- Email: [email protected]
A prospective multicenter evaluation of preoperative hemostatic screening tests
A Prospective Multicenter Evaluation of Preoperative Hemostatic Screening Tests Sidney Houry, MD, Paris, Constantin Georgeac, MD, Jean-Marie Hay, MD, Colombes, Abe Fingerhut, MD, FACS, Poissy, Marie-Jeanne Boudet, MD, Cr&ei/, France, on behalf of The French Associations for Surgical Research’
BACKGROUND: Several retrospective and four prospective reports have questioned the need for routine preoperative hemostatic screening tests (PHST) in general surgery. PATIENTSAND METHODS: The results of four standard tests (prothrombin time, activated partial thromboplastin time, platelet count, and bleeding time) were prospectively compared with patient history and clinical data in a multicenter study of 3,242 consecutive patients. The patients were divided into four groups: group A (n = 1,951) had no clinical or PHST abnormalities; group B (n = 340) had no clinical and one or more PHST abnormalities; group C (n = 779) had one or more clinical and no PHST abnormalities; group D (n = 172) had both clinical and PHST abnormalities. RESULTS: Preoperative modifications of guidelines (postponed operations and ordering of additional hemostatic tests) were significantly more frequent in both groups of patients with PHST abnormalities (groups B and D), but specific treatment to correct hemostatic disorders was prescribed only when clinical abnormalities were also present (group D). Intraoperatively, modifications of anesthetic and surgical vigilance (planning of increased number of blood units, vascular catheter placement, and number of patients requiring transfusion) were significantly more frequent in group D. Postoperatively, all groups had similar incidences of hematoma or bruises, volumes of blood loss per drainage, reoperations to control hemorrhage, and mortality due to bleeding (n = 5).
From the Departments of Surgery, HBpital Tenon (SH), Paris; Hd pital Louis Mourier (CG, J-MH), Colombes; Centre Hospitalier Intercommunal (AF), Poissy; and HBpital Hen+Mondor (M-JB), Cr&eil, France. This work was conducted under the auspices of the French Associations for Surgical research, composed of: Association des Chirurgiens de I’Assistance Publique pour les Evaluations M6diiles (ACAPEM); Association de Recherche en Chirurgie (ARC); and Association de Recherche en Chirurgie de I’lle-de-France (ARCIF). The participants in this research study are listed in the appendix. Requests for reprints should be sent to Jean-Marie Hay, MD, Associa0on de Recherche en Chirurgie, 8 Avenue des Peupliers, 92272 Bois Colombes, France. Manuscript submitted February 25, 1994 and accepted in revised form July 5, 1994.
CONCLUSIONS: Our results suggest that PHST should not be performed routinely, but only in patients with abnormal clinical data. Such a policy necessitates a thorough history-including answers to a specific questionnaire like those used in prospective studies-and a rigorous, well-conducted physical examination. Am J Surg. 1995;170:19-23.
H
emostatic tests are usually performed routinely prior to elective or emergency surgery in order to detect patients likely to have abnormal bleeding. Standard screening for hemostatic abnormalities usually includes several of the following tests: prothrombin time (PT), activated partial thromboplastin time (AP’TT), platelet count (PC), and bleeding time (BT). As has also occurred in the cases of some other preoperative screening techniques,ld4 several reports have questioned the need for routine preoperative hemostatic screening tests (PHST).5-1’ Most of these studstudies ies were retrospective. 5,6,9-11Only four prospective have been published to date. 7-8~12~‘3 In the study by Eika et al,s PHST were evaluated without any relationship to clinical data. Blery et a1,7 and VelanovichlJ carried out studies on several preoperative tests including hemostasis. Only Blery et al and Rohrer et al” studied all four standard hemostatic tests. Some studies included bloodless minor surgery,7s9 while others considered only one specific type of surgery. 14,15In other reports, varying numbers of PHST were performed either selectively7J” or routinely.5,8,‘4 BT has been investigated in only four studies.6-8J2 The aim of our study was to determine prospectively the value of four PHST (PT, APTT, PC, and BT) in predicting the risk for bleeding, and to compare these results to data gathered by a specific questionnaire, history taking, and clinical examination.
PATIENTS AND METHODS Patients From October 1988 to December 1992,3,242 patients scheduled for general surgery procedures were included in a multicenter prospective study. They were 1,590 males and 1,652 females, with a mean age of 51 + 21 years (range 16 to 99). The 17 participating centers did not all begin accumulating patients at the same time. Once a center decided to enter the study, it was asked to enroll all (consecutive) patients. Every center enrolled at least 100 patients and ended the study before the end of December 1992. Patients were classified preoperatively according to the American Society of Anesthesiologists’ (ASA) grading.16 J
THE AMERICAN
JOURNAL
OF SURGERY@
VOLUME
170
JULY 1995
19
‘REOPERATIVE HEMOSTATIC SCREENING TESTS/HOURY ET AL TABLE I Classification of Patients According to Pathology, Operations, and ASA Grading
Pathology and digestive tract operations Colorectoanal Biliary tract Gastroduodenal Hepaticopancreatic Parietal and soft tissues Gynecological Urological Thoracic Vascular Endocrine Miscellaneous ASA grading I and II Ill and IV
Group A (n = 1,951)
Group B (n = 340)
Group C (n = 779)
Group D (n = 172)
50.0’ 20.0’ 14.7 5.9 1.3 24.1 8.1 2.2 1.3 0.9 2.0 11.4
57.8 36.5 13.4 5.8 2.5 21.9 6.9 3.1 1.6 1.3 2.8 4.7
53.5+ 22.0 19.6 7.2 4.a* 17.1 7.3 2.2 0.5 0.5 2.6 16.2
60.1 14.2 23.6 8.1 14.2 17.6 4.1 2.7 0.7 2.7 2.7 9.2
93.8 6.2
91.4 8.6
88.7* 11.3’
72.0 28.0
Data reported as percentage of patients. ‘P co.01. rP <0.05. $P
TABLE ii Preoperative Modifications of Guidellnes
Postponed operations Additional hemostatic tests ordered (total patients)+ Plasma coagulation factors Fibrinogen Fibrinogen split products Euglobulin clot lysis time Specific treatment necessary to correct hemostatic disorder (total patients)+ Platelets Fresh frozen plasma Preoperative vitamin Kl Factor IX complex
Group A (n = 1,951)
Group B (n = 340)
Group C (n = 779)
Group D (n = 172)
0’ 18’ 5’ 15+ 1s 0’ 0
4 12 9 8 4 4 1
3’ 10’ 3’ 8’ 1’ 0’ 3’
13 11 26 16 11 10 26
0 0 0 0
1 0 0 0
0 1. 2 0
6 6 11 1
‘P
Both elective (n = 2,498) and emergency (n = 744) general surgery procedures were included. Intra-abdominal surgery was performed in 1,840 patients, 1,098 via a median incision and 742 via a transverse or subcostal incision. The other patients underwent parietal, soft-tissue, gy necological, urological, thoracic, vascular, or endocrine surgery. Patients undergoing surgery with local anesthesia, operative or nonoperative orthopedic procedures, surgery for varicose veins of the lower limb were excluded, as were patients who received preoperative anticoagulant therapy. On the other hand, patients with low-dose heparin, according to Kakkar et a1,17 were included. 20
group C = patients
with 2 1
Methods All patients were asked to answer a questionnaire derived from several sources.1*18,*9The questions asked about easy or excessive bruising; bleeding more than 3 minutes after brushing of teeth; nosebleeds; prolonged bleeding after cuts; severe or prolonged menstrual periods; history of blood loss through the gastrointestinal or genitourinary tract; severe bleeding after dental extraction, surgical operation, or childbirth; and history of hemophilia or inherited family hemorrhagic disorder. Also elicited were any history of renal failure, liver disease, hypersplenism, hematological or collagen vascular disease, or any medication that might interfere with hemostasis. The physical examination paid special attention
THE AMERICAN JOURNAL OF SURGERY* VOLUME 170 JULY 1995
1 PREOPERATIVE
TABLE
HEMOSTATIC
ET AL
III lntraoperative
Modifications
of Anesthetic
and Surgical Vigilance Group B (n = 340)
Group C (n = 779)
Group D (n = 172)
0 0
1 1
0’ 0’
10 9
90 (4.6) 3 1-13
21 (6.2) 3 2-10
88 (11.3)’ 3 l-30
38 (22) 4 l-8
1 6
1 0
5’ 6’
18 5
Group A (n = 1,951) Anesthetic vigilance Patients for whom Patients for whom Blood transfusion Number of patients Median number of Range
increased blood units were planned increased vascular catheters were planned (%) units
Surgical vigilance Careful hemostasis Compressive bandage on wound ‘P 10.00 I. Group A = pattents with no c//nkxi cirxcal and no PHST abnormalities,
TABLE
SCREENING TESTSlHOURY
or PHST abnormalities; group B = patients with no clinica/ and t 1 PHST abnormal&es; group D = pat/en& with both clinical and PHST abnormabties
group C = patjents
with > 1
IV Postoperative
Total mortality (%) Related to hemorrhage Number of bruises Number of hematomas Median surface size (cm*) Range (cm*) Number of patients drained Median first day postoperative blood loss per drainage Range (mL) Reoperation to obtain control of hemorrhage
Mortality and Morbidity
(mL)
Group A (n = 1,951)
Group B (n = 340)
Group C (n = 779)
Group D (n = 172)
17 (0.87). 3 89 58 12 2-150 162 80’ 9-l ,300 9
7 (2.06)
17 (2.18)’ 0 65 27 24 4-200 101 75 5-l ,020 9
11 (6.39) 2 18 11 30 6-120 30 80 5-l ,100 2
0 20 10 35 24-90 40 35 5-500 2
‘P 1 PHST abnormalities; civxcal and no PHST abnormakties; group D = patients with both c/in/ca/ and PHST abnormalities.
to the presence of purpura, hematoma, jaundice, and signs of liver cirrhosis. Clinical data were considered abnormal when one or more answers on the questionnaire or in the history taking so indicated, or when one or more physical abnormality was detected. PHST were considered abnormal when one or more screening test was abnormal. The criteria for abnormality were: PT < 70% of control data; APTT more than 1.2 times the control APTT; PC <100,000/L; and BT greater than the upper limit of the normal range according to the testing procedure.z0 Based on the results of clinical data and PHST, patients were divided into 4 groups: group A ( n = 1,951) had no clinical or PHST abnormalities; group B (n = 340) had no clinical abnormalities and 21 PHST abnormalities; group C (n = 779) had 21 clinical and no PHST abnormalities; group D (n = 172) had both clinical and PHST abnormalities (Table I). The groups were compared with respect to the following criteria: (1) preoperative modifications of guidelines (Table II); (2) intraoperative modifications of anesthetic and surgical vigilance (Table III); (3) postoperative overall and hemorrhage-related mortality and morbidity (Table IV).
group C = patients
with t 1
Statistical comparisons were made with the chi,square, Fisher’s exact, and Mann-Whitney U-test(s), as appropriate. The level of significance was set at 0.05. RESULTS Comparability of Groups The groups obviously differed in pathology, ASA status, and type of operations (Table I). More patients in groups B and D than in groups A and C, respectively, had gastrointestinal disease (P ~0.01, P ~0.05). Colorectal and anal disease were more common in group B than group A (P
THE AMERICAN
JOURNAL
OF SURGERY”
VOLUME
170
JULY 1995
21
‘REOPERATIVE
HEMOSTATIC
SCREENING
TESTS/HOURV
ET AL
hemostatic disorder, compared to 26 of 172 (15%) with abnormal clinical data (P
COMMENTS Our results showed that routine PHST were not warranted in patients with normal clinical data who were undergoing general surgical procedures. Modifications of preoperative guidelines and intraoperative anesthetic and surgical vigilance were significantly higher in patients with abnormal clinical data. Ours is the largest prospective study carried out to date evaluating both morbidity and modification of anesthetic and surgical vigilance. The numbers of patients in previous prospective studies were 100,8 282,12 and 520.13 Blery et aL7 who studied 3,866 patients prospectively, tested the hy pothesis that PHST were not necessary, but did not compare the results of PHST to outcome. The large number of patients included in our study lends weight to our conclusions and makes comparisons relevant. Rohrer et alI2 investigated changes in management according to the results of PHST, but did not specify what the modifications were or whether they took place pre- or intraoperatively. Moreover, in their study, 3 of 21 abnormal PHST were subsequently proved to be false positives, and the other 18 were ignored by the surgeon without adverse consequences. In our study, overall postoperative mortality was higher in patients with abnormal PHST. In accordance with the results of Velanovich,i3 who found a correlation between the ASA score and postoperative complications, this difference was most likely related to preoperative ASA status and other risk factors linked to patient pathology rather than to the hemorrhagic risk in itself. On the other hand, in our study, there was no relationship between the results 22
THE AMERICAN
JOURNAL OF SURGERYa
VOLUME 170
of PHST and postoperative hemorrhage-related mortality (Table IV). Studies that examine only the relationship between hemostatic disorders and postoperative mortality are subject to criticism. The prevalence of bleeding disorders is low, and the proportion of them that are clinically occult is small. Hereditary coagulation disorders affect an estimated l/10,000 to l/40,000 individuals.3~21~22 At most, one third of cases are asymptomatic.10,21 Most acquired bleeding disorders are likely to be detected by clinical examination.23 As a result, Kaplan et a124 have estimated that approximately 20,000 patients would be necessary to detect a significant increase in perioperative mortality due to hemostatic disorders, and emphasized that “the consequences of missing abnormalities that might have been revealed by these tests represent potential patient harm which is judged at most to be 1 death per 100 years (per test)” at their institution. Mortality is an unreliable predictive factor not only because of the low prevalence of coagulation disorders, but also because of routine compensation of intra- and postoperative blood 10ss.~J~ For this reason, we evaluated the relationship between hemostatic disorders and postoperative morbidity as well as the modifications of preoperative guidelines and intraoperative anesthetic and surgical vigilance. Only objective criteria of morbidity as related to hemostatic disorders were evaluated. It is generally acknowledged that morbidity can be correctly assessed only in prospective studies. Two7J3 of the four prospective studies of PHST published to date7sJ2*13 specifically investigated morbidity, but neither stated the consequences in terms of modification of guidelines and vigilance according to PHST results. Like Velanovich,13 who studied PT, APTT, and PC, we found no significant difference in postoperative morbidity according to the results of PHST. Blery at al7 showed that in 3,866 patients undergoing 3,849 operations, there were no adverse effects related to the omission of PT, APTT, PC, or BT. The tests were not ordered in 76%, 75%, 92%, and 99% of patients, respectively. The only significant difference in postoperative morbidity noted in our study was that the median quantity of blood lost through drainage during the first day was paradoxically lowest in patients with normal clinical data and abnormal PHST. The more voluminous drainage did not, however, lead to any modifications, further transfusion, or repeat surgery. Blery et al7 found that perioperative blood loss was also sig nificantly higher in patients initially scheduled for “bloodless surgery.” All patients who bled in that study, however, had undergone plastic surgery of the head and neck, procedures that were not included in our study. The authors did not state whether the blood loss occurred intra- or postoperatively or whether it modified intra- or postoperative blood transfusion requirements. Moreover, they found no link between perioperative morbidity and the omission of PHST. The results of routine PHST are most likely related to the incidence of the abnormalities each test is designed to detect and its predictive value. Routine PT and APTT have been found to be poor predictors of intraoperative hemorrhage because most bleeding is related to surgical technique9*13225 and because, given the low incidence of clotting abnormalities in patients with normal clinical data, an abJULY 1995
1 PREOPERATIVE
HEMOSTATIC
SCREENING
TESTSRlOURY
ET AL
I
normal result is likely to be falsely positive.9,25,2h Indeed, Golub et a127found PT and APTT abnormalities in 2% and 6% of 325 patients with normal clinical data, and Eisenberg et al9 found that either PT or APTT was abnormal in 2.7% of 480 such patients. No patient in either of these studies had abnormal bleeding or required transfusion. Results of preoperative PC were abnormal in 6% patients studied by Goluh et al, but none had any bleeding complications or required any transfusion. The results of BT are not readily reproducible even with the Ivy technique.ZR In the study by Rohrer et al,” 4 of 107 patients had prolonged BT, of which 1 was found to be a false-positive result when the dosage was repeated. Neither this patient nor any of the 3 with truepositive results experienced any untoward consequences. Of 27 inirially increased BTs in the study by Barber et at6 12 were false and 15 true positives. As Velanovich also concluded,13 prolonged BT is of poor predictive value for surgical complication& and can be omitted. The results of our study further support the suggestion of Velanovich” and Blery at al’ that the preoperative detection of hemostatic disorders in patients without abnormal clinical data (7 I % in our population) is of no practical use. Abnormal PHST were found in only 18.1% of patients with abnormal clinical data, even with the use of a complete and rigorous physical examination and a specific questionnaire. Patients who are unreliable historians or too ill to give a correct history and those who suffer from liver disease” need laboratory testing as part of their preoperative evaluation. When PHST are ordered, BT can be left out. Selective use of PHST most likely requires a policy like that advocated by Fowkes and Roberts.29 The surgical and anesthesiology teams must be convinced of the necessity to limit the number of routine PHST. Then, they must establish an accurate prescription protocol and validate it in the short and medium terms. Finally, the protocol must be monitored to he sure that it is applied correctly. APPENDIX The following surgeons participated in this study: J.M. Hay, MD, Y. Flamanr, MD, G. Zeitoun, MD, (Colombes); A. Elhadad, MD, D. Brassier, MD, (Aulnay-sous-Bois); B. C&out, MD, E Hoche, MD, M. Veyrikres, MD, (Pontoise); E. Dewulf, MD, (Bruxelles); J. Francin, MD, (Dinan); M. Rodary, MD, (Orsay); H. Hennet, MD, (Romorantin); I? Laigneau, MD, S. Segrk, MD, (Meaux); M. Kalfon, MD, (Chatellerault); I? Le Picard, MD, (Charenton); R. Khayar, MD, Y. Soulier, MD, (Montmorency); JL. Sicard, MD, (Nice); PL. Fagniez, MD, N. Rotman, MD, J. Allouch, MD, (Cr&eil); CV. Devien, MD, JP. Pujol, MD, (Saint-Cloud); JI? Berhoux, MD, (Paris); E Lacaine, MD, S. Houry, Y. Katsenis, MD, M. Huguier, MD, (Paris); M. Benhamida, MD, (Sens); JM Avila, MD, (Montpellier); M Sage, MD, (Auxerre); F Yazdani, MD, (Annecy).
The Results and Discussion sections confirm the common knowledge that an excellent history and physical examination is the best preoperative test for all possible hematologic disorders associated with surgical procedures. Intru- or postoperative hemorrhage of this origin is certainly uncommon. Clinical suspicion is still the best discriminator. REFERENCES 1. Bowie EJ, Gwen CA Jr. The significance of abnormal hemostatic tests. Prog Hemost Thr&. 1980;5:179-209.
preoperative
2. National Srudy by Royal College of Radiologists. Preoperative chest radiology. Limcet. 1979;ii:83+36. 3. Robbins JA, Mushlin AI. Preoperative evaluation of the healthy patient. Med Cfn North Am. 1979;63:1145-1156. 4. Roizen ME Routine preoperative evaluation. In: Miller R, ed. Anaesthesia. 2nd ed. New York: Churchill Livingstone; 1986225-254. 5. Baranetsky NG, Weinstein I? Partial thromhoplastin time for screening. Ann Intern Med. 1979;91:498--499. 6. Barber A, Green D, Galluzzo T, Ts’ao CH. The bleeding time as a preoperative screening test. Am J Med. 1985;78:761-764. 7. Blety C, Charpak Y, Szatan M, er al. Evaluation a protocol for selective ordering of preoperative tests. Lancet. 1986;i:139-141. 8. Eika C, Having 0, Godal HC. The value of preoperative hemostatic screening. Scand .I &em&. 1978;21:349-354. 9. Eisenberg JM, Clarks JR, Sussman SA. Prothrombin and partial thromboplastin times as preoperative screening tests. Arch Surg. 1982; 117:48-51. 10. Kaplan EB, Sheiner LB, Boeckmann AJ, et al. The usefulness of preoperative laboratory screening. JAMA. 1985;253:35763581. 11. Robbins JA, Rose SO. Partial thromboplastin time as a screening test. Ann Intern Med. 1979;90:796-797. 12. Rohrer MJ, Michelotti MC, Nahrworld DL. A prospective evaluation of the efficacy of preoperative coagulation testing. Ann Surg. 1988;208:554-557. 13. Velanovich V. The value of routine preoperative laboratory testing in predictmg postoperative complication: a multivariate analysis. Surgery. 1991;109:236-243. 14. Marengo-Rowe AJ, Lambert CJ, Leveson JE, et al. The evaluation of hemorrhage in cardiac patients who have undergone extracorporeal circulation. Tmns&siun. 1979;19:426-433. 15. Rader ES Hematologic screening tests in patients with operative prostatic disease. Urology. 1978;11:243-246. 16. Owen WD, Felts A, Apitznagel EL Jr. ASA physical status classifications: a study of consistency of ranngs. Anesthesiology. 1978;49: 239-243. 17. Kakkar W, Field ES, Nicolaides AN, et al. Low doses of heparin in prevention of deep-vein thrombosis. Lancet. 1971;ii:669-671. 18. Marengo-Rowe A], Leveson JE. Evaluation of the bleeding patient. Postgmd Med. 1977;621:71-77. 19. Rappaport SI. Preoperative hemostatic evaluation: which tests, if any? Blood. 1983;61:229-231. 20. Duke W’X The relation of blood platelets to hemorrhagic disease. Description of a method for determining the bleeding time and coagulation time and report of three cases of hemorrhagic disease relieved by transfusion. JAMA. 1910;55:1185-1192. 21. Samama M. L’exploration simplifiee prC-opbratoire de l’hemostase. Anesth Ann Rban. 1975;32:569-73. 22. Tuddenham EC. Inherited bleeding disorders. Int Anesthzsiol C/in. 1985;23:61-72. 23. Silverman SG, Mueller PR, Pfister RC. Hemostatic evaluation before abdominal interventions. An overview and proposal. NR. Am J Rcentgenol. 1990;154:233-238. 24. Kaplan EB, Boekmann AJ, Roizen MF, et al. Elimination of unnecessary preoperative laboratory tests. Anesrhesio~gy.1982;57:A445. 25. Suchman AL, Mushlin AI. How well does the activated partial thromboplastin time predict postoperative hemorrhage. JAMA. 1986; 2531750-753. 26. Clarke JR, Eisenberg JM. A theoretical assessment of the value of the MT as a preoperative screening test in adults. Med De& Making. 1981;1:40-43. 27. Goiub R, Cantu R, Sorrento JJ, Stein HD. Efficacy of preadmission testing in ambulatory surgical patients. Am J Surg. 1992;163: 565-57 1. 28. Bums ER, Lawrence C. Bleeding time: a guide to its diagnostic and clinical utility. Arch P&oI Lab Med. 1989;113:1219-1224. 29. Fowkes FGR, Roberts CS. Introducing guidelines into clinical practice. Eflecrive Hrulthcare. 1984;1:313-324.
THE AMERICAN
of
JOURNAL OF SURGERYa
VOLUME 170
JULY 1995
23
BACKGROUND: Several retrospective and four prospective reports have questioned the need for routine preoperative hemostatic screening tests (PHST) in general surgery. PATIENTSAND METHODS: The results of four standard tests (prothrombin time, activated partial thromboplastin time, platelet count, and bleeding time) were prospectively compared with patient history and clinical data in a multicenter study of 3,242 consecutive patients. The patients were divided into four groups: group A (n = 1,951) had no clinical or PHST abnormalities; group B (n = 340) had no clinical and one or more PHST abnormalities; group C (n = 779) had one or more clinical and no PHST abnormalities; group D (n = 172) had both clinical and PHST abnormalities. RESULTS: Preoperative modifications of guidelines (postponed operations and ordering of additional hemostatic tests) were significantly more frequent in both groups of patients with PHST abnormalities (groups B and D), but specific treatment to correct hemostatic disorders was prescribed only when clinical abnormalities were also present (group D). Intraoperatively, modifications of anesthetic and surgical vigilance (planning of increased number of blood units, vascular catheter placement, and number of patients requiring transfusion) were significantly more frequent in group D. Postoperatively, all groups had similar incidences of hematoma or bruises, volumes of blood loss per drainage, reoperations to control hemorrhage, and mortality due to bleeding (n = 5).
From the Departments of Surgery, HBpital Tenon (SH), Paris; Hd pital Louis Mourier (CG, J-MH), Colombes; Centre Hospitalier Intercommunal (AF), Poissy; and HBpital Hen+Mondor (M-JB), Cr&eil, France. This work was conducted under the auspices of the French Associations for Surgical research, composed of: Association des Chirurgiens de I’Assistance Publique pour les Evaluations M6diiles (ACAPEM); Association de Recherche en Chirurgie (ARC); and Association de Recherche en Chirurgie de I’lle-de-France (ARCIF). The participants in this research study are listed in the appendix. Requests for reprints should be sent to Jean-Marie Hay, MD, Associa0on de Recherche en Chirurgie, 8 Avenue des Peupliers, 92272 Bois Colombes, France. Manuscript submitted February 25, 1994 and accepted in revised form July 5, 1994.
CONCLUSIONS: Our results suggest that PHST should not be performed routinely, but only in patients with abnormal clinical data. Such a policy necessitates a thorough history-including answers to a specific questionnaire like those used in prospective studies-and a rigorous, well-conducted physical examination. Am J Surg. 1995;170:19-23.
H
emostatic tests are usually performed routinely prior to elective or emergency surgery in order to detect patients likely to have abnormal bleeding. Standard screening for hemostatic abnormalities usually includes several of the following tests: prothrombin time (PT), activated partial thromboplastin time (AP’TT), platelet count (PC), and bleeding time (BT). As has also occurred in the cases of some other preoperative screening techniques,ld4 several reports have questioned the need for routine preoperative hemostatic screening tests (PHST).5-1’ Most of these studstudies ies were retrospective. 5,6,9-11Only four prospective have been published to date. 7-8~12~‘3 In the study by Eika et al,s PHST were evaluated without any relationship to clinical data. Blery et a1,7 and VelanovichlJ carried out studies on several preoperative tests including hemostasis. Only Blery et al and Rohrer et al” studied all four standard hemostatic tests. Some studies included bloodless minor surgery,7s9 while others considered only one specific type of surgery. 14,15In other reports, varying numbers of PHST were performed either selectively7J” or routinely.5,8,‘4 BT has been investigated in only four studies.6-8J2 The aim of our study was to determine prospectively the value of four PHST (PT, APTT, PC, and BT) in predicting the risk for bleeding, and to compare these results to data gathered by a specific questionnaire, history taking, and clinical examination.
PATIENTS AND METHODS Patients From October 1988 to December 1992,3,242 patients scheduled for general surgery procedures were included in a multicenter prospective study. They were 1,590 males and 1,652 females, with a mean age of 51 + 21 years (range 16 to 99). The 17 participating centers did not all begin accumulating patients at the same time. Once a center decided to enter the study, it was asked to enroll all (consecutive) patients. Every center enrolled at least 100 patients and ended the study before the end of December 1992. Patients were classified preoperatively according to the American Society of Anesthesiologists’ (ASA) grading.16 J
THE AMERICAN
JOURNAL
OF SURGERY@
VOLUME
170
JULY 1995
19
‘REOPERATIVE HEMOSTATIC SCREENING TESTS/HOURY ET AL TABLE I Classification of Patients According to Pathology, Operations, and ASA Grading
Pathology and digestive tract operations Colorectoanal Biliary tract Gastroduodenal Hepaticopancreatic Parietal and soft tissues Gynecological Urological Thoracic Vascular Endocrine Miscellaneous ASA grading I and II Ill and IV
Group A (n = 1,951)
Group B (n = 340)
Group C (n = 779)
Group D (n = 172)
50.0’ 20.0’ 14.7 5.9 1.3 24.1 8.1 2.2 1.3 0.9 2.0 11.4
57.8 36.5 13.4 5.8 2.5 21.9 6.9 3.1 1.6 1.3 2.8 4.7
53.5+ 22.0 19.6 7.2 4.a* 17.1 7.3 2.2 0.5 0.5 2.6 16.2
60.1 14.2 23.6 8.1 14.2 17.6 4.1 2.7 0.7 2.7 2.7 9.2
93.8 6.2
91.4 8.6
88.7* 11.3’
72.0 28.0
Data reported as percentage of patients. ‘P co.01. rP <0.05. $P
TABLE ii Preoperative Modifications of Guidellnes
Postponed operations Additional hemostatic tests ordered (total patients)+ Plasma coagulation factors Fibrinogen Fibrinogen split products Euglobulin clot lysis time Specific treatment necessary to correct hemostatic disorder (total patients)+ Platelets Fresh frozen plasma Preoperative vitamin Kl Factor IX complex
Group A (n = 1,951)
Group B (n = 340)
Group C (n = 779)
Group D (n = 172)
0’ 18’ 5’ 15+ 1s 0’ 0
4 12 9 8 4 4 1
3’ 10’ 3’ 8’ 1’ 0’ 3’
13 11 26 16 11 10 26
0 0 0 0
1 0 0 0
0 1. 2 0
6 6 11 1
‘P
Both elective (n = 2,498) and emergency (n = 744) general surgery procedures were included. Intra-abdominal surgery was performed in 1,840 patients, 1,098 via a median incision and 742 via a transverse or subcostal incision. The other patients underwent parietal, soft-tissue, gy necological, urological, thoracic, vascular, or endocrine surgery. Patients undergoing surgery with local anesthesia, operative or nonoperative orthopedic procedures, surgery for varicose veins of the lower limb were excluded, as were patients who received preoperative anticoagulant therapy. On the other hand, patients with low-dose heparin, according to Kakkar et a1,17 were included. 20
group C = patients
with 2 1
Methods All patients were asked to answer a questionnaire derived from several sources.1*18,*9The questions asked about easy or excessive bruising; bleeding more than 3 minutes after brushing of teeth; nosebleeds; prolonged bleeding after cuts; severe or prolonged menstrual periods; history of blood loss through the gastrointestinal or genitourinary tract; severe bleeding after dental extraction, surgical operation, or childbirth; and history of hemophilia or inherited family hemorrhagic disorder. Also elicited were any history of renal failure, liver disease, hypersplenism, hematological or collagen vascular disease, or any medication that might interfere with hemostasis. The physical examination paid special attention
THE AMERICAN JOURNAL OF SURGERY* VOLUME 170 JULY 1995
1 PREOPERATIVE
TABLE
HEMOSTATIC
ET AL
III lntraoperative
Modifications
of Anesthetic
and Surgical Vigilance Group B (n = 340)
Group C (n = 779)
Group D (n = 172)
0 0
1 1
0’ 0’
10 9
90 (4.6) 3 1-13
21 (6.2) 3 2-10
88 (11.3)’ 3 l-30
38 (22) 4 l-8
1 6
1 0
5’ 6’
18 5
Group A (n = 1,951) Anesthetic vigilance Patients for whom Patients for whom Blood transfusion Number of patients Median number of Range
increased blood units were planned increased vascular catheters were planned (%) units
Surgical vigilance Careful hemostasis Compressive bandage on wound ‘P 10.00 I. Group A = pattents with no c//nkxi cirxcal and no PHST abnormalities,
TABLE
SCREENING TESTSlHOURY
or PHST abnormalities; group B = patients with no clinica/ and t 1 PHST abnormal&es; group D = pat/en& with both clinical and PHST abnormabties
group C = patjents
with > 1
IV Postoperative
Total mortality (%) Related to hemorrhage Number of bruises Number of hematomas Median surface size (cm*) Range (cm*) Number of patients drained Median first day postoperative blood loss per drainage Range (mL) Reoperation to obtain control of hemorrhage
Mortality and Morbidity
(mL)
Group A (n = 1,951)
Group B (n = 340)
Group C (n = 779)
Group D (n = 172)
17 (0.87). 3 89 58 12 2-150 162 80’ 9-l ,300 9
7 (2.06)
17 (2.18)’ 0 65 27 24 4-200 101 75 5-l ,020 9
11 (6.39) 2 18 11 30 6-120 30 80 5-l ,100 2
0 20 10 35 24-90 40 35 5-500 2
‘P
to the presence of purpura, hematoma, jaundice, and signs of liver cirrhosis. Clinical data were considered abnormal when one or more answers on the questionnaire or in the history taking so indicated, or when one or more physical abnormality was detected. PHST were considered abnormal when one or more screening test was abnormal. The criteria for abnormality were: PT < 70% of control data; APTT more than 1.2 times the control APTT; PC <100,000/L; and BT greater than the upper limit of the normal range according to the testing procedure.z0 Based on the results of clinical data and PHST, patients were divided into 4 groups: group A ( n = 1,951) had no clinical or PHST abnormalities; group B (n = 340) had no clinical abnormalities and 21 PHST abnormalities; group C (n = 779) had 21 clinical and no PHST abnormalities; group D (n = 172) had both clinical and PHST abnormalities (Table I). The groups were compared with respect to the following criteria: (1) preoperative modifications of guidelines (Table II); (2) intraoperative modifications of anesthetic and surgical vigilance (Table III); (3) postoperative overall and hemorrhage-related mortality and morbidity (Table IV).
group C = patients
with t 1
Statistical comparisons were made with the chi,square, Fisher’s exact, and Mann-Whitney U-test(s), as appropriate. The level of significance was set at 0.05. RESULTS Comparability of Groups The groups obviously differed in pathology, ASA status, and type of operations (Table I). More patients in groups B and D than in groups A and C, respectively, had gastrointestinal disease (P ~0.01, P ~0.05). Colorectal and anal disease were more common in group B than group A (P
THE AMERICAN
JOURNAL
OF SURGERY”
VOLUME
170
JULY 1995
21
‘REOPERATIVE
HEMOSTATIC
SCREENING
TESTS/HOURV
ET AL
hemostatic disorder, compared to 26 of 172 (15%) with abnormal clinical data (P
COMMENTS Our results showed that routine PHST were not warranted in patients with normal clinical data who were undergoing general surgical procedures. Modifications of preoperative guidelines and intraoperative anesthetic and surgical vigilance were significantly higher in patients with abnormal clinical data. Ours is the largest prospective study carried out to date evaluating both morbidity and modification of anesthetic and surgical vigilance. The numbers of patients in previous prospective studies were 100,8 282,12 and 520.13 Blery et aL7 who studied 3,866 patients prospectively, tested the hy pothesis that PHST were not necessary, but did not compare the results of PHST to outcome. The large number of patients included in our study lends weight to our conclusions and makes comparisons relevant. Rohrer et alI2 investigated changes in management according to the results of PHST, but did not specify what the modifications were or whether they took place pre- or intraoperatively. Moreover, in their study, 3 of 21 abnormal PHST were subsequently proved to be false positives, and the other 18 were ignored by the surgeon without adverse consequences. In our study, overall postoperative mortality was higher in patients with abnormal PHST. In accordance with the results of Velanovich,i3 who found a correlation between the ASA score and postoperative complications, this difference was most likely related to preoperative ASA status and other risk factors linked to patient pathology rather than to the hemorrhagic risk in itself. On the other hand, in our study, there was no relationship between the results 22
THE AMERICAN
JOURNAL OF SURGERYa
VOLUME 170
of PHST and postoperative hemorrhage-related mortality (Table IV). Studies that examine only the relationship between hemostatic disorders and postoperative mortality are subject to criticism. The prevalence of bleeding disorders is low, and the proportion of them that are clinically occult is small. Hereditary coagulation disorders affect an estimated l/10,000 to l/40,000 individuals.3~21~22 At most, one third of cases are asymptomatic.10,21 Most acquired bleeding disorders are likely to be detected by clinical examination.23 As a result, Kaplan et a124 have estimated that approximately 20,000 patients would be necessary to detect a significant increase in perioperative mortality due to hemostatic disorders, and emphasized that “the consequences of missing abnormalities that might have been revealed by these tests represent potential patient harm which is judged at most to be 1 death per 100 years (per test)” at their institution. Mortality is an unreliable predictive factor not only because of the low prevalence of coagulation disorders, but also because of routine compensation of intra- and postoperative blood 10ss.~J~ For this reason, we evaluated the relationship between hemostatic disorders and postoperative morbidity as well as the modifications of preoperative guidelines and intraoperative anesthetic and surgical vigilance. Only objective criteria of morbidity as related to hemostatic disorders were evaluated. It is generally acknowledged that morbidity can be correctly assessed only in prospective studies. Two7J3 of the four prospective studies of PHST published to date7sJ2*13 specifically investigated morbidity, but neither stated the consequences in terms of modification of guidelines and vigilance according to PHST results. Like Velanovich,13 who studied PT, APTT, and PC, we found no significant difference in postoperative morbidity according to the results of PHST. Blery at al7 showed that in 3,866 patients undergoing 3,849 operations, there were no adverse effects related to the omission of PT, APTT, PC, or BT. The tests were not ordered in 76%, 75%, 92%, and 99% of patients, respectively. The only significant difference in postoperative morbidity noted in our study was that the median quantity of blood lost through drainage during the first day was paradoxically lowest in patients with normal clinical data and abnormal PHST. The more voluminous drainage did not, however, lead to any modifications, further transfusion, or repeat surgery. Blery et al7 found that perioperative blood loss was also sig nificantly higher in patients initially scheduled for “bloodless surgery.” All patients who bled in that study, however, had undergone plastic surgery of the head and neck, procedures that were not included in our study. The authors did not state whether the blood loss occurred intra- or postoperatively or whether it modified intra- or postoperative blood transfusion requirements. Moreover, they found no link between perioperative morbidity and the omission of PHST. The results of routine PHST are most likely related to the incidence of the abnormalities each test is designed to detect and its predictive value. Routine PT and APTT have been found to be poor predictors of intraoperative hemorrhage because most bleeding is related to surgical technique9*13225 and because, given the low incidence of clotting abnormalities in patients with normal clinical data, an abJULY 1995
1 PREOPERATIVE
HEMOSTATIC
SCREENING
TESTSRlOURY
ET AL
I
normal result is likely to be falsely positive.9,25,2h Indeed, Golub et a127found PT and APTT abnormalities in 2% and 6% of 325 patients with normal clinical data, and Eisenberg et al9 found that either PT or APTT was abnormal in 2.7% of 480 such patients. No patient in either of these studies had abnormal bleeding or required transfusion. Results of preoperative PC were abnormal in 6% patients studied by Goluh et al, but none had any bleeding complications or required any transfusion. The results of BT are not readily reproducible even with the Ivy technique.ZR In the study by Rohrer et al,” 4 of 107 patients had prolonged BT, of which 1 was found to be a false-positive result when the dosage was repeated. Neither this patient nor any of the 3 with truepositive results experienced any untoward consequences. Of 27 inirially increased BTs in the study by Barber et at6 12 were false and 15 true positives. As Velanovich also concluded,13 prolonged BT is of poor predictive value for surgical complication& and can be omitted. The results of our study further support the suggestion of Velanovich” and Blery at al’ that the preoperative detection of hemostatic disorders in patients without abnormal clinical data (7 I % in our population) is of no practical use. Abnormal PHST were found in only 18.1% of patients with abnormal clinical data, even with the use of a complete and rigorous physical examination and a specific questionnaire. Patients who are unreliable historians or too ill to give a correct history and those who suffer from liver disease” need laboratory testing as part of their preoperative evaluation. When PHST are ordered, BT can be left out. Selective use of PHST most likely requires a policy like that advocated by Fowkes and Roberts.29 The surgical and anesthesiology teams must be convinced of the necessity to limit the number of routine PHST. Then, they must establish an accurate prescription protocol and validate it in the short and medium terms. Finally, the protocol must be monitored to he sure that it is applied correctly. APPENDIX The following surgeons participated in this study: J.M. Hay, MD, Y. Flamanr, MD, G. Zeitoun, MD, (Colombes); A. Elhadad, MD, D. Brassier, MD, (Aulnay-sous-Bois); B. C&out, MD, E Hoche, MD, M. Veyrikres, MD, (Pontoise); E. Dewulf, MD, (Bruxelles); J. Francin, MD, (Dinan); M. Rodary, MD, (Orsay); H. Hennet, MD, (Romorantin); I? Laigneau, MD, S. Segrk, MD, (Meaux); M. Kalfon, MD, (Chatellerault); I? Le Picard, MD, (Charenton); R. Khayar, MD, Y. Soulier, MD, (Montmorency); JL. Sicard, MD, (Nice); PL. Fagniez, MD, N. Rotman, MD, J. Allouch, MD, (Cr&eil); CV. Devien, MD, JP. Pujol, MD, (Saint-Cloud); JI? Berhoux, MD, (Paris); E Lacaine, MD, S. Houry, Y. Katsenis, MD, M. Huguier, MD, (Paris); M. Benhamida, MD, (Sens); JM Avila, MD, (Montpellier); M Sage, MD, (Auxerre); F Yazdani, MD, (Annecy).
The Results and Discussion sections confirm the common knowledge that an excellent history and physical examination is the best preoperative test for all possible hematologic disorders associated with surgical procedures. Intru- or postoperative hemorrhage of this origin is certainly uncommon. Clinical suspicion is still the best discriminator. REFERENCES 1. Bowie EJ, Gwen CA Jr. The significance of abnormal hemostatic tests. Prog Hemost Thr&. 1980;5:179-209.
preoperative
2. National Srudy by Royal College of Radiologists. Preoperative chest radiology. Limcet. 1979;ii:83+36. 3. Robbins JA, Mushlin AI. Preoperative evaluation of the healthy patient. Med Cfn North Am. 1979;63:1145-1156. 4. Roizen ME Routine preoperative evaluation. In: Miller R, ed. Anaesthesia. 2nd ed. New York: Churchill Livingstone; 1986225-254. 5. Baranetsky NG, Weinstein I? Partial thromhoplastin time for screening. Ann Intern Med. 1979;91:498--499. 6. Barber A, Green D, Galluzzo T, Ts’ao CH. The bleeding time as a preoperative screening test. Am J Med. 1985;78:761-764. 7. Blety C, Charpak Y, Szatan M, er al. Evaluation a protocol for selective ordering of preoperative tests. Lancet. 1986;i:139-141. 8. Eika C, Having 0, Godal HC. The value of preoperative hemostatic screening. Scand .I &em&. 1978;21:349-354. 9. Eisenberg JM, Clarks JR, Sussman SA. Prothrombin and partial thromboplastin times as preoperative screening tests. Arch Surg. 1982; 117:48-51. 10. Kaplan EB, Sheiner LB, Boeckmann AJ, et al. The usefulness of preoperative laboratory screening. JAMA. 1985;253:35763581. 11. Robbins JA, Rose SO. Partial thromboplastin time as a screening test. Ann Intern Med. 1979;90:796-797. 12. Rohrer MJ, Michelotti MC, Nahrworld DL. A prospective evaluation of the efficacy of preoperative coagulation testing. Ann Surg. 1988;208:554-557. 13. Velanovich V. The value of routine preoperative laboratory testing in predictmg postoperative complication: a multivariate analysis. Surgery. 1991;109:236-243. 14. Marengo-Rowe AJ, Lambert CJ, Leveson JE, et al. The evaluation of hemorrhage in cardiac patients who have undergone extracorporeal circulation. Tmns&siun. 1979;19:426-433. 15. Rader ES Hematologic screening tests in patients with operative prostatic disease. Urology. 1978;11:243-246. 16. Owen WD, Felts A, Apitznagel EL Jr. ASA physical status classifications: a study of consistency of ranngs. Anesthesiology. 1978;49: 239-243. 17. Kakkar W, Field ES, Nicolaides AN, et al. Low doses of heparin in prevention of deep-vein thrombosis. Lancet. 1971;ii:669-671. 18. Marengo-Rowe A], Leveson JE. Evaluation of the bleeding patient. Postgmd Med. 1977;621:71-77. 19. Rappaport SI. Preoperative hemostatic evaluation: which tests, if any? Blood. 1983;61:229-231. 20. Duke W’X The relation of blood platelets to hemorrhagic disease. Description of a method for determining the bleeding time and coagulation time and report of three cases of hemorrhagic disease relieved by transfusion. JAMA. 1910;55:1185-1192. 21. Samama M. L’exploration simplifiee prC-opbratoire de l’hemostase. Anesth Ann Rban. 1975;32:569-73. 22. Tuddenham EC. Inherited bleeding disorders. Int Anesthzsiol C/in. 1985;23:61-72. 23. Silverman SG, Mueller PR, Pfister RC. Hemostatic evaluation before abdominal interventions. An overview and proposal. NR. Am J Rcentgenol. 1990;154:233-238. 24. Kaplan EB, Boekmann AJ, Roizen MF, et al. Elimination of unnecessary preoperative laboratory tests. Anesrhesio~gy.1982;57:A445. 25. Suchman AL, Mushlin AI. How well does the activated partial thromboplastin time predict postoperative hemorrhage. JAMA. 1986; 2531750-753. 26. Clarke JR, Eisenberg JM. A theoretical assessment of the value of the MT as a preoperative screening test in adults. Med De& Making. 1981;1:40-43. 27. Goiub R, Cantu R, Sorrento JJ, Stein HD. Efficacy of preadmission testing in ambulatory surgical patients. Am J Surg. 1992;163: 565-57 1. 28. Bums ER, Lawrence C. Bleeding time: a guide to its diagnostic and clinical utility. Arch P&oI Lab Med. 1989;113:1219-1224. 29. Fowkes FGR, Roberts CS. Introducing guidelines into clinical practice. Eflecrive Hrulthcare. 1984;1:313-324.
THE AMERICAN
of
JOURNAL OF SURGERYa
VOLUME 170
JULY 1995
23