This retrospective single-center study investigated the clinical, histopathological, and laboratory findings, as well as treatment regimens and outcomes of patients with biopsy-proven AIN. The most common etiology of AIN was drug use (45.2% NSAID and antibiotics) in our cohort. Higher baseline hemoglobin levels, percentage of globally sclerotic glomeruli, interstitial fibrosis, tubular atrophy, and granuloma formation on renal biopsy, in addition to a longer interval between first symptom to initiation of immunosuppressives, were associated with poor renal outcomes. Additionally, patients who developed ESKD had higher baseline lymphocyte counts and lower PLR levels than non-ESKD patients.

Anemia can be observed in AIN [2]. Defective erythropoietin (EPO) synthesis in peritubular interstitial cells due to tubulointerstitial infiltration may be responsible for this phenomenon. Interestingly, higher baseline hemoglobin levels, regardless of creatinine or eGFR levels, were associated with poor renal outcomes in our cohort. Although baseline EPO levels and hematinic parameters, including ferritin, transferrin saturation, folic acid, and vitamin B12 levels, were unknown and none of these patients were under EPO treatment; our finding could be related to paradoxical severe interstitial infiltration. Secondary polycythemia due to EPO secretion could occur in the setting of severe leukemia and lymphoma infiltration, as seen in the cases reported by Osumi et al. [7] and Bhat et al. [8]. These patients had mild renal dysfunction, bilateral renomegaly, and increased EPO levels [7, 8]. Therefore, it could be speculated that hemoglobin and EPO levels might have a U-shaped pattern in interstitial nephritis according to the degree of interstitial inflammatory cell infiltration. In addition, the percentage of female patients was higher in the responder group compared to non-responders (88.2% vs. 71.4%), which could have contributed to the lower hemoglobin levels in the responders; however, one of the most probable explanations for this observation might be statistical error due to limited number of patients in the study group.

The remaining laboratory results, including the baseline CBC and systemic inflammation indices, were similar between responders and non-responders; however, patients who developed ESKD had higher baseline lymphocyte counts and lower PLR than non-ESKD patients. Higher levels of peripheral lymphocytes may represent more severe tubulointerstitial lymphocyte infiltration. On the other hand, no statistical difference in the degree of lymphocyte infiltration in ESKD and non-ESKD patients were observed, which might be related to the limited number of patients. To the best of our knowledge, the role of PLR in AIN prognosis has not yet been studied. Higher preoperative and early postoperative PLR levels were previously shown to be associated with AKI [9]. Another study showed that PLR has a U-shaped pattern for mortality in critically ill patients with AKI [10]. The lower PLR might be a novel marker for ESKD development in patients with AIN, according to the results of our study. Nevertheless, the limited number of patients with ESKD in our cohort might have caused this finding to be inaccurate. Further studies are required to elucidate these.

It is not surprising that patients with severe tubular atrophy, global glomerular sclerosis, and granuloma formation in renal biopsies show poor treatment responses. Similar findings have previously been reported in literature [11,12,13]. Histopathological findings generally represent renal damage and the possibility of the treatment response in nephrology practice. In contrast, biomarkers could be crucial for patients with AIN in situations restraining a renal biopsy, such as coagulopathy. A recent study showed that the urinary CXCL9-to-creatinine ratio could be a specific marker for AIN diagnosis [14]. Therefore, an AIN diagnosis can be established without renal biopsy. Therefore, new prognostic serum or urine markers may be beneficial for AIN. A previous study showed that urine IL‑9 and TNF‑α might be prognostic markers in AIN [11, 15]. We evaluated baseline laboratory results, including inflammation indices, as potential prognostic markers in our study.

The optimal treatment modality for patients with AIN is still not well known. In drug-related AIN, the cessation of the accused drug may be therapeutic. An older retrospective observational study showed no additional benefit of corticosteroids in AIN compared to supportive treatment [16]. In contrast, refractory cases may require immunosuppressive treatment, mainly corticosteroids. The optimal steroid dose for AIN treatment has not yet been determined. A previous study showed no difference between the efficacy of high-dose pulse therapy and oral steroids in drug-induced AIN [17]. In addition, longer intervals to corticosteroid treatment might be associated with poor prognosis [12, 13]. A prospective open-label trial (PRAISE study) is ongoing to compare the efficacy of steroids and supportive care in AIN [18]. In our cohort, 93.5% of patients were treated with immunosuppressives, mainly corticosteroids, immediately after biopsy-proven AIN diagnosis (two patients were not treated with immunosuppressives due to the improvement of renal functions after the cessation of the accused drugs for AIN). No statistical differences were observed between the efficacy of pulse therapy and oral steroids in our responder and non-responder patients. Despite the high rates of immunosuppressant use, renal recovery rates were not encouraging in our cohort, similar to previous studies [12, 13]. It could be related to chronic irreversible damages, such as tubular atrophy, interstitial fibrosis, and global glomerulosclerosis in significant number of patients. This was possibly related to the delayed diagnosis of AIN due to its asymptomatic clinical presentation. Typical clinical findings of AIN, such as fever, eosinophilia, and rash, which were described in early publications, are rare in AIN, according to previous studies [1, 2]. Although no statistically significant difference was present in terms of time from first symptom to biopsy between the responder and non-responder groups in our study, this interval was longer in the non-responder group. Besides the paucity of typical clinical findings of AIN, late admission or late referral of patients to our center could be another reason. Our hospital is a tertiary healthcare center, and most of the patients in our study group were referred from various hospitals. Also, the interval between first symptom to onset of immunosuppressive treatment was statistically longer in non-responders than responders. Hence, delayed diagnosis and late initiation of immunosuppressants might cause poor renal prognosis despite the high rates of immunosuppressive treatment in our cohort.

The study has limitations, including its retrospective nature, heterogeneous causes of AIN, limited number of patients, and single-center study design. In addition, COVID-19 or COVID-19 vaccine-related, IgG 4-related disease-associated, and immune checkpoint inhibitor-related AIN cases were not present among our cohort. Another limitation is that the etiology of AIN cannot be fully determined in some patients. Owing to the retrospective structure of our study, data were extracted from patient files, and thorough history, such as exposure to herbal remedies, could not be evaluated. On the other hand, long-term follow-up and outcome data, detailed treatment modalities, and laboratory and histopathological findings were the major strengths of our study. To the best of our knowledge, this is the first study to investigate the role of inflammatory indices, including NLR, PLR, and SII in AIN. In our cohort, the baseline PLR was lower in patients who progressed to ESKD, compared to the other markers, regardless of baseline serum creatinine and eGFR levels; however, due to the limited number of patients, these findings should be interpreted with caution.

In conclusion, AIN is a rare heterogeneous disease. Therefore, most studies have been retrospective case series. Compared to the biopsy series spread over a long-term period [13, 16, 19,20,21], we evaluated 31 cases elaborately in terms of clinical, histopathological, and laboratory findings, as well as treatment modalities, in this single-center study in a relatively short period (2006–2021); however, treatment responses were low in AIN, and 16.1% of patients progressed to ESKD despite high rates of immunosuppressive therapy, possibly due to delayed diagnosis and late initiation of immunosuppressants. We think our findings could provide new insights into the management of AIN patients. Nevertheless, studies including larger cohorts should be conducted to investigate the predictors of renal recovery and the optimal treatment for AIN.