Printer Friendly

Highly conservative phlebotomy in adult intensive care--a prospective randomized controlled trial.


Anaemia in critically ill patients is common and phlebotomy associated blood loss may contribute towards this anaemia. The aims of this study were twofold. Firstly, a survey was conducted to provide a summary of current phlebotomy practices within Australian intensive care units. A standardized telephone survey was aimed at Australian intensive care units registered with Australia and New Zealand Intensive Care Society (ANZICS) and questions regarding phlebotomy procedures directed at nursing staff. Secondly, a prospective randomized controlled trial aimed to assess the impact of a highly conservative phlebotomy procedure on haemoglobin concentration in intensive care patients.

Patients admitted to our own intensive care unit were randomized using a sealed envelope technique to either a highly conservative phlebotomy group, or standardized controls. Blood was taken according to strict protocols and recorded along with haemoglobin concentration daily.

The survey demonstrated that 16% of Australian units return deadspace volumes from in-line arterial sets and no unit routinely used paediatric-sized blood collection tubes. Using our highly conservative protocol, median phlebotomy-associated blood loss was reduced by over 80% (40 ml vs 8 ml P<0.001). Mean haemoglobin fell from 13.7 g/dl to 11.7 g/dl in controls (P=0.002) and from 12.7 g/dl to 11.5 g/dl (P=0.074) in our study group. We conclude that highly conservative phlebotomy is feasible in a critical care unit and is associated with a reduction in blood loss.

Key Words: anaemia, intensive care, phlebotomy, haemoglobin, transfusion, blood wastage

Anaemia is common to critically ill patients during intensive care admission (1). Possible causes include sepsis, failure of erythropoiesis, iron deficiency and overt or covert blood loss. In neonatal intensive care units, phlebotomy-associated blood loss (PBL) is the leading cause of anaemia in pre-term and low birth-weight infants (2). It is possible that blood loss from routine phlebotomy contributes to this anaemia in critically ill adults. It has been shown that volumes of blood reaching laboratories are hugely in excess of the actual amount necessary to perform each test (3). Attempts to reduce this would seem a logical strategy to minimize PBL.

PBL in the intensive care unit (ICU) is higher than in ward-based patients and higher still in patients with in-dwelling arterial cannulae, mean PBL being in the region 40-50 ml per day with total PBL per patient occasionally exceeding 1000 ml (4). It has been shown that it is technically possible to perform laboratory tests using small volume (paediatric) collection tubes in adults (5).

The routine charting of cumulative daily PBL can reduce the number of tests ordered by medical staff (6). Other methods to reduce PBL include staff education (7) and blood-conserving arterial line systems (8), both leading to around 50% less blood wastage. The blood withdrawn to clear the arterial line before sampling is often discarded. A survey performed in the United Kingdom found that two-thirds of neonatal units routinely return this blood volume to patients (9). In contrast, only 8% of adult units followed suit. This survey also showed that just 14% of adult intensive care units used paediatric collection tubes for their standard laboratory requests. We are not aware of any published data regarding this practice in Australia to date. To address these issues we conducted a survey of phlebotomy practice in Australian intensive care units and a single centre prospective randomized controlled trial to determine the impact of a highly conservative phlebotomy protocol on PBL in adult critically ill patients.


Survey of Australian Practice

A telephone survey was conducted throughout the Australian adult intensive care units registered with the Australian and New Zealand Intensive Care Society (ANZICS). A standardized questionnaire comprised five simple questions; are paediatric-sized blood tubes routinely used in adults, are dead space volumes from arterial lines returned to the patient, what are the volumes of deadspace discarded, how frequently blood is taken and whether phlebotomy blood loss is charted. The anonymity of individual units was protected.

Randomized controlled study

This study was performed in the Intensive Care Unit of the Logan Hospital, Queensland, Australia. This is a metropolitan secondary referral centre on the outskirts of the state capital, Brisbane, admitting around 400 patients per year. Forty-nine consecutively admitted patients were studied over a four-month period from August to November 2003. Patients were randomized (using a sealed envelope technique) at insertion of infra-arterial cannulae to either a highly conservative phlebotomy protocol (Group CONSERV) or a standardized control group (Group CONTROL). Exclusion criteria included patients with overt haemorrhage, those who immediately prior to their ICU admission received blood transfusion and patients in whom an arterial cannula was inserted after day one of their ICU stay. Jehovah's Witnesses were not included, standard practice in our unit being highly conservative for this group of patients. Regional ethical approval was obtained prior to patient recruitment. The review board waived the need for informed written consent.

Blood was withdrawn by nursing staff in accordance with strict protocols and all volumes charted on cumulative flow sheets. The standard discard volume used was 5 ml in both groups and returned to patients in group CONSERV Blood was aseptically withdrawn and returned via the central venous line or a peripheral cannula. In the group CONTROL this volume was discarded and charted as PBL. Daily bloods included full blood count (FBC), urea and electrolytes, glucose and liver function tests (CHEM) and coagulation studies (COAG). Samples were sent prior to the ward round and subsequent tests were requested on serum already in our laboratory. Patients whose infra-arterial cannulae were removed prior to their ICU discharge had blood withdrawn and replaced using the central and peripheral venous lines in the same manner.

Volumes for tests in group CONSERV were 0.5 ml (FBC), 1.0 ml (CHEM) and 1.6 ml (COAL). All sample tubes used in this group were our standard paediatric sized containers (Microtainer[R], Becton Dickinson, U.S.A. Microtubes[R], Sarstedt, Germany). CONTROL samples were collected in our standard evacuated adult tubes (Vacuette[R], Interpath, Australia). Volumes used were 4.0 ml (FBC), 5.0ml (CHEM), 3.5 ml (COAG) and 5.0 ml clotted samples for cardiac markers and drug levels.

Blood culture volumes were 10 ml (5 ml aerobic, 5 ml anaerobic) in both groups due to ethical committee concerns regarding yield. Arterial blood gas samples (ABGs) were collected (1 ml) in pre-heparinized syringes in each group. Blood glucose analysis was performed on ABGs. The incidence of re-testing (due to insufficient serum availability) was charted.

The Acute Physiology and Chronic Health Evaluation (APACHE II) score during the first 24 hours was recorded for each patient. The haemoglobin (Hb) was recorded on admission to and at discharge from the ICU. PBL was measured including discarded volumes where appropriate and this figure expressed per day and per ICU stay. Transfusion was at the discretion of the ICU consultant who was not blinded to patient study group. The number of units of packed cells transfused was recorded. The 28-day mortality and length of ICU stay was compared between groups.

Statistical analysis

A computer software package (SAS version 8.02) was used for statistical analysis. Statistical comparisons of primary outcomes were performed using a Wilcoxon 2-sample test for non-parametric data and a two-tailed paired t-test for parametric data. Two-way analysis of variance was used to compare population characteristics between groups.



All 84 of the adult Australian units registered with ANZICS at the time of the study participated in the telephone survey. No unit routinely used small volume (paediatric) collection containers for phlebotomy. Only 16% of units returned otherwise discarded blood back to the patient and in all cases this was via a specially formulated in-line transducer set. The median discard volume was 4 ml (range 2-10 ml). Thirteen units (15.5%) routinely sampled blood twice daily or more. No unit recorded PBL on the ICU flowsheet.

Randomized controlled study

Two Jehovah's Witnesses were excluded according to pre-agreed criteria. One patient (group CONSERV) was excluded due to severe haemorrhage while on continuous veno-venous haemofiltration (CVVH). Baseline characteristics were similar in both groups (Table 1) and median length of stay identical. Presenting diagnoses were too wide-ranging for meaningful comparison.

CONSERV median daily PBL was 8 ml (range 7-10 ml) and cumulative PBL 25 ml (range 14-33 ml). In contrast, daily and total PBL in group CONTROL was 40 ml (range 28-43 ml) and 141 ml (range 80-202 ml) respectively (P<0.001, Figure 1). Mean haemoglobin fell from 13.7 g/dl to 11.7 g/dl in CONTROLS (reduction of 2.0 g/dl, P=0.002) and from 12.7 g/dl to 11.5 g/dl (reduction of 1.3 g/dl, P=0.074) in CONSERV


Absolute numbers of tests per patient per day were not significantly different in either group (P=0.42) and on no occasion was there insufficient serum available for retrospective tests. On two occasions blood in paediatric-sized COAG containers clotted due to inadequate mixing. Two patients in CONSERV were transfused. One developed disseminated intravascular coagulation secondary to sepsis and received three standard packed cell units. The other patient received two packed cell units (PCU) postoperatively after a haemoglobin concentration of 8.6 g/dl. Three patients in group CONTROL were transfused; two long-stay patients (12 and 18 days, with haemoglobin concentrations of 6.8 and 7.3 g/dl), received two and one PCU respectively and one patient with urosepsis and ischaemic heart disease (haemoglobin concentration 9.3 g/dl), received one PCU. Mortality was not significantly different in either group. APACHE II scores were positively correlated with daily PBL (P<0.001) and multiple linear regression showed a mean difference in daily PBL between groups of 28 ml per day after adjustment for APACHE 11 score. No other significant correlation was found between baseline and outcome variables.


Phlebotomy for diagnostic laboratory tests is an important part of daily care of intensive care patients. It is clinically possible to undertake all routine blood tests on paediatric-sized samples. We have shown that conservative phlebotomy can reduce PBL by over 80%. There appears to be a small but statistically significant difference in haemoglobin reduction, although our study was not powered to detect this. This difference may contribute to the prevention of the decrease in haemoglobin that often accompanies critical illness.

The size and staffing arrangements of our unit prevented a double-blind methodology. We attempted to minimize bias with the performance of a standard panel of daily tests. Extra samples were requested retrospectively where possible. Our exclusion criteria selected out a sub-population (patients deteriorating slowly prior to the insertion of invasive monitoring) that may have introduced error if large numbers were studied although the effect of this is unlikely to be significant in our data.

Paediatric-sized COAG containers were very small and on two occasions blood clotted, probably due to inadequate mixing. This may have arisen due to the surface tension of liquid anticoagulant (citrate) within small diameter containers and did not occur again after staff education regarding proper inversion of samples. Paediatric tubes are between two to three times more costly and laboratory processing more time-consuming, due to lack of automated facilities for low-volume samples. Each sample must be manually withdrawn and introduced into the appropriate analyser, a process that takes around twice the time compared with an adult sample. This may not currently be cost-effective if employed on a large scale and has infection risks (to both nursing and laboratory staff) associated with excessive manual handling of human blood. There is a risk of line-to-line bacterial contamination when replacing blood via the central line and perhaps it would be prudent to return via the line from which it was drawn. Air embolism should be avoided by careful inspection of the blood-filled syringe before injection.

A unit of packed red cells approximately equates to 450 ml whole blood volume (10). If our highly conservative strategy were applied hypothetically (assuming no physiological variables or natural haemolysis) it would take 56 days to lose this red cell mass, compared to just 18 days in our controls. The median length of stay in our study was three days and this may have accounted for the small haemoglobin difference that we found. Patients rarely stay in intensive care for more than 18 days and, although not universally accepted, lower transfusion triggers have probably influenced transfusion rates in critical illness. It would require a much larger study to detect any clinically significant difference in transfusion requirements that could be attributed to phlebotomy volumes alone.


Phlebotomy associated blood loss can be reduced using precise, patient-tailored, transfusion thresholds (11), staff education, strict departmental procedures and blood-conserving arterial lines. In this population of critically ill patients we have demonstrated a highly conservative approach to phlebotomy can reduce PBL. Our data suggest only patients with an ICU length of stay of greater than 18 days would have a reduced need for allogenic blood transfusion. This, together with the additional costs incurred may limit the widespread application of this approach. Further research to determine if this approach leads to clinically significant reductions in transfusion requirements is warranted.


We would like to thank Queensland Health Pathology staff working at Logan Hospital for their work in processing laboratory samples.

Accepted for publication on May 10, 2006.


(1.) Vincent JL, Baron JF, Reinhart K et al. Anaemia and blood transfusion in critically ill patients. JAMA 2002; 288:1499-1507.

(2.) Lin JC, Strauss RG, Kulhavy Jcet al. Phlebotomy overdraw in the neonatal intensive care nursery. Pediatrics 2000; 106:E19.

(3.) Dale JC, Ruby SG. Specimen collection volumes for laboratory tests, a college of American pathologists study of 140 laboratories. Arch Pathol Lab Med 2003; 127:162-168.

(4.) Smoller BR, Kruskall MS. Phlebotomy for diagnostic laboratory tests in adults. Pattern use and effect on transfusion requirements. N Engl J Med 1986; 314:1233-1235.

(5.) Smoller BR, Kruskall MS, Horowitz GL. Reducing adult phlebotomy blood loss with the use of pediatric-sized blood collection tubes. Am J Clin Path 1988; 91:701-703.

(6.) Foulke GE, Harlow DJ. Effective measures for reducing blood loss from diagnostic laboratory tests in intensive care unit patients. Crit Care Med 1989; 17:1143-1145.

(7.) Henry ML, Garner WL, Fabri PJ. Iatrogenic anaemia. Am J Surg 1986; 151:362-363.

(8.) Silver MJ, Li YH, Gragg LA, Jubran F, Stoller JK. Reduction of blood loss from diagnostic sampling in critically ill patients using a blood conserving arterial line system. Chest 1993; 104:1711-1715.

(9.) O'Hare D, Chilvers RJ. Arterial blood sampling practices in intensive care units in England and Wales. Anaesthesia 2001; 56:568-584.

(10.) Australian Red Cross Blood Service Website 2004. Available online at: (Accessed August 2005).

(11.) Herbert PC, Wells G, Blajchman MA et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. N Engl J Med 1999; 340:409-417.

C. R. HARBER *, K. J. SOSNOWSKI ([dagger]), R. M. HEGDE ([double dagger])

Logan Hospital Intensive Care Unit, Brisbane, Queensland, Australia

* B.Med.Sci., B.M., B.S., Senior House Officer.

([dagger]) R.N., B.N., Grad.Dip.(Crit.Care), Clinical Nurse.

([double dagger]) M.B., B.S., M.D., E.D.I.C.M., F.J.F.I.C.M., Clinical Director.

Address for reprints: Dr C. R. Harber, 92 Abbey Road, West Bridgford, Nottingham NG2 5NB, United Kingdom.
Baseline characteristics

 group CONSERV group CONTROL

Females 12 (50%) 15 (60%)

Age in years

Mean (SD) 60 (15.7) 54 (20.6)

APACHE-II scores
Median (LQ-UQ) 17 (13-22) 18 (12-24)

Length of stay
days (LQ-UQ) 3 (3-4.8) 3 (2-4)

Baseline Hb
concentration in g/dl

Mean (Sd) 12.8 (1.9) 13.7 (2.1)
COPYRIGHT 2006 Australian Society of Anaesthetists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2006 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Harber, C.R.; Sosnowski, K.J.; Hegde, R.M.
Publication:Anaesthesia and Intensive Care
Date:Aug 1, 2006
Previous Article:Forceps dilatational percutaneous tracheostomy: safe and short.
Next Article:Anaesthesia-related mortality in developing countries.

Related Articles
Optimal diabetes care poses long-term challenge. (Follow-Up After Major Clinical Trial).
New study: psychotherapy shows mixed survival effects in metastatic cancer.
Self-monitoring flops for some diabetics.
Smoking cessation intervention for inpatients pays off.
Effect of heparin-bonded central venous catheters on the incidence of catheter-related thrombosis and infection in children and adults.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters