HERACLIS‐HDV cohort for the factors of underdiagnosis and prevalence of hepatitis D virus infection in HBsAg‐positive patients

Hepatitis D virus (HDV) underdiagnosis remains common. We assessed the HDV screening and prevalence rates in HBsAg‐positive patients seen at tertiary liver centres throughout Greece as well as factors affecting HDV diagnosis.


| INTRODUC TI ON
There is scarcity of reliable data on the current epidemiology of infections with hepatitis viruses, which may have changed over the last decades. Such epidemiological changes may be related to several factors including improvements of social-economic status of several areas over the last decades, introduction of universal infant vaccination against hepatitis B virus (HBV) in most countries in the nineties, development and wide use of effective antiviral therapy for HBV over two decades and of direct acting antivirals for hepatitis C virus (HCV) over the last decade, but also the migrations of great numbers of individuals often from areas of high to areas of low HBV and HCV endemicity. [1][2][3] The lack of strong epidemiological data has a negative impact on the implementation of effective prevention and screening strategies which put at risk the World Health Organization (WHO) targets for viral hepatitis elimination. 4 In such an uncertain epidemiological status, the epidemiology of hepatitis D virus (HDV) is even more undefinable, as HDV screening rates among diagnosed hepatitis B surface antigen (HΒsAg) positive individuals seem to vary widely usually ranging at low levels even in epidemiologically well organized countries. 5,6 Thus, the global HDV prevalence has been debatable and was declared uncertain by the 2017 WHO Global Hepatitis Report. 7 The global prevalence of HDV infection, which is usually based on the detection of antibodies against HDV (anti-HDV), has been reported to vary widely in the estimations of different systematic reviews, 5 since the global number of HDV infected individuals has been suggested to range from 12 to more than 60 millions. 8-10 At the same time, it is well accepted that HDV underdiagnosis remains Correspondence George Papatheodoridis, Department of Gastroenterology, Medical School of National and Kapodistrian University of Athens, General Hospital of Athens "Laiko", Athens, Greece. Email: gepapath@med.uoa.gr

Funding information Gilead Sciences
Handling Editor: Alessio Aghemo Abstract Background and Aims: Hepatitis D virus (HDV) underdiagnosis remains common. We assessed the HDV screening and prevalence rates in HBsAg-positive patients seen at tertiary liver centres throughout Greece as well as factors affecting HDV diagnosis.
Methods: All adult HBsAg-positive patients seen within the last 5 years were included.
Non-screened patients who visited or could be recalled to the clinics over a 6-month period were prospectively tested for anti-HDV.
Results: Of 5079 HBsAg-positive patients, 53% had anti-HDV screening (41% before and 12% after study initiation). Pre-study (8%-88%) and total screening rates (14%-100%) varied widely among centres. Screening rates were associated with older age, known risk group, elevated ALT, centre location and size and period of first visit. Anti-HDV prevalence was 5.8% without significant difference in patients screened before (6.1%) or after study initiation (4.7%, p = 0.240). Anti-HDV positivity was associated with younger age, parenteral drug use, born abroad, advanced liver disease and centre location. Overall, HDV RNA detectability rate was 71.6% being more frequent in anti-HDV-positive patients with elevated ALT, advanced liver disease and hepatitis B therapy.

Conclusions:
Anti-HDV screening rates and recall capabilities vary widely among Greek liver clinics being higher in HBsAg-positive patients of known risk group with active/advanced liver disease seen at smaller centres, while non-medical factors are also important. Anti-HDV prevalence varies throughout Greece being higher in patients born abroad with younger age, parenteral drug use and advanced liver disease.
Viremia is more frequently but not exclusively detected in anti-HDV-positive patients with elevated ALT and advanced liver disease.

K E Y W O R D S
HDV RNA, hepatitis D, prevalence, underdiagnosis

Lay Summary
• Anti-HDV screening rates and recall capabilities vary widely among Greek liver clinics.
• Anti-HDV prevalence varies throughout Greece being higher in patients born abroad with younger age, parenteral drug use and advanced liver disease.
• HDV viremia is more frequently but not exclusively detected in anti-HDV-positive patients with elevated ALT and advanced liver disease. common varying widely among different countries and even among different areas of the same country, which may face different barriers for HDV diagnosis. 5,11 Given that HDV requires for infection the helper functions of HBV and in particular the presence of HBsAg, HDV diagnosis is restricted in HBsAg-positive cases. [12][13][14] HDV infection, either as co-infection with HBV or as superinfection in individuals with pre-existing chronic HBV infection, is associated with increased morbidity and mortality. [15][16][17] In particular, HDV infection may cause severe and even fulminant hepatitis during the acute phase and is associated with frequent and rapid progression to advanced stages of chronic liver disease during the chronic phase, which increase further the risk for hepatocellular carcinoma (HCC). [15][16][17][18][19] Therefore, the prompt diagnosis of HDV infection is of great importance for the optimal patients' management. The clinical relevance of the prompt HDV diagnosis has become even more important after the recent development and approval of effective therapeutic options against HDV. 17,20 The aim of this study was to assess the anti-HDV screening and prevalence rates in HBsAg-positive patients seen at tertiary liver centres throughout Greece and to identify factors and barriers affecting HDV diagnosis. In addition, the rates and predictors of HDV RNA viremia were also evaluated.

| Patientpopulation
Τhis retrospective-prospective cohort study included all adult HBsAgpositive patients who had at least one visit within the last 5 years before  Figure S1) participated in this HEllenic multicentre ReAllife CLInical Study for HDV (HERACLIS-HDV). The study protocol was approved by each Hospital's ethics committee. All patients who were seen during the prospective period of the study signed informed consent, while the informed consent was not considered to be necessary for the anonymous retrospective collection of data.
Epidemiological, clinical and laboratory data were retrieved anonymously from the patients' records using a specifically designed case report form (CRF). HBsAg-positive patients were considered as cases of known risk group for HBV infection if they had another first degree relative with chronic HBV infection, reported parenteral drug use, were men having sex with men or had history of multiple transfusions of blood or blood products or of other parenteral exposure to biologic material of chronic HBV patients. Based on alanine aminotransferase (ALT) activity during periods without treatment, patients were classified into cases with persistently normal ALT (normal ALT: ALT<40 IU/L) and cases with persistently or transiently elevated ALT (elevated ALT). The diagnosis of cirrhosis was based on histological and/or ultrasonographic (nodules in the hepatic parenchyma, spleen >12 cm, portal vein >16 mm) and/ or endoscopic (oesophageal or gastric varices, portal gastropathy) and/ or clinical findings (ascites, variceal bleeding, encephalopathy, nonobstructive jaundice). Patients who progressed to cirrhosis or underwent liver transplantation at any time before the completion of the study were considered as cirrhotics or liver transplant patients, regardless of their initial stage of disease. According to the study protocol, patients without screening for anti-HDV up to September 2021 who visited the clinics or could be recalled to visit the clinics during the next 6 months (from October 2021 to March 2022) were tested for anti-HDV.
In order to assess the potential effects of different periods on the probability of anti-HDV screening, patients were stratified ac- HBsAg-positive patients seen over the study period, respectively.

| ME THODS
The diagnosis of HDV infection was based on the detection of anti-HDV in serum, while the diagnosis of active infection was based on the detection of serum HDV RNA. The presence of anti-HDV was determined at each participating centre by commercially available enzyme immunoassays. Detection of serum HDV RNA in the past was performed at different labs by different assays, whereas the current detection of serum HDV RNA for this study was based on the same assay performed at a central lab. In particular, serum HDV RNA was measured by a commercially available quantitative real-time polymerase chain reaction (PCR) assay (Hepatitis delta virus, genesiq® Advanced Kit, Primerdesign™ Ltd) with lower limit of quantitation 575 IU/mL and lower limit of detection 100 IU/mL.

| Statisticalanalysis
All data were entered into a specifically designed database. Statistical analyses were performed using SPSS Statistics for Windows (Version 28.0, IBM Corp Armonk, NY, 2021). Categorical variables were summarized as frequencies and percentages and their associations were determined using corrected chi-squared or two-sided Fisher's exact test. All quantitative variables were presented as mean ± standard deviation (SD) or as median values [interquartile range (IQR)] depending on their distribution type. Quantitative variables comparisons between two groups were performed by the t-test or non-parametric Mann-Whitney test, whenever appropriate. Multivariable logistic regression analysis was used to identify independent factors associated with screening for anti-HDV, anti-HDV seropositivity and HDV RNA detectability among anti-HDV positive cases; their results were presented as odds ratio (OR) with 95% confidence interval (CI). Only variables who had an association or a trend for association (p < 0.10) with the dependent variable were included in the multiple logistic regression models. Statistical significance was set at p value <0.05.

| RE SULTS
In total, 5079 HBsAg-positive patients were included. The patients' baseline characteristics are presented in Table 1. Their mean age was 56 ± 16 years, whereas 59% were males and 66% were born in Greece. The vast majority (97%) of patients were negative for hepatitis B e antigen (HBeAg) and non-cirrhotic (92%). Most of them were followed in large centres (85%) located in Athens (54%). The first visit of the majority (82%) of patients was after May 2010 (after the official onset of the Greek austerity crisis), whereas 12% of the patients had their first visit within the last 1.5 year.

| Anti-HDVscreening
Of the 5079 HBsAg-positive patients, 2696 (53%) cases eventually had anti-HDV screening; 2105 (41%) cases were screened before the study initiation (up to September 2021) and another 591 (12%) cases were screened during the 6-month period after the initiation of the study (October 2021 to March 2022). The anti-HDV screening rates before the study initiation and overall varied widely among centres ranging between 8%-88% and 14%-100%, respectively ( Figure 1).
The probabilities of anti-HDV screening before (in all 5079 patients) and after study initiation (in 2974 previously non-screened patients) are shown in Table 2. Almost all evaluated factors were associated with these probabilities in the univariable analyses. In the multivariable analysis, the probability of anti-HDV screening before the study initiation was independently associated with older age (adjusted

| HDVviremiainanti-HDVpositivepatients.
Of the 157 anti-HDV positive patients, 135 (86%) were tested for serum HDV RNA levels during the screening visit of this study,  TA B L E 2 Factors affecting the probabilities of anti-HDV screening before and after study initiation among HBsAg-positive patients seen at 14 tertiary liver centres throughout Greece.

Logistic regression analysis
Anti-HDV screeningn/N(%) p what was reported from Greece earlier (1997-2003: 57%). 26 It should be noted that we observed a great variation among the centres with screening rates ranging from 8% to 88%. Given that only clinics from tertiary liver centres were included in our study, we can reasonably assume that the overall anti-HDV screening rate is much lower in our country when HBsAg-positive persons seen from less experienced and even primary care physicians are also considered.

Logistic regression analysis
The probability of anti-HDV screening before our study initiation was independently associated with older age, known risk group,    services, especially when non-reimbursed tests such as anti-HDV were included.
There are several strategies aiming to improve the anti-HDV screening rates in HBsAg-positive patients. One of them can be based on the search of HBsAg-positive patients' records and the recall of cases who have not been previously screened. Approximately 20% of our non-screened patients could be screened over a 6-month period after study initiation. In three large centres with low screening rates which collected such information, the two most common reasons for non-screening during the 6-month recall period were failures to contact the patients (72%) and patients' unwillingness to revisit the centre or just to undergo anti-HDV testing in time (20%).
Thus, it seems that such a policy may not be effective for the majority of non-screened cases. Whether a longer recall period might have resulted in the screening of larger proportion of unscreened cases cannot be answered by our study. The probability of anti-HDV screening during the 6-month period was associated with younger age, known risk group, elevated ALT, advanced liver disease, centre location and smaller centre size and recent first patient visit. Some of the above associations may be related to higher probabilities of visits and response to recall strategy by some patients' subgroups, such as cases with younger age, known risk group, elevated ALT and advanced liver disease as well as more recent first visit. In addition, the centre characteristics were found to affect the probability of anti-HDV screening during the 6-month prospective period with larger and busier centres again having lower screening rates. It should be mentioned, however, that, although the screening rates were lower for large centres, such centres screened much higher absolute numbers of HBsAg-positive patients during either the initial or the 6month prospective period.
The overall anti-HDV prevalence among screened patients was 5.8%, which is higher to what has been previously reported for HBsAg-positive patients seen at Greek liver centres mainly between in the previous decades nineties (4.2%). 26 Since not all HBsAgpositive patients were screened, we cannot exclude that physicians might have screened more frequently cases at higher risk of HDV infection, a hypothesis also supported by the factors associated with anti-HDV screening. It was quite interesting, however, that the anti-HDV prevalence did not differ significantly in the patients screened before this study initiation (6.1%) and during the 6-month prospec- 12%, p < 0.001), but the centre location remained an independent factor of higher anti-HDV prevalence in multivariable analysis after adjustment for other factors. Such data further support the heterogeneity of anti-HDV prevalence among HBsAg-positive cases seen even at areas of the same country. 6,10 It is well known that not all anti-HDV positive patients have HDV viremia. 10 In our study, the overall HDV RNA detectability rate was approximately 72%, which is higher than what has been usually reported. 10 Whether repeating HDV RNA determination at another visit and/or the use of a more sensitive PCR assay would have resulted in higher rates of detectable serum HDV RNA is unknown. 11,29 As expected, detection of HDV viremia was associated only with markers of liver disease activity and/or severity, such as elevated ALT, advanced liver disease (cirrhosis or liver transplantation) and reported use of HBV therapy. 9,10 However, serum HDV RNA could be frequently detected in all anti-HDV positive patient subgroups, such as those with normal ALT (>40%), without advanced liver disease (>60%) and without a history of HBV therapy (>40%).
In conclusion, the majority of chronic HBV patients seen at Greek