Afifty-year-old woman was admitted to the hospital with epigastric and left
upper-quadrant pain, loss of appetite, nausea and vomiting, and vague low-back
pain that radiated to the left thigh. The back pain became worse the following
day. Seven weeks earlier, the patient had been admitted for pancreatitis and
had been diagnosed with a pancreas divisum, which occurs when the embryologic
ventral and dorsal pancreatic anlagen fail to fuse. The patient had no history
of alcohol abuse, back pain, or rheumatologic disease, and she had not taken
any medication.
Palpation of the spine at the thoracolumbar junction resulted in intense
pain. Deep tendon reflexes and sensation were normal. The serum amylase level
was marginally elevated. The patient's condition improved after
marsupialization of a pancreatic pseudocyst was performed, although the lumbar
pain persisted. A radiograph of the spine revealed mild spondylosis. When the
patient was discharged a few weeks later, she still had low-back pain of
varied intensity.
She was admitted again one week later with abdominal pain, low-back pain
that radiated to the left thigh, and dysesthesia in the proximal-lateral
portion of the left thigh. Clinical examination revealed a positive
Lasègue sign at 50° on the left side. Flexion of the left hip joint
was weaker than that of the right. The weakness was classified as grade 4 (of
5) and was attributed to the presence of pain in the hip. The leg was not
diffusely weak, however, and at subsequent examinations, the strength in the
left hip appeared to be normal. Percussion of the spinous processes at the
thoracolumbar junction did not elicit pain. The deep-tendon reflexes were
normal. Clinical and neurological examinations did not reveal impingement of
the lateral femoral cutaneous nerve. Blood cultures were negative. A
technetium-99m-methylene diphosphonate bone scan showed areas of intensely
increased activity in the body of the twelfth thoracic vertebra
(Fig. 1). Orthopaedic
consultation was sought, and magnetic resonance imaging of the spine was
performed (Figs. 2-A and 2-B).
An area of altered signal intensity was observed in the posterior part of the
twelfth thoracic vertebra. The signal intensity was low on T1-weighted images
and high on T2-weighted images. These signal changes were consistent with
increased water content but were rather nonspecific otherwise. An open biopsy
of the twelfth thoracic vertebra was performed, and a clear liquid with
free-floating fatty droplets was aspirated. Cultures of the fluid specimens
demonstrated no growth of microorganisms, including Mycobacterium
tuberculosis. The amylase level in the aspirate was measured to be 32,000
IU/L (normal, <128 IU/L), and no malignant cells were found. Histologic
evaluation of the biopsy specimen showed fat necrosis of the bone marrow
(Fig. 3). The low-back pain
worsened after the biopsy. A repeat magnetic resonance imaging scan showed
progression of the primary pathologic changes and slight displacement of the
spinal cord, with signal changes in the twelfth thoracic, first and second
lumbar, and first sacral vertebrae. Endoscopic retrograde
cholangiopancreatography and cannulation of the accessory pancreatic duct were
performed. Cytologic evaluation of the pseudocystic aspirate revealed
malignant cells. The decision was made to perform a pancreatoduodenectomy (a
Whipple operation). Pathologic investigation of the resected material revealed
adenocarcinoma of the dorsal portion of the pancreas. The patient's condition
deteriorated, and she died eleven days after the operation. No autopsy was
performed.
In an autopsy study published in 1956, Scarpelli described fat necrosis of
vertebral bodies and concluded that because of its focal nature, the
pathophysiologic and diagnostic importance of that entity was probably
limited11. That
conclusion stands in contrast to the conclusions drawn from the present case
and from the case described by Allen and Jinkins in 1978, in which vertebral
osteonecrosis appears to have been associated with severe
symptoms12.
Disseminated fat necrosis is a well-known complication of acute, chronic,
and traumatic pancreatitis as well as pancreatic cancer. It often is
associated with osteolytic lesions in the metaphysis or diaphysis of long
bones as well as in the calcaneus and
talus2-9,12-16.
The clinical presentation of disseminated fat necrosis in patients with
pancreatic disease is often in the form of panniculitis, polyarthritis, and
bone
lesions2,5,10,17-19.
Intraosseous fat necrosis in patients with pancreatic disease may or may not
be painful and may present as a pathologic
fracture1,16,20.
The pathogenesis of disseminated fat necrosis has been debated and is poorly
understood; it may be more complex than a simple cause-and-effect relationship
between circulating pancreatic enzymes and fat
tissue21,22.
Pancreatic lipase has been implicated, with the pancreatic phospholipase A
initiating the process and the resident adipocyte lipases taking an active
part, stimulated by the increased norepinephrine and low insulin levels that
are usually associated with
pancreatitis23.
High levels of alpha1-antitrypsin have been found in the serum of
patients with disseminated fat necrosis, and a defect in that protein, or
another as yet unidentified protease or lipolytic inhibitor, could help to
explain the uninhibited fat
necrosis19,24.
Back pain is often associated with both chronic and acute pancreatitis,
typically in conjunction with abdominal pain. The pain may radiate to the
middle part of the back or to the scapula, and it may be relieved by leaning
forward or sitting
upright25,26.
Visceral disease, including pancreatitis, accounts for about 2% of all cases
of low-back
pain27.
The history of our patient suggests a spinal etiology of the low-back pain.
The tumor had grown into the celiac plexus and therefore may have contributed
to low-back pain. Idiopathic vertebral necrosis without vertebral body
collapse also may be associated with back
pain28. One can
only speculate, therefore, about the origin of low-back pain; in the case of
our patient, however, the pain was attributed to necrosis of the vertebral
body. One would expect the rate of progression of the osteonecrosis to be
related to the progression of the pancreatic disease, but reliable markers to
monitor the progression of osteonecrosis do not exist. Our patient had back
pain when she was admitted for pancreatitis for the second time in three
months. The biopsy that was performed on both occasions, and after magnetic
resonance imaging had been carried out, excluded conditions such as vertebral
osteomyelitis, spinal tuberculosis, metastatic tumor, and multiple myeloma of
the spine.
Percutaneous vertebroplasty should be considered for the treatment of a
nonhealing vertebral defect in a patient with a high risk of a compression
fracture. However, because our patient had pancreatitis and vertebral
osteonecrosis in addition to pancreatic adenocarcinoma, the prognosis was very
grave and we did not elect to perform this procedure.
The osteolytic complications of pancreatic disease should be kept in mind
as a possible cause of protracted and atypical low-back pain in patients with
pancreatitis, and clinicians should have a low threshold for acquiring a
magnetic resonance imaging scan. We believe that it is important for the
general orthopaedic surgeon as well as for the spinal surgeon to be aware of
these complications and their management because orthopaedic surgeons are
often consulted regarding osteolytic lesions, irrespective of etiology. This
case report adds to our knowledge regarding the differential diagnosis in
patients with pancreatitis and back pain and therefore contains important
information for those who treat seriously ill patients who have back pain.
?