Nephrogenic systemic fibrosis (NSF), previously known as nephrogenic fibrosing dermopathy, is an idiopathic condition seen in patients with renal disease that is characterized by cutaneous sclerosis that can often result in contractures, pain, and functional disability as well as systemic complications. Recent reports have suggested a possible link with exposure to gadolinium, a commonly used radiocontrast agent. No current therapy has clearly demonstrated efficacy for NSF, although case reports suggest that extracorporeal photopheresis (ECP) may be of benefit. The purpose of this study was to explore the plausibility of a gadolinium linkage with NSF as well as to assess the efficacy of ECP in the treatment of a cohort of patients with NSF.
We report our experience with 8 consecutive patients with NSF seen at the Stanford Medical Center, Palo Alta, California, from 2004 to 2006. Of the 8 patients, 6 had a history of arterial or venous thrombotic disease and 7 had a documented exposure to gadolinium within 1 week to several months prior to the onset of NSF. Specifically, all patients were exposed to gadodiamide. We treated 5 of the patients with ECP. After a mean number of 34 treatment sessions over a mean of 8.5 months, 3 patients experienced a mild improvement in skin tightening, range of motion, and/or functional capacity.
Our data support the hypothesis that exposure to gadolinium, perhaps specifically gadodiamide, plays a role in the pathogenesis of NSF. Larger epidemiologic studies will be needed to confirm this association. In addition, our experience suggests that, if used for extended periods, ECP might have some mild benefit for patients with NSF. Larger, randomized, placebo-controlled trials of ECP should be performed to more specifically assess the benefit of ECP in the treatment of NSF.
Nephrogenic systemic fibrosis (NSF), previously called nephrogenic fibrosing dermopathy, is an acquired, progressive, systemic fibrosing disorder of unknown etiology that develops in the setting of renal disease.1,2 Nephrogenic systemic fibrosis is clinically characterized by plaquelike, woody induration of the skin that frequently initially manifests with peau d’orange surface changes and may involve the deeper underlying tissues including the subcutis, fascia, and even skeletal muscle.1 Systemic sites of involvement that have been reported include skeletal muscle (including myocardium and diaphragm), pericardium, lung, and dura mater, as well as dystrophic calcification of other organs such as the kidneys and testes.1,3- 5 Symptoms include burning pain, pruritus, weakness, and joint contractures that can impair range of motion and limb function.
There are approximately 200 cases that have been reported to the international NSF Registry to date (http://www.icnfdr.org),6 and data from an analysis of 175 of these patients demonstrate a male-female ratio of 1:1. Nephrogenic systemic fibrosis has been shown to affect all ages, ranging from 8 to 87 years, including at least 10 cases in the pediatric age range.7,8 This disease has been reported worldwide and has been described in white, Indian, Middle Eastern, Asian, Hispanic, and African American subjects.1,8
To our knowledge, no case has been identified prior to 1997, and the first case series was published in 2000.9 The fact that this disease appeared within the past 10 years suggests that the etiology of NSF may be linked to a factor that was not present before that time, such as a new medication, chemical, or infectious agent.1 Potentially associated factors have been implicated, including recent surgery or other vascular procedures, local trauma, thrombotic events, hypercoagulable states, pulmonary fibrosis, hepatic disease, and high-dose erythropoietin.2,10,11 One current model for the development of NSF involves a blood-borne cell of bone marrow origin, the circulating fibrocyte, which is involved in the normal wound healing process and is thought to be recruited to the sites of involvement and to subsequently cause the fibrotic changes that are associated with this disease.1 Cowper and colleagues12 suggest that recruitment of circulating fibrocytes to the tissues is a normal occurrence in patients with NSF, representing a physiologic response to stimuli that trigger a normal wound healing response (ie, clotting and endothelial injury) or possible passively in some cases (ie, secondary to edema). The authors contend that preexisting deposition of allergens, medications, or radiographic contrast agents might then serve as surrogate targets for these cells, resulting in their activation and initiation of the fibrotic process.1,12 Consistent with this notion, gadolinium, the contrast agent used for magnetic resonance imaging (MRI), has been implicated as an associated agent in 2 case series.13- 15
Treatment of patients with this disease is challenging, and even among the few patients who regain normal kidney function, many do not experience resolution of their symptoms. Many treatments under investigation include plasmapheresis, extracorporeal photopheresis (ECP), systemic and topical steroids, topical calcipotriene, selective histamine blockade, thalidomide, psoralen-UV-A (PUVA) light, cyclophosphamide, cyclosporine, methotrexate, interferon alfa, intravenous immunoglobulin, oral retinoids, and aggressive physical therapy.2 Unfortunately, none of these has been shown to be broadly efficacious, and many patients do not improve with these therapies. Extracorporeal photopheresis has been reported to be effective in reversing some of the symptoms of NSF in affected patients even in the face of persistent renal insufficiency.7,16,17
The purpose of our study was to document our data pertaining to gadolinium exposure as well as the effects of ECP in our patients with NSF.
Between January of 2004 and October of 2006, we treated 8 patients with NSF who had been seen at the Stanford Medical Center, Palo Alto, California. Diagnostic criteria included the presence of renal insufficiency at the time of onset of the disease, a biopsy finding consistent with NSF (hypercellular dermis with an increase in CD34+, spindle-shaped, fibroblastlike cells and dermal histiocytes, thickened collagen bundles, and a variable increase in mucin deposition), and characteristic clinical distribution of lesions over the distal extremities and trunk with sparing of the head and neck. With 1 exception (see “Results” section), the patients did not have paraproteinemia, anticentromere antibodies, anti-Scl antibodies, Raynaud phenomenon, any known underlying malignancy, or a history of an allogeneic transplant. These data made the following diagnoses unlikely: scleromyxedema, scleroderma, paraneoplastic systemic sclerosis, and sclerodermoid graft-vs-host disease. During laboratory evaluation, 1 patient (patient 8) was found to have a monoclonal gammopathy consisting of IgG-λ. The patient's clinical presentation and onset of cutaneous disease shortly after acute onset renal failure, however, were more consistent with NSF. This conclusion was also reached by multiple clinicians at a second academic institution (Department of Dermatology Grand Rounds, University of California, San Francisco, June 2005).
Extracorporeal photopheresis is an immunomodulatory therapy that involves ex vivo treatment of a patient's leukocyte rich plasma with a photosensitizing agent and UV-A radiation followed by reinfusion of the treated blood product. The ECP treatment was performed in an outpatient setting using the UVAR XTS photopheresis machine with UVADEX (Therakos Inc, Exton, Pennsylvania) according to standard procedures in the manufacturer's guidelines. Four of the patients were treated with 1 cycle consisting of 2 treatments on consecutive days. Cycles were repeated every 2 to 3 weeks. One patient (patient 7), who was receiving peritoneal dialysis, was given a single ECP treatment every week to avoid undesired excess volume with each treatment.
The degree of response to treatment was determined by using a global physician assessment that took into account quantitative skin induration (using a modified Rodnan score of affected areas),18 range of motion (when applicable), and patient perception of disability. All patients were assessed by the same physician (Y.K.) to assure consistency in response measurement. The patients' disease (compared with baseline) was scored as worsened, stable, mildly improved, or markedly improved.
All patients had renal insufficiency or renal failure of varying causes at the time of disease onset, and all patients received some form of dialysis (Table 1). Of the 8 patients, 6 had a history of arterial or venous thrombotic disease. Of the 4 patients undergoing evaluation for a hypercoagulable state, 3 were found to have elevated homocysteine levels, and 1 patient was also positive for anticardiolipin antibodies (Table 2). Two patients had possible extracutaneous disease: 1 patient (patient 4) had pleural nodularity on chest radiography (without diagnostic biopsy) and a second patient (patient 1) had yellow scleral plaques, a finding previously documented in NSF.19
In light of recent data implicating gadolinium exposure with NSF, we attempted to obtain complete documentation regarding gadolinium exposure in our cohort. Of 8 patients, 7 had documented radiographic imaging using gadolinium contrast prior to the onset of symptoms, and all received gadodiamide (Table 1). Patients 1, 2, 3, 7, and 8 received the contrast agent during MRI. Patient 6 was exposed to gadodiamide during a concurrent head MRI/magnetic resonance angiography. Patient 4 developed symptoms consistent with NSF shortly after a radiology-directed liver biopsy in which gadodiamide was used. For the 5 patients with available records, the mean time to onset of symptoms was 3 to 4 weeks (range, 1-8 weeks) following gadodiamide exposure. Records regarding the exact date of onset of symptoms were not obtainable for 2 patients with documented gadolinium exposure, but the onset of symptoms occurred within months of the specified procedure. One patient (patient 4) had transient acute renal failure that lasted only 10 days, during which time she was given gadodiamide contrast. Although her renal failure subsequently resolved, she developed NSF within 8 weeks of gadodiamide exposure but continues to slowly improve without therapy. Another patient (patient 3) underwent a successful renal transplantation in 2004 and, despite normal renal function, continued to experience stable cutaneous disease (for over a year) before therapy.
We treated 5 of our patients with ECP (Table 3). At the time of treatment, all of these patients had progressive disease (based on patient history and physician global assessment of skin sclerosis) that had been present for a mean of 2.75 years. After a mean number of 34 treatment sessions (17 courses) over a mean of 8.5 months, 3 patients experienced mild improvement in both skin tightening and/or objective range of motion despite persistent renal insufficiency (Table 3; data not shown). The mean number of treatments before improvement was observed was 20 (range, 6-30). These patients were able to perform activities limited to them before ECP treatment, such as walking without mechanical assistance (patients 2 and 7) or climbing a flight of stairs unassisted (patient 3), leading to a significant improvement in their quality of life.
We report our experience with NSF in a cohort of 8 patients at the Stanford Medical Center. Consistent with other reports, all of our patients had renal insufficiency or failure (1 transient and 7 chronic), and as in the majority of documented cases, all were receiving some form of dialysis.1 Consistent with previously published data on this disease process, improvement in renal function does not necessarily result in remission of NSF; of our patients who had transient renal failure (patient 4), significant improvement in renal function (patient 8), or a successful renal transplantation (patient 3), all continue to manifest symptoms of NSF. Only 1 of these patients (patient 4) demonstrated improvement without therapy.
Interestingly, 6 of 8 patients had a history of arterial or venous thrombosis. This has been described in patients with NSF,10 and circulating anticardiolipin antibodies were a consistent finding in at least 1 cohort,20 although this is not a consistent finding.21,22 Of the 4 patients in this series evaluated for anticardiolipin antibodies, only 1 was found to have a detectable level. Serum homocysteine level, another risk factor for thrombosis, was elevated in 3 of the 4 patients tested. This result is not surprising, since multiple reports have linked even mild renal disease with elevated homocysteine levels, and the prevalence of hyperhomocysteinemia in hemodialysis patients is between 80% and 100%.23,24 It is unclear if a hypercoagulable state is associated with NSF or simply represents the known association with renal failure.25,26 It is tempting to speculate that some form of endothelial damage might play a role in the pathogenesis of NSF, perhaps leading to the physiologic recruitment of circulating fibrocytes to the skin.1,12
Our data are consistent with a potential causative link between gadolinium contrast and NSF. All patients for whom we have a definitive set of complete records demonstrated a temporal association between the administration of gadolinium contrast and the development of cutaneous symptoms. The patient without documented gadolinium exposure is deceased, although there are no hospital records at our institution of any procedures in which the gadolinium was administered. Of note, this patient received care outside of our institution and had multiple medical problems for which an MRI/magnetic resonance angiography may have been administered. A discussion with the patient's family members, primary care physician, and nephrologist, however, failed to uncover records of any such procedure.
There are 5 different gadolinium-based contrast agents that are approved by the Food and Drug Administration (FDA) for medical use in the United States including gadopentetate dimeglumine (Magnevist; Bayer HealthCare Pharmaceuticals Inc, Wayne, New Jersey), gadoteridol (ProHance; Bracco Diagnostic Inc, Princeton, New Jersey), gadodiamide (Omniscan; Nycomed Inc, Princeton), gadoversetamide (OptiMARK; Mallinckrodt, Hazelwood, Missouri), and gadobenate dimeglumine (MultiHance; Bracco Diagnostics Inc). These were approved by the FDA in 1988, 1992, 1993, 1999, and 2004, respectively, for use in MRI.
Interestingly, all of our patients were exposed specifically to gadodiamide. This is consistent with previous studies.13,14,27 Gadodiamide is a nonionic chelate of gadolinium and diethylenetriaminepentaacetic acid bis-(methylamide),28 and is characterized by having excess chelate and being less stable. This is due to the high propensity for this particular substance to undergo transmetallation with endogenous ions.29 It is possible that either free gadopentetate ion or the chelate itself could bind to endogenous ions in the tissues.30 This ion is poorly soluble and it can form salt precipitates with phosphates and other anions that are frequently elevated in patients with renal failure, which may lead to increased tissue deposition and inflammation.13 Indeed, gadolinium has been detected in the skin of patients with NSF, most likely in intracellular deposits.15,30 This is consistent with the hypothesis that this might be a trigger for resident CD34+ cells to synthesize extracellular matrix components, including collagen. According to the FDA Web site, NSF has been associated with 3 of the 5 FDA-approved gadolinium-containing agents, and they warn that all of the available agents might have the potential for this association.27
Case reports suggest that ECP might be an effective treatment for some patients with NSF. In one report of 3 patients, all demonstrated significant improvement in terms of joint mobility and both objective and subjective skin softening following at least 4 cycles of ECP. One patient developed complete resolution of symptoms after 16 cycles of treatment.16,17 Auron et al7 report 1 case of a pediatric patient who experienced symptomatic improvement following 1 year of 2-day sessions occurring every other week. Finally, although the report by Gilliet et al17 suggests that significant improvement can be seen as early as 8 weeks, other patients have required much longer treatment periods to show benefit. Although none of our patients had as dramatic a response as those in the previous reports, results were encouraging. Of the 5 patients treated with ECP, 3 had some form of objective response. In all of the responding patients, the decreased sclerosis was primarily seen in the upper extremities. Such “minor” objective improvement can offer significant hope and benefit in the quality of life of patients, since all patients had a notable improvement in function. The 2 patients without objective improvement demonstrated stabilized disease while receiving ECP; because the disease was progressive before therapy, this might also indicate a therapeutic benefit. It is unclear whether technical differences in the ECP treatment are responsible for the varied outcomes between our cohort and those previously reported. In addition, the time lapse between disease onset and the start of ECP treatment was not mentioned in the previous reports. Our patients had a mean disease duration of 2.75 years before they received ECP. It is possible that patients treated early in their disease course, before extensive sclerosis has developed, would benefit most from this therapy. It should also be noted that acitretin therapy was used as adjuvant therapy in many of our patients, based on positive personal experience from the authors (Y.K. and D.F.) in other sclerosing disorders of the skin, such as scleromyxedema and chronic graft-vs-host disease (and limited data in NSF2). For 2 of our patients responding to ECP (patients 2 and 7), the contribution of the retinoid cannot be discounted.
Correspondence: David Fiorentino, MD, PhD, Department of Dermatology, Stanford University, 900 Blake Wilbur Dr, Stanford, CA 94305 (firstname.lastname@example.org).
Financial Disclosure: None reported.
Accepted for Publication: April 20, 2007.
Author Contributions: Dr Fiorentino had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Ms Richmond and Dr Zwerner contributed equally to this work. Study concept and design: Richmond, Kim, and Fiorentino. Acquisition of data: Richmond, Zwerner, Kim, and Fiorentino. Analysis and interpretation of data: Richmond, Zwerner, Kim, and Fiorentino. Drafting of the manuscript: Richmond. Critical revision of the manuscript for important intellectual content: Zwerner, Kim, and Fiorentino. Administrative, technical, and material support: Richmond, Zwerner, and Fiorentino. Study supervision: Kim and Fiorentino.
Funding/Support: Dr Fiorentino is partially supported by a Career Development Award from the Dermatology Foundation.
This article was corrected for typographical errors on 10/10/2007.
Richmond H, Zwerner J, Kim Y, Fiorentino D. Nephrogenic Systemic FibrosisRelationship to Gadolinium and Response to Photopheresis. Arch Dermatol. 2007;143(8):1025-1030. doi:10.1001/archderm.143.8.1025