Characterization of HBV DNA+/HBsAg blood donors in Poland identified by triplex NAT


  • Potential conflict of interest: Nothing to report.


Nucleic acid testing (NAT) for hepatitis B virus (HBV) has been performed in Poland since 2005 on samples seronegative for hepatitis B surface antigen (HBsAg), anti–hepatitis C virus (anti-HCV), and anti–human immunodeficiency virus (anti-HIV). Tools included 24-donation pool testing (PT) using Cobas Amplicor or in individual donations (ID) by Procleix Ultrio. Seven of 761,666 (1:108,800) and 21/250,191 (1:11,900) HBV DNA–positive donations were identified and confirmed by alternative methods. HBV DNA load ranged between 11.6 and 4.6 × 104 IU/mL in 11 samples and could not be quantified in 17 samples. HBV genotypes A (56%) and D (4%) were found. The analysis of combined results from index, follow-up, and look-back samples identified four groups: (1) Two cases tested HBsAg positive with alternative, more sensitive, assays; (2) Four cases were in the pre-seroconversion window period; (3) Eight cases had a fluctuating pattern of HBV DNA and anti-HBs detection (recovered infection); and (4) twelve cases carried anti-HBc without anti-HBs, which might correspond to either chronic or recovered “occult” HBV infection. One donor with no HBV markers in the follow-up was excluded, and another was in the window period preceding anti-HBs. HBV NAT identified more confirmed positive donors than HCV or HIV NAT, and 1:250,000 could not be detected by anti-HBc screening. Serological and molecular studies on follow-up and look-back samples are important to classify donors. In conclusion, further studies are needed to determine whether the considerably higher yield of HBV DNA detection obtained with individual donation screening improves blood safety compared with anti-HBc screening. (HEPATOLOGY 2006;44:1666–1674.)

The introduction of molecular methods to screen blood donors for hepatitis C virus (HCV), human immunodeficiency virus (HIV), and hepatitis B virus (HBV) has improved blood transfusion safety.1–3 In addition, detection of molecular markers of infection in asymptomatic individuals with no serological markers, which are used for blood donor screening [hepatitis B surface antigen (HBsAg), anti-HCV, anti-HIV], provides new information on the natural course of these infections. Several aspects of the biology of HBV remain unclear. One of them is the detection of HBV DNA in the absence of HBsAg (occult HBV infection) and the determination of its frequency and characteristics.4 Answering some of these questions became achievable in 2005 when HBV DNA screening of all seronegative blood donations was introduced in Poland. This report presents the analysis of such cases in asymptomatic blood donors identified during the first year of screening in a country with a medium level of HBV endemicity.

The aim of this investigation was to analyze the frequency of HBV DNA–positive/HBsAg-negative donations and to use serological and molecular analysis of index samples and, when possible, follow-up and look-back samples to determine the HBV infection status of the donors.


HCV, hepatitis C virus; HIV, human immunodeficiency virus; HBV, hepatitis B virus; HBsAg, hepatitis B surface antigen; NAT, nucleic acid testing; ALT, alanine aminotransferase.

Materials and Methods

Donors and Sample Collection.

The study was conducted in 2005, during which all blood donations collected in Poland were tested for anti-HCV (Ortho Clinical Diagnostics, Raritan, NJ), HBsAg (Ortho Clinical Diagnostics, Raritan, NJ, or Biomerieux, Lyon, France) (sensitivity 0.1 ng/mL and 0.28 ng/mL, respectively), anti-HIV1/2 (Vironostica Uni-Form II Ag/Ab, BioMerieux, Boxtel, the Netherlands) and alanine aminotransferase (ALT) with either a kinetic or enzymatic method. Only donors negative for serological viral markers and ALT level within the normal range for blood donors (below twice the normal range for the method used) were analyzed in this study. All donors were voluntary and non-remunerated. They were medically assessed and denied any known risk factors for viral infection listed in a questionnaire.

Donors found to carry HBV DNA were re-questioned for risk factors by a physician. They were also asked whether and when they were vaccinated to HBV. Recalled donors were also asked to give a follow-up blood sample for further HBV studies. In addition, previous samples from repeat donors kept stored at −30°C or lower were retrieved from archives, considered ‘look-back’ samples, and analyzed for HBV DNA and other HBV markers. The Bioethics Commission of the Institute of Hematology and Blood Transfusion, Warsaw, approved this study.

Nucleic Acid Testing, Confirmation, and Quantification.

Screening for HBV DNA was performed at 21 regional blood transfusion centers (RBTC) by one of two test formats: (1) plasma samples of 761,666 seronegative donors were tested in 24 donation pools by Cobas Amplicor/Ampliscreen HCV v. 2.0 (Roche Diagnostics, Manheim, Germany) together with an internal control and a positive control included in each run (100 IU HBV WHO Standard). The analytical sensitivity was estimated elsewhere as a 95% detection limit of 6.7 IU/mL.5 When applied to pools of 24 donor plasmas, the limit of detection became 161 IU/mL. HBV DNA reactive pools were resolved by smaller pool testing, and HBV DNA positive donations were identified. (2) Plasma samples from 250,191 individual donations were tested by Ultrio TMA HIV1/HCV/HBV assay (Chiron, Emeryville, CA) in single-donation format able to detect 11 (7.3–22) IU/mL at the 95% limit.6 Reactive donations were retested by discriminatory assays for HCV RNA, HBV DNA, and HIV RNA. Donations reactive with HBV discriminatory assay were considered positive for HBV DNA. Ultrio positive/discriminatory assay negative donations were considered discrepant. All HBV DNA positive and discrepant samples identified at regional blood transfusion centers were sent to the Institute of Hematology and Blood Transfusion for further analysis. They were tested at the Institute of Hematology and Blood Transfusion by Cobas Ampliscreen HBV (Roche Diagnostics, Mannheim, Germany) or RealArt HBV RG polymerase chain reaction Kit (Artus GmbH, Hamburg, Germany) (95% detection limit 19 IU/mL) and considered confirmed HBV DNA positive if the result was positive. Viral load was quantified with the RealArt assay (100% detection limit 1,000 copies/mL; ∼140 IU/mL) or with Amplicor HBV Monitor. To enhance the sensitivity of HBV DNA detection, we combined procedures of RealArt and Cobas with DNA extracted from 1 mL plasma by Multiprep procedure with a concentration step (Roche) or from 2 mL plasma with Nuclisens Extractor (BioMerieux, Organon Technika, Boxtel, The Netherlands).7 This procedure was expected to increase the detection limit for both assays to 2 IU/mL and 7 IU/mL for RealArt and Amplicor, respectively. If HBV DNA was not detected by quantitative assay, testing was repeated with Cobas Ampliscreen HBV and, when positive, the concentration of HBV DNA was indicated as less than 10 IU/mL. All available look-back and follow-up samples were tested by the same methods as volume permitted.

Whenever HBV DNA load permitted, sequencing of the pre-S and S regions was performed and the HBV genotype was obtained by phylogenetic comparison with references obtained from GenBank as previously described.8

HBV Serological Markers.

Confirmed DNA-positive index samples as well as available follow-up and look-back samples of the corresponding donors were tested for serological markers of HBV infection: HBsAg, using different tests than those used for screening [Ortho Antibody to HBsAg ELISA Test System 3 Ortho-Clinical Diagnostics, Raritan, NJ and chemoluminescence-based assay HBsAg Reagent Pack Ortho-Clinical Diagnostics, NJ (sensitivity, 0.08 ng/mL)], anti-HBc (Monolisa anti-HBc Plus, BIO-RAD, France, Monolisa HBc IgM Plus, BIO-RAD, France); anti-HBe [Monolisa HBe, BIO-RAD, France detection kit for hepatitis B e antigen (HBeAg) and HBeAb] and quantitative anti-HBs (Hepanostika anti-HBs, BioMerieux, Boxtel, The Netherlands). All assays were performed according to the manufacturers' instructions.


A total of 1,011,857 samples from HBsAg, anti-HIV, anti-HCV seronegative blood donations were tested for HBV DNA; 761,666 in pools of 24 with the Roche assay, 250,191 in single donation samples with the Chiron assay. A total of 28 samples were confirmed HBV DNA positive (Table 1); seven from the group tested in pools (approximately 1:108,800), 21 from the group tested in individual samples (∼1:11,900), P < .02 (Table 1). Among the latter 21 samples, 12 corresponded to screening/discriminatory discrepant samples that were confirmed with alternative methods. HBV DNA was quantified in 11 samples (viral load ranging between 11.6 and 4.6 × 104 IU/mL) but in the other 17 samples it was not quantifiable and indicated as less than 10 IU/mL. All HBV DNA–positive/HBsAg-negative donations were anti-HCV and anti-HIV negative. Within the same period, six HCV RNA positive/anti-HCV negative donations (1 with Chiron assay and 5 with Roche assay) were identified [HCV nucleic acid testing (NAT) yield rate 1:170,000]. No HIV RNA–positive seronegative donation was found.

Table 1. Molecular and Serological HBV Markers in Index Samples
DonorNATConfirmation AssayHBV DNA LoadHBsAg EIAHBV Antibody MarkersGenotype
HBV DNA IsolationHBV DNA DetectionBioMOrth EIAOrt chAnti-HBcAnti-HBeAnti-HBs IU/mL
  • Abbreviations: C, Chiron, R, Roche;

  • *

    Discrepant results with Ultrio (see text); BioM, BioMerieux; Orth EIA, Ortho Diagnostics EIA; Ort ch, Ortho chemiluminescence assay (see Materials and Methods); ND, not done.

  • The first letter indicates the genotype; the next block indicates a region amino acid substitutions from the genotype consensus sequence.

  • See Fig. 1.

1CCobas Monitor1.8E4NegPosPosNegPos16,000A
2RQiagenArthus3.2E3ND2NegPosPosPosNegD P127T
3CQiagenArthus15, 37NegNegNegNegNegNegA
5CQiagenArthus3.9E2NegNegNegNegNegNegA P142L
7C*NucliSensCobas Ampliscreen<10NDNegNegPosNegNeg 
8C*Authomatic TNAICobas Ampliscreen (6 pos 4 neg)<10NDNegNegPosNegNeg 
9C*Authomatic TNAICobas Ampliscreen (3 pos 7 neg)<10NegNegNegPosNegNeg 
10C*Authomatic TNAICobas Ampliscreen (7 pos 3 neg)<10NDNegNegPosNeg11D P127T P143R S144L
11RQiagen NuclisensArthus Cobas Ampliscreen15NegNegNegPosPos8 
12C*NuclisensCobas Ampliscreen15, NegNegNegNegNegNeg1600 
13C*Authomatic TNAICobas Ampliscreen (2 pos 4 neg)<10NegNegNegPosNegNeg 
14C*Qiagen NuclisensArthus Cobas Ampliscreen11, 6NegNegNegNegNeg59A
15RQiagenArthus3.5E2NDNegNegPosNegNegA G145A
17RQiagenArthus45NDNegNegPosPosNegA M133I
19C*NuclisensCobas Ampliscreen<10NDNegNegPosNegNeg 
20RCobas Monitor21, 7NDNegNegPosPosNegD
22CNuclisensCobas Ampliscreen<10NegNegNegPosPosNegD
24CQiagen MultiprepArthus Cobas Ampliscreen<10NegNegNegPosPosNegD
25RNuclisens MultiprepArthus Cobas Ampliscreen<10NegNegNegPosPosNegD
26C*MultiprepCobas Ampliscreen<10NegNegNegPosNegNeg 
27C*NuclisensCobas Ampliscreen<10NDNegNegPosPosNeg 
28C*Qiagen MultiprepArthus Cobas Ampliscreen<10NegNegNegNegNegNeg 

A total of 16 samples had sufficiently high viral load to amplify and sequence the S region and determine HBV genotype. Genotype A was dominant (9/16), and genotype D was present in seven strains (Table 1). Two donors carrying HBV genotype A (case 15) or D (case 22) had been vaccinated, and the amino acid sequence of the ‘a’ region was available. Case 15 strain contained the G145A substitution. In case 22, the sequence between amino acid 107 and 149 is shown in Fig. 1. Although clearly a genotype D strain by phylogenetic analysis, this particular area of the S protein sequence was unique. It included two cysteine substitutions C124I and C139S as well as a D144G mutation in the fourth loop. On either side of amino acids 104 and 183, case 22 strain sequence was similar to the consensus sequence. Within the area spanning the ‘a’ region, mutations that caused nine amino acid substitutions on the S reading frame caused only four substitutions in the P reading frame compared with the consensus sequence defined in the legend of Fig. 1. In all 16 genotyped samples, the pre-S region was essentially wild type and in particular did not include deletions.

Figure 1.

Case 22 “a” region amino acid substitutions from genotype D consensus sequence. “A” region spans amino acids 107 and 149 of the S gene product. It includes four loops (boxed) separated by cysteines. The consensus sequence was obtained from the other genotype D Polish strains in this study (cases 2, 24, and 25) and a reference sequence (accession number X72702). Substituted consensus amino acids are indicated as gray letters in a black circle. The substituted amino acids from the sequence of case 22 are shown as a contiguous gray circle with a black letter.

HBV DNA carrying donors were mostly males (25/28) ranging in age between 19 and 59 years (median, 41) (Table 2). A quarter of these donors were first time donors and three quarters were repeat donors. This distribution was not significantly different from the ratio in the general donor population. As predicted from the screening protocol, all had normal ALT level and denied any risk factors when filling the pre-transfusion questionnaire. However, on recall interview, the source of HBV infection was identified in 13 donors. Three donors had received HBV vaccine (see above).

Table 2. Clinical and Epidemiological Data of HBV DNA–Positive/HBsAg–Negative Blood Donors
DonorFirst Time/RepeatSexAgeALT IU/LPotential Source of Infection/Risk FactorVaccination
1RM2919Surgery 2004; tattoos 2001–2004NA
2FtM2216Wounded in a fight, sexual promiscuity; IVDU?NA
3RM3818Heavy alcohol useNo
4FtM2215Heterosexual promiscuityNo
5RM2218New sex partner 1 month before donationNA
6FtM2911Noninjection drug use 6 months before donation 2 sex partners within a month before donationNA
7RM3223 No
8RM5024Tattoos in prison 1978; hepatitis B within familyNo
9RM4119 NA
10RM4620 NA
11RM 21 NA
12RM2745Hepatitis B within familyNA
13RM4216 NA
14FtM4422 NA
15RM4422Surgery 6 months before first HBV DNA + Vaccination 3 × Engerix before surgeryYes
16RF4010Not knownNA
17RM4638 NA
18FtM3427 NA
19 M4610 NA
20RF4818 NA
21RM3229Family member with Hepatitis B—sharing razors; 3 sex partners in last 3 months; non-injection drug user 2002NA
22RM5410CABG 2 months before donation3 injections 1999
23RM4637Brother with acute HBV 2003; tooth operation in 1999NA
24FtM5919 NA
25RM4620 NA
26RM5923 NA
27FtF199Flu-like symptoms 6 months before donationYes–2004
28RM3335 NA

Table 1 indicates the results of molecular and serological investigations performed on each of the 28 index samples found to contain HBV DNA but no HBsAg with the routine screening assays in use at each participating blood center. In cases 1 and 2, HBV DNA load was 1.8E4 and 3.2E3, respectively, and, upon serological re-testing, more sensitive HBsAg assays were reactive. Recovery from HBV infection was confirmed in case 1 when anti-HBs was present in the index sample, and 35 weeks later, both HBsAg and HBV DNA had become undetectable. In case 2, 2 weeks after collecting the index sample, the serological pattern (anti-HBc pos; anti-HBe pos) and HBV DNA load remained at similar level. In neither case was the ‘a’ region amino acid sequence substantially different from wild-type references (Table 1).

In cases 3 through 6 (14.3%), no serological marker of HBV was detectable except in case 6, who had anti-HBc. They were classified as pre-HBsAg window phase based on the occurrence of HBsAg in the follow-up study. In cases 3 through 5, initial low viral load in the index donation increased by 2 to 4 logs in the follow-up samples. Case 6 had a high viral load of 45,000 IU/mL but was negative with two Ortho HBsAg assays. The negative HBsAg result was not explained by escape mutations because the ‘a’ region was wild type. HBsAg was detected in the follow-up sample from this donor when the viral load was 2.9 × 107 IU/mL.

The classification derived from the sole index sample of cases 7 through 28 (Table 1) was improved and sometimes modified by the availability of follow-up samples and, in some cases, of archived samples (Table 3). In case 28, despite confirmed presence of low-level HBV DNA in the index sample, no detectable serological markers were seen in the index or the follow-up sample collected 8 weeks later in which no HBV DNA was detected. This sample was considered doubtful and excluded from further analysis. The follow-up sample of case 27 obtained 9 weeks after the index samples contained a very high level of anti-HBs (30,000 IU/L) and was HBV DNA negative. This pattern suggested a primary immune response to the HBs antigen after the secondary window period. In the 20 cases (7–26) constituting the bulk of index samples, 15 cases had follow-up samples or both look-back and follow-up samples available for testing. The data regarding HBV DNA and anti-HBs status of these cases are shown in Table 3. All additional samples were HBsAg negative, and when anti-HBc or anti-HBe was present in the index samples, it was confirmed. The anti-HBe status did not provide additional information for classification because it can be found in both recovered and chronic HBV infection, although the marker may be more likely present in tail-end carriers than in late-stage recovered infection. Based on HBV DNA and anti-HBs status cumulated from the index, follow-up, and look-back sample results, two groups of donors emerged (Tables 1 and 3). A first group of eight donors (7–14) (28.6%) had HBV DNA or anti-HBs detectable in different samples. These two markers may be simultaneously present in a given sample; however, in a number of samples it was either/or neither. In all except donor 12, anti-HBs levels were low, below 60 IU/L. This pattern of fluctuating detectability of viral genome and neutralizing antibody was seen in samples from donors 7 through 13 (Table 3). A second large group of donors (15–22) was characterized by the repeated detection of HBV DNA at low concentration (<1,000 IU/mL; in most cases, <200 IU/mL) without detectable anti-HBs. Anti-HBc, often associated with anti-HBe, was consistently detected. This pattern was clear in donors 15 through 17, where both look-back and follow-up samples were available. In the other donors (18–22), only one follow-up sample was available and, in all cases, HBV DNA was undetectable. Classification of these samples was therefore somewhat uncertain, although the consistent presence of anti-HBc in both the index and the follow-up samples limited the likelihood of HBV DNA contamination. In four more index samples from donors 23 through 26, a similar pattern of HBV marker was found but the unavailability of look-back or follow-up samples prevented further attempts at classification. This group of 12 donors (15–26) constitutes 42.9% of the HBV DNA yield samples.

Table 3. HBV DNA and Anti-HBs Levels in Look-back, Index and Follow-up Samples of Occult HBV Carriers
ID Look-back SamplesIndex SampleFollow-up Samples
  1. Abbreviations: ND, not done; Neg, negative; Pos, positive. For HBV DNA and anti-HBs, Pos indicates reactivity without quantification being performed.

7Weeks to index −38−33−17−7 +16 
 HBV DNA (IU/mL) NDNDNegNeg<10Neg 
 Anti-HBs (IU/L) 65.8NegNegNeg8.5 
8Weeks to index −35−26−18−10 +4 
 HBV DNA (IU/mL) NDND<10Neg<10Neg 
 Anti-HBs (IU/L) 289.5914Neg14 
9Weeks to index  −42−23  +12 
 HBV DNA (IU/mL)  NDND <10Neg 
 Anti-HBs (IU/L)  2818 Neg39 
10Weeks to index  −42   +4 
 HBV DNA (IU/mL)  <10  <10Neg 
 Anti-HBs (IU/L)  14  11Neg 
11Weeks to index    −10   
 HBV DNA (IU/mL)    <1015  
 Anti-HBs (IU/L)    108  
12Weeks to index   −25    
 HBV DNA (IU/mL)   Neg <10  
 Anti-HBs (IU/L)   6,300 1,600  
13Weeks to index      +12 
 HBV DNA (IU/mL)     <10Neg 
 Anti-HBs (IU/L)     Neg6.5 
15Weeks to index−66−53−39−26−13 +10+26
 HBV DNA (IU/mL)NegPosPosPos2203501001050
 Anti-HBs (IU/L)NegNegNegNegNegNegNegNeg
16Weeks to index    −14 +2 
 HBV DNA (IU/mL)    Neg3.24.5 
 Anti-HBs    NegNegNeg 
17Weeks to index   −27−13 +2 
 HBV DNA (IU/mL)   1952045120 
 Anti-HBs (IU/L)   NegNegNegNeg 
18Weeks to index      +2 
 HBV DNA (IU/mL)     <10Neg 
 Anti-HBs (IU/L)     NegNeg 
19Weeks to index      +3 
 HBV DNA (IU/mL)     <10Neg 
 Anti-HBs (IU/L)     NegNeg 
20Weeks to index       +36
 HBV DNA (IU/mL)     14 ND
 Anti-HBs (IU/L)     Neg Neg
21Weeks to index      +9 
 HBV DNA (IU/mL)     <10Neg 
 Anti-HBs (IU/L)     NegNeg 
22Weeks to index      +6 
 HBV DNA (IU/mL)     <10Neg 
 Anti-HBs (IU/L)     NegNeg 


The frequency of HBsAg-negative/DNA-positive donations identified in the Polish HBV/HIV/HCV NAT screening program is considerably higher than previously published from neighboring Germany.1, 7 In Poland, the incidence of HBV acute hepatitis tends to be higher than in other European countries.9 Recently, the implementation of the hepatitis B control program may have contributed to the observed decrease in incidence (4.7/100,000).10 Blood donors are not routinely screened for anti-HBc; however, such a study was recently conducted in Krakow and revealed 7% anti-HBc in 920 consecutive plasma samples (Kusmierczyk J, personal communication).

In Germany, HBV NAT is performed in pools of 96, although this dilution factor decreasing test sensitivity is partially compensated by viral particle concentration by centrifugation.1 According to the difference obtained in this study comparing pools of 24 and single-donation screening (1:108,800 and 1:11,900, respectively), the decreased sensitivity introduced by pooling is clearly explaining the observed yield discrepancies as previously shown.11 However, such differences may not be only related to sensitivity but also can be explained in part by the fact that blood centers testing in single donations were located in areas of higher prevalence of HBV infection. These elements led to a recent reduction of the pool size to six samples.

HBV DNA levels observed in most index samples is close to or below 10 IU/mL and any dilution would have provided a negative result (Table 1). Obtaining HBV DNA–positive results at the limit of sensitivity of the assays is further illustrated by the relatively large number of samples reactive with the Chiron screening assay (95% detection limit approximately 11 IU/mL) but not reactive with the discriminatory assay, although alternative commercial assays such as Cobas Ampliscreen (95% detection limit approximately 6 IU/mL) and nucleic acid extraction from 2 mL plasma instead of 0.5 mL confirmed the presence of HBV DNA in some of these samples. There seems to be a difference in sensitivity between the screening assay, which emits the same signal for HIV, HCV, or HBV reactivity, and the discriminatory assay intended to identify the viral genome responsible for the reactivity of the screening assay. The discriminatory assay could not be considered confirmatory, and proper confirmation required using alternative assays. In addition to testing undiluted samples, the sensitivity of the assay, the volume of plasma extracted, the potential concentration of viral particles by centrifugation, and the repetition of testing in individual samples are all playing a role in the final number of HBV DNA–containing blood donations identified.

Despite low viral load, genotyping was performed in more than half of the index samples, and both genotypes A and D were found, consistent with previously described data (Table 1).12 Although the amino acid sequence of the “a” region of most samples was wild type, samples from two vaccinated donors showed mutant features. One was genotype A and had the Gly>Ala mutation in the 145 position, the same as the classical Gly145 Arg substitution described in vaccinated patients.13 The genotype D strain from donor 22 was highly unusual (Fig. 1); in particular, substitution of cysteines at position 124 and 139 is predicted to have a dramatic effect on the presentation of the HBsAg conformational epitopes.14

This study confirmed that the introduction of HBV DNA screening in Poland did identify HBsAg-negative blood donations carrying HBV DNA (overall 1:36,137 but possibly 1:11,900). Such yield is considerably higher than that obtained in antibody-negative samples tested for HCV antigen or RNA or HIV RNA, which also have been implemented.15–17 Anti-HBc screening, the alternative approach to HBV blood safety used in countries such as France, Germany, the United States, and Canada, could not be considered in Poland because the deferral of over 5% of anti-HBc–reactive donations would compromise the national blood supply.

The availability of both follow-up and look-back samples in a majority of donors identified as HBV DNA carriers in the index samples provided an opportunity to better understand the significance of this condition and to attempt classifying these individuals into different groups on the basis of the cumulated information collected. In two donors (cases 1 and 2), HBsAg was detected with more sensitive tests than used for screening, confirming the importance of improving assays for HBsAg detection. The “a” region amino acid sequence of the S gene product did not reveal substitutions that could explain the discrepancies between assay results (Table 1). Four donors were clearly in the window period characterized by an absence of serological HBV markers and the occurrence of HBsAg a few weeks after the detection of HBV DNA (cases 3–6, Table 1). A frequency of 1:250,000 window period donations (1:80,000 in the ID-NAT group) is not surprising in a population where the prevalence of HBV contact is greater than 5%. Compared with the yield of HCV or HIV window period cases obtained or predicted at the time NAT testing was implemented in Western European countries, HBV NAT screening to detect window period cases appears well justified.18, 19 Window period donations are known to be highly infectious and would not be detected by anti-HBc screening.20 Three pre-HBsAg window phase donations had viral loads below approximately 300 IU/mL (approximately 1,500 copies/mL); however, one donation had a viral load of 45,000 IU/mL (approximately 225,000 copies/mL). Such donations with high viral load without detectable HBsAg have been observed in Japan and in the United States.21, 22

Most HBV DNA cases identified in this study belong to an entity called “occult” HBV infection defined as HBsAg-negative/HBV DNA–positive samples outside the window period.4 The analysis of a complete set of data including two or more samples from most of these cases identified two separate patterns (Table 3). The first pattern, seen in eight cases, is characterized by the presence of a low level of anti-HBs detected either in the index sample or in other samples collected from the same individuals. At such low levels of either DNA or anti-HBs, it is not surprising that at any given time, either marker may or may not be detectable. Scant data on the potential infectivity of such cases are available. One case of a blood donor with fluctuating anti-HBs and HBV DNA followed up over a 7-year period was described, and none of nine recipients of components including one unit containing detectable HBV DNA was infected.23

The largest group (43%) in this study consists of donors carrying viral DNA without detectable antibodies to the surface antigen. The consistent presence of anti-HBc and, in most cases, anti-HBe suggests individuals in the non-replicative phase of chronic HBV infection characterized by low viral load, normal ALT level, and anti-HBe. Contrary to the previous group of donors who had evidence of recovery from the infection and in whom a low level of viral replication seems to persist, this group, at least those where additional samples were available, seem to have a rather stable serological pattern of reactivity. Some of them may have recovered from HBV infection, but anti-HBs is no longer detectable. Experimental approaches to differentiate results recovered from chronic cases in anti-HBc–only carriers might be either to monitor the antibody response to a dose of HBV vaccine24, 25 or to identify the presence of memory T cells responding to stimulation with the S antigen.26 Compared to the previous group, the potential infectivity by transfusion is probably higher and has been described in some anecdotal cases.27–29 In the Japanese study, a donor with occult HBV infection undetected in pools of 50 donations, but positive in individual donation testing and low level anti-HBc, donated blood 78 times and infected at least four immunodeficient recipients, two of whom developed acute hepatitis B.28 This case illustrates the fact that the immunocompetence of the recipients might make a substantial difference in terms of both susceptibility to infection from occult HBV and clinical expression. In the current study, no evidence of HBV transmission related to the identified donors was reported to the public health authorities.

This study was not designed to determine the virological outcome of transfusion recipients. It should be noted, however, that the difficulties with identifying all cases of occult HBV infectivity via blood components in the Polish health service may be the result of a variety of factors related to look-back procedures, i.e., lack of reporting does not mean absence of transmission. Until proven otherwise, components prepared from such blood units should be considered infectious even when transfused to immunocompetent recipients.


The NAT screening for HCV RNA, HIV RNA, HBV DNA in blood donors was performed in Molecular Biology Laboratories of Regional Blood Transfusion Centers in Białystok, Bydgoszcz, Gdańsk, Kalisz, Kraków, Lublin, Poznań, Racibórz, Rzeszów, Szczecin, and Warszawa. All of them, as well as other RBTC in Słupsk, Olsztyn, Opole, Zielona Góra, Łódź, Katowice, Kielce, Radom, Wałbrzych, and Wrocław performed the medical and laboratory examinations of identified infected donors and their follow-up studies. We thank all involved for their excellent cooperation. Drs. D. Candotti and A. Amiri, Laboratory of Molecular Virology, Department of Haematology, University of Cambridge, and Drs. M. Koppelman and H. Zaaijer from Sanquin Diagnostic Services, Amsterdam, the Netherlands, provided the S sequences and genotype of some samples. We thank them for their contribution. We also thank Dr. Nico Lelie for valuable discussions.

Study participants Polish Blood Transfusion Service Viral Study Group: Institute of Haematology and Blood Transfusion: D. Kubicka-Russel, A. Kopacz, A. Gronowska; Regional Blood Transfusion Centers: B. Boczkowska-Radziwon (Białystok), M. Dobrowolna, B. Zalewicz (Bydgoszcz), B. Wyrwińska (Gdańsk), M. Tarnawska (Kalisz), S. Dyla̧g (Katowice), Z. Sitarz-Żelazna (Kielce), J. Kuśmierczyk, J. Raś (Kraków), H.Radwan-Wieczorek, J. Korzeniowska (Lublin), B. Sękowska, A. Rumas (Łódź), G. Kula (Olsztyn), I. Rajca-Biernacka (Opole), M. Krug-Janiak (Poznań), J. Kulbaka-Myrczek (Racibórz), H. Kaczor (Radom), A. Krygowska (Rzeszów), A. Dobrecka (Słupsk), D. Stępień-Razzu, A. Świder-Michalska (Szczecin), U. Zarówna (Wałbrzych), J. Gawęda, D. Malka (Warszawa), E. Świa̧tek (Wrocław), M. Fabisz-Kołodzińska (Zielona Góra).