Transfusion-transmitted cytomegalovirus: behaviour of cell-free virus during blood component processing. A study on the safety of labile blood components in Switzerland

Abstract

Background

Nowadays, most blood products are leukocyte-reduced. After this procedure, the residual risk for transfusion transmitted cytomegalovirus (TT-CMV) is mostly attributed to cell-free viruses in the plasma of blood donors following primary infection or viral reactivation. Here, objectives are: 1) to study the behaviour of cell-free CMV through the blood component processing; 2) to determine the anti-CMV seroprevalence, the level of viremia, the window-period in blood donor population; and 3) to identify cases of TT-CMV in bone marrow transplant (BMT) recipients.

Materials and methods

Cell-free CMV was injected into blood bags originating from regular donors. Blood components were processed according to either the CompoSelect^® or the CompoFlow^® (Fresenius Kabi AG) techniques. Samples were analysed at each step for presence of virus DNA using quantitative polymerase chain reaction (PCR). The anti-CMV seroprevalence in our donor population was taken from our donor data system. The viremia was assessed in pooled plasmas samples from routine donations by quantitative PCR. Medical charts of 165 BMT anti-CMV seronegative recipients/anti-CMV seronegative donors who received CMV-unscreened blood products were reviewed.

Results

Cell-free CMV passes without any decrease in viral load through all stages of blood processing. The anti-CMV seroprevalence was 46.13%. Four DNA positive samples out of 42,240 individual blood donations were identified (0.009%); all had low levels of viremia (range 11–255 IU/mL). No window-period donation was identified. No TT-CMV was found.

Discussion

Cell-free CMV remains a concern with current blood component processing as it passes through all the processes. However, since low levels of CMV DNA were identified in the donations tested, and no BMT recipients had TT-CMV, the residual threat of TT-CMV after leukocyte reduction appears to be very low.

INTRODUCTION

Human cytomegalovirus (CMV) is a large worldwide endemic beta herpesvirus with a 230–240 kb double-stranded DNA genome, responsible for severe opportunistic and congenital diseases. Infection in immunocompetent individuals is mostly asymptomatic or may present as non-specific viral symptoms. Classical transmission occurs through saliva, urine, genital secretions and breast milk. During primary infection, replication takes place in myeloid cells, hepatocytes, fibroblasts, and neuronal precursor among other cells. Lifelong latency is then established mainly in myeloid cells, particularly monocytes, where viral replication can take place¹. Re-infection with another CMV strain or reactivation of latent CMV leads to a new episode of viral shedding via body fluids; for example, cell-free CMV is released in the blood stream by endothelial cells².

First described in the 1960s³, transfusion transmitted cytomegalovirus (TT-CMV) can seriously affect the outcome of immunocompromised patients, who form a major group of transfusion recipients, notably after solid or BMT or during foetal life. CMV is highly prevalent in the general population, with a range of infection between 30% and 100% worldwide; prevalence is higher in older people, those with lower economic status, and women^4,5.

Before the implementation of leucocyte reduction, the risk of CMV infection in anti-CMV negative BMT recipients receiving standard blood products ranged between 28% and 57%, with 30% of the infected patients presenting life-threatening symptoms⁶. Consequently, measures such as serological testing of the donors and leucocyte reduction of the blood products were successfully adopted to reduce the risk of TT-CMV. For example, leucocyte reduction decreased the prevalence of TT-CMV to 2.4%, 0.23% and 4% respectively^7–9. A meta-analysis compared the use of anti-CMV seronegative without leucocyte reduction or leucocyte-reduced blood products without anti-CMV serostatus screening. The results showed a slightly better reduction of risk of TT-CMV using anti-CMV seronegative components, particularly for BMT recipients (93.1% vs. 92.3; p<0.05)^10–12.

At the moment, there are no standardised international guidelines regarding the best strategy to prevent TT-CMV. Window-phase donations are blood donations given during the highly viremic state following infection and before any anti-CMV antibodies are detected or before they are able to sufficiently neutralise CMV. Window-phase donations are believed to be the major cause of the residual TT-CMV infections in the post-leucocyte depletion era. Other causes of TT-CMV infection arise from blood collected during the viremic reactivation phase¹³, along with cases due to inadequate leucocyte depletion, reported to be around 0.2%¹⁴. It is now assumed that residual TT-CMV disease is mostly caused by cell-free viruses during primary infection or reactivation. In this study, therefore, we raised the question as to whether our current leucocyte reduction and production processes could also retain cell-free CMV viruses. We also looked at anti-CMV seroprevalence in the blood donors from a western region of Switzerland (canton Vaud) and for the presence of CMV DNA in around 42,000 blood donors. Finally, we actively investigated TT-CMV in the high-risk population of our BMT recipient patients. The data generated by these studies, together with other published data, will help us propose an algorithm for CMV-safe blood products in Switzerland, with the aim of harmonising practices in the currently very heterogeneous algorithms in Swiss hospitals and blood transfusion services.

MATERIALS AND METHODS

Cell-free cytomegalovirus through blood component processing

In-house quantitative cytomegalovirus polymerase chain reaction

A quantitative in-house CMV polymerase chain reaction (PCR) assay was set-up using published primers targeted to the CMV UL54 gene (61bp amplicon)¹⁵. A bovine herpes virus 1 glycoprotein B gene was co-amplified in each reaction and served as an internal control using previously published primers (97 bp amplicon)¹⁶. The CMV PCR assay was validated against semi-logarithmic dilutions of the 1^st World Health Organization (WHO) International Standard for Human Cytomegalovirus in CMV-negative human plasma (NIBSC Code 09/162, National Institute for Biological Standards and Control, Potters Bar, UK). The 95% limit of detection (LOD) was calculated by probit analysis.

Cytomegalovirus spiking

The CMV culture supernatant from CMV AD 169 strain grown on MRC-5 fibroblast cellular cultures, in a Dulbecco-MEM with glutamine, penicillin and streptomycin but without foetal calf serum (FCS), was kindly provided by Dr. M. Engels (Institute of Virology, University of Zürich, Switzerland). A 1:100 dilution of the cell-free culture supernatant in phosphate-buffered saline (PBS) was used for the spiking experiments (10 mL / spike / bag). Twenty-three blood bags were collected from both anti-CMV seropositive and seronegative blood donors (data not shown) who had been rejected from blood donation for various reasons. Twelve bags were spiked as indicated in Figure 1, and blood components were processed according to the CompoSelect^® (Fresenius Kabi, Oberdorf, Switzerand) technique (bag 45), to prepare the two blood components: erythrocyte concentrate (EC) and plasma. An additional eleven bags were spiked as illustrated in Figure 2 and prepared according to the CompoFlow^® (Fresenius Kabi) technique (bag 43), to prepare the three blood components: EC, plasma and pooled platelet concentrates. Both techniques are routinely used in Swiss blood transfusion centres.

Figure 1.

Description of the whole blood processing according to the CompoSelect^® (Fresenius Kabi) technique (bag 45), to prepare erythrocyte concentrate (EC) and plasma, with the CMV-spiking and sampling at every step (A-F)

PBS: phosphate buffered solution.

Figure 2.

Description of the whole blood processing according to the CompoFlow^® (Fresenius Kabi) technique (bag 43), to prepare erythrocyte concentrate (EC), plasma and pooled platelet concentrates, with the cytomegalovirus (CMV)-spiking and sampling at every step (1–9)

BC: buffy coat; TC: thrombocyte concentrate; erys: erythrocytes; leuk: leucocytes; thrombos: thrombocytes; Hct: haematocrit; B-CH: Swiss national guidelines for the preparation of blood products.

Sample collection and analysis

Samples were taken at each step of the two different processes, i.e. six samples for each bag 45 and nine samples for each bag 43. Samples were diluted if necessary, depending on the amount of cells, cell debris or plasma proteins (see Table I for the protocol) before total nucleic acid was extracted from 0.5 mL plasma using the QIAamp 96 Virus BioRobot kit (QIAGEN AG, Hombrechtikon, Switzerland) on a Tecan robotic platform. The total nucleic acid was eluted in an 80 μL AVE elution buffer and 10 μL was analysed in duplicate with the validated in-house quantitative CMV PCR assay. The results were expressed in cycle-threshold (Ct) values; the Ct values were compared to determine a reduction during the production of the various samples. The above described dilution and the dilution due to the addition of intersolution (to stabilise the EC) and amotosalen (for pathogen inactivation) were included in the calculations. The corrected Ct were the values calculated for the rest plasma in the samples collected (i.e. approx. 55% plasma in whole blood; 100% plasma for the plasma fraction, etc.).

Table I.

Passage of cytomegalovirus DNA through all blood processing steps

Results Bag 45 (average of all 12 bags)
	Sample	Dilution for extraction (in PBS)	PCR corrected CT (±SD)	% virus recovered
A	Baseline, not spiked; whole blood	1:2 (500 μL+500 μL)	0	0
B	Spiked; whole blood	1:2 (500 μL+500 μL)	24.4 (±1.0)	100
C	Spiked; whole blood	1:2 (500 μL+500 μL)	24.1 (±0.4)	131
D	Spiked; EC	1:2 (500 μL+500 μL)	28.1 (±0.6)	6
E	Spiked; BC waste	1:10 (100μL+900μL)	26.6 (±1.5)	21
F	Spiked; plasma	undiluted	24.1 (±0.4)	119
Results Bag 43 (average of all 11 bags)
	Sample	Dilution for extraction (in PBS)	PCR corrected Ct (±SD)	% virus recovered
1	Baseline, not spiked; whole blood	1:2 (500 μL+500 μL)	0	0
2	Spiked; whole blood	1:2 (500 μL+500 μL)	23.5 (±0.7)	100
3	Spiked; EC	1:2 (500 μL+500 μL)	28.0 (±1.0)	4.4
4	Spiked; EC (final product)	1:2 (500 μL+500 μL)	30.5 (±1.9)	0.7
5	Spiked; BC pool	1:10 (100 μL+900 μL)	22.8 (±0.7)	164
6	Spiked; TC	undiluted	22.7 (±0.9)	174
7	Spiked; TC (final product)	undiluted	25.7 (±0.9)	22
8	Spiked; plasma	undiluted	23.3 (±0.7)	113
9	Spiked; BC waste	1:10 (100 μL+900 μL)	25.7 (±0.8)	21

PCR: polymerase chain reaction; PBS: phosphate buffered saline; Ct: cycle threshold; SD: Standard Deviaiton; EC: erythrocyte concentrate; BC: buffy coat; TC: thrombocyte concentrate.

Cytomegalovirus serology status and viremia among Swiss donors and estimation of the window-period

Cytomegalovirus total antibody seroprevalence

We assessed the CMV serology status of 25,000 blood donors resident in canton Vaud. The data were collected from the medical records of the blood transfusion service, and had previously been determined for every donor using a total CMV antibody test (Enzygnost anti-CMV/IgG + IgM; Siemens Healthcare GmbH, Erlangen, Germany). As each donor can give blood four times a year and the samples were collected during three periods of three months in 2012, it was confirmed that no donor was included in the study twice to avoid duplication bias.

Ctyomegalovirus DNA detection in pooled donor plasma

Blood donations were screened with our in-house quantitative CMV real-time PCR assay in pools of up to 96 plasma samples from routine donors in several Swiss regions (Bern, Lausanne, Sion, Geneva, St. Gallen and Basel) between October 2017 and February 2018. Total nucleic acid was extracted and CMV PCR assay conducted as described above. CMV positive DNA pools were resolved down to the individual blood donation. Quantification of the positive samples was conducted with a dilution series of the WHO CMV International Standard (NIBSC code: 09/162) in CMV negative plasma. The 95% limit of detection of the PCR assay was 6 IU CMV DNA/mL. CMV IgG and IgM serology was performed on the positive samples (Abbott Architect, Wiesbaden, Germany).

Look-back for transfusion transmitted cytomegalovirus infection in allogeneic hematopoietic stem cell transplant recipients

Data were extracted from a BMT recipient cohort in Geneva Hospital. We reviewed all CMV viremia and diagnoses of all anti-CMV seronegative patients who between 2005 and 2018 had received anti-CMV seronegative BMT and CMV-unscreened leucocyte-reduced blood products.

RESULTS

Cell-free cytomegalovirus passes through all blood processing steps

Cytomegalovirus DNA was recovered on average from 119% of the 100%-spiked plasma with bag 45, and 113% of the 100%-spiked plasma in bag 43. Complete results are shown in Table I. The reduced or enhanced corrected Ct values in the various fractions are more likely to be a consequence of inaccurate estimation of the rest plasma volume than to a real reduction or enhancement; we base this primarily on the fact that the Ct value in the final plasma fraction is essentially 100% of that in the spiked whole blood before it went through the production process.

Cytomegalovirus status and viremia among Swiss donors and estimation of the window period

Cytomegalovirus IgG positive value was detected in 44.2%, 47.4% and 46.8% of the donor samples over the three 3-month sample periods, respectively. Viremia was assessed in 440 pools of 96 samples (i.e. 42,240 individual blood donations). Four CMV DNA positive samples (0.009%) were identified. Quantification of the individual CMV positive donations identified from these pools showed low levels of viral load: 225 IU/mL, 95 IU/mL, 77 IU/mL and 11 IU/mL. All four samples were CMV IgG positive; one was CMV IgM positive, 1 weakly CMV IgM positive (1.1, positive s/co ≥1) and 2 CMV IgM negative, suggesting that 2 donors had primary infection and 2 others viral reactivation or recent seroconversion¹⁷. No window-period donation could be identified in our samples. The low CMV IU/mL identified in the individual donations after pool resolution highlights the fact that further low level CMV donations may have been missed using 96-pools, suggesting pools with less donations should be considered in future studies.

Look-back for transfusion transmitted cytomegalovirus infection in allogenic hematopoietic stem cell transplant recipients

One hundred and sixty-five anti-CMV-seronegative patients were transplanted with an anti-CMV seronegative graft and transfused with leucocyte-reduced blood products not screened for CMV IgG, with a mean of 1.02 red blood cell transfusions per patient. All patients were tested for CMV infection by PCR every week during at least the first three months, and all of their results were negative.

DISCUSSION

In the post-leucocyte depletion era, residual TT-CMV is mostly attributed to cell-free viruses found in high levels in recently infected individuals and, to a limited degree, in those undergoing virus reactivations, which occurs more frequently in older donors^13,18–20. Cell-free DNA was found in 0.02–0.13% of anti-CMV seronegative blood units due to primary seroconversion^21–23, and in the plasma of at most 0.3% of the anti-CMV seropositive blood donations²⁰. This raises a question as to the utility of CMV seronegative product, as the added value in terms of safety is low while the cost of maintaining an anti-CMV negative blood supply is high. The seroconverting phase is also hazardous. For example, cell-free CMV has been found in 25–40% of healthy adolescents during primary infection, with persistence for up to 48 weeks after viral detection²⁴, while detection of CMV antibodies, generally appearing within 6–8 weeks, can be delayed by up to ten months¹⁷. Some authors, therefore, argue for using blood from donors who seroconverted for CMV IgG more than one year previously^13,21,23.

One point that has not yet been studied is the behaviour of cell-free CMV during the blood component processing after blood donation. We found that, during the processing of both the bag systems used in Switzerland, CMV was not retained in the filter and passed straight through the processing procedure with the remaining plasma fraction. No virus was retained in leucocytes or eliminated at any step of the process. This suggests that cell-free CMV, which circulates during primary CMV infections as well as during CMV reactivation, is a direct risk for transfusion. In a study on HHV-8, another cell-associated transfusion-transmitted herpes virus, plasma viremia was also not removed by filtration²⁵. However, we have no information on the infectivity of the CMV DNA positive blood units we measured in the blood products. The blood donations used in our spiking experiments came from both anti-CMV seropositive and seronegative donors. Since there was no difference in the PCR results of the various fractions, regardless of the serological status of the donor, virus-antibody complexes do not appear to affect the PCR result. In contrast, we have no proof of the integrity of the virus after processing. For this purpose, future studies should perform viral culture experiments. One specific point to mention is the difference between Ct values in samples six and seven from bag 43, i.e. before and after pathogen reduction. Assuming the rest plasma volumes are correct, there is a reduction after pathogen reduction. Such a difference between these samples was unexpected; it is likely due to the rest plasma volume estimations that are not totally correct in the two fractions and thus the virus calculations are not representative. On the other hand, PCR has been shown to be inhibited by the crosslinking of amotosalen²⁶. However, as the CMV PCR product is only 61bp in size, it is not expected that there would be a crosslink in this area. In future spiking studies, it would be interesting to analyse larger CMV PCR products.

The next step in our approach was to identify CMV DNA in plasma from regular donors in order to assess the risk of TT-CMV. The anti-CMV seroprevalence was approximately 45%, which is in line with that in similar populations described in the literature²⁷. However, only 0.009% of the 42,240 individual donations tested positive for CMV DNA. All of these were CMV IgG positive, and two were also CMV IgM positive. In our case, none of these viremic samples would have been used if anti-CMV negative products had been chosen. Viremia levels were low (95 IU/mL, 255 IU/mL, 11 IU/mL and 77 IU/mL, respectively) and infectivity of such low doses is unknown. CMV nucleic acid testing (NAT) of positive plasma from blood donors in various reports were found to give similar levels: 0.01%²⁸, 0.024%, 0.006%, 0.007% (Pichl L and Weber-Schehl M, 2019, unpublished data). Several reports in Germany suggest that window-period donations are rarely observed in these CMV DNA-positive blood products. For example, Vollmer et al.²⁸ reported that 4 out of 6 positive CMV DNA samples were also CMV IgG positive, one was CMV IgM only, and only one both IgG/IgM negative. In contrast, Ziemman et al.¹³ described 2 window-period cases out of 10 CMV DNA positive plasma samples, the rest being CMV IgG positive. In an unpublished report by Weber-Schehl et al., among their 55 CMV DNA positive samples, 53 cases were IgG positive, 2 cases were IgM positive only, and none were IgG and IgM negative. It is striking that all these rates of DNA positive samples are much lower than the 1% annual seroconversion rate reported to occur in the German donor population, with a peak in donors in the 30–35-year old age group (1.33%)^18,21,27. This might be due to the low sensitivity of the method in pooled plasma samples, or to an overestimation of the free-cell CMV viremic phase, or to minor unspecific symptoms during CMV infection preventing the donor giving blood in the real-life setting. An important point to mention is that the current level of TT-CMV is unknown. To our knowledge, none of the published studies present solid data to allow the source of CMV infection to be assessed, for example, by demonstrating that the same viral strain was found in the infected patient and in the blood bag. A thorough examination in a transfused newborn population concluded that most cases were due to breast milk contamination and none from transfusion²³. The elevated seroconversion rate in the general population does represent another possible source of contamination, although this is improbable given the drastic measures implemented to prevent infections in transplant setting. Furthermore, the leucocyte depletion techniques currently used have been improved considerably since most of the data were published. The current filtration practice yields <10⁶ leucocytes per unit; a level of reduction much greater than the first-generation filtration procedures used in many previous studies. Finally, at present, no TT-CMV has been recorded in the Swiss haemovigilance system since the introduction of leucocyte depletion technology. Of course, it is possible that there has been some under-reporting since it is difficult to identify a single blood component as the origin of the infection. Our observation is congruent with three recent studies of 145 anti-CMV seronegative stem cell recipients, many of whom had received high-risk T-depleted grafts and many transfusions with CMV-unscreened leucocyte-reduced blood without detecting any case of TT-CMV^29–31. In a similar population, Kekre et al.³² reported one CMV infection among 77 patients transfused with CMV-unscreened leucocyte-reduced blood, and 3 cases of CMV infection, 2 of which resulted in death, related to a leucocyte-reduced anti-CMV seronegative blood transfusion. Of note, in this last study, again, none of the donors nor the back-up sample from the implicated donations had been assessed for CMV seroconversion or the presence of CMV DNA.

In our stem cell graft recipients, we also did not find active CMV infection or viremia during follow up among any of the 165 anti-CMV seronegative patients transfused with CMV-unscreened leucocyte-reduced blood products.

Because of the lack of satisfactory data, international guidelines are not standardised regarding TT-CMV prevention strategies³³. Due to the small overall risk posed by TT-CMV, and the ethical and logistical concerns in using non-leucocyte-reduced products, the extremely large multicentre study that is required to finally determine the best strategy based on concrete data is unlikely to be forthcoming.

In Switzerland, the hospitals of Bern, Zürich, Neuchâtel, Jura and Fribourg currently use CMV-unscreened blood products, including for in utero transfusion. In contrast, the hospitals of Geneva, Lausanne, Basel and St. Gallen still use anti-CMV seronegative red blood cell products for special indications: mostly neonatal, foetal and pregnancy situations. In other countries, for example, France and Italy, CMV screening has been totally abandoned, whereas it is still maintained in some parts of Germany and in Australia¹⁴. Leucocyte reduction has not been universally adopted, although it has many other advantages such as, reduction of febrile inflammatory reaction and HLA-immunisation occurrence. In Switzerland, it was implemented in 1999, during the variant Creutzfeldt-Jakob epidemic. Furthermore, since 2011, all Swiss platelet units have undergone pathogen reduction treatment with amotosalen hydrochloride in conjunction with ultra violet A irradiation, a method which was shown to be effective at preventing TT-CMV³⁴.

CONCLUSIONS

The widely used strategies of both leucocyte-reduced and anti-CMV seronegative blood products for high-risk patients is very safe³⁵ but there is no evidence of clinical benefit in combining these two strategies³³. To our knowledge, no TT-CMV cases have been proven with this policy. Theoretically, the strategy only carries the risk of cell-free CMV DNA in a window period and during reactivation. But as shown in various papers, plasma CMV DNA is very rarely found during the window period^{10,13,20,28,36}, and even then, is at much lower levels than in early anti-CMV seroconverters^22,28. On the other hand, CMV DNA has been found in anti-CMV seropositive donors at a similar level as that found in the blood donations from anti-CMV-negative donors, and no cases of proven TT-CMV have been reported with the CMV-unscreened leucocyte-reduced strategy either, with solid data regarding the safety of these products, as described above. The data from the studies presented here support the current consensus that none of the current strategies to prevent TT-CMV can totally eliminate the risk of infection. Furthermore, with the high costs and challenges involved in maintaining an adequate anti-CMV-seronegative blood supply, together with the concerns surrounding the behaviour of cell-free CMV through blood component processing techniques, we suggest abandoning the CMV serology screening strategy and implementing a rapid bedside CMV NAT analysis for those few very high-risk patients, such as foetal and neonatal cases. This will require both a cost-effectiveness evaluation with determination of a CMV NAT threshold and identification of a suitable commercially available test.