Current practice in laboratory diagnostics of autoimmune diseases in Croatia. 
Survey of the Working group for laboratory diagnostics of autoimmune diseases of the Croatian Society of Medical Biochemistry and Laboratory Medicine

Introduction

Laboratory medicine of autoimmune diseases is a growing discipline of laboratory medicine in general, and many laboratories expand their panel of tests with new methodologies for autoantibody detection. Methods in laboratory medicine of autoimmune diseases differ in many aspects like the antigen profiles, sensitivity and specificity (1). Limitations like the lack of standardized reference samples of defined specificities and different sources of antigens (recombinant vs. native) are the main source of results discrepancy between different methods. Very heterogeneous manner of result reporting is the consequence of both lack of nomenclature uniformity and consensus regarding result interpretation. Finally, the lack of experience and education of laboratory personnel in the diagnostics of certain autoimmune diseases, as well as an unawareness of a method limitation often results in confusion for clinicians. Therefore, standardization and harmonization in this field of laboratory diagnostic is essential (2). Efforts were made to provide the recommendations for humoral autoimmune diagnostics that recently resulted with International recommendations for the assessment of anti-nuclear antibody testing published by the European Autoimmunity Standardization Initiative (EASI) (3). For most systemic and organ-specific autoimmune diseases diagnostic criteria or classification criteria have been introduced (4-12). Autoantibodies included in those criteria are an important tool for diagnosis of autoimmune diseases. This prompted the initiative from Croatian Society of Medical Biochemistry and Laboratory Medicine (CSMBLM) to form a Working group (WG) for laboratory diagnostics of autoimmune diseases with following goals: to screen the current practice in performing laboratory medicine of autoimmune diseases in Croatia, to identify the main problems and accordingly propose the recommendations that would improve and harmonize laboratory medicine of autoimmune diseases. The aim of the study was to gather information about the methodology, interpretation and algorithms used in laboratory medicine of autoimmune diseases in Croatia. WG designed two-step questionnaires in order to obtain that information. This study summarises the outcome of these questionnaires and provides the insight into the main problems and sources of heterogeneity in performance of autoantibody testing.

Materials and methods

In order to obtain information about the current practice in laboratory diagnostics of autoimmune diseases in Croatia two surveys were conducted using SurveyMonkey application (SurveyMonkey Inc., Palo Alto, California, USA). The first survey was a questionnaire that contained 17 questions with predefined answers that aimed to identify laboratories which perform diagnostics of autoimmune diseases and to obtain general information like laboratory organization, number of tests per month, turnaround time (TAT), etc. The contacts of laboratory managers of all medical biochemistry laboratories in Croatia were obtained from Croatian Chamber of Medical Biochemists. The survey was sent to all laboratories within secondary healthcare facilities (general and county hospitals), tertiary healthcare facilities (university hospital/university hospital centre), private laboratories and specialized health institutions (total number of 88 laboratories). The survey was sent to the e-mail addresses of the laboratory managers with a suggestion that a laboratory professional responsible for laboratory diagnostics of autoimmune diseases should complete the survey. In case that laboratory does not perform any of the tests listed in question two of the first survey, survey participants were instructed to skip all following questions and to conclude the survey. The link for the survey was distributed via e-mails on September 26^th 2014 and was closed on April 22^nd 2015. The laboratories within primary healthcare facilities were excluded from the survey considering that they do not perform the tests for laboratory diagnostics of autoimmune diseases. After the collection of the responses, from 33 laboratories that declared to perform diagnostic from the scope, 19 were selected for the second survey based on the tests they pleaded to perform. This survey was created in the same application and contained 25 questions in order to obtain more detailed information regarding detection of selected tests. The link for the second survey was also sent via e-mails on October 22^nd 2015 and was closed on November 28^th 2015.

For more transparent review, questions and offered answers from both questionnaires are presented together in Table 1.

In total, both questionnaires included 42 questions that could roughly be divided in 6 categories: laboratory organisation, definition of the tests from the scope and organisation of routine work (N = 8), antinuclear antibodies (ANA), antibodies to double stranded DNA (anti-dsDNA) and antibodies to extractible nuclear antigens (anti-ENA) testing (N = 20), anti-neutrophil cytoplasmic antibodies (ANCA) testing (N = 6), antiphospholipid (aPL) antibodies testing (N = 3), celiac disease (N = 3), autoimmune liver diseases testing (N = 2).

Results

Discussion

Results of conducted survey confirmed not only high heterogeneity in the performance of autoantibody testing among laboratories in Croatia, but also obsolete in some cases. This includes issues with result reporting, determination of serological markers that are not recommended, methods of determination of certain markers that do not comply with existing international recommendations or the application of very heterogeneous algorithms among laboratories.

Most of the laboratories perform tests for immunodiagnostic of systemic autoimmune rheumatic diseases (SARD), predominantly serological markers of rheumatoid arthritis, RF and anti-CCP. This result was expected taking into account a few facts: a) RA is the most common systemic autoimmune disease, b) both markers are part of RA classification criteria (5), c) both markers are determined using methods set to automatic analysers, which already exist in most laboratories and do not require additional equipment. These automated methods for determining RF and anti-CCP are responsible for the fact that many laboratories (12 of 14) report the analysis results in this field of laboratory diagnostics within 24 hours. After RF and anti-CCP, ANA is the most commonly performed test (17 / 80 laboratories in total). In comparison with the study conducted by the EASI group (13) regarding the implementation of the ANA test in 12 European countries, the number of laboratories performing ANA regarding the total population in Croatia is comparable with France, the Netherlands and Switzerland. Half of the laboratories in Croatia that determine ANA belong to University hospital or University hospital centres while in most European countries to general hospitals or private institutions (with the exception of Norway and Sweden, where the ratio is equal). In order to meet the increasing demand for ANA determination, reducing costs and compensating the lack of experienced personnel for interpretation of ANA fluorescence, the preference to the alternative solid phase automated assays with very diverse sensitivity and specificity is often given over IIF screening method. The main reason for this diversity lies in the fact that commercial kits from different manufactures use different methods, mixture of purified extracts or recombinant antigens, and number of antigens included in mixture (14, 15). Using IIF method on Hep-2, larger number of nuclear, nucleolar and cytoplasmic antigens could be detected when compared to solid phase assays where detection is possible only for antigens included in the mixture. Sensitivity and specificity of ANA-screen using enzyme-based immunoassay (EIA) were examined for nine manufactures kits in study of Tan et al. (16). Results showed good sensitivity and specificity with the need for improvement in detection of anti-dsDNA and anti-Sm in some kits. Sensitivity and specificity using different methods and comparing them to IIF, show different results depending on used method (14-17). Due to significant number of false negative results when using these methods, IIF method on Hep-2 cells is still considered a reference method for ANA screening (3, 14). However, one should keep in mind that choice of ANA detection method is actually disease-dependent. For example, ANA is the serologic marker of autoimmune hepatitis but without clearly associated antigen specificity and therefore IIF method is the only one applicable in this case (10). On the other hand, myositis is associated with Jo-1 antibodies which can be accurately detected only with solid phase assays (2, 14). Since laboratory staff is mostly unaware of suspected diagnosis, some authors support the opinion that multiple screening tests should be used in order to improve the diagnostic of autoimmune diseases but with appropriate comments included for interpretation of discrepant results (18, 19). With the share of only 7 / 17 laboratories that perform ANA- screening with IIF method, Croatia is next-to-last in comparison to European countries that participated in EASI group Survey (13). In fact, in most European countries IIF method is applied in > 90% of laboratories with the exception of the Netherlands, Portugal, Norway and Ukraine (with 65%, 64%, 50% and 40% share respectively). The manner of conducting ANA-screen with IIF in Croatia is rather homogenous, in the sense that all laboratories recognize and report at least 4 regular fluorescence patterns and all except one provide titre and interpretative comments. Given that low ANA titres are common in the older population as well as in some other diseases in addition to SARD, qualitative ANA result does not give sufficient clinical information (14). Beside four regular nuclear patterns, cytoplasmic fluorescence patterns are very common and usually address the presence of other antibodies such as antimitochondrial (AMA) or smooth muscle antibodies (SMA) associated with autoimmune liver diseases (3). Also, cytoplasmic fluorescence can reflect the presence of other important SARD-associated antibodies such as those targeting ribosomal-P proteins or Ro52 (14). Therefore, negligence of cytoplasmic fluorescence can have clinical consequences such as delay in proper diagnosis. Although initial (screening) dilutions most commonly present throughout literature are 1 / 80 or 1 / 160, it is generally recommended that the laboratory appoints the initial dilution that corresponds to 95th percentile of local healthy population (with respective shares regarding age and gender) (1, 3). However, very often the IIF reagent kit manufacturer proposed initial dilution is applied. Given the fact that Croatia is a small market, the same reagent manufacturer probably recommends the initial dilution 1 / 100 used by 6 / 7 laboratories. Distinct heterogeneity still exists in the titration level of positive samples with only two laboratories following the recommendation of reporting the last dilution which still shows reactivity (3). In addition, heterogeneity exists in reporting different patterns with different fluorescence intensity observed in one sample. Positive finding is that all except one laboratory performing ANA-screen by IIF method consider the opportunity of more than one observer in the case of a doubtful result. Surprisingly poor result of the survey was that half (5 / 10) of laboratories that use solid phase assays for ANA-screening do not specify antigens included in the assay (3). If the clinician is not aware that applied method covers only the limited number of antigens, one can consider it equivalent to IIF screening assay. Commercial solid phase assays usually include clinically most relevant antigens such as SS-A/Ro60, SS-B/La, Sm, U1RNP, dsDNA, Scl-70, Jo-1 and CENP. However, some antigens clinically relevant for systemic sclerosis, such as RNA-polymerase III or some nucleolar antigens such as PM-ScL or fibrillarin are rarely included in these assays (3, 8). The same refers to histones as the target antigens for ANA associated with drug induced lupus, which are rarely included in ANA-screen solid phase assays so that the false negative ANA result can lead to misdiagnose (3).

Most laboratories that perform ANA testing also perform anti-ENA testing. However, the approach to ANA and anti-ENA testing is quite heterogeneous and even pointless in some cases (ANA solid assay followed with ENA-screen assay). According to current practice, anti-ENA determination in Croatia is somewhat lower (77%) in comparison to other European countries (except Ukraine) where more than 83% of laboratories that perform ANA testing also perform anti-ENA testing (13, 20). At least, the most relevant ENA antigens (SS-A/Ro60, SS-B/La, Sm, U1RNP, Scl-70, and Jo-1) are included in all ENA-screen assays. Regarding the identification of ENA specificities, all laboratories determine antibodies to SS-A/Ro60, SS-B/La, RNP and Scl-70 while one laboratory does not test for Sm and Jo-1 antibodies. In general, the extent of ENA-specificities determination is lower in Croatia in comparison to other European countries but the covered specificities are comparable (13). Distinction between anti-SS-A/Ro60 and anti-Ro52 (TRIM21) antibodies was found in 42% (5 / 12) of Croatian laboratories which determines ENA specificities, in comparison to only 20–25% of laboratories in other European countries. On the one hand it could be attributed to the higher awareness of different clinical relevance regarding monoreactivity to Ro52 antigen (in myositis and systemic sclerosis) or SS-A/Ro60 antigen (in subacute lupus) (21). On the other hand, some laboratories do not use single tests for specificity confirmation but in order to reduce the cost and time for analysis, rather use the panel of different specificities. Therefore, it could be that number of individual specificities is simply dictated by the available commercial products. Another interesting finding is that 3 / 12 laboratories determine RNP70 specificity either with or without U1RNP specificity (encompasses proteins A, C, and 70kD). Although the anti-U1RNP immune response targets all three protein components (70kDa, A, C), individual reactivity to 70kDa protein may be more specific for mixed connective tissue disease (MCTD) as it is less frequent in SLE than those to proteins A or C (22). However, the best sensitivity and specificity for MCTD seems to be achieved with identification of both U1RNP and RNP70 (23).

In spite of recommendation to use additional method in the case of result discrepancy between anti-ENA and ANA or with clinical suspicion, only one laboratory has the opportunity of using alternative method for anti-ENA (3).

Regarding anti-dsDNA testing, none of laboratories use Farr assay (radioimmunoassay, RIA) that is based on precipitation of antibody-antigen complexes using ammonium sulphate and the quantification of antibody in the precipitate using a radio-labelled dsDNA antigen. Farr assay is considered to be the gold standard but it is also minimally present in laboratories throughout Europe, most probably due to use of radioisotopes and laboriousness of the method (24). CLIFT as a second choice according to specificity is variable present (26–94%) in European laboratories while in Croatia only one laboratory uses this method (13). However, no quantitative method for detection of anti-dsDNA is in use in this laboratory so that the patient cannot be followed. Although all except one laboratory report quantitative result for anti-dsDNA there is considerable heterogeneity in reporting units. Since commercial tests are calibrated against the 1st International Standard for anti-dsDNA coded Wo/80, results should be given in International Units (IU/mL).

The majority of laboratories do not apply algorithms for testing anti-ENA or anti-dsDNA based on ANA-screen result but strictly adhere to the requested tests. The most probable explanation is that the Croatian health reimbursement policy does not consider reflex testing.

Most laboratories do not apply any rules regarding minimal retesting interval. This issue primarily refers to frequent requests for ANA retesting or for retesting of previously confirmed ENA-specificity. Repeat testing for ANA titre does not have confirmed value in assessing the disease activity and retesting for ENA-specificity can be justified only in the case of the appearance of new symptoms (2). Our data are comparable to the results of Canadian study that pointed out unnecessary repeat testing and simultaneous ordering of multiple tests (ANA, ENA, dsDNA) (25).

General impression of the Survey results regarding ANA testing is that the consultative role of laboratory experts in immunodiagnostic workout is poorly involved or limited only to an interpretative comment, eventually.

Testing for antiphospholipid antibodies is performed in 15% (12 / 80) of laboratories within tertiary and secondary health care facilities and private health care institutions. The minority of laboratories covers determination of all antibodies included in APS-classification criteria (9). It can partially be explained with the policy in some hospitals to perform determination of LA within the framework of transfusion medicine instead of medical biochemistry laboratory. There is considerable heterogeneity regarding reporting of the units for anticardiolipin antibodies with almost half of laboratories not following the International guidelines on anticardiolipin testing (26).

Assessment of antibodies associated with systemic vasculitis and glomerulonephritis is performed in 17% (14 / 80) of included laboratories. The current approach of ANCA determination in Croatian laboratories implies quantitative (in all but one laboratory) determination of PR3 and MPO specificities only, while ANCA-screen is performed in minority of laboratories. All three laboratories that use IIF method for ANCA-screen apply the same algorithm with obligatory testing for PR3/MPO specificity in ANCA IIF positive samples and also recognize and report the fluorescence patterns in line with International consensus statement on testing and reporting of ANCA (27, 28). Use of qualitative test for PR3/MPO specificities is limited to diagnosis purpose in emergency cases and, unlike quantitative test, has no value in monitoring the inflammatory activity of vasculitis.

Assessment of autoantibodies associated with autoimmune liver diseases is poorly present in Croatian laboratories and is mostly limited to laboratories within University hospitals. There is considerable heterogeneity in the approach to determination of SMA and LKM antibodies by IIF method concerning reporting titre. Namely, half of laboratories do not report titre for SMA while for anti-LKM all except one do not report titre. This policy is not in line with the guidelines of International Autoimmune Hepatitis Group which includes a Scoring System for diagnosis of autoimmune hepatitis based on the titre of relevant autoantibodies (ANA, SMA, LKM) (29-31). For the same reason, use of solely solid assays for detection of anti-LKM antibodies is not suitable, but should rather be used for confirmation of LKM-1 reactivity in samples with LKM-typical fluorescence pattern. Although, determination of AMA titre has no approved clinical value, half of laboratories report titre.

Among antibodies associated with gastrointestinal autoimmune diseases, determination of those associated with celiac disease are most frequently performed, with anti-tTG IgA, as the first line test, being the most prominent. Antibodies to native gliadin (AGA) are still sporadically determined although all the relevant guidelines discourage use of these antibodies in diagnosis of celiac disease (12, 32). Algorithms applied in serologic diagnosis of celiac disease are heterogeneous and almost half laboratories do not follow the guidelines on the determination of IgG class of antibodies or total IgA (12).

Regarding antibodies associated with inflammatory bowel diseases (IBD) it is important to point out that ANCA associated with these diseases targets a wide range of possible antigens but not PR3 and MPO (33). IBD associated ANCA are commonly detected only by IIF method without confirmation of specificity. Using the term “ANCA” for specific anti-PR3 or anti-MPO tests can be misleading for clinician seeking for ANCA in IBD. ANCA specificities relevant for IBD are not detected in the Croatian public medical laboratories and also rarely in private laboratories.

Determination of antibodies associated with autoimmune diseases affecting nervous system is most rarely performed and are limited only to antibodies in paraneoplastic syndrome.

Considering participation in EQA, the results clearly show that not enough laboratories participate in EQA schemes. This could partially be explained by the fact that the national program for external control provided by the Croatian Centre for Quality Assessment in Laboratory Medicine (CROQALM) does not cover tests from this scope. Participation in EQA is especially important for autoantibody testing since these methods are not standardized and use exclusively manufacturer - dependent quality control samples which makes the objective evidence of laboratory competence limited (34). Participation in EQA offers the opportunity for laboratory to compare own results with those obtained with other participants that use same or different methodologies (35). The efforts to motivate or obligate laboratories to participate in EQA schemes should be made with a help of national professional associations. This process could also be facilitated with the decision to participate in the accreditation process according to HRN EN ISO 15189 in the field of humoral laboratory diagnostic of autoimmune diseases.

Considering all challenges that our study indicates, the Italian multicentre study showed how development and implementation of diagnostics algorithms and recommendations contribute to diagnostic specificity of autoimmune rheumatic diseases (36). Limitation of this study could be attributed to all the disadvantages of using online survey to collect data about current practice in laboratory diagnostics of autoimmune diseases in Croatia. The main disadvantage is that the answers to the questions depend on respondents’ honesty regarding the real situation in laboratory practice of autoimmune diseases diagnostics. The other limitation of the study is limited number of organ-specific autoimmune diseases included.

In conclusion, the survey results indicate the heterogeneity in the manner of conducting the humoral immunodiagnostic of autoimmune diseases throughout laboratories in Croatia. This heterogeneity refers not only to the use of different methods, analytical platforms or algorithms but also to the use of the same method and to the policy of results reporting. Such heterogeneity on the level of laboratory diagnostics most certainly has an impact on the clinical level. Therefore, presented data stressed out the necessity for recommendations on national level with the aim to harmonize this specific branch of laboratory diagnostic. In issuing recommendations, currently used methodologies as well as the national health reimbursement policy, which for example, does not consider reflex testing should be taken into the consideration.