A practical approach to evaluating postoperative thrombocytopenia

Thrombocytopenia, frequently defined as a platelet count of <150 × 10⁹/L, occurs after major surgery in 30% to 60% of patients.^1,2  However, if thrombocytopenia is defined as any significant decline in platelet count (eg, 20% or 30%) compared with a preoperative baseline value, then virtually all patients undergoing major surgery experience “thrombocytopenia.”

The differential diagnosis of postoperative thrombocytopenia is broad and includes pseudothrombocytopenia (eg, spurious platelet clumping, platelet satellitism, blood drawn from above an IV infusion site), hemodilution, increased platelet clearance due to consumption or destruction, sequestration (eg, hypersplenism), or decreased platelet production.³  We aim to provide a practical approach to differentiating various pathological conditions from normal postoperative thrombocytopenia, with a focus on the timing of the platelet count decrease, its magnitude, and other associated features. We will also briefly review the distinct presentations of thrombocytopenia with artificial surface exposure such as after mechanical device implantation.

A platelet count decrease is normal and expected within 4 days of surgery, resulting from the combined effects of hemodilution along with accelerated platelet consumption related to surgical hemostasis (Figure 1). The majority of cardiac surgery patients have a nadir platelet count on postoperative days 2 to 3 (POD2 to POD3), with the platelet count returning to baseline by day 5.^4,5  In a prospective study of 581 cardiac surgery patients, 97% reached a platelet count nadir between day 1 and day 4.^1,6  Hemodilution-associated thrombocytopenia is proportional to the amount of crystalloid, colloid, and non–platelet-containing blood products given. Similar initial reductions in hemoglobin, hematocrit, and white blood cell count may also be seen.³  Dilutional thrombocytopenia manifests within minutes to a few hours following surgery, with the platelet count continuing to decrease over the following 1 to 3 days due to a combination of fluid administration and ongoing platelet consumption. Furthermore, the thrombopoietin response to acute thrombocytopenia takes 3 to 4 days to increase platelet production by the bone marrow megakaryocytes (Figure 1).

Different types of surgery may affect the degree of thrombocytopenia and subsequent platelet count recovery, likely reflecting different amounts of hemodilution and platelet consumption, including consumption from cardiac pulmonary bypass (CPB) in certain surgeries.^1,7,8  For example, 56% of cardiac surgery patients had a platelet count <150 × 10⁹/L postoperatively, compared with 28% of hip surgery patients, reflecting the greater magnitude of platelet count decrease seen in cardiac vs orthopedic surgery patients (∼50% vs ∼30%).^1,2 

The thrombopoietin response to platelet dilution and consumption results in a physiological “overshoot” in the platelet count. Postoperative platelet counts peak at two- to threefold the patient’s preoperative platelet count at approximately POD14, before gradually returning to the patient’s baseline value over the following 14 days.^2,9  This overshoot occurs because of a delay between an increase in thrombopoietin that stimulates megakaryocytes and the release of new platelets from the bone marrow. Figure 1 highlights normal thrombopoietin physiology. Healthy volunteers administered a thrombopoietin agonist showed an increased platelet count beginning at least 3 days later, with peak platelet counts achieved on days 12 to 16.⁹  Given the expected platelet count increase from the time of postoperative nadir (usually seen on POD2 and POD3 [range, POD1 to POD4]) to approximately day 14, any confirmed platelet count decline that occurs between days 4 and 14 requires investigation.

“Pseudothrombocytopenia” refers to various explanations for spurious causes of thrombocytopenia. EDTA-related platelet clumping can occur, or less commonly platelets can bind to leukocytes, resulting in the morphological appearance of platelet rosetting around polymorphonuclear leukocytes (ie, satellitism). These phenomena result from naturally occurring immunoglobulin G (IgG) antibodies that recognize EDTA-specific epitopes on platelet glycoprotein IIb/IIIa (GPIIb/IIIa) and neutrophil FcγIII receptors, respectively. Platelet clumping/satellitism occurs ex vivo, as the causative antibodies react under low calcium concentrations caused by the chelating anticoagulant, EDTA, and can be time-dependent. Examination of a peripheral blood film and redrawing the sample into a tube containing sodium citrate usually helps distinguish spurious from true thrombocytopenia: it is important to remember that the platelet count drawn into citrate will be spuriously reduced (by ∼10%) due to the dilutional or anticoagulant effect of the citrate anticoagulant solution into which the blood is drawn.¹⁰ 

Specimen collection and handling can also cause spurious thrombocytopenia. A traumatic venipuncture with ex vivo thrombin-induced platelet clumping results in a single reduced platelet value. Other causes of spurious thrombocytopenia include sample collection above a running IV fluid administration site, incorrect sample handling such as high or low temperatures, or sample movements in vacuum tube transport systems. If any unexpected platelet count occurs, a repeat platelet count value on the same day is recommended to confirm or refute the decline, along with other tests to investigate a true thrombocytopenia (see “Investigations”).

Thrombocytopenia related to existing bone marrow disorders or splenomegaly may be present prior to surgery. Immediate postoperative thrombocytopenia will be more severe and the expected postoperative thrombocytosis may not be evident.

Unexpected postoperative thrombocytopenia is caused by platelet consumption or platelet destruction. Platelet consumption refers to those disorders in which physiological processes are pathologically accelerated, such as thrombin generation or platelet/von Willebrand factor interactions in disseminated intravascular coagulation (DIC) and thrombotic microangiopathies (TMAs), respectively. Platelet destruction refers to nonphysiological processes such as direct immune-mediated platelet clearance, as seen in immune thrombocytopenia (ITP), drug-induced ITP (DITP), or posttransfusion purpura (PTP), among others (Table 1). Heparin-induced thrombocytopenia (HIT) is a unique prothrombotic disorder in which antibody-mediated activation of platelets occurs with parallel marked activation of coagulation.

The timing of thrombocytopenia is the most important feature when differentiating among the causes of postoperative thrombocytopenia. Table 1 summarizes the classification of acute thrombocytopenia in postoperative or posttrauma patients, based on early (within 4 days) vs late (POD5 or later) thrombocytopenia.

Immediate postoperative thrombocytopenia occurs because of hemodilution, consumption, and occasionally spurious results. Multiorgan failure occurs early in some postoperative patients; usually, these patients will meet criteria for DIC and/or shock. Clinical features of shock include hypotension or decreased urine output and laboratory features such as lactic acidemia, liver or kidney dysfunction, and bone marrow stress response (leukoerythroblastic picture with immature white blood cells and nucleated red blood cells). Early bacteremia and sepsis, or ongoing bleeding postoperatively, may also cause a more profound early postoperative thrombocytopenia.¹¹ 

Patients with severe DIC, ongoing shock, and associated liver dysfunction are at high risk of developing microvascular thrombosis in peripheral limbs manifesting as symmetrical peripheral gangrene. Ischemic hepatocellular injury leads to lack of protein C and antithrombin production and results in dysregulated thrombin generation. The clinical picture includes lactic acidemia (reflecting circulatory shock), circulating nucleated red cells, severe thrombocytopenia, elevated fibrin D-dimer (reflecting increased fibrin formation/fibrinolysis in DIC), and preceding acute transaminitis (indicating “shock liver”). Critical limb ischemia/necrosis usually begins or progresses 2 to 5 days after the onset of shock and shock liver, which reflects the time required to achieve critical depletion of protein C and antithrombin.¹²  In a patient who had recently received heparin, the clinical picture can suggest a (wrong) diagnosis of HIT.

Cardiac patients with implanted devices (eg, ventricular assist devices, intra-aortic balloon pumps) or indwelling catheters (eg, extracorporeal membrane oxygenation) may have additional reasons for early consumption and will be discussed in "Postoperative thrombocytopenia with artificial surface exposure."

Development of HIT prior to POD5 is due to presence of HIT antibodies resulting from recent prior heparin exposure. One scenario is a course of heparin given a few days before cardiac surgery; here, the characteristic onset of thrombocytopenia beginning 5 to 10 days after an immunizing exposure to heparin may coincide with the early postcardiac surgery period.¹³  Another scenario is administration of intraoperative heparin followed by a marked intraoperative or early postoperative platelet count decline because of “rapid-onset HIT” after an immunizing exposure to heparin in the previous 100 days.¹⁴ 

DITP rarely occurs within POD1 to POD4. However, early DIPT is possible after a single dose of GPIIb/IIIa inhibitors (due to naturally occurring antibodies), or because of reexposure to a medication to which a patient is sensitized due to previous exposure. Unlike HIT, such latent antibodies can be present for years postexposure. For example, DITP after perioperative antibiotic prophylaxis can occur in patients who previously have been exposed to the same antibiotic in the remote past (Figure 2A).

Some rare thrombocytopenic disorders, such as TMA entities like thrombotic thrombocytopenic purpura (TTP) or atypical hemolytic uremic syndrome (aHUS) can be triggered by surgery or trauma.^15,16  Features that should prompt consideration of postoperative TMA include a platelet count nadir that is below what is expected, plus the presence of fragmentation hemolysis (Figure 2B). Evidence of end-organ dysfunction with TTP or aHUS (eg, neurological or renal dysfunction) may also be present. Sometimes, patients with antiphospholipid antibody syndrome (APS), either known or previously undiagnosed, develop catastrophic APS (CAPS) following surgery or trauma.^17,18  Additional features of end-organ dysfunction (eg, renal or neurological dysfunction), or venous, arterial, or small-vessel thrombosis may be present. Severe thrombocytopenia with clinical evidence of microvascular thrombosis in several organs should prompt consideration of CAPS.

Immune causes of thrombocytopenia that reflect new antibody production usually take at least 5 days to manifest, explaining why disorders such as DITP, HIT, or PTP usually become manifest during the second postoperative week. A typical presentation includes 2 successive episodes of thrombocytopenia: the first (expected) early postoperative thrombocytopenia, followed by partial or complete platelet count recovery with a subsequent unexpected platelet count fall pointing to an immune-mediated disorder.^5,19 

DITP usually occurs after drug administration for 5 or more days, such as when medications are started in the postoperative period. Typically, severe thrombocytopenia with petechiae and purpura develop ∼1 week after starting a new medication.²⁰  Drugs most likely to be implicated for postoperative DITP include antibiotics such as vancomycin, piperacillin, cephalosporins, and trimethoprim-sulfamethoxazole; anticonvulsants such as carbamazepine; anti-inflammatories such as naproxen and ibuprofen; as well as other medications such as amiodarone, ranitidine, and haloperidol.^3,21,22  A DITP is unlikely to be present unless thrombocytopenia is severe (platelet count <20 × 10⁹/L). Exceptions to the timing of DITP are from glycoprotein IIb/IIIa inhibitors, where thrombocytopenia usually occurs within hours of the first dose. Immune thrombocytopenia without prior drug exposure is less common but possible, particularly in a patient with known ITP.

Thrombocytopenia may occur after blood product transfusion. PTP is an alloimmune reaction that causes severe thrombocytopenia ∼1 week after a blood product transfusion.³  Passive alloimmune thrombocytopenia (PAT) has a similar etiology to transfusion-associated acute lung injury where platelet-reactive alloantibodies present within a blood product cause thrombocytopenia abruptly after administrating the blood product, such as packed red blood cells or frozen plasma. Thrombocytopenia is generally severe when alloantibodies reactive against human platelet alloantigen 1a (HPA-1a) are implicated, but can be less severe if the alloantibodies against the HPA-5a/b system are responsible (Figure 2C).²³  Bacterial contamination of a blood product and associated sepsis can also cause thrombocytopenia soon after transfusion.

If heparin was administered intraoperatively or postoperatively, then HIT should be considered, especially in the absence of bleeding and infection. Most patients with HIT are postsurgical, likely reflecting the immunizing effects of combining heparin with release of platelet factor 4 (PF4) from perioperative platelet activation. Sometimes, the preceding immunizing heparin exposure is trivial (eg, heparin “flushes”) or obscure (eg, heparin administration during an interventional radiology procedure). Patients with HIT have a characteristic platelet count drop that typically begins 5 to 10 days after the preceding immunizing heparin exposure, with a >50% decrease from highest postoperative platelet count value that immediately precedes the platelet count decline.⁷  Patients with HIT may have associated venous and arterial thromboembolism (ratio of venous to arterial thrombosis of 4:1) and less commonly necrotizing skin reactions at heparin injection sites and anaphylactoid reactions within 30 minutes of receiving IV bolus heparin.^2,24  Adrenal hemorrhage or infarction is possible and presumed to be secondary to adrenal vein thrombosis. The 4Ts score is a helpful framework to identify patients at risk of HIT. If the pretest probability is high enough, defined as a 4Ts score ≥4 points, then instituting therapeutic anticoagulation with a nonheparin anticoagulant while awaiting laboratory evaluation of HIT is recommended.^25,26  Patients with HIT have evidence of overt DIC (with decreased fibrinogen or increased international normalized ratio [INR]) in ∼10% to 15% of cases.¹⁹ 

The platelet count fall of HIT can begin or worsen despite stopping heparin. Known as “delayed-onset HIT,” this condition is now classified as 1 of the “autoimmune HIT” disorders, which are characterized by the presence of pathologic HIT antibodies that can activate platelets even in the absence of heparin.^27,28  It can take several weeks for platelet count recovery in such patients (“persisting or refractory HIT”). HIT can also occur after surgery in the absence of perioperative heparin exposure. This has been reported in ∼12 patients after knee replacement surgery and could reflect immunization triggered by release of PF4/polyanion complexes released abruptly after tourniquet removal.²⁹  Patients with such “spontaneous HIT syndrome” post–knee replacement also have HIT antibodies with heparin-independent platelet-activating properties, thus classifying this condition as an autoimmune HIT disorder. Recommended treatment of this syndrome includes a nonheparin anticoagulant and adjunctive therapy with IV immunoglobin, which abruptly inhibits HIT antibody-induced platelet activation thereby deescalating HIT-associated hypercoagulability.³⁰ 

Thrombocytopenia after POD10 includes late infection, late administration of medications or blood products, or posttransfusion graft-versus-host disease. The later the onset, the more likely a fungal infection should be considered. Posttransfusion graft-versus-host disease is rare, and usually presents 3 weeks or later postoperatively. Sometimes patients are recognized as having thrombocytopenia ∼1 month postsurgery; this may simply reflect the nadir value as the platelets gradually fall from the day 14 postoperative peak.³¹ 

As discussed earlier, the severity of the platelet count decrease is an important clue that helps differentiate among the various causes of thrombocytopenia, particularly when the platelet count is <20 × 10⁹/L (Table 2). Causes of profound thrombocytopenia include septic shock; multiorgan failure; concomitant ITP, DITP, PTP, and PAT from blood transfusions; and TTP. TTP often causes a lower platelet count <20 × 10⁹/L, when compared with other TMAs such as aHUS where it is often >50 × 10⁹/L (Table 2). Unlike DITP and PTP, where platelet count nadir values are usually <20 × 10⁹/L, HIT typically causes a moderate degree of thrombocytopenia, and 80% to 90% of patients have a platelet count nadir between 20 × 10⁹/L and 150 × 10⁹/L (median platelet count nadir, ∼60 × 10⁹/L).³²  HIT may be unrecognized in postoperative patients because a >50% platelet count fall in the setting of postoperative thrombocytosis may not result in absolute thrombocytopenia (ie, a platelet count <150 × 10⁹/L) but can still be a sign of HIT. The platelet count nadir in HIT can be <20 × 10⁹/L when there is overt, decompensated DIC.

The pace of platelet count fall may also help to differentiate causes of postoperative thrombocytopenia. Immune-mediated causes of thrombocytopenia, such as HIT, may have a more rapid platelet count fall compared with non–immune-mediated causes of thrombocytopenia, such as infection or DIC.¹⁰ 

When approaching a patient with unexpected postoperative thrombocytopenia, it is important to repeat a complete blood count the same day to confirm the abnormality. Timely screening for serious conditions such as DIC and TMAs includes a peripheral blood smear, INR, partial thromboplastin time, fibrinogen, and a fibrin-specific marker (eg, D-dimer). When evaluating for a suspected TMA, testing for markers of hemolysis includes reticulocyte count, bilirubin, haptoglobin, and lactate dehydrogenase (LDH). In addition to LDH, additional markers such as aspartate aminotransferase, alanine aminotransferase, and creatine phosphokinase may be ordered to interpret whether the elevated LDH is in isolation (eg, hemolysis) or related to another cause (eg, transaminitis). In the right clinical context, an elevated alanine aminotransferase may be related to concomitant shock liver or creatine phosphokinase may be related to postoperative myocardial infarction or ischemic limb injury. If the laboratory profile points to presence of DIC or TMA, it is important to repeat these tests at least once daily as both DIC and TMAs are dynamic illnesses.

Depending on the clinical context, selected additional studies may include blood cultures, HIT testing (eg, rapid automated assays such as the latex immunoturbidimetric assay or IgG-specific chemiluminescence assay, PF4/polyanion enzyme-immunoassay, serotonin-release assay, or other platelet activation assay), antinuclear antibody or antiphospholipid antibody testing, a direct antiglobulin test (to evaluate concomitant immune-mediated hemolysis), ADAMTS13 activity and inhibitor assay and complement genetic testing (if TTP or aHUS is suspected), HIV testing, hepatitis C antibody levels, abdominal ultrasound to assess for spleen size, or a bone marrow aspirate and biopsy.^33,34 

Some postoperative patients require a device that results in exposure of blood to artificial surfaces, including an intra-aortic balloon pump, ventricular assist device, or extracorporeal membrane oxygenation. Early postoperative thrombocytopenia in patients with mechanical circulatory support is exacerbated by increased platelet consumption of critical illness, as well as increased platelet consumption on the devices. Furthermore, thrombocytopenia can persist with platelet count recovery often paralleling device discontinuation as critical illness resolves.³⁵  Heparin anticoagulation is usually required for temporary mechanical circulatory devices or a bridge to warfarin therapy. When heparin is continued at therapeutic doses for a week or more, the risk of HIT can be as high as 10%.^36,37  However, nonpathogenic anti-PF4/heparin antibodies form commonly, so HIT should be suspected only when thrombocytopenia suggestive of HIT occurs and platelet-activating antibodies are demonstrated in a serotonin release assay or other sensitive platelet activation assay.^36,38  Patients managed with mechanical circulatory support have a high risk for bleeding due to large surgeries, postoperative thrombocytopenia and device-related platelet dysfunction and coagulopathy from hyperfibrinolysis, DIC, hepatic dysfunction, anticoagulation, or acquired von Willebrand syndrome secondary to depletion of high-molecular-weight multimers from a high shear state.³⁹ 

Understanding normal thrombopoietin physiology is the crucial first step to evaluating the different causes of thrombocytopenia, with timing and magnitude of thrombocytopenia manifesting during the postoperative period as the most important features during evaluation. Although there are many potential causes of postoperative thrombocytopenia, the timing and severity of the platelet decline provide useful guidance for the clinician in formulating the differential diagnosis and planning appropriate investigations and treatment.

L.S. received support from CanVECTOR, the Canadian Thromboembolism Research Network.

Contribution: All authors contributed to writing the manuscript.

Conflict-of-interest disclosure: L.S. received honoraria from Leo Pharma and research funding from CSL Behring. L.B.K. provided consultation services for CSL Behring and the Department of Health and Human Services Vaccine Injury Compensation Program. M.A.C. received honoraria from Pfizer, CSL Behring, and Diagnostica Stago; provided consultation services to/served on advisory boards for Servier Canada, Asahi Kasei, and Precision Biologics; was on the data safety monitoring board for Bayer; has stock ownership in Alnylam; and holds the Leo Pharma Chair in Thromboembolism research, the funding for which is held in perpetuity at McMaster University (the interest is used to support M.A.C.’s research activities). T.E.W. received lecture honoraria from Alexion and Instrumentation Laboratory, and royalties from Informa (Taylor & Francis); provided consulting services to Aspen Global, Bayer, CSL Behring, Ergomed, and Octapharma; received research funding from Instrumentation Laboratory; and provided expert witness testimony relating to HIT and non‐HIT thrombocytopenic and coagulopathic disorders.

Correspondence: Leslie Skeith, University of Calgary, C210 Foothills Medical Centre, 1403 29th St NW, Calgary, AB T2N 2T9, Canada; e-mail: laskeith@ucalgary.ca.