We investigated whether the inclusion of the PFA-100 in the preoperative screening of neurosurgical patients might reduce perioperative bleeding complications. Patients with intracranial space-occupying lesions who were scheduled for neurosurgery underwent routine preoperative PFA-100 testing. In case of an abnormal PFA test, patients received prophylactic treatment with desmopressin. 93 consecutive patients were compared to 102 consecutive patients with comparable characteristics operated before introduction of the PFA-100 testing. 2 patients (2.2%) in the PFA group and 2 patients (2.0%) in the non-PFA group experienced clinically relevant intracranial bleeding confirmed by computed tomography (OR 1.05, 95% CI 0.39–2.82;
Hemorrhage is one of the most threatening and feared complications of surgery. In neurosurgery, in particular, hemorrhage can lead to devastating consequences for the patient [
In addition, perioperative hemorrhage often requires the administration of blood products in order to replenish a deficit of red cells and/or to correct resulting or underlying coagulopathy. The risks of blood transfusion still cause significant morbidity and mortality [
In recent years it has become apparent that a significant number of patients scheduled for surgery exhibit disorders of hemostasis, which might increase the risk of perioperative bleeding. It has been shown that disorders of primary hemostasis are much more prevalent in these patients than disorders of secondary hemostasis [
For decades, the bleeding time used to be the only diagnostic tool to evaluate primary hemostasis. However, this test is cumbersome and difficult to standardize and, consequently, has shown only very limited value in correctly predicting clinically significant disorders of primary hemostasis [
A large body of evidence has accumulated that disorders of primary hemostasis can often successfully be treated with desmopressin (1-deamino-8-D-arginine-vasopressin (DDAVP)). Desmopressin is a selective vasopressin receptor type 2 agonist and leads to the release of von-Willebrand factor from the endothelium into the circulation. Apart from its role in von-Willebrand disease [
We assumed that, by the introduction of the PFA-100 in the routine preoperative evaluation of elective patients with intracranial space-occupying lesions who were scheduled for complex neurosurgical interventions, we could prevent bleeding complications and reduce blood product consumption by administering desmopressin to patients where an abnormal preoperative PFA-100 result was found. It was the aim of this study to analyze whether this approach indeed produced the presumed beneficial effects.
The PFA-100 was introduced in April 2005 into the routine preoperative screening of neurosurgery patients after a decision of a clinical board consisting of the heads and deputy heads of the Department of Transfusion Medicine and Hemostaseology and the Department of Neurosurgery. It was concluded that elective patients with intracranial space-occupying lesions who were scheduled for complex neurosurgical interventions were most likely to benefit from this approach, and thus PFA-100 testing was restricted to those patients. In addition, patients suffering from vascular malformations or intracranial aneurysms were excluded. This decision was based on the notion that in these patients bleeding is primarily a result of the intracranial pathology and the concern that certain side effects of desmopressin, that is, a possible rise in blood pressure or vasoconstriction in some cases, might be a risk for these patients that should not be taken.
Patients were consecutively recruited for analysis from a fixed time period of nine months after the introduction of the PFA-100 (PFA-group). They were compared to all patients meeting the clinical criteria described above that had been operated by the same team of surgeons and the same surgical techniques in the nine months before the introduction of the PFA-100 (non-PFA group). Patients operated on in April 2005 were excluded because we assumed the new regime needed some time to be adequately established in the participating departments. We preferred fixed time frames rather than identical patient numbers for the composition of both groups because we considered this approach less prone to a selection bias.
All patients were asked for previous bleeding problems upon the routine preoperative workup. Patients with an abnormal result in one or both of the commercially available PFA test modifications (see Section
The approval of the institutional review board was not assumed necessary for this retrospective study because the intervention analyzed involved a test system that had been approved officially for clinical use and was introduced as a routine practice based on current scientific evidence. The study was only performed thereafter relying on routinely collected patient data.
The principle of the PFA-100 has been described previously [
Patient baseline data were retrieved from the routine hospital patient records. All patients that had undergone neurosurgery during the study period were initially evaluated. Patient selection was performed by two authors (K. Reuter and J. Rohlfs) by mutual agreement. Patient selection was based on initial diagnosis and the
The following outcome data were retrieved from the patient records: Glasgow Coma Scale (GCS) [ cerebrospinal fluid block, midline deviation, intraventricular bleeding, considerable deterioration of consciousness together with CT-documented bleeding but none of the above.
Transfusion requirements (including plasma derivatives) had been centrally documented in an electronic data processing system.
The statistical analysis was performed with the statistic analysis software program SAS version 9.1 (SAS Institute, Cary, NC). Parametric variables were compared using the
Finally, 102 and 93 patients were included in the non-PFA group and the PFA group, respectively. Baseline characteristics are listed in Table
Patients clinical baseline characteristics.
Parameter | Non-PFA group | PFA group |
|
---|---|---|---|
Age [Years] | 50 (16) | 50 (18) | 0.91 |
Gender: male/female [ |
54/48 | 45/48 | 0.57 |
Height [cm] | 172 (10) | 172 (10) | 0.74 |
Weight [kg] | 77.7 (16.5) | 79.2 (16.9) | 0.55 |
BMI | 26.3 (4.9) | 26.9 (5.0) | 0.43 |
Preoperative GCS | 15 (13–15) | 15 (5–15) | 0.27 |
ASA status | 0.39 | ||
1 | 13 | 9 | |
2 | 48 | 40 | |
3 | 33 | 39 | |
4 | 0 | 1 | |
n. d.* | 8 | 4 |
*n. d.: not determined (insufficient patient data).
Patients laboratory baseline characteristics.
Parameter | Non-PFA group | PFA group |
|
---|---|---|---|
Platelets [/nl] | 257 (64) | 254 (80) | 0.72 |
Preoperative Hb [g/L] | 144 (13.7) | 143 (16.5) | 0.72 |
Male | 149 (12.2) | 152 (13.8) | 0.19 |
Female | 139 (13.4) | 136 (14.7) | 0.28 |
Prothrombin time [INR] | 0.99 (0.88–1.30) | 0.98 (0.81–1.30) | 0.19 |
aPTT [s] | 28 (22–57) | 28 (20–38) | 0.33 |
Fibrinogen [g/L] | 2.9 (1.0–8.0) | 2.9 (1.5–7.0) | 0.13 |
PFA-ADP [s] | 99 (65–>300) | ||
PFA-Epi [s] | n. a.* | 107 (62–>300) | n. a. |
Creatinine [mg/dL] | 0.83 (0.21) | 0.79 (0.21) | 0.22 |
*n. a.: not applicable.
Distribution of primary diagnoses.
Diagnosis | Non-PFA group | PFA group |
|
---|---|---|---|
Neuroepithelial tumors |
38 | 40 | 0.79 |
Meningeal tumors |
25 | 21 | |
Metastases | 9 | 10 | |
Pituitary adenomas | 7 | 7 | |
Others | 23 | 15 |
In 15 (16.1%) out of the 93 patients in the PFA group, an abnormal PFA result was found (
Patient outcomes are listed in Tables
Outcome data (continuous variables).
Parameter | Non-PFA group | PFA group |
|
---|---|---|---|
Postoperative Hb [g/L] | 116 (17.8) | 116 (15.7) | 0.90 |
Male | 120 (17.7) | 124 (14.6) | 0.37 |
Female | 111 (16.7) | 110 (14.1) | 0.87 |
Duration of anesthesia [min] | 400 (159) | 384 (131) | 0.44 |
Duration of surgery [min] | 294 (148) | 281 (123) | 0.51 |
Transfusions per patient | |||
Red cells [units] | 0 (0–7) | 0 (0–10) | 0.39 |
Plasma [units] | 0 (0–8) | 0 (0–10) | 0.91 |
Platelets [units] | 0 (0–1) | 0 (0–2) | 0.33 |
Prothrombin complex [IU] | 0 | 0 | n. t.* |
Fibrinogen [g] | 0 (0–4) | 0 (0–3) | 0.91 |
Postoperative GCS | 15 (3–15) | 15 (8–15) | 0.69 |
GOS | 5 (1–5) | 5 (3–5) | 0.10 |
LOS** in intensive care [days] | 2 (0–33) | 2 (0–43) | 0.37 |
*n. t.: not tested; **LOS: length of stay.
Outcome data (categorical variables).
Parameter | Non-PFA group | PFA group | OR (95% CI) |
|
---|---|---|---|---|
Transfused patients [ |
||||
Red cells | 12 | 7 | 0.6 (0.2–1.6) | 0.35 |
Plasma | 17 | 16 | 1.0 (0.5–2.2) | 1.0 |
Platelets | 2 | 4 | 2.3 (0.4–12.6) | 0.43 |
Fibrinogen | 3 | 3 | 1.1 (0.2–5.6) | 1.0 |
Patients receiving hemostatic treatment [ |
||||
Desmopressin | 5 | 13 | 3.2 (1.1–9.2) | 0.045 |
Aprotinin | 2 | 5 | 2.8 (0.5–15.0) | 0.14 |
Tranexamic acid | 0 | 2 | 5.7 (0.3–119.6) | 0.22 |
Clinical data [ |
||||
Intraoperative bleeding tendency | 4 | 6 | 1.7 (0.5–6.2) | 0.42 |
Clinically relevant bleeding complication | 2 | 2 | 1.1 (0.4–2.8) | 1.0 |
Diagnostic tools, in order to inform clinical management, need to be evaluated with respect to their influence on patient outcome. If a change in clinical management driven by the result of a diagnostic procedure does not improve patient outcome, this diagnostic procedure only produces additional costs and should be abandoned. It was the aim of this study to analyze the diagnostic utility of the PFA-100 for the improvement of the postoperative outcome of a clearly defined subgroup of neurosurgical patients. It could be shown that introduction of the PFA-100 into the routine of preoperative screening of elective patients with space-occupying lesions undergoing complex neurosurgical procedures did not result in an improved patient outcome. Several aspects of patient outcome had been studied, and none was significantly different between patients with and without preoperative PFA testing. There was not even a trend for a favorable outcome in any of the analyzed parameters for the patients in the PFA group. However, the patients in the PFA group received more often desmopressin than patients in the non-PFA group but obviously did not benefit from this treatment. Thus, this treatment must be regarded as ineffective in this situation, therefore unnecessarily subjecting these patients to potential side effects of the drug. Furthermore, it has been shown in experimental settings that vasopressin receptors play a role in the development of posttraumatic brain injury, even though type 2 receptors appear not to be involved in this process so far [
However, previous studies have shown that disorders of primary hemostasis, and platelet function in particular, are common among patients undergoing surgery [
Two-thirds of the patients in this study displayed only an abnormal PFA-Epi test. The PFA-Epi is very sensitive to the effect of aspirin and can still be highly abnormal several days after ingestion of the last aspirin dose [
Moreover, perioperative bleeding is often caused by factors that cannot be anticipated preoperatively. One factor is obviously the risk inherent in the difficulty and complexity of the surgical procedure itself. In addition, coagulation disturbances that usually do not prevail or cannot be detected preoperatively (like increased fibrinolysis [
In summary, and to answer the second question, most of the preoperatively detected disorders of primary hemostasis in the study patients appear not to have been clinically relevant. In addition, other factors obviously have influenced bleeding complications and transfusion requirements. This may be exemplified by the fact that one of the patients in the PFA group with a bleeding complication did not have an abnormal PFA-test. This patient may of course have been suffered from a platelet defect not detectable by PFA-testing, which still is to be considered a failure of the analyzed management approach.
The foregoing considerations may explain why there was not a noticeable beneficial effect of desmopressin. It seems that the hemostatic balance and the clinical circumstances of the patients studied are far too complex as to allow a simple one-dimensional approach to be widely successful.
Finally, the PFA-100 still has limited performance characteristics with respect to the diagnosis of clinically relevant disorders of primary hemostasis [
Our study undoubtedly bears one major limitation. The retrospective design is prone to impaired data quality and variables important for patient selection or outcome determination might not have been documented at all. In addition, by selection of a historical control group comparison is always susceptible to time effects that can mask or counteract effects attributable to the intervention studied. We tried to address these issues in several ways. Since PFA screening was introduced for a clearly defined patient group, selection of the control group was therefore quite straightforward and left little room for any subjective influence. This assumption is supported by the fact that both patient groups were highly comparable with respect to all analyzed baseline variables. It is hardly conceivable that any important baseline data with a significant impact on the study outcomes had been missed or would have been different between the groups. We tried to control any time effects by analyzing as short a study period as possible. A period of no more than in total 1.5 years was considered justifiable because data were eagerly needed in order to make a substantiated decision whether to continue or to give up this clinical approach. Furthermore, time effects owing to continuous but not always recognizable improvements in health care usually favor better outcomes of the patients treated later, that is, the PFA group in our case. However, since no better outcome of the patients in the PFA group was observed, this even strengthens the notion that there indeed seems to be no beneficial effect of routine preoperative PFA testing in the patient group studied. Thus, notwithstanding the shortcomings of our study design, we are convinced that the conclusions drawn from the data at hand are based on substantial evidence and can serve as a starting point for future investigations of this topic. However, given the low number of bleeding complications in this study, randomized trials designed to evaluate measures aimed at reducing the perioperative bleeding complication rates may be difficult to perform. The statistical proof of even a relative risk reduction of 50%, that is, from a 2% to a 1% bleeding complication rate, would need nearly 2,000 study participants per group.
This is the first study analyzing a systematic management approach to reduce perioperative bleeding complications and transfusion requirements in a cohort of vulnerable neurosurgical patients. According to our data, the correction of a platelet function deficit determined by a commercially available test of platelet function could not reduce bleeding complications and transfusion requirements but resulted in increased administration of desmopressin, a hemostatic drug which itself has potential unfavorable side effects that might compromise patient outcome. Further studies are needed to determine which patients may actually benefit from routine preoperative testing of platelet function and how this testing has to be tailored in order to detect only clinically significant defects.