Subdural hemorrhage - a serious complication post-intrathecal chemotherapy. A case report and review of literature.

We need to have a high index of suspicion for subdural hemorrhage (SDH) post-lumbar puncture in hematological patients given their increased risk and the significant morbidity and mortality associated with SDHs.

Introduction

Subdural hemorrhage (SDH) is a rare but known serious complication of lumbar punctures (LP), resulting in significant morbidity and mortality 1,2. The mechanism of SDH formation post-LP is postulated to be due to intracranial hypotension from cerebrospinal fluid (CSF) leakage from the LP site. This causes traction on bridging subdural veins with subsequent hemorrhage 1. The risk of SDH is increased in patients undergoing intensive chemotherapy for hematological conditions, due to the prolonged periods of profound thrombocytopenia and treatment regimens comprising multiple intrathecal injections. In fact, postmortem studies on bone marrow transplant (BMT) patients have shown an incidence of SDH of 6.66% 3 and 5% in leukemia patients undergoing BMT in ante mortem studies 4. A literature review revealed 50 cases of SDH post-LP in patients undergoing chemotherapy for hematological conditions 4–10. Of these 50, 34 (68%) were bilateral, which is in keeping with the proposed mechanism of SDH formation through intracranial hypotension, implicating the LP as being causative. A summary of these cases is provided in Table1.

Table 1

Summary of published cases of post-LP SDH cases in patients undergoing chemotherapy for hematological conditions.

Study	Case numbers	Patient characteristics	Findings/outcomes
Pomeranz et al. 4	13 of 471 BMT	All leukemia patients	All diagnosed on CT (2 had initial normal CT)
	patients	Age range 9–46 years	9 bilateral SDH
		All had diagnostic	7 required surgical drainage
		LP +/− IT chemo	No long term morbidity/mortality
		5 of 13 SDH patients had
		post-LP headache
Jourdan et al. 5	5 of 86 AML	Age range 33–60 years	All diagnosed on CT 1–15 days post-LP
	patients	All had LP; 4 had IT chemo	2 bilateral SDH
		All had post-LP headache	1 required drainage
			No long term morbidity/mortality
Hentsche et al. 6	3 of 272 BMT	All CML patients	All diagnosed on CT 22–29 days post-LP
	patients	Age range 34–49 years	All bilateral and requiring drainage
		All received IT MTX	No long-term morbidity/mortality
		All had post-LP headache
Colosimo et al. 7	17 of 657 BMT	Age range 25–61 years	13 diagnosed on CT, 4 diagnosed on MRI
	patients	16 had IT MTX, 1 had antecedent	Diagnosed 6–248 days post-LP
		minor head trauma.	11 bilateral SDH
		13 had post-LP headache	4 requiring drainage
			No mortality. 1 with residual neurological deficit
Kannan et al. 8	Case series of 2 SDH	1 with T-cell lymphoma; 1 AML	Both had initially normal CT (18–34 days post-LP),
	in BMT patients	Age range 33–46 years	then later diagnosed on repeat CT (31–38 days post-LP)
		Both had IT MTX	Both bilateral and requiring drainage.
		Both had post-LP headache	No morbidity/mortality from SDH
Openshaw et al. 9	17 of 4812 BMT	Age range 15–65 years	Of the 8 who had LPs:
	patients	8 had LP (7 had IT chemo),	SDH was diagnosed 5 – 112 days post-LP
		3 had post-LP headache	5 hematomas (2 bilateral), 3 hygromas (all bilateral);
		Of the 9 without LPs,2 had	2 required drainage
		antecedent head trauma	No morbidity/mortality from SDH
		54% of SDH patients had LP,	Of the 9 without LP:
		higher than average of all BMT	All 9 hematomas, 4 requiring drainage.
		patients (21%)	2 fatal
Patel et al. 10	3 of 10 patients receiving	All Philadelphia chromosome	2 diagnosed on CT, 1 on MRI
	imatinib + systemic and	positive ALL	Diagnosed 3 days to >3 months post-LP
	IT chemo	Age range 35–47 years	2 bilateral
		All received IT chemo	2 received surgical drainage,1 was not fit for
		2 had post-LP headache	surgery and subsequently died

Report of a Case

We report a case of bilateral SDH in a 73-year-old man with T-cell acute lymphoblastic leukemia who received multiple intrathecal methotrexate (IT MTX) injections as part of his induction chemotherapy (Phase II UK-ALL protocol). He received IT MTX injections on Day 1, 8, and 15. Lumbar puncture on D8 was noted to be a difficult procedure with multiple passes. Post-LP headache was first reported on D9, initially intermittent then becoming a mild persistent occipital headache. On D13, he reported mild altered sensation in bilateral feet and in his right 5th finger, without any other neurological signs or symptoms. The headache resolved on D14 and he received D15 IT MTX. Of note, platelet count on D15 was 26 × 109/L, below the standard practice threshold of 50 × 109/L for LPs. This low platelet count was not replaced due to an oversight by the treating team. This procedure was straightforward with a single clean pass, without any immediate complications. The platelet count was more than 50 × 109/L on all earlier LPs (D1 and D8). On D19 the headache returned without any new neurological signs or symptoms. On D22 computed tomography (CT) imaging revealed bilateral subacute frontoparietal hematomas measuring 10 mm on the left, and 8 mm on the right, with associated local mass effect. Neurosurgical consult was obtained and the decision made for conservative management, based on his stable clinical status and the subdural hematoma size that was not greater than 10 mm with no midline shift. His platelet count was replaced to above 50 × 109/L and no further LPs were performed. Serial imaging showed no progression of the subdural hematomas and the patient remains well 11 months post-SDH with no residual symptoms. The oversight of performing an LP without recognizing his severe thrombocytopenia was noted in an internal incident report as part of our risk reduction program. Suggestions to reduce the risk of future similar incidents include emphasizing checking platelet count on the lumbar puncture protocol and a reminder via internal memo to medical staff regarding this risk.

Discussion

Pomeranz et al. 4 and Kannan et al. 8 both report patients with initially normal imaging (CT) up to 34 days post-LP that are later diagnosed with SDH up to 38 days post-LP, suggesting that SDH could occur weeks after LP. This is in keeping with the presumed mechanism of ongoing CSF leak post-LP causing SDH formation over a period of time, suggesting that the platelet count post-LP is as important as the platelet count during the LP itself. This is particularly relevant to our case as the thrombocytopenia on the D15 LP would otherwise be assumed to be the sole cause of the SDH. Furthermore, the patient reported post-LP headache and neurological symptoms prior to the D15 LP, suggesting the SDH may have already occurred. Unfortunately, while prolonged thrombocytopenia and coagulopathies are known risk factors 7, the long period of potential SDH formation makes it impractical to attempt to maintain a threshold platelet count over this period.

Brain CT revealing bilateral subacute frontoparietal subdural hematomas, measuring 10mm on the left and 8mm on the right, with associated local sulcal effacement.

In stratifying risk of SDH post-LP, presence of headache is one of the most important factors. Colosimo et al. 7 reported that out of 19 patients with headache post-LP, 14 had SDH (73.7%), compared with three of 175 (1.7%) patients without headache. Furthermore, 33 of 50 (66%) of published post-LP SDH cases reported post-LP headache. Another possible risk factor is IT MTX, which appears to increase risk compared to diagnostic LP 6,7, though it is difficult to compare with other intrathecal chemotherapy due to low case numbers.

Our case illustrates the importance of having a high index of suspicion for SDH in this patient group. Importantly, suspicion must remain high even if normal brain imaging is performed days-to-weeks post-LP. The most important risk factor is post-LP headache 7, with the majority of patients (73%) presenting with this. While post-LP headache is common 2, features that should arouse suspicion include persisting or worsening headache post-LP and neurological symptoms.

Conflict of Interest

None declared.