Alterations in HbA1c resulting from the donation of autologous blood for elective surgery in patients with diabetes mellitus

Abstract

Background

The aim of this study was to confirm the change in haemoglobin A1c consequent to pre-operative donation of autologous blood for elective surgery in patients with diabetes mellitus.

Material and methods

For enrolment in this prospective study, patients had to be scheduled for multiple autologous blood donations at different times and have a haemoglobin A1c level more than 5.8% at the first donation. The values of four factors, haemoglobin, haemoglobin A1c, glycated albumin, and glycated albumin/haemoglobin A1c ratio were determined. Changes in the values of these four factors between before and after the blood donations were calculated.

Results

In all 24 patients studied, haemoglobin and haemoglobin A1c decreased as a result of the autologous blood donations. The group with a reduced glycated albumin/haemoglobin A1c ratio had short intervals between blood donations. Correlations were observed between donation interval and change in haemoglobin A1c (r=−0.63, P=0.003), and between donation interval and change in the glycated albumin/haemoglobin A1c ratio (r=0.489, P=0.045).

Discussion

Haemoglobin A1c levels are likely to be underestimated after autologous blood donation by patients with diabetes mellitus, so glycated albumin may be a better indicator of these patients’ glycaemic control.

Introduction

The lifespan of human red blood cells (RBC) is recognised as being around 120 days in vivo ¹, and these cells are affected by glycosylation with aging. With the maturation of RBC, haemoglobin glycation becomes irreversible. Haemoglobin A1c (HbA1c) is a marker of diabetes mellitus (DM), and haemoglobin data are utilized as the monitoring index for DM control based on the notion that the HbA1c level reflects the average level of glycaemia in the preceding 1 month. In addition, HbA1c reflects the aging of RBC, since the level of glycated haemoglobin rises as RBC age. Bunn et al. showed that there is a linear increase in labelled HbA1c during the first 80 days after ⁵⁹Fe-labeled transferrin infusion². Clinically, HbA1c decreases under conditions which shorten the life-span of RBC, e.g. haemolytic anaemia, ineffective erythropoiesis, and major bleeding, which are associated with a hyperkinetic haemoglobin state. In contrast, HbA1c increases under conditions which prolong the cells’ lifespan, e.g. iron deficiency anaemia, which is associated with a hypokinetic haemoglobin state. Since autologous blood donation is considered to represent a model of artificial bleeding, the haemoglobin in RBC in this situation will be in a hyperkinetic state.

Autologous blood donation and transfusion is a strategy devised as a way of supplying blood to compensate for surgical bleeding³. Patients who are scheduled to receive autologous blood donation will inevitably include some with DM⁴. In view of the behaviour of HbA1c described above, we speculated that the HbA1c level of pre-operative patients with DM will decrease due to the blood donation and that this change in HbA1c could affect the management of DM patients. To the best of our knowledge, no studies have investigated HbA1c changes in patients undergoing a pre-operative autologous blood donation programme. We, therefore, performed this prospective study to confirm the existence of changes in HbA1c in patients with DM following the donation of autologous blood for scheduled elective surgery.

Materials and methods

Patients

This study was planned and conducted with a prospective design. The targeted patients attended the outpatient clinic for autologous blood donation of the Department of Blood Transfusion, Kobe University Hospital, between January 2009 and December 2011. The candidates had to meet the following criteria: (i) the patient had to be scheduled for more than two autologous blood donations for elective surgery; (ii) the patient’s HbA1c level had to be more than 5.8 % at the time of the first autologous blood donation, and this level was recorded as the “initial HbA1c”. This minimum value was decided in accordance with the upper limit of the Japan Diabetes Society standard (normal range: 4.3–5.8%); and (iii) the entire study had to be conducted with the patients’ informed consent and their full cooperation. With these selection criteria, 24 patients were enrolled into this study. Information concerning the patients, i.e. age, gender, and data on oral iron medication or erythropoietin administration, was confirmed by checking data in electronic medical records and by interview at the medical examination. The volume and intervals of autologous blood donations were decided by the doctor in accordance with the schedule for the surgical operation and the estimated bleeding volume. The blood volume for each donation was set between 250 mL and 400 mL (the maximum blood volume at each donation was 400 mL).

Blood sampling and tests

When an autologous blood donation was made, a sample was collected by removing 10 mL of blood at venipuncture. The first sample from each patient was obtained at the initial donation, and the second sample at the final donation (which could be the second or third donation) according to the patient’s schedule (Figure 1). Haemoglobin (Hb) count and HbA1c level were measured with a haematology analyser (XS-800i; Sysmex, Kobe, Japan), and with the Adams A1c HA-8160 (high performance liquid chromatography method) (Arkray, Kyoto, Japan), respectively. Glycated albumin (GA), which is also a useful indicator of the glycaemic status of DM patients irrespectively of haemoglobin level or erythrokinetic values, was adopted as the control index for this study⁵. GA levels were measured by an outsourcing company (SRL, Inc., Tokyo, Japan). The GA/HbA1c ratio was calculated by dividing the GA value by the HbA1c value; it is generally accepted that a normal value of this ratio is around three⁶. The values of HbA1c, GA, and the GA/HbA1c ratio at the first sampling were designated “initial HbA1c”, “initial GA”, and “initial GA/HbA1c”, respectively (Tables I and II). Changes in the values of Hb, HbA1c, GA, and GA/HbA1c between the first and second samples (i.e. 2^nd sample values - 1^st sample values) were expressed as ΔHb, ΔHbA1c, ΔGA, and Δ(GA/HbA1c), respectively. To seek relationships between changes in the GA/HbA1c ratio and other factors, i.e. age, gender, total volume of blood donation, number of donations, interval between donations, oral iron medication, erythropoietin administration, Hb, initial HbA1c, initial GA, and initial GA/HbA1c ratio, the study subjects were divided into two groups according to whether their Δ(GA/HbA1c) was ≤0 (group A) or >0 (group B).

Figure 1.

Schema of autologous blood donations and intervals between donations and blood sampling.

The black arrows show the time of blood donation and the white arrows that of the surgical operation. The horizontal axis shows time progression.

Table I.

Patients’ clinical features and relevant data at admission.

Number of patients (M/F)	24 (14/10)
Oral iron medication (n)	19
Erythropoietin administration (n)	4
Age (years)	63.5±9.4#
Total autologous blood donation, volume (mL)	477±167#
Blood volume per donation (mL)	358±48#
Interval between donations (days)	10.8±4.1#
Hb (g/dL)	13.4±1.8#
Initial HbA1c (%)	6.8±0.8#
Initial glycated albumin (%)	19.4±4.2#
GA/HbA1c ratio	2.82±0.40#

^#Data show are mean value ± standard deviation.

Hb: haemoglobin; HbA1c: haemoglobin A1c; GA: glycated albumin; M: male: F: female.

Table II.

Patients’ characteristics classified according to glycated albumin/hemoglobin A1c (GA/HbA1c) ratio.

The cohort was divided into two groups according to the degree of change in Δ(GA/HbA1c), i.e. Δ(GA/HbA1c) ≤0 (group A), and Δ (GA/HbA1c)>0 (group B).

	Group A (n =9)	Group B (n =15)	p value
Number of patients (M/F)	6/3	8/7	0.68 (a)
Oral iron medication (n)	8	11	0.70 (a)
Erythropoietin administration (n)	2	2	0.61 (a)
Age (years)#	60.3±13.4	65.3±5.8	0.42 (b)
Total autologous blood donation volume (mL)#	456±151	490±180	0.88 (b)
Patients who made three donations (n)	2	6	0.66 (a)
Interval between donations (days)#	8.8±2.7	12.0±4.4	*0.04 (c)
Hb (g/dL)#	13.1±1.4	13.6±2.0	0.54 (c)
Initial HbA1c (%)#	6.9±0.8	6.8±0.8	0.89 (c)
Initial glycated albumin (%)#	19.4±5.7	19.4±3.3	0.99 (c)
Initial GA/HbA1c ratio#	2.79±0.57	2.84±0.28	0.81 (c)

^#Data are mean value±standard deviation;

(a): Fisher’s exact test, (b): Wilcoxon’s rank sum test, (c): t-test,

^*P<0.05.

Statistical analysis

All of the statistical analyses were performed with R version 2.14 (R Foundation for Statistical Computing, Vienna, Austria). Differences between values for each item for every patient were calculated by using the paired t-test or Wilcoxon rank sum test depending on distribution. Values are expressed as mean±standard deviation. Fisher’s exact test was used for categorical variables (gender, oral iron medication, erythropoietin administration, and patients receiving three donations). Comparisons between group A and group B in terms of values of initial HbA1c, initial GA, initial GA/HbA1c ratio, Hb, and interval between donations were performed with the t-test. Correlations between the interval between donations and each of the values examined, i.e. ΔHb, ΔHbA1c, ΔGA, or Δ(GA/HbA1c), were analysed by means of Pearson’s product-moment correlation (r). To investigate which factors influenced these four parameters, linear regression analyses were performed. The parameters were also subjected to linear regression analyses adjusted for age, gender, and interval between blood collections. P values <0.05 were considered statistically significant for all analyses. P values for multiple comparisons for the correlation analysis were adjusted with Holm’s method. The relationships between interval between blood donations and each of the four values [ΔHb, ΔHbA1c, ΔGA, or Δ(GA/HbA1c)] were also analysed with the paired t-test. For this comparative analysis, the subjects were divided into two groups according to whether the donation interval was ≤10 days (n =14) or >10 days (n =10).

Results

The 24 patients investigated in this study all met the criteria for enrolment detailed in the “Patients” section. Their clinical features at the first blood sampling are shown in Table I. The mean values for ΔHb, ΔHbA1c, ΔGA, and Δ(GA/HbA1c) were −0.78±0.58, −0.22±0.32, −0.24±0.78, and 0.06±0.16, respectively. We confirmed by means of a paired t-test that the values of Hb and HbA1c were significantly reduced as a result of autologous blood donation (mean ΔHb: −0.78, mean ΔHbA1c: −0.22; P <0.001, and P <0.01, respectively). The mean values of GA and HbA1c showed no statistically significant changes.

Following the finding that the mean Δ(GA/HbA1c) index increased during autologous blood donation, we next investigated which factor(s) influenced this increase. To do this, two groups were formed on the basis of the degree of change in Δ(GA/HbA1c and compared with regards to the following factors: age, gender, total volume of donation blood, number of donations, interval between donations, oral iron medication, erythropoietin administration, Hb, initial HbA1c, initial GA, and initial GA/HbA1c. The results are shown in Table II. We found that the interval between donations in group B (12.0 days) was significantly longer than that in group A (8.8 days) (P =0.04). However, there were no other statistically significant differences in factors between the two groups.

In the next step, we investigated whether there was any correlations between donation interval and each of four factors: ΔHb, ΔHbA1c, ΔGA, or Δ(GA/HbA1c). Figure 2 shows that there was a strong correlation between donation interval and ΔHbA1c while interval and Δ(GA/HbA1c) also correlated. On the other hand, there was no correlation between interval and ΔHb, or between interval and ΔGA. Linear regression analyses adjusted for age and sex showed that ΔHbA1c and Δ(GA/HbA1c were significantly influenced by interval (β coefficient ± SE: −0.43±0.15 and −0.18±0.08, respectively). The comparison between the two groups divided according to length of interval between blood donations showed that Δ(GA/HbA1c) tended to be different, but the difference was not statistically significant (P=0.07). Finally, there were no significant difference between the two groups in terms of ΔHbA1c (P=0.10), ΔHb (P=0.60), or ΔGA (P=0.80).

Figure 2.

Correlation analysis of interval between blood donations and each of four parameters: (a) ΔHbA1c, (b) ΔHb, (c) ΔGA, and (d) Δ(GA/HbA1c ratio).

Data were analysed with Pearson’s product-moment correlation (r).

Discussion

In this study we focused on changes in HbA1c during autologous blood donation in patients with DM. Since the effect of autologous blood donation on HbA1c had not been discussed previously, we prospectively investigated changes in HbA1c resulting from autologous blood donation. We were able to show that the HbA1c level of DM patients was reduced by the donation of autologous blood. A possible explanation for this reduction can be found in the hyperdynamic state of haemoglobin as a result of blood volume loss associated with donating blood. The turnover of RBC will be accelerated, resulting in a decrease in HbA1c. Our finding is compatible with a report that HbA1c is not a satisfactory indicator of glycaemic control and the recommendation to measure GA in order to assess DM in haemorrhagic states⁶.

Interestingly, the tendency of HbA1c to decrease was enhanced by a long interval between autologous blood donations. Our result indicates that there is a marked change in HbA1c when the interval is between 14 and 20 days (Figure 2). We estimate that it takes more than 14 days for newly synthesised RBC to supply the peripheral blood after blood loss. Based on the finding that there was no correlation between the interval between donations and the degree of change in GA, the possibility that HbA1c values do not accurately reflect glycaemic control in DM patients undergoing autologous blood donation should not be ignored, especially in cases in which the interval between the donations is long. Singbartl⁷ recommends that the interval between the last autologous blood donation and scheduled surgery should be long to allow for sufficient RBC regeneration, because it was found that less than 90% of the total RBC mass deposited had been regenerated more than 4 weeks after the last blood donation. For this reason, it is possible that the HbA1c bias becomes larger around the time of the scheduled surgery. Our study found no correlation between total blood donation volume and HbA1c decrease (Table II). In theory, the production of new RBC in bone marrow resulting from the accumulated volume of donated blood may reduce the HbA1c. The results of our study support this hypothesis but, given the small number of patients investigated, should be confirmed in a larger series of patients.

One limitation of this study is that the autologous blood saved in anticoagulant citrate dextrose solution A (ACD-A) was validated until 35 days after donation. As patients can make blood donations from 34 until 4 days before the scheduled surgical operation, the longest sampling interval in this study could be 30 days. However, the sampling interval for this study was up to 20 days, because donating schedules was decided on the basis of the clinical situation. We could not, therefore analyse the changes in HbA1c more than 21 days after the start of autologous blood donation. This point warrants further investigation. On the other hand, we used GA as the marker of glycaemic control. GA is not influenced by disorders of haemoglobin metabolism, but reflects the status of glycaemic control over a shorter term than does HbA1c. GA values showed no statistically significant changes between samples from the first and second donations, which indicates that the increase in the Δ(GA/HbA1c) index was not caused by a deterioration of glycaemic control.

Some studies have reported detecting changes in the GA/HbA1c ratio in clinical settings. The HbA1c value is affected and is thus not a satisfactory indicator in cases of DM type 1⁶, haemolytic anaemia⁸, viral hepatitis⁹, or in patients undergoing haemodialysis combined with erythropoietin injections¹⁰.

Our investigation showed that the interval between donations of autologous blood correlated with Δ(GA/HbA1c), which indicates the usefulness of monitoring the GA/HbA1c ratio before or during donation (Figure 2). If the GA/HbA1c index increases during donation, the HbA1c value will be influenced by bias.

It is possible that oral iron medication and erythropoietin administration may lead to a hyperkinetic state of haemoglobin, resulting in a decrease in HbA1c. However, we could not analyse this aspect satisfactorily. Because the small sample size makes bias likely, further investigations using larger samples are needed.

In conclusion, when a DM patient makes an autologous blood donation, his or her measured HbA1c level is likely to be an underestimate. GA may be a more precise indicator than HbA1c of glycaemic control in DM patients.