Logistics, risks, and benefits of automated red blood cell exchange for patients with sickle cell disease

A 16-year-old Black male weighing 65  kg with homozygous SS sickle cell disease (SCD) developed an abnormal transcranial doppler at age 8 while treated with hydroxyurea (HU). He has been treated with monthly transfusions and HU for 8 years. His pretransfusion hemoglobin on HU is about 9-10  g/dL. He has received a mixture of simple and manual exchange transfusions (depending on pretransfusion hemoglobin). He is prescribed iron chelation but frequently misses doses and has evidence of iron overload. His peripheral access has been difficult, and some attempts at manual exchange have been aborted due to failed access, resulting in higher-than-target hemoglobin S (HbS) levels. Automated RBC exchange (aRCE) has been discussed, but patient has been resistant to surgery for placement of a central venous catheter (CVC). While he is trying to make his decision, he asks if he could stop chelation if he transitioned to aRCE and what changes in his HbS he could expect. This review will discuss the logistics, risks, and benefits of aRCE in SCD in order to address these and other clinical questions.

Red blood cell transfusion was the first disease-modifying therapy for individuals with SCD and remains an integral component of the management of SCD today, both acutely to treat and prophylactically to prevent severe SCD complications. Transfusions provide benefit by increasing Hb and, therefore, oxygen delivery to tissues, but also by decreasing abnormal HbS RBCs that contribute to SCD vaso-occlusive pathophysiology. Transfusions can be administered either as a simple transfusion by infusing normal HbA RBCs to dilute the HbS cells or as an exchange transfusion to remove HbS RBCs and replace them with HbA RBCs. Exchange transfusions can be accomplished manually through alternating cycles of phlebotomy and transfusion or as an automated procedure using an apheresis device. Whole blood is drawn from the patient into the device where blood components are separated in a centrifuge according to density. The most dense RBCs will layer on the bottom, followed by white blood cells, platelets, and then plasma. The machine will divert a proportion of the RBC layer to a waste bag to remove RBCs while simultaneously returning most of the patient's white blood cells, platelets, and plasma along with donor RBCs to replace the removed RBCs. There is some loss of other components in the waste bag, and transient decreases in white cells and platelets are expected.¹ 

In order to calculate the total volume/number of RBC units required for aRCE, the patient's data and goals of an individual procedure can be entered into the apheresis device software (or an application downloaded to a smartphone² ). The patient's height, weight, and sex are needed to calculate total blood volume (TBV); in addition, the patient's hematocrit as well as an estimated hematocrit of the RBC units that will be used for the exchange are entered. Citrate-phosphate-dextrose-adenine units typically have a hematocrit of ~75%, while units with additive solutions such as AS-1 usually have a hematocrit of ~55%; however, institution-specific averages should be obtained from local blood banks.³  Next, the desired post-aRCE hematocrit as well as the pre- and target post-aRCE HbS levels are entered. With this information, the software can calculate the exact mL of RBCs required to achieve the desired final hematocrit and HbS. The machine cannot differentiate between HbS and HbA RBCs; therefore, the machine actually calculates the fraction of patient's cells remaining (FCR). Assuming the units are sickle cell negative with no HbS, HbS reflects the patient's RBCs and HbA reflects the transfused RBCs. Therefore, FCR is proportional to HbS:

• Pre-aRCE HbS  =  75%

• Post-aRCE HbS  =  30%

• FCR  =  30/75  =  40%

If some data inputs are unknown or access to the software is not possible, the volume of RBC needed for an exchange can also be estimated by calculating 1.5 times the patient's RBC volume that would be removed/replaced. Over 1 RBC volume is used in the estimate because over the course of the exchange, HbA RBCs will be removed as well. At the beginning of the exchange, primarily HbS RBCs are removed. However, because the machine cannot differentiate HbS from HbA, HbA RBCs are removed as well, and as the exchange progresses, a larger proportion of the RBC volume removed will include the HbA RBCs transfused earlier in the procedure. Estimating 1.5 RBC volume to be removed/replaced accounts for this dynamic. For children, TBV can be estimated to be 70  mL/kg, with the exception of neonates/infants, who have larger blood volume per unit of weight, and obese patients, who have smaller blood volume per unit of weight. Formulas that include height and sex, such as the Nadler equation, can be used to calculate TBV for adults.⁴ 

An example calculation of 1.5 RBC volume for a 30  kg child with a hematocrit of 25% (therefore RBC volume is 25% of TBV) with no recent transfusion follows.

• TBV  =  70  mL/kg  ×  30  kg  =  2100  mL

• RBC volume  =  2100  mL  ×  0.25  =  525  mL

• 1.5  ×  RBC volume  =  787.5  mL

• RBC units have an average volume of approximately 300 mL. Assuming adsol units are available for this exchange, the RBC volume of available units would be approximately 165 mL (300 mL * hematocrit 55%).

• 1.5  ×  patient's RBC volume is estimated to remove the majority of patient's RBCs to decrease HbS from ~100% (no recent transfusion) to as low as possible. For a 30-kg patient on chronic aRCE with a lower pre-transfusion HbS, 3 units are likely sufficient.

The success of aRCE depends on adequate intravenous (IV) access. Two lines are required: 1 draw line to remove whole blood from the patient and 1 return line. The draw line must be capable of supporting adequate flow rates without collapsing against the negative pressure created to pull blood into the device. The flow rate is dependent upon the patient's size. A rate of only 20-35  mL/min is likely sufficient for smaller children (~25-35  kg), whereas a larger adult (80  kg) might require ≥50  mL/min. Our policy is to require a CVC for all acute/urgent aRCE and to attempt peripheral access for all chronic/prophylactic aRCE. We prefer peripheral access for chronic aRCE to avoid complications related to CVCs. In order to optimize chances for success via peripheral IV (PIV) access, we educate all patients about adequate hydration in the 1-2 days prior to aRCE and ensure that patients are warm and calm prior to attempting PIV placement. In addition, apheresis nurses are trained to use ultrasound to guide PIV placement, allowing slightly deeper peripheral veins to be accessed and overall improving the rate of successful PIV placement.

For the draw line, a larger bore PIV capable of supporting the flow and negative pressure is required. A slightly smaller PIV can be placed for the return, but it must be sufficient to withstand the positive pressure of the returning blood. The exact size will depend on the patient size (Table 1). We are able to establish peripheral access in approximately 60% of our chronic aRCE patients, which includes both children and adults. For patients with inadequate peripheral access, a variety of options for chronic aRCE are available. Most commonly, a double lumen implanted port is used. At the time of this writing, only 2 double lumen ports capable of apheresis are available in the US, the 9.5 Fr Bard Power Port Duo and the 11.4 Fr Angiodynamics Vortex, both of which can be used in adults or children whose weight >~45  kg (depending on expertise of surgeon) but not in smaller children. Other options include a single lumen port and PIV return, 2 single lumen implanted ports, a temporary CVC placed and removed on the day of an aRCE, and arterio-venous fistulas or grafts. No option has been demonstrated to be superior, and the choice depends on the patient's preference and anatomy as well as local resources and expertise of the multi-disciplinary team caring for the patient.

For patients who require urgent aRCE, we require a temporary CVC, as acutely ill patients may be dehydrated and thus peripheral access could be more difficult to obtain. The CVC must be an appropriate size for the patient (Table 1) and also rigid enough to withstand pressure and flow rates during the procedure. There are a variety of appropriate rigid lines specifically designed for apheresis or dialysis, typically called power lines. The package insert will confirm flow rates supported by a line to ensure appropriateness for aRCE.

One caveat to this discussion of 2 lines for aRCE is the recent introduction of software and tubing connectors to allow a single intravenous access line to function as both draw and return. Rather than simultaneously removing/returning blood, which occurs in the double-needle procedure, the single- needle (SN) procedure involves alternating cycles of first drawing whole blood and exchanging RBCs in the device and then returning blood to the patient through the same line. These discontinuous cycles repeat multiple times until the target parameters are met, resulting in a longer procedure. One group has reported that in their experience with SN procedures using a single lumen port, no difference was seen in pre- or post-HbS levels when compared to double-needle procedures. The group was able to increase flow rates to avoid longer procedure times.⁵  In our experience using the SN procedure in an adult with 1 PIV, the procedure times are longer than those for the patient's historical double-needle aRCEs. Because SN procedures are not yet widely used, further experience is necessary to define outcomes.

An anticoagulant must be used during aRCE to prevent clotting of blood in the machine. Citrate is the most commonly used anticoagulant, as it primarily functions as an anticoagulant in the machine with minimal systemic effects. Citrate binds the calcium ions necessary for activation of calcium-dependent coagulation proteins and is added to the draw line so whole blood is anticoagulated as it enters the machine.⁶  Upon return of blood to the patient, citrate is quickly metabolized by the liver, kidneys, and skeletal muscle.⁶ 

The benefits of aRCE are due to the removal of patient RBCs, rather than only dilution of patient RBCs, as well as the simultaneous return of blood with the removal.

Blood viscosity increases with increasing hemoglobin and patients with SCD have increased viscosity compared to non-SCD controls even at the same hemoglobin level due to lower deformability of HbS RBCs.^7,8  Therefore transfusion to a hemoglobin level significantly higher than patient’s baseline can result in symptoms of increased viscosity including hypertension, hemorrhagic stroke and even death.^9-14  The ability to remove blood (and decrease HbS) while transfusing RBCs allows an isovolemic transfusion and limits the risk of viscosity.^15,16 

An RBC unit contains approximately 250 mg of iron,¹⁷  which accumulates in the liver and, to a lesser degree, in the heart, pituitary and pancreas over time.¹⁸  Without chelation, excessive iron can result in liver or heart failure and contribute to mortality in SCD.^18-21  Removal of iron-containing RBCs with aRCE can mitigate iron loading, and multiple studies have shown less iron overload with aRCE than with simple transfusion or manual RCE.^22-25  Most of these studies were retrospective, and the ability to control for compliance with chelation medication was difficult or impossible. A secondary analysis of the Silent Cerebral Infarct Trial data was able to compare iron loading, as measured by rise in ferritin, that occurred prospectively in the first year of chronic transfusions before chelation was necessary.²⁶  This analysis clearly showed the ability of aRCE to attenuate rise in ferritin (Figure 1). It is important to note that iron balance depends on the difference between the pre- and post-aRCE hemoglobin/hematocrit. If the ordered post-aRCE hematocrit is significantly higher than the pre-aRCE hematocrit, the patient will remain in positive iron balance and chelation will likely be required.

Exchange transfusion can decrease HbS to a significantly greater extent than can simple transfusion. An apheresis device can accomplish a larger volume exchange and therefore more efficiently lowers HbS than can manual RCE.²⁷  As described earlier in this article, a specific posttransfusion HbS level can be targeted with an individual aRCE procedure. For patients receiving an acute transfusion, the post-HbS level can be more efficiently and reliably decreased with aRCE. For patients treated with chronic transfusions, a specific pretransfusion HbS level is targeted, particularly for patients who undergo transfusion for stroke prevention. The Stroke Prevention Trial in Sickle Cell Anemia demonstrated transfusions that occurred approximately once a month to maintain a pretransfusion HbS of <30% were associated with a 92% reduction (P < 0.001) in stroke risk compared to standard care, making 30% HbS the standard for stroke prevention.²⁸  While most case series have shown lower pretransfusion HbS with chronic aRCE compared to simple or manual RCE,^24,25,29,30  this association has not consistently been shown.²⁷  The pretransfusion HbS is harder to predict because it depends on (1) how low the post-HbS level was after the last transfusion, (2) the interval between transfusions, and (3) the patient's endogenous suppression/production of HbS. For an individual patient, the pretransfusion HbS must be monitored. If the HbS level is above target, consider 1 of the following modifications: (1) increase the number of RBC units during aRCE to decrease posttransfusion HbS levels, (2) decrease the transfusion interval, or (3) increase the post-aRCE hematocrit to suppress endogenous HbS production. However, option 3 might not be possible for patients with higher baseline pretransfusion hemoglobin/hematocrit due to viscosity risk. In addition, significantly increasing hemoglobin levels after the procedure with respect to preprocedure levels will negate the potential mitigation of iron loading.

In general, aRCE is well tolerated, even in the acute setting. Possible adverse affects are summarized in Table 2 and are more likely to be related to CVCs than the actual procedure. Because aRCE exposes patients to significantly more blood than simple transfusion, there has been concern of increased risk of transfusion-related adverse events, though aRCE has actually been associated with lower risk of alloimmunization and transfusion reactions.^31-33 

There have been no randomized trials comparing simple transfusion to aRCE to define when one should be used instead of the other. Recommendations are based on case reports/series or expert opinion (Table 3). In general, in an acute setting, if the hemoglobin level is >2  g/dL lower than patient's baseline, a simple transfusion can be given to increase hemoglobin and improve tissue oxygenation. If the goal of the acute transfusion is to significantly decrease HbS, aRCE should be performed. Automated RBC exchange is most commonly indicated for acute stroke, severe acute chest syndrome, and multiorgan failure. Other indications should be evaluated on a case-by-case basis considering the risks and benefits in the individual patient. Automted RBC exchange should be considered in all patients treated with long term chronic transfusion therapy to mitigate iron loading.^34,35 

The cost of an individual aRCE compared to simple or manual RCE is higher due to the cost of the kit, anticoagulation and tubing for the machine, increased blood units required, any central line–related costs, and skilled nursing time. Cost has been reported as a barrier to implementing aRCE programs.³⁶  However, studies that have balanced the cost savings created by decreased chelation, and SCD hospitalizations have actually shown chronic aRCE programs to be cost-effective.^24,37,38 

The patient presented in the initial clinical case had a double lumen port placed and was transitioned to aRCE. His course demonstrates the potential benefit of aRCE, as both his HbS and ferritin levels significantly decreased (Figure 2). He did intermittently have issues related to his port, such as repeated need for local thrombolytic medication, highlighting potential adverse effects related to the required vascular access. These risks and benefits must be weighed to make patient-specific decisions for individuals with SCD who require chronic transfusions. Areas of future investigation should include further research regarding the SN procedure; comparing the effects of aRCE with chronic simple transfusion on clinical outcomes, such as prevention of stroke and other severe SCD complications; and examining mechanisms of decreased alloimmunization despite significantly increased RBC exposure with aRCE.

Shannon Kelly: no competing financial interests to declare.

Shannon Kelly: nothing to disclose.