Phlebotomy or bloodletting: from tradition to evidence‐based medicine

Phlebotomy or bloodletting is one of the oldest traditional procedures in medicine dating back to ancient Mesopotamian and Egyptian civilizations. In the early 19th century, illness was thought to be due to an imbalance in the four humors (i.e., blood, phlegm, yellow bile, and black bile) and it was believed that the balance could be restored by the withdrawal of blood, often by using devices such as lancets, scarificators, and fleams.1 Our current understanding of bloodletting is that it is ineffective, if not harmful, and modern medicine has all but abandoned this practice except to prevent or treat the effects of iron overload as in hereditary hemochromatosis (HH) or decrease red blood cell (RBC) excess as in polycythemia vera. The first case series for phlebotomy in HH was reported more than 50 years ago and it remains the principal treatment for HH today.2 Although our understanding of the basic science underlying iron metabolism has increased significantly over this half-century with discoveries such as hepcidin, ferroportin, and the transferrin receptor, the methods and evidence for treating HH have not evolved as quickly.3 The use of erythrocytapheresis (EA) in HH instead of manual therapeutic phlebotomy was first published in 1983; the evidence supporting its use published in the past 30 years has been in the form of case series.4-12 However, in this issue of TRANSFUSION, Rombout-Sestrienkova and colleagues13 report on their randomized controlled clinical trial comparing the effectiveness in removing iron in patients with homozygous C282Y hereditary HH with EA versus phlebotomy. In doing so, they take us from tradition to evidence-based medicine. Genetic HH is a disorder of iron metabolism where iron is absorbed in greater amounts than is needed.14 Iron then accumulates in the parenchyma of solid organs, particularly in the liver, pancreas, and heart. Clinical findings are variable ranging from asymptomatic—to mild with fatigue and arthropathy—to severe with liver cirrhosis, hepatocellular carcinoma, diabetes, and cardiomyopathy. HH is a disorder that can be caused by a number of rare gene mutations such as ferroportin (FPN),15 hemojuvelin (HJV),16 hepcidin (HAMP),17 and transferrin receptor 2 (TfR2).18 However, more than 80% of cases of HH are associated with homozygosity for the C282Y genetic polymorphism in the HFE gene.19 HH is the most common genetic inherited disorder in whites with a prevalence of 0.5%, but it is uncommon in other ethnic populations such as Asians and African Americans. The clinical syndrome has a low rate of penetrance—50% of individuals with HH have iron overload and 10% to 30% have iron overload morbidity.20, 21 Thus, unknown factors are likely to influence disease severity. Currently, the diagnosis is made by genetic testing for C282Y mutation in HFE along with the identification of an elevated serum ferritin or increased transferrin saturation. Whereas bloodletting in the past was performed according to superstition or tradition, Davis and Arrowsmith in 195022 provided evidence supporting the use of manual phlebotomy in HH; he described three patients with HH who were subjected to repeated phlebotomies of 500 mL of whole blood and each of the patients reported improvement in subjective energy and working ability, had improved carbohydrate metabolism and liver function, and showed decreased iron pigment. For more than 60 years, manual phlebotomy has been the mainstay of therapy for these patients, typically initiated when the serum ferritin rises above the normal range.2 Phlebotomy is generally performed weekly until the serum ferritin is reduced to less than 50 µg/L and then phlebotomy is performed periodically to maintain it at less than 50 µg/L. Evidence suggests that early diagnosis and phlebotomy treatment may improve clinical symptoms and outcome. It is important to note, however, that a randomized clinical trial comparing phlebotomy to no treatment has never been performed and would now be unethical to conduct. Moreover, patients without cirrhosis undergoing phlebotomy may have similar survival to published causes of death in a normal population.23, 24 Manual phlebotomy may not be advisable in some patients, for example, a patient with severe heart disease. Other patients may find that they cannot tolerate the procedure due to vasovagal reactions, hypotension, dizziness, or poor venous access. In these cases EA, which was first proposed in 1983 by Conte and colleagues may be useful. Conte and colleagues4 reported on three patients who underwent EA of 250 mL of RBCs twice a week. A significant reduction in the serum ferritin was achieved within 6 to 10 months. Since then, additional case reports inclusive of a total of 86 patients in 10 case series from 1983 to 2007 have been reported.4-12 These case series may or may not have phlebotomy controls and had variable intervals for EA, different volume of RBCs removed per procedure, and different target goals. However, the overall message from these studies is that iron unloading appears to be achieved in a shorter duration of time and with fewer visits than phlebotomy and more iron can be removed per procedure. In the most recent study by Fernandez-Mosteirin and coworkers,6 for example, single-needle procedures with 275 mL of RBCs removed per session decreased the iron load in one-third the time as phlebotomy. As noted in many of these studies, one of the main drawbacks of EA has been the specialized equipment and higher labor and facility costs. In this issue of TRANSFUSION, Rombout-Sestrienkova and coworkers13 from the Sanquin Blood Bank in the Netherlands report on the first randomized trial comparing phlebotomy to EA in patients with C282Y HH. Thirty-eight newly diagnosed HH patients with C282Y homozygous mutations were randomly assigned to a once weekly phlebotomy of 500 mL of whole blood arm or a biweekly EA of 300 to 800 mL of RBCs. The primary endpoints were the number and the duration of treatments needed to reduce the serum ferritin to 50 µg/L. Even after accounting for the lower initial serum ferritin levels in the EA group at the start of the study, their data show a significant reduction in treatment procedures (9 vs. 27) and treatment duration (20 months vs. 34 months) in the EA compared to the phlebotomy group, respectively. The number of adverse events was 5 of 19 (26.3%) patients in 10 of 513 (1.9%) procedures in the phlebotomy group versus 3 of 19 (15.8%) patients who reported eight events per 171 procedures (4.7%) in the EA group, which was not significantly different. These reactions consisted of one “short-lasting collapse” and mild dizziness in the phlebotomy group and mild citrate toxicity, vasovagal reaction, and dizziness in the EA group. One of the main arguments against the use of EA over phlebotomy is the cost of therapy. Although increased cost should be taken into consideration, patient compliance and outcome are important variables to consider. In a survey conducted in 1999, patients reported that although the benefits of phlebotomy outweighed the risks, 12% of respondents had a negative attitude toward phlebotomy because of poor venous access and time involved and that the blood was discarded. Rombout-Sestrienkova and coworkers13 include in their cost analysis the cost of the procedure and the loss of time and productivity. Interestingly, the costs are less for EA, but even with their analysis questions remain about adequate access to apheresis. In a survey of over 2851 HH patients in the United States, United Kingdom, Canada, and Australia, 73% of patient responded that they received their phlebotomy in a doctor's office or hospital and 25% in a blood bank.25 If the procedure cannot be offered in their physician's office, will the time required to travel to specialized centers performing EA nullify the benefits of time saved with shorter number and duration of procedures? One major limitation of the study is the absence of clinical outcomes. Evidence supports decreased survival of HH patients with cirrhosis and treatment with phlebotomy can reverse liver fibrosis and cardiac abnormalities.23, 26 Understandably, the progression of disease can be slow and not all people with HH develop disease. Such a study would also require long-term follow-up and much larger numbers of subjects would be required to detect a difference. We look forward to seeing a report on whether clinical symptoms, end organ deterioration, or survivals are different with EA versus phlebotomy from either this group or others. In conclusion, Rombout-Sestrienkova and colleagues13 have raised the quality of the studies available in the use of EA to treat HH. The move from tradition in transfusion medicine to evidence-based medicine has begun. With this evidence there is potentially a change in the categorization of EA to Category I for HH in the next edition of clinical applications guidelines published by the American Society for Apheresis.27 In a field where much of the evidence is in the form of tradition and case series, we hope to see more investigators pursuing randomized trials in transfusion medicine. None.