[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.161.241.199. Please contact the publisher to request reinstatement.
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Download PDF
Figure 1.
Incidence of Complications Stratified by Different Combinations of ΔHb and Nadir Hb Levels
Incidence of Complications Stratified by Different Combinations of ΔHb and Nadir Hb Levels

Of the patients with a hemoglobin (Hb) level of 7 g/dL or greater (to convert to grams per liter, multiply by 10.0), those with a delta hemoglobin (ΔHb) level of 50% or greater were at higher risk of complications (P < .001).

Figure 2.
Spline Function Plot of the Relationship Between Decrease in Hemoglobin Level and Morbidity
Spline Function Plot of the Relationship Between Decrease in Hemoglobin Level and Morbidity

A, Relationship stratified by sex. B, Relationship stratified by type of surgery.

Table 1.  
Preoperative Factors Associated With the Percentage of Decrease in Hb Levels
Preoperative Factors Associated With the Percentage of Decrease in Hb Levels
Table 2.  
Factors Associated With Perioperative Morbidity Among the Entire Cohort and Patients Who Received Transfusions
Factors Associated With Perioperative Morbidity Among the Entire Cohort and Patients Who Received Transfusions
Table 3.  
Factors Associated With Perioperative Ischemic Complication Among the Entire Cohort and Patients Who Received Transfusions
Factors Associated With Perioperative Ischemic Complication Among the Entire Cohort and Patients Who Received Transfusions
1.
Carson  JL, Grossman  BJ, Kleinman  S,  et al; Clinical Transfusion Medicine Committee of the AABB.  Red blood cell transfusion: a clinical practice guideline from the AABB. Ann Intern Med. 2012;157(1):49-58.
PubMedArticle
2.
Fleisher  LA, Beckman  JA, Brown  KA,  et al; American College of Cardiology; American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2002 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery); American Society of Echocardiography; American Society of Nuclear Cardiology; Heart Rhythm Society; Society of Cardiovascular Anesthesiologists; Society for Cardiovascular Angiography and Interventions; Society for Vascular Medicine and Biology.  ACC/AHA 2006 guideline update on perioperative cardiovascular evaluation for noncardiac surgery: focused update on perioperative beta-blocker therapy: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2002 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery) developed in collaboration with the American Society of Echocardiography, American Society of Nuclear Cardiology, Heart Rhythm Society, Society of Cardiovascular Anesthesiologists, Society for Cardiovascular Angiography and Interventions, and Society for Vascular Medicine and Biology. J Am Coll Cardiol. 2006;47(11):2343-2355.
PubMedArticle
3.
Ferraris  VA, Brown  JR, Despotis  GJ,  et al; Society of Thoracic Surgeons Blood Conservation Guideline Task Force; Society of Cardiovascular Anesthesiologists Special Task Force on Blood Transfusion; International Consortium for Evidence Based Perfusion.  2011 Update to the Society of Thoracic Surgeons and the Society of Cardiovascular Anesthesiologists blood conservation clinical practice guidelines. Ann Thorac Surg. 2011;91(3):944-982.
PubMedArticle
4.
Napolitano  LM, Kurek  S, Luchette  FA,  et al; American College of Critical Care Medicine of the Society of Critical Care Medicine; Eastern Association for the Surgery of Trauma Practice Management Workgroup.  Clinical practice guideline: red blood cell transfusion in adult trauma and critical care. Crit Care Med. 2009;37(12):3124-3157.
PubMedArticle
5.
Frank  SM, Resar  LM, Rothschild  JA, Dackiw  EA, Savage  WJ, Ness  PM.  A novel method of data analysis for utilization of red blood cell transfusion. Transfusion. 2013;53(12):3052-3059.
PubMedArticle
6.
Frank  SM, Savage  WJ, Rothschild  JA,  et al.  Variability in blood and blood component utilization as assessed by an anesthesia information management system. Anesthesiology. 2012;117(1):99-106.
PubMedArticle
7.
Ejaz  A, Spolverato  G, Kim  Y, Frank  SM, Pawlik  TM.  Variation in triggers and use of perioperative blood transfusion in major gastrointestinal surgery. Br J Surg. 2014;101(11):1424-1433.
PubMedArticle
8.
Ejaz  A, Spolverato  G, Kim  Y, Frank  SM, Pawlik  TM.  Identifying variations in blood use based on hemoglobin transfusion trigger and target among hepatopancreaticobiliary surgeons. J Am Coll Surg. 2014;219(2):217-228.
PubMedArticle
9.
Carson  JL, Carless  PA, Hébert  PC.  Outcomes using lower vs higher hemoglobin thresholds for red blood cell transfusion. JAMA. 2013;309(1):83-84.
PubMedArticle
10.
Carson  JL, Carless  PA, Hebert  PC.  Transfusion thresholds and other strategies for guiding allogeneic red blood cell transfusion. Cochrane Database Syst Rev. 2012;4:CD002042.
PubMed
11.
Villanueva  C, Colomo  A, Bosch  A,  et al.  Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med. 2013;368(1):11-21.
PubMedArticle
12.
Karkouti  K, Wijeysundera  DN, Yau  TM, McCluskey  SA, van Rensburg  A, Beattie  WS.  The influence of baseline hemoglobin concentration on tolerance of anemia in cardiac surgery. Transfusion. 2008;48(4):666-672.
PubMedArticle
13.
Hogervorst  E, Rosseel  P, van der Bom  J,  et al.  Tolerance of intraoperative hemoglobin decrease during cardiac surgery. Transfusion. 2014;54(10, pt 2):2696-2704.
PubMedArticle
14.
McLean  E, Cogswell  M, Egli  I, Wojdyla  D, de Benoist  B.  Worldwide prevalence of anaemia, WHO Vitamin and Mineral Nutrition Information System, 1993-2005. Public Health Nutr. 2009;12(4):444-454.
PubMedArticle
15.
van Straten  AH, Külcü  K, Özdemir  HI, Elenbaas  TW, Soliman Hamad  MA.  Preoperative hemoglobin level as a predictor of mortality after aortic valve replacement. J Cardiothorac Vasc Anesth. 2013;27(4):716-722.
PubMedArticle
16.
Hare  GM, Freedman  J, David Mazer  C.  Risks of anemia and related management strategies: can perioperative blood management improve patient safety? Can J Anaesth.2013;60(2):168-175.
PubMedArticle
Original Investigation
October 2015

Effect of Relative Decrease in Blood Hemoglobin Concentrations on Postoperative Morbidity in Patients Who Undergo Major Gastrointestinal Surgery

Author Affiliations
  • 1Department of Surgery, Johns Hopkins Hospital, Baltimore, Maryland
  • 2Department of Anesthesiology and Critical Care Medicine, Johns Hopkins Hospital, Baltimore, Maryland
JAMA Surg. 2015;150(10):949-956. doi:10.1001/jamasurg.2015.1704
Abstract

Importance  Delta hemoglobin (ΔHb), defined as the difference between the preoperative hemoglobin (Hb) level and the nadir Hb level during a patient’s hospitalization, may be associated with adverse outcomes even if the absolute level of Hb remains greater than the transfusion threshold of 7 g/dL.

Objective  To evaluate the association between ΔHb and morbidity in patients who undergo major gastrointestinal surgery as an independent factor or combined with the nadir Hb concentration.

Design, Setting, and Participants  Retrospective review of prospectively collected data on patients who underwent pancreatic, hepatic, or colorectal resection from January 1, 2010, through April 30, 2014, at Johns Hopkins Hospital were included in the study. Data regarding the ΔHb concentration following surgery, nadir Hb level, and overall perioperative blood use were obtained and analyzed. Multivariable-adjusted logistic regression models were used to identify the preoperative factors associated with ΔHb and the effect of ΔHb on perioperative morbidity. The study and data analysis took place from January 22 through February 20, 2015.

Interventions  Major gastrointestinal surgery and packed red blood cell transfusion.

Main Outcomes and Measures  Overall morbidity and ischemic-specific complications.

Results  Of the 4669 patients who underwent major gastrointestinal surgery, the median ΔHb level after surgery was 40%. Patients with multiple comorbidities (American Society of Anesthesiologists Physical Status score of 3-4: odds ratio [OR], 1.96; 95% CI, 1.30-2.97; P <.001; ≥3 coexisting medical conditions: OR, 1.62; 95% CI, 1.08-2.42; P = .001) and those who underwent pancreatic surgery (OR, 1.98; 95% CI, 1.18-3.33; P = .01) were at increased risk of having a ΔHb of 50% or greater. Compared with patients who had a ΔHb level of less than 50% and a nadir Hb level of 7 g/dL or greater, patients with a ΔHb level of 50% or greater whose nadir Hb level was less than 7 g/dL were at a high risk of developing postoperative complications (OR, 6.60; 95% CI, 4.34-10.03; P < .001); in particular, a ΔHb level of 50% or greater was strongly correlated with a risk of ischemic complications, even if the nadir Hb level was 7 g/dL or greater (OR, 5.68; 95% CI, 1.44-22.39; P = .01).

Conclusions and Relevance  A ΔHb level of 50% or greater following gastrointestinal surgery was associated with complications, especially ischemic adverse events, even if the nadir Hb level remained at 7 g/dL or greater.

Introduction

Although recent guidelines report that the use of packed red blood cells (PRBC) is usually not indicated until the hemoglobin (Hb) level decreases to less than 7 to 8 g/dL (to convert to grams per liter, multiply by 10.0),14 to our knowledge, the ideal “trigger” for transfusion has not been determined. In patients who do not present with signs and symptoms of anemia or hemodynamic instability, laboratory indicators, such as Hb level, typically drive practice patterns around transfusion use. In the past decade, the introduction of transfusion triggers and “targets” to assess the initiation and end point of transfusion based on Hb concentrations has aided in preventing the overuse of PRBC.5,6 Our group and others have previously demonstrated that the use of a restrictive transfusion trigger resulted in equivalent or improved postoperative morbidity compared with the liberal use of blood products.711 The use of Hb triggers and targets do not, however, take into account individual variability or the possibility that a subset of patients may experience inadequate tissue oxygenation even when the Hb concentration is higher than 7 or 8 g/dL.

Aside from the use of Hb triggers, some investigators have suggested that the percentage of change in the Hb level should also be used to guide transfusion practices and help dictate when it is appropriate to administer blood products to surgical patients. For example, among patients undergoing cardiac surgery, a large percentage change or “delta” in the Hb level (ΔHb) of 50% or greater was associated with worse outcomes, even when the absolute Hb level remained greater than the commonly used threshold for PRBC transfusions (<7-8g/dL).12 Given this fact, PRBC transfusions may be indicated in certain situations based on the absolute ΔHb level as opposed to the nadir Hb level. In a cohort of patients undergoing cardiac surgery, patients who had a decrease of 50% or more in the Hb level from baseline but did not reach the transfusion trigger threshold of 7 g/dL had a greater risk of in-hospital mortality, stroke, myocardial infarction, and renal failure.13 To our knowledge, the effect of ΔHb has never been examined in a cohort of patients undergoing noncardiac operations. As such, the objective of this study was to define how the ΔHb level, relative to the Hb trigger, was associated with outcomes among patients who underwent pancreatic, hepatic, or colorectal resection at a tertiary academic hospital. Specifically, we sought to characterize how the ΔHb and nadir Hb levels (Hb level, <7 g/dL) correlated with overall morbidity, as well as ischemic-specific complications, in patients who underwent major gastrointestinal surgery.

Method
Patient Selection

Patients who underwent a pancreatic, hepatobiliary, or colorectal operation at Johns Hopkins Hospital from January 1, 2010, through April 30, 2014, were identified using the appropriate International Classification of Diseases, Ninth Revision (ICD-9), procedure codes. A retrospective review of prospectively collected data was performed; the study and data analysis took place from January 22 through February 20, 2015. The Johns Hopkins Institutional Review Board approved this study, and a patient consent waiver was signed. Standard data on demographic, clinicopathological, and therapy-related variables were collected. The presence of coexisting medical conditions was classified based on the American Society of Anesthesiologists (ASA) Physical Status classification and Charlson comorbidity index scores. Patients’ ICD-9 diagnosis codes at discharge were used to identify in-hospital perioperative complications.7,8 Cerebrovascular accidents, transient ischemic attack, and myocardial infarction were defined as ischemic adverse events.13

To avoid a bias of preexisting anemia or polycythemia, we excluded patients with abnormal Hb levels and included patients with preoperative Hb levels of 12.0 to 16.0 g/dL in women and 13.0 to 18.0 g/dL in men.14

Data Collection

Perioperative transfusion and Hb levels were obtained through the IMPACT Online Blood Management Business Intelligence Portal (Hemonetics) that integrates patients’ and cases’ clinicopathological, operative, and laboratory data. Data on the location of the transfusion (eg, operative room, intensive care unit, or department) were not available. All data in the system are collected prospectively, updated monthly, and undergo routine quality reviews to verify accuracy. For the purpose of this study, the percentage of change in the Hb level was defined as (first-nadir Hb)/first Hb × 100, and the transfusion trigger was defined as the lowest measured Hb (nadir Hb) in the patients who received a PRBC transfusion during hospitalization.5,7,8 The cutoff used for ΔHb (50% decrease) and nadir Hb levels (<7 g/dL) was based on the previous findings from patients undergoing cardiac surgery.13

Statistical Analysis

Continuous variables were presented as mean (SD) or median (interquartile range [IQR] [25th and 75th percentiles]), when appropriate, and categorical variables were presented as proportions. Differences in medians and proportions among Hb categories were assessed based on the nonparametric Mann-Whitney test and Fisher exact or χ2 tests, as indicated. Multivariable-adjusted logistic regression models were used to examine the effect of preoperative factors associated with ΔHb and the associations of ΔHb with perioperative morbidity, with results reported as odds ratios (ORs) (95% CIs). Analyses were further adjusted for PRBC transfusion to observe the effect of Hb level changes independent of this factor. Variations in ΔHb levels according to important potential effect modifiers were studied. To evaluate the factors associated with adverse outcomes related to PRBC transfusion, a planned subanalysis of patients who received at least 1 U of PRBC was also performed. Two-tailed P < .05 was considered statistically significant. All analyses were conducted in SAS, version 9.4 (SAS Institute Inc).

Results
Study Population and Baseline Characteristics

A total of 4669 patients who underwent a pancreatic (1880 [40.3%]), hepatobiliary (752 [16.1%]), or colorectal (2037 [43.6%]) operation and met the inclusion criteria were identified (eTable in the Supplement). At the time of surgery, a radical pancreaticoduodenectomy was the most common type of pancreatic resection (1248 [26.7%]). The most common liver resection was a partial hepatectomy (673 [14.4%]), and the most common colorectal resection was a right hemicolectomy (632 [13.5%]). The median estimated blood loss was 300 mL (IQR, 150-600 mL); this value was higher for pancreatic (500 mL) vs liver (300 mL) or colorectal (150 mL) resections (P = .001). Nearly one-third of patients received at least 1 U of PRBC (1473 [31.5%]) during the hospital stay, while the median volume of crystalloid fluid infused was 5000 mL (IQR, 3300-6350 mL). Of the patients who received transfusions, a median of 2 U (IQR, 2.0-4.0 U) of PRBC was transfused. The median preoperative Hb level for all patients was 12.8 g/dL (IQR, 11.5-14.0 g/dL), and the median nadir Hb level was 8.8 g/dL (IQR, 7.6-10.2 g/dL), resulting in an average ΔHb level of 40%.

After categorizing the cohort based on a ΔHb level using the median value of 50% as a cutoff, several patient-related factors were noted to differ between the 2 groups (eTable in the Supplement). Specifically, male patients (99 [62.7%] vs 2172 [48.1%]; P = .002) and patients with an ASA classification of 3 or 4 (101 [73.7%] vs 2188 [61.3%]; P = .003) were more likely to have a ΔHb level of 50% or more. Similarly, patients with congestive heart failure (8 [5.1%] vs 94 [2.1%]; P = .01) or pulmonary diseases (27 [17.1%] vs 491 [10.9%]; P = .02) and those who underwent a radical pancreaticoduodenectomy (54 [34.2%] vs 1194 [26.5%]; P = .03) were more likely to have a ΔHb level of 50% or greater. Perhaps as expected, the preoperative Hb level (14.5 g/dL vs 12.7 g/dL; P <.001) and estimated blood loss (650 mL vs 300 mL; P <.001) were higher among those who experienced a larger ΔHb level. With regard to PRBC transfusion, patients who had a ΔHb level of 50% or greater were more likely to receive a PRBC transfusion (141 [89.2%] vs 1332 [29.5%]; P <.001) as well as more crystalloid fluids (6371 mL vs 5000 mL; P <.001) vs patients who had a ΔHb level of less than 50%. On unadjusted analysis, patients with a ΔHb level of 50% or greater had a higher incidence of perioperative complications (75 [47.5%] vs 746 [16.5%]; P <.001) and a longer length of stay (15 vs 7 days; P <.001) vs patients with a ΔHb level of less than 50%.

Factors Associated With 50% or Greater ΔHb Levels

The effect of different preoperative factors on a decreased ΔHb level of 50% or more for patients undergoing hepatopancreaticobiliary and colorectal surgery was then evaluated. On univariable analysis, male patients (OR, 1.81; 95% CI, 1.30-2.51; P < .001), patients with multiple comorbidities (ASA 3-4: OR, 1.77; 95% CI, 1.21-2.61; P = .004; ≥3 coexisting medical conditions: OR, 1.47; 95% CI, 1.05-2.06; P = .03), and patients who underwent a pancreatic resection (OR, 1.65; 95% CI, 1.00-2.71; P = .049) were at an increased risk of having a ΔHb level of 50% or greater (Table 1). After adjusting for competing risk factors on multivariable analysis, patients with an ASA classification of 3 or 4 (OR, 1.96; 95% CI, 1.30-2.97; P < .001), 3 or more coexisting medical conditions (OR, 1.62; 95% CI, 1.08-2.42; P = .02), and pancreatic resection (OR, 1.98; 95% CI, 1.18-3.33; P = .01) remained independently associated with a higher risk of experiencing a ΔHb level of 50% or greater.

Association of ΔHb Level With Perioperative Morbidity

Morbidity associated with different combinations of ΔHb and nadir Hb levels is shown in Figure 1. Of the patients with an Hb level of 7 g/dL or greater, those with a ΔHb level of 50% or more were at higher risk of complications (P < .001).

The effect of different clinicopathological factors on the morbidity of patients undergoing pancreatic, hepatic, and colorectal surgery was then analyzed (Table 2). On the univariable analysis, the following factors were associated with increased risk of postoperative complications: an age of 65 years or older, male sex, an ASA classification of 3 or 4, a Charlson comorbidity index score of 3 or higher, 3 or more coexisting medical conditions, colorectal and pancreatic surgery, receipt of a transfusion, and the patient’s first Hb level (Table 2 and Figure 2). After adjusting for demographic, clinical, and pathological characteristics, different combinations of ΔHb and nadir Hb levels were evaluated. Compared with patients who did not experience a ΔHb level of 50% or greater and had a nadir Hb level of 7 g/dL or higher (reference group) (Table 2), patients with a ΔHb level of 50% or greater whose nadir Hb level was less than 7 g/dL had a higher risk of complications (OR, 6.60; 95% CI, 4.34-10.03; P < .001). Patients with a ΔHb level of 50% or more whose nadir Hb level remained at 7 g/dL or greater also were more likely to experience a postoperative complication (OR, 5.16; 95% CI, 2.98-8.93; P < .001). Patients who had a ΔHb level of less than 50% and a nadir Hb level of 7 g/dL were slightly less likely to have postoperative complications (OR, 3.65; 95% CI, 2.84-4.70; P < .001). In addition, when separately evaluating the effect of having a ΔHb level of 50% or greater and a Hb nadir level of less than 7 g/dL on morbidity, each factor was an independent predictor of adverse events (ΔHb level of ≥50%: OR, 4.56; 95% CI, 3.30-6.29; Hb level of <7 g/dL: OR, 4.78; 95% CI, 3.91-5.85; P < .001).

To evaluate factors associated with adverse outcomes related to PRBC transfusion, a planned subanalysis of patients who received at least 1 U of PRBC was conducted (1473 [31.5%]). Of the patients with a ΔHb level of less than 50% and a nadir Hb level of 7 g/dL or greater, 987 (23.8%) received at least 1 U of PRBC (median, 2 U; IQR, 1-3 U). Most patients (94 [95.9%]) with a ΔHb level of 50% or greater and a nadir Hb level of less than 7 g/dL received transfusions (median, 4 U; IQR, 2-10 U). Of the patients with a ΔHb level of 50% or greater and a nadir Hb level of 7 g/dL or higher, 47 (78.3%) received transfusions (median, 2 U; IQR 2-6 U). Finally, of the patients who had a ΔHb level of less than 50% and a nadir Hb level of less than 7 g/dL, transfusion was common (345 [93.2%]; median, 3 U; IQR 2-6 U).

Of the patients who received PRBC transfusions, those who had a ΔHb level of 50% or greater were more likely to experience postoperative complications independent of the nadir Hb level. Specifically, patients with a ΔHb level of 50% or greater whose nadir Hb level was less than 7 g/dL were almost 3 times more likely to experience complications vs patients with a ΔHb level of less than 50% and a nadir Hb level of 7 g/dL or greater (OR, 2.92; 95% CI, 1.84-4.64; P < .001) (Table 2). Patients with a ΔHb level of 50% or greater whose nadir Hb level remained at 7 g/dL or higher were more than 2 times more likely to experience a postoperative complication vs patients with a ΔHb level of less than 50% and an Hb level of 7 g/dL or higher (OR, 2.29; 95% CI, 1.18-4.47; P = .02). In addition, patients with a ΔHb level of less than 50% and a nadir Hb level of less than 7 g/dL were also at higher risk of morbidity (OR, 2.11; 95% CI, 1.58-2.80; P < .001).

Finally, the effect of a decrease in the Hb level and nadir Hb level on ischemic complications was evaluated (Table 3). Both a ΔHb level of 50% or greater (OR, 5.53; 95% CI, 2.28-13.41; P < .001) and an Hb level of less than 7 g/dL (OR, 4.77; 95% CI, 2.42-9.39; P < .001) were associated with anemia-related complications (Table 3). Compared with patients who had a ΔHb level of less than 50% and a nadir Hb level of 7 g/dL or greater, patients who had a ΔHb level of 50% or greater had the highest risk of ischemia-related complications, even if the nadir Hb level remained at 7 g/dL or greater (OR, 5.68; 95% CI, 1.44-22.39; P = .01). In addition, of the patients who received at least 1 U of PRBC, only patients who had a ΔHb level of 50% or greater and whose nadir HB level was less than 7 g/dL were at higher risk of anemia-related complications (OR, 2.87; 95% CI, 1.19-6.92; P = .02) (Table 3).

Discussion

Symptoms of anemia as well as intravascular volume are important parameters to consider when determining the need for PRBC transfusion to avoid anemia-related ischemic injury. Because these parameters are sometimes subjective and difficult to assess quantitatively, nadir Hb levels have been proposed and adopted at many centers as an objective measure to guide transfusion practices in patients without signs and symptoms of anemia or hemodynamic instability.511 However, the exclusive use of Hb triggers to guide transfusion practices does not consider patient-specific factors such as sex, age, or the absolute change in the patient’s Hb level. While data from cardiac studies have suggested that the ΔHb level was associated with the risk of postoperative morbidity and the need for transfusion,13 to our knowledge, the ΔHb level has not been evaluated in a general surgery population. This study was important because we studied a large cohort of patients who underwent gastrointestinal surgery, including pancreatic, hepatobiliary, and colorectal procedures, and specifically defined how ΔHb relative to Hb triggers was associated with outcomes. Particular strengths of our study included the large cohort size (N = 4669) and the prospective high-quality transfusion and Hb data that were available through our institutional IMPACT Online Blood Management Business Intelligence Portal. We were able to examine patterns in Hb levels among patients and determine the use of PRBC and transfusion practices. Perhaps more important, we demonstrated that a ΔHb level of 50% or more was associated with complications, especially adverse ischemic events, even if the nadir Hb level remained higher than 7 g/dL.

Several factors may be associated with changes in the Hb level, including operative blood loss and the amount of crystalloid administered in the perioperative period.5,6 We noted that factors associated with ΔHb levels included having an ASA classification of 3 or 4 and having 3 or more coexisting medical conditions. In turn, these factors were associated with an increased risk of transfusion, both among patients who had a nadir Hb level greater than 7 g/dL and those who did not. Of note, most patients who had a ΔHb level of 50% or greater and a nadir Hb level of less than 7 g/dL received transfusions (94 [95.9%]), while only about one-fourth (987 [23.8%]) of patients with a ΔHb level of less than 50% and a nadir Hb level of 7 g/dL or greater received a transfusion. In contrast, patients who had an Hb level of 7 g/dL were likely (345 [93.2%]) to receive a transfusion, even if the ΔHb level was less than 50%. These data, as well as data from 2 other studies,12,13 suggest that future studies investigating the relationship between transfusion strategy and perioperative outcomes should focus not only on Hb triggers but also on the change in Hb level associated with any procedure or hospital stay.

The emerging emphasis on the ΔHb level is related to data suggesting that changes in the Hb level—independent of the nadir Hb level—may be associated with an increase in adverse outcomes. While van Straten and colleagues15 have reported that the preoperative Hb level itself can predict mortality after surgery, other investigators have emphasized the importance of the change in Hb level rather than the initial Hb level.12,13,15 Hogervorst et al13 noted that a ΔHb level of 50% or greater during cardiac surgery was associated with adverse outcomes, even if the absolute Hb level remained higher than the commonly used transfusion threshold of 7 g/dL. Specifically, compared with patients whose Hb level remained at 7 g/dL or greater and did not decrease by 50% or more, the risk of in-hospital mortality, stroke, myocardial infarction, and renal failure was higher among patients who had an Hb level remain at 7 g/dL or greater but experienced a ΔHb level of 50% or more.13 Karkouti and colleagues12 similarly analyzed the degree of acute anemia that patients could safely tolerate during cardiac surgery. The authors noted that a ΔHb level of 50% or more during surgery was associated with more complications in cardiac patients. In this study, we noted that patients with a ΔHb level of 50% or greater—regardless of whether the nadir Hb level was less than 7 g/dL or 7 g/dL or greater—had a 5- to 6-fold higher risk of complications (Table 2). Collectively, data from the current study, as well as from previous cardiac studies,12,13 strongly suggest that the ΔHb level may be an important factor in stratifying patients with regard to the risk of complication and need for a transfusion.

As noted by others,13,16 an acute ΔHb level may have particular adverse consequences on end-organ function, especially oxygen-dependent organs, such as the kidney, heart, and brain. As such, large decreases or changes in the ΔHb level may be particularly associated with ischemic complications, such as myocardial infarction, transient ischemic attack, and cerebrovascular accidents. Hogervorst et al13 and Karkouti et al12 noted an increased risk of ischemic complications with ΔHb levels of 50% or more; however, these studies were conducted among cardiac patients who were at a higher baseline risk of cardiovascular or ischemic complications. This study expands on these findings because it examined the risk of ischemic complications among a more general population of patients undergoing gastrointestinal surgery. Patients undergoing pancreatic, hepatobiliary, or colorectal procedures who had a ΔHb level of 50% or greater had more ischemic-related complications, such as cerebrovascular accidents, transient ischemic attack, and myocardial infarction, compared with patients who did not, even if the nadir Hb level remained higher than 7 g/dL. As such, when assessing the risk of ischemic-related morbidity, the ΔHb level should be considered rather than relying solely on an absolute Hb trigger of 7 to 8 g/dL.

Several limitations should be considered when interpreting the data. Patients in the analytic cohort were limited to individuals who did not have preexisting anemia or polycythemia and who had a preoperative Hb level of 12.0 to 16.0 g/dL (women) or 13.0 to 18.0 g/dL (men). While we used these criteria to avoid bias and to be consistent with previous cardiac studies that examined the ΔHb level,12,13 whether our findings are generalizable to patients with preexisting anemia requires further study. In addition, perioperative morbidity data were obtained from medical record abstraction using ICD-9 codes, and thus the overall complication rate may be underestimated. However, this detection bias is likely random in nature and unlikely to have a significant effect on our overall results and conclusions. Finally, data on whether transfusion occurred intraoperatively or postoperatively were not available. The clinical decision to give a transfusion may be based on different parameters in the operating room, such as estimated blood loss and hemodynamics.

Conclusions

Our data imply that an exclusive focus on a nadir Hb level of 7 to 8 g/dL as the recommended best-practice trigger for PRBC transfusion may not always be appropriate. Future randomized clinical trials should examine the effect of both nadir Hb and ΔHb levels when investigating the association of transfusion practices with the risk of perioperative complications.

Back to top
Article Information

Accepted for Publication: April 3, 2015.

Corresponding Author: Timothy M. Pawlik, MD, MPH, PhD, Department of Surgery, Johns Hopkins Hospital, 600 N Wolfe St, Blalock 688, Baltimore, MD 21287 (tpawlik1@jhmi.edu).

Published Online: July 29, 2015. doi:10.1001/jamasurg.2015.1704.

Author Contributions: Drs Frank and Pawlik had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: All authors.

Acquisition, analysis, or interpretation of data: Kim, Ejaz, Frank.

Drafting of the manuscript: Spolverato, Frank.

Critical revision of the manuscript for important intellectual content: Kim, Ejaz, Frank, Pawlik.

Statistical analysis: Kim, Ejaz.

Administrative, technical, or material support: Spolverato, Pawlik.

Study supervision: Frank, Pawlik.

Conflict of Interest Disclosures: None reported.

References
1.
Carson  JL, Grossman  BJ, Kleinman  S,  et al; Clinical Transfusion Medicine Committee of the AABB.  Red blood cell transfusion: a clinical practice guideline from the AABB. Ann Intern Med. 2012;157(1):49-58.
PubMedArticle
2.
Fleisher  LA, Beckman  JA, Brown  KA,  et al; American College of Cardiology; American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2002 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery); American Society of Echocardiography; American Society of Nuclear Cardiology; Heart Rhythm Society; Society of Cardiovascular Anesthesiologists; Society for Cardiovascular Angiography and Interventions; Society for Vascular Medicine and Biology.  ACC/AHA 2006 guideline update on perioperative cardiovascular evaluation for noncardiac surgery: focused update on perioperative beta-blocker therapy: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2002 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery) developed in collaboration with the American Society of Echocardiography, American Society of Nuclear Cardiology, Heart Rhythm Society, Society of Cardiovascular Anesthesiologists, Society for Cardiovascular Angiography and Interventions, and Society for Vascular Medicine and Biology. J Am Coll Cardiol. 2006;47(11):2343-2355.
PubMedArticle
3.
Ferraris  VA, Brown  JR, Despotis  GJ,  et al; Society of Thoracic Surgeons Blood Conservation Guideline Task Force; Society of Cardiovascular Anesthesiologists Special Task Force on Blood Transfusion; International Consortium for Evidence Based Perfusion.  2011 Update to the Society of Thoracic Surgeons and the Society of Cardiovascular Anesthesiologists blood conservation clinical practice guidelines. Ann Thorac Surg. 2011;91(3):944-982.
PubMedArticle
4.
Napolitano  LM, Kurek  S, Luchette  FA,  et al; American College of Critical Care Medicine of the Society of Critical Care Medicine; Eastern Association for the Surgery of Trauma Practice Management Workgroup.  Clinical practice guideline: red blood cell transfusion in adult trauma and critical care. Crit Care Med. 2009;37(12):3124-3157.
PubMedArticle
5.
Frank  SM, Resar  LM, Rothschild  JA, Dackiw  EA, Savage  WJ, Ness  PM.  A novel method of data analysis for utilization of red blood cell transfusion. Transfusion. 2013;53(12):3052-3059.
PubMedArticle
6.
Frank  SM, Savage  WJ, Rothschild  JA,  et al.  Variability in blood and blood component utilization as assessed by an anesthesia information management system. Anesthesiology. 2012;117(1):99-106.
PubMedArticle
7.
Ejaz  A, Spolverato  G, Kim  Y, Frank  SM, Pawlik  TM.  Variation in triggers and use of perioperative blood transfusion in major gastrointestinal surgery. Br J Surg. 2014;101(11):1424-1433.
PubMedArticle
8.
Ejaz  A, Spolverato  G, Kim  Y, Frank  SM, Pawlik  TM.  Identifying variations in blood use based on hemoglobin transfusion trigger and target among hepatopancreaticobiliary surgeons. J Am Coll Surg. 2014;219(2):217-228.
PubMedArticle
9.
Carson  JL, Carless  PA, Hébert  PC.  Outcomes using lower vs higher hemoglobin thresholds for red blood cell transfusion. JAMA. 2013;309(1):83-84.
PubMedArticle
10.
Carson  JL, Carless  PA, Hebert  PC.  Transfusion thresholds and other strategies for guiding allogeneic red blood cell transfusion. Cochrane Database Syst Rev. 2012;4:CD002042.
PubMed
11.
Villanueva  C, Colomo  A, Bosch  A,  et al.  Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med. 2013;368(1):11-21.
PubMedArticle
12.
Karkouti  K, Wijeysundera  DN, Yau  TM, McCluskey  SA, van Rensburg  A, Beattie  WS.  The influence of baseline hemoglobin concentration on tolerance of anemia in cardiac surgery. Transfusion. 2008;48(4):666-672.
PubMedArticle
13.
Hogervorst  E, Rosseel  P, van der Bom  J,  et al.  Tolerance of intraoperative hemoglobin decrease during cardiac surgery. Transfusion. 2014;54(10, pt 2):2696-2704.
PubMedArticle
14.
McLean  E, Cogswell  M, Egli  I, Wojdyla  D, de Benoist  B.  Worldwide prevalence of anaemia, WHO Vitamin and Mineral Nutrition Information System, 1993-2005. Public Health Nutr. 2009;12(4):444-454.
PubMedArticle
15.
van Straten  AH, Külcü  K, Özdemir  HI, Elenbaas  TW, Soliman Hamad  MA.  Preoperative hemoglobin level as a predictor of mortality after aortic valve replacement. J Cardiothorac Vasc Anesth. 2013;27(4):716-722.
PubMedArticle
16.
Hare  GM, Freedman  J, David Mazer  C.  Risks of anemia and related management strategies: can perioperative blood management improve patient safety? Can J Anaesth.2013;60(2):168-175.
PubMedArticle
×