OBJECTIVES. The primary aims of this study were to determine the volume of blood submitted for culture in routine clinical practice and to establish the proportion of blood cultures with a blood volume inadequate for reliable detection of bacteremia.
METHODS. The volumes of blood samples submitted for culture from infants and children up to 18 years of age were measured over a 6-month period. Blood cultures were deemed adequate submissions if they contained an appropriate (age-related) volume of blood and were submitted in the correct blood culture bottle type. During the study, an educational intervention designed to increase the proportion of adequate blood culture submissions was undertaken.
RESULTS. The volume of blood submitted in 1358 blood culture bottles from 783 patients was analyzed. Of the 1067 preintervention blood cultures, 491 (46.0%) contained an adequate blood volume and only 378 (35.4%) were adequate submissions on the basis of collection into the correct blood culture bottle type. After the intervention, there were significant increases in both the proportion of blood cultures containing an adequate blood volume (186 [63.9%] of 291 cultures) and the proportion of adequate submissions (149 [51.2%] of 291 cultures). Overall, blood cultures with an adequate blood volume were more likely than those with an inadequate blood volume to yield positive blood culture results (34 [5.2%] of 655 cultures vs 14 [2.1%] of 648 cultures). Similarly, adequate blood culture submissions were more likely than inadequate submissions to yield positive blood culture results (26 [5.1%] of 506 cultures vs 22 [2.8%] of 797 cultures).
CONCLUSIONS. In routine clinical practice, a negative blood culture result is almost inevitable for a large proportion of blood cultures because of the submission of an inadequate volume of blood. Even after an educational intervention, nearly one half of blood cultures were inadequate submissions.
Blood cultures remain the standard method for detecting bacteremia in the evaluation of sick infants and children.1,2 The isolation of an organism confers many advantages, including the optimal choice and duration of antibiotic treatment.3 The exclusion of bacteremia is also important, because it may enable the cessation of antibiotic treatment and consequently reduce the length and cost of hospital stay, as well as decreasing the development of antimicrobial resistance.4,5 The ability to exclude bacteremia on the basis of a negative blood culture result depends on the sensitivity and negative predictive value of this test. Many factors influence the yield from blood cultures but the single most important factor is blood volume. Evidence from both adult6–12 and pediatric13–15 studies shows that the rate of isolation from blood cultures increases with the quantity of blood submitted. The interpretation of negative blood culture results is based on studies using adequate volumes of blood in research settings. When the volume of blood submitted for culture is inadequate, a negative blood culture result is potentially misleading in falsely excluding significant bacteremia. The aim of this study was to determine the volume of blood inoculated into blood culture bottles in a tertiary children's hospital in routine clinical practice and to determine the proportion of blood cultures with a blood volume inadequate for reliable detection of bacteremia. The effect of an educational intervention designed to increase the rate of adequate blood culture submissions was also investigated.
All blood cultures from infants and children up to 18 years of age that were received in the microbiology laboratory at the Royal Children's Hospital Melbourne during a 6-month period were included in the study. Blood cultures were drawn almost exclusively by medical or nursing (central line only) staff members. To ensure that the results represented routine clinical practice, these staff members were not informed of the study. Blood culture bottles were weighed with an Ohaus TS400D top-loading balance, coded numerically, and distributed to clinical areas for use in the normal manner. The bottles included “pediatric” (yellow, 20 mL, BacT/Alert PF/Pediatric; Organon-Teknika, Durham, NC) and “nonpediatric” (green, 30 mL, BacT/Alert FA; Organon-Teknika) bottles used for the detection of aerobic and facultative anaerobic organisms in an automated blood culture system (BacT/Alert; Organon-Teknika).16 The manufacturers recommend the use of pediatric bottles for inoculation of blood volumes up to 4 mL and nonpediatric bottles for volumes from 5 to 10 mL.16 Anaerobic blood culture bottles were not included in the study because they represent only a small proportion (<1%) of all blood cultures submitted to the laboratory.
On arrival in the microbiology laboratory, the bottles were reweighed before subsequent normal processing for culture. The final blood culture weight was determined by subtracting the initial bottle weight (adjusted for the weight of the cap) from the bottle weight on receipt in the laboratory, with an additional adjustment for weight loss attributable to storage duration before use. To determine the weight variation over time under normal storage conditions, 20 capped blood culture bottles stored at room temperature, away from direct sunlight, were weighed at regular intervals during the study period. The final blood volume was determined by dividing the final blood weight by a factor of 1.06.17
The following details were recorded for each sample: date of blood culture, patient's date of birth, and site of blood culture. The site of blood collection was classified as central or unspecified. The age of the patient at the time of blood culture was determined.
In this study, an adequate blood culture volume was defined as ≥0.5 mL for patients <1 month of age,1,18 ≥1.0 mL for patients between 1 month and 36 months of age, and ≥4.0 mL for patients ≥36 months of age.1,4,13,14,19–21 An adequate blood culture submission was deemed to have occurred when an adequate blood volume was collected into the correct bottle type. The use of a pediatric bottle was defined as correct if the volume of blood was <5.0 mL and a nonpediatric bottle if the volume was ≥4.0 mL. The use of either bottle for volumes between 4.0 and 5.0 mL was deliberately deemed correct, to take into account the fact that estimation of collected blood volume at the point of sampling is not precise.
An intervention designed to encourage increased awareness among clinical staff members about appropriate blood culture volumes and bottle types was undertaken during the study. The intervention consisted of the introduction in all blood collection areas of laminated posters indicating the blood volumes recommended for different ages, with guidance on correct bottle use.
Because the overall rate of positive blood cultures at our institution is low, the study was not designed to assess the influence of blood culture volume on subsequent isolation rates. We did, however, analyze the blood culture results to determine whether blood cultures with an adequate blood volume or an adequate submission were more likely to yield positive bacterial culture results. For this analysis, coagulase-negative staphylococci (except from neonates), diphtheroid species, Micrococcus spp, and nonpneumococcal α-hemolytic streptococci were classified as contaminants. Statistical significance testing was undertaking by using the χ2 test or Fisher's exact test for categorical data and the Mann-Whitney test for continuous data.
During the study period, 1358 blood culture bottles collected from 783 patients were included in the study. Of those patients, 618 (78.9%) had a single blood culture performed, with the rest having 2 (n = 77; 9.9%), 3 (n = 23; 2.9%), or ≥4 (n = 65; 8.3%) blood cultures. In almost all instances, these represented blood cultures undertaken on different occasions, because it is normal practice in our institution to inoculate only a single blood culture bottle. The median patient age was 3.9 years (range: 0.0–18.7 years). Of 1358 blood cultures, 133 (9.8%) were collected from patients <1 month of age, 469 (34.5%) from patients between 1 and 36 months of age, and 756 (55.7%) from patients ≥36 months of age. Pediatric blood culture bottles were used for 859 blood cultures (63.3%) and nonpediatric bottles for the rest.
Of the 1358 blood cultures, 1067 (78.6%) were collected before the intervention and the rest after. Although there was no significant difference in age distribution between the 2 groups overall (P = .06, Mann-Whitney test), there were small but significant differences between the preintervention and postintervention groups in the proportions falling within 2 of the 3 age categories, with a smaller proportion of blood cultures from patients in the <1-month group (8.9% vs 13.1%; P = .04, Fisher's exact test) and a correspondingly larger proportion of blood cultures from patients in the ≥36-month group (57.3% vs 49.8%; P = .03) in the preintervention group.
The blood volumes submitted for culture in each age category are shown in Table 1 and Fig 1. Of the 1067 preintervention blood cultures, 491 (46.0%) contained an adequate blood volume. Of those 491 blood cultures, 378 (77.0%) were deemed adequate blood culture submissions on the basis of collection into the correct blood culture bottle. Therefore, overall 378 (35.4%) of 1067 submissions were classified as adequate blood culture submissions (Table 1).
After the intervention, there were increases in all age groups in both the proportion of blood cultures with adequate blood volume and the proportion of these collected into the correct blood culture bottle. Specifically, of the 291 blood cultures taken after the intervention, 186 (63.9%) contained an adequate blood volume, of which 149 (80.1%) were collected into the correct blood culture bottle. Therefore, overall 149 (51.2%) of 291 blood cultures were considered adequate blood culture submissions after the intervention (P < .0001).
Culture results were available for 1303 (95.9%) of the 1358 blood cultures in the study (1021 [95.7%] of 1067 preintervention cultures and 282 [96.9%] of 291 postintervention cultures; P = .3). During the study period, bacteria were cultured from 96 (7.4%) of these 1303 blood cultures. Of these 96 blood cultures, 48 (50.0%) grew likely contaminants and the rest grew potentially pathogenic organisms, meaning that there were 48 (3.7%) true-positive blood cultures overall. Blood cultures with an adequate blood volume were more likely than those with an inadequate blood volume to yield positive blood culture results (34 [5.2%] of 655 cultures vs 14 [2.2%] of 648 cultures; P = .005). Similarly, adequate blood culture submissions were more likely than inadequate submissions to yield noncontaminant positive blood culture results (26 [5.1%] of 506 cultures vs 22 [2.8%] of 797 cultures; P = .03). The increased likelihood of a positive culture result with an adequate blood volume remained significant even when coagulase-negative staphylococci were excluded from the analysis (29 [4.4%] of 655 cultures vs 13 [2.0%] of 648 cultures; P = .02). Of the 1358 blood cultures overall, 169 (12.4%) were submitted with a volume of <0.5 mL. Notably, within this group, a pathogenic organism (Enterobacter cloacae) was cultured from only 1 blood culture.
Of the 859 blood cultures submitted in a pediatric blood culture bottle, 77 (8.9%) were deemed inadequate because the volume exceeded the recommended maximum of 5 mL. These were defined as inadequate because an excess volume of blood, relative to broth, may diminish the sensitivity of blood culture.22 Reanalysis of the data with these 77 blood cultures categorized as adequate did not alter the main findings of our study. In this alternative analysis, 418 (39.2%) of the 1067 preintervention blood cultures were adequate submissions and there was still a significant increase in the proportion of adequate submissions after the intervention (178 [61.2%] of 291 cultures; P < .0001). In addition, adequate blood culture submissions were still more likely overall than inadequate submissions to yield positive blood culture results (29 [5.1%] of 574 cultures vs 19 [2.6%] of 729 cultures; P = .03).
Interestingly, under normal storage conditions, there was a noteworthy loss of weight in the control, capped, blood culture bottles. The mean ± SD weight loss from the 20 bottles during the study period was 0.72 ± 0.0100 g. This weight loss was taken into account, on a pro rata basis, when the final blood volumes were calculated.
The finding of a negative blood culture frequently influences management.23 Our finding that, in routine clinical practice at a tertiary children's hospital, over half of blood cultures contained a volume of blood that was inadequate to enable a negative result to exclude bacteremia reliably has important implications. A negative result is almost invariably interpreted without regard for the volume of blood that was submitted and therefore without a true appreciation of the test's sensitivity or negative predictive value in a given patient. In many instances in our study, the blood culture submitted did not just constitute a test with decreased sensitivity but was equivalent to no meaningful test having occurred, because the blood volume submitted was too small to have a reasonable chance of leading to the detection of even high-level bacteremia. Specifically, of 1358 blood cultures, 169 (12.4%) were submitted with <0.5 mL of blood, and this proportion increased to 40 (30%) of 133 cultures for patients <1 month of age. It is worrying that a negative blood culture result is potentially misleading for up to one third of neonatal cultures. Although our study was not designed to assess the influence of blood volume on subsequent isolation rates, blood cultures submitted with an adequate blood volume were more likely to yield a positive bacterial culture than were those with an inadequate volume. Conversely, a positive culture result was obtained for only 1 of the 169 blood cultures submitted with <0.5 mL of blood.
For small infants and children, phlebotomy can be difficult and in certain situations potentially harmful.24,25 Although smaller-volume blood samples are therefore submitted from children, this is offset to an extent by the fact that the level of bacteremia is usually higher in infants and young children.7,26 In one study, the average bacterial count for children with Haemophilus influenzae bacteremia was 6000 colony-forming units (CFUs) per mL of blood and that for children with Streptococcus pneumoniae was 50 CFUs per mL of blood.19 In another study, 60% of children with significant isolates had bacterial counts of >10 CFUs per mL.20 For a series of 30 newborns with Escherichia coli bacteremia, a high bacterial count (>1000 CFUs per mL) was found in 11 cultures (31%). In that study, however, a significant proportion (23%) of cultures had colony counts between 0 and 4 CFUs per mL,26 which suggests that the level of bacteremia in small infants and children may sometimes be low and therefore more difficult to detect in small-volume blood cultures.3,23 A large study of infants from birth to 2 months of age showed that two thirds of isolates detected had colony counts of <10 CFUs per mL.3 Furthermore, in an in vitro study, 0.5 mL of adult blood spiked with pediatric pathogens was inadequate for the detection of low-level bacteremia (<4 CFUs per mL).1
Currently, there are limited data about the optimal volume for blood culture in children. Much of the available information has been extrapolated from adult studies, from which it is clear that isolation rates increase proportionally with the volume of blood submitted for culture, with optimal culture volumes of 10 to 30 mL being recommended.8,21 In one study, 17% more clinically significant isolates were detected when 13 to 16 mL of blood was cultured, in contrast to 6.5 to 8 mL.7 Several studies have confirmed the increase in detection rates that results from increased blood culture volume, reporting figures of 0.6% to 4.7% increased yield for each extra 1 mL of blood cultured.8,9,27 In the pediatric population, a study of 300 children attending an emergency department reported an increased yield with a single 6-mL blood culture, compared with 2 separate 2-mL cultures.13 In that study, up to 10 mL of blood was drawn from each individual patient, allowing a within-patient comparison of different blood volumes. In a recent study of Kenyan children, there was an increase in the blood culture isolation rate with increasing volume, from 5.6% at 1 mL to 6.8% at 2 mL and 7.9% at 3 mL.28 In a study of immunocompromised children, the introduction of a policy to increase blood culture volume led to an increase in the number of significant isolates recovered.14 In contrast, there was no difference in overall isolation rates between 0.5-mL and 1.5-mL blood cultures (13.4% vs 13.1%) in a recent study that investigated 2 blood culture procedures in a PICU.29 However, coagulase-negative staphylococci were isolated more commonly with the low-volume blood culture system, and it was not possible to determine whether they were significant or contaminant isolates.
Despite these studies, it is difficult to establish the optimal volume to detect circulating bacteria in infants and children. Although few definitions of an “adequate” blood culture volume exist, minimal volumes of 0.5 mL to 1 mL for infants18,30 and 1 to ≥30 mL for older children3,4,13,14,31 have been recommended. Our definition of adequate volume was deliberately designed to be conservative, thereby giving a minimal estimate of the rate of inadequate blood cultures. A more-stringent definition requiring a minimum of 1 mL for children <36 months of age and 6 mL for older children would have increased the proportion of blood cultures with inadequate volume before the intervention to 719 (67.3%) of 1067 cultures overall (70 [73.6%] of 95 cultures for the group of patients <1 month of age and 506 [82.8%] of 611 cultures for patients ≥36 months of age).
Our study showed that, even when the blood culture volume is adequate, the sensitivity of blood cultures is reduced by the frequent use of an incorrect bottle type. Blood culture bottles are designed for the incubation of a specific range of blood volumes. The inoculation of inappropriately large or small amounts of blood into blood culture bottles may result in decreased isolation rates as a result of altered blood/broth ratios.2,12,22
In addition, we showed that a relatively simple and reproducible intervention to educate staff members about the importance of the correct collection of blood cultures was associated with an increase in the proportion of adequate blood culture submissions.
The gradual decrease in weight of capped blood culture bottles during storage under normal conditions has not been reported previously. Evaporation of medium through the cap is the most likely explanation. By taking this weight loss into account when calculating final blood volumes, we avoided underestimating the volumes collected.
Possible limitations to our study include the inability to determine the clinical indications for the blood cultures, because frequently clinical details were not provided on request forms. Therefore, we were unable to ascertain which patients were thought to be at high risk for bacteremia or had clinical sepsis. However, several studies have shown that almost 75% of blood cultures taken from patients admitted to the ICU with a clinical diagnosis of sepsis fail to recover an organism.32,33 In addition, although we found an association between adequate submission and isolation rates, our study was not designed primarily to determine the influence of blood culture volumes on subsequent isolation rates, because of our relatively small sample size.
Despite these limitations, our study highlights the fact that blood cultures are frequently performed inadequately. We believe that the results of our study are likely to be reflected at other institutions. To our knowledge, this is the first study in a pediatric population investigating the volume of blood submitted for culture in routine practice outside a study setting. It is worrying that, for a large proportion of blood cultures, a negative result is almost inevitable because of inadequate blood volume.
We thank Dr Andrew Daley and the microbiology laboratory staff members, including Chris Pearce, for their cooperation and contribution to this study.
- Accepted January 2, 2007.
- Address correspondence to Nigel Curtis, FRCPCH, PhD, Department of Pediatrics, University of Melbourne, Royal Children's Hospital Melbourne, Flemington Rd, Parkville, VIC 3052, Australia. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Kellogg JA, Manzella JP, Bankert DA. Frequency of low-level bacteremia in children from birth to fifteen years of age. J Clin Microbiol.2000;38 :2181– 2185
- ↵Byl B, Clevenbergh P, Jacobs F, et al. Impact of infectious diseases specialists and microbiological data on the appropriateness of antimicrobial therapy for bacteremia. Clin Infect Dis.1999;29 :60– 66
- ↵Plorde JJ, Tenover FC, Carlson LG. Specimen volume versus yield in the BACTEC blood culture system. J Clin Microbiol.1985;22 :292– 295
- Li J, Plorde JJ, Carlson LG. Effects of volume and periodicity on blood cultures. J Clin Microbiol.1994;32 :2829– 2831
- Tenney JH, Reller LB, Mirrett S, Wang WL, Weinstein MP. Controlled evaluation of the volume of blood cultured in detection of bacteremia and fungemia. J Clin Microbiol.1982;15 :558– 561
- ↵Salventi JF, Davies TA, Randall EL, Whitaker S, Waters JR. Effect of blood dilution on recovery of organisms from clinical blood cultures in medium containing sodium polyanethol sulfonate. J Clin Microbiol.1979;9 :248– 252
- ↵Szymczak EG, Barr JT, Durbin WA, Goldmann DA. Evaluation of blood culture procedures in a pediatric hospital. J Clin Microbiol.1979;9 :88– 92
- ↵Thorpe TC, Wilson ML, Turner JE, et al. BacT/Alert: an automated colorimetric microbial detection system. J Clin Microbiol.1990;28 :1608– 1612
- ↵Trudnowski RJ, Rico RC. Specific gravity of blood and plasma at 4 and 37°C. Clin Chem.1974;20 :615– 616
- ↵Sullivan TD, LaScolea LJ Jr, Neter E. Relationship between the magnitude of bacteremia in children and the clinical disease. Pediatrics.1982;69 :699– 702
- ↵Auckenthaler R, Ilstrup DM, Washington JA II. Comparison of recovery of organisms from blood cultures diluted 10% (volume/volume) and 20% (volume/volume). J Clin Microbiol.1982;15 :860– 864
- ↵Buttery JP. Blood cultures in newborns and children: optimising an everyday test. Arch Dis Child Fetal Neonatal Ed.2002;87 :F25–F28
- ↵McKay RJ Jr. Diagnosis and treatment: risks of obtaining samples of venous blood in infants. Pediatrics.1966;38 :906– 908
- ↵Nexo E, Christensen NC, Olesen H. Volume of blood removed for analytical purposes during hospitalization of low-birthweight infants. Clin Chem.1981;27 :759– 761
- ↵Brown DF, Warren RE. Effect of sample volume on yield of positive blood cultures from adult patients with haematological malignancy. J Clin Pathol.1990;43 :777– 779
- ↵Neal PR, Kleiman MB, Reynolds JK, Allen SD, Lemons JA, Yu PL. Volume of blood submitted for culture from neonates. J Clin Microbiol.1986;24 :353– 356
- Copyright © 2007 by the American Academy of Pediatrics