This open-label, two-group, randomized study enrolled donors aged 15–65 years if they were eligible for donor screening for PBPC apheresis. All the donors received standard intravenous calcium gluconate prophylaxis. Donors were randomly assigned in a 1:1 ratio to take oral calcium (Cohort A) or placebo (Cohort B) drinks before starting apheresis. Cohort B was allowed to take oral calcium drinks if symptoms occurred. The adjustment factor for minimization was being female. This prospective analysis was conducted in accordance with the Declaration of Helsinki and approved by the Okayama University Certified Review Board. Written informed consent was obtained from all the participants.

PBPCs were harvested using the two-needle method with a continuous cell separator, Spectra Optia (Terumo BCT, Inc., Lakewood, CO, USA), and harvesting was completed using the continuous mononuclear cell collection protocol. The target cell count, ranging from 150 to 300 mL/kg, was set to be collected in one day with as much processed blood volume as possible. Cases involving PBPC collection on two consecutive days were included in the study only on the first day. Acid citrate-dextrose A (ACD-A; Terumo Corporation, Tokyo, Japan) was used as an anticoagulant. The blood-to-ACD-A ratio was initially set at 12:1, with changes to 13:1 in some cases. The packing factor was set to 4.0.

All donors received continuous intravenous administration of calcium gluconate hydrate (Calcicol, Nichi-Iko Pharmaceutical Co., Ltd., Toyama, Japan; calcium gluconate 8.5 % [wt/vol]; 0.39 mEq/mL of calcium; pH 6.0–8.2) at 22 mL/h during apheresis. Donors in Cohort A were administered 250 mL of calcium drinks (Calgen mini: Calgen Pharmaceuticals, Inc., Osaka, Japan) containing 66.3 kcal of energy, 16.3 g of carbohydrates, 6.8 mg of sodium, 19.8 mg of potassium, 130 mg of calcium, 11.5 mg of phosphorus, and 0.01 mg of vitamin B2 per 125 mL. Conversely, donors in Cohort B were administered 250 mL of placebo drinks (Calpis water: Calpis Co., Ltd., Tokyo, Japan) containing 45 kcal of energy, 11 g of carbohydrates, about 10 mg of potassium, and less than 10 mg of calcium per 100 mL. Constant monitoring for hypocalcemia symptoms was performed during PBPC apheresis. Additionally, constant monitoring of heart rate for arrhythmia was performed using electrocardiography devices. If mild paresthesia developed, all donors in Cohort B, except those who refused to drink, consumed a calcium drink. If the symptoms persisted, the infusion rate of calcium gluconate hydrate was increased to 25–30 mL/h.

Experienced doctors assessed hypocalcemic symptoms such as paresthesia, cramping, and nausea. Then the symptoms were graded using a system based on a study by Bolan et al.11 The doctor assigned a grade of (i) ‘0’ if no symptoms were noted; (ii) ‘1’ if symptoms were barely noticeable (mild paresthesia); (iii) ‘2’ if symptoms were irritating (moderate paresthesia); (iv) ‘3’ if symptoms were uncomfortable (severe paresthesia, nausea, nervousness, and anxiety); and (v) ‘4’ if symptoms were unbearable (arrhythmia, tetany, and seizures).

One mL of whole blood was collected from the upstream venous line of the calcium gluconate line before commencing PBPC apheresis, at intervals of 0, 15, 30, 45, 60, and 90 min, and at the end of apheresis. Ionized calcium levels in whole blood were measured with added heparin using Stat Profile pHOx (Nova Biomedical K.K., Tokyo, Japan) within 10 min of collection.

The primary endpoint was the incidence of Grade 1 or higher citrate-related reactions. Secondary endpoints included the incidence of Grade 3 or higher citrate-related reactions, changes in ionized calcium levels in the early period (15 and 30 min) after commencing apheresis, and safety. Changes in ionized calcium levels were calculated using the following mathematical formula:

Change in ionized calcium (%) = [(ionized calcium level at start) – (ionized calcium level at each timepoint)] / (ionized calcium level at start) × 100.

The sample size of 42 was calculated as follows: assuming a 78 % probability of Grade 1 or higher citrate-related reactions in the placebo group and a 40 % probability in the calcium drink group based on the authors’ previous report,10 an alpha error of 0.05 (two-sided) and a beta error of 0.2, the required number of participants was calculated as 40. Assuming a 5 % dropout, 42 (21 per group) participants were selected. Stata (Stata Corporation, version 16.1, College Station, TX, USA) and GraphPad Prism 8 software (GraphPad Inc., San Diego, CA, USA) were used for all statistical analyses. All p-values were two-sided, and those less than 0.05 were considered statistically significant. Means and standard deviations are used to describe normally distributed continuous variables. In comparison, medians and interquartile ranges are used to describe non-normally distributed continuous variables. Categorical variables are described as frequencies and percentages.

A time series of ionized calcium levels was drawn for each cohort and symptom, and the differences at each timepoint were tested using Welch's t-test. The differences in covariates between Cohorts A and B are represented as standardized mean differences (SMDs) and variance ratios (VRs). The absolute SMD values of less than 0.25 and VRs in a range of 0.5–2.0 were considered balanced.12,13 As this study was randomized; the covariates should be balanced in Cohorts A and B. Therefore, the outcomes between groups were compared in the analysis without adjusting for covariates (crude model). However, if they were found to be unbalanced, the inverse probability weighted regression adjustment (IPWRA) model was used to address potential confounding factors, chosen based on consensus among the investigators. Using this model introduced double robustness into the analyses.14–16 The treatment model (IPW model) predicted the probability of drinking calcium drinks. In comparison, the outcome model (regression adjustment model) used a statistical model to predict the outcome. The variables used for these predictions in both models were age, sex, height, weight, total blood volume, maximum blood flow, ACD-A injection rate, pre-apheresis CD34+ count, total processing time, the total amount of ACD-A, and total processing volume. Using a weighted outcome model, the average treatment effect estimate was calculated as the difference between the average predicted outcomes.

Between January 2021 and July 2022, 42 healthy donors (21 each in Cohorts A and B) underwent randomization. Before IPW matching, the donor characteristics were greater than 0.25 in SMDs, including age, height, weight, maximum blood flow, ACD-A injection rate, total processing time, total amount of ACD-A, and total processing volume. After using IPW matching, they were greater than 0.25 in SMDs in the total amount of ACD-A and total processing volume (Table 1).

Regarding citrate-related reactions observed in the 21 donors in Cohort A, 15 donors were asymptomatic (Grade 0), three developed Grade 1 symptoms, two developed Grade 2 symptoms, one developed Grade 3 symptoms, and none developed Grade 4 symptoms. Of the 21 donors in Cohort B, 12 were asymptomatic (Grade 0), seven developed Grade 1 symptoms, two developed Grade 2 symptoms, and none developed Grade 3 or Grade 4 symptoms (Table 2). Grade 1 or higher symptoms were 14.3 p.p. less likely in Cohort A than in Cohort B in the crude model and 9.8 p.p. less likely in the doubly robust estimation. Grade 3 symptoms were 4.8 p.p. more likely in Cohort A in the crude model and 10.5 p.p. more likely in the doubly robust estimation; however, this difference was insignificant (Table 3). Grade 4 symptoms were not observed in either group. Serious adverse events, including citrate-related reactions that required apheresis interruption or discontinuation, were not observed.

Figure 1 shows the changes in ionized calcium levels at 0, 15, 30, 45, 60, and 90 min after the start and at the end of apheresis. Ionized calcium levels decreased at 15 min compared with those at 0 min in both cohorts. They increased transiently at 30 min, then slowly declined until the apheresis ended. Changes in ionized calcium levels were calculated in the early period (at 15 and 30 min) and the maximum change during apheresis was calculated. Due to missing values, the number of donors for the early analysis period was 40. The changes at 15 min and 30 min and the maximum change during apheresis were smaller in Cohort A than in Cohort B; however, this difference was insignificant (Table 4).

Figure 1.

The transition of the level of ionized calcium in Cohorts A (calcium drinks) and B (placebo drinks) The black line indicates Cohort A. The gray line indicates Cohort B. Tinted area indicates reference ranges of ionized calcium (1.12–1.32 mmol/L). Values represent means ± standard error of the mean.

(0.07MB).

No difference in changes in ionized calcium levels was observed between Cohorts A and B. However, when all donors with and without symptoms were retrospectively compared, donors with symptoms showed decreased ionized calcium levels at 60 and 90 min (Figure 2A). Compared to those with and without symptoms in Cohort A, the ionized calcium levels of donors with symptoms declined at 60 min (Figure 2B). In Cohort B, there was no difference in ionized calcium levels between symptomatic and asymptomatic donors (Figure 2C).

Figure 2.

The transition of the level of ionized calcium between donors with and without symptoms in each cohort Cohort A took oral calcium drinks, and Cohort B took placebo drinks. (A) Combined for both Cohorts A and B. (B) Cohort A alone. (C) Cohort B alone. The black line indicates the group with symptoms. The gray line indicates the group without symptoms. Tinted area indicates reference ranges of ionized calcium (1.12–1.32 mmol/L). Values represent means ± standard error of the mean. *p-value < 0.05, **p-value < 0.01.

(0.17MB).