British Committee for Standards in Haematology, Blood Transfusion Task Force (2003) produced guidelines suggesting a platelet count of ≥50 × 109/l to safely proceed with lumbar puncture (LP). This has led us to review our institute's policy of a platelet count of ≥30 × 109/l as a safety cutoff for LP procedures and administration of intrathecal chemotherapy. There is obvious concern because a low platelet count may be associated with devastating neurological complications due to spinal bleeding. Furthermore, on presentation if a traumatic tap occurs, leukaemic cells circulating in the blood may be introduced into the cerebrospinal fluid (CSF), thereby possibly worsening the patient's prognosis (Gajjar et al, 2000).
Howard et al (2000) reported a large retrospective analysis of 4309 LPs performed in 959 children, including 941 procedures with platelet count of <50 × 109/l with no neurological haemorrhagic complications. This article recommended that a platelet count of 10 × 109/l or higher would be adequate to perform routine LP. The authors (Howard et al, 2002), found that platelet counts of below 100 × 109/l were associated with increased risk of traumatic or bloody tap, but this risk was not different across platelet categories once the platelet count reduced below this threshold.
Our aim in this retrospective cohort study was to review the available literature, ensure the safety of our practice and determine if the platelet count had any influence on the rate of traumatic LP performed. We report a review of 54 consecutive paediatric patients, 29 male and 25 female, age range 3 months to 17 years (median 7 years), with haematological malignancy who underwent a total of 738 LP procedures at Royal Manchester Children's Hospital between March 2001 and June 2005. All patients underwent LP and intrathecal chemotherapy under general anaesthetic conditions and procedures were performed by experienced clinicians (Consultant or Specialist registrar) according to the national guidelines. Patients underwent between 1 and 25 (median 11) procedures LPs from diagnosis. For each LP, the platelet count and number of red blood cells (RBCs) within CSF was recorded. Traumatic LP was defined as an LP in which cerebral fluid contained at least 10 RBCs/μl and bloody LP which the CSF contained at least 500 RBCs/μl.
Of the 738 LPs performed, 65 (9%) were traumatic and 30 (4%) were bloody. No bleeding or neurological complications observed in any of the patients. Regression curve analysis showed no correlation with platelet count and red cell count in the CSF (R2 = 0.004). There appeared to be no increase in the incidence of traumatic or bloody taps in patients with a lower pre-procedure platelet count (Fig 1).
Within this cohort of patients with platelet count <50 × 109/l we observed no bleeding complications, and no increase in traumatic or bloody CSF. These findings support previous published data. These results indicate a significantly less traumatic rate than previously reported (Howard et al, 2000).
At our institute we attribute the low rate of red cell contamination to experienced clinicians, dedicated general anaesthetic lists, and specific LP 24-guage needles. This study and existing evidence in the literature would support the safety of performing LPs with platelet counts ≥30 × 109/l. This policy enabled 27 of 738 patients (3.6%) to avoid a platelet transfusion, and thus reduced donor exposure, platelet-associated infections, delay in procedure, as well as reducing costs and making better use of platelet resources.