T R A N S F U S I O N P R A C T I C E

Early changes in hemoglobin and hematocrit levels after packed red cell transfusion in patients with acute anemia

J.I. Elizalde, J. Clemente, J.L. Marin, J. Pan& B. Aragon, A. Mas, J.M. Piquk, and J. Terh

BACKGROUND: Equilibration of hemoglobin concentra- tion after transfusion has been estimated to take about 24 hours, but some studies have shown that earlier measurements reflect steady-state values in persons who have not bled recently.This study was aimed at as- sessing the changes over time in hemoglobin concentra- tion after transfusion in acutely anemic patients because of recent bleeding. STUDY DESIGN AND METHODS: Thirty-two normovolemic patients recovering from an acute bleed- ing episode who were no longer thought to be bleeding and who received a 2-unit red cell transfusion were studied. At baseline and 15, 30, 60, and 120 minutes and 24 hours after transfusion, hemoglobin concentra- tion and hematocrit values were measured. RESULTS: The administration of 2 units of packed red cells elicited a 24-hour increase of 22.4 i 6.8 g per L in hemoglobin concentration. Hemoglobin values were not different at any of the defined posttransfusion times. He- matocrit levels experienced similar changes over time. Agreement between 15-minute and 24-hour values was excellent, as only 6 percent of patients exhibited a clini- cally significant difference (>6 g/L) between the hemo- globin measurements. CONCLUSION: Hemoglobin and hematocrit values rap- idly equilibrate after transfusion in normovolemic pa- tients who are recovering from an acute bleeding epi- sode. This fact would allow a rapid assessment of the effects of transfusion and of the recurrence of bleeding in patients remaining at risk.

he clinical use of blood components has experi- enced a sustained increase during the last 50 years.' Up to now, controversy has existed about T the usefulness of monitoring the effects on hemo-

globin (Hb) or hematocrit (Hct) levels of the administration of packed red cells (RBCS).~ Although an increase of 10 to 15 g per L in Hb per unit of blood transfused is e ~ p e c t e d , ~ the time between blood transfusion and reflection of the incre- ments in the blood measurements is unknown. Most text- books state that the equilibration of Hb concentration after transfusion takes about 24 hours ,"~~ but there is no evidence supporting this idea. Conversely, some studies conducted in children6 and adults7 have provided data indicating that measurements performed as early as 15 minutes after the end of transfusion reflect steady-state values. However, the only study performed in an adult population excluded pa- tients who had had a recent bleeding episode, even though it is exactly those patients in whom a rapid assessment of the effects of transfusion could help in early detection of recur- rent bleeding. Moreover, in that study, the duration of trans- fusion was variable but, in any instance, longer than the usual time that most centers require to perform the proce- dure. In addition, only data related to Hb concentration were a s ~ e s s e d . ~ Hct levels are both rapid and easy to determine, and they provide clinical information similar to that of Hb concentration, with the possibility that the results can even be obtained at the bedside.

ABBREVIATIONS: Hb = hemoglobin; Hct = hematocrit, RBC(s) = red cell(s).

From the Gastroenterology Department, the Liver Unit, and the Emergency Laboratory Unit, Hospital Clinic i Provincial, Univer- sity of Barcelona, Barcelona, Spain.

Supported by Grant FISS 96/0241 from the Fondo de Investi- gaciones Sanitarias de la Seguridad Social and by Grant FIAP/96- 9105 from Comissionat per a Universitats i Recerca de la Generalitat de Catalunya (JIE).

tober 16,1996, and accepted December 5,1996. Received for publication August 5,1996; revision received Oc-

TRANSFUSION 1997;37:573-576.

Volume 37, June 1997 TRANSFUSION 573

ELIZALDE ET AL.

The present study was done to clarify the following is- sues: first, whether early measurements of Hb after blood transfusion adequately reflect Hb levels after prolonged equilibration periods in patients with acute anemia; second, whether Hct values really do reflect such changes; and third, whether the calculated increase in Hb (assuming an increase of 10 g/L of Hblunit of blood) differs clinically from that re- vealed in the observed data.

MATERIALS AND METHODS Subjects Patients on a gastroenterologic general ward or a gastro- enterologic intensive care unit of a 1000-bed teaching ter- tiary-care center who were selected to receive a 2-unit trans- fusion of packed RBCs because of acute blood losses in the previous 5 days were considered for this study. Only patients whose bleeding was believed to have stopped at least 24 hours before the transfusion (as judged by the stabilization of their Hb levels) were included. A normovolemic state, as defined by systolic blood pressure higher than 100 torr, right atrial pressure higher than 2 cm of H,O, heart rate of less than 100 beats per minute, and urine output equal to or greater than 0.5 mL per kg per hour, was also required at the time of inclusion. Exclusion criteria included age younger than 18 years, history of congestive heart failure, current renal failure as defined by serum creatinine levels higher than 2 mg per dL, previous inclusion in the study, or refusal to give informed consent. This study was in accordance with the standards of the Ethics Committee of the Hospital Clinic i Provincial de Barcelona.

Transfusion time was adjusted in all cases to 240 min- utes. Hb concentration and Hct values were measured be- fore and 15 minutes, 30 minutes, 1 hour, 2 hours, and 24 hours after the end of transfusion. For that purpose, blood from an arterial or central venous line was drawn into stan- dard 3-mL heparinized tubes (Vacutainer, Becton Dickinson Company, Rutherford, NJ), and complete blood counts were measured on automated analyzers (H1 System, Technicon Instruments, Tarrytown, NY) by standard laboratory meth- ods within 1 hour of blood extraction. Care was taken to obtain the samples from the same line in each individual case. Body surface area was calculated from body weight and height by using standard nomograms.

Statistics We calculated sample size by taking into account an ex- pected increase in Hb concentration of 20 f 8 g per L7 and by considering as clinically significant those changes in Hb concentration greater than 6 g per L.8 With the probability of a type I error established at a = 0.05, a sample of 32 pa- tients would yield 90-percent power to detect a change of 6 g per L in the Hb level.

Data are presented as mean * SD, and, for some analy- ses, 95-percent CIS are also provided. The changes in Hb or

Hct levels over time were examined by repeated-measures ANOVA using a statistical package (Statview 4.5, Abacus Concepts, Berkeley, CA) on a computer (Macintosh LC475, Apple Computer, Cupertino, CA). Sex, age, weight, body sur- face area, urine output, creatinine clearance, and the time since hemostasis were included in the ANOVA to explore possible confounding effects on Hb equilibration. The agreement between Hb levels measured at 24 hours and those measured at 15 minutes or calculated on the basis of the expected increase in Hb concentration was also as- sessed. For that purpose, the mean difference and the stan- dard deviation of the differences between 15-minute and 24- hour Hb concentration data and between expected8 and real 24-hour Hb values were calculated and compared to the accepted, clinically relevant differences for Hb concentra- t i ~ n . ~ The same approach was used to assess the agreement between 15-minute and 24-hour Hct values. Significance was established at a two-sided p value of 0.05.

RESULTS From August 1995 to June 1996,35 consecutive patients ful- filling the inclusion criteria were included in this study. Among them, 23 patients had had an acute gastrointestinal bleeding episode in the previous 5 days, the source of which was a bleeding peptic ulcer in 10 patients, stress ulcers in 4, and gastroesophageal varices in 9. In the remaining 12 pa- tients, transfusion was indicated because of anemia second- ary to blood losses during or after surgery. Once included, two patients experienced recurrence of bleeding, and, in another, blood transfusion had to be interrupted because of the development of a transfusion reaction. Thus, 32 patients (21 men and 11 women, aged 61 f 3 years) were included in the data analysis. At entry, mean arterial blood pressure and heart rate were 81 f 3 torr and 83 * 2 beats per minute, re- spectively, for the whole group.

The administration of 2 units of packed RBCs elicited a 24-hour increase in Hb concentration of 22.4 f 6.8 g per L, a response that was not different (p = 0.2) from that expected according to the published data (20 f 8 glL). The values of Hb concentration were not different (p = 0.4) when mea- sured at 15 minutes, 30 minutes, l hour, 2 hours, or 24 hours after the end of transfusion (Fig. 1A). None of the parameters included in the ANOVA as possible confounding factors (sex, age, weight, body surface area, urine output, creatinine clearance, and time since hemostasis) were found to influ- ence Hb equilibration. Hct levels did show similar changes over time after transfusion (Fig. l B ) , with increases of 6.4 percent (95% CI, 6.1-7.30) at 15 minutes and of 6.8 percent (95% CI, 6.14-7.40%) at 24 hours.

The differences between 15-minute and 24-hour Hb concentration data and between expected (determined by adding 20 g/L to the baseline values) and observed 24-hour Hb values were calculated. The 95-percent CI for these dif-

574 TRANSFUSION Volume 37, June 1997

TRANSFUSION MONITORING IN ACUTE ANEMIA

Baseline 15 rnin 30 min 60 rnin 120 rnin 24 hours

Baseline 15 rnin 30 min 60 rnin 120 rnin 24 hours

Time Fig. 1. A) Hb concentration at baseline and at various intervals after the transfusion of 2 units of packed RBCs. Results are ex- pressed as mean i 1 SD. B) Hct values at baseline and at vari- ous intervals after the transfusion of 2 packed RBC units. Re- sults are expressed as mean f 1 SD.

ferences were -9.26 to 14.02 g per L for expected and ob- served 24-hour Hb values, and -5.94 to 7.86 g per L for 15- minute and 24-hour Hb concentrations. The last figure com- pares favorably with the previous, and with the generally accepted significant change in Hb level, which is between 6.6 and 10 g per L.8 Moreover, the difference between ex- pected and observed 24-hour Hb values was less than 6.6 g per L in 62.5 percent of cases; the difference was less than 6.6 g per L in 93.8 percent when 24-hour values were plot- ted against 15-minute data (p = 0.006). Similarly, the agree- ment between Hct values obtained at 15 minutes and 24 hours after transfusion was excellent, as 94 and 75 percent of patients exhibited a difference equal to or less than 2 and 1 percent, respectively, between the determinations.

DISCUSSION Gastrointestinal bleeding is a medical emergency in which prompt and timely intervention may be 1ife-saving.IO Recur- rence of bleeding, which has been identified as one of the major determinants of a fatal outcome, usually constitutes an indication for surgical or endoscopic treatment. Most

groups include a drop in the Hb level in the definition of rebleeding, because other signs, such as changes in heart rate or blood pressure, may be delayed, absent, or attributed to intercurrent condition^.*^-'^ Moreover, external bleeding may not occur in several situations, including some duode- nal ulcers in which monitoring of gastric aspirates is useless. The measurement of Hb or Hct levels two or three times a day is generally used to monitor patients who are at risk for rebleeding, as a way of detecting clinically unsuspected re- currences, and a rapid assessment of the effects of blood transfusion in these patients could be of great importance in their monitoring.

Even though more than 12 million units of packed RBCs are transfused annually in the United States,I4 there is still a considerable amount of confusion about the best way to monitor the effects of this therapy tool. For instance, there is no reason to explain why many centers delay the measure- ment of Hb levels for up to 24 hours after transfusion. Re- cently, an equilibration study has shown that the platelet counts 10 and 60 minutes after platelet transfusion are equivalent,I5 which enables investigators to decrease the time that patients spend in the clinic. Similarly, some data have been provided, demonstrating that the equilibration rate following packed RBC administration in patients with- out recent bleeding is very fast, because measurements per- formed as early as 15 minutes after transfusion really do reflect steady-state values.’

The main aim of the present study was to assess whether the changes over time in Hb concentration after RBC transfusion in acutely anemic patients were similar to those reported in patients with chronic anemia. With that purpose, patients who had been rendered anemic because of a recent bleeding episode and who were selected to re- ceive a 2-unit packed RBC transfusion were included. Be- cause an equilibration study requires a steady state, only patients who had stopped bleeding at least 24 hours earlier and were normovolemic at the moment of inclusion were considered. Our results disclose that Hb measurements per- formed as early as 15 minutes after the end of transfusion in acutely anemic patients reflect the effect of transfusion on Hb concentration, because the values are not different from those obtained at 24 hours. This result cannot be as- cribed to a long transfusion time, as all patients were trans- fused at the usual rate employed in the clinical practice for patients without hypovolemic shock or cardiac or renal fail- ure. Moreover, in nearly 40 percent of patients, the 24-hour increase in Hb differed by 6.6 g per L or more from the ex- pected increase, according to a theoretical 10 g per L in- crease per unit of blood.3 This result points out the need to monitor the effects of transfusion, as response variability exists among patients. In this study, no factor was identified as being responsible for such variability, but this was not in the scope of the study and such a determination would probably require a larger number of patients. The difference

Volume 37, June 1997 TRANSFUSION 575

ELIZALDE ET AL.

between 15-minute and 24-hour Hb values was greater than 6.6 g per L in only 6 percent of patients. This finding chal- lenges the generalized concept that equilibration of blood Hb concentration after transfusion requires a long time in acutely anemic patients with recent bleeding.

Hb concentration is generally preferred to Hct measure- ment in assessing the extent of anemia, because it is mea- sured directly, not calculated.8 However, determination of the Hct value is a cheaper and easier procedure.16 This study suggests that Hct values could be as useful as Hb measure- ments at monitoring the effects of RBC transfusion. How- ever, a larger number of patients would be necessary to con- firm this assessment.

In summary, the results of the present study confirm previous work showing that Hb concentration rapidly equili- brates after a 2-unit packed RBC transfusion and extends this concept to patientswith acute anemia, even when trans- fusion time is as short as that usually employed in the clini- cal practice.

ACKNOWLEDGMENTS

The authors thank the nursing staff of the Gastroenterology and Liver Intensive Care Units. They are also indebted to Maite Gueto for secretarial support.

REFERENCES 1. Wallace EL, Surgenor DM, Hao HS, et al. Collection and trans-

fusion of blood and blood components in the United States, 1989. Transfusion 1993;33:139-44.

2. McCullough 1. The nation’s changing blood supply system.

3. Simon TL. Red cell transfusion. In: Rossi EC, Simon TL, Moss GS, eds. Principles of transfusion medicine. Baltimore: Will- iams& Wilkins, 1991:97-140.

4. Barnes A. Blood component therapy. In: Bick RL, ed. Hematol- ogy: clinical and laboratory practice. St. Louis: Mosby, 1993:1653-70.

5. Jones 1. The transfusion of red cells in anemia. In: Mollison PL, Engelfreit CP, Contreras M, eds. Blood transfusion in clinical medicine. 9th ed. Oxford: Blackwell, 1993:420-33.

6. Sekhsaria S, Fomufod A. Readjustment of hematocrit values after packed red cell transfusion in the neonate. J Perinatol 1991; 1 1: 161-3.

7. Wiesen AR, Hospenthal DR. Byrd JC, et al. Equilibration of he- moglobin concentration after transfusion in medical inpatients not activelybleeding. Ann Intern Med 1994;121:278-80.

JAMA 1993;269:2239-45.

8.

9.

10.

11.

12.

13.

14.

15.

16.

Wintrobe MM. Clinical hematology. 8th ed. Philadelphia: Lea &

Febiger, 1981:8. Bland JM, Altman DG. Statistical methods for assessing agree- ment between two methods of clinical measurement. Lancet

Panos MZ, Walt RP. Current management of bleeding peptic u1cer.A review. Drugs 1993;46:269-80. Turner IB, Jones M, Piper DW. Factors influencing mortality from bleeding peptic ulcers. Scand J Gastroenterol

Zimmerman J, Siguencia J, Tsvang E, et al. Predictors of mortal- ity in patients admitted to hospital for acute upper gastrointes- tinal hemorrhage. Scand J Gastroenterol1995;30:327-31. Rockall TA, Logan RF, Devlin HB, Northfield TC. Variation in outcome after upper gastrointestinal haemorrhage. The Na- tional Audit of Acute Upper Gastrointestinal Haemorrhage. Lancet 1995;346:346-50. Surgenor DM, Wallace EL, Hao ES, Chapman RH. Collection and transfusion of blood in the United States, 1982-1988. N Engl J Med 1990;322:1646-51. O’Connell B, Lee EJ, Schiffer CA. The value of 10-minute post- transfusion platelet counts. Transfusion 1988;28:66-7. Jiranek GC, Kozarek RA. A cost-effective approach to the pa- tient with peptic ulcer bleeding. Surg Clin North Am

19861 ~307- 10.

1991;26:661-6.

1996;76:83-103.

AUTHORS

J. Ignasi Elizalde, MD, Assistant Professor, Gastroenterology Depart- ment, Hospital Clinic i Provincial, Universitat de Barcelona, Villar- roe1 170,08036 Barcelona, Spain. [Reprint requests]

Josefina Clemente, RN, Staff Nurse, Liver Unit, Hospital Clinic i Provincial, Universitat de Barcelona.

Jose Luis Marin, MD, Assistant Professor, Emergency Labora- tory Unit, Hospital Clinic i Provincial, Universitat de Barcelona.

Julian Pan& MD, Assistant Professor, Gastroenterology De- partment, Hospital Clinic i Provincial, Universitat de Barcelona.

Blanca Aragbn, RN, Staff Nurse, Liver Unit, Hospital Clinic i Provincial, Universitat de Barcelona.

Antoni Mas, MD, Assistant Chief Professor, Liver Unit, Hospital Clinic i Provincial, Universitat de Barcelona.

Josep M. Pique, MD, Assistant Chief Professor, Gastroenterol- ogy Department, Hospital Clinic i Provincial, Universitat de Barce- lona.

Josep Teres, MD, Chief Professor, Gastroenterology Depart- ment, Hospital Clinic i Provincial, Universitat de Barcelona.

576 TRANSFUSION Volume 37, June 1997