Obstetric neuraxial anaesthesia in the context of maternal immune thrombocytopenia: secondary analysis of a retrospective cohort study

Editor—Immune thrombocytopenia (ITP) is an autoimmune condition affecting 1–10 in 10 000 pregnancies.¹ The bleeding risk associated with significant thrombocytopenia poses a challenge peripartum, particularly for placement of neuraxial anaesthesia, where evidence to guide practice remains scant.² Our aim is to describe our experience with neuraxial anaesthesia in the setting of ITP in pregnancy at two tertiary-level academic institutions. The report is a secondary analysis of a retrospective study of pregnant women with ITP at two tertiary-level Canadian academic institutions: Mount Sinai Hospital (MSH), Toronto and McMaster University Medical Centre, Hamilton (MUMC) from January 2000 to August 2014.³ Research ethics board approval was granted at both hospital sites.

Briefly, potential participants were identified via search of the hospital medical records, applying ICD-10 codes for thrombocytopenia, and cross-referenced with hospital-specific electronic databases. Those with a diagnosis of ITP, confirmed by a history of thrombocytopenia (platelets <100 × 10⁹ litre⁻¹) predating the pregnancy or with onset in the first trimester, were included. Patients who had thrombocytopenia with a different aetiology, including pre-eclampsia spectrum and sepsis, were excluded. Likewise, patients without a history of thrombocytopenia outside pregnancy and with platelets >70 × 10⁹ litre⁻¹, which normalized postpartum, were considered to reflect gestational thrombocytopenia and excluded.⁴ Information on neuraxial (spinal and epidural) anaesthesia was collected on all patients as part of data collection for the original study by a subset of study authors. Data management and statistical analyses were performed using SAS 9.3 (SAS Institute, Inc., Cary, NC, USA) and R 2.15 (www.r-project.org). Differences in the rate of neuraxial anaesthesia placement between the two institutions were estimated using a probability linear model.

Of 689 pregnancies with thrombocytopenia (platelets <100 × 10⁹ litre⁻¹) or history of ITP, 454 were excluded, resulting in inclusion of 235 pregnancies in 195 women with current or past diagnosis of ITP.³ Of 234 pregnancies for which data were available, neuraxial anaesthesia was provided in 136 (58%), of which a neuraxial anaesthetic was placed at platelet counts <80 × 10⁹ litre⁻¹ in 24 (18%) patients. Two patients had neuraxial anaesthesia at platelet counts <50 × 10⁹ litre⁻¹ (45 and 48 × 10⁹ litre⁻¹, respectively). Specific platelet count ranges at the time of neuraxial anaesthesia placement are shown in Table 1. No neurological or bleeding complications associated with neuraxial anaesthesia were reported. Institutional variation was noted with respect to the proportion of neuraxial anaesthetic initiated at platelet count ranges of 50–79 and 80–99 × 10⁹ litre⁻¹ (Table 1). The overall adjusted odds ratio of neuraxial anaesthesia placement at MSH vs MUMC, accounting for the level of maternal platelets at delivery, was 2.22 (95% confidence interval 1.15, 4.27).

Table 1

Proportion of neuraxial anaesthetics initiated at specific platelet count ranges in the present study based on site and as compared with the prior report by Tanaka and colleagues.⁶ MSH, Mount Sinai Hospital; MUMC, McMaster University Medical Centre. *Difference is the difference in the rate of neuraxial anaesthesia between the two hospitals (MSH vs MUMC). ^†N/A, the confidence interval (CI) for the difference in rate cannot be obtained owing to the rarity of events

Platelet count (×109 litre−1)				Neuraxial anaesthesia [n (%)]
	Epidural	Spinal	Total (MSH+MUMC)	MSH	MUMC	P-value	Difference (95% CI)*	Tanaka and colleagues6
≥100	56/99	25/99	81/99 (82)	68/83 (82)	13/15 (87)	0.99	−4.7 (−23.8, 14.4)	–
90–99	8/16	4/16	12/16 (75)	26/30 (87)	5/10 (50)	0.03	36.7 (3.4, 70.0)	34/37 (92)
80–89	15/24	4/24	19/24 (79)
70–79	6/20	3/20	9/20 (45)	17/38 (45)	5/25 (20)	0.04	24.7 (2.5, 47.0)	13/27 (48)
60–69	9/25	2/25	11/25 (44)
50–59	1/18	1/18	2/18 (11)
<50	1/33	1/33	2/33 (6)	2/20 (10)	0/13 (0)	0.51	10.0 (N/A)†	0/11 (0)

Literature on ITP in pregnancy is sparse, and bleeding risk assessment is hindered by reports analysing ITP alongside thrombocytopenic disorders with potentially impaired platelet function (e.g. pre-eclampsia),⁵ impeding an accurate estimation of the degree of risk attributable strictly to ITP. Irrespective of platelet counts, we found no complications from neuraxial anaesthesia in 136 pregnancies with ITP in one of the largest cohorts reported to date, with the lowest platelet count under which neuraxial anaesthesia was initiated being 45 × 10⁹ litre⁻¹. Our findings are similar to those of Tanaka and colleagues,⁶ who studied non-pre-eclamptic thrombocytopenic women at MSH between 2001 and 2006 (Table 1).

Precise platelet counts predicting increased risk of neuraxial anaesthetic complications have not been determined,⁷ and given the rarity of epidural haematoma, with an estimated incidence of 1 in 168 000,⁸ our sample size does not allow us conclusively to establish the safety of neuraxial anaesthesia in the context of thrombocytopenia. Using the ‘rule of three’,⁹ the absence of adverse events reported in our study for platelet counts <100×10⁹ litre⁻¹ yields an upper 95% confidence interval limit of the population risk of 5.5% (3/n=3/55).

Our results also highlight the institutional variation in placement of neuraxial anaesthesia at low platelet counts (Table 1), notwithstanding the setting of large tertiary-care academic centres with arguably more exposure to this condition than community hospitals. This lack of consensus ultimately leaves many eligible women without access to satisfactory pain relief in labour. Our experience provides valuable additive data to a sparsely reported subject for a commonly recognized clinical condition and highlights the need for further study of neuraxial anaesthesia in parturients with bleeding disorders.^10,11

Authors’ contributions

Interpretation of data, manuscript revision, and approval of final version: all authors.

Acknowledgements

We thank Dr Xiang Y. Ye for his assistance with the statistical analysis.

Declaration of interest

None declared.

Funding

Canadian Blood Services Small Projects Fund; Canadian Institute of Health Research/Canadian Blood Services New Investigator Award (to N.S.).