- Email: [email protected]
Central line placement in patients with and without prophylactic plasma
Journal of Critical Care (2012) 27, 529.e9–529.e13
Central line placement in patients with and without prophylactic plasma Gerardo P. Carino MD, PhDa , Arseniy V. Tsapenko MDa , Joseph D. Sweeney MDb,⁎ a
Department of Critical Care, The Miriam Hospital, Providence, RI 02906, USA Department of Transfusion Medicine, The Miriam Hospital, Providence, RI 02906, USA
b
Keywords: Prophylactic plasma; Venous catheterization
Abstract Central line placement (CLP) is a common life-saving intervention in critically ill patients, and patients with coagulation abnormalities as identified by an abnormal international normalized ratio (INR) may receive prophylactic plasma transfusion before the procedure despite previously published data that such a practice is not efficacious. Over a 14-month period, 287 CLPs were performed in the intensive care unit. The use of plasma preprocedure in patients with an elevated INR was generally discouraged but left to the discretion of the operator. A total of 100 lines were placed in patients with a preprocedure INR greater than 1.5, 27 of whom received prophylactic fresh frozen plasma. Only 1 case of bleeding was observed in a patient with an INR of 3.9, who received fresh frozen plasma preprocedure (0/73 vs 1/27; P = .6). The occurrence of bleeding was very low overall with CLP (0.3%; 95% confidence interval, 0%-2%), and no benefit of prophylactic plasma was observed. © 2012 Elsevier Inc. All rights reserved.
1. Introduction Critically ill patients in the intensive care unit (ICU) frequently require central venous catheters for drug administration or hemodynamic monitoring. Common comorbidities such as liver disease, disseminated intravascular coagulation, and vitamin K antagonism or deficiency may be associated with laboratory evidence of coagulation abnormalities. These patients have historically been often perceived to be at increased risk for bleeding during invasive diagnostic procedures or for the placement of central lines. On account of this, it has been a practice to transfuse either fresh frozen plasma (FFP) or 24-hour frozen plasma (FP24) ⁎ Corresponding author. Blood Bank, The Miriam Hospital, Providence, RI 02906, USA. Tel.: +1 401 793 4810; fax: +1 401 351 5928. E-mail address: [email protected] (J.D. Sweeney). 0883-9441/$ – see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcrc.2011.12.016
prophylactically [1]. Specifically with regard to central line placement (CLP), the use of such FFP or FP24 is intended to avoid bleeding complications such as insertional site bleeding, hematoma, or even hemothorax. However, this practice of prophylactic FFP administration is not based on high-grade evidence such as randomized trials [1] or even observational studies [2-5] and would appear to be driven largely by expert opinion, practice guidelines based on consensus [6], historical precedent, and perhaps even concern regarding litigation in the event of a complication. Since the 1960s, several observational studies examined bleeding events in patients with coagulopathies undergoing a variety of invasive procedures, including CLP. Low overall risk was demonstrated [2-5], and no protective effect of plasma was shown in those studies in which it was transfused [4,5]. The now widespread utilization of ultrasound guidance for CLP may further reduce this already low risk [7,8]. It is
529.e10
G.P. Carino et al.
well recognized that plasma transfusion can be associated with significant morbidity and mortality such as transfusionrelated acute lung injury, transfusion-associated circulatory overload, and allergic or febrile reactions [9-11]. In 1 retrospective study in ICU patients, the use of prophylactic FFP transfusion before line placement did not decrease hemorrhage yet was associated with an increased risk of pulmonary injury within 48 hours of the transfusion (18% vs 4%). These findings clearly suggest that a conservative approach to prophylactic plasma transfusion is appropriate as the risk-benefit ratio appears unfavorable. Based on the previously published observational studies in patients with coagulopathy [2-5] and other reports on the lack of known efficacy of FFP transfusion is this context [12,13], we adopted a practice in our ICU to discourage prophylactic FFP transfusion in nonbleeding patients before CLP when the international normalized ratio (INR) is less than 3.0 and not encourage the use of plasma when the INR exceeds 3.0. However, the decision to transfuse plasma was left to the operator. The goal of this retrospective observational study was to compare the prevalence of bleeding in patients who underwent CLP with and without prophylactic FFP transfusions after adoption of this practice.
determining the presence of an abnormality in the extrinsic system and for a decision regarding plasma transfusion [6]. All blood components given either preprocedure (b12 hours) or postprocedure (within 24 hours) were also recorded as was the postprocedure hemoglobin level. Any lines changed over a wire were excluded from analysis as the bleeding risk from such procedures was deemed too low. Clinically evident bleeding events were considered to be a documented hematoma at the insertion site, any need for blood transfusion within 24 hours postprocedure attributable to the procedure, or any need for an intervention beyond local manual pressure. A postprocedure roentgenogram was obtained after each line placement and reviewed for anatomically correct line location. Any new opacity in the hemithorax was considered a new bleeding event. Data were entered into an Excel spreadsheet and exported to a statistical application (Epistat, Richardson, Tex) for analysis. Data were analyzed by descriptive statistics, and independent t tests and χ2 analyses were performed, as appropriate. Statistical significance was defined as P b .05.
2. Materials and methods
A total of 287 CLPs were performed over this 14-month period in 130 males and 151 females. Six females had catheters inserted on separate occasions on the contralateral side, 2 of whom had the 2 insertions performed on the same day, Only 1 procedure-related bleeding event occurred. Once available, ultrasound (US) guidance was used extensively for CLP placement and was used for 44% of the total IJ lines. The right IJ was the most preferred site (57%) presumably because most operators are right handed and the teaching of the IJ route is emphasized more at our institution. Ultrasound guidance for the subclavian site is possible; however, it is much more technically difficult and, therefore, was not used [7]. Stratification of this population based on INR is shown in Table 1. Not unexpectedly, there is a trend to use prophylactic plasma as the INR increases and approximately 50% of operators choose to administer prophylactic plasma when the INR was greater than 3.0. Administration of plasma was rare when the INR was less than 1.5, a commonly used threshold from guidelines. Sixty-six procedures were performed in patients with INR values of 1.5 to 2.9 without prophylactic plasma. Historically, INRs in this range would commonly have been associated with plasma transfusion in our institution. No clear rationale for FFP transfusion (other than the INR result) could be found for the 35 patients who received FFP. During this 14-month period, there were numerous attending physicians, fellows, and residents who rotated between different training hospitals. It is possible that the decision to transfuse FFP was related to prior experience in other ICUs in which prophylactic FFP transfusion in patients with elevated INRs could have been routine. Data regarding post-FFP transfusion INRs (b6 hours after
This retrospective cohort study was approved by the Institutional Review Board of the Miriam Hospital, Providence, RI. The Miriam Hospital is an academic community hospital, and the ICU is a 16-bed combined medical and surgical unit. The period of this study was a 14-month period between July 1, 2007, and August 30, 2008. All central line insertions were placed by operators with varying levels of experience but always under direct supervision of attending physicians in the case of residents and interns. The choice of anatomical access site, either internal jugular (IJ) or subclavian, and side of insertion was made by the operator considering patient preference, anatomical factors, and any history of previous difficulty with line placement. An ultrasound device (Sonosite S-ICU; Sonosite, Inc, Bothell, Wash) became available in early 2008 and was used for ultrasonic guidance in some of these procedures. In all cases, the operator used his/her discretion regarding the prophylactic use of plasma but was encouraged not to do so for INR values less than 3.0. Where plasma was used, a weight-based dose of 10 to 15 mL/kg was recommended. All plasma administered was freshly thawed FFP: neither FP24 nor liquid stored plasma was given. The data extracted retrospectively from the patients' chart included age, sex, clinical diagnosis, preprocedure activated partial thromboplastin time, INR, hemoglobin level, platelet count, creatinine, catheter location, use of ultrasound, and the occurrence of any clinically evident bleeding complication. The INR was chosen rather than the prothrombin time as the INR has come into more common use as a basis for
3. Results
CLP in patients with and without prophylactic plasma
529.e11
Table 1 Blood component therapy after separation of patients into 2 populations defined by the use of prophylactic plasma with stratification by INR INR
No. of catheters
No. of units FFP transfused
No. of patients receiving RBC transfusion (%)
Total no. of units RBC transfused
No. of procedurerelated bleeds
I. Patients who did not receive FFP preprocedure Unknown 40 0 b1.5 138 0 1.5-2.9 66 0 ≥3.0 6 0 Total 250 0
4 (10%) 28 (20%) 13 (20%) 0 45 (18%)
8 72 33 0 113
0 0 0 0
II. Patients who received FFP preprocedure b1.5 7 26 1.5-2.9 19 74 ≥3.0 9 56 Total 35 156
6 (86%) 10 (53%) 6 (67%) 22 (63%)
34 36 22 92
0 0 1
transfusion) were available in only a few patients and, hence, are not included in the table. More red cell transfusions were given to the FFP transfused patients overall (22/35 vs 45/250; P b .01) and for each of the INR stratifications, except the INR greater than 3.0 group, although this likely reflect the small numbers in this substratification (INR less than 1.5, P b .01; INR 1-5, 2.9, P b .01; INR greater than 3.0; P = .09). This is not to infer that the plasma precipitated red cell transfusion in the FFP treated patients, although it could have caused hemodilution in already anemic patients. It is possible that these patients were more critically ill and that this influenced the decision to administer plasma, but no direct verification of this was available. Those patients who received prophylactic plasma were transfused with an average of 3.8 U, which is close to the recommended dose of 10 to 15 mL/kg. Taken together
with the absence of any clinically evident procedurerelated bleeding in the nonplasma transfusion groups, this supports the position that prophylactic plasma would appear to confer no benefit when the INR is elevated in this range. There was no information regarding preprocedure INR in several cases (n = 40), none of whom received plasma. This was because the test was not performed, as the patient was deemed to be at low risk for coagulopathy or because the emergent nature of line insertion precluded waiting for a test result. However, this group is included as we believe that it likely mirrors actual practice. It is intriguing that 1 potential implication of this inclusion is that the mere performance of the prothrombin time preprocedure is a risk factor for FFP transfusion. Table 2 shows the patient population separated by the use or otherwise of US guidance and prophylactic plasma. There
Table 2 Laboratory data and red cell transfusion in the patient population separated by the use of prophylactic plasma and ultrasonic guidance Received plasma
n Age (y) INR aPTT (s) Creatinine (mg/dL) Platelet count (109/L) Hemoglobin level pre-CLP (g/dL) Hemoglobin level post-CLP (g/dL) Patients receiving RBC No. of RBC units transfused No. of procedure-related bleeds
Did not receive plasma
US-guided
Non–US-guided
US-guided
Non–US-guided
13 60 ± 23 2.5 ± 1.1 ⁎ 42.9 ± 12.5 ⁎ 1.6 ± 1.0 149 ± 62 ⁎ 9.6 ± 1.9 9.5 ± 1.8 7 (54%) 24 1
22 67 ± 15 2.1 ± 0.9 † 45.5 ± 24.0 † 1.7 ± 1.1 152 ± 117 † 9.5 ± 2.7 † 9.5 ± 1.9 15 (68%) 68 0
78 66 ± 17 1.6 ± 0.8 33.2 ± 8.4 2.1 ± 2.0 234 ± 108 10.4 ± 1.6 9.6 ± 1.5 11 (14%) 31 0
174 67 ± 15 1.4 ± 0.7 34.7 ± 15.4 1.8 ± 1.6 228 ± 109 10.7 ± 2.3 9.9 ± 1.9 34 (20%) 82 0
Data are the mean ± 1 SD. aPTT indicates activated partial thromboplastin time. ⁎ P b .05 between the US-guided groups that did or did not receive plasma. † P b .05 between the non–US-guided groups that did or did not receive plasma.
529.e12 is no evidence that US guidance was more likely to be used in patients with higher INRs or lower platelet counts. This suggests that the decision to give plasma was driven by the INR by those operators who elected to give plasma and the availability of US guidance did not influence this decision supporting the notion that the concern regarding procedurerelated bleeding is driven by the perceived systemic disorder of hemostasis (high INR) and not concern regarding correct anatomical placement of the cannula. Sixteen patients had a platelet count of less than 100 × 109/L, of whom 5 received a total of 7 doses of platelets. All of the patients who received platelets had a preprocedure platelet count less than 40 × 109/L. These patients would have received plasma in the platelet product (approximately 1-2 U equivalents by volume). The FV (Factor V) and FVIII levels would have been lower than FFP [14], but we cannot exclude that the clotting factors contained in the anticoagulated plasma of the platelet products transfused could have improved hemostasis and reduced the risk of a procedure-related bleed. Because of concerns that those patients who received red cell transfusions could have been transfused because of procedure-associated bleeding, which was not clinically evidence, the population were further examined into those who received red cell transfusions (n = 67) and those who did not (n = 220). Each patient who received red cells was examined for cause. Not unexpected, the most common reasons for red cell transfusion were gastrointestinal bleeding (33%); bleeding from an undetermined site, but not intrathoracic, (30%); and anemia (hemoglobin level b 7 g/dL, 37%). No complications were reported nor observed in association with the transfusion of any of these blood components. However, given the number of components transfused (156 U FFP, 205 U red blood cell [RBC], and 7 doses of platelets), this is not unexpected. There was only 1 patient who experienced a bleeding complication in this entire cohort, giving a point estimate of 0.3% (95% confidence interval, 0%-2%). This patient had a left IJ catheter placed with US guidance with a preprocedure INR of 3.9 and received 5 U of FFP preprocedure. Postprocedure a large hematoma was noted, and both local compression and an additional 5 U of FFP were used to stop the bleeding. Of note, the patients' creatinine was 1.7 mg/L, and platelet count was 118 × 109/L before the procedure. Technical difficulties with placing the line and multiple cannulation attempts were documented and ultimately placement of the line failed.
4. Discussion Central venous catheterization provides vital secure venous access and, therefore, a potential life-saving intervention in many critically ill patients. Bleeding risk appears to be linked to the number of sticks required for
G.P. Carino et al. successful venous access or inadvertent arterial punctures, which has been shown to be increased in lines placed using solely the landmark method [15,16]. In nonemergent situations, bleeding risk can be more fully assessed, and ultrasound guidance can be used to decrease the number of attempts required for successful cannulation, complications, and overall failure rate [15-17]. In our ICU, lines are commonly placed emergently in patients with coagulation abnormalities. In most cases, this is performed without prophylactic FFP transfusion even when the INR is known to be greater than 3.0. Only 1 patient in this retrospective series developed bleeding after their procedure, and there was operator difficulty with line placement in this patient. Interestingly, even with a recommendation to withhold FFP transfusion unless the INR was less than 3.0, 7 (5%) of 145 patients with an INR less than 1.5, and 19 (22%) of 84 patients with the INR between 1.5 and 3.0 received FFP transfusions. This is likely due to continued physician belief that there is some benefit for prophylactic transfusions or perceived litigation mitigation [18]. More extensive education about transfusion recommendations will likely be necessary to reduce the number of these nonefficacious and potentially dangerous transfusions [9,12,13]. We recognize that there are several previous reports regarding the risk of bleeding and the practice of prophylactic plasma transfusions in the context of central venous catherization [2-5]. These previous reports had either a large number of patients with thrombocytopenia or a large fraction of patients with trauma [5] or patients restricted to liver disease or liver transplantation [2-4]. Furthermore, these previous studies expressed the prothrombin time as either an absolute clotting time [3,5] or as a percentage of normal [2] or included patients who received plasma without clearly separating those who received plasma from those who did not, resulting in difficulty in deciphering an effect or otherwise of plasma transfusion [4,5]. Hence, we cannot combine the data in a meta-analysis separating prophylactic plasma recipients from nonplasma recipients with stratification by INR. We believe that our study population in a nontrauma center with a mixed medical-surgical population is closer to the patient population in a large community hospital and our observations may be more generalizable to such institutions. There are several important limitations to this retrospective observational study. First, as with all such studies, there may be important confounders, which might have minimized bleeding risk such as the experience level of the operators or close supervision of inexperienced operators, as this can be an important determinant of bleeding risk [5,7]. Second, the nature of the coagulopathy was not clarified prospectively or retrospectively by clotting factor assays, but in practice, this is unlikely to be performed routinely in ICU patients, and hence, the INR will be the coagulation test, which is most commonly used in practice to guide plasma transfusion decisions. Third, we could not
CLP in patients with and without prophylactic plasma clarify as to why FFP was transfused by the operators in some cases but not in others but believe that this is consistent with common practice reflecting intrainstitutional variation. Fourth, 67 patients received red cell transfusion immediately before, concurrent with, or immediately after the procedure. It is possible that some of these allogeneic red cell transfusions could have been in response to CLP associated bleeding. We sought to address this by examining these patients by reason for red cell transfusion and were able only to find a single case where the red cell transfusion was causally linked to intrathoracic bleeding. We believe that our data are consistent with previous reports with regard to the low observed bleeding risk but differ from previous reports in the clear separation of FFP recipients vs non-FFP recipients coupled with INR stratification. Our results support the position that prophylactic plasma transfusion in this context is not efficacious [19] and potentially dangerous [20] and should be actively discouraged when the INR is less than 3.0 and used highly selectively in patients with higher INRs.
References [1] Stanworth SJ, Brunskill SJ, Hyde CJ, et al. Is fresh frozen plasma clinically effective? A systematic review of randomized controlled trials. Br J Haematol 2004;126:139-52. [2] Foster PF, Moore LR, Sankary HN, et al. Central venous catheterization in patients with coagulopathy. Arch Surg 1992;127:273-5. [3] Doerfler ME, Kaufman B, Goldenberg AS. Central venous catheter placement in patients with disorders of hemostasis. Chest 1996;110: 185-8. [4] Fisher NC, Mutimer DJ. Central venous cannulation in patients with liver disease and coagulopathy—a prospective audit. Intensive Care Med 1999;25:481-5. [5] DeLoughery TG, Liebler JM, Simonds V, et al. Invasive line placement in critically ill patients: do hemostatic defects matter? Transfusion 1996;36:827-31. [6] Malloy PC, Grassi CJ, Kundu S, et al. Consensus guidelines for periprocedural management of coagulation status and hemostasis risk in percutaneous image-guided interventions. J Vasc Interv Radiol 2009;20:S240-9.
529.e13 [7] Gualtieri E, Deppe SA, Sipperly ME, et al. Subclavian venous catheterization: greater success rate for less experienced operators using ultrasound guidance. Crit Care Med 1995;23:692-7. [8] Tercan F, Ozkan U, Oguzkurt L. US-guided placement of central vein catheters in patients with disorders of hemostasis. Eur J Radiol 2008;65:253-6. [9] Dara SI, Rana R, Afessa B, et al. Fresh frozen plasma transfusion in critically ill medical patients with coagulopathy. Crit Care Med 2005;33:2667-71. [10] Napolitano LM, Kurek S, Luchette FA, et al. Clinical practice guideline: red blood cell transfusion in adult trauma and critical care. Crit Care Med 2009;37:3124-57. [11] Silliman CC, Fung YL, Ball JB, et al. Transfusion-related acute lung injury (TRALI): current concepts and misconceptions. Blood Rev 2009;23:245-55. [12] Segal JB, Dzik WH, et al. Paucity of studies to support that abnormal coagulation test results predict bleeding in the setting of invasive procedures: an evidence-based review. Transfusion 2005; 45:1413-25. [13] Holland LL, Brooks JP. Toward rational fresh frozen plasma transfusion: the effect of plasma transfusion on coagulation test results. Am J Clin Pathol 2006;126:133-9. [14] Sweeney JD, Kouttab NM, Holme SH, et al. Prestorage pooled whole blood derived leukoreduced platelets stored for seven days preserve acceptable quality and do not show evidence of a mixed lymphocyte reaction. Transfusion 2004;44:1212-9. [15] Miller AH, Roth BA, Mills TJ, et al. Ultrasound guidance versus the landmark technique for the placement of central venous catheters in the emergency department. Acad Emerg Med 2002;9:800-5. [16] Wigmore TJ, Symthe JF, Hacking MB, et al. Effect of the implementation of NICE guidelines for ultrasound guidance on the complication rates associated with central venous catheter placement in patients presenting for routine surgery in a tertiary referral centre. Br J Anaesth 2007;99:662-5. [17] Balls A, LoVecchio F, Kroeger A, et al. Ultrasound guidance for central venous catheter placement: results from the Central Line Emergency Access Registry Database. Am J Emerg Med 2010;28: 561-7. [18] Vlaar AP, In Der Maur AL, Binnekade JM, et al. A survey of physicians' reasons to transfuse plasma and platelets in the critically ill: a prospective single-centre cohort study. Transfus Med 2009;19: 207-12. [19] Stanworth SJ, Walsh TS. Fresh frozen plasma: friend or faux pas in critical illness? Crit Care Med 2005;33:2714-6. [20] Narick C, Triulzi DJ, Yazer MH. Transfusion associated circulatory overload after plasma transfusion. Transfusion 2011 [Published online ISSN: 1537-2995].
Central line placement in patients with and without prophylactic plasma Gerardo P. Carino MD, PhDa , Arseniy V. Tsapenko MDa , Joseph D. Sweeney MDb,⁎ a
Department of Critical Care, The Miriam Hospital, Providence, RI 02906, USA Department of Transfusion Medicine, The Miriam Hospital, Providence, RI 02906, USA
b
Keywords: Prophylactic plasma; Venous catheterization
Abstract Central line placement (CLP) is a common life-saving intervention in critically ill patients, and patients with coagulation abnormalities as identified by an abnormal international normalized ratio (INR) may receive prophylactic plasma transfusion before the procedure despite previously published data that such a practice is not efficacious. Over a 14-month period, 287 CLPs were performed in the intensive care unit. The use of plasma preprocedure in patients with an elevated INR was generally discouraged but left to the discretion of the operator. A total of 100 lines were placed in patients with a preprocedure INR greater than 1.5, 27 of whom received prophylactic fresh frozen plasma. Only 1 case of bleeding was observed in a patient with an INR of 3.9, who received fresh frozen plasma preprocedure (0/73 vs 1/27; P = .6). The occurrence of bleeding was very low overall with CLP (0.3%; 95% confidence interval, 0%-2%), and no benefit of prophylactic plasma was observed. © 2012 Elsevier Inc. All rights reserved.
1. Introduction Critically ill patients in the intensive care unit (ICU) frequently require central venous catheters for drug administration or hemodynamic monitoring. Common comorbidities such as liver disease, disseminated intravascular coagulation, and vitamin K antagonism or deficiency may be associated with laboratory evidence of coagulation abnormalities. These patients have historically been often perceived to be at increased risk for bleeding during invasive diagnostic procedures or for the placement of central lines. On account of this, it has been a practice to transfuse either fresh frozen plasma (FFP) or 24-hour frozen plasma (FP24) ⁎ Corresponding author. Blood Bank, The Miriam Hospital, Providence, RI 02906, USA. Tel.: +1 401 793 4810; fax: +1 401 351 5928. E-mail address: [email protected] (J.D. Sweeney). 0883-9441/$ – see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcrc.2011.12.016
prophylactically [1]. Specifically with regard to central line placement (CLP), the use of such FFP or FP24 is intended to avoid bleeding complications such as insertional site bleeding, hematoma, or even hemothorax. However, this practice of prophylactic FFP administration is not based on high-grade evidence such as randomized trials [1] or even observational studies [2-5] and would appear to be driven largely by expert opinion, practice guidelines based on consensus [6], historical precedent, and perhaps even concern regarding litigation in the event of a complication. Since the 1960s, several observational studies examined bleeding events in patients with coagulopathies undergoing a variety of invasive procedures, including CLP. Low overall risk was demonstrated [2-5], and no protective effect of plasma was shown in those studies in which it was transfused [4,5]. The now widespread utilization of ultrasound guidance for CLP may further reduce this already low risk [7,8]. It is
529.e10
G.P. Carino et al.
well recognized that plasma transfusion can be associated with significant morbidity and mortality such as transfusionrelated acute lung injury, transfusion-associated circulatory overload, and allergic or febrile reactions [9-11]. In 1 retrospective study in ICU patients, the use of prophylactic FFP transfusion before line placement did not decrease hemorrhage yet was associated with an increased risk of pulmonary injury within 48 hours of the transfusion (18% vs 4%). These findings clearly suggest that a conservative approach to prophylactic plasma transfusion is appropriate as the risk-benefit ratio appears unfavorable. Based on the previously published observational studies in patients with coagulopathy [2-5] and other reports on the lack of known efficacy of FFP transfusion is this context [12,13], we adopted a practice in our ICU to discourage prophylactic FFP transfusion in nonbleeding patients before CLP when the international normalized ratio (INR) is less than 3.0 and not encourage the use of plasma when the INR exceeds 3.0. However, the decision to transfuse plasma was left to the operator. The goal of this retrospective observational study was to compare the prevalence of bleeding in patients who underwent CLP with and without prophylactic FFP transfusions after adoption of this practice.
determining the presence of an abnormality in the extrinsic system and for a decision regarding plasma transfusion [6]. All blood components given either preprocedure (b12 hours) or postprocedure (within 24 hours) were also recorded as was the postprocedure hemoglobin level. Any lines changed over a wire were excluded from analysis as the bleeding risk from such procedures was deemed too low. Clinically evident bleeding events were considered to be a documented hematoma at the insertion site, any need for blood transfusion within 24 hours postprocedure attributable to the procedure, or any need for an intervention beyond local manual pressure. A postprocedure roentgenogram was obtained after each line placement and reviewed for anatomically correct line location. Any new opacity in the hemithorax was considered a new bleeding event. Data were entered into an Excel spreadsheet and exported to a statistical application (Epistat, Richardson, Tex) for analysis. Data were analyzed by descriptive statistics, and independent t tests and χ2 analyses were performed, as appropriate. Statistical significance was defined as P b .05.
2. Materials and methods
A total of 287 CLPs were performed over this 14-month period in 130 males and 151 females. Six females had catheters inserted on separate occasions on the contralateral side, 2 of whom had the 2 insertions performed on the same day, Only 1 procedure-related bleeding event occurred. Once available, ultrasound (US) guidance was used extensively for CLP placement and was used for 44% of the total IJ lines. The right IJ was the most preferred site (57%) presumably because most operators are right handed and the teaching of the IJ route is emphasized more at our institution. Ultrasound guidance for the subclavian site is possible; however, it is much more technically difficult and, therefore, was not used [7]. Stratification of this population based on INR is shown in Table 1. Not unexpectedly, there is a trend to use prophylactic plasma as the INR increases and approximately 50% of operators choose to administer prophylactic plasma when the INR was greater than 3.0. Administration of plasma was rare when the INR was less than 1.5, a commonly used threshold from guidelines. Sixty-six procedures were performed in patients with INR values of 1.5 to 2.9 without prophylactic plasma. Historically, INRs in this range would commonly have been associated with plasma transfusion in our institution. No clear rationale for FFP transfusion (other than the INR result) could be found for the 35 patients who received FFP. During this 14-month period, there were numerous attending physicians, fellows, and residents who rotated between different training hospitals. It is possible that the decision to transfuse FFP was related to prior experience in other ICUs in which prophylactic FFP transfusion in patients with elevated INRs could have been routine. Data regarding post-FFP transfusion INRs (b6 hours after
This retrospective cohort study was approved by the Institutional Review Board of the Miriam Hospital, Providence, RI. The Miriam Hospital is an academic community hospital, and the ICU is a 16-bed combined medical and surgical unit. The period of this study was a 14-month period between July 1, 2007, and August 30, 2008. All central line insertions were placed by operators with varying levels of experience but always under direct supervision of attending physicians in the case of residents and interns. The choice of anatomical access site, either internal jugular (IJ) or subclavian, and side of insertion was made by the operator considering patient preference, anatomical factors, and any history of previous difficulty with line placement. An ultrasound device (Sonosite S-ICU; Sonosite, Inc, Bothell, Wash) became available in early 2008 and was used for ultrasonic guidance in some of these procedures. In all cases, the operator used his/her discretion regarding the prophylactic use of plasma but was encouraged not to do so for INR values less than 3.0. Where plasma was used, a weight-based dose of 10 to 15 mL/kg was recommended. All plasma administered was freshly thawed FFP: neither FP24 nor liquid stored plasma was given. The data extracted retrospectively from the patients' chart included age, sex, clinical diagnosis, preprocedure activated partial thromboplastin time, INR, hemoglobin level, platelet count, creatinine, catheter location, use of ultrasound, and the occurrence of any clinically evident bleeding complication. The INR was chosen rather than the prothrombin time as the INR has come into more common use as a basis for
3. Results
CLP in patients with and without prophylactic plasma
529.e11
Table 1 Blood component therapy after separation of patients into 2 populations defined by the use of prophylactic plasma with stratification by INR INR
No. of catheters
No. of units FFP transfused
No. of patients receiving RBC transfusion (%)
Total no. of units RBC transfused
No. of procedurerelated bleeds
I. Patients who did not receive FFP preprocedure Unknown 40 0 b1.5 138 0 1.5-2.9 66 0 ≥3.0 6 0 Total 250 0
4 (10%) 28 (20%) 13 (20%) 0 45 (18%)
8 72 33 0 113
0 0 0 0
II. Patients who received FFP preprocedure b1.5 7 26 1.5-2.9 19 74 ≥3.0 9 56 Total 35 156
6 (86%) 10 (53%) 6 (67%) 22 (63%)
34 36 22 92
0 0 1
transfusion) were available in only a few patients and, hence, are not included in the table. More red cell transfusions were given to the FFP transfused patients overall (22/35 vs 45/250; P b .01) and for each of the INR stratifications, except the INR greater than 3.0 group, although this likely reflect the small numbers in this substratification (INR less than 1.5, P b .01; INR 1-5, 2.9, P b .01; INR greater than 3.0; P = .09). This is not to infer that the plasma precipitated red cell transfusion in the FFP treated patients, although it could have caused hemodilution in already anemic patients. It is possible that these patients were more critically ill and that this influenced the decision to administer plasma, but no direct verification of this was available. Those patients who received prophylactic plasma were transfused with an average of 3.8 U, which is close to the recommended dose of 10 to 15 mL/kg. Taken together
with the absence of any clinically evident procedurerelated bleeding in the nonplasma transfusion groups, this supports the position that prophylactic plasma would appear to confer no benefit when the INR is elevated in this range. There was no information regarding preprocedure INR in several cases (n = 40), none of whom received plasma. This was because the test was not performed, as the patient was deemed to be at low risk for coagulopathy or because the emergent nature of line insertion precluded waiting for a test result. However, this group is included as we believe that it likely mirrors actual practice. It is intriguing that 1 potential implication of this inclusion is that the mere performance of the prothrombin time preprocedure is a risk factor for FFP transfusion. Table 2 shows the patient population separated by the use or otherwise of US guidance and prophylactic plasma. There
Table 2 Laboratory data and red cell transfusion in the patient population separated by the use of prophylactic plasma and ultrasonic guidance Received plasma
n Age (y) INR aPTT (s) Creatinine (mg/dL) Platelet count (109/L) Hemoglobin level pre-CLP (g/dL) Hemoglobin level post-CLP (g/dL) Patients receiving RBC No. of RBC units transfused No. of procedure-related bleeds
Did not receive plasma
US-guided
Non–US-guided
US-guided
Non–US-guided
13 60 ± 23 2.5 ± 1.1 ⁎ 42.9 ± 12.5 ⁎ 1.6 ± 1.0 149 ± 62 ⁎ 9.6 ± 1.9 9.5 ± 1.8 7 (54%) 24 1
22 67 ± 15 2.1 ± 0.9 † 45.5 ± 24.0 † 1.7 ± 1.1 152 ± 117 † 9.5 ± 2.7 † 9.5 ± 1.9 15 (68%) 68 0
78 66 ± 17 1.6 ± 0.8 33.2 ± 8.4 2.1 ± 2.0 234 ± 108 10.4 ± 1.6 9.6 ± 1.5 11 (14%) 31 0
174 67 ± 15 1.4 ± 0.7 34.7 ± 15.4 1.8 ± 1.6 228 ± 109 10.7 ± 2.3 9.9 ± 1.9 34 (20%) 82 0
Data are the mean ± 1 SD. aPTT indicates activated partial thromboplastin time. ⁎ P b .05 between the US-guided groups that did or did not receive plasma. † P b .05 between the non–US-guided groups that did or did not receive plasma.
529.e12 is no evidence that US guidance was more likely to be used in patients with higher INRs or lower platelet counts. This suggests that the decision to give plasma was driven by the INR by those operators who elected to give plasma and the availability of US guidance did not influence this decision supporting the notion that the concern regarding procedurerelated bleeding is driven by the perceived systemic disorder of hemostasis (high INR) and not concern regarding correct anatomical placement of the cannula. Sixteen patients had a platelet count of less than 100 × 109/L, of whom 5 received a total of 7 doses of platelets. All of the patients who received platelets had a preprocedure platelet count less than 40 × 109/L. These patients would have received plasma in the platelet product (approximately 1-2 U equivalents by volume). The FV (Factor V) and FVIII levels would have been lower than FFP [14], but we cannot exclude that the clotting factors contained in the anticoagulated plasma of the platelet products transfused could have improved hemostasis and reduced the risk of a procedure-related bleed. Because of concerns that those patients who received red cell transfusions could have been transfused because of procedure-associated bleeding, which was not clinically evidence, the population were further examined into those who received red cell transfusions (n = 67) and those who did not (n = 220). Each patient who received red cells was examined for cause. Not unexpected, the most common reasons for red cell transfusion were gastrointestinal bleeding (33%); bleeding from an undetermined site, but not intrathoracic, (30%); and anemia (hemoglobin level b 7 g/dL, 37%). No complications were reported nor observed in association with the transfusion of any of these blood components. However, given the number of components transfused (156 U FFP, 205 U red blood cell [RBC], and 7 doses of platelets), this is not unexpected. There was only 1 patient who experienced a bleeding complication in this entire cohort, giving a point estimate of 0.3% (95% confidence interval, 0%-2%). This patient had a left IJ catheter placed with US guidance with a preprocedure INR of 3.9 and received 5 U of FFP preprocedure. Postprocedure a large hematoma was noted, and both local compression and an additional 5 U of FFP were used to stop the bleeding. Of note, the patients' creatinine was 1.7 mg/L, and platelet count was 118 × 109/L before the procedure. Technical difficulties with placing the line and multiple cannulation attempts were documented and ultimately placement of the line failed.
4. Discussion Central venous catheterization provides vital secure venous access and, therefore, a potential life-saving intervention in many critically ill patients. Bleeding risk appears to be linked to the number of sticks required for
G.P. Carino et al. successful venous access or inadvertent arterial punctures, which has been shown to be increased in lines placed using solely the landmark method [15,16]. In nonemergent situations, bleeding risk can be more fully assessed, and ultrasound guidance can be used to decrease the number of attempts required for successful cannulation, complications, and overall failure rate [15-17]. In our ICU, lines are commonly placed emergently in patients with coagulation abnormalities. In most cases, this is performed without prophylactic FFP transfusion even when the INR is known to be greater than 3.0. Only 1 patient in this retrospective series developed bleeding after their procedure, and there was operator difficulty with line placement in this patient. Interestingly, even with a recommendation to withhold FFP transfusion unless the INR was less than 3.0, 7 (5%) of 145 patients with an INR less than 1.5, and 19 (22%) of 84 patients with the INR between 1.5 and 3.0 received FFP transfusions. This is likely due to continued physician belief that there is some benefit for prophylactic transfusions or perceived litigation mitigation [18]. More extensive education about transfusion recommendations will likely be necessary to reduce the number of these nonefficacious and potentially dangerous transfusions [9,12,13]. We recognize that there are several previous reports regarding the risk of bleeding and the practice of prophylactic plasma transfusions in the context of central venous catherization [2-5]. These previous reports had either a large number of patients with thrombocytopenia or a large fraction of patients with trauma [5] or patients restricted to liver disease or liver transplantation [2-4]. Furthermore, these previous studies expressed the prothrombin time as either an absolute clotting time [3,5] or as a percentage of normal [2] or included patients who received plasma without clearly separating those who received plasma from those who did not, resulting in difficulty in deciphering an effect or otherwise of plasma transfusion [4,5]. Hence, we cannot combine the data in a meta-analysis separating prophylactic plasma recipients from nonplasma recipients with stratification by INR. We believe that our study population in a nontrauma center with a mixed medical-surgical population is closer to the patient population in a large community hospital and our observations may be more generalizable to such institutions. There are several important limitations to this retrospective observational study. First, as with all such studies, there may be important confounders, which might have minimized bleeding risk such as the experience level of the operators or close supervision of inexperienced operators, as this can be an important determinant of bleeding risk [5,7]. Second, the nature of the coagulopathy was not clarified prospectively or retrospectively by clotting factor assays, but in practice, this is unlikely to be performed routinely in ICU patients, and hence, the INR will be the coagulation test, which is most commonly used in practice to guide plasma transfusion decisions. Third, we could not
CLP in patients with and without prophylactic plasma clarify as to why FFP was transfused by the operators in some cases but not in others but believe that this is consistent with common practice reflecting intrainstitutional variation. Fourth, 67 patients received red cell transfusion immediately before, concurrent with, or immediately after the procedure. It is possible that some of these allogeneic red cell transfusions could have been in response to CLP associated bleeding. We sought to address this by examining these patients by reason for red cell transfusion and were able only to find a single case where the red cell transfusion was causally linked to intrathoracic bleeding. We believe that our data are consistent with previous reports with regard to the low observed bleeding risk but differ from previous reports in the clear separation of FFP recipients vs non-FFP recipients coupled with INR stratification. Our results support the position that prophylactic plasma transfusion in this context is not efficacious [19] and potentially dangerous [20] and should be actively discouraged when the INR is less than 3.0 and used highly selectively in patients with higher INRs.
References [1] Stanworth SJ, Brunskill SJ, Hyde CJ, et al. Is fresh frozen plasma clinically effective? A systematic review of randomized controlled trials. Br J Haematol 2004;126:139-52. [2] Foster PF, Moore LR, Sankary HN, et al. Central venous catheterization in patients with coagulopathy. Arch Surg 1992;127:273-5. [3] Doerfler ME, Kaufman B, Goldenberg AS. Central venous catheter placement in patients with disorders of hemostasis. Chest 1996;110: 185-8. [4] Fisher NC, Mutimer DJ. Central venous cannulation in patients with liver disease and coagulopathy—a prospective audit. Intensive Care Med 1999;25:481-5. [5] DeLoughery TG, Liebler JM, Simonds V, et al. Invasive line placement in critically ill patients: do hemostatic defects matter? Transfusion 1996;36:827-31. [6] Malloy PC, Grassi CJ, Kundu S, et al. Consensus guidelines for periprocedural management of coagulation status and hemostasis risk in percutaneous image-guided interventions. J Vasc Interv Radiol 2009;20:S240-9.
529.e13 [7] Gualtieri E, Deppe SA, Sipperly ME, et al. Subclavian venous catheterization: greater success rate for less experienced operators using ultrasound guidance. Crit Care Med 1995;23:692-7. [8] Tercan F, Ozkan U, Oguzkurt L. US-guided placement of central vein catheters in patients with disorders of hemostasis. Eur J Radiol 2008;65:253-6. [9] Dara SI, Rana R, Afessa B, et al. Fresh frozen plasma transfusion in critically ill medical patients with coagulopathy. Crit Care Med 2005;33:2667-71. [10] Napolitano LM, Kurek S, Luchette FA, et al. Clinical practice guideline: red blood cell transfusion in adult trauma and critical care. Crit Care Med 2009;37:3124-57. [11] Silliman CC, Fung YL, Ball JB, et al. Transfusion-related acute lung injury (TRALI): current concepts and misconceptions. Blood Rev 2009;23:245-55. [12] Segal JB, Dzik WH, et al. Paucity of studies to support that abnormal coagulation test results predict bleeding in the setting of invasive procedures: an evidence-based review. Transfusion 2005; 45:1413-25. [13] Holland LL, Brooks JP. Toward rational fresh frozen plasma transfusion: the effect of plasma transfusion on coagulation test results. Am J Clin Pathol 2006;126:133-9. [14] Sweeney JD, Kouttab NM, Holme SH, et al. Prestorage pooled whole blood derived leukoreduced platelets stored for seven days preserve acceptable quality and do not show evidence of a mixed lymphocyte reaction. Transfusion 2004;44:1212-9. [15] Miller AH, Roth BA, Mills TJ, et al. Ultrasound guidance versus the landmark technique for the placement of central venous catheters in the emergency department. Acad Emerg Med 2002;9:800-5. [16] Wigmore TJ, Symthe JF, Hacking MB, et al. Effect of the implementation of NICE guidelines for ultrasound guidance on the complication rates associated with central venous catheter placement in patients presenting for routine surgery in a tertiary referral centre. Br J Anaesth 2007;99:662-5. [17] Balls A, LoVecchio F, Kroeger A, et al. Ultrasound guidance for central venous catheter placement: results from the Central Line Emergency Access Registry Database. Am J Emerg Med 2010;28: 561-7. [18] Vlaar AP, In Der Maur AL, Binnekade JM, et al. A survey of physicians' reasons to transfuse plasma and platelets in the critically ill: a prospective single-centre cohort study. Transfus Med 2009;19: 207-12. [19] Stanworth SJ, Walsh TS. Fresh frozen plasma: friend or faux pas in critical illness? Crit Care Med 2005;33:2714-6. [20] Narick C, Triulzi DJ, Yazer MH. Transfusion associated circulatory overload after plasma transfusion. Transfusion 2011 [Published online ISSN: 1537-2995].