- Email: [email protected]
Prehospital Transfusion for Gastrointestinal Bleeding
Air Medical Journal xxx (2017) 1e5
Contents lists available at ScienceDirect
Air Medical Journal journal homepage: http://www.airmedicaljournal.com/
Original Research
Prehospital Transfusion for Gastrointestinal Bleeding Maile E. Parker, MD 1, Mohammad A. Khasawneh, MBBS 1, Cornelius A. Thiels, DO 1, Kathleen S. Berns, APRN, CNS, MS 2, James R. Stubbs, MD 3, Donald H. Jenkins, MD 4, Scott P. Zietlow, MD 5, Martin D. Zielinski, MD 5, * 1
Department of Surgery, Mayo Clinic Rochester, Rochester, MN Department of Nursing, Mayo Clinic Rochester, Rochester, MN 3 Department of Laboratory Medicine and Pathology, Mayo Clinic Rochester, Rochester, MN 4 Division of Trauma and Emergency Surgery, The University of Texas Health Science Center at San Antonio, San Antonio, TX 5 Division of Trauma, Critical Care, and General Surgery, Mayo Clinic Rochester, Rochester, MN 2
a b s t r a c t Objective: Gastrointestinal (GI) bleeding is a common medical emergency with significant morbidity and mortality. Many patients are coagulopathic, which may perpetuate bleeding. Remote damage control resuscitation, including early correction of coagulopathy and anemia, may benefit exsanguinating patients with GI bleeding. Methods: We conducted a retrospective review of patients with acute GI bleeding who received packed red blood cells (pRBC) and/or plasma during transportation to our institution between 2010 and 2014. A comparison group of patients who were not transfused en route was selected, and demographics, outcomes, and response to resuscitation were compared. Results: A total of 112 patients with GI bleeding received pRBC (82%, n ¼ 92 pRBC, mean 1.7 ± 0.9 units), plasma (62%, n ¼ 69, mean 1.7 ± 0.8 units) or both (44%, n ¼ 49) en-route. The comparison group comprised 49 patients transported by helicopter who were not transfused en-route. Demographics, crystalloid resuscitation, transfusion prior to transfer, rate of intervention, ICU days, length of stay, and mortality were similar between groups. Patients transfused en route had a significant increase in hemoglobin from 8.3 ± 2.2 to 8.9 ± 2.1 (P ¼ .03) and decrease in INR from 2.0 ± 1.0 to 1.6 ± 1.4 (P ¼ .01), whereas those not transfused en route experienced stable hemoglobin (8.7 ± 2.8 to 9.4 ± 2.5; P ¼ .21) and INR values (1.9 ± 1.0 to 1.6 ± 1.4; P ¼ .32). Both groups had a significant improvement in hemodynamic parameters with resuscitation. Conclusion: Prehospital damage control resuscitation with pRBC and/or plasma resulted in the improvement of hemodynamic instability, coagulopathy and anemia in patients with acute GI bleeding. Almost all patients required additional inpatient interventions and/or transfusions, suggesting that prehospital transfusion is being utilized for appropriately selected patients. Copyright © 2017 by Air Medical Journal Associates
Patients presenting with acute gastrointestinal (GI) hemorrhage experience significant morbidity and mortality.1-3 Notably, many patients with GI bleeding are coagulopathic at presentation, and those with coagulopathy have been shown to have an increased risk of mortality and rebleeding.4-6 Early hemostatic resuscitation with blood products, including plasma, is instrumental in correcting acidosis and coagulopathy and decreases mortality in hemorrhaging trauma patients.7,8 The importance of a balanced transfusion * Address for correspondence: Martin D. Zielinski, MD, Division of Trauma, Critical Care, and General Surgery, Mayo Clinic Rochester, Rochester, MN. E-mail address: [email protected] (M.D. Zielinski). 1067-991X/$36.00 Copyright © 2017 by Air Medical Journal Associates http://dx.doi.org/10.1016/j.amj.2017.06.002
protocol (approximately equal parts packed red blood cells [PRBCs], plasma, and platelets) in optimizing outcomes for massively transfused trauma patients has been shown.9-13 Initiating transfusion in the prehospital setting has been shown to improve selected outcomes in trauma patients suffering from hemorrhagic shock.14,15 Data on the use of blood products for medical bleeding in the prehospital setting are limited, and we are not aware of any institutional protocols specific for this purpose. Current institutional prehospital transfusion protocols are largely based on military practice and were therefore designed for trauma patients. Whether these same protocols are applicable to nontrauma hemorrhaging patients is unknown. Nontrauma patients account for the majority of prehospital
2
M.E. Parker et al. / Air Medical Journal xxx (2017) 1e5
Blood products should be administered if an adult paent has 2 or more of the following aer traumac injury or evidence of bleeding: 1. 2. 3. 4. 5. 6.
Hypotension (single reading of systolic blood pressure ≤90 mm Hg) Tachycardia (single reading of heart rate ≥120) Penetrang mechanism of injury Point of care lactate ≥ 5 mg/dL Base deficit ≥ -5 mmol/L StO2 ≤ 65%
Blood products should be administered for paents with medical (non-traumac) intracranial bleeding and are on Coumadin: 1. 2.
INR ≥1.5 or unknown: Iniate plasma transfusion INR ≤1.5: no plasma transfusion
Using clinical judgment, consider blood products for known large volume blood loss and/or decreasing hemoglobin/hematocrit.
-Blood product transfusion should begin with plasma followed by packed red blood cells (PRBC) based on the paent’s clinical status and hemodynamics. PRBC and plasma can be given concurrently. -One gram of calcium chloride should be given if the paent has received 6 units of blood product. -Transfusion should connue unl the paent’s condion stabilizes or unl available blood products are depleted. Crystalloids may be given at this point. -If a paent on Coumadin is bleeding but hemoglobin/hematocrit is stable, plasma can be transfused and hemoglobin monitored. Figure 1. Current institutional prehospital blood transfusion guidelines.
transfusions by our institution's critical care transport team and differ in many ways from the population of trauma patients we serve. We have previously described the use of transfusion for medical bleeding by our air medical transport teams, including a subset of patients with GI bleeding, and we sought to further examine treatment and outcomes in this group.16 We have previously shown improvement in severe anemia after prehospital transfusion of PRBCs in patients transported for hemorrhage, including a subset of GI bleed patients.17 Although initial civilian remote damage control resuscitation was limited to PRBC transfusion, we have incorporated thawed plasma as part of the prehospital critical care resuscitation protocol in order to treat coagulopathy as early as possible. As a result of this evolution in our protocol, we also aimed to determine whether plasma in combination with PRBCs is beneficial to patients with exsanguinating GI bleeding, hypothesizing that hemodynamics, coagulopathy, and anemia will improve. Materials and Methods After obtaining approval from the institutional review board, we conducted a retrospective review of our institution's prospectively maintained transport database. Our rural tertiary care center is the only large tertiary hospital within 60 miles of 30 referring centers. Patients transported to our institution by critical care air or ground transport for the management of presumed acute GI bleeding between January 2010 and February 2014 were identified. GI bleeding was defined as hematochezia, melena, or hematemesis either by patient report or witnessed by referring providers or prehospital providers. Our institution began a program for prehospital PRBC transfusion in 1988 and incorporated prehospital plasma transfusion in 2009.17,18 Our air medical transport teams currently carry 3 units of O negative PRBCs and 3 units of A positive thawed plasma. These teams sometimes are required to travel by ground rather than air because of inclement weather. Transfusion is initiated according to our institutional critical
care transport guideline (Fig. 1). Hemodynamic instability is defined in the protocol as systolic blood pressure 90 mm Hg and/or heart rate 120 beats/min. Our institution's transport teams document an indication for transfusion in the transport records and provide a description of the patient's clinical presentation. A performance improvement (PI) review is performed on every single transport by the clinical nurse specialist and medical director. Any discrepancies are discussed at monthly flight team meetings, and either protocol changes or education is completed. Flight teams also have the capability to check point-of-care laboratory values using the Abbott i-STAT 1 (Abbott Laboratories, Abbott Park, IL). Additional data were collected from the electronic medical record and transport database. We compared demographics, transfusion volumes, laboratory values, hospital course, and outcomes for adult patients who received transfusions of PRBCs and/or plasma during transportation by our flight crews with those who were transported by visiting helicopters (which do not carry PRBCs or plasma) during the same time period. Patients transported to other institutions or by basic life support crews were excluded as were prisoners and pregnant patients. Laboratory values before and after transfer (within 6 hours of arrival) were compared. The bicarbonate level was used as a surrogate measure for acidosis because only a handful of patients had an arterial or venous blood gas obtained before or during transfer. Anticoagulation was considered to include warfarin, clopidogrel, therapeutic dosing of lowemolecular-weight heparin, and rivaroxaban but excluded aspirin. Normally distributed data are reported as the mean with standard deviation and compared using the Student 2-tailed t test. Non-normally distributed data are reported as the median with interquartile range (IQR) and compared using the MannWhitney U test. Categoric variables were compared using the Fisher exact text. JMP version 10.0 (SAS Institute, Inc., Cary, NC) was used for statistical analysis. A P value < .05 was considered significant.
M.E. Parker et al. / Air Medical Journal xxx (2017) 1e5 Table 1 Demographics, Resuscitation, Hospital Course, and Clinical Outcomes
Patient characteristics Male (%) Mean age (years) Chronic anticoagulation (%) Hemodynamic instability (%) Resuscitation Crystalloid before transfer (L) PRBCs before transfer (%) Mean units PRBCs Plasma before transfer (%) Mean units plasma Vitamin K before transfer (%) Mean crystalloid en route (L) Transfusions en route PRBCs (%) Mean units PRBCs Plasma (%) Mean units plasma PRBCs and plasma (%) Hospital course Transfusion after admission (%) PRBCs (%) Median units Plasma (%) Median units Received vitamin K (%) ICU admission (%) Operation (%) Interventional radiology procedure (%) Endoscopy (%) EGD (%) Colonoscopy (%) Endoscopic hemostasis (%) Median time to endoscopy (hours) Upper GI bleed (%) Lower GI bleed (%) Outcomes Median LOS (days) Median ICU stay (days) Median ventilator days Mortality, 30 day (%) Rebleeding, 30 day (%)
Transfusion En Route (n ¼ 112)
No Transfusion En Route (n ¼ 49)
P Value
58 67 ± 15 28 49
59 64 ± 15 22 47
1.0 .23 .56 .86
1.6 61 2.8 14 2.9 14 1.6
± 1.0
± 1.2
± 1.3
.2 .6 .8 .1 .3 .44 .1
89 86 4.0 (2.0-6.0) 54 2.0 (2.0-4.0) 8 91 12 16
86 82 3.0 (2.0-6.0) 35 2.0 (1.0-4.5) 10 92 6 12
.60 .49 .84 .03 .43 .76 1.00 .39 .64
87 78 23 58 7.0 (3.0-14.1)
88 76 18 61 9.4 (3.8-16.7)
1.00 .84 .54 .73 .68
63 21
71 20
.29 1.00
5.0 (4.0-8.0) 2.0 (1.0-2.0) 1.0 (1.0-2.0) 13 11
6.0 (4.0-8.0) 2.0 (1.0-3.0) 1.5 (1.0-5.0) 12 20
.52 .69 .65 1.00 .15
± 3.0 ± 2.5 ± 1.0
2.0 65 3.0 32 2.1 8 1.2
± 2.4 ± 0.8
80 1.6 ± 1.0 61 1.7 ± 0.8 40
EGD ¼ esophagogastroduodenoscopy; GI ¼ gastrointestinal; ICU ¼ intensive care unit; LOS ¼ length of stay; PRBCs ¼ packed red blood cells.
Results Demographics and Initial Resuscitation for Patients Receiving Prehospital Transfusion The transfusion of PRBCs and/or plasma was administered to 112 patients with acute GI bleeding during either air medical (n ¼ 92, 82%) or ground critical care transport (n ¼ 20, 18%) from referring institutions, with a mean transport time of 37 ± 26 minutes (Table 1). Fifty-eight percent of patients were men, with a mean age of 68 ±15 years. Twenty-eight percent were chronically anticoagulated with warfarin, clopidogrel, therapeutically dosed lowemolecular-weight heparin, or rivaroxaban. The indications for transfusion documented by critical care transport staff varied and were often multiple (Fig. 2). The PI process did not identify any case of inappropriate transfusion. Patients received a mean of 1.6 (± 1.0) L crystalloid en route. Eleven percent (n ¼ 12) had point-of-care hemoglobin drawn during transport, with a mean hemoglobin of 7.6 (± 1.2). Fortynine percent (n ¼ 55) of patients were hemodynamically unstable during transport as determined by the institutional guideline.
3
Initiation of the first unit of transfusion occurred during transport for 17% (n ¼ 19) of patients, whereas the remaining 83% had transfusion initiated at the referring institution and had additional units given during transport. Sixty-one percent (n ¼ 68) received PRBCs before transfer with a mean of 2.8 (± 3.0) units given. Fourteen percent (n ¼ 16) received plasma at the referring institution with a mean of 2.9 (± 2.5) units transfused. A mean of 1.6 (± 1.0) L crystalloid was given during initial resuscitation at the referring hospitals. Three patients received tranexamic acid, 2 during transport and 1 at the referring hospital before transfer. Two patients received cryoprecipitate before transfer. In-hospital Interventions and Outcomes for Patients Receiving Prehospital Transfusion Eighty-six percent (n ¼ 96) of patients transfused en route went on to receive additional PRBCs, with a median of 4.0 units (IQR, 2.0-6.0) transfused during their hospital stay. Fifty-four percent (n ¼ 61) went on to receive additional plasma, with a median of 2.0 units (IQR, 2.0-5.0) transfused during their hospital stay. Eleven percent (n ¼ 12) of patients received no further PRBCs or plasma transfusion after admission. Ninety-one percent (n ¼ 102) of patients were admitted to the intensive care unit (ICU) with a median ICU length of stay of 2 days. Eighty-seven percent (n ¼ 97) underwent endoscopy (78% esophagogastroduodenoscopy and 23% colonoscopy) with 58% requiring hemostatic intervention. Therapeutic endoscopic interventions for hemostasis included clipping, thermocoagulation, epinephrine injection, or insertion of a Minnesota tube. The source of hemorrhage was upper GI bleeding proximal to the ligament of Treitz in 63% (n ¼ 70), lower GI bleeding in 20% (n ¼ 22), and indeterminate source of bleeding in 18% (n ¼ 20). Radiologic interventions were performed in 16% (n ¼ 18) of patients and included angiography with or without embolization, transjugular intrahepatic portosystemic shunt or transjugular intrahepatic portosystemic shunt revision, iliac artery stenting in a GI bleed associated with a pancreas transplant, and percutaneous sclerotherapy for bleeding peristomal abdominal wall varices. Twelve percent (n ¼ 13) of patients underwent operative exploration for their GI bleed. Operative procedures included segmental colectomy, small bowel resection, hemorrhoidectomy, antrectomy, repair of a gastric perforation, repair of aortoenteric fistula, and other abdominal explorations including 1 inoperable duodenal lymphoma, 1 inoperable superior mesenteric artery to duodenal fistula, and abdominal exploration with enteroscopy. Mortality at 30 days was 13%. The rate of rebleeding requiring re-endoscopy or readmission to the hospital within 30 days of the index admission was 11%. One case of presumed transfusionassociated acute lung injury occurred. There were no other documented transfusion reactions of any type associated with prehospital transfusion. Comparison Group We identified 49 patients with GI bleeding transported by visiting helicopters and therefore not transfused en route. Demographics, rate of outpatient anticoagulation, hemodynamic instability, and pretransport laboratory values were similar between groups. Crystalloid usage during resuscitation (before transfer and en route) and the amounts of PRBCs and plasma transfused before transfer were not significantly different between groups. Eighty-two percent (n ¼ 40) of the comparison group received PRBCs after admission, with a median of 3 units transfused, which is similar to the patients who received prehospital transfusion. Fewer of the comparison group patients (35%, n ¼ 17) received additional plasma after admission to the hospital (P ¼ .03), but the amount transfused was similar with a median of 2.0 units transfused after admission. Fourteen percent (n ¼ 7) did not receive any
4
M.E. Parker et al. / Air Medical Journal xxx (2017) 1e5
Figure 2. Indications for prehospital transfusion.
Table 2 A Comparison of the Laboratory and Hemodynamic Response to Resuscitation for Patients Who Received Prehospital Transfusion Versus Patients Who Did Not
Transfusion en route Hb (g/dL) INR HCO3 (mmol/L) Hemodynamic instability (%) No transfusion en route Hb (g/dL) INR HCO3 (mmol/L) Hemodynamic instability (%)
Pretransport
Post-transport
P Value
8.25 ± 2.23 2.01 ± 1.51 23.20 ± 5.14 49
8.90 ± 2.11 1.56 ± 0.83 22.41 ± 4.35 18
.03 .01 .27 <.001
8.65 ± 2.79 1.94 ± 0.97 23.67 ± 6.89 47
9.35 ± 2.53 1.62 ± 1.37 21.90 ± 4.15 18
.21 .32 .29 .005
Hb ¼ hemoglobin; HCO3 ¼ bicarbonate; INR ¼ international normalized ratio.
additional transfusion. The rate of procedural intervention, mortality, ICU days, ventilator days, length of stay, and rebleeding were similar between groups. Response to Prehospital Transfusion Forty-nine percent (n ¼ 55) of patients transfused en route were hemodynamically unstable during transport, improving to 18% (n ¼ 20) at hospital admission (Table 2). Ninety-seven percent (n ¼ 109) had laboratory values available before transfer, and 62% (n ¼ 69) had the international normalized ratio (INR) drawn before transfer. Hemoglobin significantly increased from a mean of 8.3 mg/dL before transport to 8.9 mg/dL (P ¼ .03) on hospital arrival. Patients transfused en route showed a significant decrease in INR from a mean of 2.01 before transport to 1.56 on arrival to the hospital (P ¼.01). Patients were not significantly acidotic before transfer and did not show a significant change in the bicarbonate level at hospital admission. Like the patients who received prehospital transfusion, the comparison group of patients who did not receive blood products en route showed a significant improvement in hemodynamic instability on admission to our institution after transport. Eightyeight percent (n ¼ 43) of the comparison group had laboratory values available before transfer, and only 37% (n ¼ 18) had INR drawn before transfer. Patients in the comparison group experienced an increase in hemoglobin from 8.7 to 9.3, which did not reach statistical significance. This group also showed a decrease in
INR from 1.94 to 1.62, which did not reach statistical significance. Like the group of patients who were transfused en route, the comparison patients showed no difference in the bicarbonate level before and after resuscitation and were not significantly acidotic before transfer. Discussion Patients who received prehospital PRBCs and plasma transfusion during transport for GI bleeding were critically ill, coagulopathic, and severely anemic and had a high rate of procedural intervention for hemostasis. Many showed evidence of hemodynamic instability during transport. All but 11% required additional transfusion after admission to the hospital. We found that our patients whose continuum of resuscitation included transfusion of PRBCs or plasma by critical care transport teams en route to definitive care showed a statistically significant correction of coagulopathy and anemia, which was not true for the comparison group. Notably, we did not detect a difference in any clinical outcome between groups. Patients who received transfusion en route were significantly more likely to be transfused with additional plasma after admission to our institution, which could reflect a more severe coagulation dysfunction not detected by INR or could simply be a product of a small sample size. Although it is difficult to ascertain a definite clinical benefit from the current study, our data suggest that patients are being appropriately selected for prehospital transfusion, and these interventions appear to be having the desired effect on the early correction of anemia and coagulopathy. The examination of pretransport and admission laboratory values showed a significant decrease in INR and an increase in hemoglobin after prehospital transfusion of plasma and PRBCs, respectively. Patients were anemic before transport; however, it should also be noted that the hemoglobin level likely does not accurately reflect the degree of blood loss in hypovolemic acutely bleeding patients and should be interpreted cautiously in combination with clinical variables.19 Hemoglobin level is not a strict component of our transfusion protocol but may be used along with hemodynamic status and clinical judgment in order to determine the need for transfusion. It is known that coagulopathy in upper GI bleed patients is associated with increased mortality and rebleeding4-6; therefore, we would predict that the early correction of
M.E. Parker et al. / Air Medical Journal xxx (2017) 1e5
coagulopathy may be of clinical benefit to our upper GI bleed patients. It remains to be determined whether these laboratory changes translate to a clinically significant difference in outcomes such as rebleeding and mortality. It is worth noting that in the general population of hemodynamically stable acute upper GI bleed patients, minimizing the transfusion of PRBCs is associated with improved outcomes.19,20 A randomized controlled trial found that a lower threshold of 7 g/dL for the transfusion of PRBCs compared with 9 g/dL was associated with decreased mortality and a decreased rebleeding rate in certain upper GI bleed patients, especially those with portal hypertension and variceal bleeding.21 Notably, those authors excluded patients with massive exsanguinating bleeding and hemodynamic instability, so it is challenging to directly apply these data to our patient population, which included many hemodynamically unstable patients. To our knowledge, similar studies of early transfusion or transfusion threshold have not been undertaken specifically for patients with lower GI bleeding, and it remains to be seen how, and to what degree, coagulopathy clinically affects the lower GI bleed population. The existing data highlight the importance of being judicious with the administration of blood products and avoiding copious or unnecessary PRBC transfusion in GI bleeding. We identified a subgroup of patients transfused en route who did not require further transfusion or intervention after admission, which could indicate the unnecessary use of prehospital transfusion. On review of this particular subset of patients, several clinical presentations with hypotension secondary to severe sepsis or dehydration appear to have been initially interpreted as hemorrhagic shock from GI bleeding, leading to transfusion. Several other patients had appropriate endoscopic hemostasis for ongoing GI bleeding immediately on arrival to our institution and required no further transfusion after hemostasis was achieved. Two patients refused intervention, including transfusion after transfer to our institution. One patient had undergone an operation before transfer and received transfusion en route with blood products that had been ordered by the referring institution based on clinical parameters before transfer. Finally, 1 patient had a diverticular bleed that stopped spontaneously. Given the overall clinical picture and limited information available to the prehospital critical care transport team, these transfusions were deemed warranted through the PI process previously mentioned. The appropriate use of a prehospital transfusion protocol depends in part on the clinical acumen of critical care transport personnel to rapidly determine which patients are exsanguinating and avoid unnecessary transfusion for those at lower risk. Continual analysis and re-evaluation of prehospital transfusion practices are essential to ensure appropriate use. Limitations This retrospective assessment of a single tertiary referral center experience serves to demonstrate the appropriate use of our prehospital transfusion protocol for GI bleed patients. Our data may not be generalizable to centers with different referral populations. In addition, we were not able to obtain outside records from all transferring hospitals, and, therefore, missing pretransport laboratory values are a potential source of bias in our study. Lastly, some of the INR values recorded before transport were point-of-care measurements, whereas values after arrival at our hospital were automated plasma INR. Although the majority of the literature suggests that these methods correlate well, some studies have suggested variability between these tests.22,23
5
Conclusions Our data suggest that our institution's prehospital transfusion protocol is being used appropriately and confers early improvement in hemodynamic parameters, anemia, and coagulopathy in patients with acute GI hemorrhage. Further investigation, including prospective analysis to assess whether GI bleed patients experience clinically significant improvement in outcomes with the current protocol, is warranted and whether these protocols should be altered for nontrauma patients.
References 1. Loperfido S, Baldo V, Piovesana E, et al. Changing trends in acute upper-GI bleeding: a population-based study. Gastrointest Endosc. 2009;70:212e224. 2. Velayos FS, Williamson A, Sousa KH, et al. Early predictors of severe lower gastrointestinal bleeding and adverse outcomes: a prospective study. Clin Gastroenterol Hepatol. 2004;2:485e490. 3. Abougergi MS, Travis AC, Saltzman JR. The in-hospital mortality rate for upper GI hemorrhage has decreased over 2 decades in the United States: a nationwide analysis. Gastrointest Endosc. 2015;81:882e888. 4. Jairath V, Kahan BC, Stanworth JS, et al. Prevalence, management, and outcomes of patients with coagulopathy after acute nonvariceal upper gastrointestinal bleeding in the United Kingdom. Transfusion. 2013;53:1069e1076. 5. Loffroy R, Guiu B, D'Athis P, et al. Arterial embolotherapy for endoscopically unmanageable acute gastroduodenal hemorrhage: predictors of early rebleeding. Clin Gastroenterol Hepatol. 2009;7:515e523. 6. Restellini S, Kherad O, Jairath V, Martel M, Barkun AN. Red blood cell transfusion is associated with increased rebleeding in patients with nonvariceal upper gastrointestinal bleeding. Aliment Pharmacol Ther. 2013;37: 316e322. 7. Gonzalez EA, Moore FA, Holcomb JB, et al. Fresh frozen plasma should be given earlier to patients requiring massive transfusion. J Trauma. 2007;62:112e119. 8. Mitra B, Mori A, Cameron PA, Fitzgerald M, Paul E, Street A. Fresh frozen plasma (FFP) use during massive blood transfusion in trauma resuscitation. Injury. 2010;41:35e39. 9. Riskin DJ, Tsai TC, Riskin L, et al. Massive transfusion protocols: the role of aggressive resuscitation versus product ratio in mortality reduction. J Am Coll Surg. 2009;209:198e205. 10. Bhangu A, Nepogodiev D, Doughty H, Bowley DM. Meta-analysis of plasma to red blood cell ratios and mortality in massive blood transfusions for trauma. Injury. 2013;44:1693e1699. 11. del Junco DJ, Holcomb JB, Fox EE, et al, PROMMTT Study Group. Resuscitate early with plasma and platelets or balance blood products gradually: findings from the PROMMTT study. J Trauma Acute Care Surg. 2013;75:S24eS30. 12. Novak DJ, Bai Y, Cooke RK, et al, PROPPR Study Group. Making thawed universal donor plasma available rapidly for massively bleeding trauma patients: experience from the Pragmatic, Randomized Optimal Platelets and Plasma Ratios (PROPPR) trial. Transfusion. 2015;55:1331e1339. 13. Kutcher ME, Kornblith LZ, Narayan R, et al. A paradigm shift in trauma resuscitation: evaluation of evolving massive transfusion practices. JAMA Surg. 2013;148:834e840. 14. Brown JB, Sperry JL, Fombona A, Billiar TR, Peitzman AB, Guyette FX. Pretrauma center red blood cell transfusion is associated with improved early outcomes in air medical trauma patients. J Am Coll Surg. 2015;220:797e808. 15. Kim BD, Zielinski MD, Jenkins DH, Schiller HJ, Berns KS, Zietlow SP. The effects of prehospital plasma on patients with injury. J Trauma Acute Care Surg. 2012;73:S49eS53. 16. Thiels CA, Aho JM, Fahy AS, et al. Prehospital blood transfusions in non-trauma patients. World J Surg. 2016;40:2297e2304. 17. Berns KS, Zietlow SP. Blood usage in rotor-wing transport. Air Med J. 1998;17: 105e108. 18. Jenkins D, Stubbs J, Williams S, et al. Implementation and execution of civilian remote damage control resuscitation programs. Shock. 2014;41:84e89. 19. Valeri CR, Dennis RC, Ragno G, Macgregor H, Menzoian JO, Khuri SF. Limitations of the hematocrit level to assess the need for red blood cell transfusion in hypovolemic anemic patients. Transfusion. 2006;46:365e371. 20. Villanueva C, Colomo A, Bosch A, et al. Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med. 2013;368:11e21. 21. Wang J, Bao YX, Bai M, Zhang YG, Xu WD, Qi XS. Restrictive vs liberal transfusion for upper gastrointestinal bleeding: a meta-analysis of randomized controlled trials. World J Gastroenterol. 2013;19:6919e6927. 22. Goodman MD, Makley AT, Hanseman DJ, Pritts TA, Robinson BR. All the bang without the bucks: defining essential point-of-care testing for traumatic coagulopathy. J Trauma Acute Care Surg. 2015;79:117e124. 23. van den Besselaar AM. Accuracy, precision, and quality control for point-ofcare testing of oral anticoagulation. J Thromb Thrombolysis. 2001;12:35e40.
Contents lists available at ScienceDirect
Air Medical Journal journal homepage: http://www.airmedicaljournal.com/
Original Research
Prehospital Transfusion for Gastrointestinal Bleeding Maile E. Parker, MD 1, Mohammad A. Khasawneh, MBBS 1, Cornelius A. Thiels, DO 1, Kathleen S. Berns, APRN, CNS, MS 2, James R. Stubbs, MD 3, Donald H. Jenkins, MD 4, Scott P. Zietlow, MD 5, Martin D. Zielinski, MD 5, * 1
Department of Surgery, Mayo Clinic Rochester, Rochester, MN Department of Nursing, Mayo Clinic Rochester, Rochester, MN 3 Department of Laboratory Medicine and Pathology, Mayo Clinic Rochester, Rochester, MN 4 Division of Trauma and Emergency Surgery, The University of Texas Health Science Center at San Antonio, San Antonio, TX 5 Division of Trauma, Critical Care, and General Surgery, Mayo Clinic Rochester, Rochester, MN 2
a b s t r a c t Objective: Gastrointestinal (GI) bleeding is a common medical emergency with significant morbidity and mortality. Many patients are coagulopathic, which may perpetuate bleeding. Remote damage control resuscitation, including early correction of coagulopathy and anemia, may benefit exsanguinating patients with GI bleeding. Methods: We conducted a retrospective review of patients with acute GI bleeding who received packed red blood cells (pRBC) and/or plasma during transportation to our institution between 2010 and 2014. A comparison group of patients who were not transfused en route was selected, and demographics, outcomes, and response to resuscitation were compared. Results: A total of 112 patients with GI bleeding received pRBC (82%, n ¼ 92 pRBC, mean 1.7 ± 0.9 units), plasma (62%, n ¼ 69, mean 1.7 ± 0.8 units) or both (44%, n ¼ 49) en-route. The comparison group comprised 49 patients transported by helicopter who were not transfused en-route. Demographics, crystalloid resuscitation, transfusion prior to transfer, rate of intervention, ICU days, length of stay, and mortality were similar between groups. Patients transfused en route had a significant increase in hemoglobin from 8.3 ± 2.2 to 8.9 ± 2.1 (P ¼ .03) and decrease in INR from 2.0 ± 1.0 to 1.6 ± 1.4 (P ¼ .01), whereas those not transfused en route experienced stable hemoglobin (8.7 ± 2.8 to 9.4 ± 2.5; P ¼ .21) and INR values (1.9 ± 1.0 to 1.6 ± 1.4; P ¼ .32). Both groups had a significant improvement in hemodynamic parameters with resuscitation. Conclusion: Prehospital damage control resuscitation with pRBC and/or plasma resulted in the improvement of hemodynamic instability, coagulopathy and anemia in patients with acute GI bleeding. Almost all patients required additional inpatient interventions and/or transfusions, suggesting that prehospital transfusion is being utilized for appropriately selected patients. Copyright © 2017 by Air Medical Journal Associates
Patients presenting with acute gastrointestinal (GI) hemorrhage experience significant morbidity and mortality.1-3 Notably, many patients with GI bleeding are coagulopathic at presentation, and those with coagulopathy have been shown to have an increased risk of mortality and rebleeding.4-6 Early hemostatic resuscitation with blood products, including plasma, is instrumental in correcting acidosis and coagulopathy and decreases mortality in hemorrhaging trauma patients.7,8 The importance of a balanced transfusion * Address for correspondence: Martin D. Zielinski, MD, Division of Trauma, Critical Care, and General Surgery, Mayo Clinic Rochester, Rochester, MN. E-mail address: [email protected] (M.D. Zielinski). 1067-991X/$36.00 Copyright © 2017 by Air Medical Journal Associates http://dx.doi.org/10.1016/j.amj.2017.06.002
protocol (approximately equal parts packed red blood cells [PRBCs], plasma, and platelets) in optimizing outcomes for massively transfused trauma patients has been shown.9-13 Initiating transfusion in the prehospital setting has been shown to improve selected outcomes in trauma patients suffering from hemorrhagic shock.14,15 Data on the use of blood products for medical bleeding in the prehospital setting are limited, and we are not aware of any institutional protocols specific for this purpose. Current institutional prehospital transfusion protocols are largely based on military practice and were therefore designed for trauma patients. Whether these same protocols are applicable to nontrauma hemorrhaging patients is unknown. Nontrauma patients account for the majority of prehospital
2
M.E. Parker et al. / Air Medical Journal xxx (2017) 1e5
Blood products should be administered if an adult paent has 2 or more of the following aer traumac injury or evidence of bleeding: 1. 2. 3. 4. 5. 6.
Hypotension (single reading of systolic blood pressure ≤90 mm Hg) Tachycardia (single reading of heart rate ≥120) Penetrang mechanism of injury Point of care lactate ≥ 5 mg/dL Base deficit ≥ -5 mmol/L StO2 ≤ 65%
Blood products should be administered for paents with medical (non-traumac) intracranial bleeding and are on Coumadin: 1. 2.
INR ≥1.5 or unknown: Iniate plasma transfusion INR ≤1.5: no plasma transfusion
Using clinical judgment, consider blood products for known large volume blood loss and/or decreasing hemoglobin/hematocrit.
-Blood product transfusion should begin with plasma followed by packed red blood cells (PRBC) based on the paent’s clinical status and hemodynamics. PRBC and plasma can be given concurrently. -One gram of calcium chloride should be given if the paent has received 6 units of blood product. -Transfusion should connue unl the paent’s condion stabilizes or unl available blood products are depleted. Crystalloids may be given at this point. -If a paent on Coumadin is bleeding but hemoglobin/hematocrit is stable, plasma can be transfused and hemoglobin monitored. Figure 1. Current institutional prehospital blood transfusion guidelines.
transfusions by our institution's critical care transport team and differ in many ways from the population of trauma patients we serve. We have previously described the use of transfusion for medical bleeding by our air medical transport teams, including a subset of patients with GI bleeding, and we sought to further examine treatment and outcomes in this group.16 We have previously shown improvement in severe anemia after prehospital transfusion of PRBCs in patients transported for hemorrhage, including a subset of GI bleed patients.17 Although initial civilian remote damage control resuscitation was limited to PRBC transfusion, we have incorporated thawed plasma as part of the prehospital critical care resuscitation protocol in order to treat coagulopathy as early as possible. As a result of this evolution in our protocol, we also aimed to determine whether plasma in combination with PRBCs is beneficial to patients with exsanguinating GI bleeding, hypothesizing that hemodynamics, coagulopathy, and anemia will improve. Materials and Methods After obtaining approval from the institutional review board, we conducted a retrospective review of our institution's prospectively maintained transport database. Our rural tertiary care center is the only large tertiary hospital within 60 miles of 30 referring centers. Patients transported to our institution by critical care air or ground transport for the management of presumed acute GI bleeding between January 2010 and February 2014 were identified. GI bleeding was defined as hematochezia, melena, or hematemesis either by patient report or witnessed by referring providers or prehospital providers. Our institution began a program for prehospital PRBC transfusion in 1988 and incorporated prehospital plasma transfusion in 2009.17,18 Our air medical transport teams currently carry 3 units of O negative PRBCs and 3 units of A positive thawed plasma. These teams sometimes are required to travel by ground rather than air because of inclement weather. Transfusion is initiated according to our institutional critical
care transport guideline (Fig. 1). Hemodynamic instability is defined in the protocol as systolic blood pressure 90 mm Hg and/or heart rate 120 beats/min. Our institution's transport teams document an indication for transfusion in the transport records and provide a description of the patient's clinical presentation. A performance improvement (PI) review is performed on every single transport by the clinical nurse specialist and medical director. Any discrepancies are discussed at monthly flight team meetings, and either protocol changes or education is completed. Flight teams also have the capability to check point-of-care laboratory values using the Abbott i-STAT 1 (Abbott Laboratories, Abbott Park, IL). Additional data were collected from the electronic medical record and transport database. We compared demographics, transfusion volumes, laboratory values, hospital course, and outcomes for adult patients who received transfusions of PRBCs and/or plasma during transportation by our flight crews with those who were transported by visiting helicopters (which do not carry PRBCs or plasma) during the same time period. Patients transported to other institutions or by basic life support crews were excluded as were prisoners and pregnant patients. Laboratory values before and after transfer (within 6 hours of arrival) were compared. The bicarbonate level was used as a surrogate measure for acidosis because only a handful of patients had an arterial or venous blood gas obtained before or during transfer. Anticoagulation was considered to include warfarin, clopidogrel, therapeutic dosing of lowemolecular-weight heparin, and rivaroxaban but excluded aspirin. Normally distributed data are reported as the mean with standard deviation and compared using the Student 2-tailed t test. Non-normally distributed data are reported as the median with interquartile range (IQR) and compared using the MannWhitney U test. Categoric variables were compared using the Fisher exact text. JMP version 10.0 (SAS Institute, Inc., Cary, NC) was used for statistical analysis. A P value < .05 was considered significant.
M.E. Parker et al. / Air Medical Journal xxx (2017) 1e5 Table 1 Demographics, Resuscitation, Hospital Course, and Clinical Outcomes
Patient characteristics Male (%) Mean age (years) Chronic anticoagulation (%) Hemodynamic instability (%) Resuscitation Crystalloid before transfer (L) PRBCs before transfer (%) Mean units PRBCs Plasma before transfer (%) Mean units plasma Vitamin K before transfer (%) Mean crystalloid en route (L) Transfusions en route PRBCs (%) Mean units PRBCs Plasma (%) Mean units plasma PRBCs and plasma (%) Hospital course Transfusion after admission (%) PRBCs (%) Median units Plasma (%) Median units Received vitamin K (%) ICU admission (%) Operation (%) Interventional radiology procedure (%) Endoscopy (%) EGD (%) Colonoscopy (%) Endoscopic hemostasis (%) Median time to endoscopy (hours) Upper GI bleed (%) Lower GI bleed (%) Outcomes Median LOS (days) Median ICU stay (days) Median ventilator days Mortality, 30 day (%) Rebleeding, 30 day (%)
Transfusion En Route (n ¼ 112)
No Transfusion En Route (n ¼ 49)
P Value
58 67 ± 15 28 49
59 64 ± 15 22 47
1.0 .23 .56 .86
1.6 61 2.8 14 2.9 14 1.6
± 1.0
± 1.2
± 1.3
.2 .6 .8 .1 .3 .44 .1
89 86 4.0 (2.0-6.0) 54 2.0 (2.0-4.0) 8 91 12 16
86 82 3.0 (2.0-6.0) 35 2.0 (1.0-4.5) 10 92 6 12
.60 .49 .84 .03 .43 .76 1.00 .39 .64
87 78 23 58 7.0 (3.0-14.1)
88 76 18 61 9.4 (3.8-16.7)
1.00 .84 .54 .73 .68
63 21
71 20
.29 1.00
5.0 (4.0-8.0) 2.0 (1.0-2.0) 1.0 (1.0-2.0) 13 11
6.0 (4.0-8.0) 2.0 (1.0-3.0) 1.5 (1.0-5.0) 12 20
.52 .69 .65 1.00 .15
± 3.0 ± 2.5 ± 1.0
2.0 65 3.0 32 2.1 8 1.2
± 2.4 ± 0.8
80 1.6 ± 1.0 61 1.7 ± 0.8 40
EGD ¼ esophagogastroduodenoscopy; GI ¼ gastrointestinal; ICU ¼ intensive care unit; LOS ¼ length of stay; PRBCs ¼ packed red blood cells.
Results Demographics and Initial Resuscitation for Patients Receiving Prehospital Transfusion The transfusion of PRBCs and/or plasma was administered to 112 patients with acute GI bleeding during either air medical (n ¼ 92, 82%) or ground critical care transport (n ¼ 20, 18%) from referring institutions, with a mean transport time of 37 ± 26 minutes (Table 1). Fifty-eight percent of patients were men, with a mean age of 68 ±15 years. Twenty-eight percent were chronically anticoagulated with warfarin, clopidogrel, therapeutically dosed lowemolecular-weight heparin, or rivaroxaban. The indications for transfusion documented by critical care transport staff varied and were often multiple (Fig. 2). The PI process did not identify any case of inappropriate transfusion. Patients received a mean of 1.6 (± 1.0) L crystalloid en route. Eleven percent (n ¼ 12) had point-of-care hemoglobin drawn during transport, with a mean hemoglobin of 7.6 (± 1.2). Fortynine percent (n ¼ 55) of patients were hemodynamically unstable during transport as determined by the institutional guideline.
3
Initiation of the first unit of transfusion occurred during transport for 17% (n ¼ 19) of patients, whereas the remaining 83% had transfusion initiated at the referring institution and had additional units given during transport. Sixty-one percent (n ¼ 68) received PRBCs before transfer with a mean of 2.8 (± 3.0) units given. Fourteen percent (n ¼ 16) received plasma at the referring institution with a mean of 2.9 (± 2.5) units transfused. A mean of 1.6 (± 1.0) L crystalloid was given during initial resuscitation at the referring hospitals. Three patients received tranexamic acid, 2 during transport and 1 at the referring hospital before transfer. Two patients received cryoprecipitate before transfer. In-hospital Interventions and Outcomes for Patients Receiving Prehospital Transfusion Eighty-six percent (n ¼ 96) of patients transfused en route went on to receive additional PRBCs, with a median of 4.0 units (IQR, 2.0-6.0) transfused during their hospital stay. Fifty-four percent (n ¼ 61) went on to receive additional plasma, with a median of 2.0 units (IQR, 2.0-5.0) transfused during their hospital stay. Eleven percent (n ¼ 12) of patients received no further PRBCs or plasma transfusion after admission. Ninety-one percent (n ¼ 102) of patients were admitted to the intensive care unit (ICU) with a median ICU length of stay of 2 days. Eighty-seven percent (n ¼ 97) underwent endoscopy (78% esophagogastroduodenoscopy and 23% colonoscopy) with 58% requiring hemostatic intervention. Therapeutic endoscopic interventions for hemostasis included clipping, thermocoagulation, epinephrine injection, or insertion of a Minnesota tube. The source of hemorrhage was upper GI bleeding proximal to the ligament of Treitz in 63% (n ¼ 70), lower GI bleeding in 20% (n ¼ 22), and indeterminate source of bleeding in 18% (n ¼ 20). Radiologic interventions were performed in 16% (n ¼ 18) of patients and included angiography with or without embolization, transjugular intrahepatic portosystemic shunt or transjugular intrahepatic portosystemic shunt revision, iliac artery stenting in a GI bleed associated with a pancreas transplant, and percutaneous sclerotherapy for bleeding peristomal abdominal wall varices. Twelve percent (n ¼ 13) of patients underwent operative exploration for their GI bleed. Operative procedures included segmental colectomy, small bowel resection, hemorrhoidectomy, antrectomy, repair of a gastric perforation, repair of aortoenteric fistula, and other abdominal explorations including 1 inoperable duodenal lymphoma, 1 inoperable superior mesenteric artery to duodenal fistula, and abdominal exploration with enteroscopy. Mortality at 30 days was 13%. The rate of rebleeding requiring re-endoscopy or readmission to the hospital within 30 days of the index admission was 11%. One case of presumed transfusionassociated acute lung injury occurred. There were no other documented transfusion reactions of any type associated with prehospital transfusion. Comparison Group We identified 49 patients with GI bleeding transported by visiting helicopters and therefore not transfused en route. Demographics, rate of outpatient anticoagulation, hemodynamic instability, and pretransport laboratory values were similar between groups. Crystalloid usage during resuscitation (before transfer and en route) and the amounts of PRBCs and plasma transfused before transfer were not significantly different between groups. Eighty-two percent (n ¼ 40) of the comparison group received PRBCs after admission, with a median of 3 units transfused, which is similar to the patients who received prehospital transfusion. Fewer of the comparison group patients (35%, n ¼ 17) received additional plasma after admission to the hospital (P ¼ .03), but the amount transfused was similar with a median of 2.0 units transfused after admission. Fourteen percent (n ¼ 7) did not receive any
4
M.E. Parker et al. / Air Medical Journal xxx (2017) 1e5
Figure 2. Indications for prehospital transfusion.
Table 2 A Comparison of the Laboratory and Hemodynamic Response to Resuscitation for Patients Who Received Prehospital Transfusion Versus Patients Who Did Not
Transfusion en route Hb (g/dL) INR HCO3 (mmol/L) Hemodynamic instability (%) No transfusion en route Hb (g/dL) INR HCO3 (mmol/L) Hemodynamic instability (%)
Pretransport
Post-transport
P Value
8.25 ± 2.23 2.01 ± 1.51 23.20 ± 5.14 49
8.90 ± 2.11 1.56 ± 0.83 22.41 ± 4.35 18
.03 .01 .27 <.001
8.65 ± 2.79 1.94 ± 0.97 23.67 ± 6.89 47
9.35 ± 2.53 1.62 ± 1.37 21.90 ± 4.15 18
.21 .32 .29 .005
Hb ¼ hemoglobin; HCO3 ¼ bicarbonate; INR ¼ international normalized ratio.
additional transfusion. The rate of procedural intervention, mortality, ICU days, ventilator days, length of stay, and rebleeding were similar between groups. Response to Prehospital Transfusion Forty-nine percent (n ¼ 55) of patients transfused en route were hemodynamically unstable during transport, improving to 18% (n ¼ 20) at hospital admission (Table 2). Ninety-seven percent (n ¼ 109) had laboratory values available before transfer, and 62% (n ¼ 69) had the international normalized ratio (INR) drawn before transfer. Hemoglobin significantly increased from a mean of 8.3 mg/dL before transport to 8.9 mg/dL (P ¼ .03) on hospital arrival. Patients transfused en route showed a significant decrease in INR from a mean of 2.01 before transport to 1.56 on arrival to the hospital (P ¼.01). Patients were not significantly acidotic before transfer and did not show a significant change in the bicarbonate level at hospital admission. Like the patients who received prehospital transfusion, the comparison group of patients who did not receive blood products en route showed a significant improvement in hemodynamic instability on admission to our institution after transport. Eightyeight percent (n ¼ 43) of the comparison group had laboratory values available before transfer, and only 37% (n ¼ 18) had INR drawn before transfer. Patients in the comparison group experienced an increase in hemoglobin from 8.7 to 9.3, which did not reach statistical significance. This group also showed a decrease in
INR from 1.94 to 1.62, which did not reach statistical significance. Like the group of patients who were transfused en route, the comparison patients showed no difference in the bicarbonate level before and after resuscitation and were not significantly acidotic before transfer. Discussion Patients who received prehospital PRBCs and plasma transfusion during transport for GI bleeding were critically ill, coagulopathic, and severely anemic and had a high rate of procedural intervention for hemostasis. Many showed evidence of hemodynamic instability during transport. All but 11% required additional transfusion after admission to the hospital. We found that our patients whose continuum of resuscitation included transfusion of PRBCs or plasma by critical care transport teams en route to definitive care showed a statistically significant correction of coagulopathy and anemia, which was not true for the comparison group. Notably, we did not detect a difference in any clinical outcome between groups. Patients who received transfusion en route were significantly more likely to be transfused with additional plasma after admission to our institution, which could reflect a more severe coagulation dysfunction not detected by INR or could simply be a product of a small sample size. Although it is difficult to ascertain a definite clinical benefit from the current study, our data suggest that patients are being appropriately selected for prehospital transfusion, and these interventions appear to be having the desired effect on the early correction of anemia and coagulopathy. The examination of pretransport and admission laboratory values showed a significant decrease in INR and an increase in hemoglobin after prehospital transfusion of plasma and PRBCs, respectively. Patients were anemic before transport; however, it should also be noted that the hemoglobin level likely does not accurately reflect the degree of blood loss in hypovolemic acutely bleeding patients and should be interpreted cautiously in combination with clinical variables.19 Hemoglobin level is not a strict component of our transfusion protocol but may be used along with hemodynamic status and clinical judgment in order to determine the need for transfusion. It is known that coagulopathy in upper GI bleed patients is associated with increased mortality and rebleeding4-6; therefore, we would predict that the early correction of
M.E. Parker et al. / Air Medical Journal xxx (2017) 1e5
coagulopathy may be of clinical benefit to our upper GI bleed patients. It remains to be determined whether these laboratory changes translate to a clinically significant difference in outcomes such as rebleeding and mortality. It is worth noting that in the general population of hemodynamically stable acute upper GI bleed patients, minimizing the transfusion of PRBCs is associated with improved outcomes.19,20 A randomized controlled trial found that a lower threshold of 7 g/dL for the transfusion of PRBCs compared with 9 g/dL was associated with decreased mortality and a decreased rebleeding rate in certain upper GI bleed patients, especially those with portal hypertension and variceal bleeding.21 Notably, those authors excluded patients with massive exsanguinating bleeding and hemodynamic instability, so it is challenging to directly apply these data to our patient population, which included many hemodynamically unstable patients. To our knowledge, similar studies of early transfusion or transfusion threshold have not been undertaken specifically for patients with lower GI bleeding, and it remains to be seen how, and to what degree, coagulopathy clinically affects the lower GI bleed population. The existing data highlight the importance of being judicious with the administration of blood products and avoiding copious or unnecessary PRBC transfusion in GI bleeding. We identified a subgroup of patients transfused en route who did not require further transfusion or intervention after admission, which could indicate the unnecessary use of prehospital transfusion. On review of this particular subset of patients, several clinical presentations with hypotension secondary to severe sepsis or dehydration appear to have been initially interpreted as hemorrhagic shock from GI bleeding, leading to transfusion. Several other patients had appropriate endoscopic hemostasis for ongoing GI bleeding immediately on arrival to our institution and required no further transfusion after hemostasis was achieved. Two patients refused intervention, including transfusion after transfer to our institution. One patient had undergone an operation before transfer and received transfusion en route with blood products that had been ordered by the referring institution based on clinical parameters before transfer. Finally, 1 patient had a diverticular bleed that stopped spontaneously. Given the overall clinical picture and limited information available to the prehospital critical care transport team, these transfusions were deemed warranted through the PI process previously mentioned. The appropriate use of a prehospital transfusion protocol depends in part on the clinical acumen of critical care transport personnel to rapidly determine which patients are exsanguinating and avoid unnecessary transfusion for those at lower risk. Continual analysis and re-evaluation of prehospital transfusion practices are essential to ensure appropriate use. Limitations This retrospective assessment of a single tertiary referral center experience serves to demonstrate the appropriate use of our prehospital transfusion protocol for GI bleed patients. Our data may not be generalizable to centers with different referral populations. In addition, we were not able to obtain outside records from all transferring hospitals, and, therefore, missing pretransport laboratory values are a potential source of bias in our study. Lastly, some of the INR values recorded before transport were point-of-care measurements, whereas values after arrival at our hospital were automated plasma INR. Although the majority of the literature suggests that these methods correlate well, some studies have suggested variability between these tests.22,23
5
Conclusions Our data suggest that our institution's prehospital transfusion protocol is being used appropriately and confers early improvement in hemodynamic parameters, anemia, and coagulopathy in patients with acute GI hemorrhage. Further investigation, including prospective analysis to assess whether GI bleed patients experience clinically significant improvement in outcomes with the current protocol, is warranted and whether these protocols should be altered for nontrauma patients.
References 1. Loperfido S, Baldo V, Piovesana E, et al. Changing trends in acute upper-GI bleeding: a population-based study. Gastrointest Endosc. 2009;70:212e224. 2. Velayos FS, Williamson A, Sousa KH, et al. Early predictors of severe lower gastrointestinal bleeding and adverse outcomes: a prospective study. Clin Gastroenterol Hepatol. 2004;2:485e490. 3. Abougergi MS, Travis AC, Saltzman JR. The in-hospital mortality rate for upper GI hemorrhage has decreased over 2 decades in the United States: a nationwide analysis. Gastrointest Endosc. 2015;81:882e888. 4. Jairath V, Kahan BC, Stanworth JS, et al. Prevalence, management, and outcomes of patients with coagulopathy after acute nonvariceal upper gastrointestinal bleeding in the United Kingdom. Transfusion. 2013;53:1069e1076. 5. Loffroy R, Guiu B, D'Athis P, et al. Arterial embolotherapy for endoscopically unmanageable acute gastroduodenal hemorrhage: predictors of early rebleeding. Clin Gastroenterol Hepatol. 2009;7:515e523. 6. Restellini S, Kherad O, Jairath V, Martel M, Barkun AN. Red blood cell transfusion is associated with increased rebleeding in patients with nonvariceal upper gastrointestinal bleeding. Aliment Pharmacol Ther. 2013;37: 316e322. 7. Gonzalez EA, Moore FA, Holcomb JB, et al. Fresh frozen plasma should be given earlier to patients requiring massive transfusion. J Trauma. 2007;62:112e119. 8. Mitra B, Mori A, Cameron PA, Fitzgerald M, Paul E, Street A. Fresh frozen plasma (FFP) use during massive blood transfusion in trauma resuscitation. Injury. 2010;41:35e39. 9. Riskin DJ, Tsai TC, Riskin L, et al. Massive transfusion protocols: the role of aggressive resuscitation versus product ratio in mortality reduction. J Am Coll Surg. 2009;209:198e205. 10. Bhangu A, Nepogodiev D, Doughty H, Bowley DM. Meta-analysis of plasma to red blood cell ratios and mortality in massive blood transfusions for trauma. Injury. 2013;44:1693e1699. 11. del Junco DJ, Holcomb JB, Fox EE, et al, PROMMTT Study Group. Resuscitate early with plasma and platelets or balance blood products gradually: findings from the PROMMTT study. J Trauma Acute Care Surg. 2013;75:S24eS30. 12. Novak DJ, Bai Y, Cooke RK, et al, PROPPR Study Group. Making thawed universal donor plasma available rapidly for massively bleeding trauma patients: experience from the Pragmatic, Randomized Optimal Platelets and Plasma Ratios (PROPPR) trial. Transfusion. 2015;55:1331e1339. 13. Kutcher ME, Kornblith LZ, Narayan R, et al. A paradigm shift in trauma resuscitation: evaluation of evolving massive transfusion practices. JAMA Surg. 2013;148:834e840. 14. Brown JB, Sperry JL, Fombona A, Billiar TR, Peitzman AB, Guyette FX. Pretrauma center red blood cell transfusion is associated with improved early outcomes in air medical trauma patients. J Am Coll Surg. 2015;220:797e808. 15. Kim BD, Zielinski MD, Jenkins DH, Schiller HJ, Berns KS, Zietlow SP. The effects of prehospital plasma on patients with injury. J Trauma Acute Care Surg. 2012;73:S49eS53. 16. Thiels CA, Aho JM, Fahy AS, et al. Prehospital blood transfusions in non-trauma patients. World J Surg. 2016;40:2297e2304. 17. Berns KS, Zietlow SP. Blood usage in rotor-wing transport. Air Med J. 1998;17: 105e108. 18. Jenkins D, Stubbs J, Williams S, et al. Implementation and execution of civilian remote damage control resuscitation programs. Shock. 2014;41:84e89. 19. Valeri CR, Dennis RC, Ragno G, Macgregor H, Menzoian JO, Khuri SF. Limitations of the hematocrit level to assess the need for red blood cell transfusion in hypovolemic anemic patients. Transfusion. 2006;46:365e371. 20. Villanueva C, Colomo A, Bosch A, et al. Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med. 2013;368:11e21. 21. Wang J, Bao YX, Bai M, Zhang YG, Xu WD, Qi XS. Restrictive vs liberal transfusion for upper gastrointestinal bleeding: a meta-analysis of randomized controlled trials. World J Gastroenterol. 2013;19:6919e6927. 22. Goodman MD, Makley AT, Hanseman DJ, Pritts TA, Robinson BR. All the bang without the bucks: defining essential point-of-care testing for traumatic coagulopathy. J Trauma Acute Care Surg. 2015;79:117e124. 23. van den Besselaar AM. Accuracy, precision, and quality control for point-ofcare testing of oral anticoagulation. J Thromb Thrombolysis. 2001;12:35e40.