- Email: [email protected]
Microangiopathic Hemolytic Anemia as a Complication of Diabetes Mellitus
Microangiopathic Hemolytic Anemia as a Complication of Diabetes Mellitus SAM H. JAMES, MD,
ANTHONY M. MEYERS, MD
ABSTRACT: A mature-onset diabetic patient who developed microangiopathic hemolytic anemia (MBA) is presented. Although numerous causes of hemolysis are reported in the literature, MBA is a rare complication of diabetes. The proposed mechanism of hemolytic anemia is thought to be related to the abnormal formation of cell membranes in the diabetic environment. The ratio of cholesterol to phospholipid in the core of the membrane is altered in diabetics; as a result, the red blood cell wall becomes rigid and nondeformable. The abnormal cells becomes disrupted as they circulate through the microangiopathic blood vessels. The mechanism of action of the antiplatelet agents is to enhance cell membrane compliance. With improved cell-wall compliance, one can expect a reduction in hemolysis, as occurred in our patient. The literature on diabetes mellitus-related microangiopathic hemolytic anemia is reviewed. KEY INDEXING TERMS: Antiplatelet agents; Diabetes; Diabetic micro angiopathy; Microangiopathic hemolytic anemia. [Am J Med Sci 1998;315(3): 211-215.]
D
iabetic patients are known to suffer from microvascular disease. This is associated with abnormal blood flow through the microcirculation. 1 Microhemorheologic abnormalities occur as a result of the abnormal blood vessels that develop in diabetes. I Although these changes are due mainly to vascular tone and fixed vascular lesions due to atheroma in larger blood vessels, the microvasculature is affected by alterations in the formation of blood vessel walls in diabetics. I - 3 The changes that occur in From the University of the Witwatersrand, Johannesburg, South Africa. Received June 18, 1997; accepted in revised form October 17, 1997. Correspondence: Sam H. James, MD, University Medical Center #6413, 1501 North Campbell Avenue, Tucson, AZ 85724.
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES
the vessel wall membranes also occur in red blood cells, affecting their compliance and reducing the cell's ability to mold in the circulation. 4 - s Abnormalities of the phospholipidic core of cell membranes affect all cells in the blood5 and result in an alteration of cell wall function. s Less wellknown is the association of microangiopathic hemolytic anemia (MHA) with diabetes mellitus as a result of the abnormal structure of vessel walls and the red blood cells themselves. 7,s A case of diabetes-associated MHA is presented, and the literature on hemolysis in diabetes is reviewed. Based on the literature review and our patient's response to therapy, the mechanism of diabetes related hemolysis is postulated. The improvement of the microhemorheologic disturbances when using antiplatelet agents is discussed. Case Report A 47-year-old black male with type 2 diabetes of 16 years' duration presented to our hospital in a hypoglycemic state. At time of admission, he was noted to be anemic with a hematocrit of 26%. A peripheral blood smear revealed the presence of severe MHA (Fig. 1). Red cell indices confirmed the presence of hypochromic microcytic anemia. Both white cells and platelets were normal. A source of blood loss was investigated; both gastroscopy and colonoscopy were normal. There was calcification within the pancreas, and endoscopic retrograde cholangiopancreatography (ERCP) revealed mild chronic pancreatitis. Intravascular hemolysis was confirmed by the presence of an elevated lactase dehydrogenase of 502 IU per liter (90-180 lUI L). The haptoglobin was 0.2 (1-3 gm/L) and the hemopexin was 545 (500- 1150 mg/L). A cause for the MHA was sought. Past medical history was significant for hypertension, but not malignant phase hypertension. Physical examination showed that the patient had background diabetic retinopathy. His cardiovascular system was normal. Analysis of the urine revealed 8,000 red blood cells per milliliter, no white blood cells, and 1 + of hyaline casts. The 24-hour urine contained 5.8 gm of protein, and the creatinine clearance was 49 mL per minute. The patient left the hospital against medical advice, but returned 2 months later in an overtly nephrotic state. His serum creatinine had risen to 2 mg/dL and he was passing 9 grams of protein in a 24-hour urine collection. Hematocrit had dropped to 19% from a post-transfusion level of 38%. The corrected reticulocyte count was 5.4. Coombs test was negative. Prothrombin time and partial thromboplastin time were normal. Fibrin degradation product was not elevated. The antinuclear antibody test and rheumatoid factor were negative. Both complement C3 and C4 levels were in the normal range. Vascular malformations, particularly
211
Hemolytic Anemia in Diabetes
3.4 mg/dL. Proteinuria remained elevated. The patient did not return to the clinic and was lost to follow-up.
Discussion
Figure 1. Admission peripheral blood smear demonstrating fragmented red blood cells (hematoxylin and eosin X 796).
in the lungs and liver, were excluded by radiolabeled red blood cell studies. Limited search for an occult neoplasm by means of the cell markers for CEA, AFP, and bHCG, and chest and abdominal radiographs proved negative. A shortened red blood cell survival was confirmed by chromiumlabeled studies; mean cell life was 39 days. During the study, surface counts were measured over the liver, spleen, heart, and kidneys. There was no accumulation over any of these sites, confirming that the hemolysis was intravascular. Ham's acid-serum test and sucrose screening test were both normal. As his platelet count was 77,000, indium-labeled platelet survival studies indicated a shortened life span of 1 day. Platelet aggregation was hyperactive, though circulating platelet aggregates were not found. A renal biopsy confirmed the presence of diffuse and early nodular glomerulosclerosis and microaneurysm formation (Fig. 2). There was evidence of intertubular capillary thrombosis (Fig. 3). It was concluded that the patient had MHA on the basis of his diabetic disease. The patient was treated with oral pentoxifylline 400 mg 3 times per day; additionally, he was given ferrous sulfate 325 mg twice per day. Edema was treated with furosemide 250 mg, and his insulin was adjusted to obtain good sugar control. After diuresis, the hematocrit increased to 24%. Thereafter, there was no further response to the iron therapy despite a rise in percent saturation from 14% to 23%, with the serum iron increasing from 39 to 57 JIg/dL. The hematocrit did not seem to respond to the pentoxiphylline at first; after 3 months, however, there was a marked reduction in red cell fragmentation and a steady rise in the hematocrit from the baseline of 24% to 30%, where it plateaued for the next 9 months. Renal function stabilized with a serum creatinine of
212
Different causes of hemolysis in diabetic patients have been reported in the literature (Table 1), and will be discussed in this review. (It should be noted that other causes of hemolysis may occur in diabetics but are not discussed because they have not been reported in the literature.) Several interesting observations are evident in this patient. As a poorly controlled diabetic, he showed classic changes of diabetic nephropathy in the renal biopsy, which also showed the presence of micro angiopathy and intracapillary thrombosis. Poorly controlled diabetes and microangiopathy are considered pivotal to the pathogenesis of the MHA and are discussed later. Despite normal coagulation studies, there was evidence of increased platelet activity associated with platelet hyperaggregability. Intravascular thrombosis with platelet hyperaggregability can occur in several situations. In this patient, there was no evidence of thrombotic thrombocytopenic purpura, nor had he received heparin. It is possible that he had an associated antiphospholipid syndrome, as has been described on one occasion in the literature. 9 Alternately, increased intravascular coagulation associated with nephrotic syndrome could explain the thrombi seen on renal biopsy. 10 Because no other cause of hemolysis was found, it was felt that the MHA was probably related to the development of the abnormal cell membrane that occurs in diabetes. 4 - 6 The patient's anemia was out of proportion to the degree of renal failure, and it was felt that the MHA contributed to the severity of the anemia.. Iron deficiency alone was an unlikely explanation for the anemia, since 3 months of iron therapy did not correct it. The response to pentoxiphylline, however, suggests a possible relationship between the pathogenesis of the microangiopathy and the hemolytic anemia. There was clear evidence of reduced hemolysis in the peripheral blood seen after treatment. It has been shown that abnormalities of the cholesterol to phospholipid ratios in the membrane core result in increased microviscosity of the cell membrane in diabetic patients. 3 - 5 ,11 Increased microviscosity can also occur as a result of nonenzymatic glycosylation with increased advanced glycosylation end products. 3 ,12 In this patient and in the patient reported by Dupont et al.,8 there appeared to be a response to the use of antiplatelet agents. They postulate that the benefit of reducing hemolysis is achieved with the platelet-inhibiting activity of these agents. 8 Intravascular thrombosis secondary to increased platelet activity results in MHA. Thus, altering platelet activity with antiplatelet agents 13 ,14 March 1998 Volume 315 Number 3
James and Meyers
Figure 2. Renal biopsy demonstrating diabetic nodular sclerosis {white arrow} and microaneurysms of the capillary loops (black arrow) (x 796).
will reduce intravascular thrombosis and therefore decrease MHA. On the other hand, in diabetic patients it is thought that the increased rigidity of red blood cells
circulating through microangiopathic vessels results in mechanical cell disruption, with the production of microspherocytes and cell fragments seen in the peripheral blood. 7 Another explanation for the im-
Figure 3. Renal biopsy demonstrating microthrombus in a capillary (x 796).
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES
213
Hemolytic Anemia in Diabetes
Table 1. Causes of MHA in Diabetes Mellitus Diabetic microangiopathy 7,8 Sulphonylurea drugs 18- 19 Hypophosphatemia20 Antiphospholipid syndrome 9 ,21 Glucose-6-phosphate deficiencr 3 Disseminated intravascular coagulation24
provement of MHA when using pentoxiphylline is thought to be through its action on the red blood cell membrane. 14,15 By acting on the phospholipidic core ofthe red blood cell membrane, pentoxiphylline normalizes the cell's ability to mold itself as it passes through the capillary system. Improved red cell flexibility is considered to be part of the explanation for the improved circulation when using drugs like pentoxiphylline in patients with peripheral vascular disease. 14 It is postulated that the improved flexibility of the red blood cells reduces the likelihood of mechanical disruption when the cell pass through the capillaries. 7 We postulate that the response of our patient and those reported by Dupont et al. 8 to pentoxiphylline was on the basis of the microhemorheologic improvement. The effect of agents such as pentoxiphylline is to reduce aggregation of platelets and reduce their release time. Bleeding time is prolonged and platelet adhesiveness is reduced. Normalization of platelet function has been shown to reduce hemolysis. 8 Our patient's platelet count had normalized within 2 weeks of initiating treatment with pentoxiphylline. Yet the hematocrit only began to improve 3 months after the onset of therapy. Although this does not rule out the possibility that the response was due to normalized platelet activity with decreased intravascular thrombosis, it makes that explanation less likely. Other causes of MHA in diabetics can be seen in Table 1. Sulfonylurea drugs have been reported to induce hemolysis by different mechanisms. 16 It would appear that hemolysis is a class-related complication, since episodes have been reported with numerous agents including glyburide, tolbutamide, and chlorpropamide. 16 - 19 Hemolysis has been reported to occur by a drug-induced autoimmune mechanism I6 or may be hapten-mediated. 17 Either way, it is a rare side effect of sulfonylurea drugs. When glycosylated hemoglobin levels are disproportionately lower than blood glucose levels, it is recommended that patients using sulphonylurea drugs be check for hemolysis. 16 Hypophosphatemia is a well-known cause of hemolysis. 20 The hypophosphatemia that occurs with diabetes is usually mild to moderate, and corrects with the correction of the ketoacidosis. On occasion, hypophosphatemia may be severe and result in 214
acute hemolysis with fragmentation. 20 It would appear prudent to monitor the phosphate of diabetic keto acidotic patients and replace phosphate more actively when low. There are 2 case reports in the literature of antiphospholipid syndrome in diabetes 9,2\ one had MHA. 9 Antiphospholipid synrlrome was first reported in the late 1980s. 22 It is possible that some of the diabetic patients with MHA reported in the literature prior to that time may have had the syndrome. 7 ,8 Our patient had evidence of intravascular thrombosis on renal biopsy, transient thrombocytopenia, and MHA. It is possible that he had the antiphospholipid syndrome. On the other hand, his partial thromboplastin time was normal and the antinuclear antibody blood test was negative. Unfortunately, he was not tested for anticardiolipin antibodies. It would seem appropriate to include testing for the APA in such cases in the future. The presence of a prolonged partial prothrombin time, low platelet count, false positive serology, and recurrent spontaneous abortions suggests the presence of the APA. The necessary tests to confirm the diagnosis are anticardiolipin antibodies or antiphosphatidylserine IgG antibodies. Glucose 6-phosphate deficiency is relatively common in American black male patients. Therefore, it is reasonable to expect this condition to coexist in diabetics. 23 It should be noted that it is unnecessary to prescribe drugs that catalyze oxidation in order to provoke a hemolytic episode. 23 Hemolysis in this situation tends to occur in episodes. Therefore, as in our case, chronic hemolysis makes the diagnosis of glucose 6-phosphate deficiency unlikely. Chronic disseminated intravascular coagulation causing MHA has also been reported in diabetes. 24 This would be difficult to distinguish from our patient's presentation without appropriate coagulation studies. When patients present acutely with an underlying disease such as septicemia complicated by acute hemolysis, the diagnosis of disseminated intravascular coagulation is easily made. MHA occurs as a result of intravascular thrombosis. Hemolysis occurs in about 30% of patients and is usually mild when compared with hemolytic uremic syndrome and related diseases.24 In conclusion, it would appear that persistent MHA is a rare complication of diabetes. When discovered, antidiabetic medications should be studied and withdrawn if being used. Once all other causes ofMHA have been excluded, the possibility of diabetes-induced MHA should be considered. Although not proven, antiplatelet agents such as pentoxiphylline can be administered in an attempt to ameliorate the anemia. References 1. Tooke JE. Microvascular haemodynamics in diabetes mellitus. Clin Sci. 1986; 70:119-25. March 1998 Volume 315 Number 3
James and Meyers
2. Zatz R, Brenner BM. Pathogenesis of diabetic microangiopathy: the hemodynamic view. Am J Med. 1986;80:443-52. 3. Barnett All. Origin of the microangiopathic changes in diabetes. Eye. 1993;7(2):218-22. 4. Przybylska M, Bryszewska M, Chapman IV. Thermal properties and fluidity of human erythrocyte membranes in diabetes mellitus. Int J Radiat BioI. 1993;63(3):419-24. 5. Labrouche S, Freyburger G, Gin H, Boisseau MR, Cassagne C. Changes in phospholipid composition of blood cell membranes (erythrocyte, platelet, and polymorphonuclear) in different types of diabetes: clinical and biological correlations. Metabolism. 1996;45(1):57-71. 6. Ruiz-Gutierrez V, Stiefel P, Villar J, Garcia-Donas MA, Acosta D, Carneado J. Cell membrane fatty acid composition in type I (insulin-dependent) diabetic patients: relationship with sodium transport abnormalities and metabolic control. Diabetalogia. 1993; 36(9):850-6. 7. Brunning RD, Jacob HS, Brenckman WD, JimenezPasquau F, Goetz FC. Fragmentation haemolysis in patients with severe diabetic angiopathy. Br J Haematol. 1976;34:283-9. 8. Dupont AG, Van Der Niepen P, Sennesael J, Somers G. Diabetic microangiopathic hemolytic anemia: beneficial effect of an antiplatelet agent? Diabetes Care. 1985;8(2):169-71. 9. Morita H, Suwa T, Daidah H, Takeda N, Ishizuka T, Yasuda K. Case report: diabetic microangiopathic hemolytic anemia and thrombocytopenia with antiphospholipid syndrome. Am J Med Sci. 1996;311(3):148-51. 10. Harris RC, Ismail N. Extrarenal complications of the nephrotic syndrome. Am J Kidney Dis. 1994;23(4):477-97. 11. Bryszewska M, Watala C, Torzecka W. Changes in fluidity and composition of erythrocyte membranes and in composition of plasma lipids in Type I diabetes. Br J Haematol. 1986;62:111-6. 12. Brownlee M, Vlassara H, Cerami A. Nonenzymatic glycosylation and the pathogenesis of diabetic complications. Ann Intern Med. 1984;101:527-37. 13. Panak E, Maffrand JP, Picard-Fraire C, Vallee E, Blanchard R, Roncucci R. Ticlopidine: a promise for the
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES
14. 15.
16. 17. 18. 19. 20. 21.
22.
23. 24.
prevention and treatment of thrombosis and its complications. Haemostasis. 1983; 13(suppll): 1-54. Aviado DM, Dettelbach HR. Pharmacology of pentoxifylline: a hemorheologic agent for the treatment of intermittent claudication. J Vasc Dis. 1984; 35(7):407 -17. Solerte SB, Ferrari E. Diabetic retinal vascular complications and erythrocyte filtrability: results of a 2-year followup study with pentoxifylline. Pharmatherapeutica. 1985; 4(6):341-50. Abbate SL, Hoogwerf BJ. Hemolytic anemia associated with sulfonylurea use. Diabetes Care. 1990; 13(8):904-5. Malacarne P, Castaldi G, Bertusi M, Zavagli G. Tolbutamide-induced hemolytic anemia. Diabetes. 1977;26(2):1568. Pavlic GJ. Possible potentiation of hemolytic anemia by tolbutamide. JAMA. 1966; 197(1):57 -9. Or R, Merin E, Stupp Y, Matzner Y. Chlorpropamide-induced hemolytic anemia. Drug Intell Clin Pharm. 1984; 18:981-2. Shilo S, Werner D, Hershko C. Acute hemolytic anemia caused by severe hypophosphatemia in diabetic ketoacidosis. Acta Haemat. 1985;73:55-7. Manzanares JM, Conget I, Rodriguez-Villar C, Tassies D, Fernandez-Fernandez F, Cervera R, et al. Antiphospholipid syndrome in a patient with type I diabetes presenting as retinal occlusion [letter]. Diabetes Care. 1996; 10(1):92-4. Henriksen R, Sogaard PE, Grennert L, Hansen BU, Manthorpe R, Nilsson 1M. Autoimmune antibodies and pregnancy outcome in women with false-positive syphilis test results: a retrospective controlled investigation of women from 5170 deliveries. Acta Obstet Gynecol Scand. 1989;68:537-40. Gellady AM, Greenwood RD. G-6-PD hemolytic anemia complicating diabetic ketoacidosis. J Pediatr. 1972; 80(6):1037-8. Aoki Y, Yazaki K, Shirotori K, Oguchi H, Kiyosawa K, Furuta S. Insulin-dependent diabetes mellitus showing microangiopathic hemolytic anemia and chronic disseminated intravascular coagulation. Intern Med. 1992;31(11):1310-2.
215
ANTHONY M. MEYERS, MD
ABSTRACT: A mature-onset diabetic patient who developed microangiopathic hemolytic anemia (MBA) is presented. Although numerous causes of hemolysis are reported in the literature, MBA is a rare complication of diabetes. The proposed mechanism of hemolytic anemia is thought to be related to the abnormal formation of cell membranes in the diabetic environment. The ratio of cholesterol to phospholipid in the core of the membrane is altered in diabetics; as a result, the red blood cell wall becomes rigid and nondeformable. The abnormal cells becomes disrupted as they circulate through the microangiopathic blood vessels. The mechanism of action of the antiplatelet agents is to enhance cell membrane compliance. With improved cell-wall compliance, one can expect a reduction in hemolysis, as occurred in our patient. The literature on diabetes mellitus-related microangiopathic hemolytic anemia is reviewed. KEY INDEXING TERMS: Antiplatelet agents; Diabetes; Diabetic micro angiopathy; Microangiopathic hemolytic anemia. [Am J Med Sci 1998;315(3): 211-215.]
D
iabetic patients are known to suffer from microvascular disease. This is associated with abnormal blood flow through the microcirculation. 1 Microhemorheologic abnormalities occur as a result of the abnormal blood vessels that develop in diabetes. I Although these changes are due mainly to vascular tone and fixed vascular lesions due to atheroma in larger blood vessels, the microvasculature is affected by alterations in the formation of blood vessel walls in diabetics. I - 3 The changes that occur in From the University of the Witwatersrand, Johannesburg, South Africa. Received June 18, 1997; accepted in revised form October 17, 1997. Correspondence: Sam H. James, MD, University Medical Center #6413, 1501 North Campbell Avenue, Tucson, AZ 85724.
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES
the vessel wall membranes also occur in red blood cells, affecting their compliance and reducing the cell's ability to mold in the circulation. 4 - s Abnormalities of the phospholipidic core of cell membranes affect all cells in the blood5 and result in an alteration of cell wall function. s Less wellknown is the association of microangiopathic hemolytic anemia (MHA) with diabetes mellitus as a result of the abnormal structure of vessel walls and the red blood cells themselves. 7,s A case of diabetes-associated MHA is presented, and the literature on hemolysis in diabetes is reviewed. Based on the literature review and our patient's response to therapy, the mechanism of diabetes related hemolysis is postulated. The improvement of the microhemorheologic disturbances when using antiplatelet agents is discussed. Case Report A 47-year-old black male with type 2 diabetes of 16 years' duration presented to our hospital in a hypoglycemic state. At time of admission, he was noted to be anemic with a hematocrit of 26%. A peripheral blood smear revealed the presence of severe MHA (Fig. 1). Red cell indices confirmed the presence of hypochromic microcytic anemia. Both white cells and platelets were normal. A source of blood loss was investigated; both gastroscopy and colonoscopy were normal. There was calcification within the pancreas, and endoscopic retrograde cholangiopancreatography (ERCP) revealed mild chronic pancreatitis. Intravascular hemolysis was confirmed by the presence of an elevated lactase dehydrogenase of 502 IU per liter (90-180 lUI L). The haptoglobin was 0.2 (1-3 gm/L) and the hemopexin was 545 (500- 1150 mg/L). A cause for the MHA was sought. Past medical history was significant for hypertension, but not malignant phase hypertension. Physical examination showed that the patient had background diabetic retinopathy. His cardiovascular system was normal. Analysis of the urine revealed 8,000 red blood cells per milliliter, no white blood cells, and 1 + of hyaline casts. The 24-hour urine contained 5.8 gm of protein, and the creatinine clearance was 49 mL per minute. The patient left the hospital against medical advice, but returned 2 months later in an overtly nephrotic state. His serum creatinine had risen to 2 mg/dL and he was passing 9 grams of protein in a 24-hour urine collection. Hematocrit had dropped to 19% from a post-transfusion level of 38%. The corrected reticulocyte count was 5.4. Coombs test was negative. Prothrombin time and partial thromboplastin time were normal. Fibrin degradation product was not elevated. The antinuclear antibody test and rheumatoid factor were negative. Both complement C3 and C4 levels were in the normal range. Vascular malformations, particularly
211
Hemolytic Anemia in Diabetes
3.4 mg/dL. Proteinuria remained elevated. The patient did not return to the clinic and was lost to follow-up.
Discussion
Figure 1. Admission peripheral blood smear demonstrating fragmented red blood cells (hematoxylin and eosin X 796).
in the lungs and liver, were excluded by radiolabeled red blood cell studies. Limited search for an occult neoplasm by means of the cell markers for CEA, AFP, and bHCG, and chest and abdominal radiographs proved negative. A shortened red blood cell survival was confirmed by chromiumlabeled studies; mean cell life was 39 days. During the study, surface counts were measured over the liver, spleen, heart, and kidneys. There was no accumulation over any of these sites, confirming that the hemolysis was intravascular. Ham's acid-serum test and sucrose screening test were both normal. As his platelet count was 77,000, indium-labeled platelet survival studies indicated a shortened life span of 1 day. Platelet aggregation was hyperactive, though circulating platelet aggregates were not found. A renal biopsy confirmed the presence of diffuse and early nodular glomerulosclerosis and microaneurysm formation (Fig. 2). There was evidence of intertubular capillary thrombosis (Fig. 3). It was concluded that the patient had MHA on the basis of his diabetic disease. The patient was treated with oral pentoxifylline 400 mg 3 times per day; additionally, he was given ferrous sulfate 325 mg twice per day. Edema was treated with furosemide 250 mg, and his insulin was adjusted to obtain good sugar control. After diuresis, the hematocrit increased to 24%. Thereafter, there was no further response to the iron therapy despite a rise in percent saturation from 14% to 23%, with the serum iron increasing from 39 to 57 JIg/dL. The hematocrit did not seem to respond to the pentoxiphylline at first; after 3 months, however, there was a marked reduction in red cell fragmentation and a steady rise in the hematocrit from the baseline of 24% to 30%, where it plateaued for the next 9 months. Renal function stabilized with a serum creatinine of
212
Different causes of hemolysis in diabetic patients have been reported in the literature (Table 1), and will be discussed in this review. (It should be noted that other causes of hemolysis may occur in diabetics but are not discussed because they have not been reported in the literature.) Several interesting observations are evident in this patient. As a poorly controlled diabetic, he showed classic changes of diabetic nephropathy in the renal biopsy, which also showed the presence of micro angiopathy and intracapillary thrombosis. Poorly controlled diabetes and microangiopathy are considered pivotal to the pathogenesis of the MHA and are discussed later. Despite normal coagulation studies, there was evidence of increased platelet activity associated with platelet hyperaggregability. Intravascular thrombosis with platelet hyperaggregability can occur in several situations. In this patient, there was no evidence of thrombotic thrombocytopenic purpura, nor had he received heparin. It is possible that he had an associated antiphospholipid syndrome, as has been described on one occasion in the literature. 9 Alternately, increased intravascular coagulation associated with nephrotic syndrome could explain the thrombi seen on renal biopsy. 10 Because no other cause of hemolysis was found, it was felt that the MHA was probably related to the development of the abnormal cell membrane that occurs in diabetes. 4 - 6 The patient's anemia was out of proportion to the degree of renal failure, and it was felt that the MHA contributed to the severity of the anemia.. Iron deficiency alone was an unlikely explanation for the anemia, since 3 months of iron therapy did not correct it. The response to pentoxiphylline, however, suggests a possible relationship between the pathogenesis of the microangiopathy and the hemolytic anemia. There was clear evidence of reduced hemolysis in the peripheral blood seen after treatment. It has been shown that abnormalities of the cholesterol to phospholipid ratios in the membrane core result in increased microviscosity of the cell membrane in diabetic patients. 3 - 5 ,11 Increased microviscosity can also occur as a result of nonenzymatic glycosylation with increased advanced glycosylation end products. 3 ,12 In this patient and in the patient reported by Dupont et al.,8 there appeared to be a response to the use of antiplatelet agents. They postulate that the benefit of reducing hemolysis is achieved with the platelet-inhibiting activity of these agents. 8 Intravascular thrombosis secondary to increased platelet activity results in MHA. Thus, altering platelet activity with antiplatelet agents 13 ,14 March 1998 Volume 315 Number 3
James and Meyers
Figure 2. Renal biopsy demonstrating diabetic nodular sclerosis {white arrow} and microaneurysms of the capillary loops (black arrow) (x 796).
will reduce intravascular thrombosis and therefore decrease MHA. On the other hand, in diabetic patients it is thought that the increased rigidity of red blood cells
circulating through microangiopathic vessels results in mechanical cell disruption, with the production of microspherocytes and cell fragments seen in the peripheral blood. 7 Another explanation for the im-
Figure 3. Renal biopsy demonstrating microthrombus in a capillary (x 796).
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES
213
Hemolytic Anemia in Diabetes
Table 1. Causes of MHA in Diabetes Mellitus Diabetic microangiopathy 7,8 Sulphonylurea drugs 18- 19 Hypophosphatemia20 Antiphospholipid syndrome 9 ,21 Glucose-6-phosphate deficiencr 3 Disseminated intravascular coagulation24
provement of MHA when using pentoxiphylline is thought to be through its action on the red blood cell membrane. 14,15 By acting on the phospholipidic core ofthe red blood cell membrane, pentoxiphylline normalizes the cell's ability to mold itself as it passes through the capillary system. Improved red cell flexibility is considered to be part of the explanation for the improved circulation when using drugs like pentoxiphylline in patients with peripheral vascular disease. 14 It is postulated that the improved flexibility of the red blood cells reduces the likelihood of mechanical disruption when the cell pass through the capillaries. 7 We postulate that the response of our patient and those reported by Dupont et al. 8 to pentoxiphylline was on the basis of the microhemorheologic improvement. The effect of agents such as pentoxiphylline is to reduce aggregation of platelets and reduce their release time. Bleeding time is prolonged and platelet adhesiveness is reduced. Normalization of platelet function has been shown to reduce hemolysis. 8 Our patient's platelet count had normalized within 2 weeks of initiating treatment with pentoxiphylline. Yet the hematocrit only began to improve 3 months after the onset of therapy. Although this does not rule out the possibility that the response was due to normalized platelet activity with decreased intravascular thrombosis, it makes that explanation less likely. Other causes of MHA in diabetics can be seen in Table 1. Sulfonylurea drugs have been reported to induce hemolysis by different mechanisms. 16 It would appear that hemolysis is a class-related complication, since episodes have been reported with numerous agents including glyburide, tolbutamide, and chlorpropamide. 16 - 19 Hemolysis has been reported to occur by a drug-induced autoimmune mechanism I6 or may be hapten-mediated. 17 Either way, it is a rare side effect of sulfonylurea drugs. When glycosylated hemoglobin levels are disproportionately lower than blood glucose levels, it is recommended that patients using sulphonylurea drugs be check for hemolysis. 16 Hypophosphatemia is a well-known cause of hemolysis. 20 The hypophosphatemia that occurs with diabetes is usually mild to moderate, and corrects with the correction of the ketoacidosis. On occasion, hypophosphatemia may be severe and result in 214
acute hemolysis with fragmentation. 20 It would appear prudent to monitor the phosphate of diabetic keto acidotic patients and replace phosphate more actively when low. There are 2 case reports in the literature of antiphospholipid syndrome in diabetes 9,2\ one had MHA. 9 Antiphospholipid synrlrome was first reported in the late 1980s. 22 It is possible that some of the diabetic patients with MHA reported in the literature prior to that time may have had the syndrome. 7 ,8 Our patient had evidence of intravascular thrombosis on renal biopsy, transient thrombocytopenia, and MHA. It is possible that he had the antiphospholipid syndrome. On the other hand, his partial thromboplastin time was normal and the antinuclear antibody blood test was negative. Unfortunately, he was not tested for anticardiolipin antibodies. It would seem appropriate to include testing for the APA in such cases in the future. The presence of a prolonged partial prothrombin time, low platelet count, false positive serology, and recurrent spontaneous abortions suggests the presence of the APA. The necessary tests to confirm the diagnosis are anticardiolipin antibodies or antiphosphatidylserine IgG antibodies. Glucose 6-phosphate deficiency is relatively common in American black male patients. Therefore, it is reasonable to expect this condition to coexist in diabetics. 23 It should be noted that it is unnecessary to prescribe drugs that catalyze oxidation in order to provoke a hemolytic episode. 23 Hemolysis in this situation tends to occur in episodes. Therefore, as in our case, chronic hemolysis makes the diagnosis of glucose 6-phosphate deficiency unlikely. Chronic disseminated intravascular coagulation causing MHA has also been reported in diabetes. 24 This would be difficult to distinguish from our patient's presentation without appropriate coagulation studies. When patients present acutely with an underlying disease such as septicemia complicated by acute hemolysis, the diagnosis of disseminated intravascular coagulation is easily made. MHA occurs as a result of intravascular thrombosis. Hemolysis occurs in about 30% of patients and is usually mild when compared with hemolytic uremic syndrome and related diseases.24 In conclusion, it would appear that persistent MHA is a rare complication of diabetes. When discovered, antidiabetic medications should be studied and withdrawn if being used. Once all other causes ofMHA have been excluded, the possibility of diabetes-induced MHA should be considered. Although not proven, antiplatelet agents such as pentoxiphylline can be administered in an attempt to ameliorate the anemia. References 1. Tooke JE. Microvascular haemodynamics in diabetes mellitus. Clin Sci. 1986; 70:119-25. March 1998 Volume 315 Number 3
James and Meyers
2. Zatz R, Brenner BM. Pathogenesis of diabetic microangiopathy: the hemodynamic view. Am J Med. 1986;80:443-52. 3. Barnett All. Origin of the microangiopathic changes in diabetes. Eye. 1993;7(2):218-22. 4. Przybylska M, Bryszewska M, Chapman IV. Thermal properties and fluidity of human erythrocyte membranes in diabetes mellitus. Int J Radiat BioI. 1993;63(3):419-24. 5. Labrouche S, Freyburger G, Gin H, Boisseau MR, Cassagne C. Changes in phospholipid composition of blood cell membranes (erythrocyte, platelet, and polymorphonuclear) in different types of diabetes: clinical and biological correlations. Metabolism. 1996;45(1):57-71. 6. Ruiz-Gutierrez V, Stiefel P, Villar J, Garcia-Donas MA, Acosta D, Carneado J. Cell membrane fatty acid composition in type I (insulin-dependent) diabetic patients: relationship with sodium transport abnormalities and metabolic control. Diabetalogia. 1993; 36(9):850-6. 7. Brunning RD, Jacob HS, Brenckman WD, JimenezPasquau F, Goetz FC. Fragmentation haemolysis in patients with severe diabetic angiopathy. Br J Haematol. 1976;34:283-9. 8. Dupont AG, Van Der Niepen P, Sennesael J, Somers G. Diabetic microangiopathic hemolytic anemia: beneficial effect of an antiplatelet agent? Diabetes Care. 1985;8(2):169-71. 9. Morita H, Suwa T, Daidah H, Takeda N, Ishizuka T, Yasuda K. Case report: diabetic microangiopathic hemolytic anemia and thrombocytopenia with antiphospholipid syndrome. Am J Med Sci. 1996;311(3):148-51. 10. Harris RC, Ismail N. Extrarenal complications of the nephrotic syndrome. Am J Kidney Dis. 1994;23(4):477-97. 11. Bryszewska M, Watala C, Torzecka W. Changes in fluidity and composition of erythrocyte membranes and in composition of plasma lipids in Type I diabetes. Br J Haematol. 1986;62:111-6. 12. Brownlee M, Vlassara H, Cerami A. Nonenzymatic glycosylation and the pathogenesis of diabetic complications. Ann Intern Med. 1984;101:527-37. 13. Panak E, Maffrand JP, Picard-Fraire C, Vallee E, Blanchard R, Roncucci R. Ticlopidine: a promise for the
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES
14. 15.
16. 17. 18. 19. 20. 21.
22.
23. 24.
prevention and treatment of thrombosis and its complications. Haemostasis. 1983; 13(suppll): 1-54. Aviado DM, Dettelbach HR. Pharmacology of pentoxifylline: a hemorheologic agent for the treatment of intermittent claudication. J Vasc Dis. 1984; 35(7):407 -17. Solerte SB, Ferrari E. Diabetic retinal vascular complications and erythrocyte filtrability: results of a 2-year followup study with pentoxifylline. Pharmatherapeutica. 1985; 4(6):341-50. Abbate SL, Hoogwerf BJ. Hemolytic anemia associated with sulfonylurea use. Diabetes Care. 1990; 13(8):904-5. Malacarne P, Castaldi G, Bertusi M, Zavagli G. Tolbutamide-induced hemolytic anemia. Diabetes. 1977;26(2):1568. Pavlic GJ. Possible potentiation of hemolytic anemia by tolbutamide. JAMA. 1966; 197(1):57 -9. Or R, Merin E, Stupp Y, Matzner Y. Chlorpropamide-induced hemolytic anemia. Drug Intell Clin Pharm. 1984; 18:981-2. Shilo S, Werner D, Hershko C. Acute hemolytic anemia caused by severe hypophosphatemia in diabetic ketoacidosis. Acta Haemat. 1985;73:55-7. Manzanares JM, Conget I, Rodriguez-Villar C, Tassies D, Fernandez-Fernandez F, Cervera R, et al. Antiphospholipid syndrome in a patient with type I diabetes presenting as retinal occlusion [letter]. Diabetes Care. 1996; 10(1):92-4. Henriksen R, Sogaard PE, Grennert L, Hansen BU, Manthorpe R, Nilsson 1M. Autoimmune antibodies and pregnancy outcome in women with false-positive syphilis test results: a retrospective controlled investigation of women from 5170 deliveries. Acta Obstet Gynecol Scand. 1989;68:537-40. Gellady AM, Greenwood RD. G-6-PD hemolytic anemia complicating diabetic ketoacidosis. J Pediatr. 1972; 80(6):1037-8. Aoki Y, Yazaki K, Shirotori K, Oguchi H, Kiyosawa K, Furuta S. Insulin-dependent diabetes mellitus showing microangiopathic hemolytic anemia and chronic disseminated intravascular coagulation. Intern Med. 1992;31(11):1310-2.
215