Prevalence and sociodemographic correlates of antinuclear antibody testing by indirect immunofluorescence or solid-phase assays in a Spanish population: the Camargo Cohort

Autoantibodies are the hallmark of autoimmunity, and specifically, antinuclear antibodies (ANA) are one of the most relevant antibodies present in systemic autoimmune diseases (AID). The presence of ANA, together with anti-dsDNA antibodies and extractable nuclear antigens (ENA), is important for the classification, diagnosis, and monitoring of patients with systemic AID. The first evidence of the existence of ANA was the description of the LE-cell phenomenon by Hargraves et al. in 1948 [1]. However, it was not until the early 1950s, due to the research performed by Coons, Kaplan, and Weller, when indirect immunofluorescence (IIF), using cryopreserved sections of rodent tissues as substrate, began to be used for the detection of ANA. Later on, in the mid-70s, it was discovered that human tissue culture cells, such as human epithelial type-2 cells (HEp-2 cells) derived from laryngeal carcinoma, were better than primary organ sections for the detection of ANA. This is so because the production of this type of cell in large numbers was easier and they had bigger nuclei and expressed antigens in various stages of the cellular cycle [2].

Despite being characteristic of systemic AID, ANA are not specific since they may be present in organ-specific AID [3], infections [4, 5], inflammatory disorders [6], neoplasms [7], or even in healthy individuals [8, 9].

Currently, the IIF assay using HEp-2 cells as substrate, due to its high sensitivity, continues to be one of the recommended methods for the study of ANA, as expressed in the position statement of the American College of Rheumatology (ACR), published in 2010, and also by the European League Against Rheumatism, EULAR in 2019 [10]. Likewise, as demonstrated in a recent publication by the Spanish group of autoimmunity (GEAI), it continues to be the most widely used method for ANA detection [11]. Nevertheless, in recent years, several publications have questioned this position and question whether IIF should be replaced by solid-phase assays (SPA) [12‐16]. Despite all this technological development for the study of ANA, the 2019 EULAR/ACR classification criteria for SLE maintain the presence of ANAs using HEp-2 cells or an equivalent positive test as an entry criterion for SLE patients [17]. However, it should be noted that the IIF assay using HEp-2 cells is a laborious process, that needs skilled operators, and therefore, the reliability of the results is directly related to the experience of the technician.

In comparison with the IIF assay, SPAs are less sensitive but more specific for the detection of ANA in systemic AID [18, 19]. In addition, different studies have shown that combining the results obtained with both techniques (IIF and SPA) is more accurate than performing either of the two tests separately [14, 16, 20]. In fact, the single presence of ANA by IIF is not specific of AID, since these autoantibodies appear in up to 30% of healthy individuals, a proportion that increases with age, especially in women [8, 21‐23].

However, all of this evidence generally comes from well-characterized cohorts of patients with systemic AIDs, and yet there are scarce data on the value of ANAs detected by IIF and SPA in unselected populations and especially in cohorts with a low pretest probability. This is a very important aspect to consider due to how the request for ANA has evolved over the years. Historically, only rheumatologists and clinical immunologists ordered ANA for diagnosis of systemic AID. Over time, several clinicians, such as nephrologists, internists, and gastroenterologists joined in requesting ANA, and currently, almost any clinician and even general practitioners should request them [24]. Initially, as a consequence of this increase in the number of requested tests, the pre-test probability for ANA detection was very high, and in recent years, it has been significantly reduced, reaching currently a very low pre-test probability. Therefore, due to this low pre-test probability, the post-test probability for ANA positivity is also low. Taking into account this aspect, we consider that it is of great importance to study the presence of ANA through IIF and SPA in a low pre-test probability population for systemic AID, such as the Camargo Cohort [25] since ANA detection by IIF may be better when the pre-test probability is high, whereas SPA techniques are more useful in populations with an overall low pre-test probability for systemic AID.

A less studied aspect is whether there is a relationship between the prevalence of ANA and the sociodemographic features of a population. Although some reports suggest a higher prevalence of ANA in females and older individuals [26, 27], the association between ANA and other sociodemographic factors is largely unknown. Therefore, in the present work, we evaluate the relationship between ANA and sociodemographic and biobehavioral factors in the Camargo Cohort, a well-established population with a low pre-test probability for systemic AID, which could reflect the current trend regarding the request for ANA.

The determination of ANA is very important for the diagnosis of patients with suspected systemic AID. However, less is known about the true value of these autoantibodies in populations with a low pre-test probability for systemic AID, and even less about their association with sociodemographic features in these populations.

The present study provides an estimation of the prevalence of ANA, using an IIF assay and two different SPAs, according to different sociodemographic factors and with selected biobehavioral measures. Our results show an overall prevalence of ANA of 25.8% by IIF at 1/160 serum-dilution, 13.6% using ANA Screen, and 7.9% using CTD Screen, which is higher in the case of IIF in comparison to other published reports in healthy selected populations [35‐37]. These differences likely relate to the characteristic of the different study populations and the variations in ANA assessment across laboratories. The higher prevalence of ANA detected by IIF assay using HEp-2 cells, in comparison with SPA, could be due to the fact that HEp-2 cells have approximately 100–150 possible autoantigens, which can be considered the bigger array of autoantigens, while not all specific autoantigens are included in SPAs. Likewise, independently of the methodology used for ANA detection, the frequency of positive results was higher in women than in men. Besides, considering the age of the participants, the older the subject the higher the prevalence of ANA, regardless of the method, although this increase was more evident when SPAs (ANA Screen and CTD Screen) were used [22, 38, 39]. This linear tendency was not observed using IIF, probably due to the presence of ANA patterns that are not associated with any detectable antigenic specificity, as reported in some studies [40]. Furthermore, in line with the results published by other authors, ANA prevalence did not vary with the education level [35].

Considering the association between ANA with selected biobehavioral variables, unlike other studies in which it was observed that the presence of ANA is less common in overweight and obese individuals than in normal-weight individuals [35], we did not observe differences in the prevalence of ANA according to the BMI categories. However, the prevalence of these autoantibodies varied depending on the physical activity and the alcohol intake of the participants. Thus, the prevalence was higher in those individuals with a sedentary lifestyle and in ex- and non-alcohol users compared to current alcohol consumers. The results related to physical activity could be important because the presence of ANA has been associated with decreased carotid elasticity, suggesting that mechanisms resulting in ANA production may be involved in the development of early atherosclerosis [41]. Other authors have also shown that moderate alcohol consumption has been associated with decreased systemic lupus erythematosus (SLE) risk [42]. Likewise, it has also been described in several studies that alcohol acts as a protective factor against the development of other AID, such as diabetes [43], multiple sclerosis, [44] or especially, rheumatoid arthritis (RA) [45, 46]. In the case of RA, upon alcohol exposure, different factors have been identified to be reduced, such as antigen-presentation, T cell activation capacity of antigen-presenting cells, B cell maturation and proliferation, IL-21 production by T follicular helper cells, antigen-specific IgG, and proinflammatory cytokines. Moreover, the Th2 immune response, M2 macrophage function, and anti-inflammatory cytokines (IL-10, TGF-β) [47] seem to be increased. All these factors could be associated with a reduced risk of RA. In the present study, we observed an increasing, albeit non-significant, trend in the antibody load in non-alcohol consumers compared to current drinkers.

In addition, despite some studies suggesting smoking as a risk factor for SLE, RA, and other AID [48], we found a higher frequency of ANA-positive results in non-smoker subjects independently of the screening method used. Finally, using IIF and CTD Screen, a significant increase in the prevalence of ANA was found in those subjects with higher serum CRP levels, which could indicate the involvement of systemic inflammation in the development of systemic AID. In the same way, this finding could be related to the processes of immunosenescence and autoimmunity that are observed in aging people, in whom, the development of autoreactive T cells, as well as an increase in the production of autoantibodies, has been described [49‐51]. Moreover, there is a relationship between immunosenescence and the development of a senescence-associated secretory phenotype (SASP), which has been associated with malignancy progression as well as with autoimmune disorders [51‐54].

One could speculate that our findings regarding the association of ANA with different sociodemographic and behavioral factors might be related to the sample size or even by chance. However, in our study, the greater the specificity of the test [55], the greater the statistical significance of some relevant sociodemographic variables analyzed, such as age, alcohol intake, physical activity, and serum CRP levels (as a subrogate marker of inflammation). Due to the scarcity of data and these potential associations, further studies are needed to confirm our findings.

To our knowledge, this is the first study in which the relevance of the autoantibody load is considered, as has been described in antiphospholipid syndrome (APS). In the APS, depending on the number, the type, and the titer of antiphospholipid antibodies, it is possible to establish different risk profiles [55]. In our study, focusing on the relevance of the antibody load using ANA Screen and their associations with selected sociodemographic features and biobehavioral measures showed that among subjects with antibody loads 1 (one autoantibody) and ≥2 (two or more autoantibodies), in addition to the prevalence of ANA, the prevalence of the antibody load also increases with age.

Our study has some limitations. Firstly, those inherent to a cross-sectional study. Secondly. the evaluation of antigenic specificities in those patients with ANA positive results in the screening assay using SPA was only possible using the ANA screen and not the CTD screen. Thirdly, not all types of autoantibodies were assessed. Finally, the small simple size of certain subgroups may have limited our power to detect differences in ANA prevalence for some factors. Nevertheless, our study also has several strengths, such as the large number of subjects included belonging to a well-established population cohort with a low pre-test probability for systemic AID, and a long-term follow-up, which could reflect the current trend regarding the request for ANA and allow us to draw more robust conclusions. There is no other study in which so many subjects have been included, with such a long follow-up period, where the association between the presence of ANA and selected sociodemographic features and biobehavioral measures had been evaluated.

In conclusion, the prevalence of ANA is higher using IIF assay than SPA in a population-based cohort with a low pretest probability for AIDs. However, independently of the technique used, the prevalence of ANA is higher in females and older subjects. Considering biobehavioral measures, ANA-positive results were frequent in those individuals with a sedentary lifestyle, as well as in ex-and non-alcohol consumers. Finally, it is important to highlight the study of the antibody load, since it may be relevant when establishing the risk profile of population-based subjects.