Factor VIII and fibrinogen recovery in plasma after Theraflex methylene blue-treatment: effect of plasma source and treatment time

Abstract

Background

The quality of fresh-frozen plasma is affected by different factors. Factor VIII is sensitive to blood component storage processes and storage as well as pathogen-reduction technologies. The level of fibrinogen in plasma is not affected by the collection processes but it is affected by preparation and pathogen-reduction technologies.

Materials and methods

The quality of plasma from whole blood and apheresis donations harvested at different times and treated with a pathogen-reduction technique, methylene blue/light, was investigated, considering, in particular, fibrinogen and factor VIII levels and recovery.

Results

The mean factor VIII level after methylene blue treatment exceeded 0.5 IU/mL in all series. Factor VIII recovery varied between 78% and 89% in different series. The recovery of factor VIII was dependent on plasma source as opposed to treatment time. The interaction between the two factors was statistically significant. Mean levels of fibrinogen after methylene blue/light treatment exceeded 200 mg/dL in all arms. The level of fibrinogen after treatment correlated strongly with the level before treatment. There was a negative correlation between fibrinogen level before treatment and recovery. Pearson’s correlation coefficient between factor VIII recovery and fibrinogen recovery was 0.58.

Discussion

These results show a difference in recovery of factor VIII and fibrinogen correlated with plasma source. The recovery of both factor VIII and fibrinogen was higher in whole blood plasma than in apheresis plasma. Factor VIII and fibrinogen recovery did not appear to be correlated.

Introduction

The quality of fresh-frozen plasma is affected by numerous factors such as the source of the plasma (whole blood/apheresis), preparation time and treatment with a pathogen-reduction technique^1–8. Pathogen-reduction techniques are used to reduce the residual risk of transmission of infections after serological and nucleic acid testing⁹. Treatment with methylene blue and light (MB) is a well-known procedure for inactivating blood-borne viruses in fresh-frozen plasma.

The aim of this study was to investigate the effect of plasma source and preparation time on the quality of MB-treated fresh-frozen plasma in terms of factor VIII (FVIII) and fibrinogen levels and recovery. Moreover, we aimed to study a possible correlation between FVIII and fibrinogen recovery.

Material and methods

Plasma collection and processing

Whole blood units were collected into “top and bottom” configuration quadruple bags sets (Composelect®, Fresenius Kabi, Bad Homburg, Germany) with 63 mL of CPD anticoagulant (sodium citrate 26.3 g/L, citric acid 3.3 g/L, dextrose 25.5 g/L, sodium biphosphate 2.5 g/L, pH 5.3–5.9). The centrifuged units were processed in accordance with standard procedures to obtain red blood cell concentrate, buffy-coat and plasma (T-ACE II, Terumo BCT Europe, Zaventem, Belgium). Apheresis plasma collections were harvested in accordance with standard procedures using Autopheresis-C™ (Fenwal, Mont Saint Guibert, Belgium) equipment. A container for plasma collection was docked to the disposable set. The blood was anticoagulated with a 4% citrate solution (sodium citrate dehydrate 40g/L, pH 6.4–7.5) at a ratio of 1:16 (6%).

Plasma selection is based on gender (mainly male donors), no history of past transfusions and/or pregnancy and volume. By visual in-process control, lipaemic or haemolytic specimens were excluded from the MB treatment process. All plasma collections were treated using the MacoPharma Theraflex MB-Plasma Bag system with PLAS4 filter and Bluexflex filter (MacoPharma, Tourcoing, France). Plasma units were connected to the Theraflex kit and filtered by gravity. Anhydrous MB chloride (85 μg) in the form of a dry pill was sufficient to obtain a concentration of 1 μM based on a plasma unit having a volume of 266 mL (range, 230–318 mL). The pill, in the transfer line, dissolved in the plasma during transfer. The packs were illuminated using the MacoTronic (Macopharma) plasma illumination system. The peak emission wavelength for the lamps was 590 nm. This procedure delivers the required dose of 180 J/cm². Illuminated plasma was filtered through a Blueflex filter to eliminate at least 90% of MB and its photoresidues. The whole procedure was performed in accordance with the manufacturer’s instructions.

Plasma collected by apheresis was split into two or three units and leucodepleted with the Plasmaflex PLAS4 filter before MB-treatment.

All plasma units (MB-treated or not) were frozen to −30 °C with the same procedure within 60 minutes using a shock freezer (MBP42, Dometic, Hosingen, Luxembourg) in respect of the Council of Europe recommendations¹⁰. WB plasma collections were frozen within 18 hours of collection; apheresis plasmas within 6 hours of collection. They were stored for 2 to 20 days (mean 7 days for WB plasmas and 12 days for apheresis plasmas) at −30 °C. They were thawed in a water bath at 30 °C for 60 minutes and immediately treated by the MB procedure.

The study comprised six study arms each involving 30 plasmas (Table I). In one branch, whole blood-derived (WB) plasmas were separated and treated within 6 hours of collection (D0), separated after whole blood storage on eutectic plates at 20 °C and processed within 18 hours of collection (D1) or treated after freezing and thawing (F-T). In the other branch, apheresis plasmas (Aph) were treated within 6 hours (D0), stored on eutectic plates at 20 °C and treated within 18 hours (D1) or treated after freezing and thawing (F-T).

Table I.

Sampling design.

	WB plasma			Apheresis plasma
Study arm	A	B	C	D	E	F
Treatment time	Day 0	Day 1	Frozen-thawed	Day 0	Day 1	Frozen-thawed
Number units	30	30	30	30	30	30
Group O	10	12	0	13	13	10
Group A, B, AB	20	18	30	17	17	20
Volume before treatment (mL)	272±16	279±13	273±13	267±35	244±6	276±40

Sampling taken before and after treatment.

Plasma sampling

For each plasma unit, samples were taken just before MB-treatment and immediately after treatment. All samples were snap-frozen and stored at a temperature below −70 °C before FVIII and fibrinogen were assessed.

Fibrinogen and factor VIII measurement

The concentration of fibrinogen (factor I) was measured following Clauss’s method with thrombin reagent (Dade Behring) on a CA-7000 automated coagulation analyser (Sysmex Ltd., Milton Keynes, UK). FVIII was quantified using one-stage clotting times with Actin-FS (Dade-Behring) on a CA-1500 automated instrument (Sysmex Ltd.). Clotting factor-deficient plasmas consisted of human immune-adsorbed plasmas and were obtained from Dade-Behring.

Statistical analysis

The values were analysed using the Anderson-Darlin test for normal distribution. All results are expressed as the mean±standard deviation (SD) and range for each study arm. Comparisons of FVIII and fibrinogen recovery between the various groups were performed using analysis of variance (ANOVA) for independent groups and two factors: source (WB or apheresis) and treatment time (D0, D1, F-T). Subsequent statistical analyses were performed with Student’s t-test to uncover which groups were different. Correlations were examined using Pearson’s correlation coefficient (r). The rates of FVIII exceeding 0.5 UI/mL and fibrinogen exceeding 200 mg/dL between different groups were evaluated with Pearson’s chi-square association test. Differences and correlations were considered to be statistically significant when the P value was lower than 0.05. All analyses were performed using Minitab^® Statistical Software (State College, Pennsylvania, USA).

Results

Plasma factor VIII

Detailed results including mean SD, minimum and maximum, are presented in Table II for both FVIII and fibrinogen. The means of FVIII after MB treatment from all arms of the study exceeded 0.50 IU/mL. The percentage of plasmas in which the FVIII level exceeded 0.5 IU/mL varied between 83% for the WB and frozen-thawed plasma and 100% for the WB plasma on day 0. The Pearson’s chi-square result (0.258, p=0.879) showed no difference in the proportion of plasma with FVIII exceeding 0.50 UI/mL between the different groups.

Table II.

Plasma factor recovery in MB-treated apheresis and WB plasma.

	Whole blood			Apheresis
Study arm	WB D0	WB D1	WB F-T	Aph D0	Aph D1	Aph F-T
	FVIII (IU/mL)
Before treatment	1.20±0.30 (0.67–1.79)	0.79±0.19 (0.44–1.10)	0.81±0.16 (0.42–1.24)	1.04±0.27 (0.57–1.71)	0.98±0.23 (0.54–1.57)	1.04±0.33 (0.33–2.03)
After treatment	1.08±0.27 (0.57–1.67)	0.70±0.16 (0.37–0.96)	0.67±0.18 (0.30–1.22)	0.83±0.24 (0.49–1.37)	0.76±0.21 (0.44–1.30)	0.85±0.27 (0.29–1.66)
Recovery (%)	89±6 (76–100)	89±7 (76–99)	82±10*† (53–98)	80±10*† (54–97)	78±8*† (59–91)	83±5*†‡ (73–97)
% units >0.5 IU/mL	100	87	83	93	93	93
	Fibrinogen (mg/dL)
Before treatment	258±49 (143–400)	260±53 (172–386)	269±49 (171–372)	257±38 (184–353)	254±43 (198–411)	284±66 (171–414)
After treatment	235±40 (137–347)	231±47 (146–319)	229±48 (141–342)	219±35 (178–353)	219±35 (170–324)	244±53 (149–361)
Recovery (%)	92±4 (83–100)	89±6 (77–99)	85±7*§ (67–97)	85±7*§ (70–100)	87±5* (79–99)	86±4*§ (76–93)
% units >200 mg/dL	90	70	70	70	63	80

Values are the mean ± standard deviation (minimum-maximum) for six arms.

^*p<0.001 compared with WB D0;

^†p<0.01(WB FT) and p<0.001 (Aph) compared with WB D1;

^‡p<0.01 compared with Aph D1;

^§p<0.05 compared with WB D1.

Fibrinogen

ANOVA results showed a significant effect of treatment time (p<0.05) but no significant effect of the plasma source on fibrinogen data before treatment. The interaction between plasma source and treatment time was not statistically significant. Means of fibrinogen levels in the different groups before treatment were not statistically different with Student’s t-tests except for the comparison between apheresis plasma treated D1 and F-T. The mean levels of fibrinogen after MB treatment in all the study arms exceeded 200 mg/dL (Table II). The lowest percentage of plasmas in which the fibrinogen level exceeds 200 mg/dL was observed in the apheresis plasma treated on day 1 (63%).

The Pearson’s chi-square result (0.863, p=0.649) showed no difference in the numbers of plasmas with fibrinogen exceeding 200 mg/dL in different groups. ANOVA results of fibrinogen after treatment were not statistically different between different groups.

Factor VIII recovery

We compared the recovery of FVIII before and after MB treatment (Figure 1). Higher results were obtained with WB plasma treated on day 0 and day 1 (Table II). The comparison of FVIII recovery in the different arms by Student’s t-test revealed a statistically significant difference (p<0.001) in the recovery of FVIII between the WB plasma processed at day 0 (89±6%) or day 1 (89±7%) compared with frozen/thawed WB plasma (82±10%) and apheresis plasma on day 0 (80±10%), day 1 (78±8%) or frozen/thawed (83±5%. The lowest result was obtained in apheresis plasma treated on day 1 (78±8%). All treated plasmas of the six arms had a FVIII recovery above 50%. ANOVA results showed a significant effect of plasma source (p<0.001) but not of treatment time (p=0.218). The interaction between the two factors was statistically significant (p<0.001).

Figure 1.

Means of factor VIII recovery with confidence interval at 95%.

The FVIII concentration after MB treatment showed a strong correlation with the FVIII level before treatment (r=0.941, p<0.001). There was no correlation between the level of FVIII before treatment and FVIII recovery (r=0.013, p=0.859).

Fibrinogen recovery

As for FVIII, the mean recovery of fibrinogen in WB plasmas treated on day 0 (92±4%) and day 1 (89±6%) was higher than that in the plasmas from the others arms of this study (85–87%) (Figure 2). This difference was statistically significant (p<0.05) except for WB plasma on day 1 vs apheresis plasma on day 1 (Table II).

Figure 2.

Means of fibrinogen recovery with confidence interval at 95%.

Fibrinogen levels after treatment correlated strongly with those before treatment (r=0.939, p<0.001). In contrast to the findings for FVIII, the correlation between fibrinogen levels before treatment and recovery was negative (r= −0.195, p<0.01), implying that the loss was greater when the units contained higher pre-treatment levels.

Correlation between factor VIII and fibrinogen recovery

Pearson’s correlation coefficient was 0.58 (p<0.001) when all data were considered (Figure 3). It was above 0.50 for the three series with WB plasma and for the series with apheresis plasma on day 0 (Table III). In contrast, the correlation coefficient of the D1 and F-T apheresis plasma was very low (r= −0.23 and 0.21).

Figure 3.

Correlation between fibrinogen and factor VIII recovery.

Table III.

Pearson’s product-moment correlation between FVIII and fibrinogen recovery in the six arms.

Study arm	WB D0	WB D1	WB F-T	Aph D0	Aph D1	Aph F-T
Pearson’s correlation coefficient	0.59	0.80	0.89	0.57	−0.23	0.12
P value	<0.001	<0.001	<0.001	<0.001	0.212	0.543

Discussion

It is well-known that FVIII is sensitive to blood component storage conditions, pathogen-reduction technology processes and blood groups^11,12. For this reason we compared the recovery of FVIII before and after MB treatment. The percentage of plasmas in which the FVIII level after MB treatment exceeded 0.5 IU/mL was the highest in the WB plasma on day 0 (100%).

The mean recovery of FVIII after treatment with MB in the different series was above 80%, except in the apheresis day 1 arm (78%). These results are better than those obtained in the initial validation in our centre which was performed in 2003 (data not shown) (89% vs 83% for the WB plasma treated on day 1 and 83% vs 81% for the series with WB frozen/thawed plasma). This can be explained by the improved skills of technicians over years. The results of this study are notably better than those published by Rock⁸ (67%), Hornsey et al.¹³ (75%) and Garwood et al.⁴ (median between 76% and 71%). However, they are comparable to those published by Moog et al.¹⁴ (85%) and Politis et al.¹⁵ (82%). Mean FVIII recovery in WB plasma treated on day 0 and day 1 (89%) was higher than that in the plasma in the other arms of this study.

The difference between the WB plasma on day 0 and frozen-thawed plasma was also noted by Garwood et al.⁴, who found an 8% reduction due to freeze/thawing. Unfortunately, the authors were not able to identify a root cause. Zeiler et al.¹⁶ observed only minimal loss of FVIII and fibrinogen due to freezing and thawing. The difference observed between WB plasma on day 0 and day 1 on the one hand and apheresis plasma on day 0 on the other hand could be explained by the use of different anticoagulants. Preston¹ and Myllylä² noted that the type of anticoagulant and the amount of citrate were factors that influenced plasma quality. Preston¹ pointed out the importance of pH on the stability of factor VIII. Carlebjörk et al.¹⁷ observed the difference in activity and stability of FVIII in blood collected into different anticoagulants (CPD-A and ACD vs citrate). The higher pH in blood collected with citrate might give rise to lower activity and instability. Back in 1965, Weiss¹⁸ had demonstrated that FVIII was less stable at pH values higher than 7.3. Prowse et al.¹⁹ demonstrated that collecting WB into half the usual amount of citrate (0.5 CPD-A) greatly improved the stability of FVIII in plasma or WB kept at room temperature. At the same time, Rock et al.²⁰ postulated that the use of anticoagulants with lower levels of citrate was an advantage to the generation of FVIII. Beck et al.³ confirmed that reduced citrate concentrations in apheresis collections resulted in better preservation of FVIII activity. In our apheresis collections, we currently use the citrate concentration recommended by Beck et al. In our study, plasmas were collected into two different anticoagulants, CPD (citrate concentration 105 mM) for WB collection and citrate 4% (citrate concentration 136 mM) for apheresis collection. The higher pH and citrate concentration of anticoagulant for the apheresis collections could explain the lower recovery of FVIII for apheresis plasma. The routine quality control data from previous years tends to support this hypothesis.

The level of fibrinogen in the plasma before treatment is not affected by the preparation processes, unlike FVIII which is sensitive to different processes and blood groups. Even after treatment, fibrinogen is imperceptibly affected by the plasma preparation processes. The percentage of plasmas in which the fibrinogen level exceeded 200 mg/dL varied from 63% for the apheresis plasma treated on day 1 to 90% for the WB plasma treated on day 0. On the other hand, we observe a significant difference in the recovery of fibrinogen between the WB plasma processed on day 0 or 1 compared with frozen/thawed WB plasma and apheresis plasma on day 0 or frozen/thawed.

The average recovery of fibrinogen after MB-treatment was above 80%. These results were higher than those of the initial validation conducted in our centre in 2003 (data not shown) (89% vs 78% for the WB plasma on day 1 and 85% vs 76% for WB frozen/thawed plasmas). The results of our study are significantly better than those published by Rock⁸ (65%) and Garwood et al.⁴ (median 79% and 72%). They are comparable to those published by Moog et al.¹⁴ (85%) and Politis et al.¹⁵ (82%). We confirm the results of Politis et al.¹⁵ who described a negative correlation between data before treatment and recovery of fibrinogen implying that the loss was greater when the units had higher pre-treatment levels. The reason for this is unknown. The MB/light-induced reduction of functional fibrinogen, measured by the Clauss method, has been explained previously by a modification of histidine residues²¹. It can be assumed that there might be an optimal relation of total amount or concentration of MB and fibrinogen which results in a maximum effect. Interestingly, the lowest fibrinogen concentrations after treatment occurred in the plasmas with the lowest starting volume (i.e. highest MB concentration because of the addition of a constant amount of MB [85 μg] per unit).

The Pearson’s correlation coefficients between the recovery of fibrinogen and FVIII exceeded 0.50 for the three groups of WB plasma and for the apheresis plasma on day 0. Conversely, the correlation coefficient for frozen/thawed apheresis plasma was very low (0.12).

In summary, these results show a difference in recovery of FVIII and fibrinogen in relation to the source of plasma. Recovery levels of both factors were higher in WB plasma than in apheresis plasma. This observation could be related to the anticoagulant composition or ratio.

Both studied factors affect the quality of fresh-frozen plasma. FVIII recovery is influenced by both treatment time and the pathogen-reduction process. FVIII recovery is indeed representative of the preparation process as Lawrie et al.²² suggested. The level of fibrinogen in fresh-frozen plasma is a clinically significant factor²³. Nevertheless it has been shown previously that the thrombin generation of MB-treated plasma is only slightly²⁴ or not²⁵ changed.

FVIII and fibrinogen recovery did not appear to be correlated. Politis et al.¹⁵ made the same observation. The combined effects of the preparation and MB-treatment seem to be coagulation factor-dependent.

The results of the FVIII recovery comply with Belgian legislation (recovery >50%). All treated plasmas in the six arms had a recovery greater than 50%. Loss of coagulation factor activity was within the internationally accepted range. All means of FVIII after treatment in the six arms complied with Council of Europe recommendations¹⁰ (average exceeding 0.5 IU/mL).

In our routine production, 90–95% of MB-treated plasma comes from WB processed within 18 hours of collection and only 5 to 10% of MB-treated plasmas come from apheresis, frozen-thawed collection with Autopheresis-C mainly of AB, A or B MB-treated stock. This production strategy gives us confidence both in the FVIII and fibrinogen content of our MB-treated fresh-frozen plasma.