EPSTEIN-BARR VIRUS IS THE CAUSE OF RHEUMATOID ARTHRITIS

Aim. Many studies presented some evidence that EBV might play a role in the pathogenesis of rheumatoid arthritis. Still, there are conflicting reports concerning the existence of EBV in the synovial tissue of patients suffering from rheumatoid arthritis. This systematic review assesses the causal relationship between Epstein-Barr virus (EBV) and rheumatoid arthritis (RA) for gaining a better understanding of the pathogenesis of RA. Methods. This systematic review and meta-analysis aim to answer among other questions the following: Is there a cause effect relationship between Epstein-Barr virus and rheumatoid arthritis? The method of the conditio sine qua non relationship was used to proof the hypothesis without Epstein-Barr virus no rheumatoid arthritis. In other words, if rheumatoid arthritis is present, then Epstein-Barr virus has to be present too. The mathematical formula of the causal relationship k was used to proof the hypothesis, whether there is a cause effect relationship between Epstein-Barr virus and rheumatoid arthritis. Significance was indicated by a p-value of less than 0.05. Results. The studies analysed were able to provide convincing evidence that Epstein-Barr virus is a necessary condition (a conditio sine qua non) of rheumatoid arthritis. Furthermore, the studies analysed provide impressive evidence of a cause-effect relationship between Epstein-Barr virus and rheumatoid arthritis. Conclusion. EBV infection of human synovial tissues is a conditio sine qua non, a conditio per quam of rheumatoid arthritis. In other words, Epstein-Barr virus is the cause of rheumatoid arthritis.


INTRODUCTION
Rheumatoid arthritis (RA), a systemic, predominantly (1) CD4+ T helper type 1 (Th1)-driven disease characterized by an extensive synovial hyperplasia and infiltration by macrophages, monocytes, lymphocytes and fibroblasts. Rheumatoid arthritis is a destructive, chronic and debilitating arthritis and can cause systemic complications. RA affects more or less about 1% of the world's population (2). The prevalence of rheumatoid arthritis in men is twofold to fourfold less (3,4) than in women. The long-term prognosis of rheumatoid arthritis remains very poor. In particular, the average life expectancy of RA patients is reduced by 3 to 18 years (5). The direct costs of treatment of RA, the loss of employment and the indirect costs of disability due to RA are very high (6,7). At present there is no known cure for rheumatoid arthritis, an adequate use of various kinds of disease-modifying anti-rheumatic drugs may achieve complete remission in about 30-50% of RA patients. Many exposures investigated as possible risk factors for the development of rheumatoid arthritis such as dietary (antioxidants) factors (8) red meat protein (9), fat intake (10,11) breast feeding, the use of oral contraceptives or hormone replacement therapy (12) have shown no strong associations. Only cigarette smoking (13) has been found to increase the risk of rheumatoid arthritis. In the quest to uncover the unknown etiology of rheumatoid arthritis, viruses including Epstein-Barr virus (EBV), human herpesvirus-6, human herpesvirus-8, parvovirus (14) B19 (B19), HTLV-1, and human endogenous retroviruses-5 have all been hypothesized for many years to be involved in the pathogenesis of rheumatoid arthritis (15,16).
Epstein-Barr virus (EBV) is an ancient, ubiquitous virus determined by a 184 kbp-sized, double-stranded DNA genome which has infected probably more than 90% of the world's population (17). Many studies presented some evidence suggesting that especially EBV might play a role in the pathogenesis of RA. Among them Alspaugh and Tan (18)(19) were one of the first. RA patients have higher levels of serum antibodies against EBV (20)(21)(22)(23)(24) than normal individuals. However, due to conflicting reports concerning the existence of EBV in the synovial tissue of RA patients a cause or the cause of rheumatoid arthritis remains unknown.

MATERIAL AND METHODS
RA is an autoimmune disease characterized by progressive and more or less persistent inflammation of joints of human body. At present, prognosis of RA may be very poor in the absence of an appropriate early treatment (25) with disease-modifying antirheumatic drugs (DMARDs) like methotrexate, sulphasalazine, azathioprine, antimalarials, gold-containing compounds, D-penicillamine and cyclosporin. In particular, an additional short-term duration treatment with corticosteroid is expected to prevent progressive course of RA with erosive joint damage and functional impairment.

Statistical analysis
All statistical analyses were performed with Microsoft Excel version 14.0.7166.5000 (32-Bit) software (Microsoft GmbH, Munich, Germany). In order to increase the transparency, to correct some of the misprints of former publications and to simplify the understanding of this article several of the following lines are repeated sometimes word by word and taken from my former publications. Table  The meaning of the abbreviations at, bt, ct, dt, Nt  of the data table used are explained by a 2 by 2-table  (Table 1). In general it is (at+bt) = At, (ct+dt) = At, (at+ct) = Bt, (bt+dt) = Bt and at+bt+ct+dt=Nt. Equally, it is Bt+Bt = At + At = Nt. In this context, it is p(at)=p(At ÇBt), p(At) = p(at)+p(bt) or p(At)= p(At ÇBt)+ p(bt) =p(At ÇBt)+p(At ÇBt) while p(At) is not defined as p(at). In the same context, it is p(Bt) = p(at)+p(ct) = p(At ÇBt) +p(ct) and equally in the same respect p(Bt) = 1-p(Bt) =p(bt)+p(dt).

The 2x2
Furthermore, the joint probability of At and Bt is denoted in general by p(At ÇBt). Thus far, it is p(At ÇBt) = p(At) -p(bt) = p(Bt) -p(ct) or in other words it follows clearly that p(Bt) + p(bt) -p(ct) = p(At). In general, it is p(at)+p(ct)+p(bt)+p(dt) = 1.

The data of the studies analysed
The data of the studies analysed are presented by several tables (Table 2, Table 4, Table 6, Table 7,  Table 8, Table 9, Table 10, Table 11). The meaning of the abbreviations at, bt, ct, dt, Nt of tables is explained by a 2 by 2-table (Table 1) too. Some studies provided self-contradictory data ( Table 3, Table 5) and were not considered for a re-analysis.

Independence
In the case of independence of At and Bt it is generally valid that

Exclusion (At Excludes Bt and Vice Versa Relationship)
The mathematical formula of the exclusion relationship (28-48) (At excludes Bt and vice versa) of a population was defined as and used to proof the hypothesis: At excludes Bt and vice versa.  When using data to perform some analysis, several conditions must be taken into consideration. Unfortunately, not all data are appropriate for detailed analysis. Due to formal mathematical requirements it is possible to identify data as self-contradictory and it is necessary to exclude these data from further analysis. The reason for the self-contradiction of the data is marked in bold numbers/letters. These studies were not considered for further analysis even if all these studies supported the hypothesis without EBV VCA IgG sero-positivity no RA. The term #DIV/0! denote the case that there is a division by zero.  The reason for the self-contradiction of the data above is marked in bold numbers/letters. . These studies were not considered for further analysis even if most of these studies supported the hypothesis without EBV EBNA IgG sero-positivity no RA. #DIV/0! denotes the case that there is a division by zero.

Necessary Condition (Conditio Sine Qua Non)
The mathematical formula of the necessary condition relationship (28-48) (conditio sine qua non) of a population was defined as and used to proof the hypothesis: without At no Bt .

Sufficient Condition (Conditio per Quam)
The mathematical formula of the sufficient condition relationship (28-48) (conditio per quam) of a population was defined as and used to proof the hypothesis: if At then Bt .

The X² Goodness of Fit Test of a Necessary Condition
Under conditions where the chi-square goodness (8-48) of fit test cannot be used it is possible to use an approximate and conservative (one sided) confidence interval known as the rule of three. Using the continuity correction , the chi-square value of a conditio sine qua non distribution before changes to

The X² Goodness of Fit Test of the Exclusion Relationship
The chi square value with degree of freedom 2-1=1of the exclusion relationship (28-48) with a continuity correction can be calculated as The chi square Goodness of Fit Test of the exclusion relationship examines how well observed data are compared with the expected theoretical distribution of an exclusion relationship.

The Mathematical Formula of the Causal Relationship k
The mathematical formula of the causal relationship  k is defined at every single event, at every single Bernoulli trial t, as where At denotes the cause and Bt denotes the effect.
The chi-square distribution can be applied to determine the significance of causal relationship k. Pearson's (49) concept of correlation (50) is not identical with causation (28,36,37). Causation as such is not identical with correlation. This has been proven many times and is widely discussed in many publications (51).

The 95% Confidence Interval of the Causal Relationship k
A confidence interval (CI) of the causal relationship k calculated from the statistics of the observed data can help to estimate the true value of an unknown population parameter with a certain probability. Under some conditions, the 95% interval for the causal relationship k is derived (47) as

Hypergeometric distribution
The hypergeometric distribution with its own and very long history (52,53,54,55) is defined by the parameters population size, event count in population, sample size and can be used to calculate the exact probability of an event even for small samples which are drawn from relatively small populations, without replacement.
The hypergeometric distribution differs from the binomial distribution. In contrast to the hypergeometric distribution, the probability of a binomially distributed random variable is the same from trial to trial.
The probability of having exactly at (Table 1) successes or the significance of the causal relationship k can be tested under conditions of sampling without replacement by the hypergeometric distribution (56) as

Odds Ratio
The odds ratio (OR) is given57,58,59 by It is necessary to point to the case were ct=0. Under conditions were ct=0, there is a conditio sine qua non relationship between At and Bt while the Odds ratio collapses. To date, it is not generally accepted to divide by zero.
The Odds ratio cannot speak about the natural, profound and far reaching conditio sine qua non relationship but must pass over in silence on this relationship. Pagano & Gauvreau (60) are quietly returning through the back door to circumvent this fundamental problem of Odds ratio by adding (60) 0.5 to the cells at, bt, ct, dt.
This simple way to circumvent the inconsistency and spectacular methodological incompleteness of Odds ratio is fundamentally misleading. To date, a substantial amount of research is analyzed by the Odds ratio. The more serious difficulty of this point of view is that it appears to be impossible to rely on Odds ratio in principle.
Furthermore, under conditions were bt=0, a conditio per quam relationship between At and Bt is given while the Odds ratio collapses again.
For this reason, the Odds ratio is overshadowed by a deep theoretical inconsistency and appears not to be grounded on a seemingly sound piece of reasoning.  More likely, the Odds ratio (OR) is nothing more but Yule's coefficient of association (6)1 Q re-written (62) in a non-normalized form and given by Under conditions where Yule's coefficient of association (Yule, 1900) Q = 0, there is no association. Although severely and justifiably criticized especially by Karl Pearson (1857Pearson ( -1925, the long-time and rarely challenged leader of statistical science and Heron (63), Odds ratio is still regularly referred to. The standard error and 95% confidence interval of the Odds ratio (OR) can be calculated according to Altman (64). Given the severely limited character of odds ratio, the standard error of the log Odds ratio is calculated as while the term ((N-n)/(N-1)) denotes the finite population correction (67).
p-value and works when the dataset analyzed is large enough (n ~ 30 and more). An approximate and conservative (one sided) confidence interval as discussed by Rumke (69)

RESULTS
Rheumatoid arthritis is an inflammatory progressive disease with more or less a very poor prognosis. In this context, the studies (74-99) considered for a re-analysis should help us to get a better understanding of this disease.

Without EBV IgG antibody positivity no rheumatoid arthritis
EBV VCA IgG antibodies can be used to investigate the relationship between EBV and RA.

Claims Null hypothesis: (no causal relationship)
The presence of EBV VCA IgG antibodies is a necessary condition (a conditio sine qua non) of rheumatoid arthritis. In other words, the sample distribution agrees with the hypothetical (theoretical) distribution of a necessary condition.

Alternative hypothesis: (causal relationship)
The presence of EBV VCA IgG antibodies is not a necessary condition (a conditio sine qua non) of rheumatoid arthritis. In other words, the sample distribution does not agree with the hypothetical (theoretical) distribution of a necessary condition. The significance level (Alpha) below which the null hypothesis will be rejected is alpha=0.05.

Proof
The data reviewed by this article which investigated the relationship between EBV VCA IgG antibodies and rheumatoid arthritis are presented by Table 2. In total, 9 studies with 2507 cases and controls provided non self-contradictory data and were meta-analysed while the level of significance was alpha = 0.05. In particular, all studies provided significant evidence of a conditio sine qua non relationship between EBV VCA IgG antibodies and rheumatoid arthritis (X²(Calculated [conditio sine qua non]) =0.8597 and is less than X² (Critical [conditio sine

The Chi Square Distribution
The following critical values (65,66) of the chi square distribution (68) as visualized by Table 13 are used in this publication. The rule of three The Chi-square goodness of fit test (68) used to test whether a sample distribution is identical with a theoretical distribution yields only an approximate qua non]) =16.919). In fact, the presence of EBV VCA IgG antibodies is a necessary condition (a conditio sine qua non) of rheumatoid arthritis. Ultimately, for this reason, without the presence of EBV VCA IgG antibodies no rheumatoid arthritis.

Null hypothesis
The presence of EBV EBNA IgG antibodies is a necessary condition (a conditio sine qua non) of rheumatoid arthritis. In other words, the sample distribution agrees with the hypothetical (theoretical) distribution of a necessary condition.

Alternative hypothesis
The presence of EBV EBNA IgG antibodies is not a necessary condition (a conditio sine qua non) of rheumatoid arthritis. In other words, the sample distribution does not agree with the hypothetical (theoretical) distribution of a necessary condition.
The significance level (Alpha) below which the null hypothesis will be rejected is alpha=0.05.

Proof
The data reviewed by this article which investigated the relationship between EBV EBNA IgG antibodies and rheumatoid arthritis are shown in Table  3. At this point it might be important that 7 studies with 794 cases and controls provided non self-contradictory data and were considered for a meta-analysis while the level of significance was alpha=0.05. We can point to the fact that all 7 studies (Table 4) provided significant evidence of a conditio sine qua non relationship between EBV EBNA IgG antibodies and rheumatoid arthritis (X² (Calculated [conditio sine qua non]) = 3.1435 and is less than X² (Critical [conditio sine qua non]) = 14.0671). By that very fact, the presence of EBV EBNA IgG antibodies is a necessary condition (a conditio sine qua non) of rheumatoid arthritis. The last point suggests that without the presence of EBV EBNA IgG antibodies no rheumatoid arthritis.

EBV is a cause of rheumatoid arthritis
(The Study of Saal et al. (Table 10)) The presence of EBV DNA in synovial tissues is a possible method to show an etiological link be-tween EBV and the pathogenesis of rheumatoid arthritis. Several studies published convincing results on this topic.

Claims Null hypothesis: (no causal relationship)
There is no significant causal relationship between an infection by Epstein-Barr virus and rheumatoid arthritis. (k=0).

Alternative hypothesis: (causal relationship)
There is a significant causal relationship between an infection by Epstein-Barr virus and rheumatoid arthritis. (k¹0).
Conditions. Alpha level = 5%. The two tailed critical Chi square value (degrees of freedom = 1) for alpha level 5% is 3.841458821.

Proof
The data for this hypothesis test were provided by Saal et al. (Table 10) and are illustrated by the Table 10. The causal relationship k(Epstein-Barr virus, rheumatoid arthritis) was calculated as k = +0.2954 (p value (k) = 9.29228E-05; 95% CI (k) = [0.1213;0.4695]) while the level of significance was alpha=0.05. The data of Saal et al. (Table 10) provide evidence that EBV is a sufficient condition (X²(IMP) = 1.5203; X² Critical (IMP) = 3.841458821) of rheumatoid arthritis while the cause effect relationship between EBV and RA is highly significant (k = +0.2954 (p value (k) = 9.29228E-05).

EBV is a cause of rheumatoid arthritis
(The Study of Takeda et al. (Table 11))

Claims Null hypothesis: (no causal relationship)
There is no significant causal relationship between an infection by Epstein-Barr virus and rheumatoid arthritis. (k=0).

Alternative hypothesis: (causal relationship)
There is a significant causal relationship between an infection by Epstein-Barr virus and rheumatoid arthritis. (k¹0).
Conditions. Alpha level = 5%. The two tailed critical Chi square value (degrees of freedom = 1) for alpha level 5% is 3.841458821.

Proof
The data for this hypothesis test were provided by Study of Takeda et al. (Table 11) and are illustrat-ed by the  11) provide evidence that EBV is a sufficient condition (X²(IMP) = 0.0167; X² Critical (IMP) = 3.841458821) of rheumatoid arthritis while the cause effect relationship between EBV and RA is highly significant (k = +0.5470 (p value (k) = 6.07959E-06).

EBV is the cause of rheumatoid arthritis
The Study of Chiu et al. (Table 12)

Claims Null hypothesis: (no causal relationship)
There is no significant causal relationship between an infection by Epstein-Barr virus and rheumatoid arthritis. (k=0).

Alternative hypothesis: (causal relationship)
There is a significant causal relationship between an infection by Epstein-Barr virus and rheumatoid arthritis. (k¹0).
Conditions. Alpha level = 5%. The two tailed critical Chi square value (degrees of freedom = 1) for alpha level 5% is 3.841458821.

DISCUSSION
Epstein-Barr Virus discovered 1964 by Epstein et al. (100) is a widely disseminated lymphotropic herpes virus. As key results, several studies suspected that particularly Epstein-Barr virus is involved in etiology of rheumatoid arthritis. Catalano et al. (21) reported that patients with RA had a significantly higher frequency and titer of rheumatoid arthritis-associated nuclear antigen (anti-RANA) antibodies than did control subjects and confirmed the previous results of Alspaugh and Tan (18). Using the protein blot technique, Billings et al. (23) were able to provide evidence that rheumatoid arthritis nuclear antigen (RANA) and Epstein-Barr virus nuclear antigen identify the same polypeptide.
However, data about EBV burden in RA patients reported have been contradictory and the role of EBV still remains elusive. Indeed, on this matter, as with so many other major medical issues, several re-views101, 102 and meta-analysis were not able to find a definite solution on this fundamental topic. Thus far, it is not excluded that this meta-analyses is susceptible to different kind of publication bias. In its broadest sense, the studies analysed differ in various aspect. Thus, the question arises why not all patients were diagnosed according to the American Rheumatism Association 1987 revised criteria for the classification 103 of RA. While some studies considered for a meta-analysis provided no diagnostic criteria for the diagnosis of rheumatoid arthritis other studies utilised a form of the American College of Rheumatology (ACR) or American Rheumatology Association criteria. Additionally, reporting of data of some studies are to some extent unsatisfactory, because not all studies provided detailed cut-off values for EBV sero-positivity. RA patients and non-RA controls both were tested quantitatively for different antibodies against Epstein-Barr virus while using different substrates or kits or antigens and various technologies. Hence we need to take into consideration under what conditions is it appropriate to use antibodies against Epstein-Barr virus to investigate the relationship between EBV and rheumatoid arthritis? To date it is known that IgG molecules with two antigen binding sites are created and released by human plasma B cells not without any reason but i. e. to control an infection in human body. Especially IgM, IgG et cetera molecules are not existing for ever but suffer a kind of pharmacokinetics. The half-live (104) for total IgG was found to be 25.8 days. In this context EBV antibodies are major components of human humoral immunity allowing controlling an EBV infection of human body tissues through several mechanisms. A natural concern is whether EBV antibodies suffer a turnover rate with regard to the infectious status. Several factors can influence the pharmacokinetics of EBV antibodies. The half-lives for antibodies specific for Epstein-Barr virus antigens depend on EBV infection status. In the case of recent EBV infection or during the course of EBV reactivation the humoral response of human immune system against EBV antigens will lead to different changes in antibodies specific for Epstein-Barr virus antigens. An acute EBV (re-) infection is indicated by the presence of VCA IgM and VCA IgG but without EBNA-1 IgG. Typical for a past EBV infection is the presence of VCA IgG and EBNA-1 IgG but without VCA IgM (105).
At the very least, enough is published to convince our self that after a primary EBV infection, EBV persists for life in vivo in a quiescent state in resting human memory B cells (106) which circulate in the peripheral blood. This fact considerably leads to the conclusion that VCA IgG or EBNA IgG provide evidence of an EBV infection of human body and are therefore helpful in causal analysis. And yet, despite contradictory results several studies give convincing evidence of the linkage between EBV and RA. Many studies demonstrated remarkable higher levels of different serum antibodies against Epstein-Barr virus in patients suffering from rheumatoid arthritis than in healthy controls (21,22,24,76,107,108,109). Baecklund et al. (110) provided evidence that a high inflammatory activity of RA rather than the treatment of RA is a major risk determinant of lymphoma in a subset of patients with RA. Sherina Sherina et al. (99) in the context of a very large and impressive epidemiological study in patients with RA and in non-RA subjects. Sherina et al. used the presence of anti-viral antibodies as surrogate markers for viral infection.
The data of Sherina et al. (99) with a sample size of n= 1690 cases and controls concerning the relationship between parvovirus B19 and rheumatoid arthritis (Table 7) were not self-contradictory and could be used for further analysis. The data of Sherina et al. (99) (Takeda et al., 2000) provided evidence of a highly significant cause effect relationship between an infection of human joints by EBV and RA (k =+0.5470; p value (k) =6.07959E-06).
Using real-time polymerase chain reaction Balandraud et al. (116) were able to document that Epstein-Barr virus DNA load in the peripheral blood (116) of patients with rheumatoid arthritis was increased almost 10-fold .

In-situ hybridization
In-situ hybridization (ISH), has been described in the year 1969 by Joseph G. Gall (117). According to Fan & Gulley (118), In situ hybridization (ISH) to Epstein-Barr virus (EBV)-encoded RNA (EBER) is an appropriate method to detect and localize EBV DNA in biopsy samples of rheumatoid arthritis patients and healthy controls. Like any other method, even the in situ hybridization is not completely free of bias and can be labelled with some severe limita-tions. The study group of Chiu et al. (96) (Table 12) conducted a study to investigate the expression of Epstein-Barr virus-encoded small RNA1 (EBER1) by ISH in the synovial tissues taken from 23 patients with rheumatoid arthritis and 13 patients with OA. The RA patients were diagnosed according to the American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis (103). All synovial samples from RA showed positive expression of EBER1 (23/23,100%), but none of the control group patients (0/13).
Mehraein et al. (119) investigated the influence of synovial virus infections in rheumatoid arthritis, and found evidence of increased synovial persistence of EBV in 5/29 rheumatoid arthritis (RA) patients.
Mahabadi et al. (98) investigated Epstein-Barr virus DNA by PCR in synovial fluid of 50 rheumatoid arthritis patients and detected EBV DNA by PCR in 30 cases (60%). Mahabadi et al. (98) concluded that EBV may play a role in the pathogenesis of RA.A control group was not provided and it was not possible to consider the data for causal analysis.
Although it has been investigated and speculated for over 40 years that Epstein-Barr virus is a strong candidate to contribute to the cause of RA definite evidence was wanting. Considering the half-life (120) of EBV antibodies and the results of the reviews (121) mentioned, the studies re-analysed in the present article indicate a high degree of confidence that an EBV infection is the cause RA and the etiology of rheumatoid arthritis no longer remains unknown. The lack of appropriate ancient medical texts regarding rheumatoid arthritis has forced many researchers to acknowledge the first description of RA by modern medicine to Augustin Jacob Landré-Beauvais (122,123) from the year 1800 published in his dissertation. In the year 2018 and about 218 years later, the cause of rheumatoid arthritis is finally identified.

CONCLUSION
The results of the present study are consistent with the hypothesis that there is a relationship between EBV and RA and give further evidence of the linkage between EBV and RA. The data not only do support the hypothesis that EBV infection is some-