SEARCH

SEARCH BY CITATION

Keywords:

  • anticoagulation;
  • hemorrhage;
  • platelet inhibition;
  • surgery

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Review of historical data
  5. Additional points of view
  6. Definition of major bleeding in surgical studies
  7. Discussion
  8. Disclosure of Conflict of Interests
  9. References

Summary.  The definition of major bleeding varies between studies on surgical patients, particularly regarding the criteria for surgical wound-related bleeding. This diversity contributes to the difficulties in comparing data between trials. The Scientific and Standardization Committee (SSC), through its subcommittee on Control of Anticoagulation, of the International Society on Thrombosis and Haemostasis has previously published a recommendation for a harmonized definition of major bleeding in non-surgical studies. That definition has been adopted by the European Medicines Agency and is currently used in several non-surgical trials. A preliminary proposal for a parallel definition for surgical studies was presented at the 54th Annual Meeting of the SSC in Vienna, July 2008. Based on those discussions and further consultations with European and North American surgeons with experience from clinical trials a definition has been developed that should be applicable to all agents that interfere with hemostasis. The definition and the text that follows have been reviewed and approved by relevant co-chairs of the subcommittee and by the Executive Committee of the SSC. The intention is to seek approval of this definition from the regulatory authorities to enhance its incorporation into future clinical trial protocols.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Review of historical data
  5. Additional points of view
  6. Definition of major bleeding in surgical studies
  7. Discussion
  8. Disclosure of Conflict of Interests
  9. References

Major bleeding is an important safety endpoint in studies on new drugs interfering with hemostasis. In surgical studies the occurrence of major bleeding is often associated with prolonged length of stay, leading to re-intervention and utilization of additional health-care resources, and it may occasionally be fatal. In order to facilitate comparison of safety results in different studies it is crucial to harmonize the definition of major bleeding.

In 2005 the SSC Subcommittee on Control of Anticoagulation published a recommendation on a definition for major bleeding in non-surgical studies on antihemostatic agents [1]. This definition was adopted in 2006 by the European Medicines Agency (EMEA) in their update guidelines on clinical investigations in non-surgical studies [2]. That has also led to the acceptance of this definition in an increasing number of study protocols on new antithrombotic agents.

In surgical studies there is no such harmonization of the definition of major bleeding [3]. EMEA refers to the above-mentioned guidelines for non-surgical studies from the International Society on Thrombosis and Haemostasis (ISTH) [1] and suggests a number of additional criteria that should cover bleeding from the surgical site [4].

Review of historical data

  1. Top of page
  2. Abstract
  3. Introduction
  4. Review of historical data
  5. Additional points of view
  6. Definition of major bleeding in surgical studies
  7. Discussion
  8. Disclosure of Conflict of Interests
  9. References

We reviewed the publications of 34 randomized controlled trials on anticoagulant agents in surgical patients from 1997 to 2009 to identify the criteria used to define major bleeding, the interval during which data on such events was collected and the eligible study population. These were mainly studies on major joint replacements or hip fractures, and there were only three studies on general surgery. All studies had ‘bleeding in a critical organ’ as one criterion. The more recent studies also included ‘fatal bleeding’ and ‘surgical bleeding requiring re-operation’ among their criteria. ‘Bleeding warranting discontinuation of study drug’, sometimes with the additional requirement that the bleeding is overt, was used in some recent studies. In the definition of major bleeding in non-surgical studies this criterion was not recommended, because it was felt to be quite subjective. The definition of surgical wound bleeding as major depending on either quantification of blood loss, drop in hemoglobin or transfused volume demonstrated the greatest variability. The time period for collection of data was often not stated, or it was from start of surgery or from the first dose of study drug, which sometimes was postoperatively, and until approximately five half-lives after the last dose of study drug or to the day of venography or other variant. The eligible study population was mostly those who had ‘received at least one dose of the study drug’ but in a few studies ‘all patients randomized’.

Additional points of view

  1. Top of page
  2. Abstract
  3. Introduction
  4. Review of historical data
  5. Additional points of view
  6. Definition of major bleeding in surgical studies
  7. Discussion
  8. Disclosure of Conflict of Interests
  9. References

The International Surgical Thrombosis Forum published in 2008 a policy paper on both efficacy and safety endpoints in studies on major joint replacement [5]. They suggest that surgical wound bleeding data should be reported separately and not form a criterion for major (serious) bleeding. Their policy agrees with the standpoint of EMEA that the definition of major bleeding should be consistent throughout the development program of a new drug.

In the discussion with European and North American surgeons with vast experience from clinical trials it became apparent that incorporation of criteria for surgical blood loss, such as bleeding index of at least 2, transfusion of more than two units of whole blood or red cells or drop of hemoglobin of at least 20 g L−1, does not support the clinical concept of major surgical blood loss. The volume of blood loss is obviously associated with the type of surgery and its complexity. However, it was felt by several of the surgeons that a criterion covering an excessive and unexpected surgical blood loss should be used, because this is a clinically important event. Such an event should be assessed by the surgeon, taking into account revision surgery, vascular malformations or other factors that may explain an increased bleeding tendency.

A preliminary proposal for a parallel definition for surgical studies was presented at the 54th Annual Meeting of the SSC in Vienna, July 2008. During a discussion at this session a suggestion was raised that transfused volume of more than two standard deviations above the mean in historical studies could be used as a criterion for major bleeding. In our review of the published studies it became clear that reliable data only exist for primary hip or knee arthroplasty but not for revision surgery or for general surgery. Whereas arthroplasty can be considered as a relatively standardized trauma, general surgery includes a great number of different surgical procedures with large variations in blood loss. Furthermore, policies regarding transfusion as well as wound drain insertion vary between sites and surgeons. Most surgeons felt that inclusion of the time from start of surgery is important to demonstrate the overall risk, and this would also be necessary to compare across all regimens, both with pre- and postoperative start of study drug. In studies with only postoperative start of study drug separate information may preferably be provided on event rate from that time point.

Definition of major bleeding in surgical studies

  1. Top of page
  2. Abstract
  3. Introduction
  4. Review of historical data
  5. Additional points of view
  6. Definition of major bleeding in surgical studies
  7. Discussion
  8. Disclosure of Conflict of Interests
  9. References

Taking into account historical criteria and additional consultations with European and North American surgeons with experience from clinical trials and event adjudication the Subcommittee on Control of Anticoagulation has approved the following recommendation for definition of major bleeding in surgical studies.

  • 1
     Fatal bleeding, and/or
  • 2
     Bleeding that is symptomatic and occurs in a critical area or organ, such as intracranial, intraspinal, intraocular, retroperitoneal, pericardial, in a non-operated joint, or intramuscular with compartment syndrome, assessed in consultation with the surgeon, and/or
  • 3
     Extrasurgical site bleeding causing a fall in hemoglobin level of 20 g L−1 (1.24 mmol L−1) or more, or leading to transfusion of two or more units of whole blood or red cells, with temporal association within 24–48 h to the bleeding, and/or
  • 4
     Surgical site bleeding that requires a second intervention– open, arthroscopic, endovascular – or a hemarthrosis of sufficient size as to interfere with rehabilitation by delaying mobilization or delayed wound healing, resulting in prolonged hospitalization or a deep wound infection, and/or
  • 5
     Surgical site bleeding that is unexpected and prolonged and/or sufficiently large to cause hemodynamic instability, as assessed by the surgeon. There should be an associate fall in hemoglobin level of at least 20 g L−1 (1.24 mmol L−1), or transfusion, indicated by the bleeding, of at least two units of whole blood or red cells, with temporal association within 24 h to the bleeding.
  • 6
     The period for collection of these data is from start of surgery until five half-lives after the last dose of the drug with the longest half-life and with the longest treatment period (in case of unequal active treatment durations).
  • 7
     The population is those who have received at least one dose of the study drug.

It is important to note that these definitions only are valid if the study is double blind. Any other design is subject to unacceptable bias. There should be an independent, blinded adjudication committee to review all hemorrhagic complications [6].

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Review of historical data
  5. Additional points of view
  6. Definition of major bleeding in surgical studies
  7. Discussion
  8. Disclosure of Conflict of Interests
  9. References

The criteria (#1, 2 and 3) for non-surgical site bleeding are consistent with those for non-surgical studies. This is specifically the case for the drop in hemoglobin or transfusion criteria. Although it may occasionally be difficult to ascribe the proportion of such events to the non-surgical vs. the wound bleeding, the surgeons consulted felt that in most cases it will be obvious. For further detailed comments on the non-surgical site bleeding we refer to the discussion in that publication [1].

Intra-articular bleeding in the operated joint may not be considered a criterion for major bleeding (#2) because it is a normal finding. However, a large hemarthrosis that delays rehabilitation, or delayed wound healing that results in prolonged hospitalization or deep wound infection is a serious complication. If re-operation or evacuation of hemarthrosis is performed the criterion for major bleeding is fulfilled.

Reoperation (#4) does not include situations in which a few sutures are opened, for example to evacuate a small volume of blood, or arthrocentesis for diagnostic purpose. The criterion refers to the situation when the patient is brought back to the operating room and, typically under general or neuraxial anesthesia, the entire wound is opened to evacuate a hematoma and/or to control active bleeding. Needle aspiration of the knee joint to relieve a painful hemarthrosis should also count as major bleeding. In borderline cases the surgeon should assess whether the criterion is fulfilled with verification by the independent adjudication committee. Bleeding with ensuing deep infection of clinical significance will often result in reoperation and will thereby fulfill this criterion for major bleeding.

Criterion #5 is the most subjective one, but in every large surgical trial there are such cases of surgical site bleeding, which obviously fall outside the normal, are unexpected and result in utilization of additional health-care resources and therefore are important to include as major bleeds. ‘Unexpected’ implies that when all perioperative medical information except for the study drug is taken into account the bleeding is clearly outside the normal range. Conditions that need to be considered in the decision as to whether the bleeding was unexpected include the type and extent of surgery, infection, tumor invasion or other anomalous blood vessels, accidental operative blood vessel injury, revision surgery or a congenital bleeding disorder. The subjective opinion of the surgeon cannot be replaced by rigid rules in this complex situation. For the assessment of ‘prolonged’ it is important to take into account time-consuming efforts with local hemostasis or other attempts to control the bleeding.

In the assessment of the magnitude of the unexpected bleeding the surgeon should also take into account the volume of blood products, fluids and inotropic drugs given to counteract hemodynamic instability. Thus, if hypovolemic shock did not occur but would most likely have been the case with the patient left untreated, the criterion is fulfilled.

Additional information provided in the report should be on volumes of blood collected in wound drains and cell savers (in milliliters) whenever used, red cells and whole blood transfused (in units and in milliliters), and adverse and serious adverse events. These data are not part of the major bleeding criteria but are useful for comparison of secondary outcomes and should be presented. Separate information may be provided on the postoperative period when all study drugs have been initiated after surgery.

A certain degree of subjectivity in these assessments is unavoidable, both at the investigator site and at the central adjudication level. Differences in patient populations (e.g. age and co-morbidities), surgical techniques and experience of the surgeons may also influence the results. Evaluation of bleeding complications in studies with an open-label design is subjected to considerable bias and the surgeon may be inclined to ascribe the hemorrhagic events to the new antithrombotic drug. Selection of a study design with double-blind randomized comparison between the drugs will therefore always be the most valid option to reduce such effects.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Review of historical data
  5. Additional points of view
  6. Definition of major bleeding in surgical studies
  7. Discussion
  8. Disclosure of Conflict of Interests
  9. References
  • 1
    Schulman S, Kearon C. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost 2005; 3: 6924.
  • 2
    European Medicines Agency. Guideline on Clinical Investigation of Medicinal Products for Prophylaxis of Venous Thromboembolic Risk in Non-Surgical Patients. London: European Medicines Agency, 2006. ( http://www.emea.europa.eu/pdfs/human/ewp/623504en.pdf. Accessed 8 June 2009.)
  • 3
    Hull RD, Yusen R, Bergqvist D. Assessing the safety profiles of new anticoagulants for major orthopedic surgery thromboprophylaxis. Clin Appl Thromb Hemost 2009; 15: 37788.
  • 4
    European Medicines Agency. Guideline on Clinical Investigation of Medicinal Products for Prophylaxis of High Intra- and Post-Operative Venous Thromboembolic Risk. London: European Medicines Agency, 2007. ( http://www.emea.europa.eu/pdfs/human/ewp/70798en_fin.pdf. Accessed 8 June 2009)
  • 5
    Dahl OE, Borris LC, Bergqvist D, Schnack Rasmussen M, Eriksson BI, Kakkar AK, Colwell CW, Caprini JA, Fletcher J, Friedman RJ, Lassen MR, Frostick SP, Sakon M, Kwong LM, Kakkar VV. Major joint replacement. A model for antithrombotic drug development: from proof-of-concept to clinical use. Int Angiol 2008; 27: 607.
  • 6
    Bergqvist D, Clement D. Adjudication of end points in studies on substances influencing haemostasis – an example from vascular surgery. Eur J Vasc Endovasc Surg 2008; 36: 7034.