Reduction of blood loss from laboratory testing in hospitalized adult patients using small-volume (pediatric) tubes.
Abstract: Context.--Blood loss from laboratory testing (BLLT) can be significant in hospitalized patients. It is a common practice to draw full large-volume tubes of blood from adults.

Objective.--To determine whether BLLT occurred when a small-volume (pediatric) blood collection tube (SVT) was substituted for each large-volume blood collection tube and to note whether an adequate sample still was obtained.

Design.--During 2 consecutive weeks, hospital test requisitions were reviewed to collect patient demographics, tests requested, and number and type of tubes obtained. The amount of blood collected and BLLT per patient were calculated. Reduced sample requirements were calculated, and phlebotomists and ward nurses were required to use SVTs. After 2 weeks of familiarization, data were collected as previously described. Laboratory technicians logged problems related to the use of SVTs.

Results.--Baseline: 227 patients had 664 requisitions, and median BLLT per patient was 13.5 mL (interquartile range [IQR], 7.6-27.3 mL). In critical care patients, the median was 19.9 mL (IQR, 12.0-35.8 mL), and maximum BLLT was 159.8 mL. Intervention phase: 246 patients had 696 requisitions, median BLLT was 3.7 mL (IQR, 1.2-6.3 mL; P < .001). In critical care patients, the median was 5.1 mL (IQR, 2.3-10.9 mL), and maximum BLLT was 61.8 mL (P < .001). All tests requested could be performed using SVTs, and no additional blood collections were required. Use of SVTs reduced overall BLLT per patient by 73% and by 74% in critical care patients.

Conclusions.--By decreasing the size of the blood collection tube for adults, we were able to markedly reduce BLLT without noting any insufficient specimen volumes.
Subject: Blood banks
Authors: Sanchez-Giron, Francisco
Alvarez-Mora, Francisco
Pub Date: 12/01/2008
Publication: Name: Archives of Pathology & Laboratory Medicine Publisher: College of American Pathologists Audience: Academic; Professional Format: Magazine/Journal Subject: Health Copyright: COPYRIGHT 2008 College of American Pathologists ISSN: 1543-2165
Issue: Date: Dec, 2008 Source Volume: 132 Source Issue: 12
Product: Product Code: 8093000 Blood Banks & Collection Centers; E121940 Adults NAICS Code: 621991 Blood and Organ Banks SIC Code: 8099 Health and allied services, not elsewhere classified
Accession Number: 230246889
Full Text: Laboratory tests increasingly are used for patient care due to the large number of tests available and the short turnaround time due to automated laboratory instruments that make them easy to perform, as well as the need for objective data to support medical decisions. For many years it has been well known that blood loss from laboratory testing (BLLT) can be significant in hospitalized adults. Early studies report an average daily BLLT of 12.4 mL in general wards and 40 to 50 mL per day in intensive care units. (1-3) In a tertiary care teaching hospital, blood collected was, on average, 45 times the amount necessary for analysis. (4) In addition, a survey of 19 community and academic hospitals identified that blood collected for laboratory testing in adult patients was 2.5 to 10 times the amount collected for pediatric patients for the same tests, despite the fact that laboratories used the same or similar laboratory instruments. (5) Blood loss due to laboratory testing or may not lead to lower hemoglobin concentration, anemia, and increased transfusion requirements. (2-4,6) This situation continues to date. (7-10) Actual laboratory instruments use small amounts of sample, often in the range of 3 to 100 [micro]L, but it is common practice to draw full large-volume tubes, regardless of tests requested and amount of sample required.

Our objective for this study was to reduce BLLT in hospitalized adults using small-volume tubes (SVTs) and to evaluate its impact in the analytical phase.


This study was performed in a hospital laboratory in a 120-bed private academic general hospital in Mexico City, Mexico.

During 2 consecutive weeks, hospital test requisitions were reviewed in order to collect patient demographics, tests requested, and number and type of tubes obtained. Emergency department patients and pediatric patients were excluded. Monovette (Sarstedt, Germany) large-volume tubes were used for adult phlebotomy with the following volumes and additives: 2.7 mL EDTA, 3 mL sodium citrate, and 4.9 mL serum separator, heparin, and no-additive tubes. The amount of blood collected was calculated by multiplying the tube capacity by the number of tubes obtained, because the standard phlebotomy practice in our hospital is to obtain full-volume tubes. Blood loss from laboratory testing per patient was the sum of all blood collected during the study period. Instruments used were Gen-S or Max-M (Beckman-Coulter, Fullerton, Calif) for complete blood cell count (CBC); Synchron LX-20, CX-9, or Access (Beckman-Coulter) for chemistry; Sysmex 1500, Sysmex 1000 (Dade Behring, Marburg, Germany), or Mini VIDAS (bioMerieux, Hazelwood, Mo) for coagulation; and ADVIA Centaur (Bayer Diagnostics, Tarrytown, NY), Axsym (Abbott Diagnostics, Abbott Park, Ill), VITROS ECi (Ortho Clinical Diagnostics, Rochester, NY), and Immage (Beckman-Coulter) for special chemistry, therapeutic drug monitoring, and infectious disease serology. Blood cultures, send-out tests, and arterial blood gas analyses were excluded from the protocol.

Small-volume tubes used in the intervention phase were 1.2 mL EDTA, 1.4 mL citrate, 2.6 and 1.1 mL serum separator, 1.1 mL heparin, and no-additive tubes. Each instrument provider was contacted to assure that instruments could handle both full and small volumes as primary tubes in the same run. Tube rack adapters for SVTs and sample probe calibration were needed for some instruments. A bar code label of the appropriate size for the SVT was implemented, and a dedicated bar code label printer for hospital test requisitions was designated. Before beginning the study, it was experimentally determined that each instrument could handle the SVT, that the bar code label was properly recognized, and that sampling and testing could be correctly performed (data not shown).

For each single test or combination of tests identified in the baseline phase, the reduced amount of whole blood, plasma, or serum sample needed for testing was calculated, as well as the type and number of tubes needed. Plasma or serum yield was calculated assuming a hematocrit of 41%. The amount of sample needed for testing was the sample amount used to perform the test plus the dead volume (amount of sample needed for an accurate instrument sampling) plus the volume for at least 1 repetition. When multiple tests were done using the same instrument with the same type of sample, only 1 dead volume was considered. Laboratory phlebotomists and ward nurses were informed of the intended use of SVTs in adult patients and were trained on the changed sample and tube requirements. They were given a pocket card with sample amount and type and number of SVT required for the most frequent tests requested. An inventory of SVTs was delivered to each ward and to the laboratory hospital phlebotomy team station.

Two weeks were allocated for familiarization with the changed sample criteria and for reinforcement on the use of SVTs. Data from hospital test requisitions were collected during the next 2 consecutive weeks as described in the baseline phase.

For statistical analysis, nonparametric descriptive statistics were used. For statistical differences, Mann-Whitney U test for 2 different samples was used with a significance level of P < .05.


Baseline results were as follows: 664 hospital test requisitions for 227 patients were received; 125 (55.1%) were from female patients, and 102 (44.9%) were from male patients; median age was 50 years (interquartile range [IQR], 36-65 years), ranging from 18 to 94 years old. The median test requisitions per patient was 1 request (IQR, 1-3), ranging from 1 to 19 requisitions. Of all patients, 52.9% had just 1 request, whereas 6.2% had between 10 and 19 requisitions. There were 345 test requisitions (52%) from critical care units, 32.9% were from the intermediate therapy unit, 11.6% were from the intensive care unit (ICU), and 7.5% were from the coronary care unit, whereas 48% were from other hospital services, including different surgical and internal medicine specialties and subspecialties. The number of blood tubes collected was 1309 for 5088 ordered tests;4123 (81.0%) were chemistry tests, 774 (15.2%) were CBC and coagulation studies, and 191 (3.8%) were immunology tests. The median number of tests per tube was 8 for serum separator, 3 for heparin tube, and 1 for EDTA, citrate, and no-additive tubes. Median BLLT per patient was 13.5 mL (IQR, 7.6-27.3), ranging from 2.7 to 159.8 mL, whereas in critical care patients the median was

19.9 mL (IQR, 12.0-35.8). Eleven patients (4.8%) had a

BLLT greater than 71.5 mL (95th centile), and 64% of those

patients were in the ICU or intermediate therapy unit. The highest BLLT (159.8 mL) was in an ICU patient. Total collected blood was 5062.4 mL, and the amount used for testing was 450.8 mL (8.9%), whereas the remaining blood after testing was 4611.6 mL (91.1%). The most frequent test requested was CBC (11%), followed by different combinations of CBC, chemistry panel 4 (glucose, blood urea nitrogen, creatinine, uric acid), serum electrolytes, calcium, phosphorus, and magnesium, which accounted for 23.2% of all tests requested.

During the intervention phase, 696 test requisitions for 246 patients were received: 130 women (52.8%) and 116 men (47.2%) with a median age of 52 years (IQR, 37-68 years; P = .45) and ranging from 18 to 95 years. The median number of requisitions per patient was 1 requisition (IQR, 1-3), the same as in the baseline phase; 54.9% of patients had 1 requisition, whereas 4.5% had 10 to 27 requisitions; 44.2% of requisitions were from critical care units, 18.7% were from the intermediate therapy unit, 13.4% were from the ICU, and 12.1% were from the coronary care unit. There were 1346 blood tubes collected for a total of 5165 tests. There was no variation in the type of tests requested, with 81.2% for chemistry, 15.6% for CBC and coagulation, and 3.3% for immunology. Median BLLT per patient greatly decreased to 3.7 mL (IQR, 2.3-7.3 mL), ranging from 1.1 to 61.9 mL, a reduction of 73% compared with baseline (P < .001). Critical care patients had a median BLLT of 5.1 mL (IQR, 2.3-10.9 mL), a reduction of

74% compared with baseline. The highest BLLT was 61.8 mL from a critical care patient. Twelve patients (4.9%) had

a BLLT greater than 25.1 mL (95th centile), and 80% of

those patients were from an ICU or intermediate therapy

unit. Total blood collected was 1715.2 mL (P < .001), and

the amount used for testing was 447.8 mL (26.1%), which was not statistically different from baseline (P = .05). The remaining blood after testing was 1267.4 mL (73.9%), statistically different from baseline (P < .001; Table).

For the impact on the analytical phase, all tests requested were performed with the SVTs as primary tubes. It was not necessary to transfer the sample to sample cups and, when needed, repeated analysis could also be performed. There was no need to redraw blood from the patient because of insufficient sample volumes.


Use of SVTs in adults to reduce BLLT was first indicated 18 years ago by Smoller et al,11 who reported a 46.8% reduction of BLLT in 41 ICU patients. In the same year, Foulke and Harlow12 reported a 33% reduction of BLLT using SVTs, reduced syringe volumes, and reporting patients' daily BLLT to physicians. Eleven years later, in a survey of 19 hospitals (7 pediatric hospitals), Hicks5 showed that it was common practice to draw an excess of blood for laboratory testing in adult patients compared with the amount obtained in children's hospitals. Blood collected was up to 10 times the volume requested for pediatric patients, despite the fact that hospitals surveyed used the same or similar equipment.5 This was further confirmed in a Q-Probes study by the College of American Pathologists on specimen collection volumes for laboratory tests. In this study, only 20 (14.4%) of 139 participant laboratories collected smaller volumes of blood for ICU specimens than for non-ICU specimens.13 Harber et al (14) demonstrated that in Australia and New Zealand no ICUs routinely used SVTs, and by implementing a highly conservative protocol that, among other factors, included the use of pediatric phlebotomy tubes, they were able to reduce the median

BLLT by more than 80% (40 vs 8 mL). There is only 1

descriptive report in Mexico on BLLT in ICU patients, where a loss of 45 mL per day and a total of 180 mL during the ICU stay has been reported, but no intervention to reduce the BLLT was done.15

Our study shows that in a general hospital ward, 50% to 55% of the patients who need laboratory testing had only 1 request, and 6.2% of patients at baseline had 10 to 19 requests per patient. Also, 4.5% in the intervention phase had 10 to 27 requests per patient. Using SVTs, the overall median BLLT in adult hospitalized patients could be reduced by 73% and by 74% in critical care patients. Maximum BLLT was reduced from 159.8 mL at baseline to 61.8 mL in the intervention phase. Patients with the highest BLLT were in the intermediate and critical care units. It is important to note that in our study, the BLLT in the ICU and the coronary care unit is highly underestimated for 2 reasons: the first because BLLT from arterial blood gas analysis was not included, because this test is not performed by the clinical laboratory. The second reason is that discarded blood for line flushing was not considered. BLLT from arterial blood gas analysis sampling comprises 19% to 34% of the total BLLT.7 Blood loss from flushing varies between 2 and 5 mL, with a median of 2.8 mL per blood draw.13 We can speculate that the real BLLT in critical care patients would be approximately 50% higher than the actual figure found in our study. However, not only critically ill patients have significant BLLT, as 20% to 36% of patients with BLLT greater than the 95th centile were from noncritical wards. The most frequently requested chemistry, hematology, and coagulation tests can be done in 1.2-, or 1.4-mL tubes, respectively. For chemistry, a 2.6-mL tube was needed only when total iron-binding capacity plus any other analyte was ordered. All other combinations of chemistry tests could be done with a 1.1mL tube. In hematology, a 2.7-mL tube was needed only when erythrocyte sedimentation rate together with a CBC was ordered. In immunology for a single analyte or thyroid profile requisition, a 1.1-mL tube was obtained; for gynecologic hormones and rheumatoid and TORCH immunoglobulin G and/or immunoglobulin M profiles, a 2.6-mL tube is needed, and a 4.9-mL tube must be obtained only when hepatitis A, B, or C panel or when 2 or more profiles were ordered.

Although the use of SVTs is a relatively easy way to reduce BLLT, laboratories willing to use this approach must confirm that their analytical instruments are able to handle SVTs as primary tubes. Adapters for tube racks and sampling probe adjustments may be required. Also, it must be experimentally determined that the use of SVTs does not affect the analytical quality of the results. The use of an appropriately sized bar code label for SVTs is needed to allow the observation of the sample characteristics and the amount of blood obtained.

The most efficient use of SVTs to reduce BLLT in hospitalized adult patients would be to restrict their routine use to the critical, coronary care, and intermediate therapy units and to have a mechanism to identify patients in general wards with a high frequency of test requisitions so that SVTs could be used instead of large-volume tubes. Implementation and compliance with such a policy would be difficult. Other mechanisms to reduce BLLT, such as reinfusing the blood used for line flushing, (13) recording and reporting to physicians daily, or cumulative BLLT,7,12 may also be considered as a complement to the use of SVTs. Upon completion of this study, the use of SVTs for laboratory testing in adult patients is in effect at our hospital.

In summary, use of SVTs in adult patients reduced the median BLLT by 73%. Patients in critical and intermediate therapy units have the highest BLLT, which can be reduced by 74%. Most frequently requested tests can be done by drawing blood in 1.1-, 1.2-, or 1.4-mL tubes. All specimens collected in an SVT were of sufficient volume and compatible with our automated instruments.


1. Henry ML, Garner WL, Fabri PJ. Iatrogenic anemia. Am J Surg. 1986;151: 362-363.

2. Smoller BR, Kruskall MS. Phlebotomy for diagnostic laboratory tests in adults. Pattern of use and effect on transfusion requirements. N Engl JMed. 1986; 314:1233-1235.

3. Tarpey J, Lawler PG. Iatrogenic anemia?: a survey of venesection in patients in the intensive therapy unit. Anaesthesia. 1990;45:396-398.

4. Dale JC, Pruett SK. Phlebotomy--a minimalist approach. Mayo Clin Proc. 1993;68:249-255.

5. Hicks JM. Excessive blood drawing for laboratory tests. N Engl J Med. 1999; 340:1690.

6. Alazia M, Colavolpe JC, Botti G, Corda N, Ramero C, Francois G. Blood loss from diagnostic laboratory tests performed in intensive care units: preliminary study. Ann Fr Anesth Reanim. 1996;15:1004-1007.

7. Wisser D, van Ackern K, Knoll E, Wisser H, Bertsch T. Blood loss from laboratory tests. Clin Chem. 2003;49:1651-1655.

8. Thavendiranathan P, Bagai A, Ebidia A, Detsky AS, Choudhry NK. Do blood tests cause anemia in hospitalized patients?: the effect of diagnostic phlebotomy on hemoglobin and hematocrit levels. J Gen Intern Med. 2005;20:520-524.

9. Wong P, Intragumtornchai T. Hospital-acquired anemia. J Med Assoc Thai. 2006;89:63-67.

10. Chant C, Wilson G, Friedrich JO. Anemia, transfusion, and phlebotomy practices in critically ill patients with prolonged ICU length of stay: a cohort study. Crit Care. 2006;10:R140

11. Smoller BR, Kruskall MS, Horowitz GL. Reducing adult phlebotomy blood loss with the use of pediatric-sized blood collection tubes. Am J Clin Pathol. 1 989; 91:701-703.

12. Foulke GE, Harlow DJ. Effective measures for reducing blood loss from diagnostic laboratory tests in intensive care unit patients. Crit Care Med. 1989;


13. Dale JC, Ruby SG. Specimen collection volumes for laboratory tests. Arch Pathol Lab Med. 2003;127:1 62-168.

14. Harber CR, Sosnowski KJ, Hegde RM. Highly conservative phlebotomyin adult intensive care. A prospective randomized controlled trial. Anaesth Intensive

Care. 2006;34:434-437.

15. Kaneko-Wada FJT, Reyes CEJ, Hernandez MF. Loss of blood volume from routine hematological tests sampling in the Military Central Hospital intensive care unit patients. Rev Asoc Mex Med Crit Ter Int. 2000;14:5-11.

Francisco Sanchez-Giron, MD; Francisco Alvarez-Mora, MD

Accepted for publication May 20, 2008.

From the Clinical Pathology Laboratory, Medica Sur, Mexico, DF, Mexico. Dr Sanchez-Giron is currently with the National Institute of Genomic Medicine, Mexico, DF, Mexico. Dr Alvarez-Mora is currently with the Blood Bank, National Institute of Neurology and Neurosurgery, Mexico, DF, Mexico.

The authors have no relevant financial interest in the products or companies described in this article.

Reprints: Francisco Sanchez-Giron, MD, National Institute of Genomic Medicine, Research Direction, Periferico Sur 4124, Torre Zafiro II Piso 6, Col Ex-Rancho de Anzaldo, Mexico, DF, 011900, Mexico (e-mail:;
Comparative Results Summary (Large- and Small-Volume Tubes) *

                                                Large-Volume Tubes

Patients, No.                                      227
Age, median (IQR), y                                50 (36-65)
Male/female, No.                                   125/102
Tubes required, No.                               1309
Requisitions, No.                                  664
Tests, No.                                        5088
Requisitions per patient, median (IQR)               1 (1-3)
Maximum requisitions/patient, No.                   19
Overall BLLT/patient, median (IQR), mL              13.5 (7.6-27.3)
Critical care BLLT/patient, median (IQR), mL        19.9 (12-35.8)
Maximum blood loss, mL                             159.8
Total blood collected, mL                         5062
Analyzed blood, mL                                 450
Remaining blood, mL                               4612

                                                Small-Volume Tubes

Patients, No.                                      246
Age, median (IQR), y                                52 (37-68)
Male/female, No.                                   130/116
Tubes required, No.                               1346
Requisitions, No.                                  696
Tests, No.                                        5165
Requisitions per patient, median (IQR)               1 (1-3)
Maximum requisitions/patient, No.                   27
Overall BLLT/patient, median (IQR), mL               3.7 (2.3-7.3)
Critical care BLLT/patient, median (IQR), mL         5.1 (2.3-10.9)
Maximum blood loss, mL                              61.8
Total blood collected, mL                         1715
Analyzed blood, mL                                 448
Remaining blood, mL                               1267

                                                     U Test, P

Patients, No.
Age, median (IQR), y                                   .45
Male/female, No.
Tubes required, No.
Requisitions, No.
Tests, No.
Requisitions per patient, median (IQR)
Maximum requisitions/patient, No.
Overall BLLT/patient, median (IQR), mL                <.001
Critical care BLLT/patient, median (IQR), mL          <.001
Maximum blood loss, mL                                <.001
Total blood collected, mL                             <.001
Analyzed blood, mL                                     .05
Remaining blood, mL                                   <.001

* IQR indicates interquartile range; BLLT, blood loss from laboratory
Gale Copyright: Copyright 2008 Gale, Cengage Learning. All rights reserved.