Lumbar puncture in thrombocytopenic children


The risk of spontaneous- or procedure-related haemorrhage is increased in patients with thrombocytopenia. Lumbar puncture is routinely used to assess central nervous system involvement and to give intrathecal chemotherapy in children with acute leukemia. To reduce the risk of haemorrhage and neurological complications the British Committee for Standards in Haematology (BCSH) recommends the use of prophylactic platelet transfusions if the platelet count is less than 40 × 109/l in children (BCSH 2004) and if less than 50 × 109/l in adults (BCSH 2003). Spinal haemorrhage is rare and in both guidelines it is acknowledged that there is a lack of evidence to support the recommendations. The evidence for prophylactic transfusions is based on case reports of spinal haemorrhage following lumbar puncture. Most occur at platelet counts below 40 × 109/l, but spinal haemorrhage has been described at higher platelet counts and the risk is higher if there is coexistent coagulopathy, traumatic taps, rapidly falling platelet counts and hyperleucocytosis (Edelson et al, 1974). Spinal haemorrhage after lumbar puncture can also occur in the absence of any risk factors, including thrombocytopenia, but the incidence is unknown. Platelet transfusions have not been shown to reduce the risk of complications and there is evidence that lumbar punctures can be performed safely even in children with severe thrombocytopenia (Howard et al, 2000). In this study, there were no complications in 5223 lumbar punctures, of which 941 were performed at a platelet count of less than 50 × 109/l and 170 at a platelet count of 11–20 × 109/l. At the Sheffield Children's Hospital, prophylactic transfusions are not routinely given, irrespective of the platelet count, and we retrospectively investigated the occurrence of complications related to the first diagnostic and therapeutic lumbar puncture in children with newly diagnosed acute lymphoblastic leukaemia (ALL). In the period between 1989 and 2003 there were 226 ALL patients; 135 had a platelet count of less than 50 × 109/l, of which 129 had a lumbar puncture (72 without platelet transfusion). The patients were divided into three groups based on their platelet count. Group 1; <10 × 109/l platelets, group 2; 11–20 × 109/l platelets, group 3; 21–50 × 109/l platelets. The blood bank records and case notes were checked for a platelet transfusion, lumbar puncture and complications within 72 h following the lumbar puncture. In 56% of the patients transfused with a platelet count between 21 and 50 × 109/l and in 20% with a platelet count of 11–20 × 109/l, the reason for the transfusion was unclear on retrospective review. Table I shows the number of lumbar punctures in each group and the occurrence of a traumatic tap. There were no complications in any group although one patient had severe lower back pain requiring opiates for 48 h (platelet count 30 × 109/l and no transfusion). In five the outcome is unknown. Thirty (23%) had more than 10 red cells/μl in the cerebrospinal fluid (CSF) sample. Our survey is in keeping with the study of Howard et al (2000), and does not support the routine use of prophylactic platelet transfusions. Complications are rare and platelet transfusions can be complicated by infections, transfusion reactions and platelet refractoriness. The number of patients with a platelet count of less than 10 × 109/l is small and this group may benefit from transfusion. It would be useful to further investigate the incidence of complications related to lumbar punctures in a general paediatric population to establish a baseline risk and we propose a study through the British Paediatric Surveillance Unit.

Table I.  Number of lumbar punctures (LP) per group and number of red cells /μl CSF.
Platelet count <10/μl 10–100/μl 100–1000/μl >1000/μlOutcome unknownTotal number of LP
Group 1 (<10 × 109/l)
 Total LP14511122 (100%)
 LP without  platelets511119 (41%)
Group 2 (10–20 × 109/l)
 Total LP32432142 (100%)
 LP without  platelets174  122 (52%)
Group 3 (21–50 × 109/l)
 Total LP5363 365 (100%)
 LP without  platelets3632  41 (60%)
Total number of LPs9915735129 (100%)