Acute Kidney Injury from Mononuclear Cell-Predominated Interstitial Nephritis After Introduction of a Glucagon-Like Peptide-1 Receptor Agonist: A Case Report

Introduction

The prevalence of diabetes mellitus is increasing, leading to a decrease in quality of life, a shorter life expectancy, and an enormous economic burden globally [1]. Long-term complications of diabetes include macrovascular coronary, cerebral, or peripheral arterial diseases and microvascular complications. Of these complications, cardiovascular disease, heart failure, diabetic kidney disease, and cerebrovascular events pose a significant mortality risk. Therefore, diabetes management that prevents these complications is preferred, such as anti-diabetic medications with pleiotropic effects, to reduce these events [2]. Considering clinical practice guidelines on diabetes management, such as the American Diabetes Association’s Standards of Medical Care in Diabetes, sodium–glucose cotransporter 2 (SGLT2) inhibitors and glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are suggested in addition to metformin in patients with or at high risk for atherosclerotic cardiovascular disease, heart failure, or chronic kidney disease due to multiple studies on their pleiotropic effect on cardiometabolic and renal protection [3]. Moreover, GLP-1 RAs can significantly reduce blood pressure and body weight, further modifying cardiovascular risk factors [4,5]. Regarding the established benefit in delaying the progression of diabetic kidney disease and ultimately preventing end-stage kidney disease, Kidney Disease: Improving Global Outcomes (KDIGO) recommends adding GLP-1 RAs with proven cardiorenal benefit, such as injected Semaglutide, dulaglutide, or liraglutide, to achieve the glycemic target [6]. These endorse the global increase in the prescription of GLP-1 RAs in recent years.

Despite the renoprotective effect of all classes of medications suggested by the international guidelines, rare adverse effects on the kidney can complicate treatment in some patients, causing a paradoxical decline in kidney function instead of a delay in glomerular filtration rate (GFR) decline. This has been shown in cases of excess GFR decline with renin-angiotensin-aldosterone inhibitors or the osmotic nephrosis from SGLT2 inhibitors, requiring cautious monitoring of clinical condition and renal function after drug initiation or titration [7]. Considering GLP-1 RAs, most adverse events include self-remitting gastrointestinal symptoms, and reported kidney injury from medication is rare [8]. Acute kidney injury (AKI) after GLP-1 RAs prescription ranges from mild to severe, which requires dialysis in some patients. The main proposed mechanism of GLP-1 RA-induced AKI is dehydration from vomiting or poor fluid intake. However, rare cases of biopsy-proven acute interstitial nephritis (AIN) have been reported with uncertain insights into the pathophysiology [9–11]. The low incidence of reported adverse outcomes could be due, at least in part, to the short duration of widespread medication use as compared to other drugs for diabetic kidney disease, resulting in the lack of consensus on an appropriate monitoring protocol and typical manifestations of the disease. In this paper, we report a case of biopsy-proven AIN after dulaglutide administration, with some clinical features differing from those previously reported.

Case Report

A 63-year-old woman with diabetic kidney disease stage 3bA3 (baseline serum creatinine 1.7 mg/dL and estimated GFR according to the equation of the Collaboration of Chronic Kidney Disease Epidemiology Collaboration equation 31 mL/min/1.73 m2), difficultly controlled hypertension, and depressive disorder had an increase in serum creatinine to 2.7 mg/dL detected in a follow-up examination in October to November 2023. She had good glycemic control but had persistent albuminuria despite receiving the maximum dose of renin-angiotensin-aldosterone inhibitors. Two months before the event (August 2023), subcutaneous dulaglutide 0.75 mg/week was prescribed to further alleviate albuminuria instead of an SGLT2 inhibitor due to concerns about frequent urinary tract infections (Figure 1). Oliguria and hypertension were not present along with kidney injury. Her body weight, blood pressure control, and albuminuria status improved after dulaglutide administration (urine albumin creatinine ratio decreased from 7528 mg/g to 4178 mg/g). No signs of volume depletion were detected. Her other medications included azilsartan 80 mg/day, chlorthalidone 75 mg/day, verapamil (sustained release) 480 mg/day, sodium bicarbonate 1200 mg/day, ergocalciferol 20,000 unit/week, atorvastatin 20 mg/day, hydralazine 50 mg/day, Ketosteril® 4 tablets/day, folic acid 5 mg/day, calcium polystyrene sulfonate 5 g/day, intravenous iron sucrose 100 mg every 4 weeks, mirtazapine 15 mg/day, and vitamin B complex supplement. Testing revealed unremarkable liver function and no eosinophilia. Blood electrolyte profile showed sodium 137 mmol/L, potassium 5.2 mmol/L, chloride 104 mmol/L, and bicarbonate 21 mmol/L. Urinalysis revealed specific gravity 1.009, pH 7.0, protein 2+, leukocyte 3+ (numerous white blood cells per high power field), negative nitrite, no bacteria, and blood 3+ (30–50 red blood cells per high power field without significant dysmorphic change). Anti-double-stranded deoxyribonucleic acid, anti-neutrophil cytoplasmic antibodies, serum complements, and viral serologic test results were negative.

The differential diagnoses of her kidney injury without the presence of glomerular hematuria included: (1) decreased renal perfusion (prerenal azotemia) due to decreased appetite combined with the concurrent use of diuretics and the renin-angiotensin-aldosterone system inhibitor; (2) AIN, as supported by the presence of leukocyturia without bacteria; and (3) a postrenal cause such as diabetic bladder, which was later excluded by renal ultrasound, which revealed no hydronephrosis. Therefore, the azilsartan and chlorthalidone dosages were tapered off due to the possibility of causing azotemia in a mild volume depletion state, but serum creatinine was still rising to 3.66 mg/dL. At this stage, a kidney biopsy was performed and the pathology result showed diabetic changes in the glomeruli and renal vessels, together with tubulointerstitial infiltration of lymphocytes and plasma cells, with few eosinophils. Sixty percent of the glomeruli (12 of 20) were globally sclerosed, and 50% interstitial fibrosis and tubular atrophy was observed. There was no evidence of immune complex deposition in immunofluorescence and electron microscopy (Figure 2). Despite the scarcity of reports on AIN associated with dulaglutide use, the medication was discontinued due to its most recent introduction before the onset of kidney injury, and renal function began to improve gradually after drug discontinuation without prescribing steroid therapy because the patient had frequent urinary tract infections. Azilsartan and chlorthalidone could then be successfully rechallenged after almost complete recovery of kidney function due to the need for long-term preservation of renal function and albuminuria suppression. The follow-up serum creatinine and estimated GFR were then stable at around 2.0 mg/dL and 24 mL/min/1.73 m2, respectively, which were close to the baseline levels.

Discussion

We report a case of a woman with proteinuric diabetic kidney disease developing biopsy-proven AIN with lymphocytic predominance instead of predominated eosinophilic infiltration, as evidenced in previous drug-induced AIN reports. Despite the lack of a standard method for identifying the drug causing AIN, the recently challenged dulaglutide is the most likely cause in this patient due to the following reasons: (1) the subacute time course from the drug initiation and the onset of kidney injury, compatible with the immune-related process of AIN; and (2) the ability to successfully rechallenge other medications that could lead to declined GFR without causing recurrent kidney injury, such as a renin-angiotensin-aldosterone antagonist and a diuretic. To date, AIN has been rarely reported as a complication associated with GLP-1 RAs. The manifestation may range from an asymptomatic presentation to severe forms in which dialysis was initiated (Table 1). All reported cases, including our patient, have non-specific symptoms and no oliguria. There was no hypo/hypertension, and typical extrarenal manifestations of AIN, such as rash, transaminitis, arthralgia, and eosinophilia, were not present. Urinalysis in AIN can range from minimal abnormalities to leukocyturia and microscopic hematuria, as in our case [12]. AKI can develop from 2 days to 2 years after the initiation of RA with GLP-1, and the median onset was 63 days [13]. These non-specific and various manifestations require a prompt kidney biopsy in any patient developing unexplained AKI after initiation of this class of medication, even without obvious extrarenal clues.

The mechanistic study for the pathophysiology of AIN, especially from novel drugs such as GLP1-RAs, is not completely plausible due to the lack of an animal model closely resembles humans. Therefore, information from multiple case reports and extrapolation from AIN from other medication classes are still necessary to guide future management. Although volume depletion from vomiting in some patients is the most common cause of AKI after GLP-1 RAs, the medication can cause kidney injury via interstitial nephritis from drug-specific T-cell-mediated hypersensitivity [14,15]. Indeed, GLP1-RAs have been demonstrated by a previous animal study to have a beneficial anti-inflammatory effect [16,17]. Therefore, the paradoxical hypersensitivity that causes AIN is most likely idiosyncratic. Other uncertainties include the exact inflammatory pathway for different GLP1-RA. Referring to general pathophysiology of AIN in other medication classes, different cytokines and effector cells after specific antigen presentation to the naïve CD4+ T-helper cells give the unique pathological finding for each causative medication such as macrophage predomination from interferon-gamma and interleukin-4 in antibiotic-associated AIN, T-helper 17 cell predomination in proton pump inhibitor-associated AIN, or eosinophil predomination driven mainly by interleukin-5 from various medications [18]. In contrast to previous case reports of GLP1-RA-induced AIN with eosinophil predomination in kidney pathology, our patient had infiltration primarily with lymphocytes and plasma cells. Some of the previous cases also have mononuclear cell infiltration along with eosinophils [19–22]. Of interest, Komala presented the case of a patient who did not develop AIN from long-term use of exenatide but had AIN after changing from GLP1-RA to dulaglutide for 4 weeks [23]. These show the complexity in GLP1-RA-induced AIN pathogenesis and the possibility of different pathogenetic pathways for different GLP1-RAs. Therefore, if this hypothesis is proven, it may be possible to rechallenge different GLP1-RAs in indicated patients who had developed AIN from another GLP1-RA. An immunostaining study to characterize lymphocyte subtypes and an inflammatory cytokine profile study, which we did not conduct, could elucidate this question, and further studies are needed. Until there is a clear answer to this question, we do not recommend rechallenging different GLP1-RAs after AIN.

In addition to discontinuation of the causative agent, there has been no consensus on the treatment of AIN associated with GLP-1 RA, and there is concern that systemic steroid therapy can aggravate hyperglycemia in diabetic patients. However, in previous reports, the prescription of prednisolone (50 mg/day or 1 mg/kg/day) has been associated with partial-to-complete recovery of kidney function, and all cases without renal function recovery did not receive steroid therapy (Table 1). However, our patient is the first case to achieve recovery in renal function without steroid therapy, which was not administered due to the risk of immunosuppression, as she had frequent urinary tract infections. This may, at least in part, be due to the short interval from diagnosis to drug discontinuation and the non-extensive interstitial fibrosis and tubular atrophy in the pathology report. In all reported cases without recovery from GLP1-RA-induced AIN, there were more than 2 months from the start of the drug to the discontinuation, and chronicity features from renal pathology (severe interstitial fibrosis and tubular atrophy). Initial GFR and age might not be significant risk factors, since an 83-year-old patient with estimated GFR of 32 mL/min/1.73 m2 still had partial recovery of renal function [20]. Therefore, we recommend consideration of prompt steroid therapy on a case-by-case basis, considering the severity of AKI, contraindications to immunosuppression, cardiometabolic control, and the pathology results of chronicity in concordance with the AIN from other classes of medication [12]. Although we report a rare adverse effect of GLP-1 RAs, this does not impede the prescription of this drug class to selected diabetic kidney disease patients due to the overwhelming cardio-metabolic-renal benefit for most of these patients. Rather, we emphasize the need for vigilant follow-up within 1 month after each drug initiation or titration, prompt kidney biopsy, and appropriate treatment. Finally, whether this complication occurs, and to what extent, from the prescription of GLP1-RA-based therapy for other indications in patients who might not have diabetes and impaired kidney function, such as obesity and metabolic dysfunction-associated fatty liver disease, necessitates further studies, as AIN has not been reported in this population to date [24,25].

Conclusions

GLP-1 RA-associated AIN has occasionally been reported to complicate the treatment of diabetic kidney disease, with various and non-specific manifestations. The pathophysiology and standard treatment of this condition remain inconclusive due to limited evidence. However, GLP-1 RA administration is still recommended in selected patients according to clinical guidelines, despite this rare adverse effect, but vigilant monitoring, early kidney biopsy within 1 to 2 months of the onset of kidney injury, and timely drug discontinuation with judicious consideration of steroid therapy are suggested. Despite the lack of cross-reaction demonstrated in a single study, rechallenging with a different GLP1-RA requires further study.