Does hepatitis delta virus have a preference for hepatitis B virus genotype? A systematic review of the literature

Hepatitis delta virus (HDV) is a deficient virus that requires the surface proteins of Hepatitis B virus (HBV) to complete its replication. HDV is thus only found in those already infected with HBV (~5% worldwide). There are eight different HDV genotypes (1–8) and 10 HBV genotypes (A–J), each having fairly distinct geographic distributions. While their pairings may be coincidental based on epidemiological occurrence, some evidence exists regarding possible virologic basis for genotype dominance and patterns. Here we sought to determine which HBV genotype is most often linked with active HDV infection and speculate on whether this may represent a viral ‘preference’. Electronic databases with OVID Medline were comprehensively searched for studies published between 1977 and 2022 indexing the word ‘genotype’ and all permutations of ‘HDV’ (hepatitis D virus, hepatitis delta, etc.). Primary studies of patient samples reporting genotype data for either or both of HDV and HBV were tabulated. The initial search revealed 419 articles and was narrowed to 133 studies reporting genotype data for either or both HBV and HDV. We limited our search to cases with detectable HDV RNA. These represented over 5800 samples from over 70 countries. Of these, 1947 samples had paired genotype data for both viruses. The most common pairing was HDV‐1 with HBV‐D, but it remains unclear whether this represents a viral ‘preference’ or mere co‐endemicity of the two viruses. Determining if there is a virologic link between HBV and HDV genotypes may have important implications for emerging HDV and HBV research.

][10][11][12] This curious interplay between the two viruses raises the question of whether some evolutionary drive has linked them and/ or whether this pairing may have implications for emerging HBV and HDV research.
HDV was discovered in 1977 by Rizzetto et al, when a novel antigen was identified in the serum and liver of patients with more severe liver disease. 13Since then, eight different genotypes (1-8) have been defined, based on a genetic variation of 15%. 14HBV also has 10 different genotypes (A-J) based on a variation of 8%. 15Both viruses tend to have very distinct geographic distributions (Figure 1).HBV's genotype A is found in sub-Saharan and Western Africa, and Northern Europe; genotypes B and C, in Asia; genotype D, in Africa, Europe, Mediterranean countries, and India; genotype F, in South America; genotype G, while less common, is found in France, Germany and the United States; and genotype H, in Central and South America. 16For HDV, genotype 1 is predominant globally, but other genotypes appear to be more localized, whereby genotype 2 is found in Asia; genotype 3, in South America; genotype 4, in Japan and China; genotype 5, in Western Africa; and genotypes 6-8 in Central Africa. 3The coevolution of these viruses over time is intimately linked with human anthropological histories, 17 yet does not tell the whole story.
At the virologic level, HBV produces three different-sized surface proteins, named small-, medium-, and large-hepatitis B surface antigens (S-, M-, L-HBsAg).Each contains the same C-terminal segments, but M-and L-HBsAg contain additional N-terminal domains, given they are transcribed from earlier promoters in HBV's circular genome.Viral entry into hepatocytes is facilitated via the sodium taurocholate co-transporting peptide (NTCP) receptor. 18While the domains of the S-HBsAg alone are sufficient to enable assembly and viral egress of HDV, the L-HBsAg is the surface protein essential to both HBV and HDV for viral entry, as it is L-HBsAg's pre-S1 domain that facilitates the translocation of the protein across the viral lipid membrane to be presented and accessible to the NTCP receptor. 18,19That is, an HDV virion with an envelope composed of S-HBsAg alone is non-infectious. 19As the different HBV genotypes have been shown to produce differing ratios of these surface proteins, 20 it follows that a virologic basis for an HDV 'preference' may exist.
Other virologic factors besides surface protein ratios may also contribute to potential preferences.This interplay of HBV and HDV genotypes was recently explored in an excellent in vitro study looking at co-infection of each HBV genotype with each of the HDV genotypes. 21 this study, an extensive literature review of HDV cases has been performed to determine whether any patterns emerge from existing clinical data suggesting a 'preferred' HBV genotype for HDV.

| MATERIAL S AND ME THODS
Electronic databases associated with OVID Medline were comprehensively searched for studies published between 1977 and 2022.The keywords used included 'genotype' and all permutations of 'HDV' (hepatitis D virus, hepatitis delta, etc.).All studies with English indexing were searched, regardless of the language of the main text.Articles were then screened and cross-checked for relevance by both authors.Only primary studies of patient samples reporting genotype data for either or both of HDV and HBV were included (i.e.review articles and basic science papers that employed and propagated derivatives of prior known patient samples were excluded).More specifically regarding HDV status, we included papers that detected HDV RNA (not just anti-HDV antibodies) as this would imply active infection.Articles describing triple-or quadruple-infections were also included (i.e.those with HBV-HDV and another virus, such as hepatitis C virus [HCV]   or Human Immunodeficiency Virus [HIV]).Genotypes for each HBV-HDV co-infection sample (where known) were tabulated using Microsoft Excel™ (Tables S1-S3).When both HBV and HDV genotypes were successfully determined and reported in a paper for the same co-infection, this was considered 'paired' data and it is this subset around which the majority of the analyses are based.
Data from articles that had incomplete descriptions of genotypes, either because they were not performed, indeterminant or specified only for either HBV or HDV were still tabulated but included in the 'unspecified genotype' data (Table 1).In the cases where a single co-infection sample contained more than one genotype of the same virus, each was separated out as its own pairing.Any reference with HDV-2b was considered data for HDV-4, as reclassification of the viral genotypes occurred in 2006. 22Any other subgenotypes were recorded only as the main genotype.In addition, all data reporting on HBV and HDV genotypes, sample size (number of HDV positive samples regardless of genotype data available), and geographical origin (where available) were recorded (Tables S1-S3).Viral distribution was graphed based on geographical location (Figure 1).When graphing the data based on geographical location, only studies that had genotypes from 'paired' samples matched with a geographical location were included.
For the comparator HBV mono-infection data, the Supplementary files from Velkov, et al. 23 were reviewed and data combined by region (Table S4).From this, pie charts were constructed to give a proportional genotype distribution for comparison (Figure 1C).

| RE SULTS
A total of 419 articles were identified with the initial search.After screening, 133 studies contained any genotyping data (whether for HDV, HBV, or both).Of these, 69 articles contained definitive pairing data for HDV and HBV; 13 articles attempted genotyping of both viruses but did not obtain paired data; and 51 articles only performed genotyping for one of the viruses (two reported on HBV  S4).For (A) and (B), all samples with identified genotypes (of either HDV or HBV or both) were categorized geographically based on the location of the study if not otherwise stated.All immigrant samples (even if origin of birth was included) were categorized geographically based on where the sample was obtained.All data were graphed using Microsoft Excel™.The world map used was retrieved from www. freew orldm aps.net.The numbers indicate the number of samples with each HBV genotype found.alone; 49 reported on HDV alone).Data from these 133 studies were tabulated (Table 1; Supporting Information) and the geographical distribution of the viruses (where known) in Figure 1.Samples came from around the world and included North America, South America, Africa, the Middle East, Asia, Russia, Oceania/Australia and Europe.
Four papers did not have geographic data paired with the genotype data.
Regarding the HDV-4 data, there were 27 papers that had either HDV-2 or -4.Only four papers (34 samples) reported on HDV-2 and were published before 2006.Five studies reported HDV-2b (51 samples) and were therefore reclassified in this data as HDV-4.
One study included three samples that were quadruple-infected-HBV-HDV-HCV-HIV. 24

| HBV and HDV genotype pairings
Within the currently available literature, a total of 5816 HBV-HDV co-[or triple-] infection samples have been documented that report genotype data for HDV, HBV or both.These come from at least 72 different countries around the world (Table 1, Tables S1-S3).
Overall, there were 1947 samples that had paired data for both the HBV and HDV genotype (compiled from the 69 articles that were successful at genotyping both the HBV and HDV in the same sample).There were 13 samples that reported more than one genotype TA B L E 1 Summary of the hepatitis B virus and hepatitis delta virus genotype pairings.Overall, there were 5816 multi-infections found across articles with genotype data and 1947 samples where genotyping was successfully determined for both HBV and HDV.Total number of samples (in 69 papers) that had genotype data for both HDV and HBV in the multi-infection.Thirteen of the pairs had mixed genotypes. c Total number of multi-infection samples (in 133 papers) with genotype data for either or both HDV and HBV.

| Global distribution-HDV and HBV
From the studies that attempted genotyping of both viruses, there were 4738 co-or triple-infected samples where HDV genotype was linked to a patient from a specific geographic region (Figure 1A, Tables S1 and S2).Except for South America, HDV-1 was universally the predominant HDV genotype found worldwide (3018 samples).
Every region covered by the studies had reported samples of HDV-1.In Oceania, only HDV-1 was reported from the samples (with over 98% of samples in Europe and over 99% of samples from the Middle East also being HDV-1 genotype).Notably, in South America, the predominant HDV genotype was HDV-3 (386 samples).The only other region with a reported case of HDV-3 was in Europe (only one).
Africa had the most diverse number of genotypes, where the only genotype not represented was HDV-3 (although there was only one documented HDV-2 sample).Also found in Africa were the majority of HDV genotypes 5 through 8 (with a few cases of HDV-5 reported in North America, Europe, Asia, and South America; one case of HDV-7 reported in Europe; and six samples of HDV-8 reported in South America).HDV-2 and HDV-4 were primarily found in Asia (in almost equal proportions of 446 and 440 samples respectively), with 28 samples of HDV-2 reported in Russia, and 12 samples of HDV-4 reported in Africa.
With respect to the HBV genotypes from available pairings, a total of 1926 co-infected samples where HBV genotype was determined with specific geographic location were charted (Figure 1B).
HBV-D was found universally (786 samples) and predominated worldwide as the primary genotype in all regions with HBV-HDV co-infection sampling.The exception for this was South America (where HBV-F predominates) and Africa (where HBV-E is predominant).South America was almost exclusively the only location where HBV-F was found (138 samples, while two reported in Asia).In both South America, and Africa, the second most common HBV genotype was also HBV-A, with smaller numbers of HBV-D compared to other parts of the world.HBV-A (376 samples) was also found throughout, except for the Oceania region.In Asia, HBV-D and HBV-B are dominant when found with HDV (222 and 225 recorded samples, respectively).Asia was also the only region that had higher levels of HBV-B

genotype. HBV-C is found almost exclusively in Asia, with only one other sample of HDV co-infection with HBV-C reported in South
America.There were very few documented HBV-HDV co-infection samples with HBV-G, with the only three samples with geographical data recorded from Asia.HBV-C had the second lowest sample count (69 samples) within the HBV-HDV co-infections.There were no documented cases of HBV-H within a co-infection.

| Global distribution-HBV co-infection versus mono-infection
An important comparator for assessment of 'HBV preference' is to examine the baseline prevalence of genotypes to determine if the co-infection is merely a reflection of the baseline proportions.The data for this were derived from the thorough literature review by Velkov et al, 23 who documented a total of nearly 260 million cases.
Comparing the relative amounts of HBV-HDV co-infection by region, discrepancies are mainly noted in North America, Europe and Asia/Oceania (Figure 1B,C).In North America, HBV genotypes A-C collectively make up the majority of cases of HBV, but are barely registering on the HBV-HDV co-infection pie chart, outnumbered by HBV-D, and then HBV-E.In Europe, the predominant HBV genotype is HBV-D, but in the co-infection cohort, while HBV-D still makes up the majority of cases, HBV-E accounts for about a third of cases.
Lastly, in Asia/Oceania, HBV-C accounts for about half of the overall HBV genotypes but makes up only about 11% of the HBV-HDV coinfection genotypes in Asia and none of that in Oceania.

| DISCUSS ION
HDV and HBV have coevolved in humans for decades, if not centuries, though no molecular clock has yet coined its approximate beginning.The drivers for viral adaptation have likely been both opportunities relating to susceptible hosts (geographical convenience) and viral pressures.This literature review and assessment of reported clinical cases incorporating viral genotyping was performed to determine whether HDV has a 'preferred' HBV genotype with which to replicate, or rather, whether such a preferred pairing may be revealed from currently reported clinical cases.
One of the larger studies to date looking at this HBV-HDV gen- Of the 1947 dually genotyped sample pairings recorded (Table 1) the most common pairing for HBV-HDV co-infection was between HDV-1 and HBV-D (892 cases (45.8%)).The next most common pairings were HDV-1 with HBV-E, then HBV-A (190 (9.7%) and 142 (7.2%) samples, respectively).Given the worldwide distribution of HDV-1, the high pairings with HDV-1 are not fully unexpected.Another prominent pairing was that of HDV-3 with HBV-F (130 samples; 6.7%).Both these viral genotypes were predominantly found in South America, likely accounting for this pairing on a geographic basis.Another interpretation of this data is that these pairings have a selective advantage compared to other combinations allowing them to exist and propagate more effectively.
The most common HBV genotype (regardless of HDV genotype) was HBV-D (1076 of 2577 total known HBV-genotyped samples (41.8%)).The next most common HBV genotype paired was with HBV-E (567 samples; 22.0%), followed by HBV-A (468 samples; 18.2%).While our study captured an additional 3239 co-infection cases that were not genotyped, given the country of origin of many of those studies and high prevalence of HBV-D therein, it is quite possible that the numbers for HBV-D would have been even greater.
When HBV-D is compared with the total number of samples (known and unknown HBV genotypes) the percentage confirmed HBV-D was 18.5% but could be as high as 74%.
In terms of flexibility for HBV genotypes to pair with each of the HDV genotypes, HBV-A was found to pair with the greatest number of HDV genotypes (evidence for all but HDV-4).This was followed by HBV-D (all but HDV-6 and -7) and HBV-E (all but HDV-2, −3, and − 4).
Again, this likely represents geographic distribution of the HDV genotypes rather than virologic preference, as all HBV genotypes have been shown to be able to support HDV in the laboratory. 21For HBV-G, previous studies have documented HBV-G in France, Germany, and the United States. 16While nine of the documented paired co-infection cases are from France, 25,26 three additional genotyped samples of HBV-G found in this co-infection dataset originated from Asia, which may indicate a larger geographic distribution than previously reported.
Looking at the data from the perspective of the HDV genotypes, HDV-1 was the only HDV genotype found clinically to pair will all the documented co-infection HBV genotypes (again, there were no HBV-H to -J co-infection samples).Africa was the region with the most diverse number of HDV genotypes; Asia had the most non-HDV-1 genotypes (largely HDV-2 and 4).
There is an undeniable trend supporting a geographical preference for certain strains and this could be the sole explanation for overall findings (Figure 1).Some notable patterns do break from this (Figure 1B,C).North America has nearly equi-pan-genotypic distribution of HBV genotypes in the cumulative data, yet its HBV-HDV coinfection genotypes are predominantly HBV-D, followed by HBV-E.HBV-E genotype also appears to stand out in the European data with co-infection, but not as a predominant genotype in the cumulative.Finally, in Asia/Oceania, HBV-C is by far the most commonly noted HBV genotype overall, but in the co-infection data, accounts for less than one eighth of the cases.These discrepancies may still be fully attributable to migration/immigration patterns and/or to the increased case-finding in regions where co-infection is known (i.e. Mongolia).However, the question of virologic pressure and pairing from coevolution remains.In the recent study by Wang, et al. examining the assembly and infection efficacy of the 8 HDV with each of the HBV genotypes (A-H) and found that all HDV genotypes could be supported by all HBV genotypes, albeit to different degrees. 21In regards to envelopment preferences, the optimal pairing of HBV-B with HDV-6 was observed.They next went on to show HDV viral production kinetics using the HBV-B envelopes and found HDV-8 to be the most efficient. 21However, as researchers also point out, their study represents one strain of each of the genotypes and may not be representative of the full clinical picture or potential.In our review, HBV-B was only documented to pair with HDV-1, -2, and -4 in the reported genotyped studies (Table 1).Regarding the HDV-8 noted to have the greatest efficiency in viral replication, there have not yet been any identified cases 'naturally', but whether this represents insufficient sampling or an alternate virologic influence is not known.HDV-8 was found almost exclusively in Africa, where limited capacity for testing and genotyping will almost certainly limit conclusions that can be drawn around this.
Our study is of course limited by the comprehensiveness of published data from the case series and patient cohorts.By the nature of such studies, testing is limited to subsets of the population and are likely biased towards those with either more severe disease and/ or those with the highest engagement in clinical care.Sampling will also only be accounted for from countries that have the capacity to test for HDV (and HBV) and perform genotyping.Thus, many countries and regions will not be represented in this data, limiting the worldwide scope.Further, for regions (such as Oceania and North America) that document lower numbers of HBV-HDV co-infection, pairing trends may be skewed in the geographical data.
Study limitations were not only from epidemiologic circumstance, but also viral factors.It is well-described that in cases of HBV-HDV co-infection, levels of HBV are suppressed. 8,10,11,27This trend was seen in many of the studies here, where HDV viral RNA was identified and genotyped, but the levels of HBV were so low as to not be typable.This limited some of our pairing conclusions but would have been incorporated in the less specific tabulations (Table 1).Of the co-infections where HDV was genotyped, but not HBV (secondary to low levels/undetectable HBV), we note that the most common HDV genotype was HDV-1, again from the predominance of these total cases.However, if we look at the ratio of unknown to known HBV genotypes for HDV, HDV-4 followed by HDV-2 take the lead, hinting at a potential link for these as having a greater impact on HBV inhibition.This is of course within the limits of those studies and their capacity and intent to genotype the HBV.The converse question cannot be answered from this data -which HBV genotype was most vulnerable to suppression?But a recent study looking at viral loads of HBV and HDV from patients with HBV-A, -D, and -E genotypes showed the greatest impacts on HBV-A, suggesting genotypic differences in vulnerability to suppression. 28Similarly, in the case of triple infection (HIV-HDV-HBV), HIV can also influence the viral loads of the others. 28,29In our review, we identified 21 cases of HIV-HBV-HDV infection, wherein only five had paired genotyping but without a clear trend (Table S1b).For some studies, genotyping both viruses was either not the priority and/or perhaps from a region with well-known HBV or HDV genotypes regionally that they were inferred for their study.In such cases, while a high probability of being the correction assumption, we still only reported the pairings in our more precise tabulations that were definitive, but again, would have still otherwise been included in the non-specified columns.This limits how much paired data is available and explains the large numbers in overall co-infections where only one virus is genotyped.To the best of our knowledge, this review provides the most comprehensive accumulation of the data on how the two viruses tend to pair available to date.
As noted above, HDV-4 was previously classified as HDV-2b. 30is renaming may have affected some of the older data that reported on HDV-2, inflating the numbers reported if HDV-4 samples were also included.To the best of our knowledge this would account for 34 samples (20 of which fell into the paired genotype samples).

F I G U R E 1
Geographical distributions of HDV (A) and HBV (B) genotypes from HBV-HDV co-infection data.(C) Cumulative HBV genotype distribution by region for comparison to the HBV-HDV co-infected data (based on the data collated by Velkov et al. 2018; Table (i.e.where there were two different genotypes of either HDV or HBV present in the same infection).These were then counted as 26 pairings for the purposes of this data.HDV-1 had the most recorded samples (3427) compared to all other HDV genotypes.For HBV genotypes, HBV-D had the most recorded samples (1076) with HBV-E (567 samples) and HBV-A (468 samples) accounting for the majority of the remainder.The most predominant HBV-HDV pairing was between HDV-1 and HBV-D (892 samples).Further, HDV-1 was the only HDV genotype found to pair with all HBV genotypes from A to G (there were no reported HBV-H-J co-infections).HBV-A and HBV-E were found to pair with HDV-1 in a comparable number of samples (142 and 190, respectively).The two least common HDV genotypes were HDV-6 and HDV-8, with 38 and 49 samples, respectively (including unpaired data).The least common HBV genotype found (only found in paired data) was HBV-G with 12 recorded samples.HBV-G predominantly paired with HDV-1 (11 samples), with only one other paired sample being documented for HDV-7.HDV-5, HDV-6, and HDV-7 all primarily paired with HBV-E (85, 10, and 29 samples, respectively).HBV-A was determined to pair with the most diverse number of HDV genotypes, having documented pairings with all HDV genotypes except for HDV-4.In contrast, HBV-G was found to pair with the least number of HDV genotypes (only HDV-1, and HDV-7), with HBV-B and HBV-F pairing with the second least amount (3 HDV genotypes each-HBV-B with HDV-1, HDV-2, and HDV-4; and HBV-F with HDV-1, HDV-3, and HDV-5).Finally, HBV-F almost exclusively (over 92% of recorded samples) pairs with HDV-3, as 130 out of 140 paired samples of HBV-F were paired with HDV-3.
otype link clinically is by Gerber et al in 2021 examining 526 paired genotype co-infections. 25Their work derives from the French National Reference Centre for Viral Hepatitis B, C, and D (FNRC-D) the largest HDV biobank and repository worldwide.Their pairingsshowing a predominance of HDV-1 with HBV-D, are incorporated in our data, along with an additional 1421 paired cases from the literature.
Note: Bolded values in the main table highlight the most abundant HBV pairing for each HDV genotype.In the totals, both the highest and lowest numbers are indicated.
Abbreviations: HBV, hepatitis B virus; HDV, hepatitis delta virus.a Total number of samples (in 133 papers) that had genotype data for either HDV or HBV but not both paired in the multi-infection.b