This retrospective cohort study enrolled 61 patients with AoSD and 61 age and sex-matched patients with FMS. All of the AoSD patients were diagnosed initially in the Rheumatology Outpatient clinic at Pamukkale University School of Medicine from 2010 to 2018 based on Yamaguchi diagnostic criteria [7]. These criteria require the presence of in total five features, at least two of which must be major for the diagnosis of AoSD after excluding other mimicker diseases such as malignancy, infectious and inflammatory diseases. Arthralgia, fever, typical rash and leukocytosis are defined as major criteria whereas sore throat, presence of lymphadenopathy and/or splenomegaly, abnormal liver function tests and absence of rheumatoid factor and Anti-Nuclear Antibody (ANA) are minor criteria.

The data of all AoSD and FMS patients who were on regular follow-up between January 2010 and December 2018, in the tertiary health centre was collected in this study. The analysis of the collected data was done between April and July 2020. This study was performed after local Ethics Committee approval of the Pamukkale University Faculty of Medicine (date/no: 03/2020-06) and in accordance with the Helsinki declaration. Informed consent form was obtained from each patient.

Inclusion criteria: AoSD diagnosis fulfilling the Yamaguchi criteria [7] or FMS diagnosis according to the current classification criteria [8] and having regular follow-up for at least one year for both patient groups.

Exclusion criteria: It includes other inflammatory, infectious or malignant disease, using anti-inflammatory drugs at diagnosis, onset of the disease before 18-year-old and insufficient medical record for both patient groups.

NLR and PLR were calculated based on baseline laboratory values at the diagnosis before initiation of corticosteroid treatment to prevent the effects of glucocorticoids on blood leukocyte count and distribution. Pouchot score was specifically used in AoSD patients to assess disease activity based on symptoms, physical examination findings and laboratory results. Total score varied between 0 and 12, with each of the clinical manifestations being assigned as 1 point: typical rash, sore throat, fever, leukocytosis ≥15,000, hepatomegaly, splenomegaly, lymphadenopathy, myalgia, abdominal pain, pneumonia, pleuritis and pericarditis [9].

Patients were categorised in three different clinical courses as self-limited, intermittent and chronic courses at follow-up [2]. If there was a single attack, it was defined as a self-limited course. If there was more than one attack with relapses, this was evaluated as intermittent course. Persistent active disease and joint symptoms longer than one year were accepted as chronic course. In this context, remission was accepted as no disease activity according to the global assessment of the physician and laboratory parameters [10].

Primary outcome was the association between NLR, PLR, Pouchot score and the three different disease courses in AoSD patients. Secondary outcome was the relationship between Pouchot score and inflammatory markers in AoSD patients.

All statistical analyses were performed using Statistical Package for Social Sciences (SPSS) 25.0 software. Kolmogorov Smirnov and Shapiro Wilk tests were used for determination of normal distribution. Continuous variables were defined as the mean±standard deviation and median (minimum-maximum values), and categorical variables were defined by number and percentages. For independent group comparisons, we used independent samples t-test when parametric test assumptions were met and Mann-Whitney U test and Kruskal Wallis Variance Analysis were used when parametric test assumptions were not met. Differences between categorical variables were analysed using Chi-square test. A p-value ≤0.05 was considered statistically significant.

The majority of patients were female in both groups (57.4% (n=35) of patients with AoSD vs 72.1% (n=44) of patients with FMS) [Table/Fig-1]. The median age was 39 years in AoSD patients and 43 years in FMS patients (p=0.666) [Table/Fig-1].

The mean follow-up of the 61 patients with AoSD was 74 months (range, 14-169). Eighteen patients (29.5%) had a self-limited disease course, nine patients (14.8%) an intermittent disease course and 34 (55.7%) a chronic disease course. The baseline Pouchot score was ≥4 in 72.1% (n=44) of AoSD patients [Table/Fig-2].

AoSD patients had significantly higher serum NLR, PLR and lower mean platelet volume (MPV) values than FMS patients {6.68 (1.67-19.7), 1.83 (1.1-4) p=0.0001; 187 (82.9-549), 114 (72-246) p=0.0001; 8.3 (6.4-11.3), 9.3 (7.7-11.7) p=0.0001, respectively} [Table/Fig-1]. ESR and CRP were also at higher titers than in FMS [Table/Fig-1]. Baseline NLR, PLR and Pouchot score were similar between three subgroups according to the disease course [Table/Fig-3]. There was no statistical difference in terms of age and gender in these three subgroups (p>0.05 vs p>0.05, respectively). The majority of patients in the self-limited and chronic course groups had higher baseline Pouchot score. The NLR and CRP were significantly higher in AoSD patients with active disease than inactive disease {7.02 (1.8-19.7), 4.17 (1.67-14.8) p=0.06; 13 (1.9-29.5), 9 (1.6-20.9) p=0.046, respectively} [Table/Fig-4,5]. PLR and ESR did not show a statistically significant difference between active and inactive AoSD patients [Table/Fig-5].

This study was designed to reveal any association between disease activity, baseline inflammatory markers and subsquent disease course in order to find a novel measure for prediction of disease course. The majority of patients in the self-limited and chronic course groups had higher baseline Pouchot score. The baseline NLR and CRP were higher in active than inactive AoSD patients with a statistical significance. Pouchot J et al., also showed that the most common course is a chronic pattern (36%), followed by self-limited (34%) and intermittent (24%) patterns among 62 AoSD patients with 70 months mean follow-up time [9].

In a large multicentre cohort including 356 patients with 22 months median follow-up, male gender, delayed diagnosis, arthritis involving wrist/elbow joints and failure to achieve remission with initial treatment were poor prognostic factors for developing chronic disease [11]. In addition, 76.5% of patients (n=26) who developed a chronic course had baseline active disease based on Pouchot score in this study. Also, the mean NLR and PLR were higher in the vast majority of the chronic course group, though not statistically signficant. These statistical inconsistencies may be due to a very limited patient size in each subgroup, especially in the intermittent group.

NLR and PLR have been investigated in many rheumatic diseases as predictors of disease activity. Pan L et al., observed a positive correlation between these inflammatory ratios and Kerr score in patients with Takayasu disease [12]. Two studies including AoSD patients had conflicting results. First, Seo JY et al., investigated NLR as a diagnostic marker and its predictive value for relapse of AoSD. A cut-off value of 3.08 showed the greatest sensitivity (91.7%), specificity (68.4%) and AUC value (0.967) as a diagnostic tool for AoSD [13]. In the second study, Zhang M et al., noted that although PLR was not different between groups, patients with sepsis showed significantly increased NLR compared to AoSD patients [14]. In this cohort, the NLR was significantly higher in active AoSD patients compared with inactive disease in a support of previous literature.

CRP has been defined as an useful marker for monitoring disease activity in AoSD patients due to the main role of Interleukin-6 (IL-6) in the pathogenesis of the disease [15]. In addition, elevated level of IL-18 was detected in active AoSD patients and correlated with disease activity [16]. IL-18 is a member of the IL-1 family and is also well-known to be involved in haematopoietic progenitor cell growth. The other discriminating factor of active and non-active. AoSD patients was elevated CRP titers in this study. IL-18 and IL-6 could lead to this inflammatory response, although the levels could not be analysed due to the retrospective nature of the study.

Platelet size is another useful predictive biomarker of inflammatory processes that may provide important information on the course and prognosis of a variety of pro-inflammatory diseases [17]. Liu JP et al., showed MPV to be at lower titers in AoSD patients than those with sepsis (mean MPV was 9.8±1.2 fL) [18]. Zhang M et al., also showed that a lower value of MPV may be a quantitative assessment tool for the diagnosis of AoSD and other Fever of Unknown Origin (FUO) [19]. Similar to previous reports, in this study the mean MPV values were lower in AoSD patients than FMS patients as a result of inflammation.

Due to the low incidence rate and being a single centre experience, the sample size was quite limited. Simultaneously, analysis of pro-inflammatory markers that may show disease activity such as IL-18 could not obtained due to retrospective nature of study.

The majority of patients in the self-limited and chronic course groups had higher baseline Pouchot score. The baseline NLR and CRP were higher in active AoSD patients compared with inactive AoSD patients. These findings suggest that NLR may be a useful, easy and less costly index for active patients. Further well-designed prospective, randomised-controlled studies in larger homogeneous populations are required to define risk factors for accepted disease courses.