Renal Sarcoidosis: One Disease, Different Kidney Involvements

Abstract

Renal sarcoidosis has a low frequency, from 0.1% to 0.2%, considering American monocentric cohorts of about 10,000 native kidney biopsies performed in 10-year period. Acute kidney injury (AKI), occurring in <1% of patients, brings sarcoidosis to nephrologist’s attention. AKI in sarcoidosis is mainly due to hypercalcemia and sarcoid granulomatous interstitial nephritis (sGIN), the hallmark pathological finding of the disease. AKI related to hypercalcemia generally responds to steroids. At the contrary, not always all sGIN-AKI has a benign prognosis. This chapter will describe the widest casistics of renal sarcoidosis, considering the predictive value of clinical features, laboratory, radiological parameters, and histological patterns regarding induction therapy response to AKI. Rarely sarcoidosis is life-threatening: fatal events could occur during AKI or during the progression from chronic kidney disease (CKD) to end-stage renal disease (ESRD), a high-risk condition for cardiovascular, infectious, and oncological events. AKI to CKD transition due to specific injury of renal sarcoidosis is one of the most interesting aspects for nephrologists, as the reason why only a minority of sGIN cases will develop AKI: generally, sGIN is s a silent finding observed at autopsy in 7–23% of sarcoidosis patients.

1. Introduction

Sarcoidosis (S) is an idiopathic multisystemic granulomatous disorder, histologically characterized by epithelioid noncaseating granulomas which involve primarily lungs, mediastinum, lymph nodes, liver, eyes, and skin. All organs could be affected by the disease that rarely strikes kidneys, central nervous system, and heart [1]. In susceptible hosts, an abnormal T-helper 1 response to an unknown, not degradable antigen stimulates macrophages and dendritic cells to transform into epithelioid cells and to fuse together into multinucleate giant cells. In 1933 for the first time, Garland and Thomson described noncaseating sarcoid granulomas in the kidney [2]. The nodular inflammatory infiltrate in the renal interstitium characterizing sarcoid granulomatous interstitial nephritis (sGIN) includes one or more distinct aggregates of epithelioid cells as large as a glomerulus with or without multinucleate giant cells and lymphocytes in outer layers [3].

2. Renal sarcoidosis definition

sGIN is the classic pathological lesion of renal S, but there are many other forms of renal involvement. We can distinguish sarcoidosis-associated renal involvements (SARI) from renal sarcoidosis (RS): SARI is any renal manifestation of S, even functional as hypercalciuria and hypercalcemia. The term RS entails compatible renal histopathology. The spectrum of SARI can be divided into two groups, according to the presence of renal failure (Table 1). The same SARI may occur with or without renal failure, as observed in hypercalcemia and sGIN.

Without renal failure	With renal failure
Hypercalciuria Hypercalcemia Nephrolithiasis Granulomatous interstitial nephritis Tubular dysfunction	Hypercalcemia Calcium-phosphate deposition and nephrocalcinosis Tubulous Interstitial nephritis without granulomas (TIN) Granulomatous interstitial nephritis (rarely) Glomerular diseases with immune complex deposition Renovascular disease Obstructive uropathy

Table 1.

S-associated renal involvements (SARI) and renal failure.

3. Disturbances of calcium and vitamin D

3.1 Frequency of hypercalciuria

Hypercalciuria (defined as urinary calcium >300 mg/day) is the more common SARI. It results from both an increased glomerular calcium filtration and a decreased calcium tubular reabsorption due to a PTH suppression caused by calcitriol (1,25-dihydroxycholecalciferol) excess. Even 1% decrease of calcium reabsorption can provoke hypercalciuria [4]. Hypercalciuria affects from 20% up to 40% of systemic S [5, 6], and up to 50% of patients will develop hypercalciuria at some time during the course of the disease [6]. JA Lancina Martín et coll observed in a retrospective Spanish cohort (96 consequent S patients from 1978 to 1993) hypercalciuria in 26.6% of patients with S and urinary lithiasis or lithogenic risk factors (hypercalcemia and hypercalciuria) [5]. PR Studdy et coll in a UK series, found hypercalciuria in 40% (77 of 192), taking an upper limit of urinary calcium excretion as 300 mg/24 h (7.5 mmol/24 h) [6]. In some past work, hypercalciuria, defined with a lower limit of urinary calcium >200 mg/24 h, was found in 62% of patients [4]. Hypercalciuria often remains asymptomatic, and it can be remains undiagnosed without hypercalcemia or nephrolithiasis. This could be dangerous because hypercalciuria is a risk factor for nephrolithiasis.

3.2 Frequency of hypercalcemia

In general population, hypercalcemia occurs in 0.2–4%, of which primary hyperparathyroidism and malignancies represent about 80–90% of all cases [7]. Mild to severe hypercalcemia affects from 4% up to 18% of S patients [6, 8]. In the multicenter prospective study ACCESS (a case–control etiologic study of S) [8], R. Baughman and colleagues observed a low frequency of S-associated hypercalcemia 3.7% (27 of 736), perhaps for a 44% of representation of African-Americans where S-associated hypercalcemia is less common. In the ACCESS study, a significant risk factor for S-associated hypercalcemia found (odds ratio 3.6, over 470 patients analyzed) was the combination of HLA DRBI1*1101 allele and exposure to insecticides [9]. Hypercalcemia frequency was 6.3% (6 of 96) in J.A. Lancina Martín’s retrospective Spanish cohort [5], whereas P.R. Studdy and colleagues in their UK series found hypercalcemia in 18% (99 of 547, taking an upper limit of normal of 10.5 mg/dl, 2.6 mmol/l) [6]. Referring to the biggest last studies with more than 1000 S patients, hypercalcemia frequency varies from 5 to 11%: In J. Werner’s study on Sweden cohort of 1229 S patients (data between 1987 and 2018 collected in local registry at Karolinska University Hospital, Stockholm), 5.4% (66/1229) had hypercalcemia (defined as serum ionized calcium >1.33 mmol/L, reference n.v. 1.15–1.33 mmol/L) [10]. In a single-center retrospective study by R. Baughman and colleagues (n = 1606) reported that hypercalcemia appeared in about 6% of S patients [11]. In this study, a group of about 100 patients with hypercalcemia was compared to 1500 patients with S without calcium metabolism disturbances: there were not any differences in sex, age or ethnicity. In a world survey, James and colleagues noted hypercalcemia in 11% (200/1760) of S patients [12].

3.3 Hypercalcemia in renal sarcoidosis

Hypercalcemia in renal S had a very high prevalence, between 20 and 30%. In A.R. Berliner’s meta-analysis [13], the widest sGIN-AKI retrospective study, 19,5% (18/94) of patients were hypercalcemic, 7.7% (4/52) had hypercalciuria, calcitriol was elevated only in one case (1/10), that presented concomitant hypercalcemia. Three patients presented hypercalcemia and inappropriately normal calcitriol. M. Mahévas and colleagues observed a 32% (15/47) hypercalcemia frequency in their retrospective study of the French sarcoidosis Group [14]: hypercalcemia was significantly more common in white patients (44%, 14/32) than in blacks (6.5%, 1/14) (Odd ratio (OR) 8; 95% confidence interval [IC], 1.75–85, p < 0.001) and was more common in men than in women (12 vs. 3). There was a seasonal variation of hypercalcemia that is aggravated by sunlight: 50% (11/22) patients diagnosed between June and September had hypercalcemia, compared with only 16% (4/25) patients diagnosed during the other months (OR 7.8; 95% IC, 1.6–44.8, p < 0.001). All patients with hypercalcemia presented hypercalciuria, which was complicated by nephrolithiasis in three cases and nephrocalcinosis in one case.

3.4 Frequency of nephrolithiasis

Nephrolithiasis is a complication that happens in 10–14% of S patients over the course of disease, more frequently than in the general population. In A.R. Berliner’s meta-analysis 2.7% of S patients have asymptomatic stones at the diagnosis, whereas in 1% of S patients, renal-ureteral colic is the first symptom of disease [13].

Urolithiasis frequency was 14.5% (14/96) in J.A. Lancina Martín’s retrospective Spanish cohort with S patients with lithogenic risk factors (hypercalcemia and hypercalciuria) [5].

3.5 Hypercalcemia in sarcoidosis and the role of extrarenal calcitriol

In 1939 for the first time, Harrell and Fisher found an association between dysregulated calcium homeostasis and S granulomas [15]. In the late 1970s, there were observed increased calcitriol concentrations in a few cases of hypercalcemic S patients [16]. In 1981 Barbour and colleagues proved the extrarenal production of activated vitamin D through the observation of hypercalcemia and elevated calcitriol levels in an anephric S patient [17]. In 1982 hypercalcemia and elevated calcitriol levels were observed in a chronic hemodialysis patient with S [18]. Physiologically renal proximal tubular cells transform 25-hydroxycholecalciferol (calcidiol) into 1,25-dihydroxycholecalciferol (calcitriol) through 1-α hydroxylase enzyme. This is under PTH control, which is suppressed by hypercalcemia [19]. Calcitriol is an active vitamin D that increases calcemia, both increasing calcium absorption from bowel and releasing calcium from the bone through the induction of osteoclast cells. It was proved that calcitriol was a metabolite causing S-associated hypercalcemia, and its extrarenal source was found both in activated pulmonary macrophage [20, 21] and in lymph node homogenates [22]. Subsequent works revealed that calcitriol synthesis macrophages are permitted thanks to expression of a single gene, the CYP27B1, and its product 1α-hydroxylase [23]. This extrarenal ectopic production of calcitriol is resistant to negative feedback of hypercalcemia because it is PTH-independent [24]. Many studies confirmed that a high proportion of hypercalcemic S patients have elevated calcitriol [25, 26, 27]. The typical pattern of hypercalcemia in S is associated with high plasmatic calcium levels, low intact PTH, and high calcitriol, as observed in the case of AKI due to sGIN presented by Berliner (calcemia 12.9 mg/dl, calciuria 232 mg/24 h, PTH 5.0 pg./ml, calcidiol 38 ng/ml, and calcitriol 120 pg./ml) [13]. In Mahévas’s study, 7 of 10 hypercalcemic patients with PTH tested had PTH < 10 ng/l, lower than normal range (n.v. 10–65 ng/l); 7 of 9 hypercalcemic patients with calcidiol tested had calcidiol <10 nmol/L, lower than normal range (n.v. 10–40 nmol/L) [14]. Sometimes calcitriol is in normal range in S patients with hypercalcemia or hypercalciuria, although it is inappropriately normal, and it should be lower in case of hypercalcemia [28]. PTH-related protein, expressed in S granulomas, is another mediator of S-associated hypercalcemia. High series PTH-related protein level was observed in hypercalcemic S patients [29].

4. Diagnosis of renal sarcoidosis

4.1 Frequency of renal sarcoidosis

According to ACCESS Research Group, kidney involvement in S is “unusual but clinically important” [30]. There are various data about frequency of RS, because it depends on what definition is adopted. Considering patients with acute or chronic S, renal involvement is not frequent, as it emerges from geo-epidemiological big data approach to S: in 2019 a review collected large series (>100 patients) of S reported in the PubMed library in the last 20 years, renal involvement frequency was observed in 0.3% patients in Northern Europeans (6/2209 patients with extrathoracic S (ET-S), total patients 71,566), 1.8% patients in Southern European cohorts (26/1477 ET-S patients, total patients 5902), 1.1% patients in US cohorts (28/2436 ET-S patients, total patients 7263), 3.5% patients in Japanese cohorts (36/1027 ET-S patients, total patients 3315) [31].

If RS was interpreted as kidney involvement in S with renal failure, frequency is below 2% of cases: 0.7% (5 of 736) patients of ACCESS study, RS defined as “treatment responsive renal failure” or “steroid-responsive renal failure in a patient with diabetes and/or hypertension” [30]; 1.2% (10 of 818) S patients series of Royal Northern Hospital of London [32].

R. Bergner and colleagues in their German cohort of 327 S patients precisely distinguish between a probable RS and a definite RS [33]:

A patient has a probable RS if two inclusion criteria are satisfied:

• The diagnosis of S according to international guidelines as the American Thoracic Society (ATS) recommendations [34]

• At least one of three renal abnormalities: estimated glomerular filtration rate (eGFR) < 60 ml/min without an alternative explanation, proteinuria >300 mg/g creatinine or > 300 mg/24 h, or abnormal urinary sediment.

A patient has a definite RS if he has a probable RS with a compatible renal histopathology. According to these criteria, frequency of probable and definitive RS is high, respectively, 33.3% (109/327) and 27.5% (90/108) in German cohort. This observation clearly points out how RS is underestimated because asymptomatic in most of cases.

4.2 Pathological urinary sediment and renal sarcoidosis suspect

In A.R. Berliner’s meta-analysis at exordium 30% of RS patients had sterile pyuria, 20% hypercalcemia and microhematuria, 10% hypercalciuria and glycosuria [13]. There are not unanimous observation about urinary sediments: In R. Bergner’s German cohort 52% renal S patients had pathological sediment (53 of 102 with available sediment, total of 109 patients) in comparison with 29% (51 of 175) S patients without renal involvement (175 patients with available sediment, total 197 patients, p < 0.001), difference remained significant observing leukocyturia, more present in renal S patients than in S patients without renal involvement (34%, 37/109 vs. 25%, 43/175, p < 0.001) [33]. For R. Bergner and colleagues, pathological urinary sediment was an important tool to identify probable RS.

4.3 Renal biopsy, a fundamental diagnostic tool

Renal biopsy (RBx) can not only identify a further localization of a disease already diagnosed in other organs by biopsy specimens but often manages to give the diagnosis of S, both with systemic involvement since then unknown or with exclusive renal involvement. RS mostly occurs at the presentation of S: in 81% (38/47) in Mahévas’s French series [14] and 77% (30/39) in F. Rastelli’s Italian series [35]. In Italian series, only four patients (10%, 4/39) had renal S during a systemic disease lasting for more than 1 year, and five patients (13%, 5/39) had exordium symptoms less than 12 months before from RBx. These data were confirmed by Tao Zhao’s Beijing cohort, and kidney involvement was the first symptom of S in 83.3% of cases [36]. S diagnosis was possible thanks to renal biopsy in 47% (8/17 patients) in Rajakariar’s monocentric study in London [37] and 49% (23/47 patients) in Mahévas’s multicentric French study [14]: these patients had an unexplained renal impairment and were subsequently diagnosed with RS involvement and systemic disease. Rastelli and colleagues noticed that RBx permitted the diagnosis of unknown S in 74% (29/39) of Italian patients, whereas in 23% (9/39) of cases presented isolated kidney S. Unexpectedly, 44% (8/18) patients with S lung at diagnosis were without respiratory symptoms, that is 21% (8/39) of patients in Italian series. Not always lung involvement gives rise to dry cough, a frequent symptom in S patients [35]. R. Bergner’s and colleagues demonstrated in their German cohort that RBx is a precious diagnostic tool when performed in patients with high suspected renal S: 327 S patients were collected, 109 with probable or definite renal S were identified, of which 72 patients (66%) had eGFR <60mml/min, 57 patients (52%) had proteinuria >300 mg/24 h, 53 patients (49%) had pathological sediment (leukocyturia), in 108 patients RBx was performed, and 90 had confirmed renal S because they had compatible histopathology (sGIN, TIN, nephrolithiasis, and nephrocalcinosis). So RS was confirmed in 83% of cases (90 of 108 RBx performed) [33].

4.4 Phenotypes of renal sarcoidosis patients

In German cohort [33], patients with renal S had more extrapulmonary lymph nodes than other S patients [40% (44/109) vs. 26% (52/197), p < 0.05] and more liver involvement [25% (27/109) vs. 16% (31/197), p = 0.07].

Clinical phenotype of Mahévas’s patients did not differ significantly from patients without renal involvement according to the prevalence of thoracic and extrathoracic localizations [14], similar to the ones in ACCESS study [30].

4.5 Timing for renal biopsy

In the current clinical practice, a patient with systemic S can reach the nephrologist’s attention in these two main situations: an impaired renal function with no evidence of decreased kidney perfusion or obstruction and significant proteinuria >1 g/24 h, 2 conditions where there is indication of renal biopsy. Prevalence of renal S is underestimated because only in centers with great expertise in S management RBx is performed in all patients with a suspected renal S: not only in case of decreased renal function or significant proteinuria but even pathological urinary sediment, as R. Bergner and colleagues carried out in their tertiary care centers of German cities of Darmstadt, Géttingen, Ludwigshafen, Offenbach, Trier [33]. To our knowledge, this German multicentric casistic is the widest study of RS for RBx performed.

An impaired renal function can occur with these three clinical conditions: (1) acute kidney injury (AKI): rapidly elevated serum creatinine or an increase, (2) acute kidney injury superimposed on chronic kidney diseases (CKD), and (3) progression of chronic kidney diseases without other leading nephropathogenic causes.

AKI frequency due to RS is low: 0.76% according to Shas’s monocenter study considering all 2780 native kidney biopsies performed in The Ohio State University in 7 years, from January 1, 2003 to December 31, 2009: only 21 patients presented a biopsy-proven S with AKI [38]. On the contrary, considering cohort of S patients and not the general population, AKI is not a rare S presentation. In Tao Zhao’s cohort in Beijing, kidney involvement was the first symptom of S in 83.3% of cases, with half of them presenting with acute kidney injury [36]. In German cohort two-thirds of patients with renal S histologically confirmed had an eGFR<60 ml/min [33]. R. Bergner and colleagues [33] demonstrated that in patients with systemic S a reduced eGFR below 60 is mainly caused by renal manifestation of disease than other causes: among 109 patients with suspected renal S, renal S was confirmed in 83% of cases (90 cases of renal S in 108 RBx performed, with interstitial nephritis with or without granulomas in about two-thirds of the patients), wherein only 17% (18/108) there were renal findings not compatible with S: 10.2% (11/108) nephroangiosclerosis, 2.8% (3/108) diabetic nephropathy, 0.9% (1/108) tubular damage, and 2.8% (3/108) unremarkable.

4.6 Severe AKI

In Tao Zhao’s cohort in Beijing, 11/18 patients (61.1%) suffered from severe renal function impairment (eGFR <30 mL/min/1.73) at the time of the renal biopsy [36], similar to 68.5% reported in Mahévas’s study [14]. The most frequent causes of AKI in RS are hypercalcemia, sGIN, TIN or an association between hypercalcemia with interstitial nephritis in patients suffering from renal sarcoidosis could appear both as sGIN and TIN without granulomas. If hypercalcemia is only a SARI without a disease presence in renal parenchyma, the AKI associated to hypercalcemia is functional and it can be completely recovered. Hypercalcemia provokes a prerenal AKI through a reversible hemodynamic insult due to afferent arteriolar vasoconstriction, hence a decrease in renal blood flow and glomerular filtration rate. If this condition is prolonged, it can cause polyuria and dehydration for inhibition of sodium-potassium ATPase in tubular cell and urinary sodium wasting or tubular necrosis [39].

4.7 Necessity of acute hemodialysis at renal onset

Considering Berliner’s meta-analysis, 7,4% of patients (7/94) underwent hemodialysis as acute renal replacement therapy initially or shortly after onset despite steroids.

5. Granulomatous interstitial nephritis in sarcoidosis and other diseases

5.1 sGIN frequency

Sarcoid granulomatous interstitial nephritis (sGIN) is the typical renal lesion in S. Not all sGIN mean have a renal S clinically symptomatic: in fact, it is a silent finding observed at autopsy in 7-23% of S patients [40, 41]. sGIN is mainly observed at RBx performed in the case of AKI, occurring in <1% of S patients [42]. This explains its low frequency, which varies from 0.1 to 0.2% considering the widest American monocentric cohorts of about 10,000 native kidney biopsies performed in 10 year-period up to 1–5% considering RBx in patients referring to tertiary S centers: 0.1% (11/9779) at Harvard Medical School [43], 0.18% (19/10,023) at Johns Hopkins University [44], 1.1% (2/187) in American S cohort at a tertiary S referral center in Minnesota (consecutive patients diagnosed between March 2015 and May 2019), 5% (11/218) Spanish S cohort at tertiary teaching Hospital Clinic of Barcelona (consecutive patients diagnosed between 1990 and 2015) [45]. The epidemiological Japanese study confirmed the low frequency of 0.11% (16/14,191) [46]. sGIN frequency is 0.6% of renal transplant biopsies [47].

5.2 Differential diagnosis of GIN

In relation to N.Joss’s study in Glasgow, all causes of GIN represent 1% of all diagnoses on native renal biopsies [48]. According to V.Bijol’s study, drug represents the most frequent cause of GIN (45%, 17/38), whereas S is the 2° cause (29%, 11/38) [43], that is to say, that only one out of three cases of GIN is an sGIN (Table 2).

GIN due to all causes	Bijol [43]	Javaud [42]	Joss [48]
Drug	45% (17/38)	17.5% (7/40)	11% (2/18)
S	29% (11/38)	50% (20/40)	28% (5/18)
Wegeners granulomatosis	15.9% (6/38)	2 (5%)	0
idiopathic	10.5% (4/38)	12.5% (5/40)	50% (9/18)
TB	0	7.5% (3/40)	0
Leprosy	0	2.5% (1/40)	0
Mycobacterium avium infection	0	1 (2.5%)	0
Crohn	0	1 (2.5%)	0
foreign-body giant cell reaction	15.9% (6/38)	0	0
bacillus Calmette-Guérin therapy for bladder cancer	15.9% (6/38)	0	0
xanthogranulomatous pyelonephritis	15.9% (6/38)	0	0
TINU syndrome	0	0	11% (2/18)

Table 2.

GIN due to all causes.

In Tao Zhao’s cohort in Beijing, sGIN has double frequency (66%, 12/18) respect GIN by other causes (33%, 6/18) [36].

Not caseous non-necrotizing granulomas are typical for RS but not pathognomonic. Many diseases cause GIN, so renal S remain a probability diagnosis. In presence of granuloma physician is more confident of diagnosis of RS, but to have a “definite” diagnosis there are to be excluded all causes of granulomatous inflammation:

• Infectious diseases: brucellosis, fungal infections (histoplasmosis), tuberculosis, and leprosy,

• occupational diseases: chronic beryllium exposure/berylliosis,

• immune-related disease: Crohn’s disease, primary Stevens–Johnson syndrome, IgG4-related TlN,

• autoimmune diseases: vasculitis as granulomatosis with polyangiitis, autoimmune-induced TIN,

• reaction due to neoplasm: in particular breast cancer, lung cancer, Hodgkin lymphoma,

• granulomatous foreign-body reaction: heroin, cholesterol atheroembolism, and

• drug-induced interstitial nephritis: nonsteroidal anti-inflammatory drugs, allopurinol, fluoroquinolone antibiotics, diuretics, HIV-antiretroviral, and alfa-interferon.

Other two diseases provoking GIN which are considered as two distinct entities are tubulointerstitial nephritis and uveitis (TINU) syndrome and necrotizing sarcoid granulomatosis (NSG).

5.2.1 Drug-induced interstitial nephritis

Drug-induced interstitial nephritis is the most frequent cause of GIN [49] and it gives rise to interstitial granulomas in 25–50% of cases [50].

The discovery of interstitial granulomas in patients with recent drug intake orients diagnosis toward drug-induced GIN, especially when eosinophils are not present in inflammatory infiltrate [50, 51]. To rule out classical drug-induced interstitial nephritis is important not only an accurate pharmacological anamnesis but also to identify the area interested in granulomatous interstitial reaction: it is the cortical-medullary junction for classical drug-induced interstitial nephritis whereas only renal cortex in sGIN [13, 40].

5.2.2 Tuberculosis

Distinguishing between tuberculosis GIN and sGIN is not always easy in absence of caseous necrosis, typical of tuberculosis and not present in S. Epithelioid not-necrotizing granulomas can be present also in renal tuberculosis. So it is mandatory to recognize Koch bacillum at Ziehl-Neelsen staining and to undertake TB cultures. Polymerase chain reaction (PCR) test of the biopsy for mycobacterial DNA manages to detect Koch bacillum in 67% of cases [52].

Unfortunately, a negative PCR in renal tissue result cannot exclude a TB infection because kidney involvement in TB is not always direct but also indirect, via immune-mediated pathways [53]. E. Danila and E. Zurauskas, demonstrated a significant overlap in types of granulomatous inflammation between tuberculosis and S in the bronchoscopic lung or bronchial biopsies of 105 patients: of patients with tuberculosis, 76% had epithelioid cell granulomas with caseosus necrosis, and 24% had not-necrotizing epithelioid cell granulomas [54].

In Table 3 there are accuracy characteristics of not-necrotizing granulomas and caseous granulomas, respectively, in diagnosis of S and tuberculosis.

	Sensitivity	Specificity	Positive predictive value	Negative predictive value
Not-necrotizing epithelioid cell granulomas in S diagnosis	94%	60%	68%	92%
epithelioid cell granulomas with caseosus necrosis in tuberculosis diagnosis	76%	85%	69%	88%

Table 3.

Characteristics of not-necrotizing granulomas and caseous granulomas respectively in diagnosis of S and tuberculosis.

A wrong diagnosis is also a risk in nephrological field: B. Oliveira and colleagues reported four patients who were unnecessarily treated with steroids for an initial diagnosis of renal-limited S (40 mg median dose steroids) and received the right diagnosis of T tuberculosis B infection with a mean delay of 22 months (range: 1–60), one patient before the renal biopsy and three after. All patients at exordium had no evidence of lung tuberculosis, or systemic disease associated with tuberculosis infection, Ziehl Neelson (ZN) stains on the kidney biopsy were negative. Three of these patients extrarenal TB manifestations [55].

Mycobacteria and fungi are the main infective causes of GIN, above all in immunosuppressed patients, such as patients who received renal transplants [47, 56].

5.2.3 Leprosy

Leprosy, provoked by infection with either Mycobacterium leprae or Mycobacterium lepromatosis, has a worldwide incidence of 200,000 new cases each year [57]. Aiming to consider leprosy in differential diagnosis of GIN, it is fundamental to examine the prevalence of diseases in the area of interest, since 80% of new cases are reported in Brazil, India, and Indonesia [57]. The most frequent kidney involvement in leprosy is amyloidosis [42], whereas GIN is observed only in 1% of leprosy [58].

5.2.4 Fungal infections

Chronic fungal infections have to be ruled out both clinically and through fungal stains on suspected granulomas, such as Grocott-Gomori’s methenamine silver stains. In addition, serologic tests should be done in patients living in or coming from endemic areas for Histoplasma capsulatum and Coccidioides immitis.

In immunodepressed patients with chronic fungal infections, granulomas could present a “dirty” necrosis, characterized by cellular debris and neutrophilic infiltrate, mostly observed in liver during autopsy.

5.2.5 Granulomatosis with polyangiitis

Granulomatosis with polyangiitis (GPA) was also known as Wegener granulomatosis. This vasculitis gives GIN in 5% of cases [59]: the main feature leading to diagnosis is the pauci-immune crescentic glomerulonephritis associated. GPA and S could share some common mechanisms in pathogenesis: some patients presenting with GPA kidney involvement at diagnosis of vasculitis developed biopsy-confirmed lung S months later through cyclophosphamide therapy [60].

Necrotizing sarcoid granulomatosis.

NSG is a condition first described in 1973 by Averill Liebow in the James Burns Amberson lecture to American Thoracic Society. This condition has three main characteristics [61]:

• Histologically: presence of Sarcoid-like granulomas with vasculitis and necrosis.

• Radiologically: multiple lung nodules without hilar adenopathy.

• Clinically: benign course.

To recognize NSG from an S is fundamental to find out necrotizing aspects of granulomatous flogosis, either large foci of necrosis or small central areas of bland necrosis within granulomas. Besides, granulomatous flogosis has focal vasculitis features, with granulomas intruding on the walls of small arteries. It still remains controversial whether NSG is a variant of S or a distinct entity. Liebow suggested that NSG is a variant of angiocentric granulomatosis; according to this hypothesis, granulomas are sarcoid reactions to necrotizing angiitis. The other hypothesis is that NSG is a manifestation of systemic S with necrotic granulomas and may merge with the entity of nodular S [62].

5.2.6 Tubulointerstitial Nephritis and Uveitis syndrome

TINU syndrome is also called Dobrin syndrome because Dobrin described it for the first time in 1975. TINU syndrome is considered a distinct entity from S because of the absence of hypercalcemia, hypercalciuria and bihilar lymphadenopathy [63]. These patients should be evaluated for Sjőrgren’s syndrome [64].

5.2.7 Interstitial nephritis in sarcoidosis patients: not always granulomatous

Considering monocentric cohorts of TIN (all causes), in single-center study In T. Zhao’s cohort in Beijing, TIN due to renal S was 57% (30/53): sGIN was 23% (12/53) and 34% (18/53) TIN without granulomas considering all causes-GIN, sGIN was 40% (12/30) and 60% (18/30) TIN without granulomas considering RS patients [36].

Considering monocentric RS cohorts, in sGIN frequency was 33% in Rajakariar’s series (13 sGIN and 4 TIN without granulomas out of 39 RS cases) [37].

Considering multicentric RS cohorts, in sGIN frequency was 79% (37/47) in Mahévas’s series (10 TIN without granulomas and 37 sGIN, of whom 59%, 22/37, with giant cells, out of 47 RS cases) [14], 79% (31/39) in Rastelli’s series (31 sGIN and 5 TIN without granulomas, out of 39 RS cases) [35].

6. Demografic and clinic characteristics of sGIN patients

6.1 Sex

If S has a slight predominance in women (female: male ratio 1.2–1.5:1) [1]: female predominance was in ACCESS study [30], 61% (500/818) were female of S patients series of Royal Northern Hospital of London of Royal Northern Hospital of London [32]. RS has a marked predominance in men: in Bergner’s German cohort female were 41% (45/109) [33], in Javaud’s French study 60% (12/20) [42]. sGIN is more frequent in male gender: 64% in Berliner meta-analysis [13]. The only cohort with different results is Rajakariar’s study, where 53% (9/17) of renal S patients were women: probably the high numbers of patients with Afro-Caribbean origin (53%, 9/17) contributed to this different demography [37].

6.2 AGE of onset

RS, as well as sGIN, is more likely to have a later onset than systemic S: in German cohort 58 years (range 18–86) for renal S onset, 51 years (range 19–85) for S onset [33]. In Berliner meta-analysis, mean age at presentation of 94 patients was 46.9 years (range 11–80, for men 46 year, for women 48.5 year) [13]. Mahévas study, the largest retrospective study (47 patients), documented similar data [14]. In Javaud’s study, the mean age at presentation of 20 patients with renal S was 51.8 years (range 29–76) [42].

6.3 AKI due to sGIN

In the Italian series, sGIN-AKI resulted more severe than No-sGIN-AKI [35]. In the Berliner meta-analysis regarding sGIN-AKI onset, renal function at presentation was series creatinine 4.8 mg/dl and urinary proteinuria 1 g/24 h [13]. No one of the large studies observed a correlation between proteinuria level and histological findings [65]; nevertheless, a proteinuria >2.0 g/24 h is associated with a glomerular involvement of S, also in the course of sGIN.

7. Pathological characteristics of renal sarcoidosis at diagnosis

7.1 GIN vs TIN without granulomas pathological features independently of causing disease

In order to characterize pathological lesions of TIN T. Zhao and colleagues created a score for active tubulointerstitial injury (including the degeneration or necrosis of tubular epithelial cells, tubulitis, interstitial edema, inflammatory infiltration, and granuloma formation) and a score for chronic tubulointerstitial injury (including tubular atrophy and interstitial fibrosis). Independently of cause of TIN (patients with or without S), 18 patients with granulomatous TIN had higher scores for acute interstitial injury score than 18 patients with TIN without granulomas [6(4.5–7.0) vs. 4(2.0–6.0) p = 0.001)] but Tao Zhao and colleagues did not observe a higher leukocyturia in granulomatous TIN patients (there was leukocyturia in 28% of patients with granulomatous TIN and in 50% patients with TIN without granulomas, p = 0.103) [36].

7.2 Renal sarcoidosis (sGIN and TIN without granulomas) vs tubulointerstitial nephritis in not S patients

In the Chinese cohort of Beijing, chronic tubulointerstitial lesions and glomerulosclerosis were associated with kidney function impairment (eGFR<30 ml/min) in renal S when RBx was performed: renal S (sGIN and sTIN without granulomas) had more chronic tubulointerstitial lesions than non-S patients (GIN and nongranulomatous TIN) at diagnosis: chronic tubular injury score S patients 2(1.030) vs. not S patients 2(0-20), P = 0.060; chronic interstitial injury score: S patients 1.5 (0-3.0) vs. not S patients 1(0–2.0), p = 0.631. At the contrary, not S patients had total tubulointerstitial injury acute score more higher than S patients: 5(4.0–6.0) vs. 10 (6.0–12.0), P = 0.002, in details acute tubular injury score 2(040) vs. 4(20–6.0), p = 0.008, acute interstitial injury score, 5 (3.0–60) vs. 4 (25–60) p = 0.457 [36].

8. Therapy

8.1 Response to steroid therapy

Most patients were treated with prednisone 1 mg/kg as induction therapy. Steroid tapering is started after 12 months since exordium, and average period of maintenance therapy should be 24 months.

All 5 cases described by Joss improved their renal function with an adequate maintenance therapy (Mean follow-up time: 47.2 months (range from 33 to 86)), all patients [48]. In Rajakariar study, all patients showed beneficial responses to prednisolone within the 1° year. In Rajakariar study, there was no difference in response to treatment between black and nonblack patients [37].

8.2 Other studies show no benign prognosis for sGIN-AKI

In the Berliner meta-analysis, most patients improved their renal function but did not regain normal creatinine and were at varying degrees of CKD [13]. In their retrospective study collected 40 cases of GIN due to all causes, Javaud and colleagues observed that sGIN had one of the worse renal function recoveries through steroid therapy other GIN forms because only 65% sGIN patients (13/20) increased renal function after steroid therapy, in comparison with 86% (6/7) drug-induced patients and 100% (5/5) idiopathic GIN. All four patients with Mycobacterium tuberculosis and avium-related GIN and the patient with M. leprae-related GIN recovered their renal function after specific therapy. In the Javaud study, two patients had worse renal function (one patient received a kidney graft after a transient period on hemodialysis), and another with chronic insufficiency died after 3 months of follow-up [42]. In Mahévas study, a complete response to steroids at 1 year was found to be strongly correlated with the complete response at 1 month ([OR] 7; 95% [CI], 1.6-44.8, p < 0.001). Furthermore, a complete response to steroids at the end of FU was found to be strongly related to a complete response at 1 month ([OR] 7.6; 95% [CI], 2-41, p < 0.001). In Mahévas cases, patients who received high intravenous doses of methylprednisolone (MP) before oral prednisone seem to have a better response to therapy: patients receiving MP had complete response in 80% (8/10), and at the end of follow-up 50%(5/10) had eGFR>60 ml/min, compared to 30%(10/36) patients only receiving oral prednisone [14].

Rajakariar [37] hypercalcemia was strongly correlated with the complete response at 1 year ([OR] 16; 95% [CI], 1.8-137, p = 0.003). In a multivariate analysis, hypercalcemia was independently correlated with complete response (OR = 18.9, p = 0.001).

8.3 Better response for No-sGIN-AKI

In Rajakariar’s study, 75% (3/4) of patients with tubulointerstitial nephritis without granulomas, normal calcemia and extrarenal S had steroid-responsive course although advanced scarring at RBx [37]. Rastelli and colleagues confirmed these observations: all 5 No-sGIN-AKI patients had a complete response, independently of histological findings [35].

8.4 Induction with steroid-sparing agent

In Chinese cohort, the five patients who received induction therapy with steroids and steroid-sparing immunosuppressive agents had better long-term kidney recovery than those treated with steroids alone (changes in eGFR%: +221% vs. + 49%, p = 0.045) [36]. In German cohort, 35% of patients (38/109) received azathioprine, 4.5% (5/109), mycophenolate mofetil 6.4%, (7/109) methotrexate, unfortunately, outcomes are not available [33].

9. Prognosis (renal outcome and overall survival)

9.1 Pathological prognostic and predictive factors

Rajakariar [37] and Mahévas [14] observed, respectively, 24% (4/17) and 21% (10/47) patients with tubulointerstitial nephritis (TIN) without granulomas. TIN without granulomas patients have a similar response to steroid therapy than in sGIN patients. For authors, this is likely to be due to sampling effect (sparse granulomas are missed in biopsy specimens). Rajakariar [37] recognizes a further factor in advanced disease at presentation. T. Zhao and colleagues found had chronic tubular lesions as a negative predictive response factor independently of the presence of granulomas in interstitial nephritis: renal S patients with satisfactory response at 1 month had chronic tubular injury score significantly lower than S patients who unsatisfactory response: 2(0.5–20) vs. 3(2.0–3.0), p = 0.044 (satisfactory response defined as eGFR increase percent-age > 50% at 1 month after immunosuppression start). Interesting trend showed acute lesions (both tubular and interstitial) as positive predictive response factors.

Renal S patients with satisfactory response at 1 month had both acute tubular injury score and acute interstitial injury score higher than S patients who had unsatisfactory response, even if without statistical significance: respectively 4(2.5–5.5) vs. 1(0–2.0), p = 0.069 and 3(20–40) vs. 3(2.0–4.5), p = 0.930 [36]. Giant cells were present in 59% (22/37) cases of the sGIN Mahévas study [14] and in all five patients of Joss study [48]. There was no relationship between eGFR at 1 month, 1 year and the end of FU and giant cells presence in Mahévas study [14]. Mahévas [14] and colleagues did not find a correlation between the score of granulomas and calcemia [14]. There was no evidence in the literature that granulomas with distinct lymphocyte cuffs around had different prognoses than “naked” granulomas. Five patients in Rajakariar’s study [37] had evidence of intracellular calcification. The presence of intracellular calcification was not associated with differences in either presenting renal function or response to corticosteroids. In Mahévas’s [14] study, there was interstitial calcification in 8 patients, all with hypercalcemia. There was no relationship between eGFR at 1 month, 1 year and the end of FU and interstitial calcification. In Rajakariar’s study [37], there was no correlation between presenting creatinine and the degree of tubular atrophy (Spearman’s r = 0.045, P=NS). All patients had tubular fibrosis at biopsy. Final Mean eGFR at 1° year was independent of tubulointerstitial fibrosis degree. Both in Rajakariar [37] and in Mahévas [14] study, interstitial fibrosis was observed in all cases. In Mahévas study, there was an inverse relationship between response to steroids and the initial degree of interstitial fibrosis: patients with low scores of fibrosis were the ones who had the best improvement of renal function, and patients with high scores of fibrosis did not respond to therapy. So interstitial fibrosis is the only pathological finding that correlates with renal outcome. It is a paradox that renal fibrosis is observed in all patients, even in one diagnosis of kidney S is got at disease presentation, for Mahévas renal granulomas should promote fibrosis development very rapidly [14]. In their wider casistic, T. Zhao and colleagues confirmed Mahévas’s hypothesis: independently of the cause of GIN (patients with or without S), patients with granulomatous TIN had lower scores for acute tubular injury (p = 0.024) and higher scores for glomerulosclerosis (p = 0.014) and acute interstitial inflammation (p = 0.001) than patients with TIN without granulomas [36].

9.2 Development of CKD

In Berliner meta-analysis, repeated kidney biopsies in patients with bad renal outcomes after steroid therapy showed increased interstitial fibrosis as CKD sign, and some patients progressed toward end-stage kidney disease [13].

Tao Zhao and colleagues noticed that two patients who underwent repeat renal biopsy after 3 months of immunosuppressive treatment showed significant chronic changes coupled with a decreased number of interstitial granulomas, suggesting that immunosuppressive therapies might affect macrophage polarization and thus alter the disease course. In one patient in this study, 3-month immunosuppressive therapy led to the disappearance of renal interstitial M1 macrophages in the kidney biopsy, combined with a good treatment response, thus indicating the potential value of M1 macrophage evaluation in disease monitoring [36].

9.3 ESRD

Since the ‘80 of last centuries, there have been reports of sGIN patients who had chronic renal failure [66, 67]. Rajakariar’s study, one patient was lost to follow-up after 11 years and stopped therapy, subsequently presenting ESRD [37]. One patient in Javaud series reached ESRD and received a kidney graft after a transient period on hemodialysis [42]. In Mahévas’s series, 4% of patients who developed end-stage renal disease with S-related disease required initiation of renal replacement therapy (RRT) within 6 months of diagnosis [14]. T. Zhao and colleagues noticed that patients with glomerular involvement showed worse kidney outcomes, with two patients progressing to end-stage renal disease and requiring maintenance hemodialysis [36].

9.4 Relapses

In five patients of Joss’s study, relapse occurred in four patients when prednisolone was reduced, and all responded to an increased steroid dose. Three patients of whom were subsequently treated with azathioprine as a steroid-sparing agent. Three of the five patients with S developed extrarenal disease: Granulomas were identified in nasal mucosa in one patient, in lymph node in another patient whereas third patient developed uveitis and restrictive lung defect during follow-up [48]. In Rajakariar’s study, 18% (3 of 17) patients relapsed with renal function worsening after stopping steroids because of their side effects or poor compliance, which was reversed upon restarting steroids. Another two patients had multiple relapses as evidenced by AKI and got a complete response with mycophenolate mofetil and azathioprine, respectively, as steroid-sparing agents [37].

In Mahévas’s study, 12 patients had relapses (9 extrarenal and 3 renal), all treated with prednisone and additional immunosuppressive (low-dose methotrexate, azathioprine, and mycophenolate mofetil). In 3 cases with renal relapse, both mycophenolate mofetil (2 cases) and azathioprine (1 case) permitted both remission and steroid reduction. Unfavorable response to steroid at 1° month of FU was related to relapse during the disease progression up to the end of FU (p = 0.049). The median duration of relapse treatment was 48 months (range 24-76). Mahévas observed the eGFR at the end of FU of patients with renal relapse was not different from that of patients without renal relapses. Rajakariar pointed out that in advanced renal disease, steroid discontinuation appears to lead worsening of renal function, which may reach ESRD [14]. Twenty-eight percent (11/39) of patients in the Italian study had relapses [35]: nine sGIN-AKI patients, one No-sGIN-AKI patient, and one patient with sGIN and nephrotic syndrome at exordium. Thirteen percent of sGIN-AKI patients (4/31) had another AKI post-RBx. For two patients, AKI occurred after steroid stop (respectively 2 years and 3 months). For the other two patients, AKI occurred on steroid maintenance (respectively during 1° and 3° steroid therapy year). Among nine sGIN-AKI patients with relapse, a patient with isolated renal S had a renal recurrence 2 years later RBx.

9.5 Mortality

Pilar Brito-Zerón and colleagues demonstrated in their Barcelona 218 S cohort that 11 patients with renal involvement had a higher risk of acute complications and increased mortality rates than patients with other extrapulmonary involvement: in fact, between 68 patients with Charlson Comorbidity Index >1 there were patients with a high frequency of RS (8 patients, 11.8%), whereas in 150 patients with Charlson Comorbidity Index (CCI) between 0 and 1 RS was only 2%, three patients (p = 0.005). Calcium/vitamin D abnormalities, kidney involvement and death remained significantly associated with a high CCI index after adjusting by age and gender [45]. In an Italian study, three sGIN-AKI patients died during follow-up: one for lung cancer at the age of 85 years, 9 years after RBx; one for pneumonia at the age of 77 years, 6 years after RBx, one for unknown cause at the age of 82 years, 10 years after RBx [35]. In the Javaud study, one patient with renal S and CKD died after 3 months of follow-up [42].

10. Conclusion

The fact that almost 30% of S patients with suspect RS have a real renal involvement makes mandatory a diagnostic renal workout in every patient with a new diagnosis of S. Moreover, RBx should be performed not only in case of renal function impairment but even with urinary abnormalities or proteinuria, since sGIN, in particular, does not have a not benign prognosis. It is fundamental to test whether sGIN-AKI is rapidly responsive to treatment to stop renal fibrosis and avoid AKI transition to CKD. In case of absence of response at 1° line steroid, other immunosuppressors are to be rapidly considered. There are needed RTCs for new drugs. It could be important to know if an early diagnosis, hence a rapid treatment, could affect prognosis.

Acknowledgments

We express great gratitude to ACSI Onlus “Amici contro la Sarcoidosi Italia”, the Italian national society of Sarcoidosis patients, in particular to its President Filippo Martoneand the Secretary Amelia Carlucci. ACSI Onlus entirely financed this chapter publication.

Conflict of interest

The authors declare no conflict of interest.

Notes/thanks/other declarations

A particular thanksgiving to Claudia Giuliani.