Claire and the House of Miracles

| 11 May 2017

Miss. Claire

Despite the flaws and frustrations of our health care system and the challenges faced by our crowded hospitals, we should not lose sight of the miracles that are accomplished in academic health science centers every day.

Decorative Image

On the shore of Lake Ontario sits a hospital. Built 175 years ago the hospital has a storied history, including serving as the initial parliament building for Canada, before our capital moved to Ottawa. This hospital, formerly called Kingston General Hospital, now Kingston Health Sciences Centre, has seen the births and deaths of the good people of Southeastern Ontario for generations. The hospital has also been the site of many amazing and innovative therapies and served as the training ground for doctors, nurses and other healthcare professionals at Queen’s University. Decorative Image

Kingston Health Sciences Centre

Recently, morale at this and many other hospitals in Canada has sagged under the burden of overcrowded emergency rooms and congested medicine wards and the resulting system gridlock. With the rising cost of health care, the ministry of Health and Long Term Care has pumped the brakes and flat lined funding for hospitals, pinning its hopes on outpatient care. After 5 years of minimal or no funding increases, Ontario’s hospitals have been fraying at the seams. With all these forces at play, one sometimes forgets that our hospital is a house of miracles. This blog is but one reminder of this and of why the medical team (broadly defined) comes to work every day. Amid waiting list and delays there are miracles happening every day. Some of these miracles are high-tech, life-saving surgeries. Some are catheter-based interventions, such as opening blocked arteries to treat heart attacks (angioplasty-PCI), replacing the aortic valve without surgery (TAVI), removing clots from the middle cerebral artery to reverse a stroke (EVT) or eliminating life threatening heart rhythm disorders by radiofrequency catheter ablation techniques. Sometimes the miracles appear in the form of new wonder drugs that pull patients back from the brink of death. Whether medical, surgical or pharmacological, these miracles always involve passionate patient advocates: doctors, nurses, technicians, trainees and a host of supporting staff. Curiosity and drive to obtain the right diagnosis and to deliver a definitive therapy to cure the patient underlies the provision of most miracles. The miracles in our house also involve good people in laboratories, imaging services and administration, who though not at the bedside, are going the extra mile to help patients they will never meet. There are people cleaning the rooms, assigning beds, supporting information technology, managing finance, and dispensing medications. There are also administrators, who are charged with overseeing the complex organism that is the 21^st Century academic health sciences centre. The patient in her room may not get the sense of the dozens of unseen team members that will determine their fate, but they are there. Claire’s story: Let me recount one recent miracle, which inspires me and reminds me that our team is remarkable. This is the story of a young woman, Claire, and I tell it with her permission. It is also the story of her doctor, Jocelyn Garland, a nephrologist (kidney doctor) and the many colleagues who played essential roles in saving Claire’s life. The story ends with a consideration of the rising cost of these miracles. Claire’s story started innocently enough. Claire was a healthy high school student. She had just returned to Ontario after a family vacation in Europe in July 2016. Shortly after her return she ate several hamburgers and ate a lot of cookie dough while baking. This sounds an odd start to a story of such high drama. However, physicians know that hamburger meat and other food sources are occasionally home to toxic bacterial strains that can cause life-threatening diseases, including kidney failure. You may have read stories of young people eating improperly cooked or contaminated hamburger or fresh produce containing a type of bacteria called E. Coli (shown in electron micrograph images from the Centre for Disease Control, below). Decorative Image

Some E. coli make toxins that damage blood and kidney cells and can result in death from kidney failure. The only reliable means of avoiding intermittent outbreaks of “hamburger disease” is to ensure the meat is well cooked (see figure below). In Claire’s case, we don’t know the source of the problem (pathogenic E. coli can also be found in contaminated water and other foods such as flour, the latter which has recently been reported in the media as an exposure to pathogenic E. coli).

The time course of Claire’s illness is a stark reminder of how fast one can go from being entirely well to being very sick.

August 9, 2016: Claire began having severe abdominal cramping that progressively worsened through the week. On August 13^th the pain escalated and rectal bleeding started so she went the ER. Claire was admitted to the pediatrics service of Kingston General Hospital on August 15, 2016 under the excellent care of Dr. Michael Storr. Infectious disease specialists and the Acute Pain Service were consulted to help with Claire’s case. The gastroenterology service (a team of doctors that specializes in diseases of the intestines) used an endoscope to look inside Claire’s bowel. The image below is what they saw: the wall of her colon was swollen and bleeding, which is typical for infections, or, in medical parlance, infectious colitis.

Decorative Image

At this point in her story everything seemed like a fairly severe case of infectious colitis, which typically spontaneously resolves with supportive care. Claire was an otherwise healthy 17-year-old whose biggest worry was that she was stuck in hospital while her friends were enjoying their summer and she was missing out. While the doctors thought she would likely just be in hospital for a few days one feature caught Dr. Storr’s eye in her laboratory tests: her complete blood count (CBC) showed her platelet count was slightly low at 136 (normal: 150-400 X 10⁹/L). Platelets are circulating cell fragments that lack their own nucleus (they were once part of a megakaryocyte). Platelets prevent bleeding. The image below provides a nice illustration of some of the types of blood cells that became relevant in Claire’s case: in red: red blood cell, in purple: platelets and in white: white blood cell.

Decorative Image

Low platelet counts portend increased risk of bleeding. Often the fact the platelet counts are low reflects their destruction by an evil humor, such as a bacterial toxin, or inappropriate activation of the body’s immune system. August 17: By the morning of the 17th, her diarrhea had slowed down and Claire was feeling a bit better and was optimistic about going home soon. However, as the day progressed she started to feel more and more unwell. She noticed that her abdomen was becoming swollen, she was no longer hungry, her pain started getting worse and she began vomiting. Dr. Storr was worried that her bowels had stopped moving ordered an ultrasound and asked the general surgeons to see her. Labs were drawn serially during the day and with each draw became more worrisome. She was now anemic (low red blood cell levels) with hemoglobin of 108 (normal: 120-160 g/L). Her platelet count dropped to 47 (normal: 150-400 X 10⁹/L) and there were fragmented red blood cells observed on her peripheral blood film. Fragmented red blood cells occur in a variety of illnesses, and may mean that red cells are being destroyed by a process known as microangiopathic hemolytic anemia. Her urine now had a substantial amount of blood and protein it which was, a sign that her kidneys were becoming severely damaged (likely by a toxin coming from the bacteria in her bowel and its effects on her immune system). The ultrasound result was alarming. Claire now had ascites which is a collection of free fluid in her abdomen that had leaked out from her bowels, and her kidneys were enlarged and ‘echogenic’ which is a radiology word which is indicative of something called hemolytic uremic syndrome (HUS). HUS is a medical emergency with a 5-10% mortality rate, and 30% likelihood of having permanent kidney damage. Some infections release toxins that bind to the cells that line the colon, kidney and brain causing bleeding, micro-clots and platelet destruction. Late on the 17^th Claire’s platelets dropped further to 29 and she was transferred to pediatric ICU, given a transfusion of platelets and consultation requests were urgently requested from hematology, general surgery and nephrology. August 18: Unfortunately, Claire worsened on the 18^th. She had central lines inserted into her veins all the way into her heart, a suction tube was inserted from her nose into her stomach to relieve pressure, and artificial nutrition called TPN was initiated. A catheter was put into her bladder and now frank blood was seen in her urine. She started getting severe headaches and her blood pressure rose. Over the day, Claire slipped into kidney failure (her creatinine rose to 103 umol/L (normal < 85 umol/L) and her platelet count fell to a dangerously low-level of 16 (150-400 X 10⁹/L. She was transferred to the Adult Intensive Care Unit (ICU). An intensive laboratory investigation of Claire’s blood was underway. It revealed that her red blood cells were being hemolyzed (exploding) (she had an undetectable haptoglobin level and levels of lactate dehydrogenase were marked elevated at 1607 (normal: 120-315 U/L) (which continued to rise until their peak at 3400). Her body could not make red blood cells as fast as they were being destroyed. As a result, her hemoglobin level; fell to 88 g/L. This is bad because red blood cells contain the hemoglobin, which is required to carry oxygen. She also had evidence of a liver inflammation with increased liver enzymes (AST 82)(normal 10-46 u/l) and total bilirubin (33) (normal 0-17 umol/L). August 19-20: For the first time doctors learned that Claire’s complement system was involved in the mechanism of her illness. The complement system is normally a defense mechanism that when turned on can kill bacteria. However, when it is turned on excessively, or when it is indiscriminately targeted, it destroys the host-damaging own cells. Massive activation of the complement system depletes the complement proteins. Consistent with this, Claire developed hypo-complementemia, meaning that key components of the system were present in low levels (e.g. C3 = 0.76 (normal: 0.8-1.8 g/L)), C4 = 0.07 (normal: 0.13-0.4 g/L). As a result, her hematological picture worsened with platelet counts of only 25) (normal 150-400 X 10⁹/L) and a hemoglobin of only 65 (normal: 120-160 g/L) requiring transfusions. Even worse, her kidney function deteriorated, evident with the rise in her serum creatinine to 270 umol/L (normal less than 85 umol/L) and thus she required dialysis to clean her blood of toxins that her kidneys were otherwise not able to eliminate. She was losing 20 g/day of protein in her urine (normal 0.01-0.14 g/day). A variety of vasculitis tests and tests for other exotic diseases (which I won’t mention) were normal (ADAMTS13 level, PTT and INR, antiphospholipid antibody). August 18 to August 22: Given the continued deterioration in her clinical status and the fall in her complement levels, therapy with plasma exchange was initiated on August 18, 2016. This is a form of replacing the “bad blood” with good blood using essentially a dialysis-like circuit. She had daily plasma exchange from August 18 to August 22 without benefit. This suggested that the toxic factor could not be removed from the plasma and raised the possibility of rare syndrome called atypical hemolytic uremic syndrome (aHUS). Bacterial toxins, like Shiga toxin, can cause damage to the lining of blood vessels called the endothelium, resulting in red blood cell destruction (hemolysis) and platelet consumption. In addition, the effects of Shiga toxin can involve many organs in one’s body, and often attacks the kidney by a complicated mechanism that may involve complement activation. Claire’s initial stool culture was checked by the pediatric infectious disease specialist for the usual culprit Shiga toxin producing E. coli (STEC); however the first cultures were negative. Retesting the stool sample using a more sensitive molecular test sent to Toronto (called polymerase chain reaction or PCR) provided the Eureka moment: the test showed that E.coli (producing Shiga toxin) was present. STEC is the most common cause of HUS (STEC HUS). In Claire it wasn’t clear if there also was underlying aHUS that was triggered by STEC HUS August 22, 2016: As often happens in medicine, there were features of the case which were confusing. Plasma exchange is a therapy used to treat a similar condition called thrombotic thrombocytopenic purpura (TTP). Often, plasma exchange is tried as a treatment in patients who have severe STEC HUS, but the usual treatment is supportive. The manifestation of Claire’s illness was so aggressive and involved multiple organs, therefore a trial of plasma exchange was initiated. When plasma exchanged failed to make any clinical difference, Dr. Garland consulted the medical literature, and saw that rarely, STEC HUS is reported as a trigger for a disease atypical HUS. aHUS is a disease that presents similarly, except that instead of being caused by a bacterial toxin, as in STEC HUS, it is caused by a genetic defect in the alternative pathway on complement. Also, in some cases, severe STEC HUS cases have been treated successfully with a medication called Soliris. The team had to make a pivot in their therapeutic strategy. With falling levels of C3 and C4, they requested permission to use an expensive, orphan drug that specifically blocks complement activation, called Soliris (a monoclonal antibody to C5 which inhibits the terminal complement pathway). The hope was that by stopping complement activation damage to platelets and other cells could be halted. Dr. Garland, and her colleague Dr. Shamseddin, contacted me requesting Soliris and cited case reports of eculizumab being used to treat severe Shiga-toxin-producing Escherichia coli hemolytic-uremic syndrome (STEC-HUS) or atypical HUS (see references at end of blog). E. Coli make this Shiga toxin, which triggers uncontrolled complement activity through direct activation of the alternative complement pathway as well as by binding to and inactivating the regulatory protein complement factor H. Based on her compelling case for using the medicine in this patient the drug was approved by the Medicine program (Mr. Richard Jewitt and me) and the Chief of Staff (Dr. David Zelt) and supplied by pharmacy (Ms. Vero Briggs). Hospital administration and pharmacy rapidly approved and delivered the medication. Claire received her first dose of Soliris late on August 23^rd. August 25: At 4:30 am Claire’s mom found Claire standing up in her ICU room confused and repeating that she couldn’t see anything and that everything was black. Claire had been having increasing headaches and rising blood pressure on the preceding days. To the horror of her parents and physicians in the early morning of August 25th Claire was blind and disoriented and thought that she was at school. Later that morning she started having seizures. Her platelets were critically low at 12 (normal 150-400 X 10⁹/L) so a massive hemorrhagic stroke was a probable diagnosis. (See her compelling first-person account of this at the end of the blog). A MRI showed she had developed a feared complication of hypertensive crisis, called posterior reversible encephalopathy syndrome (PRES). PRES is neurotoxic statethat occurs when the blood supply to the back of the brain can’t autoregulate and lower pressure when systemic BP spikes. The resulting increased perfusion at high-pressure damages the blood brain barrier and endothelial cells lining vessels and causes a vasogenic edema, most commonly in the parieto-occipital regions of the brain (the back part which controls vision). In Claire’s case the PRES was due to from high blood pressure and endothelial cell damage (likely due both to Shiga toxin and the effects of complement activation).

Decorative Image

The areas of PRES on her brain scans were not entirely typical (unlike the example above which is typical, she had involvement of the watershed areas in the frontal cerebral cortex). This suggested that the PRES was quite possibly due to a thrombotic microangiopathy (another unusual complication). Claire suffered one small micro-hemorrhage but thankfully no other permanent structural damage occurred. She had five more plasma exchanges to treat a presumed neurological thrombotic microangiopathic crisis but this did not improve her clinical status, and unfortunately removed from her blood most of the first dose of Soliris that she had been given. In total Claire received 60 litres of donor plasma that arrived daily frozen in coolers from Ottawa. Imagine how many blood donors it took to provide this. August 29th, 2016: Today a tricky decision had to be made. Claire had a neurological crisis about 36 hours after receiving her first dose of Soliris. Was this caused by the Soliris? Or was it the natural progression of the hypertensive crisis and/or sequelae from the STEC HUS? Her labs showed some hints that possibly the Soliris hard started to work. Given that the plasma exchange was not effective the decision was made to re-start Soliris with a dose adjustment. Soliris was started again in a higher dose, as it was determined that she was likely subtherapeutic with the first dose of Soliris, as it may have been lost into the pleural space, ascites, urine if severe proteinuria and possibly, through the bowel in the setting of colitis. August 31st: After this higher dose Soliris treatment, her blood counts finally began to improve within 2-3 days and by August 31, 2016, her hemoglobin had increased to 76 (normal: 120-160 g/L), platelet rose to 70's (normal: 150-400 X 10⁹/L) and C3 levels returned to normal. Blood pressure control improved dramatically as well, and her neurological status improved. Claire was beginning the long road back to health. By this point the team was aware Claire’s condition was unique and Dr. Garland and team wanted to understand better what was happening. Assessing the complement system is difficult, as the levels of complement system components that are circulating in the blood, such as C3, and C5b-9, do not necessarily reflect complement activity at the level where it matters: the endothelial cells that line the blood vessel. With the help of the KGH laboratory, Claire’s blood was shipped to Italy for a special laboratory test where actual C5b-9 deposition on endothelial cells was quantified. This testing was carried out through the kind assistance of Dr. Marina Noris, Milan, Italy and Sara Gastoldi, Istituto di Ricerche Farmacologiche Mario Negri Centro, Bergamo, Italy. The test results demonstrated marked complement activation at baseline, pre-Soliris, and normalization of complement activation with Soliris treatment at the level of the endothelial cell membrane, explaining why anti-C5 therapy with high dose Soliris had resulted in such a dramatic improvement in her clinical condition. In summary, Miss Claire developed life-threatening, multi-organ involvement hemolytic uremic syndrome, Shiga toxin, E. coli (STEC HUS). Her clinical course only improved with C5 inhibition treatment (Soliris). The team will also perform genetic testing to make sure the STEC HUS was not a trigger unmasking a predisposition to atypical HUS. The testing in Italy should clarify this in the months to come. Claire’s case raises several important points. First, the role of the physician as a champion and advocate is crucial. Dr. Garland and her team’s advocacy and expertise were key to the success of Claire’s management. Second, medicine is a team sport and our hospital-based teams perform miracles. Think of the countless people who made significant contributions to Claire’s care: procedures, imaging, lab tests, those supplying the drugs, operating specialized plasma exchange and dialysis equipment, nursing care, room cleaning and stocking, portering her from room to room, providing food that was replenishing and appealing, providing emotional support to a teenager spending 30+ days in the ICU, family support, blood donors, volunteers, therapy dogs, and hospital patrons who made donations that furnished quiet rooms, cots for family members to remain at her bedside and even for donating the air hockey table that she enjoyed once she returned to the pediatric floor. It takes an exceptional number of willing and skilled hands to achieve the level of superb care Claire received. Kudos to all who helped. Some led, others followed, in this case the system and all its members worked well. The third point, and a corollary to the importance of hospitals as sites of miraculous care, is that we need adequate and timely access to our hospital; beds. We cannot allow tertiary care hospitals to be relegated as de facto alternate level of care (ALC) facilities (see prior blog on this topic). A fourth and thorny point is the cost of miracles; there is an interesting crisis looming regarding the affordability of new drugs (often biologicals-meaning antibodies) that are unique in their ability to treat uncommon diseases (called orphan diseases). Because their approval is governed by special rules these drugs can be approved quite quickly. This is laudable but often the understanding of the new drugs and their targets is funded by the public through entities like the Canadian Institutes of Health Research (CIHR). Then a company acquires the rights and does the final development of the product as a drug and costs skyrocket. These high costs do no therefore reflect research and development costs; rather they reflect the market forces (i.e.-scarce resource plus desperate patients equal high price!). Let’s examine this in the case of Soliris, the drug that was so helpful to Claire. Eculizumab (Soliris) is made by Connecticut-based Alexion Pharmaceuticals). Soliris is a humanized monoclonal antibody that binds to the terminal component of the complement cascade, which is called component 5 (C5). The complement cascade of proteins is meant to kill invading bacteria that invade our body (i.e. they are supposed to be on our side!). The complement proteins assemble and punch holes in the offending microbe. Sometimes, as in Claire’s case, the body itself becomes the target and complement wreaks havoc on our own cells. By inhibiting the complement cascade at the C5 level, there is a reasonable balance between turning off the undesirable actions of C5 that promote inflammation and destruction of cells without rendering the patient too immunosuppressed. Soliris is approved for the treatment of two condition so rare that even most physicians won’t know much about them, paroxysmal nocturnal hemoglobinuria (PNH) (literally peeing blood at night-due to breakdown of red blood cells) and atypical hemolytic uremic syndrome (aHUS)-breakdown of red blood cells associated with kidney failure (uremia). In small clinical trials of 17 and 20 patients Soliris improved recovery of renal function. In a study with 26 weeks of treatment, Soliris increased platelet count, improved kidney function and quality of life. Two-year follow-up data, from the extension studies, suggests that Soliris provides sustained inhibition of complement-mediated tissue microvascular disease and significant, improvement in renal function. Soliris is exorbitantly expensive. Claire continues to receive a Soliris infusion in the dialysis unit every three weeks. Each dose costs $29,500 which is over $500,000 year. Soliris is one of a growing number of orphan drugs (agents developed for rare diseases and usually sold at astronomical prices). These drugs represent a challenge for the health care system. Often these benefits are incremental but even when they are dramatic, as in this case, the cost is staggering. The incremental cost/life year and cost/QALY (quality-adjusted life-year; 1 QUALY =1 year of perfect health) with Soliris is very high –for example for PNH patients these costs are CAN$4.62 million and CAN$2.13 million, respectively (Coyle, D et al Medical Decision-Making. 34 (8): 1016 1029). Soliris is the only drug approved by Health Canada to treat aHUS. In December 2014 the Government of Ontario, Canada was in price negotiations with the Soliris manufacturer. On February 25, 2015 the Minister of Health in Ontario announced that patients who needed Soliris would receive it through the Exceptional Assess Program. However in reality this program is very difficult to access and often requests for Soliris are declined (read the story of Rachael-also from our SELHIN). In a revealing article in 2015, Kelly Crowe reported for the CBC on the financial dealings of Alexion. What is clear is that much of the intellectual property upon which the drug is based arose in the public sector (i.e. we the people paid for it). Full credit to Alexion for making and marketing the drug, but they reportedly are charging Canada amongst the highest prices in the world. Hopefully Claire’s family will be successful in receiving ongoing MOHLTC support for funding this outpatient therapy! Outcome: To date, Claire has almost completely recovered from her illness. She is about to graduate from high school and start university here in Kingston in the fall. Her future is bright. Her kidney function blood work, and all other blood tests, has normalized. The treatment outcome in Claire’s case is truly miraculous. Let me close with Claire’s own account of her experience. She recently delivered this speech to introduce Dr. Garland at the Human Touch Award in Toronto. Decorative Image My name is Claire Barber and Dr. Garland saved my life when I had hemolytic uremic syndrome. I know a lot about doctors. Both of my parents are doctors, their friends are doctors and my friends’ parents are doctors. I spent most weekends when I was younger doing hospital rounds with my parents and knew how to write physician orders when I was 5. But, until last summer I didn’t understand anything about being a patient or how important good doctors are. I was in the ICU for about 5 weeks last summer in Kingston. I had dialysis, an IV from my neck into my heart, lots of tubes, artificial feeding and received coolers full of blood and plasma. I gained 30 pounds in water when my kidneys failed and then I went blind, had seizures and a stroke when the HUS affected my brain. I was terrified and felt like I was trapped in a fishbowl without any privacy or any say about what was done to me. While I had amazing care from everyone in the hospital from the porters, to the nurses who let my dog come visit, to the cleaning staff who hung extra curtains so I could sleep better, I noticed that some of the doctors treated me like ‘an interesting medical case’ instead of as a person. Many spoke over my head, or spoke only to my parents or used medical words that I didn’t understand. They spoke about me but not to me. My parents and I nominated Dr. Garland for this award because she is an exceptional doctor. Dr. Garland started taking care of me while she was in New Brunswick on vacation. She had expertise in HUS and I noticed that the other doctors contacted her to get advice. She talked to them every day while on vacation and texted and phoned my mom every day. She came in to see me when she was not on call and even when she was on vacation to make sure that everything was ok. She knew to start me on a drug called Soliris that not only saved my life but also my kidneys. This drug costs thirty thousand dollars per dose and Dr. Garland advocated for the hospital to pay for this drug instead of transferring me to another hospital, so that I could stay close to my friends and family. When I got out of hospital she wrote letters that were five pages long, insisting that the insurance companies continue my coverage of Soliris, which I still get every 3 weeks. Dr. Garland arranged so many details from detailed medical plans to arranging for me to get chocolate milk in the ICU while she was still in New Brunswick. She sat on my bed to speak to me as a person and not a patient, she asks about my dog, and school and with Dr. Leifso she arranged for me to still be able to go see Drake in Toronto even when my PICC line became infected and I was supposed to be re-admitted. She always comes in to visit me when I get my Soliris infusions at the dialysis centre. Her nurses text me reminders about lab tests and email me lab results. She is still working hard for me by regularly talking to specialists from around the world and even located some of my frozen blood in Cincinnati and arranged to have it flown to Italy for an important test. Now my kidneys have recovered and I am healthy again. I am about to graduate from grade 12 and am going to Queen’s University next year. I am very thankful for all the people who took care of me in hospital but especially to Dr. Garland and her team. I am so glad that she won this award – she deserves it. Acknowledgement: Thanks to Dr. Kathie Doliszny, Dr. Karen Hall-Barber and Dr. Joceyln Garland and Dr. Dorothy Thomas for their editorial input. References:

Morigi, M et al (2011). "Alternative pathway activation of complement by shiga toxin promotes exuberant C3a formation that triggers microvascular thrombosis". J Immunol. 187 (1): 172–180. doi:4049/jimmunol.1100491. PMID 21642543.
Joshua M. et al (2009). "Alternative pathway of complement in children with diarrhea-associated hemolytic uremic syndrome". Clin J Am Soc Nephrol. (12): 1920–1924. doi:2215/CJN.02730409. PMC 2798880 . PMID 19820137.
Orth,D eta l (2009). "Shiga toxin activated complement and binds factor H: evidence for an active role of complement in hemolytic uremic syndrome". J Immunol. 182 (10): 6394–6400. doi:4049/jimmunol.0900151. PMID 19414792.
Noris et al (2014). “Dynamics of complement activation in aHUS and how to monitor eculizumab therapy”. Blood. 124 (11): 1715-26. doi: https://doi.org/10.1182/blood-2014-02-558296 , PMCID: PMC4162105