According to one report, platelet counts on capillary (fingerstick) blood average ±3% lower than on venous blood samples and about 25% of the capillary samples were 25% or more below venous results. Platelet-associated abnormality is most commonly produced by decreased number (thrombocytopenia) or defective function (thrombocytopathia). A bleeding tendency may also appear with a greatly increased platelet count (thrombocytosis), usually not until the count exceeds 1 million/ mm3. Clinically, purpura is the hallmark of platelet abnormality. Most other types of coagulation disorders do not cause purpura.


Decrease in number is the most common platelet abnormality. In general, such conditions may be classified according to etiology:

1. Immunologic thrombocytopenia
Drug-induced thrombocytopenia
Idiopathic thrombocytopenia
Posttransfusion thrombocytopenia
Other thrombocytopenias with an immunologic component
2. Hypersplenism
3. Bone marrow deficiency
4. Other causes

Immunologic thrombocytopenia

Drug-induced thrombocytopenia. This syndrome occurs due to idiosyncratic hypersensitivity to certain drugs. It may develop during initial, continued, or intermittent use of the drug. Once hypersensitivity begins, platelet depression follows swiftly. The bone marrow most often shows a normal or increased number of megakaryocytes, which often display degenerative changes. The most frequently associated drugs are heparin, quinidine, quinine, cimetidine, and various sulfonamide derivatives; but other drugs (potassium chloride and furosemide, among others) have been incriminated in rare instances. Of course, this effect is uncommon even with the relatively frequent offenders. Platelet antibodies have been demonstrated in many cases. Intravenous heparin causes thrombocytopenia below 100,000/mm3 (µL) in about 10%-15% of patients (range, 1%-30%). About 2% develop prolonged decrease below 100,000/mm3. It has been estimated that 33%-66% of patients who receive heparin intravenously develop some degree of platelet decrease from baseline levels. Thrombocytopenia has even been reported with heparin flushes. Decrease in platelets occurs about 5-7 days (range, 2-15 days) after start of therapy. The degree of thrombocytopenia is most often mild or moderate but in some cases may be less than 50,000/mm3. Diagnosis of immune thrombocytopenia can be assisted by platelet-associated IgG measurement in the presence of the offending drug. However, at present none of the methods is easy; none detects all cases; false positive results are sometimes reported in nonimmune thrombocytopenias, and small degrees of hemolysis may interfere.

Idiopathic thrombocytopenic purpura. ITP may exist in either an acute or chronic form. The acute form is usually seen in children, has a sudden onset, lasts a few days to a few weeks, and does not recur. The majority of cases follow infection, most often viral, but some do not have a known precipitating cause. The chronic form is more common in adults; however, onset is not frequent after age 40. There are usually remissions and exacerbations over variable lengths of time. No precipitating disease or drug is usually found. Platelet antibodies have been demonstrated in 80%-90% of patients with chronic ITP. The methods used are generally based on measurement of platelet-associated IgG. Some patients eventually are found to have systemic lupus erythematosus or other diseases.

Clinically, there is purpura or other hemorrhagic manifestations. The spleen is usually not palpable, and an enlarged spleen is evidence against the diagnosis of ITP. Bone marrow aspiration shows a normal or increased number of megakaryocytes, although not always.

Posttransfusion purpura. Platelets contain certain blood group and tissue antigens on their surface; the most important of these are ABO, HLA, and platelet-specific antigen (PLA). Posttransfusion purpura usually (not always) occurs in patients whose platelets are PLA1 negative, who have been sensitized to the PLA1 antigen by previous transfusions or by pregnancy, and who are then administered blood products containing PLA1-positive platelets. An alloantibody is formed in response to sensitization. Once the antibody is formed, it exhibits rather unusual behavior for an alloantibody since it attacks both PLA1-positive or PLA1egative platelets, whether belonging to the patient or a donor. The syndrome is uncommon in spite of the fact that about 2% of Europeans are PLA1 negative. The great majority of reported patients have been female. Onset of thrombocytopenia most often occurs about 7 days after transfusion (range 2-14 days). Thrombocytopenia is often severe. The episodes last about 3 weeks, with a range of 4-120 days. They may or may not recur with future transfusion. If massive transfusions with stored bank blood are given during a short time period, thrombocytopenia frequently develops, usually ascribed to dilutional factors and to the low functional platelet content of stored bank blood. This takes at least five units of blood and usually more than 10, given within 1-2 days’ time, and may or may not be accompanied by a bleeding tendency. However, stored blood deficiency of factors V and VIII (the unstable clotting factors) may contribute to any bleeding problem.

Other thrombocytopenias with an immunologic component

Neonatal thrombocytopenia may be due to antiplatelet antibodies in maternal blood that are produced when fetal platelets contain an antigen (most commonly PLA1) that is absent on maternal platelets. Analogous to Rh immune disease, the mother produces IgG antiplatelet alloantibody that crosses the placenta to the fetal circulation. The mother and fetus are usually ABO group compatible. Infants often respond poorly to random platelet transfusion but much better to washed maternal platelets. The mother is usually not thrombocytopenic and the maternal antibody does not react with maternal platelets. Neonatal thrombocytopenia may also be produced by maternal antibodies associated with maternal ITP. About 50% of the infants of these mothers have severe thrombocytopenia. The mother’s platelet count does not reliably predict the infant’s count. In one study, about 25% of women with autoimmune thrombocytopenia and platelet counts more than 100,000/mm3 (100 Ч 109/L) delivered infants with platelet counts less than 50,000/mm3. Neonatal thrombocytopenia may also be due to other causes, such as intrauterine viral infection or neonatal sepsis.

Narcotic addicts and clinically healthy homosexual males have a high incidence of thrombocytopenia. Most, but not all, display reactive screening tests for HIV-I infection. Some have demonstrable antiplatelet antibodies and some do not.


Hypersplenism was discussed in Chapter 5. The syndrome may be primary or secondary; if secondary, it is most commonly due to portal hypertension caused by cirrhosis. There may be any combination of anemia, leukopenia, or thrombocytopenia, but isolated thrombocytopenia is a fairly frequent manifestation. The spleen is usually palpable, but not always. Bone marrow megakaryocytes are normal or increased. The thrombocytopenia seen in lupus erythematosus is usually associated with antiplatelet antibodies, but there may be an element of splenic involvement even though the spleen is often not palpable.

Bone marrow deficiency

This condition and its various etiologies were discussed in Chapter 4, the principal causes being metastatic tumor to bone, aplastic anemia, and myelofibrosis. This group forms a large and important subgroup of the thrombocytopenias and is the reason why bone marrow examination is frequently indicated in a patient with thrombocytopenia.

Thrombocytopenia is a very frequent feature of acute leukemia and monocytic leukemia, even when the peripheral blood WBC pattern is aleukemic. It may also occur in the terminal stages of chronic leukemia.

Other causes of thrombocytopenia. A miscellaneous group remains that includes various unrelated disorders, some of which will be discussed.

Microangiopathic hemolytic anemia. This is a group of conditions that share the hematologic picture of hemolytic anemia, thrombocytopenia, a considerable number of red cell schistocytes in the peripheral blood, and a tissue histologic picture of fibrin thrombi in small blood vessels. This group includes DIC, thrombotic thrombocytopenic purpura (Moschowitz’s disease), the hemolytic-uremic syndrome, the prosthetic valve hemolytic syndrome, cancer chemotherapy (rarely, cancers without chemotherapy, mostly adenocarcinomas such as prostate or stomach), Zieve’s syndrome, sepsis, and the HELLP preeclamptic syndrome. Thrombotic thrombocytopenic purpura (TTP) will be presented here as a representative example of the group. The other conditions are discussed separately elsewhere. TTP is a very uncommon disorder that occurs most frequently in young adults, although it may occur at any age. There is a characteristic triad of severe microangiopathic hemolytic anemia (96%-98% of cases), thrombocytopenia (83%-96%), and neurologic symptoms (84%-92%) that typically are multiple and shifting. About 75% of patients have the complete triad. Some also include renal disease (76%-88%) and fever (59%-98%). Hemoglobin is less than 10 gm/100 ml in about 90% of cases. The peripheral blood usually contains many schistocytes. Nucleated RBCs are present in about 20%. The direct Coombs’ test is positive in 6% (0%-7%). The white blood cell count is increased in about 55%. Serum bilirubin levels are elevated to some degree in about 80% with the unconjugated fraction predominating. Serum LDH levels are increased, and haptoglobin levels are decreased. In textbook cases, PT, APTT, fibrinogen, and FDP are all normal. However, in several series about 18% had elevated PT; 7% had elevated APTT; 7% decreased fibrinogen; and about 25% elevated FDP.

Thrombi composed of platelets with some fibrin occur in capillaries and small arterioles. Diagnosis is most definitively made through biopsy. Renal biopsy was the original recommended procedure; however, gingival or skin biopsy (if possible, of a petechia) is more common now. However, diagnostic yield from these sources is less than 40%. Differential diagnosis includes the other microangiopathic hemolytic anemias. Systemic lupus, autoimmune hemolytic anemia, and Evan’s syndrome also enter consideration of hemolytic anemia, but these anemias usually are not microangiopathic, and the results of the Coombs’ test are usually positive.

Megaloblastic anemia. Thrombocytopenia occurs as a frequent manifestation of untreated well-established B12 and folic acid deficiency anemias, sometimes even when the anemia is mild. In chronic iron deficiency anemia, platelet counts are normal and may at times actually be somewhat increased.

Infections. Thrombocytopenia has been reported in 18%-77% of patients with bacteremia. Neonatal thrombocytopenia always raises the question of sepsis. However, thrombocytopenia is not limited to actual septicemia. In one study of patients who had surgery because of intestinal perforation with peritonitis, all patients showed platelet count decrease that reached its lowest point 3-5 days after surgery at mean platelet levels about 55,000/mm3. This was followed by slow platelet count increase that reached the 100,000/mm3 level about postoperative day 10. This thrombocytopenia did not seem to produce an increased tendency to bleed and was not related to DIC, although DIC can develop in septic patients. Thrombocytopenia can occur in nonbacterial infections. It is especially associated with the congenital rubella syndrome and with HIV-I virus infection (2.6%-90% of HIV non-AIDS patients and 11% in patients with less than 250 CD4 lymphocytes). However, thrombocytopenia may occasionally and transiently be found in the early stages of other virus infections such as Epstein-Barr infectious mononucleosis. The hemolytic-uremic syndrome is a microangiopathic hemolytic anemia with thrombocytopenia and renal failure, usually following infection. It is most often seen in younger children, usually before age 5 years, and most often following onset of gastroenteritis. Today, possibly the most common cause is verotoxin-producing Escherichia coli 0157:H7, although viruses and a number of bacteria have also been incriminated.

Rheumatoid-collagen diseases. Thrombocytopenia is reported in about 15% (range, 5%-26%) of patients with systemic lupus.

Hypertension of pregnancy.—Thrombocytopenia has been associated with preeclampsia (pregnancy-induced hypertension) in 16% of cases in one study. Preeclampsia itself is reported in about 8% of pregnancies (range, 5%-15%), about 80%-85% of which are first pregnancies. About 3%-12% of preeclamptic patients (more often severe) develop microangiopathic peripheral smear changes with relatively normal coagulation studies (HELLP syndrome), a laboratory picture similar to microangiopathic hemolytic anemia. Thrombocytopenia, usually mild and transient, has been reported in about 10% (range, 0.3%-24%) of all pregnancies.

Thyrotoxicosis. Some degree of thrombocytopenia has been reported in 14%-43% of patients with Graves’ disease. Severe thrombocytopenia with bleeding is uncommon.

Uremia. As noted in the discussion of the bleeding time test, up to 50% of patients in uremia are reported to have some degree of thrombocytopenia in addition to various platelet function defects.

Artifactual thrombocytopenia. Apparent thrombocytopenia is occasionally encountered when platelet counts are performed by particle-counting machines. Some of the causes are platelet satellitosis around neutrophils in EDTA-anticoagulated blood, many giant platelets, improperly prepared specimens (platelet aggregates), and platelet cold agglutinins. Peripheral blood smears may also falsely suggest some degree of thrombocytopenia due to platelet clumping or uneven distribution if the slide is not properly made.

Platelet function defects

Platelet function usually refers to the role of platelets in blood coagulation. To carry out this role, platelets go through a series of changes, partially morphologic and partially biochemical. Abnormality may develop at various stages in this process, and platelet function tests have been devised to detect abnormality in certain of these stages. These are special procedures not available in most laboratories and include techniques designed to evaluate platelet factor release (PF-3 or serotonin), platelet aggregation (adenosine diphosphate, thrombin, collagen, epinephrine) and platelet adhesion (glass bead retention). These tests are useful primarily to categorize platelet action abnormality rather than to predict the likelihood of bleeding. The bleeding time test is probably the best procedure to evaluate degree of clinical abnormality.

Hereditary disorders of defective platelet action with normal platelet count is uncommon, the most famous of this group being Glanzmann’s disease (hereditary thrombasthenia). Platelets in Glanzmann’s disease have abnormal aggregation and glass bead retention. The clot retraction test result is abnormal, whereas the results are normal in other thrombopathic (platelet function abnormality) disorders. Tourniquet test results are variable.

Defective platelet function has been observed in many patients with uremia and some patients with chronic liver disease, even without the thrombocytopenia that occasionally may develop. Cryoprecipitate or desmopressin can correct the bleeding time elevation in uremia. Many other conditions can sometimes be associated with platelet function test abnormalities, including leukemias and myeloproliferative disorders, dysproteinemias such as myeloma, and systemic lupus. Giant platelets may be found in certain conditions, especially in myeloid metaplasia (less often in chronic myelocytic leukemia), but this does not seem to produce a clinical bleeding tendency.

Certain drugs may interfere with platelet function, as noted earlier. Aspirin affects platelet factor release and also platelet aggregation. Other drugs may interfere with one or more platelet function stages.

von Willebrand’s disease

von Willebrand’s disease combines platelet function abnormalities with deficiency in factor VIII activity. The clinical disease produced has been called “pseudohemophilia.” Although there used to be considerable argument as to just what this disease should include, it is now generally restricted to a hereditary disorder of the von Willebrand factor portion of the factor VIII/vWf complex. As described in the section on factor VIII deficiency, the factor VIII/vWf complex consists of two components, the hemophilia A factor VIII portion and the von Willebrand factor portion. Factor VIII controls factor VIII activity within the coagulation intrinsic pathway system. In hemophilia A, VIII antigenic material is present but is nonfunctional, so that VIII:C activity is decreased. The vWf portion of the complex is normal both in quantity and function. That part of the complex comprising vWf controls at least two aspects of platelet function. In von Willebrand’s disease the vWf antigen is decreased, and platelets display decreased adhesiveness (manifested by decreased retention in glass bead columns) and also decreased platelet agglutination under the stimulus of the antibiotic ristocetin. In addition, the vWf is thought to stabilize factor VIII levels, so that a decrease in vWf leads to a decrease in factor VIII levels, both in the quantity of factor VIII as well as its activity. Thus, the entire factor VIII complex is decreased.

Classic von Willebrand’s disease (and all but one variant forms) is transmitted as an autosomal dominant trait (in contrast to hemophilia A, which is transmitted as a sex-linked recessive trait). Most patients with von Willebrand’s disease are heterozygous and have a clinically mild hemorrhagic disorder. Most of the serious bleeding episodes are induced by trauma or surgery. Patients homozygous for von Willibrand’s factor deficiency are uncommon; these patients have a severe hemorrhagic disorder. In a third variant, also uncommon, vWf is present in normal quantity but is nonfunctional or only partially functional; these patients have normal factor VIII activity and normal vWf antigen by immunoassay but low ristocetin cofactor and platelet glass bead retention activity and an abnormal bleeding time.

Acquired von Willebrand’s disease. A few patients have been reported who had clinical and laboratory findings consistent with von Willebrand’s disease but had no evidence of hereditary transmission. These patients had a variety of diseases but most often seemed to have lymphoma, carcinoma, autoimmune disorders, and conditions associated with monoclonal gammopathies.

Laboratory diagnosis. The various forms of von Willebrand’s disease as well as the expected results of the various diagnostic tests are summarized in Table 8-1. The most commonly used screening tests are the bleeding time and, to a lesser extent, the APTT. As noted previously, there is a wide spectrum of clinical severity and also degree of laboratory test abnormality in patients with von Willebrand’s disease. Therefore, sensitivity of the APTT has varied from 48%-100% in different reports. The bleeding time is more reliable, but also can be normal. In the usual (“classic”) type, factor VIII activity is variably decreased, the bleeding time is prolonged, and platelet function by glass bead column retention and ristocetin aggregation is abnormal. However, even in the classic form, some patients display bleeding times that may intermittently be normal. In some patients factor VIII:C activity is normal but the bleeding time is prolonged. In others the disease is so mild that even the bleeding time is normal, but pretreatment with aspirin can uncover the bleeding time defect. As noted in the section on factor VIII, the factor VIII/vWf complex is one of the so-called acute reaction protein group, so that vWf can be increased to some degree in surgery, infection or noninfectious inflammation, severe exercise, and severe stress of other types and in mild cases may temporarily correct the bleeding time and factor VIII/vWf assay results. Increased estrogens (pregnancy or use of estrogen-containing contraceptives) can increase factor VIII activity and vWf antigen levels in some patients with von Willebrand’s disease even though the bleeding time may continue to be abnormal. In a few cases, especially when laboratory results are conflicting or equivocal, it may be necessary to obtain multimeric analysis (analysis of the factor VIII/vWf complex structure fraction sizes by special gel electrophoresis).

Table 8-1 von Willebrand’s disease variant forms*
von Willebrand's disease variant forms

Therapy. Fresh frozen plasma or cryoprecipitate contain the factor VIII complex and can temporarily correct vWf deficiency. Administration of desmopressin can temporarily stimulate factor VIII/vWf production (to levels about twice baseline) and correct the bleeding time, as was discussed in the section on hemophilia A therapy. However, desmopressin effect on vWf lasts only about 3-6 hours, which is somewhat less prolonged than the effect on factor VIII.


Most attention given to blood platelets is focused on disorders of platelet function or decreased platelet number. However, thrombocytosis may sometimes occur. If the platelet count is greater than 900,000 or 1,000,000/mm3, there is concern for the possibility of hypercoagulability leading to venous thrombosis. In one report, about 25% of patients with thrombocytosis (platelet count >900,000/mm3) had a hematologic disorder (myeloproliferative syndrome, idiopathic thrombocythemia, severe hemolytic anemia, posthemorrhage, e.); about 25%-35% had cancer; about 20% were postsplenectomy; about 20% had acute or chronic infection or inflammatory conditions; and about 10% had collagen disease. In most cases there is no clinical problem until the platelet count exceeds 1 million/mm3. When that happens there is an increased tendency to bleed and also to develop thrombosis. The most common diseases associated with very high platelet counts are idiopathic thrombocythemia and the myeloproliferative syndromes.

Mean platelet volume. Certain automatic particle counters that count platelets also calculate the mean platelet volume (MPV). The reference range apparently varies inversely with the platelet count, unless a wide reference range is established to include all patients with platelet counts within the platelet count reference range. The MPV is said to be increased in ITP and various thrombocytopenias and in conditions associated with increased platelet size such as some of the myeloproliferative syndromes and the May-Hegglin anomaly. It is very typically increased in the Bernard-Soulier syndrome, which has large platelets. The MPV is decreased in the Wiscott-Aldrich syndrome and possibly in some patients with chronic iron deficiency or aplastic anemia. Occasionally, RBC fragments may be counted as platelets, producing artifactual increase in the platelet count. However, MPV is also increased in these patients.

Vascular defects

Senile purpura is a frequent nonhereditary type of vascular fragility problem, manifested by localized purpuric lesions or small bruises developing on the extremities of older persons. The only laboratory abnormality is a positive tourniquet test result in some cases. A somewhat similar clinical condition is the easy bruising found in some young adult persons, especially women. All results of standard laboratory tests are usually normal, except for an occasionally positive tourniquet test result. Some of these persons have abnormal platelet function test results; in the majority, however, the reason for abnormality is not known. It should be mentioned that continued or intermittent bleeding from a small localized area is most often due to physical agents (e.g., repeated trauma) or to a local condition (e.g., scar tissue) that prevents normal small-vessel retraction and subsequent closure by thrombosis.

Allergic (anaphylactoid) purpura is characterized by small blotchy hemorrhages over the extremities, frequently accompanied by ankle edema, produced by an allergic capillary or small-vessel vasculitis. Many patients also have glomerulonephritis. Henoch’s purpura is a subdivision in which the bleeding occurs mainly in the GI tract. Sch?nlein’s purpura features the skin manifestations without GI involvement. The tourniquet test result is usually positive. Platelet counts and other laboratory test results are normal. Diagnosis is made through biopsy of a fresh small purpuric area.

Embolic purpura can simulate capillary fragility defects, although the skin lesions may resemble petechiae more than the larger lesions of purpura. There may, however, be some component of capillary fragility. These disorders include subacute bacterial endocarditis, fat embolism, and some cases of septicemia (although other cases of septicemia also have thrombocytopenia). The tourniquet test result is often positive and the bleeding time is variable. Other coagulation defect test results are normal (except in septicemia complicated by DIC). When the patient is seriously ill with disease of acute onset, purpura raises the question of meningococcemia.

Laboratory investigation of purpura

The etiologic diagnosis of purpura should begin with a platelet count and a complete blood count, with special emphasis on the peripheral blood smear. If the platelet count discloses thrombocytopenia, depending on the clinical circumstances a bone marrow aspiration may be justified, using preferably both a clot section and a smear technique. The clot section affords a better estimate of cellularity. The smear permits better study of morphology. This is true for megakaryocytes as well as for other types of cells. Investigation of purpura without thrombocytopenia should include a bleeding time (for overall platelet function). In a few cases a tourniquet test might be considered to test for abnormal capillary fragility. If necessary, platelet function tests may be done. These tests are not indicated in already known thrombocytopenia, because their results would not add any useful information. The other tests for hemorrhagic disease must have previously ruled out abnormality in other areas. Occasional cases of nonthrombocytopenic purpura are caused by abnormal serum proteins, which may be demonstrated by serum protein electrophoresis (then confirmed by other tests, described in Chapter 22).

Usually a bleeding tendency does not develop in thrombocytopenia until the platelet count is less than 100,000/mm3 (100 x 109/L) (direct method) and most often does not occur until the platelet count is less than 50,000/mm3. The 20,000/mm3 value is usually considered the critical level. However, some patients do not bleed even with platelet counts near zero, whereas occasionally there may be trouble with patients with counts more than 50,000/mm3. Most likely there is some element of capillary fragility involved, but the actual reason is not known at this time.