Nephrogenic Systemic Fibrosis: A Mysterious Disease in Patients with Renal Failure—Role of Gadolinium-Based Contrast Media in Causation and the Beneficial Effect of Intravenous Sodium Thiosulfate

Patient

A 26-yr-old white woman who had end-stage renal disease (secondary to Henoch-Schönlein purpura) and was on hemodialysis (Table 1) presented with a history of severe pain (sharp, aching, throbbing) in her lower extremities for approximately 1 year with tightness, itching, joint stiffness, skin discoloration, and tenderness that began in January to February 2004. On examination, the patient was found to have yellowish scleral plaques bilaterally and lower extremities with hairless, shiny skin with bluish-brown discoloration and woody induration (Figure 1). Clubbing in both lower and upper extremities and sclerodactyly of the upper extremities (Figure 2) were also noted. These findings were reported to be insidious in onset by the patient.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Shiny, discolored, woody skin (on inspection and palpation) seen in the patient with nephrogenic systemic fibrosis (NSF). This picture was taken approximately 1 to 1.5 mo after the initiation of sodium thiosulfate (STS) therapy.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Sclerodactyly and severe clubbing seen in the patient with NSF.

During the course of the next several months, further diagnostic work-up was done (Tables 1 and 2). The patient was treated with large dosages of narcotics and gabapentin with minimal improvement of pain.

The patient was noted to have diffuse large arterial calcification on x-rays, and in view of this and the unresolved pain, a trial of intravenous STS was started (12.5 g three times a week at the end of hemodialysis) in May 2006. The patient tolerated the treatment well, with significant improvement in skin discoloration, pain, and joint stiffness within one month. The improvement in pain was remarkable to the extent that the patient’s narcotic analgesic requirements became minimal and the skin discoloration was notably better. A skin biopsy was later obtained in July 2006, which confirmed the diagnosis of NSF (Figure 3). No histologic evidence of calciphylaxis was noted on the biopsy.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Thickened dermis with haphazardly arranged collagen bundles, increased dermal mucin, and scattered pigmented macrophages seen in a skin biopsy specimen in NSF. Biopsy was done approximately 1 to 1.5 mo after the initiation of STS therapy, with improvement in pain and skin changes.

The patient’s STS therapy was subsequently stopped in late August 2006 in view of her prolonged and complicated hospitalization for septic shock secondary to hemodialysis catheter–related infection. STS therapy has been reinstituted now, and a repeat biopsy is being considered.

Retrospective review of the patient’s medical record revealed exposure to multiple Gd³⁺-based MRI and an MR venogram in 2003, which could be temporally related to the development of skin changes (Table 3). The patient was also noted to have multiple vascular access surgeries. In relation to the development of her symptoms, the patient had an arteriovenous graft placed in the right groin in March 2003, for which she underwent thrombectomy in December 2003 (approximately 1 mo before her symptoms developed). Subsequently, this access failed and the patient had several tunneled and nontunneled dialysis catheters placed at multiple sites, all after her symptoms developed. Her current hemodialysis access is a tunneled femoral catheter. Hemodialysis was complicated by chronic hypotension; therefore, antihypertensive agents, including angiotensin-converting enzyme inhibitors, were not administered.

Role of Gadodiamide in the Development of NSF

Gadodiamide (Omniscan; Amersham Health, Amersham, UK) is a nonradioactive contrast agent that is approved by the FDA for use in MRI. Since its introduction, gadodiamide has become accepted as a safe alternative to iodine-based contrast agents in patients with impaired kidney function (9,10) and is usually well tolerated except for a few minor adverse effects. However, recent case reports have suggested an association between the use of gadodiamide and the development of NSF in patients with impaired kidney function. In a recent case series by Marckmann et al. (5), the delay from exposure to first sign of the disease was 2 to 75 days.

The normal elimination half-life of gadodiamide in healthy individuals is approximately 1.5 h and can be prolonged up to 10 to 60 h in patients with ESRD, especially when dialysis is delayed. The clearance of gadodiamide with peritoneal dialysis (69% after 22 days of continuous dialysis) is prolonged compared with its clearance in a single session of hemodialysis (65%) (11).

The stability of Gd³⁺ depends on its being bound to a ligand (e.g., diethylenetriaminepentaacetic acid). The adverse effect profile of gadodiamide is likely due to the in vivo dissociation of Gd–ligand complex into Gd³⁺ ion and ligand (12). This process is accelerated in patients with renal failure because of a combination of metabolic acidosis and inadequate clearance of the Gd–ligand complex. Endogenous metals such as Zn²⁺, Cu²⁺, Ca²⁺, and Fe³⁺ also destabilize the complex by transmetallation (a process of displacing Gd³⁺ from its ligand by competitive binding), leading to its dissociation (12). Grobner (4), in a recent case report, suggested that acidosis might be an essential co-factor in the pathogenesis of NSF, but this was not the case in the study by Marckmann et al. (5), as well as with our patient (Table 2).

The dissociated Gd³⁺ ion has poor solubility and could form in vivo precipitates of salts by chelating with anions such as phosphate (which is elevated in patients with renal failure), carbonate, and hydroxyl, with eventual deposition in various tissues, such as liver, bone, skin, muscle, and the interstitium (12). Two different sequences of pathophysiologic events, as a reaction to noxious agents (e.g., tissue deposits of Gd), have been proposed in the development of NSF; the first hypothesis suggests possible infiltration of the affected tissues by CD68⁺/XIIIa⁺ dendritic cells as a host response to noxious substances. These activated dendritic cells produce TGF-β, which not only initiates fibrosis but also enhances dendritic cell function, thereby initiating a vicious cycle of events that cause excess accumulation of dendritic cells in the affected tissues and extensive tissue fibrosis (13). A second hypothesis suggests the possibility that bone marrow–derived CD45RO⁺/CD34⁺/collagen I⁺ circulatory fibrocyte is released as a response to noxious stimuli in patients with NSF, resulting in fibrosis (14).

Recent case reports suggested development of NSF after administration of gadodiamide for MR angiography (which involves administration of gadodiamide at a dosage three times higher than the dosage approved by FDA for MRI in healthy patients), but our patient became symptomatic after repeated Gd³⁺-based MRI exposures. This leads us to hypothesize a dosage-dependent effect of the association of gadodiamide with the development of NSF (G.S., personal observation).

Additional factors that have been postulated to contribute to the pathogenesis of NSF include large boluses of intravenous iron and recombinant epoetin (8,15). In addition to Gd³⁺, iron and other metal deposits have been described in skin specimens from patients with NSF. It is possible that the interactions between iron molecules and the ligand agent might contribute to the development of NSF. Our patient was receiving both epoetin (100 to 200 U/kg with each dialysis) and intravenous iron (100 mg/wk) in the months before and after the development of her symptoms.

Swaminathan et al. (15) found a higher median weekly epoetin dosage among case patients than control subjects (427 versus 198 U/kg). Of note, case patients had a higher serum ferritin and lower serum albumin (which was the case with our patient [Table 1]) than control subjects, suggesting chronic inflammation and decreased epoetin responsiveness. Although it is conceivable that the epoetin itself could be contributing to the development of NSF, another hypothesis is that patients with chronic inflammation have significantly greater endothelial dysfunction, leading to the extravascular accumulation of gadodiamide and leading to NSF. Alternatively, chronic inflammation may result in an accelerated fibrotic response. Here, higher dosages of epoetin may simply be a manifestation of resistance to epoetin secondary to the inflammation.

Possible Mechanism of Action of STS in NSF

STS (Na₂S₂O₃) is a white crystalline substance that has reducing/antioxidant and chelating properties with multiple uses, including photography, treatment of cyanide poisoning, prevention of carboplatin-related (16) and cisplatin-related (17) toxicity, and most recently in the treatment of calciphylaxis (18,19). We propose yet another potential use of STS in the treatment of the debilitating disease NSF.

STS has FDA-labeled indications for the treatment of acne, Tinea versicolor, and cyanide poisoning, but other uses are off-label. Available evidence from human studies suggests that it is a safe drug with minimal adverse effects at the dosages currently used. The most common adverse effects are nausea and vomiting (usually during the infusion). Usual dosages are 12.5 to 25 g with every dialysis.

It is important to realize that both calciphylaxis and NFD can be present in the same patients, as described in a study by Edsall et al. (20). Association between metastatic calcification and NSF has also been described in many studies (3,21–23). TGF- β/Smad signaling cascades have been proposed as common factors in the pathogenesis of both of these conditions (i.e., fibrosis in NSF and calcification in calciphylaxis/metastatic calcification [20]). Therefore, it is possible that there is a similar mechanism of action of STS in both NFD and calciphylaxis, but further studies are needed to establish this relationship.

The beneficial effect of STS in our patient was rather dramatic with marked improvement in pain and skin changes after the first few treatments. We hypothesized that the beneficial effects of STS could be due to its chelating and antioxidant properties (Table 4).

It is possible that STS may chelate Gd³⁺, like other cations, and enhance its solubility and stability in serum, thereby facilitating its excretion during dialysis. If the Gd³⁺ precipitates in the tissue are the trigger for the fibrosis in NSF, then its removal may have helped ameliorate our patient’s condition. The diffuse arterial calcification that was noted in our patient and the effect of STS in chelating cations including both Gd³⁺ and calcium could also explain the improvement in skin changes. However, the effects of STS occurred rapidly in our patient, a benefit that cannot be attributed solely to its chelating action.

Gd³⁺ has been shown to accumulate in sites adjacent to recent endothelial trauma or inflammation, eventually leading to the worsening of endothelial dysfunction in patients who have chronic kidney disease or are on dialysis (which was the case with our patient) (24). STS could restore this endothelial dysfunction by its antioxidant properties (ability to donate electrons to pair unpaired free radicals), which has a positive effect on endothelial nitric oxide (eNO) synthase uncoupling and in the production of eNO (19). eNO has several positive effects in the maintenance of a healthy, normally functioning endothelium, including scavenging of reactive oxygen species, anti-inflammatory effect, fibrinolysis, and vasodilation. It is also likely that STS has the same beneficial effect in the neuronal vascular bed—the vasa nervorum and the endoneurium of the peripheral neuronal unit (25)—which may explain the rapid improvement of pain. The chelation of calcium/Gd³⁺ may also contribute to its beneficial effect; however, this is most likely a long-term effect.