Medical errors associated with surgery and medications have long been recognized as major risks. Conversely, errors in laboratory test selection are often dismissed because of their perceived small additional cost and for being “harmless”. Current expectations are that an accurate diagnosis must be achieved rapidly and at the lowest possible cost. According to the Medical Professionalism Project published in 2002, “physicians are required to provide health care that is based on the wise and cost-effective management of limited clinical resources” . The appropriate selection of tests from the clinical laboratory has become increasingly difficult and consequently potentially wasteful, due to a quickly expanding test menu and the tremendous variability in the naming and the abbreviations for test names [2–5].
In the last decade, a number of publications have stressed the risks to the patient associated with incorrect test selection or result interpretation [6–10]. In part, this problem is fueled by often-contradictory guidelines toward establishing a diagnosis using the laboratory tests [11–13]. When the literature provides conflicting data, and particularly when experts in the field disagree in the interpretation of the literature, inconsistency in practice occurs. One example is the significant controversy about the appropriate tests to include when evaluating a patient for hypercoagulability .
The Centers for Disease Control and Prevention (CDC) convened seven institutes from 1984 to 2007 on critical issues in clinical laboratory practice (http://wwwn.cdc.gov/dls/institutes/). These institutes brought national and international experts together to focus on the role of the clinical laboratory in providing quality testing services for improved patient outcomes. The Clinical Laboratory Integration into Healthcare Collaborative (CLIHC)™, established by CDC's Division of Laboratory Science and Standards in 2008, is focusing on important “gaps” that must be filled to optimize the ability of practicing clinicians to effectively utilize laboratory services for better patient care. One major problem identified by CLIHC™ was the challenge of correct test selection, with the possible solution of widespread use of algorithms. This publication is an outcome of CLIHC™.
Methods: Algorithm Development
Three experts in diagnostic coagulation (the authors of this report) were independently asked to develop a reflex test algorithm to evaluate a patient with a prolonged partial thromboplastin time (PTT) and a normal prothrombin time (PT) based on personal experience and published data.
During regular conference calls to review each author's draft of the algorithm and reach a consensus, significant disagreement in test selection was apparent. The authors realized that more than one algorithm was required, and a second phase of the project began. They next decided on the most relevant clinical parameters to guide the development of unique test algorithms: age (adult or newborn), patient location (inpatient or outpatient), symptoms (none, bleeding or thrombosis), and timing of the abnormal PTT result (recent versus extended period of time). Ultimately, five reflex test selection algorithms were created: (1) the inpatient with a prolonged PTT on multiple occasions over an extended period of time and no history of bleeding; (2) the adult outpatient with a prolonged PTT and no bleeding or unexplained venous and/or arterial thrombosis; (3) the adult outpatient or inpatient known to have prolonged PTT over an extended period of time who is bleeding; (4) the adult outpatient with a prolonged PTT, no bleeding history, and a recent onset of significant bleeding; (5) a hospitalized newborn with a prolonged PTT.
To obtain feedback from physicians not involved in the project, the algorithms were sent to three coagulation experts (not the authors) for review, resulting in the merger of the first four algorithms into one, and revision of the fifth algorithm to apply to all children up to 6 months of age. The final versions after the incorporation of the reviewers' suggestions are shown as Figs. 1 and 2.
Results: The Prolonged PTT Algorithms
The main uses of the PTT are to monitor intravenous unfractionated heparin therapy or to investigate possible causes of a bleeding tendency. However, as at least 15 different reasons are known to produce a prolonged PTT with a normal PT, physicians may find it difficult to decide which tests to order when the PTT prolongation is unexpected (Table 1). As seen in Fig. 1, the prolonged PTT may be due to spurious causes such as heparin contamination of the blood sample (not rare in hospitalized patients) or in vitro factor VIII degradation from improper sample handling and delay in testing (often in outpatients). Lupus anticoagulant, a potentially hypercoagulable condition, congenital bleeding disorders such as hemophilia A or B, acquired or congenital von Willebrand disease (vWD), and a rare but serious condition known as acquired hemophilia are also known causes. The appropriate management of patients with each of the potential causes for the prolonged PTT differs significantly. The algorithms highlight the challenge of test selection faced by nonspecialist physicians dealing with a common clinical scenario when algorithms are not used.
Appropriate laboratory test selection is more difficult than apparent to many physicians, even experts in specialty fields of medicine. While everyone is familiar with the PT and the PTT, follow-up evaluation of a patient with one or both prolonged requires knowledge of a number of very different possibilities. It is common for physicians to order too many tests or, worse yet, miss the diagnosis by not ordering the correct ones or by misinterpreting the results of the tests chosen.
When algorithms are unavailable, nonexpert physicians in a given area need to seek timely consultation from experts who are willing to provide advice . According to the medical professionalism report, physicians have the responsibility to meticulously avoid superfluous tests and procedures to prevent harm and expense and to preserve availability of resources for all . We advocate that pathologists are perfectly positioned to guide the most efficient use of testing strategies as they work alongside clinical colleagues toward optimum patient care .
It is generally accepted that ordering physicians cannot interpret a biopsy nor read an MRI. However, it has long been assumed that every physician can interpret and evaluate abnormal results of commonly ordered screening tests. This review challenges the belief that all physicians know precisely what to do with a prolonged PTT in various clinical situations such as bleeding, thrombosis, or neither. As the PTT and other “routine” tests are ordered far more frequently than any test in anatomic pathology or in radiology, the magnitude of the wasted resources and the impact on time to diagnosis is immense.
Table I. Causes and Evaluation of Prolonged PTT
How to exclude or confirm
Heparin, unfractionated (or direct thrombin inhibitor)
Thrombin time prolonged
Heparin may be a contaminant or used therapeutically.
Intrinsic factor deficiency
Correction in mixing study followed by low factor activity [VIII, IX, XI, XII, high-molecular-weight kininogen (HMWK), or prekallikrein].
More clinically significant deficiencies (VIII and IX) are X-linked. Deficiencies of factor XII, HMWK, and prekallikrein do not cause bleeding.
Lupus anticoagulant (LA)(inhibitor of phospholipid)
Lupus anticoagulant testing demonstrates prolonged clotting times when phospholipid concentration is low (screen) and shorter clotting times when phospholipid concentration is high (confirm).
May cause falsely low intrinsic factor levels.
Specific factor inhibitor
Noncorrection in mixing study followed by low factor activity, most commonly factor VIII. Rarely, inhibitor to other factors such as IX or XI.
Inhibitor titer (Bethesda assay) may be used to check response to treatment.
PTT technical considerations:
PTT mixing study must be incubated to rule out time- and temperature-dependent inhibitor (2 hr at 37°C).
Correction in mixing study suggests factor deficiency; noncorrection suggests inhibitor.
Heparin (and direct thrombin inhibitors) must be excluded before mixing study—will act as inhibitor in the assays.
If correction in mixing study and factors VIII, IX, and XI normal, consider factor XII, HMWK, prekallikrein, and LA (LA sometimes falsely correct in mixing studies). Deficiencies of factor XII, prekallikrein, and HMWK may prolong the PTT >150 sec.
Lack of correction in the PTT mixing study in a nonbleeding patient suggests a LA and in a bleeding patient, an inhibitor (most commonly a factor VIII inhibitor).
Testing ideally performed on one blood specimen, with appropriate tests reflexively performed by the laboratory. This practice prevents the patient from having to return multiple times for extra blood draws and avoids delay in diagnosis. Once the cause of the PTT prolongation is determined, repeat testing on a fresh specimen can be useful to confirm findings.
As clinical history, age, and gender influence how likely one entity is over another, the algorithms (figures) provide a more detailed, targeted approach to test selection.
If PT also substantially prolonged, a different differential diagnosis applies.
Footnotes for Both Figures – annotated as Superscripts after the Text in the Boxes
Low-molecular-weight heparins (LMWH) do not consistently prolong the PTT. The effect is variable even when using the same PTT reagent. The presence of direct thrombin inhibitors (e.g., argatroban and dabigatran) and rivaroxaban also need to be excluded by history, and a normal thrombin time excludes direct thrombin inhibitors.
If available, order an anti-Xa assay to detect and quantitate heparin or LMWH.
A mixing study may be difficult to interpret when the PTT is less than 5 sec above the reference range. Concurrent testing for lupus anticoagulant (LA) and factors VIII, IX, XI, and XII may be informative. Correction in the PTT mixing study is not uniformly defined. Commonly used definitions include a PTT value of the mix in the reference range, to within 5 sec of the upper limit of reference range, or to within 10% of the normal plasma result. The mixing study of patients with weak LA may correct and not reveal the inhibitor. Clinical correlation will help suspect LA. To assess for a factor VIII inhibitor, a 1- or 2-hr incubation is suggested.
Unexplained bleeding may or may not be related to the prolonged PTT.
A possibility for spuriously prolonged PTT is degradation of factor VIII in vitro during handling or transport of the blood. A repeat PTT with a freshly collected specimen may be useful to show the true result.
The bleeding severity in factor VIII or IX deficiency (hemophilia A and B) is directly related to the factor level (severe < 1%, moderate 1–5%, mild > 5%). Approximately 30% of hemophiliacs have spontaneous new mutations, with negative family histories. Guidelines for the Management of Hemophilia published by the World Federation of Hemophilia may be found in this link: www.wfh.org/2/docs/Publications/…/Guidelines_Mng_Hemophilia.pdf. Factor XI deficiency is associated with bleeding in areas of high fibrinolytic activity such as mouth or urinary tract. It may also be asymptomatic. Personal or family history of bleeding helps identify patients at risk of bleeding. Factor XI deficiency is most common in Ashkenazi Jews (present in 4%) but can affect any ethnic group.
Von Willebrand disease (vWD) is diagnosed with assays for von Willebrand factor (vWF) antigen and activity, and factor VIII. vWD may be due to a quantitative defect (types 1 and 3) or qualitative abnormalities of vWF (multiple subtypes of type 2). A patient's vWF result should be interpreted based on the expected level for his/her blood type. As vWF transports factor VIII in the plasma, factor VIII often becomes decreased when vWF is decreased. The PTT becomes prolonged if factor VIII is significantly decreased due to low vWF. Type 2N vWD is an autosomal recessive disorder caused by a mutation in vWF which leads to decreased binding to factor VIII and therefore increased proteolysis of factor VIII. Acquired vWD may be difficult to differentiate from congenital depending on the patient's age. The distinction is aided by review of family history and evaluation for the presence of underlying conditions such as multiple myeloma, lymphoproliferative malignancies, mitral valve prolapse, hypothyroidism, and several drugs, which are associated with acquired VWD. The National Institutes of Health has recently published a comprehensive review of the diagnosis and management of patients with vWD .
Mild to moderate deficiencies of factors VIII, IX, and XI may be asymptomatic, particularly if the patient has never had a surgical procedure where he/she could have bled. However, as factor XII deficiency is more common, it can be excluded first in patients without a history of bleeding . If the factor XII is normal, consider measuring factors VIII, IX, and XI levels and lupus anticoagulant. Even if factor XII is low, assays for factors VIII, IX, XI, and lupus anticoagulant can be considered to exclude other possibilities (such as a combined acquired factors XI and XII deficiency due to proteinuria, or falsely low factor XII due to a lupus anticoagulant).
Patients with factor XII deficiency do not require transfusion to prevent bleeding, even if surgery is anticipated . Transient factor XII deficiency may be seen in patients awaiting tonsillectomy.
Prekallikrein (Fletcher factor) and HMWK (Fitzgerald factor) are contact factors with a poorly understood physiologic role in coagulation. Even severe deficiencies of these proteins do not cause bleeding. A PTT with prolonged incubation time may serve as a screening test for prekallikrein deficiency. Quantitative assays for prekallikrein and HMWK are typically only available from reference laboratories.
If factor VIII activity is normal, measure factors IX and XI to assess for the rare possibility of a factor IX or XI inhibitor.
Acquired hemophilia must be suspected in elderly individuals with negative bleeding history who suddenly develop soft tissue hematomas and/or persistent and significant gastrointestinal or genitourinary hemorrhage . This is a rare but serious autoimmune disorder, which can also occur in pregnancy or the postpartum period. Prompt diagnosis of acquired hemophilia is imperative because the treatment requires immunosuppressants and activated coagulation factor concentrates, since plasma is ineffective. Factors IX and XI should be performed especially if factor VIII is normal, to assess for a factor IX or XI inhibitor. Factor IX inhibitors are less common than factor VIII inhibitors, and factor XI inhibitors are very rare. Typically, the mixing study does not correct in either an immediate or incubated mix with factor IX or XI inhibitors, whereas with factor VIII inhibitors, the mix usually shows correction immediately but not after incubation. When a factor inhibitor is present, the inhibitor can interfere in the other factor assays. For example, a factor VIII inhibitor can cause falsely low factor IX and/or XI levels and the interference decreases as the sample (and the inhibitor) is diluted.
The Bethesda titer is a semiquantitative determination of concentration of the factor VIII inhibitor. A lupus anticoagulant, if present, can cause false-positive Bethesda assays, and falsely low factor assays.
Lupus anticoagulant (LA) is an antibody that binds to a plasma protein–phospholipid complex. In vitro, it prolongs phospholipid-dependent clot-based tests such as the PTT; in vivo, it can be strongly thrombophilic. LA may be transient or may indicate antiphospholipid syndrome, the most common acquired cause of hypercoagulability. The diagnosis of antiphospholipid syndrome requires two positive tests for an antiphospholipid antibody (such as LA) at least 12 weeks apart, in the setting of thrombosis or recurrent fetal loss [18, 19]. Laboratories should use two screening assays for LA due to the heterogeneity of this antibody. A prolonged result in one or both screening assays should be followed by a mixing study and confirmatory tests. As the mixing study may falsely correct in the presence of an LA, repeat testing on a separate sample or performing a confirmatory test despite correction in the mixing study may lead to a true positive result. When there is an LA in the plasma, it may cause falsely low coagulation factor activities, which may lead to the incorrect diagnosis of factor deficiency(ies). Laboratories should assess for lupus anticoagulant interference by performing factor assays at multiple dilutions.
Factor inhibitors can cause false positive LA tests. Consider measuring factors VIII, IX, XI, and XII to exclude this possibility.
While most coagulation factors are expected to be lower in children less than 6 months of age compared with adult reference ranges, factors VIII and vWF are exceptions. Factors VIII and vWF are usually higher than baseline at birth, and decline to baseline by about age 6 months. Selected reference ranges for neonates are: factor VIII: 50–178% (term) and 50–213% (premature); vWF: 50–287% (term) and 78–210% (premature); factor IX: 15–91% (term) and 19–65% (premature); and factor XI: 10–66% (term) and 8–52% (premature) [20, 21]. Guidelines for the Management of Hemophilia published by the World Federation of Hemophilia may be found in this link: www.wfh.org/2/docs/Publications/…/Guidelines_Mng_Hemophilia.pdf.
A positive LA in a child under age 6 months is usually not considered clinically significant (unless it arises from the mother via transplacental antibody transfer) and the antibody is usually expected to disappear over a couple of weeks. However, exceptions can occur.
Hemophilia carriers may be asymptomatic or have a history of bleeding during hemostatic challenges such as surgery or trauma.
Note that rare cases of factor inhibitors have been reported under age 6 months, and a PTT mixing study would assess for this possibility.
Testing should be repeated after age 6 months to determine if the result is reproducible, because in normal individuals, the adult reference range is usually reached at this time.