Howard SC, Gajjar A, Ribeiro RC, Rivera GK, Rubnitz JE, Sandlund JT, Harrison PL, de Armendi A, Dahl GV, Pui C. Safety of Lumbar Puncture for Children With Acute Lymphoblastic Leukemia and Thrombocytopenia. JAMA. 2000;284(17):2222-2224. doi:10.1001/jama.284.17.2222
Author Affiliations: Departments of Hematology-Oncology (Drs Howard, Gajjar, Ribeiro, Rivera, Rubnitz, Sandlund, and Pui), Biostatistics and Epidemiology (Ms Harrison), Pathology (Dr Pui), and Anesthesiology (Dr de Armendi), St Jude Children's Research Hospital, and Department of Pediatrics, University of Tennessee College of Medicine (Drs Gajjar, Ribeiro, Rivera, Rubnitz, Sandlund, and Pui), Memphis, Tenn; and Department of Pediatrics (Dr Dahl), Stanford University School of Medicine, Stanford, Calif.
Context Patients with thrombocytopenia are at risk for spontaneous or procedure-related
hemorrhage. Whether such patients can safely undergo lumbar puncture (LP)
without prophylactic platelet transfusion is unknown.
Objective To determine whether an association exists between thrombocytopenia
and LP complications among children with acute lymphoblastic leukemia.
Design, Setting, and Patients Retrospective review of the records of 958 consecutive children (median
age, 5.5 years) with newly diagnosed acute lymphoblastic leukemia who were
treated at a pediatric cancer center between February 1984 and July 1998.
Interventions All patients underwent a diagnostic LP followed by a median of 4 LPs
to instill intrathecal chemotherapy.
Main Outcome Measure Serious complications of LP occurring during the remission induction
and consolidation treatment periods (when thrombocytopenia is likely to occur),
defined as any neurologic, infectious, or hemorrhagic problems related to
the procedure, reported by platelet count at the time of the procedure.
Results Of the 5223 LPs evaluated, 29 were performed at platelet counts of 10
× 109/L or less, 170 at platelet counts of 11 to 20 ×
109/L, and 742 at platelet counts of 21 to 50 × 109/L. No serious complications were encountered, regardless of the platelet
count. The 95% confidence interval for the proportion of serious complications
in the 199 patients with platelet counts of 20 × 109/L or
less was 0% to 1.75% and that for the 941 patients with platelet counts of
50 × 109/L or less was 0% to 0.37%.
Conclusions In our study of children undergoing remission induction or consolidation
therapy for acute lymphoblastic leukemia, serious complications of LP were
not observed, regardless of platelet count. Prophylactic platelet transfusion
is not necessary in children with platelet counts higher than 10 × 109/L. Due to the small number of patients in our study with platelet
counts of 10 × 109/L or less, conclusions cannot yet be drawn
for such patients.
For children with leukemia, lumbar puncture (LP) is routinely used to
assess central nervous system involvement and to instill intrathecal chemotherapy.
Children undergoing such procedures frequently have thrombocytopenia. Because
of the possibility of devastating neurologic complications after LP,1- 3 many pediatric oncologists
administer prophylactic platelet transfusions before LP is performed. However,
this practice is controversial: the absolute risk of permanent neurologic
injury is low, thrombocytopenia is not a proven risk factor for permanent
neurologic injury, prophylactic platelet transfusion has not been shown to
reduce the risk of such injury, and transfusion can be complicated by infection
and transfusion reactions and can cause delays in treatment and increase cost.4,5
The indications for transfusion of blood components before invasive
procedures were recently reviewed,6 but the
platelet count considered "safe" for LP is unknown. Neurology textbooks recommend
a minimum platelet count of 30 to 50 × 109/L before LP is
performed,7,8 but these thresholds
have not been established by clinical studies. Complications of LP have been
reported in patients with thrombocytopenia,9
but whether thrombocytopenia itself increases the risk of iatrogenic complications
is unclear. Because prophylactic platelet transfusion is not routinely given
before LP for children with acute lymphoblastic leukemia (ALL) at our center,
we could determine whether an association exists between LP complications
From February 1984 to July 1998, 958 children with newly diagnosed ALL
were treated on 4 consecutive protocols10 at
St Jude Children's Research Hospital, Memphis, Tenn. The 524 boys (55%) and
434 girls (45%) ranged in age from 1 month to 18 years (median, 5.5 years).
Eighty-three percent were white, 13% were black, and 4% were of other races.
All patients received induction chemotherapy that included prednisone, asparaginase,
vincristine sulfate, daunorubicin hydrochloride, etoposide (or teniposide),
and cytarabine with or without methotrexate. In addition to the diagnostic
LP, patients received 2 to 9 (median, 4) LPs during the study period for administration
of intrathecal chemotherapy (methotrexate, hydrocortisone sodium succinate,
and cytarabine). Several types of caregivers performed LPs: pediatric oncologists,
pediatric oncology fellows, pediatric residents, and nurse practitioners.
For each day during induction and consolidation therapy on which a patient
underwent LP, we recorded the platelet count, number of red blood cells per
high-powered microscopic field in the cerebrospinal fluid (CSF), and any platelet
transfusion given. We reviewed progress notes for complications associated
with LP. Review was limited to the remission induction and consolidation treatment
periods because thrombocytopenia seldom occurs beyond these phases of therapy.
All patients were treated at a single institution, and each was seen in the
outpatient clinic at least twice weekly; therefore, it is unlikely that a
serious complication occurred but went unrecorded. Serious complications were
defined as any neurologic, infectious, or hemorrhagic problems that resulted
from LP, excluding headache, nausea, irritability, or somnolence. Traumatic
LP was defined as the presence in CSF of 500 or more red blood cells per high-powered
The 95% confidence intervals (CIs) for the probability of serious complications
at each level of platelet count were calculated using the method of Casella11 as implemented in the StatXact-4 Version 4.0.1 software
package (Cytel Software Corporation, Cambridge, Mass).
During the review period, 958 children underwent 5442 LPs. One patient
was excluded because of early death from sepsis before initiation of intrathecal
therapy and another because intracranial hemorrhage at the time of diagnosis
resulted in grossly bloody CSF at each of 11 LPs performed to relieve pressure.
Another 208 LPs were excluded because no platelet count was determined within
1 day of LP (n = 41) or because no platelet count was documented after administration
of prophylactic platelet transfusion (n = 167). No CSF results were reported
for 14 LPs, but these procedures were included because patients were evaluable
for complications. Therefore, 956 patients underwent 5223 evaluable LPs, of
which 895 were done at diagnosis.
No serious complications were associated with any of the 5223 evaluable
LPs, but this does not mean that the probability of adverse events is zero.
Table 1 shows the upper bounds
for the frequency of potential complications according to platelet count.
For 199 procedures performed at platelet counts at 20 × 109/L
or below, the 95% CI for the proportion of serious complications was 0% to
1.75%. Traumatic LP occurred in 548 procedures (10.5%) but was not associated
with adverse sequelae and, therefore, was not itself considered a complication.
Complications were also not encountered in the 208 LPs excluded from analysis.
This study suggests that serious complications are rarely associated
with LP in children with ALL and thrombocytopenia. Serious nonhemorrhagic
complications of LP include vertebral disk infection or collapse, paraplegia,
cranial nerve dysfunction, spinal nerve root herniation, meningitis, and cerebellar
tonsil herniation.2,3 Serious
hemorrhagic complications of LP include spinal subdural, subarachnoid, and
epidural hematomas,2,12- 17
which have been thought to occur at higher rates in patients with coagulopathy,
those treated with anticoagulant or antiplatelet therapy, those who have experienced
a difficult or traumatic LP, or those with thrombocytopenia. The risk of complications
increases when several risk factors are present.12
In this study, no serious complications occurred after 5223 LPs (95% CI, 0%-0.07%),
895 of which were performed at diagnosis, when platelet counts were 10 ×
109/L or less in 11 cases, 20 × 109/L or less
in 67, and 50 × 109/L or less in 306.
Because spontaneous spinal hemorrhage has been reported in patients
with hemophilia or those treated with anticoagulant or thrombolytic agents,18,19 an increased risk of hemorrhage in
these patients after LP is not surprising. Although coagulopathy occurs in
3% of children at the time of the initial diagnosis of ALL, it is generally
subclinical, and only 6 patients in this study were given fresh frozen plasma
or cryoprecipitate on the day of LP.20 In a
case-control study of 684 adults undergoing LP, Ruff and Dougherty12 compared rates of serious complications in patients
treated with systemic administration of heparin after the procedure with those
of controls not treated with heparin. Serious complications (paraparesis or
severe back or lumbosacral radicular pain lasting for more than 48 hours)
occurred in 6.7% of the heparin group and in 1.8% of the control group. In
the heparin group, the complication rate was higher if the LP was grossly
bloody, if heparin therapy was started within 1 hour after LP, or if concurrent
aspirin therapy was administered. Traumatic LP but not aspirin therapy was
associated with a higher incidence of complications in the control group,
suggesting that platelet dysfunction may not be an independent risk factor
for LP complications.
Thrombocytopenia has been associated with complications of LP in individual
cases.3,9 Owens et al14 reviewed 33 cases of spinal hematoma after LP, of
which 9 had no hemorrhagic risk factors, although 7 of them had a difficult
or bloody LP. Twenty-four patients had 1 or more hemorrhagic risk factors:
13 patients had received anticoagulants; 1 had prolonged bleeding time of
unknown cause; 1 had coagulopathy because of liver disease; 9 had low platelet
counts (1 to 44 × 109/L); 2 had been treated with antiplatelet
medications; and 7 had elevated prothrombin time or partial thromboplastin
time. In addition to these hemorrhagic risk factors, 14 patients had undergone
a difficult or grossly bloody LP. This review suggested that any hemorrhagic
risk factor might predispose patients to LP complications. However, this review
of case reports does not allow determination of the relative risk for complications
in these patients compared with patients with normal hemostasis and normal
Some evidence suggests that LP is safe for patients with mild-to-moderate
thrombocytopenia. No hemorrhagic complications were observed in 24 women with
unsuspected thrombocytopenia (platelet counts of 15 to 99 × 109/L) who received spinal anesthesia21
or in 24 children with ALL and platelet counts below 50 × 109/L
who underwent LP at the time of diagnosis,22
including 5 patients with platelet counts below 20 × 109/L.
Although these small studies preclude any firm conclusion, the fact that 170
procedures were performed at platelet counts of 11 to 20 × 109/L without serious complications in the current study suggests that
platelet transfusion is not necessary for patients with platelet counts above
10 × 109/L. Only 29 evaluable procedures were performed at
platelet counts of 10 × 109/L or less and only 6 of these
at platelet counts of 5 × 109/L or less. Thus, the safety
of LP for patients with profound thrombocytopenia (≤10 × 109/L) has not been proven. Prophylactic platelet transfusion is considered
the standard of care for nonfebrile adults with acute myeloid leukemia and
platelet counts of 10 × 109/L or less.23,24
However, we do not routinely give prophylactic platelet transfusion during
remission induction with prednisone, vincristine, and asparaginase, a combination
that induces a hypercoagulable state.25 Whether
prophylactic platelet transfusion before LP is necessary for patients at diagnosis
with a platelet count of 10 × 109/L or less requires further
No serious LP complications were found in this review of 5223 procedures
performed during remission induction or consolidation therapy for ALL, despite
the fact that 941 procedures were performed with platelet counts of 50 ×
109/L or less. However, platelet counts were 10 × 109/L or less in only 29 instances. We conclude that children with ALL
do not require prophylactic platelet transfusion for LP if the platelet count
is greater than 10 × 109/L.