Sample size and unit selection
The study took place from February to April in 2007 in Mombasa, Kenya. Three units of AD-WB from each of one static and three mobile donation sessions were provided by the Regional Blood Transfusion Centre (RBTC; total of 12 units). No instruction was given about which units from a particular session were to be entered into the study, although it was known that the study was looking at quality and storage issues. Twenty-four consecutive umbilical cord whole blood (UC-WB) donations with a minimum volume of 40 mL were used in the study.
Collection technique and storage conditions
The RBTC staff collected AD-WB units according to their standard operating procedures: blood donations were collected into single bags designed for a standard 450 mL (±10%) collection and prefilled with 63 mL of CPDA-1 (Medibag; Eastern Medikit Ltd, India); predonation screening of donor hemoglobin (Hb) was performed using the copper sulfate method (specific gravity 1.053, corresponding to a Hb of 12.5 g/dL), and the volume of blood collected was monitored at the time of collection using a spring balance.
UC-WB was collected after delivery of the placenta from consenting mothers delivering at term on the labor ward of Coast Provincial General Hospital, Mombasa. The placenta was placed fetal side down on the absorbent surface of an incontinence pad, which was clipped over a metal hoop. The hoop was attached to a metal rod like a retort stand and the clamped umbilical cord passed through a hole cut into the middle of the incontinence pad. The hoop was raised up (and with it the placenta) and clamped higher up the stand such that the cord hung down with its full length suspended and the umbilical cord vein filled with blood by gravity.
The entire cord was disinfected twice with 70% isopropyl alcohol and then the intended venipuncture site, which is at the distal end of the cord just above where it is clamped, swabbed with 2% povidone iodine tincture. Cord blood was collected by a single venipuncture of the umbilical vein with the 16-gauge needle of a 250-mL single blood collection bag containing 21 mL of CPDA-1 (Macopharma, Twickenham, UK) and drainage of the cord blood into the bag by gravity. Universal precautions and attention to asepsis were observed throughout the procedure.
Screening of all units for human immunodeficiency virus, hepatitis B and C, and syphilis was performed by the RBTC laboratory using their standard methods. In the case of cord blood, a maternal sample taken around the time of delivery was tested. AD-WB and UC-WB units were stored in the same refrigerator at a temperature of 2 to 6°C for 35 days. In the event of main power supply failure, the refrigerator was connected to the hospital diesel generator. The refrigerator temperature was monitored manually twice a day.
Sampling of blood packs
All blood packs were sampled on Days 2, 7, 14, 21, 28, and 35 after collection (day of collection is Day 1). They were removed from the refrigerator and rocked gently by hand in a plastic tray for 10 minutes to resuspend the cellular components. To reduce the risk of bacterial contamination, which can cause hemolysis, blood packs were sampled on Days 2, 7, 14, and 21 in the following manner: the blood in the pilot tube was stripped back into the main pack three times using a tube stripper, the tubing was clipped twice 15 cm from the distal end of the tube and then cut between the two clips, the two cut ends were rinsed with 70% methanol and the blood pack was placed back in the refrigerator, and the 1 to 2 mL of blood in the removed tube segment was then transferred into a plain test tube.
On Day 28, when pilot tube lengths were too short for the sampling method described above, a sterile spiked sampling coupler (Baxter, Newbury, UK) was inserted aseptically in a laminar flow hood into one of the ports on each pack, and a sample was withdrawn through this. The coupler was left in place and the Day 35 samples were obtained in the same manner.
Unit weights were established by weighing on a tared scale before sampling on Day 2. Hemograms were performed on all samples with an automated cell counter (Nihon Kohden Corp., Shinjuku-ku, Japan). Samples were centrifuged at 3000 × g for 5 minutes (Heraeus Instruments, Hanau, Germany) and the supernatant was transferred with a pipette to a different tube, which was centrifuged at 3000 × g for an additional 5 minutes. Supernatant (plasma) Hb was assayed with a plasma/low Hb photometer (Hemocue, Angelholm, Sweden) and then the remainder of the supernatant was stored at −20°C. The supernatant (plasma) potassium concentration was assayed at a later date on the thawed plasma samples (Instrumentation Laboratory, Lexington, MA).
Calculated values and statistical analysis
Donation volumes were estimated by subtracting the CPDA-1 volume from the unit weight and multiplying by 1.06. 12
Supernatant Hb (g/dL), total Hb (g/dL), and hematocrit (Hct) were used to calculate percentage hemolysis according to the formula
To adjust for the dilution effect of excess CPDA-1 in smaller donations, supernatant potassium concentrations (supernatant K, mmol/L) were standardized by donation volume (L) by estimating the total potassium and dividing by donation volume as follows:
All data were entered into an electronic database and analyzed using statistical software (Stata v9.2, StataCorp, College Station, TX). AD-WB variables were assumed to have a normal distribution and summarized by the mean and standard deviation (SD). The 24 UC-WB donations were ranked according to donation volume and allocated to six equal groups (). Nonparametric statistics were used to summarize these data and to test for the significance of observed differences between cord blood groups (Kruskal-Wallis) and between adult-donated and cord blood (Wilcoxon rank sum). Binary data were expressed as proportions and observed differences compared for statistical significance with the chi-squared test of association.
Baseline characteristics of 24 UC-WB donations with stratification by donation volume (see text)