Predicting gastrointestinal and renal involvement in adult IgA vasculitis

Background Immunoglobulin A vasculitis (IgAV) is still poorly defined in the adult population. We aimed to determine the predictors of gastrointestinal (GI) or renal involvement in adult IgAV. Methods The prospective study included histologically proven adult IgAV cases diagnosed between January 2013 and July 2019 at our secondary/tertiary rheumatology center. We evaluated the role of clinical and the laboratory parameters as markers predicting the GI or renal involvement in IgAV, using the multiple logistic regression analysis. Results During the 79-month observation period, we identified 214 new adult IgAV cases (59.3% males, median (interquartile range) age 64.6 (57.2–76.7) years). The GI tract and renal involvement developed in 58 (27.1%) and 83 (38.8%) cases, respectively (concurrently in 26 (12.1%) cases). In the multivariate logistic regression analysis, generalized purpura (OR 6.74 (95%CI 3.18–14.31)), the pre-treatment neutrophil to lymphocyte ratio (NLR) > 3.5 (OR 2.78 (95%CI 1.34–5.75)), and elevated serum IgA levels (OR 0.40 (95%CI 0.20–0.79)) were extracted as factors associated with GI complications, whereas current smoking (OR 3.23 (95%CI 1.50–6.98)), generalized purpura (OR 1.98 (95%CI 1.08–3.61)), elevated serum IgA (OR 2.25 (95%CI 1.21–4.18)), NLR > 3.5 (OR 1.96 (95%CI 1.02–3.77)), and marginally age (1.02 (95%CI 1.01–1.04)) emerged as factors associated with renal complications. Conclusion Generalized purpura and pre-treatment NLR predicted both GI and renal involvement, whereas active smoking was associated with renal involvement, and the serum IgA level had a divergent effect on renal and GI involvement in adult IgAV.


Introduction
Immunoglobulin A vasculitis (IgA) is an immune complex small vessel leukocytoclastic vasculitis that commonly affects the skin, joints, gastrointestinal (GI) tract, and kidneys [1]. IgAV is a typical childhood vasculitis, commonly with a benign, self-limiting course and complete recovery in the young population. In adults, a severe acute disease with significant GI and renal involvement is common [2,3]. GI involvement with a severe hemorrhage or bowel perforation represents the major risk of mortality in acute adult IgAV, and renal involvement is associated with an increased risk of progression to chronic kidney failure [2]. Nevertheless, the predictors of visceral involvement in acute adult IgAV have not been extensively studied. Cao et al. reported that the risk of significant renal disease increased with patient age, and the presence of necrotic or bullous skin lesions [4]. Patient age forecasted visceral involvement also in the study by Poterucha et al. Adults under the age of 40 had an increased risk of GI involvement, and those over 40 without eosinophils in a skin biopsy had an increased risk of renal involvement [5]. Contrary to St John et al., who reported a predictive value of skin lesions on the upper and lower extremities for significant long-term renal involvement [6], the distribution of skin lesions above the waist was not found to be a reliable indicator of systemic disease in adult IgAV in the study by Poterucha et al. [5]. Nagy et al. described in a retrospective study a significant association between pretreatment elevated neutrophil to lymphocyte ratio (NLR) and the development of GI or renal manifestations [7]. A similar association between an elevated NLR and severe GI or nephritis has recently been described also in childhood IgAV [8].
The aim of our study was to determine the predictors of gastrointestinal and renal involvement in a prospectively collected cohort of adult IgAV.

Setting and patient selection
This prospective study was conducted at the Department of Rheumatology, University Medical Centre Ljubljana, Slovenia. Patients with suspected vasculitis are regularly referred by their general practitioners or other subspecialists to our early intervention clinic, where they are evaluated on the day of the referral, obviating the need for immunosuppressive treatment before referral.
We included adults (i.e., persons aged ≥ 18 years), with the symptoms and signs compatible with the definition of IgAV according to the 2012 revised International Chapel Hill Consensus Conference Nomenclature of Vasculitides [1], with histologically proven IgAV, diagnosed for the first time between January 2013 and August 2019. Additionally, all patients fulfilled the classification criteria of the European League Against Rheumatism/Paediatric Rheumatology International Trials Organization/Paediatric Rheumatology European Society (EULAR/PRINTO/PRES) criteria for IgAV [9].

Clinical, laboratory, and histological data
Patients underwent a detailed clinical evaluation (including a structured history of smoking and potential triggers of IgAV, i.e., infection or the use of new medication within a month prior to an IgAV episode, and cancer history), and an extensive laboratory workup, as described previously [10]. The definitions used for the assessment of the skin, renal, and gastrointestinal involvement are provided in Table 1.
The serum immunoglobulin A (IgA) concentration, pre-treatment white blood cell count (WBC), and NLR were measured. Since concurrent infection could influence the NLR, concurrent infections were recorded and included in the analysis.
Skin or renal biopsies were evaluated using bright field microscopy, and direct immunofluorescence.

Statistical analysis
The results were expressed as a median and interquartile range (IQR) or mean and standard deviation (SD) for metric, and as percentages for categorical variables. To test the differences between IgAV subgroups, we used the Mann-Whitney test for metric and Fisher's exact test for categorical variables. A receiver operating characteristic (ROC) curve was constructed to evaluate the prognostic utility of NLR and to determine the optimal cutoff value. Potential predictors of IgAV GI and renal involvement with a p value < 0.2 were tested using a multiple logistic regression analysis. The significance threshold selected for the final model was set at 0.05.
All 214 patients fulfilled EULAR/PRINTO/PRES classification criteria for IgAV [9]. In addition to skin involvement, in all 214 patients, the skin biopsy was performed and was consistent with IgAV. Sixty-seven out of 214 (31.3%) patients had skin limited IgAV (i.e., fulfilling mandatory skin criterion and histological criterion); the remaining 68.7% fulfilled at least one additional item of EULAR/ PRINTO/PRES classification criteria for IgAV (i.e., articular or GI or renal item). Fifteen patients had daily proteinuria from 1.0 to 3.5 g. Nephrotic range proteinuria was detected in 13 patients. Acute kidney injury developed as a result of vasculitis in 23 (10.7%) patients. Overall, 30 (14.0%) patients had a severe course of kidney involvement, defined as the development of acute kidney injury or nephrotic syndrome. A renal biopsy was performed in 9 (4.2%) patients. Proliferative glomerulonephritis was noted in 5 patients, crescentic glomerulonephritis in 3 cases, and predominantly sclerotic glomerular lesions in one patient.
The characteristics of IgAV patients with and without GI or renal involvement are presented in Table 2. The subgroup of patients with renal involvement was significantly older compared to the subgroup with GI involvement (median (IQR) age 67.8 (53.6-77.4) vs. 55.5 (39.0-74.0) years, p = 0.007). The age-related distributions in patients with GI involvement and renal involvement are presented in Fig. 1.
We measured the serum concentration of IgA in all patients. The pre-treatment IgA serum concentration exceeded the upper limit of reference in 102 (47.7%) cases. The median (IQR) concentration in patients with an elevated serum IgA level was 5.26 (4.60-6.70) g/l (normal range 0.61-3.56 g/l). No patient had a subnormal serum IgA concentration. IgAV patients with renal involvement had significantly higher serum IgA concentrations compared to the subgroup with a GI involvement (median (IQR) of 4.4 (3.2-5.6) vs. 3.6 (2.4-4.6) g/l, p = 0.008).
Median (IQR) pre-treatment NLR in our IgAV cohort was 4.0 (2.6-5.9), range 0.8-55.0. The ROC curve analysis demonstrated NLR > 3.5 as the best cut-off value for the prediction of concurrent GI or renal involvement with 68.7% sensitivity, 52.5% specificity, and the area under the curve of 0.606 (95%CI 0.537-0.672).

IgAV treatment
We treated the patients in line with the local practice.
Indications for systemic immunosuppressive treatment were necrotic purpura, any GI involvement, or severe kidney involvement. IgAV spontaneously remitted in 48 (22.4%) cases. Topical steroids were the only treatment in 21 (9.8%) patients. We treated 61 (28.5%) patients with intravenous methylprednisolone pulses (MP; 125-1000 mg qd, for three consecutive days). In 17 patients, MP pulses were the only treatment, while the remaining 44 patients continued with oral glucocorticoids. We treated a total of 129 (60.3%) patients with oral MP in a
Disease duration prior to the diagnosis, patient gender, history of prior infection, use of new medication, history of cancer, and necrotic skin purpura at presentation were not significantly associated with GI or renal involvement in adult IgAV.

Discussion
In the present study, we evaluated predictors of GI and renal involvement, the two most commonly affected visceral organs in IgAV [10], in a prospectively collected cohort of adult IgAV patients. Although IgAV is usually considered a childhood disease with GI involvement reported in 50-75%, and renal involvement in 40-50% of cases [11], the disease is not uncommon in adults [12]. In addition, the disease is frequently severe in adults, compared to commonly uneventful disease course in children [2].
In the majority of cases, GI and kidney inflammation followed the development of skin lesions. Particularly, renal involvement can be delayed for several weeks and can be missed [13] and can progress to chronic kidney disease [2,3]. In addition, severe GI involvement with bowel ischemia that leads to tissue necrosis and perforation or major bleeding represents the major source of morbidity and mortality in the acute phase of IgAV [14]. The risk for this devastating complication is especially increased in the elderly and the polymorbid patients. Therefore, the identification of markers predicting the two most frequent visceral manifestations of IgAV would considerably ease the routine clinical work.
Our study shows that a generalized purpura (i.e., purpura above the waistline), and not skin necroses, predicts both renal and GI involvement in adult IgAV. As previously shown, generalized purpura in smokers also predicted the severity of kidney and GI involvement [15]. Our present findings are in contrast with those of Cao et al., and Poterucha et al., who found a correlation between necrotic purpura and visceral involvement, and the absence of a correlation with the extent of skin lesions and visceral involvement, respectively [4,5]. Nevertheless, similarly to Poterucha et al., we noticed a trend toward opposing age-related distribution between patients with GI and those with renal involvementadult patients with GI involvement being significantly younger than those with renal involvement [5].
Active smoking emerged as a potential risk factor for renal involvement in adult IgAV. The association between smoking and the progression of chronic kidney disease has been appreciated in diabetic and several non-diabetic chronic kidney diseases, including IgA nephropathy [16,17]. Cha et al. demonstrated a link between the progression of IgA nephropathy and smoking-related injury of the microvasculature [16]. In animal and in vitro models, nicotine activated inflammatory mediators, induced mesangial cell proliferation, and fibronectin production [18].
In our study, the serum IgA concentrations were divergent in patients with GI and renal involvement. Patients with the elevated serum IgA seem to have a lower risk of GI tract involvement but a higher risk of renal involvement. A lower frequency of elevated serum IgA in GI involvement has not been previously reported. Data in the literature is scarce; nevertheless, a recent study in children found significantly lower serum IgA in IgAV patients with scrotal involvement versus those without this manifestation [19]. And a study on biomarkers of childhood IgAV nephritis showed no significant differences in the serum IgA level comparing cases with and without nephritis [20]. The predictive roles of elevated pre-treatment NLR as a marker of systemic involvement in IgAV and a marker reflecting IgAV inflammatory response have been reported. Nagy et al. found that pre-treatment NLR > 3.3 correlated with the severity of the disease in patients who developed GI and renal involvement [7]. Similarly, we also found that higher NLR was associated with GI and renal involvement. The optimal cut-off in our cohort was NLR > 3.5. However, as it had poor sensitivity and specificity, it should not be interpreted in isolation.
Our study had several strengths. We prospectively investigated 214 adult patients with IgAV proven by skin biopsy and elucidated risk factors of severe involvement. However, the present study also had limitations: (1) cumulative smoking exposure, (2) a factor of passive smoking, and (3) long-term IgAV prognosis have not been analyzed due to lack of information. Because of the small number of patients who underwent renal biopsy, the histological features of kidney involvement could not be analyzed.

Conclusion
Gastrointestinal or renal involvement in adult IgAV could be predicted by smoking status, distribution of skin lesions, and the neutrophil to lymphocyte ratio. The presented findings could help physicians identify the patients at risk who need vigilant monitoring during the acute phase of the disease.