Lumbar Puncture and Subdural Hygroma and Hematomas in Hematopoietic Cell Transplant Patients

ORIGINAL ARTICLE Lumbar puncture and subdural hygroma and hematomas in hematopoietic cell transplant patients

H Openshaw1, JA Ressler2 and DS Snyder3

1Department of Neurology, City of Hope National Medical Center, Duarte, CA, USA; 2Division of Diagnostic Radiology, City of Hope National Medical Center, Duarte, CA, USA and 3Division of Hematology and Hematopoietic Cell Transplantation, City of Hope National Medical Center, Duarte, CA, USA

We reviewed records ofhematopoietic cell transplantation transplantation (HCT). Prevalence varies from 0 to 12% in (HCT) patients seen over the past 10 years who had head autopsy series,1,2 and up to2.6% in retrospectiveclinical scan documentation ofsubdural fluid collections. A total reports.3 Subdurals can be easily overlooked early in the of17 patients were identified: 13 with allogeneic and 4 transplant course. They may present with reduced sensor- with autologous HCT (0.71% ofallogeneic and 0.13% of ium without abnormal motor signs or they may present autologous HCT patients seen in this time interval). with nausea, vomiting and headache, masquerading as Although less than 20% ofHCT patients have lumbar toxicity from the preparatory regimen. puncture, 8 ofthe 17 subdural patients had lumbar Intrathecal therapy as part of the preparatory regimen puncture. The lumbar puncture was done 5–112 days has been suggested as a risk factor for subdural hemato- (median 46 days) before subdural detection. Acute mas.3,4 In a retrospective study, 17 subdural hematomas lymphocytic leukemia was the diagnosis in five ofthese were found in 657 transplant patients in Australia, and 16 eight; whereas, either acute myelogenous leukemia or of the 17 subdural patients had received intrathecal myelodysplasia was the diagnosis in seven ofthe nine methotrexate.3 An association of subdural hematomas patients without lumbar puncture. In the patient group and lumbar puncture has also been reported in nontrans- with lumbar puncture, subdurals were diagnosed earlier plant patients with leukemia.5 after HCT (median 25 days versus 5 months in the patient In the present review, we describe the known association group without lumbar puncture) and were more often of lumbar puncture, intracranial hypotension and subdural hygromas (37.5 versus 0%). These results support the formation; and we determine how often subdurals in HCT suggestion oflumbar puncture or intrathecal therapy as a patients are associated with lumbar puncture. The term risk factor for subdurals. The presumptive mechanism subdural in this report is used to include nonhemorrhagic involves lumbar cerebrospinal leak, low intracranial fluid collections (hygromas) as well as hemorrhagic fluid pressure, downward displacement ofthe brain, cerebro- collections (hematomas). spinal fluid accumulation into the inner dural layers of the cerebral convexities (hygromas) and bleeding into these fluid collections (hematomas). Methods Bone Marrow Transplantation (2008) 41, 791–795; doi:10.1038/sj.bmt.1705971; published online 4 February 2008 Medical records of patients admitted to the Hematopoietic Keywords: intracranial hypotension; post-lumbar puncture Cell Transplantation Unit at the City of Hope National headache; subdural hygroma; subdural hematoma Medical Center between 1996 and mid 2006 were reviewed for diagnosis of subdural hygroma or hematoma within 25 months of transplant. Brain imaging and medical records were reviewed and the following aspects recorded: patient’s age and sex, hematological diagnosis, the type of trans- Introduction plantation, presence or absence of lumbar puncture for diagnosis or intrathecal therapy, neurological symp- Subdural hematomas are a well-recognized, infrequent but toms at the time of the diagnostic head scan, time interval serious neurological complication of hematopoietic cell of subdural detection post transplant (and, if relevant, post- lumbar puncture), need for neurosurgery and ultimate course of the subdural.

Correspondence: Dr H Openshaw, Department of Neurology, City of Hope National Medical Center, 1500 East Duarte Road, Duarte, Results CA 91010-3000, USA. E-mail: [email protected] Received 2 March 2007; revised 20 November 2007; accepted 26 Table 1 summarizes information on 17 patients seen November 2007; published online 4 February 2008 between 1996 and mid 2006 whohad subdurals after Lumbar puncture and subdural hygroma and hematomas H Openshaw et al 792 HCT (13 allogeneic and 4 autologous). A total of 1838 none of the patients in the other group. Subdurals were allogeneic and 2974 autologous HCT were done for diagnosed earlier in the transplant course in patients with hematological malignancies at the City of Hope National lumbar puncture (median 25 days post-HCT) than patients Medical Center in this time interval. The subdural without lumbar puncture (median 5 months post-HCT). complication rate, therefore, was 0.71% in allogeneic and Minor head trauma soon after HCT was a feature in two 0.13% in autologous HCT. As shown in the two columns patients in the non-lumbar puncture group; with these two of Table 1, just less than half of subdural patients had patients excluded, the median time of subdural diagnosis in lumbar puncture. The age range was similar in the two the non-lumbar puncture group would be 11 months. On groups, there was male predominance in the group without the basis of head scan characteristics, three patients in the lumbar puncture, and there was the expected difference in lumbar puncture group had hygromas; whereas, all patients hematological diagnoses (predominance of acute lympho- in the non-lumbar puncture group had hematomas. The cytic leukemia in lumbar puncture group and acute median platelet count on the day the subdural was detected myelogenous leukemia and myelodysplasia in non-lumbar was lower in the non-lumbar puncture group. The main puncture group). Leptomeningeal disease was present in neurological problem leading to the head scan which ﬁve of the eight patients in the lumbar puncture group and diagnosed the subdural in the lumbar puncture group was confusion and reduced sensorium in four patients, headache in three patients (but with characteristics of low- pressure headache in only one of these three) and seizure in Table 1 HCT patients with subdural hematomas or hygromas one patient. Presentation was similar in the non-lumbar Lumbar puncture No lumbar puncture puncture group but none of the three headache patients had characteristics of low-pressure headache (that is, headache Patients 8 9 on assuming the upright position). Overall, patients in the Male/female 4/4 7/2 non-lumbar puncture group did worse: four of the nine Age (median) 27–60 (46) 15–65 (51) patients required surgery, twopatients died withoutsurgery as a consequence of their subdural, 14 and 44 days after Diagnosis subdural detection (both patients had recurrent hemato- AML 0 4 ALL 5 0 logical disease post-HCT) and only one patient survived CML 2 1 to resolve the subdural without surgery. In comparison, MD 0 3 there were nofatal subdurals in the lumbar puncture NHL 1 1 group, four patients resolved the subdurals without surgery, tworequired surgery and twosuccumbed to Leptomeningeal disease 5 0 hematological disease recurrence 4 and 69 days after Transplant subdural diagnosis. Allogeneic 7 6 Table 2 gives information on the eight patients who had Autologous 1 3 lumbar puncture, listed in order of the time post-HCT that Head trauma 02subdural was diagnosed. As indicated in the fourth column, all patients had the subdural diagnosed after the lumbar Time post-HCT of subdural diagnosis puncture. Patient 1 had small bilateral subdural hematomas Range (median) 0 day—9 months 13 days—25 months diagnosed 1.5 weeks prior to preparatory therapy for (25 days) (5 months) transplant. The subdural increased in size at the time of this o30 days 5 2 o60 days 6 3 patient’s death from disease recurrence 51 days after HCT. Four other patients (Patients 2, 3, 5 and 6) had lumbar Platelet count per ml (median) 5–101 k (46 k) 10–36 k (20 k) punctures prior to HCT, 14–62 days (median 27 days as shown in column 4). Three patients had lumbar puncture Type of subdural Hematoma 5 9 after HCT: on day þ 5 in Patient 4 as ongoing treatment of Hygroma 3 0 leptomeningeal disease, on day þ 155 in Patient 7 for leptomeningeal disease recurrence after HCT and on day Clinical presentation þ 158 in Patient 8 because of symptoms of meningoence- Headache 3 3 phalitis. As shown in Table 2, subdurals were diagnosed Seizure 1 3 Encephalopathy 4 3 5–112 days after the last lumbar puncture (median 46 day). All three patients with head scan characteristic of hygromas Outcome subdurala had bilateral lesions; whereas only two of the ﬁve patients Resolve 4 1 with hematomas had bilateral lesions. Magnetic resonance Neurosurgery 2 4 Present terminally 2 2 imaging (MRI) scans and clinical synopsis of the two Fatal 0 2 patients whorequired neurosurgeryare presented in Figures 1a and b. The clinical course of Patient 8 presented Abbreviations: ALL ¼ acute lymphocytic leukemia; AML ¼ acute myelo- in the legend toFigures 1c and d was typical forrecurring genous leukemia; CML ¼ chronic myelogenous leukemia; HCT ¼ hemato- post-spinal tap headache; and her MRI showed changes of poietic cell transplantation; MD ¼ myelodysplasia; NHL ¼ non-Hodgkin’s low intracranial pressure (including subdural hygroma, lymphoma. aPresent terminally, subdural present at time of death from complications leptomeningeal enhancement and downward displacement of hematological malignancy; fatal, subdural as cause of death. of the brain).

Bone Marrow Transplantation Lumbar puncture and subdural hygroma and hematomas H Openshaw et al 793 Table 2 Subdurals after lumbar puncture in HCT patients

Patient Age/sex Diagnosis Interval between Subdural type Presenting Time subdural detected Outcomeb LP and HCTa features After HCT After last LP

1 31/M CML 22 days before HCT Hematoma Headache 0 5 days Terminal day+56 CNS disease (bilateral) persistent CML 2 60/F ALL 14 days before HCT Hematoma Confusion 17 days 31 days Terminal day+53 CNS disease (unilateral) ksensorium acute GVHD, graft failure 3 45/M ALL 18 days before HCT Hygroma Confusion 18 days 36 days Resolve NoCNS disease (bilateral) ksensorium 4 48/M NHL 5 days after HCT Hygroma Confusion 22 days 17 days Neurosurgery CNS disease (bilateral) ksensorium 5 42/F ALL 62 days before HCT Hematoma Headache 28 days 90 days Neurosurgery NoCNS disease (unilateral) 6 47/F ALL 35 days before HCT Hematoma Confusion 55 days 90 days Resolve CNS disease (unilateral) ksensorium 7 34/F ALL 155 days after HCT Hematoma Seizure 7 months 55 days Resolve CNS disease (bilateral) 8 27/F CML 158 days after HCT Hygroma Low-pressure 9 months 112 days Resolve NoCNS disease (bilateral) headache

Abbreviations: ALL ¼ acute lymphocytic leukemia; CML ¼ chronic myelogenous leukemia; CNS ¼ central nervous system; HCT ¼ hematopoietic cell transplantation; LP ¼ lumbar puncture; NHL ¼ non-Hodgkin’s lymphoma. aHCT is day of cell infusion (day 0). bTerminal, subdural present at time of death from complications of hematological malignancy; fatal, subdural as cause of death.

Discussion intothe dural venoussinuses. 8 CSF then leaks intothe inner dural layers producing hygromas that are usually Of 1838 allogeneic transplant patients done at the City of bilateral. The subdural fluid collections cause the veins that Hope National Medical Center between 1996 and mid bridge the brain and venous sinuses to be stretched, and 2006, 379 patients (21%) had hematological diagnoses for subdural hematomas result when these veins bleed. Low- which intrathecal therapy is routinely used: 364 patients pressure headaches (post-spinal headaches) may provide a with ALL and 15 patients with AML M4 or M5. If warning of the early part of this process. Low-pressure subdurals occurred randomly post-HCT, one would expect headache occur in about one of four patients after about one-fifth of the subdural patients to have a history of diagnostic lumbar punctures.9 Head or spine pain usually lumbar puncture. The observed frequency of 54%—7 of the of severe degree develops on standing and promptly 13 allogeneic transplant patients in Table 1 had lumbar resolves or significantly diminishes on lying down. punctures—suggests that lumbar puncture is a risk factor Our review did not include an analysis of patients for subdurals. Also consistent with lumbar puncture as a undergoing lumbar puncture who did not develop subdur- risk factor is the earlier presentation of subdurals in the als, but we estimate the risk of symptomatic subdurals to be lumbar puncture group (median day 25 post-HCT com- 1–2% after lumbar puncture. This estimate is based on 364 pared to 5 months in the non-lumbar puncture group). This allogeneic HCT done at City of Hope for ALL in the 10.5 timing is explained by the observed short interval between years of this review, the assumption that these patients will the last lumbar puncture and subdural detection in most have had intrathecal therapy or prophylaxis at some point patients (median 46 days) and the fact that six of eight prior to transplant, and documented subdurals in only five patients in Table 1 had their lumbar puncture for ALL patients post transplant (1.3%). A prospective study intrathecal therapy before or at the time of conditioning by Staudinger et al.,10 however, suggests a much higher risk therapy or in the case of Patient 4 (Table 2) on day þ 5. of asymptomatic subdurals after lumbar puncture. Brain Our results, therefore, are confirmatory of the report of computed tomography (CT) scans were done before HCT Colosimo et al.3 suggesting that lumbar puncture for and on day 30 in 50 consecutive patients, 32 of whom had intrathecal therapy is a risk factor for subdural formation intrathecal methotrexate as part of their conditioning in transplant patients regimen. Nine patients developed subdural hygromas, all Since seven of the eight patients presented in Table 2 had in the intrathecal methotrexate group (giving an incidence intrathecal therapy, our data do not provide an answer of 28% in this group). Data from the retrospective report whether it is intrathecal therapy or merely the lumbar of Colosimo et al.3 alsosuggest a significant risk oflumbar puncture that is the risk factor. Although rare, there are puncture. For example, subdural hematomas were reported case reports of subdural hematomas developing after in 12 of 155 patients (8%) who received intrathecal routine diagnostic lumbar punctures and lumbar punctures methotrexate as part of BU and CY conditioning compared for spinal anesthesia or myelography.6,7 The mechanism to 0 of 71 conditioned with BU/CY alone; and subdural involves intracranial hypotension, downward displacement hematomas were reported in 2 of 12 patients (17%) who of the brain and avulsion of arachnoid granulations, the received intrathecal methotrexate as part of CY and TBI structures involved in cerebrospinal fluid (CSF) absorption conditioning compared to 1 of 112 (1%) conditioned with

Bone Marrow Transplantation Lumbar puncture and subdural hygroma and hematomas H Openshaw et al 794 a b

c d

Figure 1 (a) T1-weighted postgadolinium magnetic resonance imaging (MRI) of Patient 4, a 48-year-old man with human immunodeficiency virus (HIV)- associated non-Hodgkin’s lymphoma and leptomeningeal disease who underwent two lumbar punctures for intrathecal therapy the week before autologous hematopoietic cell transplantation (HCT) and one on day 5 post-HCT. Two weeks after transplantation there was confusion and reduced sensorium. MRI showed bilateral subdural fluid collections (arrows), obliteration of cortical sulci and balanced mass effect. After bilateral surgical drainage of the subdural fluid collections, there was improvement in sensorium. He had no neurological sequelae after surgery. (b) T1-weighted MRI of Patient 5, a 42-year-old woman with ALL who developed headache 2–3 weeks after allogeneic transplant. MRI showed a large, extensive left-sided subdural hematoma with a 1.5 cm midline shift. Platelet count was 27 000 per ml on the day of the MRI. She was given mannitol and underwent emergency evacuation of the subdural hematoma on day 28 after transplant. There were no neurological sequelae. Her last cerebrospinal fluid (CSF) examination with intrathecal methotrexate therapy was 90 days prior to detection of the subdural. She had a prolonged post-spinal headache after intrathecal therapy. (c, d) T1-weighted postgadolinium brain MRIs on the day prior to lumbar puncture (c) and 3.75 months later (d) of Patient 8, a 27-year-old woman with CML who had a matched unrelated donor transplant. Lumbar puncture was done for suspected viral meningomyelitis 5 months after HCT. A post-spinal, postural headache occurred, persisted for 2 days and then resolved. However, 3.75 months later, a postural posterior and vertex headache of severe degree returned on a daily basis. Brain MRI (d) showed gadolinium enhancement of pachymeninges (arrowhead) and bilateral small dural fluid collections over the cerebral convexities (small arrow), not present on earlier MRI. There was also obliteration of cortical sulci, smaller lateral ventricles and compressed suprasellar cistern (large white arrow) with downward displacement of the pituitary stalk (bright signal in suprasellar cistern). Headache resolved after a series of three lumbar epidural blood patches of 10–20 ml each.

CY and TBI alone.3 Leptomeningeal disease was present in show diffuse meningeal gadolinium enhancement, down- ﬁve of our eight lumbar puncture patients with subdurals ward displacement of the cerebellar tonsils, enlargement of and in none of the nine subdural patients reported by the pituitary, engorgement of the venous sinuses and Staudinger et al.10 whoreceived intrathecal methotrexate. decrease in the size of the lateral ventricles.12,13 Examples There is noevidence that the risk ofsubdurals is greater in of these changes from patients in our report are shown in patients with leptomeningeal disease compared to those Figure 1. receiving intrathecal therapy prophylactically. It is certainly possible that lumbar puncture was In their report of two HCT subdural patients with post- coincidental in some of the patients in Table 2, particularly lumbar puncture headache, Kannan et al.11 emphasized the those with longer intervals between the tap and subdural relative insensitivity of brain CT scans to detect subdurals detection. However, it is well known that low-pressure early in the course. Brain MRI scans are more sensitive in headache can occur or recur several weeks after lumbar detecting hygromas. In addition, brain MRI scans in puncture. Patient 8 in Table 2 (Figures 1c and d) provides patients with sustained intracranial hypotension may an example of a delayed, recurrent low-pressure headache

Bone Marrow Transplantation Lumbar puncture and subdural hygroma and hematomas H Openshaw et al 795 that resolved after a lumbar epidural blood patch, a References procedure in which an autologous blood sample is injected in the epidural space at the level of the earlier lumbar 1 Bleggi Torres LF, de Medeiros BC, Werner B, Neto JZ, Loddo puncture toseal the dural leak. Since low-pressurehead- G, Pasquini R et al. Neuropathological findings after bone aches can be delayed, it is difficult toknowprecisely marrow transplantation: an autopsy study of 180 cases. how long after lumbar puncture the patient is at risk for Bone Marrow Transplant 2000; 25: 301–307. low central nervous system (CNS) pressure. Generally, 2 Mohrmann RL, Mah V, Vinters HV. Neuropathologic an epidural blood patch has a success rate over 75% for findings after bone marrow transplantation: an autopsy study. post-spinal headache.14 Hum Pathol 1990; 21: 630–639. Subdural hygromas in hematological malignancy pa- 3 Colosimo M, McCarthy N, Jayasinghe R, Morton J, Taylor K, Durrant S. Diagnosis and management of subdural haemato- tients are particularly troublesome because of the increased ma complicating bone marrow transplantation. Bone Marrow risk of bleeding from thrombocytopenia. Standard proce- Transplant 2000; 25: 549–552. dure in our transplant unit is to keep platelet counts above 4 Hentschke P, Hagglund H, Mattsson J, Carlens S, Lonnqvist 20 000 per ml post-HCT. When subdurals are identified, B, Ljungman P et al. Bilateral subdural haematomas following management involves an increase in platelet transfusion lumbar puncture in three haematopoietic stem cell transplant parameters above 50 000–75 000 per ml; and even with recipients. Bone Marrow Transplant 1999; 24: 1033–1035. optimum management the size of the subdural may 5 Jourdan E, Dombret H, Glaisner S, Miclea JM, Castaigne S, increase, as in Patient 1, the patient whounderwent HCT Degos L. Unexpected high incidence of intracranial subdural despite having a recognized subdural hematoma. Other haematoma during intensive chemotherapy for acute myeloid Br J Haematol treatment approaches for a lumbar CSF leak, including leukaemia with a monoblastic component. 1995; 89: 527–530. continuous epidural saline infusion or epidural injection of 6 Gaucher DJ, Perez Jr JA. Subdural hematoma following 15,16 dextran or fibrin glue, are infrequently used and may lumbar puncture. Arch Intern Med 2002; 162: 1904–1905. carry increased risk in HCT patients. 7 Vos PE, de Boer WA, Wurzer JA, van Gijn J. Subdural Intracranial hypotension adversely impacts intrathecal hematoma after lumbar puncture: two case reports and review therapy. If subdurals are recognized, the planned intrathe- of the literature. Clin Neurol Neurosurg 1991; 93: 127–132. cal course may be deferred and the patient undertreated. 8 Schutta H. Diseases of the dura mater. In: Joynt RJ and Typically, patients with CSF leak and intracranial hypo- Griggs RC (eds). Clinical Neurology.vol 4 Lippincott Williams tension have low manometer opening pressures; and & Wilkins: Philadelphia, 1998, pp 1–137. traumatic taps may occur because the spinal needle is 9 Strupp M, Schueler O, Straube A, Von Stuckrad-Barre S, inadvertently advanced beyond the collapsed subarachnoid Brandt T. ‘‘Atraumatic’’ Sprotte needle reduces the incidence of post-lumbar puncture headaches. Neurology 2001; 57: 2310–2312. space. There is also a greater likelihood of injected drug 10 Staudinger T, Heimberger K, Rabitsch W, Schneider B, going into the epidural space; and because the direction of Greinix HT, Nowzad S et al. Subdural hygromas after bone CSF flow is reversed, drug that does reach the lumbar marrow transplantation: results of a prospective study. subarachnoid space tends to be lost through the leak rather Transplantation 1998; 65: 1340–1344. than distributed throughout the neuroaxis. Over time, 11 Kannan K, Koh LP, Linn YC. Subdural hematoma in two subdural hygromas may become membrane bound8 and as hematopoietic stem cell transplant patients with post-dural in Patient 4 (Figure 1a), the hygromas may produce mass puncture headache and initially normal CT brain scan. Ann effect on the cerebral hemispheres requiring surgical Hematol 2002; 81: 540–542. decompression. Lumbar puncture for intrathecal therapy 12 Fishman RA, Dillon WP. Dural enhancement and cerebral Neurology in these patients with cerebral mass effect may lead tofatal displacement secondary to intracranial hypotension. 1993; 43: 609–611. brain herniation. 13 Mokri B, Piepgras DG, Miller GM. Syndrome of orthostatic Recommendations to prevent low-pressure headache headaches and diffuse pachymeningeal gadolinium enhance- after lumbar puncture include small needle size, placement ment. Mayo Clin Proc 1997; 72: 400–413. of the needle bevel parallel to the dural fibers (that is, 14 Safa-Tisseront V, Thormann F, Malassine P, Michel H, parallel to the long axis of the spine) and replacement of the Riou B, Coriat P et al. Effectiveness of epidural blood patch stylet before withdrawing the needle.17 In one study using in the management of post-dural puncture headache. J Am Soc 22-gauge needles, the frequency of post-lumbar puncture Anesthesiol 2001; 95: 334–339. headaches was reduced from 24.4 to 12.2% by a pencil- 15 Aldrete JA. Persistent post-dural-puncture headache treated point or non-cutting needle (such as a Sprotte needle) with epidural infusion of dextran. Headache 1994; 34: 265–267. rather than the conventional cutting needle (Quincke 16 Crul BJ, Gerritse BM, van Dongen RT, Schoonderwaldt HC. 9 Epidural fibrin glue injection stops persistent postdural needle). If lumbar puncture prior to high-dose therapy is puncture headache. Anesthesiology 1999; 91: 576–577. unavoidable, it is advisable to use a 22-gauge pencil point 17 Armon C, Evans RW. Addendum to assessment: prevention of needle and consider an epidural blood patch early in the post-lumbar puncture headaches: report of the Therapeutics course of a low-pressure headache, before thrombocyto- and Technology Assessment Subcommittee of the American penia precludes a blood patch. Academy of Neurology. Neurology 2005; 65: 510–512.

Bone Marrow Transplantation