- Email: [email protected]
Cerebrospinal fluid cell count following spinal cord injury
293
Cerebrospinal Fluid Cell Count Following Spinal Cord Injury Andrew
Travlos MB, ChB, Hugh A. Anton,
MD, Peter C. Wing, MB, ChB
ABSTRACT. Travlos A, Anton HA, Wing PC. Cerebrospinal fluid cell count following spinal cord injury. Arch Pbys Med Rehabil 1994;75:293-6. l Cerebrospinal fluid (CSF) changes after spinal cord injury (SCI) were evaluated by retrospective analysis of all patients admitted to the Acute Spinal Cord Injury Unit (ASCIU) at University Hospital, Shaughnessy Site in Vancouver, British Columbia. A total of 1,917 admissions occurred during the IO-year study period with 1,151 due to acute trauma. The charts of all patients with SC1 due to trauma were reviewed to identify patients in whom a CSF puncture was performed for any reason. Traumatic SC1 is associated with elevations in the corrected white blood cell count (cWBC) that we believe reflect an inflammatory response to injury. The elevation in cWBC is greatest in the first week after injury. Elevations in cWBC more than 1 week after injury do occur but are uncommon. The white cell differential count consists primarily of lymphocytes and polymorphonucleocytes. Three weeks after injury cell counts should be normal. Protein levels are elevated after trauma and should not be used to assess the presence of infection. LP should be part of the septic workup of SC1 patients just as it would be for any other population if meningitis is a consideration. 0 1994 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and
Rehabilitation The development of dedicated spinal cord injury (SCI) units over the last 20 years has led to improved care after traumatic spinal cord injury. An important component of tertiary or quaternary care is decreased morbidity and mortality, and better diagnosis and treatment of complications. Survival after SC1 has improved significantly during the latter part of this century’.’ but infection remains a major clinical problem. Fever and/or infection occur in up to 67% of patients,’ with urinary tract and respiratory infections being most often responsible.“6 Disease of the respiratory system has been identified as a major cause of death after SC1 and 78.5% of respiratory deaths are associated with infection. Other infections were responsible for 8.8% of all deaths according to the National SC1 Database.2 In a more recent publication by Whiteneck and colleagues,7 disorders of the genitourinary system were the leading cause of death over all and urinary tract infections occurred with an annual incidence of 20%. Clinicians who manage patients with SC1 know that the potential causes for fever are multiple and many patients after investigation are labelled as having pyrexia of unknown origin. Further, the clinical evaluation of fever in such patients is frequently difficult because of their neurologic deficit, impaired thermoregulation, and immobilization in rigid external fixation. Consequently, one’s clinical examination may be unreliable and inadequate. This often leads to either overinvestigation, or underinvestigation with over treatment. Meningitis is a potential source of fever that may be particularly difficult to diagnose clinically in patients with acute From Anton).
the Departments
of Physical
and Orthopedics,
(Dr.
Medicine
Wing)
University
and Rehabilitation Hospital.
(Dr.
Shaughnessy
Twvlos,
Dr.
Site. Vancou-
ver. BC. Canada. Submitted
for publicatton
No comtnerc~al supporting
thrs article
organization Reprint
wth
0
1994
Academy
3Nl.
a benefit
27. 1993.
in the results of the research
upon
the authors
or upon
any
MB,
ChB,
Shaughnessy
Division
of Phytical
Site, Room
A 170-4500
Medicine
and
Oah Street.
Canada.
by the American of Physical
tn revised form May
interest
are associated.
Travlos.
Hospital.
tinancial
confer
the authors
UnivrrGty BC V6H
23, 1993. Accepted
a direct
has or will
which
requests to Andrew
Rehabilitation, Vancouver.
March
party having
Congress
Medicine
(~003.9993/94/7503-OU63$3.00/0
of Rehabilitation
and Rehabilitation
Medicine
and the Amencan
SCIs. The consequences of not diagnosing and treating a patient with meningitis may be devastating. A patient who already has a significant neurologic impairment due to SC1 may be placed at risk for further damage to the central nervous system with consequent increased morbidity. impairment and resultant disability. Our interest in this problem occurred after we saw a patient on our acute SCI unit in whom we were unable to exclude the diagnosis of a partially treated meningitis. The patient’s cerebrospinal fluid (CSF) white blood cell count (WBC) was 10, he was febrile and confused, and no other source for his symptoms could be found. A review of the current literature found little mention of either meningeal infection or inflammation as a possible source of fever in SCI. Nor was there data on the spectrum of CSF cell count after SCI. We believe that the lack of such data is a barrier to making informed clinical decisions. We hypothesized that SC1 might cause bleeding into the CSF coupled with an inflammatory response and resultant increase in the white cell count. The purpose of this retrospective study was to determine if the inflammatory process following traumatic SC1 causes a rise in the CSF white cell count. METHODS CSF changes after SC1 were evaluated by retrospective analysis of all patients admitted to the Acute Spinal Cord Injury Unit (ASCIU) at University Hospital, Shaughnessy Site in Vancouver, British Columbia. The ASCIU is a quarternary referral center for the management of spinal injuries in the province of British Columbia. A total of 1,9 I7 admissions occurred during the study period of April 198 1 to March 1991 with 1,1151 due to acute trauma. The charts of all patients with SC1 due to trauma were completely reviewed to identify all patients in whom a CSF puncture was performed for any reason. Inclusion criteria were: all had trauma as cause of injury; in addition either an unstable spinal fracture (and/or dislocation with or without neurologic deficit), or no bony injury with neurologic deficit.
Arch Phys Med Rehabil Vol 75, March 1994
CEREBROSPINAL
294
FLUID
Exclusion criteria were: possible diagnosis of meningitis or encephalitis, CSF leakage to the exterior, surgical fixation of the spine prior to CSF sampling, metastatic bone disease, primary tumors of the nervous system and its associated structures, and possible diagnosis of hysteria as the basis for their neurologic deficit. All laboratory analysis was done at University Hospital, Shaughnessy Site. CSF cell counts were done using a cytospin differential count8 To compensate for the possible contaminating effects of a bloody tap or traumatic hemorrhage on CSF WBC, a corrected count (cWBC) was calculated by dividing the red cell count by 7009 and subtracting this number from the total WBC count (normal = less than or equal to 5 cells/mm cubed”). Negative results were considered as zero. CSF protein levels were considered abnormal when greater than 0.45mg/dL.” Level and type of neurologic injury were classified according to the American Spinal Injury Association (ASIA) criteria based on review of clinical notes in the patients’ charts. Seventy-four patients had spinal punctures. [In all cases, the CSF punctures were done routinely as part of a myelogram. When CSF was sent for analysis, it was done so as part of routine protocol and not out of clinical expectation of abnormal result.] The interval between the injury and the myelogram, in days, was recorded. In 20 patients, no CSF was sent for analysis. In 1 patient, the CSF was sent but the result could not be found. A total of 53 patients therefore had CSF results and were entered into the study. Forty-five had both cell counts and protein levels documented. An additional 6 patients had cell counts available and were included in the analysis. Another 2 patients had only the protein levels measured and were not included in the analysis. Very few patients had glucose levels done and none had simultaneous blood glucose levels.
CELL COUNT,
Travlos
were very rare staff cells and eosinophils. The mean cWBC for all patients was 28/mm3 (SD, 91). Twenty eight spinal punctures were done within 7 days of injury. The mean cWBC for this group was 48.9/mm3 (SD, 118.1). The highest cWBC values were from the specimens obtained within 3.5 days of injury (fig). ri99. Among the 7 patients with an abnormal cWBC level, the range was from 6 to 446cells/mm3. Twenty five spinal punctures were done after day 7. Seventeen of the 25 were done between the first and third weeks. The mean cWBC was 2.4/mm3 (SD = 2.6). Three patients had raised cWBC counts measured at 7, 8, and 17cells/mm’. Differential counts showed lymphs’, polys’, and monocytes predominating in one each of the three patients, respectively. No statistically significant association was found between cWBC and interval between injury and specimen collection (p = 0.24). However, the power of detecting a 50% difference at the 0.05 level was only 21% due to the small number of patients. CSF protein levels were available for 47 patients. In 34 patients (72%) including all 10 in whom the cWBC was elevated, protein levels were abnormal. In all 7 patients sampled within 7 days whose cWBC was elevated, CSF protein levels were raised (mean, 125; SD, 54.3). In all three patients sampled after day 7 whose cWBC was elevated, the mean protein level was 261.3 (SD, 103.1). There was a significant correlation between the presence of elevated protein and elevation of cWBC (p = 0.02). Elevation of CSF protein was not limited to patients with an elevated cWBC, having also been elevated in 22 patients of the 35 with a normal cWBC. The degree of CSF obstruction on the myelograms varied and ranged from none to complete block. The latter was present in 6 patients and was associated with some of the highest CSF protein levels without association with cell count. DISCUSSION
STATISTICAL
ANALYSIS
Results for cWBC and CSF protein were categorized as normal or abnormal as previously defined. For each specimen the interval between injury and collection was categorized as being within the first 7 days or on day eight or later.’ l,” Fisher’s exact test was used to evaluate the possibility of an association between specimen timing and cWBC, as well as between cWBC and CSF protein levels. RESULTS A total of 41 men (77%) and 12 women (23%) met the inclusion criteria. The mean age of the group was 36.4 years (standard deviation [SD] 16.8 years). Cervical injuries made up 60% of the group and the two most common levels of injury were C5 and C6. The time between injury and CSF sampling ranged from 4 hours to 111 days with a median time of 6 days (mean, 15.2; SD, 24.3). CSF WBC results were found in 51 patients. In 32% of specimens analyzed (16 patients) the uncorrected WBC was abnormal with 20% of specimens ( 10 patients) remaining abnormal after correction. The differential cell count revealed a mean polymorphonucleocytes count (polys’) of 57.8 (SD, 33.2), a mean lymphocyte (lymphs’) count of 23.4 (SD, 25. I), and a mean monocyte count of 19.6 (SD, 18.2). There
Arch Phys Med Rehabil Vol75,
March 1994
Infection is a common clinical problem after traumatic SCI. The differential diagnosis of fever is broad and meningitis is a relatively uncommon cause of fever after SCI. However, the potential morbidity and neurologic consequences of meningitis are significant. Several factors make the diagnosis of meningitis after SC1 difficult. Patients are often immobilized in external traction or rigid orthoses. Patients may have spinal pain due to their injury if sensation is intact. Sensory impairment may make pain and the usual clinical signs of meningeal irritation such as Brudzinski’s, Kemig’s, neck stiffness, and rigidity unreliable. Temperature regulation problems are common after high spinal injury and fever may be an unreliable indicator of CNS infection. Mental status changes may reflect the increased use of high dose corticosteroid therapy after acute injury.13 Patients with SC1 are often on broad spectrum antibiotic regimens for fever and culture-positive urinary tract infections or other infections.14 Meningitis, if present, may therefore be partially treated, further masking clinical signs of meningeal irritation. The indications for doing a lumbar puncture are well known to most physicians.’ Yet, despite the unreliability of clinical evaluation for meningitis after SCI, lumbar puncture (LP) is performed relatively infrequently in this population.
CEREBROSPINAL
FLUID
CELL COUNT,
Travlos
295
.
.
. .
Relationship of corrected CSF cell count to days post injury (Y-axis is exponential).
1 +.,L0
!
zd
v
+
7
21
Given that meningitis may occur after SC1 and that clinical evaluation is unreliable, it can be argued that CSF analysis is especially important in the diagnosis of meningitis after SCI. However, few guidelines are available to help the clinician in assessing the significance of changes in CSF after SCI. In the current English literature there are only three reports of cells in the CSF after SCI. PostI and AshbyIh refer to samples taken between 15 days to several years of injury. Frankel” looked at an unusual group of seven patients with an ascending level of injury who at the time of their myelogram had CSF analysis performed. None of these studies addressed the question of the relationship of cell count to acute injury in the general SC1 population. Our study patients were similar in age and sex distribution to the typical population with SC1 when compared with the National SC1 Database with 82% and 18% for gender distribution and 29.7 years for age.’ A somewhat larger percent of the total (60% as compared to the National SC1 Database figure of 54.6%) had cervical injuries. That may have reflected the relatively more frequent need to perform computed tomographic myelography to adequately image the cervical spine compared with lower levels of injury. We have demonstrated that the WBC count may be significantly raised within the first week after injury with a predominance of lymphocytes and polymorphonucleocytes. CSF changes in WBC should therefore be interpreted cautiously early after spinal cord injury. It is less common for the CSF cWBC to be elevated more than 1 week after injury, although this did occur. CSF protein levels were elevated in most of our patients, with the degree of elevation related to the presence of a complete block on the myelogram. Protein levels in the CSF did show a relationship to the raised cell count but not to the normal counts. Protein levels are unlikely to be of value in the diagnosis of possible infection because of this. Balentine conducted experimental SC1 in adult rats and found that by 8 hours of injury, necrosis of the spinal cord was almost complete. Days 3 to 7 were associated with increased phagocytosis and by the second to third week, phagocytosis was nearly complete.” Weirich and coworkers.” using a similar model to Balentine but adding magnetic
40
60 TIME
100
80
120
IN DAYS
resonance imaging (MRI) scans, revealed maximal changes at 48 to 72 hours on MRI, a progression from hemorrhage to necrosis, and eventual cavitation at 1 week. It thus followed that we used 1 week and 3 weeks as reference times to look at the CSF. We believe the observed changes in CSF cWBC after SC1 in our patients demonstrate an inflammatory response to injury early after SCI. We also believe that the raised protein levels reflect inflammation, myelin breakdown, and breakdown of other proteins associated with hemorrhage. With the emergence of MRl as a less invasive alternative to myelography, routine CSF studies as part of a myelogram to image the cord after SC1 have become uncommon. If data were available from other centers, it might be possible to combine it with ours to increase the statistical power and to further evaluate the time of return of the cell count to normal. Although we did not find a statistically significant difference between abnormal analyses taken before and after day 7, one may still exist because the power of detecting a 50% difference with our small study population was only 2 1% at the 0.05 level. We fully acknowledge the shortcomings of a retrospective study such as this. We believe that a prospective design that included serial CSF evaluations in order to characterize the course of CSF white blood cell counts would better address the questions we have raised. However, lumbar punctures can be associated with a number of side effects such as headache. diplopia, backache, paresthesias and motor weakness, and local or diffuse meningeal infection. ‘.” We therefore did not believe that it was ethical, nor in our patients’ interests, to perform such invasive and potentially noxious research. In summary, traumatic SC1 is associated with elevations in cWBC that we believe reflect an inflammatory response to injury. The elevation in cWBC is greatest in the first week after injury. Elevations in cWBC more than 1 week after injury do occur but are uncommon. The white cell differential count consists primarily of lymphocytes and polymorphonucleocytes. Three weeks after injury cell counts should be normal. Protein levels are raised after trauma and should not be used to assess the presence of infection. LP should be part of the septic workup of SC1 patients just as it would be for any other population if meningitis is a consideration.
Arch
Phys
Med
Rehabil
Vol 75, March
1994
CEREBROSPINAL
296 Acknowledgment: The authors for his review of the manuscript.
FLUID CELL COUNT,
would like to thank J.R. Busser, MD.
References 1. Allen RJ. Infectious complications in patients with spinal cord injury. JAMA 1982;248:83-4. 2. Stover SL, Fine PR, Go BK, et al. Spinal cord injury: the facts and figures. University of Alabama at Birmingham, 1986. 3. Sugarman B, Brown D, Musher D. Fever and infection in spinal cord injury patients. JAMA 1982;248:67-70. 4. Anderson JM, Schutt AH. Spinal injury in children. Mayo Clinic Proc 1980;55:499-504. 5. Khella L, Stoner EK. 101 cases of spinal cord injury: 15 year followup study in a large city hospital. Am J Phys Med 1977;56:21-32. 6. Wilcox NE, Stauffer ES. Follow-up of 423 consecutive patients admitted to the spinal cord center, Ranch0 Los Amigos Hospital. 1 Jan-Dee 1967. Paraplegia 1972;10:115-22. 7. Whiteneck GG, Charlifire SW, Silver JR, et al. Mortality. morbidity, and psychosocial outcomes of persons spinal cord injured more than 20 years ago. Paraplegia 1992;30:617-30. 8. Haywood RA. Shipiro MF, Oye RK. Laboratory testing on cerebrospinal fluid. Lancet 1987; January: 1-4.
Arch Phys Med Rehabil Vol75, March 1994
Travlos
4th ed. New York: 9. Adams RD. Victor M. Principals of neurology. McGraw-Hill 1989:10-4. 511-3, 554-62. 10. Gorelick PB, Biller J. Lumbar puncture: techniques, indications and complications. Postgrad Med 1986;79:257-68. Il. Balentine JD. Pathology of experimental spinal cord trauma. Lab Invest 1978;39:236-66. correlation 12. Weirich SD, Cotler HB. Harris JH, et al. Histopathologic of magnetic resonance imaging signal patterns in a spinal cord injury model. Spine 1990;15:630-8. 13. Travlos A, Hirsch G. Steroid psychosis: a cause of confusion on the acute spinal cord injury unit. Arch Phys Med Rehabil 1993;74:312-5. 14. Sugarman B. Fever in recently injured quadriplegic persons. Arch Phys Med Rehabil 1982;63:639-40. 15. Post RM. Goodwin FK, Gordon E. Amine metabolites in human cerebrospinal fluid. Science 1973;179:897-9. 16. Ashby P, Vertier M, Warsh J, Price K. Spinal reflexes and the concentrations of 5.HIAA, MHPG, and HVA in lumbar cerebrospinal fluid after spinal lesions in man. J Neural Neurosurg Psychiatry 1976;39:1191-200. 17. Frankel HL. Ascending cord lesions in the early stages following spinal injury. Paraplegia 1968; scientific meeting: 11 l-l 8. 18. Dripps RD. Vandam LD. Hazards of lumbar puncture. JAMA 1951;147:1118-21.
Cerebrospinal Fluid Cell Count Following Spinal Cord Injury Andrew
Travlos MB, ChB, Hugh A. Anton,
MD, Peter C. Wing, MB, ChB
ABSTRACT. Travlos A, Anton HA, Wing PC. Cerebrospinal fluid cell count following spinal cord injury. Arch Pbys Med Rehabil 1994;75:293-6. l Cerebrospinal fluid (CSF) changes after spinal cord injury (SCI) were evaluated by retrospective analysis of all patients admitted to the Acute Spinal Cord Injury Unit (ASCIU) at University Hospital, Shaughnessy Site in Vancouver, British Columbia. A total of 1,917 admissions occurred during the IO-year study period with 1,151 due to acute trauma. The charts of all patients with SC1 due to trauma were reviewed to identify patients in whom a CSF puncture was performed for any reason. Traumatic SC1 is associated with elevations in the corrected white blood cell count (cWBC) that we believe reflect an inflammatory response to injury. The elevation in cWBC is greatest in the first week after injury. Elevations in cWBC more than 1 week after injury do occur but are uncommon. The white cell differential count consists primarily of lymphocytes and polymorphonucleocytes. Three weeks after injury cell counts should be normal. Protein levels are elevated after trauma and should not be used to assess the presence of infection. LP should be part of the septic workup of SC1 patients just as it would be for any other population if meningitis is a consideration. 0 1994 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and
Rehabilitation The development of dedicated spinal cord injury (SCI) units over the last 20 years has led to improved care after traumatic spinal cord injury. An important component of tertiary or quaternary care is decreased morbidity and mortality, and better diagnosis and treatment of complications. Survival after SC1 has improved significantly during the latter part of this century’.’ but infection remains a major clinical problem. Fever and/or infection occur in up to 67% of patients,’ with urinary tract and respiratory infections being most often responsible.“6 Disease of the respiratory system has been identified as a major cause of death after SC1 and 78.5% of respiratory deaths are associated with infection. Other infections were responsible for 8.8% of all deaths according to the National SC1 Database.2 In a more recent publication by Whiteneck and colleagues,7 disorders of the genitourinary system were the leading cause of death over all and urinary tract infections occurred with an annual incidence of 20%. Clinicians who manage patients with SC1 know that the potential causes for fever are multiple and many patients after investigation are labelled as having pyrexia of unknown origin. Further, the clinical evaluation of fever in such patients is frequently difficult because of their neurologic deficit, impaired thermoregulation, and immobilization in rigid external fixation. Consequently, one’s clinical examination may be unreliable and inadequate. This often leads to either overinvestigation, or underinvestigation with over treatment. Meningitis is a potential source of fever that may be particularly difficult to diagnose clinically in patients with acute From Anton).
the Departments
of Physical
and Orthopedics,
(Dr.
Medicine
Wing)
University
and Rehabilitation Hospital.
(Dr.
Shaughnessy
Twvlos,
Dr.
Site. Vancou-
ver. BC. Canada. Submitted
for publicatton
No comtnerc~al supporting
thrs article
organization Reprint
wth
0
1994
Academy
3Nl.
a benefit
27. 1993.
in the results of the research
upon
the authors
or upon
any
MB,
ChB,
Shaughnessy
Division
of Phytical
Site, Room
A 170-4500
Medicine
and
Oah Street.
Canada.
by the American of Physical
tn revised form May
interest
are associated.
Travlos.
Hospital.
tinancial
confer
the authors
UnivrrGty BC V6H
23, 1993. Accepted
a direct
has or will
which
requests to Andrew
Rehabilitation, Vancouver.
March
party having
Congress
Medicine
(~003.9993/94/7503-OU63$3.00/0
of Rehabilitation
and Rehabilitation
Medicine
and the Amencan
SCIs. The consequences of not diagnosing and treating a patient with meningitis may be devastating. A patient who already has a significant neurologic impairment due to SC1 may be placed at risk for further damage to the central nervous system with consequent increased morbidity. impairment and resultant disability. Our interest in this problem occurred after we saw a patient on our acute SCI unit in whom we were unable to exclude the diagnosis of a partially treated meningitis. The patient’s cerebrospinal fluid (CSF) white blood cell count (WBC) was 10, he was febrile and confused, and no other source for his symptoms could be found. A review of the current literature found little mention of either meningeal infection or inflammation as a possible source of fever in SCI. Nor was there data on the spectrum of CSF cell count after SCI. We believe that the lack of such data is a barrier to making informed clinical decisions. We hypothesized that SC1 might cause bleeding into the CSF coupled with an inflammatory response and resultant increase in the white cell count. The purpose of this retrospective study was to determine if the inflammatory process following traumatic SC1 causes a rise in the CSF white cell count. METHODS CSF changes after SC1 were evaluated by retrospective analysis of all patients admitted to the Acute Spinal Cord Injury Unit (ASCIU) at University Hospital, Shaughnessy Site in Vancouver, British Columbia. The ASCIU is a quarternary referral center for the management of spinal injuries in the province of British Columbia. A total of 1,9 I7 admissions occurred during the study period of April 198 1 to March 1991 with 1,1151 due to acute trauma. The charts of all patients with SC1 due to trauma were completely reviewed to identify all patients in whom a CSF puncture was performed for any reason. Inclusion criteria were: all had trauma as cause of injury; in addition either an unstable spinal fracture (and/or dislocation with or without neurologic deficit), or no bony injury with neurologic deficit.
Arch Phys Med Rehabil Vol 75, March 1994
CEREBROSPINAL
294
FLUID
Exclusion criteria were: possible diagnosis of meningitis or encephalitis, CSF leakage to the exterior, surgical fixation of the spine prior to CSF sampling, metastatic bone disease, primary tumors of the nervous system and its associated structures, and possible diagnosis of hysteria as the basis for their neurologic deficit. All laboratory analysis was done at University Hospital, Shaughnessy Site. CSF cell counts were done using a cytospin differential count8 To compensate for the possible contaminating effects of a bloody tap or traumatic hemorrhage on CSF WBC, a corrected count (cWBC) was calculated by dividing the red cell count by 7009 and subtracting this number from the total WBC count (normal = less than or equal to 5 cells/mm cubed”). Negative results were considered as zero. CSF protein levels were considered abnormal when greater than 0.45mg/dL.” Level and type of neurologic injury were classified according to the American Spinal Injury Association (ASIA) criteria based on review of clinical notes in the patients’ charts. Seventy-four patients had spinal punctures. [In all cases, the CSF punctures were done routinely as part of a myelogram. When CSF was sent for analysis, it was done so as part of routine protocol and not out of clinical expectation of abnormal result.] The interval between the injury and the myelogram, in days, was recorded. In 20 patients, no CSF was sent for analysis. In 1 patient, the CSF was sent but the result could not be found. A total of 53 patients therefore had CSF results and were entered into the study. Forty-five had both cell counts and protein levels documented. An additional 6 patients had cell counts available and were included in the analysis. Another 2 patients had only the protein levels measured and were not included in the analysis. Very few patients had glucose levels done and none had simultaneous blood glucose levels.
CELL COUNT,
Travlos
were very rare staff cells and eosinophils. The mean cWBC for all patients was 28/mm3 (SD, 91). Twenty eight spinal punctures were done within 7 days of injury. The mean cWBC for this group was 48.9/mm3 (SD, 118.1). The highest cWBC values were from the specimens obtained within 3.5 days of injury (fig). ri99. Among the 7 patients with an abnormal cWBC level, the range was from 6 to 446cells/mm3. Twenty five spinal punctures were done after day 7. Seventeen of the 25 were done between the first and third weeks. The mean cWBC was 2.4/mm3 (SD = 2.6). Three patients had raised cWBC counts measured at 7, 8, and 17cells/mm’. Differential counts showed lymphs’, polys’, and monocytes predominating in one each of the three patients, respectively. No statistically significant association was found between cWBC and interval between injury and specimen collection (p = 0.24). However, the power of detecting a 50% difference at the 0.05 level was only 21% due to the small number of patients. CSF protein levels were available for 47 patients. In 34 patients (72%) including all 10 in whom the cWBC was elevated, protein levels were abnormal. In all 7 patients sampled within 7 days whose cWBC was elevated, CSF protein levels were raised (mean, 125; SD, 54.3). In all three patients sampled after day 7 whose cWBC was elevated, the mean protein level was 261.3 (SD, 103.1). There was a significant correlation between the presence of elevated protein and elevation of cWBC (p = 0.02). Elevation of CSF protein was not limited to patients with an elevated cWBC, having also been elevated in 22 patients of the 35 with a normal cWBC. The degree of CSF obstruction on the myelograms varied and ranged from none to complete block. The latter was present in 6 patients and was associated with some of the highest CSF protein levels without association with cell count. DISCUSSION
STATISTICAL
ANALYSIS
Results for cWBC and CSF protein were categorized as normal or abnormal as previously defined. For each specimen the interval between injury and collection was categorized as being within the first 7 days or on day eight or later.’ l,” Fisher’s exact test was used to evaluate the possibility of an association between specimen timing and cWBC, as well as between cWBC and CSF protein levels. RESULTS A total of 41 men (77%) and 12 women (23%) met the inclusion criteria. The mean age of the group was 36.4 years (standard deviation [SD] 16.8 years). Cervical injuries made up 60% of the group and the two most common levels of injury were C5 and C6. The time between injury and CSF sampling ranged from 4 hours to 111 days with a median time of 6 days (mean, 15.2; SD, 24.3). CSF WBC results were found in 51 patients. In 32% of specimens analyzed (16 patients) the uncorrected WBC was abnormal with 20% of specimens ( 10 patients) remaining abnormal after correction. The differential cell count revealed a mean polymorphonucleocytes count (polys’) of 57.8 (SD, 33.2), a mean lymphocyte (lymphs’) count of 23.4 (SD, 25. I), and a mean monocyte count of 19.6 (SD, 18.2). There
Arch Phys Med Rehabil Vol75,
March 1994
Infection is a common clinical problem after traumatic SCI. The differential diagnosis of fever is broad and meningitis is a relatively uncommon cause of fever after SCI. However, the potential morbidity and neurologic consequences of meningitis are significant. Several factors make the diagnosis of meningitis after SC1 difficult. Patients are often immobilized in external traction or rigid orthoses. Patients may have spinal pain due to their injury if sensation is intact. Sensory impairment may make pain and the usual clinical signs of meningeal irritation such as Brudzinski’s, Kemig’s, neck stiffness, and rigidity unreliable. Temperature regulation problems are common after high spinal injury and fever may be an unreliable indicator of CNS infection. Mental status changes may reflect the increased use of high dose corticosteroid therapy after acute injury.13 Patients with SC1 are often on broad spectrum antibiotic regimens for fever and culture-positive urinary tract infections or other infections.14 Meningitis, if present, may therefore be partially treated, further masking clinical signs of meningeal irritation. The indications for doing a lumbar puncture are well known to most physicians.’ Yet, despite the unreliability of clinical evaluation for meningitis after SCI, lumbar puncture (LP) is performed relatively infrequently in this population.
CEREBROSPINAL
FLUID
CELL COUNT,
Travlos
295
.
.
. .
Relationship of corrected CSF cell count to days post injury (Y-axis is exponential).
1 +.,L0
!
zd
v
+
7
21
Given that meningitis may occur after SC1 and that clinical evaluation is unreliable, it can be argued that CSF analysis is especially important in the diagnosis of meningitis after SCI. However, few guidelines are available to help the clinician in assessing the significance of changes in CSF after SCI. In the current English literature there are only three reports of cells in the CSF after SCI. PostI and AshbyIh refer to samples taken between 15 days to several years of injury. Frankel” looked at an unusual group of seven patients with an ascending level of injury who at the time of their myelogram had CSF analysis performed. None of these studies addressed the question of the relationship of cell count to acute injury in the general SC1 population. Our study patients were similar in age and sex distribution to the typical population with SC1 when compared with the National SC1 Database with 82% and 18% for gender distribution and 29.7 years for age.’ A somewhat larger percent of the total (60% as compared to the National SC1 Database figure of 54.6%) had cervical injuries. That may have reflected the relatively more frequent need to perform computed tomographic myelography to adequately image the cervical spine compared with lower levels of injury. We have demonstrated that the WBC count may be significantly raised within the first week after injury with a predominance of lymphocytes and polymorphonucleocytes. CSF changes in WBC should therefore be interpreted cautiously early after spinal cord injury. It is less common for the CSF cWBC to be elevated more than 1 week after injury, although this did occur. CSF protein levels were elevated in most of our patients, with the degree of elevation related to the presence of a complete block on the myelogram. Protein levels in the CSF did show a relationship to the raised cell count but not to the normal counts. Protein levels are unlikely to be of value in the diagnosis of possible infection because of this. Balentine conducted experimental SC1 in adult rats and found that by 8 hours of injury, necrosis of the spinal cord was almost complete. Days 3 to 7 were associated with increased phagocytosis and by the second to third week, phagocytosis was nearly complete.” Weirich and coworkers.” using a similar model to Balentine but adding magnetic
40
60 TIME
100
80
120
IN DAYS
resonance imaging (MRI) scans, revealed maximal changes at 48 to 72 hours on MRI, a progression from hemorrhage to necrosis, and eventual cavitation at 1 week. It thus followed that we used 1 week and 3 weeks as reference times to look at the CSF. We believe the observed changes in CSF cWBC after SC1 in our patients demonstrate an inflammatory response to injury early after SCI. We also believe that the raised protein levels reflect inflammation, myelin breakdown, and breakdown of other proteins associated with hemorrhage. With the emergence of MRl as a less invasive alternative to myelography, routine CSF studies as part of a myelogram to image the cord after SC1 have become uncommon. If data were available from other centers, it might be possible to combine it with ours to increase the statistical power and to further evaluate the time of return of the cell count to normal. Although we did not find a statistically significant difference between abnormal analyses taken before and after day 7, one may still exist because the power of detecting a 50% difference with our small study population was only 2 1% at the 0.05 level. We fully acknowledge the shortcomings of a retrospective study such as this. We believe that a prospective design that included serial CSF evaluations in order to characterize the course of CSF white blood cell counts would better address the questions we have raised. However, lumbar punctures can be associated with a number of side effects such as headache. diplopia, backache, paresthesias and motor weakness, and local or diffuse meningeal infection. ‘.” We therefore did not believe that it was ethical, nor in our patients’ interests, to perform such invasive and potentially noxious research. In summary, traumatic SC1 is associated with elevations in cWBC that we believe reflect an inflammatory response to injury. The elevation in cWBC is greatest in the first week after injury. Elevations in cWBC more than 1 week after injury do occur but are uncommon. The white cell differential count consists primarily of lymphocytes and polymorphonucleocytes. Three weeks after injury cell counts should be normal. Protein levels are raised after trauma and should not be used to assess the presence of infection. LP should be part of the septic workup of SC1 patients just as it would be for any other population if meningitis is a consideration.
Arch
Phys
Med
Rehabil
Vol 75, March
1994
CEREBROSPINAL
296 Acknowledgment: The authors for his review of the manuscript.
FLUID CELL COUNT,
would like to thank J.R. Busser, MD.
References 1. Allen RJ. Infectious complications in patients with spinal cord injury. JAMA 1982;248:83-4. 2. Stover SL, Fine PR, Go BK, et al. Spinal cord injury: the facts and figures. University of Alabama at Birmingham, 1986. 3. Sugarman B, Brown D, Musher D. Fever and infection in spinal cord injury patients. JAMA 1982;248:67-70. 4. Anderson JM, Schutt AH. Spinal injury in children. Mayo Clinic Proc 1980;55:499-504. 5. Khella L, Stoner EK. 101 cases of spinal cord injury: 15 year followup study in a large city hospital. Am J Phys Med 1977;56:21-32. 6. Wilcox NE, Stauffer ES. Follow-up of 423 consecutive patients admitted to the spinal cord center, Ranch0 Los Amigos Hospital. 1 Jan-Dee 1967. Paraplegia 1972;10:115-22. 7. Whiteneck GG, Charlifire SW, Silver JR, et al. Mortality. morbidity, and psychosocial outcomes of persons spinal cord injured more than 20 years ago. Paraplegia 1992;30:617-30. 8. Haywood RA. Shipiro MF, Oye RK. Laboratory testing on cerebrospinal fluid. Lancet 1987; January: 1-4.
Arch Phys Med Rehabil Vol75, March 1994
Travlos
4th ed. New York: 9. Adams RD. Victor M. Principals of neurology. McGraw-Hill 1989:10-4. 511-3, 554-62. 10. Gorelick PB, Biller J. Lumbar puncture: techniques, indications and complications. Postgrad Med 1986;79:257-68. Il. Balentine JD. Pathology of experimental spinal cord trauma. Lab Invest 1978;39:236-66. correlation 12. Weirich SD, Cotler HB. Harris JH, et al. Histopathologic of magnetic resonance imaging signal patterns in a spinal cord injury model. Spine 1990;15:630-8. 13. Travlos A, Hirsch G. Steroid psychosis: a cause of confusion on the acute spinal cord injury unit. Arch Phys Med Rehabil 1993;74:312-5. 14. Sugarman B. Fever in recently injured quadriplegic persons. Arch Phys Med Rehabil 1982;63:639-40. 15. Post RM. Goodwin FK, Gordon E. Amine metabolites in human cerebrospinal fluid. Science 1973;179:897-9. 16. Ashby P, Vertier M, Warsh J, Price K. Spinal reflexes and the concentrations of 5.HIAA, MHPG, and HVA in lumbar cerebrospinal fluid after spinal lesions in man. J Neural Neurosurg Psychiatry 1976;39:1191-200. 17. Frankel HL. Ascending cord lesions in the early stages following spinal injury. Paraplegia 1968; scientific meeting: 11 l-l 8. 18. Dripps RD. Vandam LD. Hazards of lumbar puncture. JAMA 1951;147:1118-21.