
c For 7 cases missing data.
d For 6 cases missing data.
Symptoms
Most patients (n = 82) experience sensory symptoms at first presentation, 53 patients reported pain
and only 27 patients had motor dysfunction. In 40 cases swelling was reported. Triggering
of the wrist was presented in six cases. In the neural category, 22 patients experienced
sensory symptoms (63%), 15 reported pain (43%), 6 had motor dysfunction (17%), 20
swelling (57%), and 1 triggering (3%). In the nonneural category, 58 patients had
sensory symptoms (92%), 36 patients experienced pain (57%), 20 had motor dysfunction
(32), 20 swelling (32%), and in 5 patients the lesion caused triggering of the wrist
(7.9%).
The tumor was located in the proximal (fore)arm in 28 cases. A total of 15 out of
these 28 patients experienced pain (54%).
Treatment
In 58 cases the mass was completely resected and in only 4 patients a partial resection
was possible due to a strong adherence of some areas of the tumor to the median nerve
and due to the inability to distinguish the tumor from the surrounding tissue.
The most performed procedure was an open median nerve release (52 patients) whether
or not combined with resection of the mass causing compression. In one case an endoscopic
carpal tunnel release was performed. In three cases the patient rejected a surgical
approach. Treatment in case of neural masses: 22 open median nerve releases (69%),
complete resection in 15 cases (47%), 1 partial resection (3.1%), and 3 resection
biopsies (9.4%) have been performed. Treatment for nonneural masses consisted of 28
performed open median nerve releases (47%), 41 complete resections (68%), 3 partial
resections (5%), and 4 resection biopsies (6.7%).
Outcome of Treatment
The mean follow-up period was 11 months. Not all studies described the outcome of
treatment and eventual recurrence. Information about the recovery of 43 out 71 patients
was reported in the studies of which 28 out of 71 patients reported an improvement.
In 47 out of 50 patients no signs of recurrence were present at follow-up. In 50 cases
the recurrence rate was not reported.
Causes of Median Nerve Compression
[Table 2] shows all 35 different histological types of 100 lesions that can cause median nerve
compression. There are three main categories of tumors causing median nerve compression:
neural origin (n = 35), nonneural origin (n = 63), and a combination of causes (n = 2). The most common cause of neural origin is the LFH (19 patients, 19%). Lipomas
are the most common cause in the nonneural origin category (16 patients, 16%). In
two patients, a combination of causes was present. A total of nine tumors were malignant.
The pie chart in [Fig. 1] provides a quick overview of all causes.
Fig. 1 Pie chart of tumors causing median nerve compression. The exploded slices show neural
causes. Causes in the group “other” occurred only once and contains: hamartoma, glomus
tumor, extraskeletal chondroma, synovial osteochondromatosis, synovial hemangioma,
osteophyte trapezium, metaplastic synovial cyst, melorheostosis, macrodystrophia lipomatosa,
hibernoma, and calcifying aponeurotic fibrolipoma. Causes in the group “other (malignant)”
occurred only once and contains: epithelioid sarcoma, non-Hodgkin lymphoma, neuroleukemiosis,
and metastasis of breast cancer.
Table 2
Tumors causing median nerve compression
|
Tumor
|
Frequency (n)
|
Percent of total (%)
|
Percent of category (%)
|
|
Neural origin
(n = 35)
|
Lipofibromatous hamartoma
|
19
|
19
|
54
|
|
Schwannoma
|
8
|
8
|
23
|
|
Intraneural lipoma
|
2
|
2
|
5.7
|
|
Intraneural perineurioma
|
3
|
3
|
5.7
|
|
Neurothekeoma
|
2
|
2
|
5.7
|
|
Plexiform neurofibroma
|
1
|
1
|
2.9
|
|
Nonneural origin
(n = 63)
|
Lipoma
|
16
|
16
|
26
|
|
Ganglion
|
6
|
6
|
9.5
|
|
Fibroma
|
6
|
6
|
9.5
|
|
Tenosynovial giant cell tumor
|
5
|
5
|
3.2
|
|
Deep leiomyoma
|
2
|
2
|
3.2
|
|
Heterotrophic ossification
|
2
|
2
|
3.2
|
|
Myoepithelioma
|
2
|
2
|
3.2
|
|
Myxofibrosarcoma[a]
|
2
|
2
|
3.2
|
|
Osteochondroma
|
2
|
2
|
3.2
|
|
Solitary calcified nodule
|
2
|
2
|
3.2
|
|
Synovial sarcoma[a]
|
2
|
2
|
3.2
|
|
Calcifying aponeurotic fibrolipoma
|
1
|
1
|
1.6
|
|
Epithelioid sarcoma[a]
|
1
|
1
|
1.6
|
|
Extraskeletal chondroma
|
1
|
1
|
1.6
|
|
Glomus tumor
|
1
|
1
|
1.6
|
|
Hamartoma
|
1
|
1
|
1.6
|
|
Hibernoma
|
1
|
1
|
1.6
|
|
Macrodystrophia lipomatosa
|
1
|
1
|
1.6
|
|
Malignant tenosynovial giant cell tumor[a]
|
1
|
1
|
3.2
|
|
Melorheostosis
|
1
|
1
|
1.6
|
|
Metaplastic synovial cyst
|
1
|
1
|
1.6
|
|
Metastasis of breast cancer[a]
|
1
|
1
|
1.6
|
|
Neuroleukemiosis[a]
|
1
|
1
|
1.6
|
|
Non-Hodgkin lymphoma[a]
|
1
|
1
|
1.6
|
|
Osteophyte trapezium
|
1
|
1
|
1.6
|
|
Synovial hemangioma
|
1
|
1
|
1.6
|
|
Synovial osteochondromatosis
|
1
|
1
|
1.6
|
|
Combined
|
Intraneural lipoma and AVM
|
1
|
1
|
50
|
|
Lipofibromatous hamartoma and osteochondroma
|
1
|
1
|
50
|
Abbreviation: AVM, arteriovenous malformation.
a Malignant tumors.
Location of Compression
Tumors causing the compression of the median nerve can occur at every point of the
trajectory. [Fig. 2] shows the results of the found locations of the tumors. Compression due to a tumor
occurs most often in the carpal tunnel (n = 32), 21 lesions were found in the palm of the hand and 13 lesions were located
in the distal forearm. The other sites of compression ranged from biceps to the proximal
phalanx of the index finger. In a few cases, the lesion was so extensive that it took
up more than one location, among others a lesion in the distal forearm reaching to
the palm of the hand.
Fig. 2 Pie chart of the anatomic locations of median nerve compression due to a tumor.
Discussion
Overview of Tumors Causing Carpal Tunnel Syndrome
Neural (n = 35)
Lipofibromatous Hamartoma
The most common cause of median nerve compression of neural origin in this review
is the LFH (19 cases). The LFH is a benign and slow growing peripheral nerve tumor.
It has predilection for the median nerve and it is associated with macrodactyly (more
than one-third).[4]
[5] This type of tumor is mostly reported in young patients.
Fibrofatty tissue infiltrates the nerve which separates and compresses the nerve,
so complete excision is not possible without damaging the nerve. In case of CTS, decompression
surgery has shown to be effective. Different surgical approaches are advocated, among
which microsurgical intraneural dissection and excision with nerve grafting. If excision
is not possible, a biopsy can confirm the diagnosis.[6]
Schwannoma
Schwannomas also called neurilemmomas, are benign, slow growing neural tumors originating
from Schwann cells of a peripheral nerve. The incidence in adults is only 5%.[7] Typical finding at physical examination is a mass which is mobile in longitudinal
plane but not in transversal plane.[7] The nerve fascicles are pushed aside while the tumor grows. Therefore, complete
enucleation is in most cases possible.[8] Recurrence is rarely reported, even in incomplete excision.[7]
Intraneural Perineurioma
In three patients an intraneural perineurioma caused median nerve compression. This
is a slow growing benign tumor. The presentation is in most cases as a slow progressive
mononeuropathy causing hypoesthesia and motor weakness.[9] No consensus about therapy exists, varying from conservative management[10] to resection with interposing nerve graft.[11]
[12] Uerschels et al promotes interfascicular neurolysis in combination with decompression
surgery.
Intraneural Lipoma
An intraneural lipoma is encapsulated but does not contain neural elements. Therefore,
a complete excision is possible without damaging the nerve. In only two cases an intraneural
lipoma caused median nerve compression.
Neurothekeoma
A neurothekeoma is a benign, small, and slow growing tumor which originates from the
endoneurium of peripheral nerves. It mostly occurs before 30 years of age with a preference
for the female gender. It has a high tendency of local recurrence (11%).[13]
[14] Complete surgical excision is the treatment of first choice.
Plexiform Neurofibroma
Neurofibromas are benign nerve sheath tumors occurring in patients between 20 and
30 years. Tumor mass and nerve fascicles are intertwined and therefore complete excision
may sacrifice the nerve. With microdissection of the neurofibroma, nerve fascicles
can be spared. Plexiform neurofibromas are associated with neurofibromatosis-1 and
surgical removal is often impossible due to their serpiginous growth, significant
vascularization, and multiple nerve fiber involvement. In previous publications, a
recurrence rate of 50% is reported.[15]
In this review, all neural causes were benign. Malignant peripheral nerve sheath tumors
were not described in the reviewed articles, but must also be considered in the differential
diagnosis. Other benign neural causes, not described in this review, are neuroma and
osteoid osteoma.
Nonneural (n = 63)
Benign lesions are often easily resected with minimal nerve morbidity and with a very
good outcome.[16]
Soft Tissue
Lipoma
In 16 cases a lipoma was found as a cause of median nerve compression. Lipomas are
soft encapsulated masses, easily separated from surrounding tissue. Chen et al found
three lipomas in 23 space occupying lesions and Cavit et al reported 20 out of 402
hand tumors to be lipomas (4.9%).[17]
Ganglion Cyst
Ganglion cysts can communicate with tendon (sheats), nerves, or joints and can arise
from joint synovial tissue. Cavit et al published a review in which ganglion cysts
were the most common soft tissue tumors (125 out of 402, 31%). In this review, ganglion
cysts caused median nerve compression in 6 out of 63 nonneural tumors. Recurrence
rate in case of aspiration is higher compared with a surgical excision.[18]
[19]
[20]
Fibroma
A fibroma is a well-circumscribed, lobulated, and sometimes encapsulated benign mass,
which is generally attached to tendon sheaths. Fibromas are easily excised. Evangelisti
and Nojiri describe a 24% recurrence rate in case of inadequate excision.[22]
[30]
Tenosynovial Giant Cell Tumor
Also called pigmented villonodular synovitis, caused by a benign, but invasive synovial
proliferation. Treatment requires extensive synovectomy and in some cases adjuvant
radiotherapy depending on the extensiveness of the lesion. Despite aggressive therapy,
a high recurrence is reported (8–56%).[21]
Other uncommon causes found in this review were a deep leiomyoma (n = 2), myoepithelioma (n = 2), solitary calcified nodule (n = 2), glomus tumor (n = 1), hamartoma (n = 1), hibernoma (n = 1), metaplastic synovial cyst (n = 1), synovial hemangioma (n = 1), and calcified aponeurotic fibrolipoma (n = 1).
Bone and Cartilage
A total of eight bony and cartilage tumors were found in this review. All of these
tumors were benign in nature. The following histological bony and cartilage tumors
were found as a cause of median nerve compression: osteochondroma (n = 2), heterotopic ossification (n = 2), osteophyte (n = 1), extraskeletal chondroma (n = 1), melorheostosis (n = 1), and synovial osteochondromatosis (n = 1).
Malignant
Merely 9 out of 100 reviewed lesions causing median nerve compression were malignant,
involving epithelioid sarcoma, non-Hodgkin lymphoma, neuroleukemiosis, and a metastasis
of breast cancer. There were only nonneural malignant masses. Especially, fast growing
tumor or a tumor with more than 5 cm in diameter has to be suspected for malignancy.[23]
Combined (n = 2)
In two patients a combined cause of median nerve compression was found.
The first case described treatment of proximal median nerve compression with full
release of median nerve in the forearm and epineurotomy.
During the operation, a lipomatous tumor was found and left undisturbed. Postoperative
imaging showed a LFH and an osteochondroma.
Due to recurrent symptoms and diagnosed osteochondroma, a second operation consisting
of excision of the osteochondroma was performed. The median nerve was pushed anteriorly
by the osteochondroma and internal compression took place due to the LFH.
The second patient was diagnosed with an intraneural lipoma combined with an arteriovenous
malformation during preoperative imaging. Both lesions were totally excised in combination
with a carpal tunnel decompression. The patient had full recovery and there were no
signs of tumor recurrence during follow-up. The authors recommend preoperative imaging
in case of unilateral CTS and palpable tumors.
Both papers make no statement about incidence or coincidence of finding both tumors
in one patient.
The most described cause of median nerve compression due to a mass lesion in previous
publications is a LFH. The second and third most common published cause is a lipoma
and a ganglion cyst, respectively. In this review, 8 patients already received treatment
for median nerve compression but experienced persistent or recurrent symptoms. In
all 8 cases unilateral symptoms were present and initial treatment through carpal
tunnel release or lidocaine/corticosteroid injection was performed without preoperative
imaging. Therefore, a tumor causing CTS was initially missed and after a period of
persistence of symptoms after initial treatment, the tumor was found on imaging. Owing
to the high incidence of idiopathic CTS, tumors are often missed. In all cases of
diagnosing the tumor preoperatively, no recurrence of symptoms was reported, except
for tenosynovial giant cell tumors. For these tumors, a high recurrence rate was already
reported.[21]
After analysis of all cases included in this review, we conclude imaging could be
helpful for correct diagnosis and preoperative planning, especially in cases of unilateral
median nerve compression to prevent multiple operations.
This review consists of case reports and case series published in the English literature,
which are unfortunately known for a low level of evidence. However, these cases provide
useful information for clinical practice. This study is limited by publication bias.
Case reports will only be published in case of an interesting or rare topic, such
as rare causes of median nerve compression. For actual incidence rates of common causes
of median nerve compression, other research articles must be consulted.
The carpal tunnel is a preferred location for median nerve compression, but compression
can occur along the complete trajectory of the median nerve. It can be difficult to
distinguish a tumor as a cause of median nerve compression from the idiopathic cause
of CTS, which is more common. However, a palpable swelling was found in 40 patients,
suggesting a lesion as the cause of the complaints. Besides, in 6 cases, triggering
of the wrist was a symptom at first presentation which is an indication of a lesion.
These findings are essential not only for diagnostic process, but also for adequate
preoperative planning of patients. A total of 53 out of 100 patients experienced pain
as a symptom. From the 28 cases of proximal median nerve compression due to a tumor,
15 patients experienced pain (54%).
In case of the presence of swelling or wrist triggering combined with symptoms of
median nerve compression, a tumor must be suspected and therefore imaging is recommended.
Landi et al provided a thorough and clear overview of rare causes of CTS. However,
they used studies up to 1998 and only wrote about median nerve compression at the
site of the carpal tunnel. In our review, we analyzed studies about median nerve compression
due to tumors, from 2000 till present, without restriction of a compression site.[31]
The causes were divided into three groups (neural, nonneural, and combination of causes),
to provide a clear overview. However, the treatment cannot be grouped and should always
be conducted according to the corresponding guidelines per country and/or hospital.
Malignant tumors causing median nerve compression are uncommon. However, leukemia,
lymphomas, and even metastases can cause median nerve compression if located in the
arm.[24]
[25]
[26]
[27]
[28]
Conclusion
Rare tumors causing median nerve compression can be divided into three groups: neural,
nonneural, and a combined origin. The most common cause of compression in our review
is due to LFH, a neural tumor. Complete excision of LFH is not possible without damaging
the nerve but decompression surgery can relief symptoms.[6] Lipomas form the most frequent cause of uncommon median nerve compression in the
nonneural category. In general, complete excision of a lipoma is possible.
A thorough physical examination of the complete upper extremity must be performed
and documented to find any swelling along the course of the median nerve. Based on
the data found in this review, in case of swelling or triggering, a lesion must be
suspected and therefore early imaging diagnostics are recommended.
To choose the adequate treatment, it is essential to have a preoperative diagnosis
before the operation. Complete excision of neural masses is often more difficult in
comparison to nonneural masses.[29] Malignant tumors as a cause of median nerve compression is extremely rare. Therefore,
treatment should be adapted to the national guidelines and with multidisciplinary
consultation.
