Karen Lusky
February 2020—In vivo hemolysis accounts for less than two percent of cases of hemolysis, but don’t assume all hemolysis is in vitro.
That was the point of one of the patient cases presented at CAP19 in a session on curbside consults in clinical pathology. William E. Winter, MD, Brenda J. Grossman, MD, MPH, and Diana S. Desai, MD, MBA, presented the cases, two of which are reported here, and spoke with CAP TODAY recently.
In the first case, a 15-year-old female had been well until the prior week when she developed lower abdominal pain, cramping, and bloody diarrhea, followed by vomiting, said Dr. Desai, associate director of blood bank and transfusion medicine, Staten Island University Hospital-Northwell Health, in introducing the case. “She was taken to the emergency department where her abdominal pain was non-localizing. She was afebrile and had mild tachycardia and a normal blood pressure.” Blood was drawn for a complete blood count and basic metabolic panel. The patient was uncooperative with the phlebotomy and had to undergo a second venipuncture. A point-of-care pregnancy test ordered in the emergency department was negative.
Dr. Winter, professor of pathology and pediatrics at the University of Florida College of Medicine, shared the laboratory results (Fig. 1) and pointed to a mild leukocytosis. “She has thrombocytopenia but not profound. We are not going to think about her bleeding excessively.” Schistocytes, which are red cell fragments, were noted on her peripheral smear. The patient’s hemoglobin, red cell count, and hematocrit were at the lower limit of their reference intervals, so she had hematologic aberrations, he said.
The basic metabolic panel (Fig. 1) showed hyperkalemia, a mild acidosis, creatinine at the upper limit of the reference interval, a slightly increased blood urea nitrogen, and elevated glucose. “Of these abnormalities, we are going to focus on her potassium and the history” of a difficult phlebotomy, Dr. Winter said, asking, “Should the patient’s hyperkalemia be attributed to hemolysis from phlebotomy?” He would argue, he said, that additional data is needed in this case. “Most of the time, hyperkalemia is preanalytical, but we don’t know.” Hyperkalemia can be endogenous or exogenous.
“If you run a lab, potassium is probably the bane of your life because undoubtedly it’s five o’clock on Friday,” and a patient whose blood was drawn at 8 AM has left the facility. “You find that the potassium is 6.5 mEq/L,” and someone has to ask the patient to return to repeat the phlebotomy to confirm or deny the presence of hyperkalemia.
In the differential diagnosis of in vitro hyperkalemia are its hemolytic causes: difficult phlebotomy, fragile cells, placement of cells in a hypotonic solution. “The causes of in vitro hyperkalemia can be a small-gauge needle,” Dr. Winter said. “You could have excessive mixing with red cell damage. You need to mix a tube a few times if it has an anticoagulant in it—and not do the Rumba.” Hyperkalemia from fragile cells can be seen with very high white counts in some forms of leukemia.
As for nonhemolytic causes, “If you used a heparinized tube, you should use lithium or sodium, not potassium heparin, if you are measuring potassium,” he noted. “Delayed centrifugation can cause hyperkalemia, because even if the red cells don’t lyse, when they run out of glucose, the sodium-potassium ATPase pump doesn’t run, and the potassium will leak out of the cells. Recentrifugation can cause hyperkalemia when potassium that has leaked out of cells mixes with the serum above the gel barrier.” With thrombocyte counts of 1 million or more, for example, there could be hyperkalemia in serum, but not in plasma.
Dr. Winter again noted the patient’s difficult blood draw. They didn’t think the patient had leukemia, he said, at least according to her total white count—although, unless extremely high, the white blood cell count is not a sensitive test for leukemia. He hoped that the lab used a green-top (heparin-containing) tube lacking potassium.
In the differential diagnosis of in vivo hyperkalemia are hemolysis (not excluded), crush injury (no history of trauma), severe renal disease/failure (creatinine not frankly elevated), acidosis (mild), and uncommon inborn errors (absent). “It could be hemolysis in the patient; people do have hemolytic anemias,” Dr. Winter said. It could be a transfusion reaction, a crush injury, or an autoimmune hemolytic anemia. In severe renal disease or renal failure, with acidosis, “hydrogen ion enters cells and potassium comes out.”
Of severe renal disease or failure, Dr. Winter said, “The creatinine was at the upper limit of the reference interval but not elevated.” What if he told the attendees this 15-year-old patient weighed 90 pounds? he said. “What do you think about her creatinine of 1.2 mg/dL? It’s within the reference interval. Do you think it’s normal?” he asked. “You’ve got to be concerned that the reference intervals are for a large population, so if you have a small patient, you can be within the reference interval and yet really have elevated creatinine.”
What laboratory tests should be ordered to look for hemolysis? “You’d certainly want to get lactate dehydrogenase, bilirubin, haptoglobin, and a reticulocyte count,” he said. He then revealed the patient’s test results, asking attendees if the results helped them “sort out” in vitro from in vivo hyperkalemia. If it’s in vivo, “that’s going to help you diagnostically,” he said. If it’s in vitro, the phlebotomy needs to be repeated.
The lactate dehydrogenase was increased, the haptoglobin was low, and the reticulocyte count and bilirubin were elevated. “Should we ascribe this hemolysis to her difficult stick?” he asked. The audience members said no; Dr. Winter agreed. “The bilirubin, haptoglobin, and reticulocyte count suggest that this is in vivo, not in vitro. So this is true hyperkalemia in the patient.”
Had the hemolysis been due to a difficult phlebotomy, “the haptoglobin would not be low, because when haptoglobin binds hemoglobin, it must be removed by the liver. That takes time,” Dr. Winter said in an interview. “If somebody is hemolyzing and the hemolysis is in vitro, you wouldn’t necessarily have anemia, and you wouldn’t expect to have a high reticulocyte count. The retic count shows the person is anemic and that their bone marrow was responding.” An elevated reticulocyte count is characteristically seen in acute hemolysis or acute hemorrhage. “It takes time to convert hemoglobin, to break it down to bilirubin, and if it was a case of in vitro hemolysis, you wouldn’t expect an elevated bilirubin.”
“On the other hand, the lactate dehydrogenase doesn’t tell you whether it’s in vivo or in vitro because that is released once the red cell breaks,” Dr. Winter adds. “The red cells have about 240 times as much LD as the plasma does, so whether it’s in vivo or in vitro hemolysis, the LD will be elevated.”
What is the cause of hemolysis and thrombocytopenia, with a low normal hemoglobin and hematocrit, or even obvious anemia? Dr. Winter asked. “What should you consider in the differential diagnosis? It’s kind of like, ‘Why did Willie Sutton rob banks?’” The audience responded: “That’s where the money is.”
“The money here,” he said, “is this is suggestive of a microangiopathic hemolytic anemia, also known as a thrombotic microangiopathy.” In microangiopathic hemolytic anemia, the red cells are being hemolyzed. He said he thinks about it this way: “You take a water balloon and throw it through a set of wires, and the balloon is going to burst. And the microvasculature is the location of the hemolysis.” Platelets are usually damaged at the same time, so the patient can have thrombocytopenia.
Microangiopathic hemolytic anemia usually results from massive or severe endothelial injury, Dr. Winter said, and can be organ-specific or generalized. It could also be caused by ultra-high molecular weight von Willebrand factor, which is a deficit of ADAMTS13 or von Willebrand factor cleaving protease. The differential diagnosis of microangiopathic hemolytic anemia could be primary, Dr. Winter said, which is thrombotic thrombocytopenic purpura (TTP) or hemolytic uremic syndrome (HUS), or it could be secondary to infection (DIC), cancer, toxemia or HELLP syndrome, severe hypertension, autoimmune disorders, or bone marrow or organ transplantation. In this patient’s case, he said, the secondary causes are less probable.
That the patient didn’t have a fever does not exclude infection. There was no evidence of cancer. The patient wasn’t pregnant, so they could rule out toxemia and HELLP syndrome. Her blood pressure was reported as normal in the emergency department. “We don’t know about autoimmune disorders, but she didn’t have any previous history of significant illness.” There was no transplant.
Dr. Winter compared the key findings in the two primary diagnoses: TTP and HUS (Fig. 2). “TTP is a problem with metabolizing ultra-high molecular weight von Willebrand factor to high-molecular weight,” he said. The patient can have von Willebrand factor strands in the microvasculature. Most cases of TTP are autoimmune. The person makes the autoantibody that interferes with the cleavage protease. About 10 percent of people with TTP have a congenital form, in which they are born without normal levels of ADAMTS13.
“HUS is really a microangiopathy within the glomerulus, typically due to Shiga toxin,” he said, which enters the circulation and binds to the endothelium, particularly in the glomerulus. “And that’s where the fibrin strands set up in the glomerulus and damage the endothelium, which damages the red cells.”
HUS and TTP features include hemolytic anemia and thrombocytopenia, Dr. Winter said. In the patient’s case, there was no fever or central nervous system changes, which are features of TTP. “There was a history of bloody diarrhea. That would favor Shiga toxin HUS as the most likely cause of her microangiopathic hemolytic anemia. And the hemolysis is in vivo, and it is leading, among other things, to the hyperkalemia in this case.” There is no specific treatment for Shiga toxin HUS; all therapy is supportive. (Some forms of atypical congenital HUS do have specific therapies, he noted.)
In vivo hemolysis accounts for less than two percent of cases of hemolysis, “but when it’s there, it should attract your attention,” Dr. Winter said. The takeaway: Don’t presume hemolysis is always in vitro.
The Shiga toxin stool test, if it were positive, would support a diagnosis of Shiga toxin-induced HUS. In an interview, Dr. Winter said his hospital laboratory now performs molecular diagnosis of the E. coli infection on the stool instead. Dr. Desai reports that Northwell has a clinical algorithm for HUS and an immunoassay for Shiga toxin in a stool sample.
Dr. Winter’s message to pathologists: “It’s our role as consultants to clinicians, both generalists as well as experts, to be able to deal with something that is as common as hyperkalemia as well as something that is as rare as HUS.”
“The main thing,” Dr. Desai says, “is that so many basic testing results and correlation of clinical symptoms can get you far in the diagnosis of many disorders.” In the HUS case, “a lot of the basic labs guided you to the diagnosis.”
[dropcap]I[/dropcap]n a second case, Dr. Desai reported that a 30-year-old woman’s only clinical history was abdominal surgery two weeks earlier, requiring a transfusion of four units of red blood cells. “Now she is being admitted through the emergency department with complaints of fatigue. She is found to have a hemoglobin of 6 grams per deciliter with normal RBC indices.” The white blood cell and platelet counts are within reference ranges, as are her electrolytes, BUN, and creatinine. Total bilirubin and lactate dehydrogenase values are elevated. The antibody screen result on admission is negative. An anemia workup reveals that the direct antiglobulin test, or DAT, is positive for IgG and complement. The haptoglobin is undetectable and the indirect bilirubin is elevated.
Dr. Grossman, professor of pathology and immunology and of medicine at Washington University in St. Louis, who managed this patient’s case, said she was going to begin with a question. “This is a patient who has a positive DAT. Given the information you have on her, what’s the most likely cause of her anemia?” The possibilities: hemolysis, rebleeding after surgery, intrinsic red cell abnormality, and bone marrow suppression.
The patient could have rebleeding because she’s two weeks post-operative, and as her surgical site remodels, she could rebleed, said Dr. Grossman, who is also medical director of transfusion medicine services at Barnes Jewish Hospital. Data from Dr. Grossman’s laboratory pointed to hemolysis. “With bone marrow suppression, she should have abnormal white cells and platelet counts. An intrinsic red cell abnormality or recent blood transfusion could cause hemolysis,” but the question was the most likely cause, “so that’s why hemolysis is the correct answer.”
Dr. Grossman said that when a patient has hemolytic anemia, they like to divide it into immune-mediated and nonimmune-mediated, and then extravascular or intravascular. The most frequent causes of immune-mediated hemolytic anemia are autoimmune disease, drug-induced hemolysis, and transfusion reactions. Nonimmune-mediated causes are intrinsic RBC abnormalities, which are uncommon, and trauma and mechanical problems. “There are osmotic [causes] when you give people hypo-osmotic fluids, and there are infections that can cause hemolysis, such as botulism.”
Dr. Grossman asked which of the following laboratory data are most in keeping with an immune etiology of the anemia: hemoglobin, reticulocyte count, RBC morphology, bilirubin, haptoglobin, lactate dehydrogenase levels, or the direct antiglobulin test? The answer: a positive DAT.
The methodology used in the direct antiglobulin test, or Coombs test, was first described in animal studies in 1908. “It wasn’t until 1945 that Coombs, Mourant, and Race actually discussed the concept of an incomplete antibody, which could not be seen in vitro in the context of hemolytic disease of the newborn until what they call Coombs reagent was added to the reaction,” Dr. Grossman said. It took two more decades for Kaplan and Garraty to demonstrate the antibody’s relationship to immune-mediated hemolysis and show how in that setting, the DAT has an 83 percent positive predictive value of an immune process.
In a direct antiglobulin test, “if your patient has IgG antibodies coating their red cells, we can’t tell that because the red cells have negative charges that cause them to repel each other. IgG is not large enough to bring the red cells together to agglutinate, so we wash these red cells to remove any unbound immunoglobulin, and what we are left with is the bound immunoglobulin.”
Next, the Coombs reagent or Coombs antisera is added, “which is an IgM that is directed against the Fc portion of human IgG or complement.” When that is added, “it binds to the Fc receptors of the IgG on the red cells and brings them together, so we can see visible agglutination. What we are asking with the DAT is, are there antibodies coating my patient’s red cells in vivo, which could be causing a problem?”
In someone without evidence of hemolysis, a positive DAT may not mean anything, Dr. Grossman cautions. Normally, everyone has a small amount of IgG on their red cells, so there will be normal people who may be at the “higher edge” of the number of IgG molecules that a DAT can detect. “Some people will have positive DATs with no consequences. So a positive DAT does not necessarily mean you are going to have shortened red cell survival.”
When the patient does have immune-mediated hemolysis, he or she will have reduced RBC survival that could be intravascular or extravascular hemolysis. If complement is activated, the patient will end up with an intravascular hemolysis, she said. If the complement is not activated, when the red cells go through the liver and spleen, they are removed from circulation. Some of the red cells have only part of their membrane removed. “Therefore, the usual discoid red cell becomes more like a spherocyte. So you see these spherocytes that are basically round without the central pallor of normal red cells.”
All of this may be observed before the patient has anemia, Dr. Grossman said. “Bone marrow has a tremendous capacity to compensate, and not until you start running out of that compensation will you start seeing anemia,” she explained. “We look for other laboratory evidence that Dr. Winter talked about for hemolysis to determine whether someone has immune-mediated hemolysis during that early period.”
A positive DAT indicates only that the hemolysis may have an immune basis, Dr. Grossman said. “But you have to look at other things also because a positive DAT may just be coincidental.” In the differential diagnosis for a positive DAT, she said, are autoimmune hemolytic anemia, drug-induced anemia, transfusion reaction, and “nonspecific acquired antibodies, such as in the situation where someone is hypergammaglobulinemic, either from exogenous IgG or from a disease like chronic lymphocytic leukemia,” and IgG is coating the red cells.
The next steps are to obtain a transfusion history and a drug history. “And in the laboratory, we will do an eluate,” Dr. Grossman said, referring to treating the patient’s red cells with something that will decrease the antibody-antigen reaction and elute the antibodies. “In our case, we usually use acid elution. So you spin it down, remove the red cells, and what you are left with is the supernatant, which should have antibody in it, if it is present.”
At that point, they perform an indirect antiglobulin test using the eluate. “We run an antibody panel against that eluate, and then it gives us a certain result,” she said. There can be two different positive eluates. One is where the eluate reacts with every cell in the red cell panel, suggesting a possible autoantibody. “Or you can have an alloantibody where the red cell panel shows a very specific antigen. And that’s what you usually see in a transfusion reaction,” because the red cells that have the antibody causing the positive DAT will be the transfused cells still in circulation.
The eluate could be negative because in a drug reaction, the drug has to be present, and there won’t be drug in the eluate. “So even though you have an antibody, there is nothing for that antibody to react to. In addition, a negative result could occur in the setting of a transfusion reaction if all transfused cells have been destroyed,” Dr. Grossman said. Finally, a non-O patient who received out of group platelets could have a positive DAT. “And since all of our antibody panel cells are O, if [the patient] has that history, you need to add an A and a B cell to that eluate panel because you could potentially have IgG from an O patient coating red cells of a non-O patient.”
Returning to the case of the 30-year-old female, Dr. Grossman said the woman’s antibody screen was negative on initial admission and readmission. On the second admission, she had a positive DAT and “laboratory evidence of hemolysis” after a recent transfusion. “Early on, a positive DAT may be the only hint that an immune process may be the mechanism of hemolysis, especially in the setting of a Kidd system antibody. In this patient, the anti-Jka was not detected in the serum on admission.” However, when working up the DAT, they noted a new anti-Jka.
Thus, they added to the patient’s history that in future transfusions she must get red cells that lack the Jka antigen. The patient had a transfusion of one unit of Jka-negative RBCs before discharge. “She was given folate and iron and sent on her way and did well,” Dr. Grossman said.
Karen Lusky is a writer in Brentwood, Tenn.