| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
,
* From the Department of Obstetrics and Gynecology, St. Joseph's Hospital and McMaster University; the
Department of Clinical Epidemiology and Biostatistics, McMaster University; the
Division of Critical Care, McMaster University, Hamilton, Ontario, Canada; the
Department of Obstetrics and Gynecology, Monash University, Clayton, Victoria, Australia; and the
¶ Prehospital Care Program, Sunnybrook and Women's College Health Sciences Center, Toronto, Ontario, Canada.
Address correspondence to: Dr. R.J. McDonagh, Department of Obstetrics and Gynecology, St. James's Hospital, 301 James Street South, 2nd Floor-Fontbonne Building, Hamilton, Ontario L8N 4A6, Canada. Phone: 905-522-4941, ext. 4258; Fax: 905-521-6089; E-mail: rorymcdonagh{at}hotmail.com
 |
Abstract |
|---|
 TOP Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Methods: This retrospective subgroup analysis, taken from a cohort of 2051 hypertensive pregnant women, comprises 87 individuals who underwent both a BT and platelet count prior to delivery. We calculated the correlation between the platelet count and BT at three platelet cut-off points with respect to prolonged BT of eight minutes or more.
Results: There was a significant negative correlation between platelet count at delivery and BT [r= -0.45, 95% confidence interval (CI) -0.26 to -0.60; P <0.0001]. All three platelet cut-off points had a sensitivity of less than 66% with negative predictive values below 75% for an abnormal BT. A platelet count 
75 x 109/L was specific for the presence of an abnormal BT (specificity 97.8%, 95% CI 91.7–100.0), with a positive predictive value of 95.5% (95% CI 83.1–100.0) and a positive likelihood ratio of 24 (95% CI 3.3–168).
Conclusions: In a group of hypertensive parturients, the platelet count appears to be very specific for predicting a prolonged BT. The platelet count may aid the anesthesiologist in determining the risk of bleeding from regional anesthesia. Given the study's potential for bias future research is needed to validate these findings.
|
Introduction |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Since its introduction nearly a century ago arguments have been made for15,16and against17–19 the use of the BT to predict hemorrhagic risk, including that associated with regional anesthesia.18 The association between the platelet count and BT has also been debated. For example, a hallmark study by Harker and Slichter found a strong correlation between platelet count and BT,20 a finding corroborated by a study of women with preeclampsia.21 However, other reports contradict these findings.3,17 A recent statement by the College of American Pathologists and the American Society of Clinical Pathologists concluded that the BT cannot accurately predict the risk of surgical hemorrhage, and that a normal BT does not exclude the possibility of excessive hemorrhage during an invasive procedure.11 This statement also addressed the use of the BT in situations where bleeding risk cannot be evaluated by history alone.11 Thus, there remains ongoing controversy about the best way to determine bleeding tendency among individuals at higher than average risk, such as a pregnant woman with preeclampsia and non-protenuric gestational hypertension.
Recent evidence suggests that the platelet count is a practical and cost-effective means for excluding the presence of a coagulopathy in women with toxemia.22 In addition, previous survey data have identified a heavy reliance on the platelet count in the evaluation of bleeding risk in the preeclamptic population.23 Accordingly, we examined the association between BT and platelet count in a cohort of pregnant women with hypertension during pregnancy. Our primary goal was to determine the test properties of the platelet count in predicting an abnormal BT among parturients with both proteinuric and non-proteinuric forms of hypertension during pregnancy. We further discuss the practical limitations of both the BT and platelet count in assessing bleeding risk in the particular context of neuraxial anesthesia.
|
Methods |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Participants were consecutively recruited into MOS HIP at the time of delivery and accounted for 98% of all women with hypertension who delivered at our hospital during this period. Data were collected from the mother's hospital chart by two authors (R.B. and E.B.) within three months after delivery, abstracted to a standard data collection form and subsequently entered into a database (SPSS, SPSS Inc, Chicago, IL). Prior to delivery, a complete blood count, including automated platelet count, was conducted in all women. If the platelet count was below normal (i.e., 150 x 109/L), a manual count was performed. In addition, prothrombin time (PT), aPTT, and fibrinogen level were assessed.
The subgroup of women in MOS HIP who underwent BT evaluation was included in the current analysis. BTs were completed by a hematology technician using a surgicutt BT device according to the method of Mielke.25,26 The timing and indications for the BT were not typically documented in the patient's chart and the test was done at the discretion of the attending physician. We did not collect data on the type of anesthesia or analgesia provided at delivery, or complications related to neuraxial anesthesia or maternal hemorrhage.
Baseline maternal and neonatal characteristics, among those who had a BT done, were compared with those who did not; similar comparisons were made among those whose BT was normal (i.e., less than eight minutes) vs abnormal (i.e., greater than or equal to eight minutes); (Table I
). Baseline characteristics were compared using one-way ANOVA for continuous variables and a chi-square test for categorical data. All P values were two-sided, and a significance level of 0.05 was chosen. Statistical analyses were performed using SAS Version 6.12 (SAS Institute Inc., Cary, North Carolina, USA).
|
We also evaluated BT (i.e., below vs above eight minutes) with respect to three different platelet count cut-off points (i.e., 50 x 109/L, 75 x 109/L, and 100 x 109/L). The use of the cut-off point of eight minutes to define an abnormal BT was based on data from other studies,31 textbooks,32 and the upper limit of normal of our own centre. Data were entered into a 2 x 3 table, and the sensitivity and specificity, positive (PPV) and negative predictive (NPV) values, as well as positive and negative likelihood ratios (LR) were calculated, together with their respective 95% CI. These test properties were then re-calculated after excluding women who had received aspirin within one week of delivery.
The McMaster University Medical Ethics Committee was consulted at the commencement of the study in 1986. They decided that formal consent was not required because all maternal and newborn personal identifiers were omitted upon database entry, which preserved patient anonymity and prevented future patient contact.
|
Results |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
). The rate of eclampsia varied from approximately 0.1% in the gestational hypertension group to 3.9% among those with preeclampsia. None received anticoagulants prior to delivery and less than 5% of all participants received magnesium sulphate prophylaxis. The characteristics of all MOS HIP participants are outlined in Table I. Of the 87 women studied, 41 (47.1%) had a prolonged BT. There were several significant differences between groups, including the observation that women with a prolonged BT were more likely to have a diagnosis of preeclampsia or HELLP syndrome (92.7%) compared to those whose BT was normal (71.7%) or who did not undergo a BT (32.1%).
Hematological profile
The hematological profile of study participants is shown in Table I. The mean platelet count at delivery was significantly lower among those with a prolonged BT (88 x 109/L) than those with a normal BT (200 x 109/L). A platelet count 75 x 109/L at delivery was observed among 76 of 2051 study participants (3.7%), and was more prevalent among women with an abnormal BT (51.2%) than those with a normal BT (2.2%) or those in whom a BT was not done (2.7%). The mean PT and aPTT were slightly higher in the group with a prolonged BT, but all mean values were within normal limits for our hospital laboratory.
Correlational analyses
A significant negative correlation between platelet count at delivery and BT (r= -0.45, 95% CI -0.26 to -0.60; P <0.0001) was observed (Figure). The magnitude of the relationship was greater after excluding women receiving aspirin (r= -0.58, 95% CI -0.41 to -0.71; P <0.0001). The partial correlation coefficients, controlling for PT and aPTT, were similar: = -0.42 (P <0.0001) among all 87 women and r= -0.57 (P <0.0001) among those not receiving aspirin.
|
presents the test properties of the different platelet count cut-off points in relation to the BT. Among women with a platelet count below 50 x 109/L, none had a normal BT; accordingly, a value of "1" was assigned to this group in order to estimate the test properties at that platelet count cut-off point. All three platelet cut-off point values had a low sensitivity for the detection of an abnormal BT, varying from as low as 32% for a platelet count 50 x 109/L, to nearly 66% for a count 100 x 109/L. The corresponding NPVs were all below 75% (Table II).
|
). The test properties for a platelet count 50 x 109/L were largely the same as those for a count 75 x 109/L. A platelet count cut-off point of 100 x 109/L had a somewhat lower specificity (80.4%, 95% CI 67.9–90.5), PPV (75.0%, 95 % CI 61.0–89.0) and positive LR (3.4, 95% CI 1.8–6.3). After excluding seven women who were taking aspirin, the test properties of the various platelet cut-off points remained essentially unchanged (data available upon request). |
Discussion |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The decision to also include pregnant women with non-proteinuric forms of hypertension (i.e., chronic and gestational hypertension)24 was made because we did not know why these women had a BT ordered, and thus, could not be certain that they were no more hypertensive or ill than women with overt proteinuria (i.e., preeclampsia). Similarly, the finding of non-proteinuric hypertension and thrombocytopenia in a parturient may have represented a variant of preeclampsia in advance of the manifestation of overt proteinuria, which often develops only at a later stage.24
There are several potential sources of bias within our study that require comment.
First, this was a post hoc subgroup analysis from a larger cohort study and, accordingly, we did not collect data on or control for certain relevant factors, such as the indication for the BT. Similarly, those who ordered a BT were not masked to the patient's platelet count or severity of toxemia; the higher rate of thrombocytopenia and preeclampsia among the women who underwent a BT supports this point. Finally, we did not assess whether the BT was measured prior to or during the first stage of labour; a condition of heightened anxiety which may slightly augment the BT.17 Accordingly, we must interpret our findings with caution, particularly the apparently high specificity and PPV of the platelet count for a prolonged BT.
The absence of a normotensive pregnant control group presents another limitation to our study, especially given the paucity of literature surrounding BT variations during uncomplicated pregnancy.2,17 Thus, our definition of an abnormal BT (i.e., greater or equal to eight minutes) may have been inappropriately applied, leading to another source of bias in our estimation of the platelet count-BT test properties. Finally, in the absence of data on hemorrhagic events in this population, we cannot draw inferences about the predictive value of the platelet count for this outcome.
Notwithstanding the above limitations, these findings do add to a scarce body of literature on the utility of the platelet count in estimating, but not defining, the risk for hemorrhage among women with hypertension in pregnancy before delivery. The CIs around the estimates were quite narrow, and, assuming that the high specificity and PPV test properties of a low platelet count are only somewhat biased, it is likely that these data can be of some practical use to clinicians.21
It has been demonstrated that a prolonged BT predicts a decline in the hematocrit value, and could be indirect evidence of excess hemorrhage .33,34 However, this concept has been countered by the observation that anemia itself, in the face of a normal platelet count, can prolong the BT.17 Furthermore, it is known that the BT may be prolonged due to a variety of conditions that are independent of platelet quantity or function.17
We have observed that a platelet count of 75 x 109/L or less may be quite specific for the presence of an abnormal BT. When a woman's platelet count is below this value, it is unlikely that a BT measurement can add much to the estimation of bleeding risk, since our data suggest that the BT will likely be prolonged. However, a platelet count of 100 x 109/L or greater is less specific (80%), while remaining insensitive (66%), for a prolonged BT. At this point, the clinician may require another method to estimate the probability that a bleeding diathesis is present, as well as to evaluate the risk for a major hemorrhagic event.
Although the risk of traumatic neuraxial hemorrhage remains a major issue for the obstetrical anesthesiologist,8,9 there are few established risk factors for this serious but rare event.8,9,35 The development of a spinal hematoma is not guaranteed in the face of a prolonged BT and would likely remain a rare event even in the presence of a bleeding diathesis. Increasing experience and the use of more modern techniques for neuraxial blockade, including the selective use of a single spinal block instead of insertion of an epidural catheter, could translate into a lower baseline risk for traumatic hemorrhage. Although our data may aid in clinical decision-making, we encourage further discussion and the development of more precise estimates of bleeding risk in this population.
A large prospective cohort study might optimally determine which risk factors and hematological markers are important predictors of bleeding risk among both normotensive and hypertensive pregnant women. By comparing these women to those with gestational thrombocytopenia,36 one may shed further light on bleeding abnormalities peculiar to women with preeclampsia. Finally, although spinal hematomas are extremely rare, neuraxial anesthesia is used frequently enough that a large international registry could provide a valuable mechanism to identify risk factors for spinal hematoma.
|
Footnotes |
|---|
Revision received February 5, 2001. Accepted for publication December 4, 2000.
|
References |
|---|
|
TOPAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
2 Ivankovic M, Pereira J, Germain A, Opazo A, Mezzano D. Pregnancy and the bleeding time (Letter). Thromb Haemost 1992; 68: 375.[Medline]
3
Kelton JG, Hunter DJS, Neame PB. A platelet function defect in preeclampsia. Obstet Gynecol 1985; 65: 107–9.
4 Giles C, Inglis TCM. Thrombocytopenia and macrothrombocytosis in gestational hypertension. J Obstet Gynecol 1981; 88: 1115–9.
5 Redman CWG, Bonnar J, Beilin L. Early platelet consumption in pre-eclampsia. Br Med J 1978; 1: 467–9.
6 Sibai BM, Caritis SN, Thom E, Shaw K, McNellis D, and the National Institute of Child Health and Human Development Maternal-Fetal Medicine Network. Low-dose aspirin in nulliparous women: safety of continuous epidural block and correlation between bleeding time and maternal-neonatal bleeding complications. Am J Obstet Gynecol 1995; 172: 1553–7.[Medline]
7
Wildsmith JAW, McClure JH. Aspirin, bleeding time and central neural block (Letter). Brit J Anaesth 1993; 70: 112.
8 Horlocker TT. Complications of spinal and epidural anesthesia. Anesthesiol Clin North America 2000; 18: 461–85.[Medline]
9
Wulf H. Epidural anesthesia and spinal haematoma. Can J Anaesth 1996; 43: 1260–71.
10
Cohen MM, Wade J, Woodward C. Medical-legal concerns among Canadian anaesthetists. Can J Anaesth 1990; 37: 102–11.
11
Peterson P, Hayes TE, Arkin CF, et al. The preoperative bleeding time test lacks clinical benefit: College of American Pathologists' and American Society of Clinical Pathologists' position article. Arch Surg 1998; 133: 134–9.
12
Simon L, Santi TM, Sacquin P, Hamza J. Pre-anaesthetic assessment of coagulation abnormalities in obstetric patients: usefulness, timing and clinical implications. Br J Anaesth 1997; 78: 678–83.
13 Kopacz DJ, Gibbs CP. Training anesthesia personnel to perform bleeding times. Reg Anesth 1991; 16: 161–3.[Medline]
14 Rodgers A, Sage D, Futter M, MacMahon S. Attitudes and practices of New Zealand anaesthetists with regard to epidural and subarachnoid anaesthesia. Anaesth Intensive Care 1996; 24: 79–86.[Medline]
15 Mattix H, Singh AK. Is the bleeding time predictive of bleeding prior to a percutaneous renal biopsy? Curr Opin Nephrol Hypertens 1999; 8: 715–8.[Medline]
16 Davis CL, Chandler WL. Thromboelastography for the prediction of bleeding after transplant renal biopsy. J Am Soc Nephrol 1995; 6: 1250–5.[Abstract]
17 Rogers RPC, Levin J. A critical reappraisal of the bleeding time. Semin Thromb Hemost 1990; 16: 1–20.[Medline]
18
O'Kelly SW, Lawes EG, Luntley JB. Bleeding time: is it a useful clinical tool? Br J Anaesth 1992; 68: 313–5.
19 O'Brien JR. The bleeding time in normal and abnormal subjects. J Clin Pathol 1951; 4: 272–85.
20 Harker LA, Slichter SJ. The bleeding time as a screening test for evaluation of platelet function. N Eng J Med 1972; 287: 155–9.
21 Ramanathan J, Sibai BM, Vu T, Chauhan D. Correlation between bleeding times and platelet counts in women with preeclampsia undergoing cesarean section. Anesthesiology 1989; 71: 188–91.[Medline]
22
Barron WM, Heckerling P, Hibbard JU, Fisher S. Reducing unnecessary coagulation testing in hypertensive disorders of pregnancy. Obstet Gynecol 1999; 94: 364–70.
23 Voulgaropoulos DS, Palmer CM. Coagulation studies in the preeclamptic parturient: a survey. J Clin Anesth 1993; 5: 99–104 .[Medline]
24 Anonymous. National high blood pressure education program working group report on high blood pressure in pregnancy. Am J Obstet Gynecol 1990; 163: 1689–1712.
25 Sutor AH, Bowie EJW, Thompson JH, Didisheim P, Mertens BF, Owen CA Jr. Bleeding from standardized skin punctures: automated technic for recording time, intensity, and pattern of bleeding. Am J Clin Path 1971; 55: 541–50.[Medline]
26 Mielke CH. Measurement of the bleeding time. Thromb Haemostas 1984; 52: 210–1.[Medline]
27 Jorgensen KA, Dyerberg J, Olesen AS, Stoffersen E. Acetylsalicylic acid, bleeding time and age. Thromb Res 1980; 19: 799–805.[Medline]
28 Marshall PW, Williams AJ, Dixon RM, et al. A comparison of the effects of aspirin on bleeding time measured using the SimplateTM method and closure time measured using the PFA-100TM, in healthy volunteers. Br J Clin Pharmacol 1997; 44: 151–5.[Medline]
29 Falck G, Lundgaard H, Jareld T, et al. Effect of magnesium infusion on bleeding time in healthy male volunteers. Scand J Clin Lab Invest 1999; 59: 425–30.[Medline]
30 Ames WA, McDonnell N, Potter D. The effect of ionised magnesium on coagulation using thromboelastography. Anaesthesia 1999; 54: 999–1006.[Medline]
31 Burnes ER, Lawrence C. Bleeding time. A guide to its diagnostic and clinical utility. Arch Pathol Lab Med 1989; 113: 1219–24.[Medline]
32 Wallach J. Interpretation of Diagnostic Tests, 5th ed. Boston: Little Brown and Co., 1992.
33 Hellem AJ, Borchgrevink CF, Ames SB. The role of red cells in hemostasis: the relation between haematocrit, bleeding time and platelet adhesiveness. Br J Haematol 1961; 7: 42–50.[Medline]
34 Fernandez F, Goudable C, Sie P, et al. Low haematocrit and prolonged bleeding time in uraemic patients: effect of red cell transfusions. Br J Haematol 1985; 59: 139–48.[Medline]
35 Armstrong RF, Addy V, Breivik H. Epidural and spinal anaesthesia and the use of anticoagulants. Hosp Med 1999; 60: 491–6.[Medline]
36 Burrows RF, Kelton JG. Incidentally detected thrombocytopenia in healthy mothers and their infants. N Engl J Med 1988; 319: 142–5.[Abstract]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
