Approach to transfusion in pregnant women with sickle cell disease: a survey of physicians

Pregnancies in women with sickle cell disease (SCD) remain characterized by frequent maternal and fetal complications (Kuo & Caughey, 2016). Prophylactic red blood cell transfusion (PT) has been proposed to improve outcomes, though studies have been conflicting (Tuck et al, 1987; Koshy et al, 1988). A Cochrane review addressing PT versus selective transfusion in SCD-affected pregnancies cited inadequate evidence to guide transfusion protocols (Okusanya & Oladapo, 2016). While our meta-analysis demonstrated that PTs may positively impact maternal and neonatal outcomes, the conclusions stemmed from relatively few studies with significant methodological limitations (Malinowski et al, 2015). However, both reviews crystalized the urgent need for a randomised controlled trial (RCT) to determine the role of PT during SCD-affected pregnancies. To achieve the necessary power for such a trial, international collaboration will be essential and appreciation of the current approach to transfusion is needed. Thus, a survey of relevant health care professionals was conducted to explore these issues (Appendix S1). Professional Maternal-Fetal Medicine and Haematological Societies in Canada and the United States were approached via e-mail, with a request to forward a formal invitation to their membership. An electronic link to the survey (FluidSurveys™) was included. Participation implied consent. Two reminder e-mails followed. Of 811 invitations, 163 involved Haematologists and 648 involved Maternal-Fetal Medicine Physicians and Obstetricians (MFM/OB). Of 94 (12%) responses, 28% (26/94) were from Haematologists, 70% (66/94) from MFM/OB, and 2% (2/94) lacked specification of specialty. Of 93 respondents, 27 (29%) had practiced for ≤10 years, 23/93 (25%) for 11–20 years, and 43/93 (46%) for >20 years. Geographic location was indicated by 92, of which 1 (1%) was from Panama, 3 (3%) from Canada and 88 (96%) from the United States, with 77/93 (83%) practicing in academic/teaching hospitals and 14/93 (15%) in community hospitals. Half the respondents (46/92) practiced in a joint Haematology/MFM clinic, with a median volume of 5 pregnant SCD patients/year [Interquartile Range 25–75 (IQR) 3-8 patients/year]. Transfusion practice varied, with 4/91 (4%) providing PT to all pregnant SCD-affected women, 46/91 (51%) providing PT for specific indications only (e.g. recurrent vaso-occlusive pain episodes (VOPE), fetal growth restriction, etc.), 37/91 (41%) providing on-demand transfusion on an as needed basis, and 4/91 (4%) indicating ‘other’ (Table 1, Question [Q]3). Transfusion technique likewise varied, with 76/79 (96%) respondents preferring simple transfusion, 23/79 (29%) choosing manual exchange transfusion, and 39/79 (49%) utilizing automated exchange transfusion. More specifically, simple transfusion, partial manual exchange transfusion, and automated exchange transfusion were used at a median (IQR) of 90% (70–100%), 20% (10–40%), and 30% (10–50%) of time respectively. Based on 35 responses, the median desired pre-transfusion Haemoglobin (Hb) A target was 50% (IQR 28–50%), and based on 46 responses the median post-transfusion Hb A target was 50% (IQR 48–70%), while 30 participants specified the use of variable targets depending on transfusion indication (Table 1, Q5). Based on 45 responses, the median desired post-transfusion Hb target was 100 g/l (IQR 90–100 g/l) and the median post-transfusion haematocrit target was 30% (IQR 26–30%). According to 88 respondents, challenges with patient acceptance of the recommendation for PT were encountered often by 6%, sometimes by 28%, seldom by 43% and never by 23%, with reasons for non-acceptance cited in Table 1, Q9. The need for a RCT to provide a conclusive answer regarding the benefits of PT in SCD-affected pregnancies was supported by 85/90 (94%) of respondents, while interest in participation was declared by 55/85 (61%), with 89% disclosing the inability to participate yet also affirming the need for the trial. In a trial setting, 78/88 (89%) would provide PT for recurrent VOPE, 33/88 (38%) for any VOPE, 34/88 (39%) for acute chest syndrome (ACS) pre-dating pregnancy, 79/88 (90%) for ACS in current pregnancy, 24/88 (27%) as primary prophylaxis for neurovascular risk and 40/88 (45%) as secondary prophylaxis, 38/88 (43%) for fetal growth restriction, 31/88 (35%) for multiple gestation and 18/88 (20%) for pre-pregnancy hydroxycarbamide use. Three respondents declared unwillingness to ever consider PT and nine indicated “other” (Table 1, Q4). Only 25% felt that improvement in VOPE should serve as the most important (MI)/very important (VI) primary outcome, while none thought improvement in fetal weight should be the MI primary outcome. Conversely, 82% considered the composite of adequate placentation (including perinatal mortality, small for gestational age size and hypertensive disorders of pregnancy) to be MI/VI and 90% considered the composite of SCD-related maternal adverse events (including VOPE, ACS and venous or arterial thromboembolism) to be MI/VI (Fig 1). The notable variation in approach to transfusion is probably multifactorial, involving the attributes of local blood banks, logistics of arranging particular transfusion types, health-care coverage, availability of local expertise, and clinician-dependent perception of risks and benefits of transfusion in this setting. However, variation may also be driven by absence of well-designed studies on which to base management decisions (Malinowski et al, 2015; Okusanya & Oladapo, 2016). Given potential transfusion-related harms (Chou, 2013; Delaney et al, 2016), benefits must be irrefutable for practice to change. Furthermore, concerted education efforts to reduce misconceptions and increase acceptance of transfusion among patients must be undertaken. The substantial interest in a trial is encouraging and a composite primary outcome seems to be favoured, with some inclination towards adoption of a composite of maternal outcome over a composite of fetal outcome. As trial design unfolds, inclusion of patients’ opinion in this regard will be essential, particularly given evidence that women place a higher value on fetal effects of interventions in preference to the impact these may have on themselves (Bates et al, 2012). Our low response rate is a limiting though not unexpected factor, as response rates have been on the decline (Sheehan, 2006) and low physician specialist response rates have been reported (Cunningham et al, 2015). The non-targeted distribution of invitations to general memberships of professional societies, a significant proportion of which may not encounter SCD-affected pregnancies, probably contributed. Acknowledging these limitations, our survey confirmed significant variation apropos current approach to PT in SCD-affected pregnancies and established interest in trial participation, highlighting the desirability of a composite outcome. J.P. has been generously supported by the University of Toronto Department of Obstetrics and Gynaecology Chair's Summer Student Award. N.S. is supported by a Canadian Institute of Health Research/Canadian Blood Services New Investigator Award. A.K.M. substantially contributed to research design, data acquisition, analysis and interpretation, drafted the paper, critically revised the paper, and approved the final version. J.P. substantially contributed to data acquisition, analysis and interpretation, critically revised the paper and approved the final version. N.S. substantially contributed to research design and data interpretation, critically revised the paper and approved the final version. R.W. substantially contributed to research design and data interpretation, critically revised the paper and approved the final version. K.H.M.K. substantially contributed to research design, data analysis and interpretation, critically revised the paper and approved the final version. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.