Author + information
- Received August 11, 2008
- Revision received December 3, 2008
- Accepted December 8, 2008
- Published online May 5, 2009.
- Andreas Kribben, MD*,* (, )
- Oliver Witzke, MD*,
- Uwe Hillen, MD†,
- Jörg Barkhausen, MD‡,
- Anton E. Daul, MD∥ and
- Raimund Erbel, MD§
- ↵*Reprint requests and correspondence:
Dr. Andreas Kribben, Department of Nephrology, University Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germany
Nephrogenic systemic fibrosis (NSF) is a newly recognized disorder occurring exclusively in patients with renal failure. Exposure to gadolinium-based magnetic resonance (MR) contrast media has been associated with subsequent development of NSF. Nephrogenic systemic fibrosis is characterized by skin induration preferentially affecting the extremities. In addition, involvement of internal organs occurs, which leads ultimately to death. Skin biopsy is important for confirmation of the diagnosis. The main therapeutic goal is restoration of renal function. To reduce the risk of NSF, renal function must be determined before exposure to gadolinium-containing MR contrast agents. Gadolinium-based MR contrast media should be avoided in the presence of advanced renal failure with estimated glomerular filtration rate below 30 ml/min/1.73 m2, unless the diagnostic information is essential and not available with noncontrast magnetic resonance imaging techniques. The recommended dose of contrast agent should not be exceeded. In addition, a sufficient period of time for elimination of the contrast agent from the body should be allowed before readministration of the contrast agent.
Nephrogenic systemic fibrosis (NSF) is a new disorder exclusively observed in patients suffering from renal failure. Because it was initially assumed that the disorder was limited to the skin, the term “nephrogenic fibrosing dermatopathy” was chosen (1). The recognition of the disorder's systemic nature with fibrotic changes in various organ systems led to the renaming of the disease as nephrogenic systemic fibrosis (2). As of May 2008, 215 cases have been described in a U.S. registry (3). Approximately 80 cases are cumulatively reported within European patients; the reports mainly originated from Austria and Denmark. A registry for Germany was opened in the summer of 2007 (4). Gadolinium-containing magnetic resonance (MR) contrast agents seem to be associated with the disease development (5,6). This overview summarizes the current knowledge about the pathogenesis, diagnosis, and therapy in order to increase awareness of this new syndrome.
Nephrogenic systemic fibrosis was only recognized after large numbers of patients were given gadolinium-based contrast agents. Before 1997 the disease was not reported. There was a change in clinical practice and use of high-dose gadolinium-enhanced magnetic resonance angiography (MRA) for improved imaging. Gadolinium-enforced MRA was used preferentially in patients with renal failure to avoid iodinated (X-ray) contrast agents. The increasing cumulative dosage of these gadolinium-containing contrast agents might have contributed to the development of NSF. Nevertheless, the combination of rapid, higher-than-approved doses of contrast agents and the clustering of cases were probably important factors that likely led to the recognition of this new syndrome.
Additional unknown risk factors might play a role in facilitating NSF. Nephrogenic systemic fibrosis is diagnosed with equal frequency in men and women and affects patients of all ages and ethnicities (7–9). In a cohort of American dialysis patients, the frequency was 3 cases in 467 patients observed over 18 months, resulting in an estimation of 4.3 cases/1,000 patient-years (10). The risk of triggering the disorder in these patients through a gadolinium application with different gadolinium-based agents was calculated to be 2.4% (10).
A recent European study found NSF in 18 (18%) of 102 patients with stage 5 chronic kidney disease (CKD) with and without dialysis after exposure to the gadolinium derivate gadodiamide (Omniscan, GE Healthcare Medical Diagnostics, Amersham, United Kingdom). Most of the patients in this study received high doses of gadodiamide (typically approximately 0.3 mmol/kg) for angiography (11). Another study using Gadoteridol (ProHance, Bracco Diagnostics, Milan, Italy) in 141 patients receiving long-term hemodialysis found no NSF (12). A further study found discrete clinical signs of NSF in 30% of hemodialysis patients after gadopentetate dimeglumine (Magnevist, Bayer Schering Pharma AG, Berlin, Germany) exposure. However, skin biopsies were not used to confirm the diagnosis of NSF in this patient cohort (13). Prince et al. (14) assessed the incidence of NSF in 2 large medical centers in the U.S. Of 83,121 patients who received gadolinium-based contrast agents, 15 (0.02%) developed NSF. All of them got a high dose of the contrast agent that exceeded the standard dosage. Most of them suffered from acute renal failure at the time of administration; 8.4% of all patients with acute renal failure receiving gadolinium developed NSF. Wertman et al. (15) recently calculated the benchmark incidence of NSF with data from 4 large U.S. health care centers. The benchmark incidence of NSF ranged from 1 of 2,131 patients to 1 of 65,000 patients. This study confirmed previous findings showing that only patients with severe renal impairment and/or stage 4/5 CKD develop NSF.
According to current knowledge, impaired renal function seems to be a conditio sine qua non for NSF. Nephrogenic systemic fibrosis was described in patients with stage 4 and 5 CKD without dialysis, patients requiring dialysis (hemodialysis and peritoneal dialysis), patients who had received renal transplants, as well as patients with acute kidney injury (AKI).
CKD is defined as either kidney damage or glomerular filtration rate (GFR) <60 ml/min/1.73 m2for ≥3 months. Kidney damage is defined as pathologic abnormalities or markers of damage, including abnormalities in blood or urine tests or imaging studies (16). Chronic kidney disease is classified into 5 stages according to the GFR. Stage 1 CKD is diagnosed as kidney damage with normal or increased GFR (>90 ml/min/1.73 m2); stage 2 is diagnosed as kidney damage with a GFR of 60 to 89 ml/min/1.73 m2. Stage 3 CKD is defined by a GFR of 30 to 59 ml/min/1.73 m2, and stage 4 CKD is defined by a GFR of 15 to 29 ml/min/1.73 m2. Stage 5 CKD is defined as established kidney failure with a GFR <15 ml/min/1.73 m2or permanent dialysis therapy (16).
The extent of renal failure that triggers NSF is not known. There are only a few cases with a GFR >15 ml/min/1.73 m2reported. To the authors' best knowledge, there is no case with a reported GFR >30 ml/min/1.73 m2. Although severe renal impairment seems to be a major condition, caution should be advised in recommending a safe range for gadolinium exposure above a specific GFR until further details of the disease are elucidated.
The association between the occurrence of NSF and preceding gadolinium exposure within the context of an MR study was first demonstrated in 2006 by Grobner (5) and his team in Austria in 5 patients and by Marckman et al. (6) in 19 patients from Denmark. An evaluation of an American patient cohort indicated gadolinium exposure for patients before the manifestation of NSF. Gadolinium was found in the skin of the patients (17,18). This association led to warnings by the Food and Drug Administration (FDA) and the European Agency for the Evaluation of Medicinal Products regarding the use of gadolinium-containing contrast agents in patients suffering from renal failure (19,20).
In normal renal function, free gadolinium is removed by the kidney with a half-life of approximately 2 hours. In impaired renal function, this half-life is significantly longer. Two consecutive hemodialysis sessions remove approximately 93% of the gadolinium, whereas the peritoneal dialysis is significantly less effectively and only removes approximately 75% of the gadolinium after 5 days (21–23).
A recent meta-analysis by Agarwal et al. (24) demonstrated a significant association between gadolinium-based contrast agent exposure and NSF. Nevertheless, at the time of this writing it cannot be decided whether or not the risk of occurrence of NSF differs for the various gadolinium-containing contrast agents (Table 1).The majority of patients with NSF cases had a prior administration of gadodiamide (Omniscan, GE Healthcare Medical Diagnostics), although the product is used in only approximately 15% of magnetic resonance imaging (MRI) studies worldwide (25). It has been hypothesized that the occurrence of NSF after the application of gadodiamide is related to the lower stability and increased occurrence of toxic free gadolinium in the gadodiamide complex (26,27). Nevertheless, in some of the studies the gadodiamide dose was very high; thus it has to be considered that the high dosage rather than gadodiamide itself facilitated the development of NSF in these cases (5,6,28–31). According to the FDA, no MR contrast agent can principally be regarded as safe, because NSF has also been observed after exposure to other gadolinium-containing contrast agents in the U.S. (19): Gadopentetane dimeglumin (Magnevist, Bayer Schering Pharma AG), gadobenate dimeglumin (MultiHance, Bracco Diagnostics), gadodiamide (Omniscan, GE Healthcare Medical Diagnostics), gadoversatemide (Optimark, Mallinckrodt, St. Louis, Missouri), and gadoteridol (ProHance, Bracco Diagnostics). However, in some patients, coadministration of gadobenate dimeglumin (MultiHance, Bracco Diagnostics) and gadodiamide had been described (27). Nephrogenic systemic fibroses have been reported after administration of Gadoterate meglumin (Dotarem, Guerbet SA, Aulnay-sur-Bois, France), gadobutrol (Gadovist, Bayer Schering Pharma AG), gadoxet acid dinatrium (Primovist, Bayer Schering Pharma AG), or gadofosveset trinatrium (Vasovist, Bayer Schering Pharma AG) (Table 1). But, the repeated gadolinium exposures seems to increase the risk of NSF by 5- to 10-fold as compared with single gadolinium exposure (32).
The mechanisms of gadolinium-related damage to the skin and other organs are not well-understood (5,6,33). Gadolinium deposits are found in the skin of NSF patients, which implies a direct toxic effect of gadolinium (17,18). Gadolinium can block ion channels in an unspecific way. This effect, however, has mainly been shown in epithelial cells (34). In addition, gadolinium chelates undergo transmetallation in vivo (33). This might be triggered by elevation in free serum iron after gadolinium exposure. Exogenous iron treatment and chronic microinflammation in renal failure have been suggested to contribute to free gadolinium release (33). Iron induces transmetallation of gadolinium chelates (35). The combination of free gadolinium and the toxic effects of iron might result in initial tissue injury, especially under conditions of inflammation and oxidative stress in patients with renal failure.
It is currently unclear what types of fibroblasts contribute to the development of NSF. Cowper and Bucala (36) suggested that spindle cells seen in skin biopsies of NSF patients are derived from circulating fibrocytes, because these cells show CD34 and procollagen 1 expression. Why these cells seem to be persistently activated after the initial trigger is removed is currently unclear. Moreover, further analysis revealed metal deposition and accumulation in fibrotic areas of vessels. This might contribute to vascular stiffening and subsequent cardiovascular events.
The use of high doses of erythropoiesis-stimulating agents seems to be associated with NSF due to vascular injury and endothelial dysfunction with subsequent development of fibrogenesis (6,33,37). Furthermore, erythropoietin as well as parathyroid hormone induces the systemic release of CD34+cells that are involved in the fibrotic process of wound-healing (38,39).
Association with thrombophilia
A significant number of patients showed acute thromboses or coagulation abnormalities (2,8). These patients had occlusion of the hemodialysis access routes shortly before the manifestation of the NSF. This involved both native vessels and implanted dialysis catheters or artificial arteriovenous grafts. Deep vein thrombosis as well as renal vein and vena cava thrombosis have been observed in association with the subsequent development of NSF (40–42). Some patients had auto-antibodies to cardiolipin (43). The link of NSF to thrombosis might indicate an inflammatory genesis as a co-factor for the development of NSF (8,44).
Association with surgical procedures
Surgical procedures, particularly those with the placement of a blood vessel access for hemodialysis, have been associated with the development of NSF. If organ transplantation and the placement of central catheters are included, approximately 90% of the NSF patients reported such an event shortly before the start of symptoms. The perioperative activation of the coagulation and the inflammation might be responsible for this reaction (8,30).
Diagnosis: Symptoms and Findings
The cutaneous lesions of NSF are distributed symmetrically over the limbs and the trunk. Involvement of the legs frequently extends upward from the talocalcanean joints to the center of the thighs; on the arms, involvement extends from the wrist to the center of the upper arm. The primary efflorescences consist of coarse skin with or without erythematous papules that might form a confluence of brown, indurated cobblestone-like plaques (Fig. 1).The lesions might possess a peau d′orange aspect. Swelling of the arms and legs is occasionally observed (Fig. 2).The palpatory indurated areas might have a wood-like texture. Because the skin lesions frequently extend beyond the joints, restriction of movement ranging through to contraction might develop, even within days or weeks. The patients complain about pain or itchiness in the affected skin areas (8).
Subtle cutaneous changes, defined as hyperpigmentation, and hardening and tethering of the skin are found in 30% of hemodialysis patients after gadolinium (gadopentetate dimeglumine) exposure. These changes are associated with a significantly increased risk of mortality as compared with a hemodialysis cohort without cutaneous involvement (13). A scoring system was proposed assessing hyperpigmentation (comparing appearance of the skin on the extremities with skin elsewhere), hardening (by gradually applying increasing pressure while palpating the skin on the extremities in a circular motion), and tethering (resistance to motion while gently pinching the skin). A binary value of either 0 or 1 was assigned for these 3 characteristics, and a score of 2 and 3 was defined for clinical evidence of NSF (13). The validity of this approach and its comparison with histological data need to be explored by further studies.
Apart from the skin involvement, the skeletal muscle and the visceral muscle of the gastrointestinal tract (esophagus, stomach) and the cardiac muscles seem to be affected. In 1 study, the amount of cardiac gadolinium deposits was shown to be as high as in affected skin (45). In addition, cardiac fibrosis was revealed in 1 case with gadolinium deposits of the heart. This study also demonstrated gadolinium, iron, and aluminum accumulation in the blood vessels and skin of these patients. Moreover, the authors reported that a form of accelerated cardiovascular disease was found in patients who died after the onset of NSF (45); nevertheless, due to the design of the study, it cannot be concluded that this association is of causal type. Fibrosis of serous membranes, including the meninges, might also occur (40–42). Involvement of the eyes with scleral plaques has also been reported (46).
The most significant laboratory findings in patients with NSF consist of impaired renal function. Other factors described in NSF, like cardiolipin antibodies, antinuclear antibodies, and elevated calcium and phosphate concentrations, are not specific (8).
Skin and tissue biopsies
The diagnosis of NSF is confirmed by means of a skin biopsy (33,41,42,45). It is important to take a sufficiently large and deep biopsy (including subcutaneous fat). Histopathologic findings depend on the point of time in the course of disease in which the biopsy is taken. Fully developed lesions show a cell-rich dermal fibrosis that might extend to the subcutis. Collagen bundles are thickened and separated from one another by clefts and interposed thin collagen bundles. The cellular infiltrate consists of CD34-positive spindle cells and procollagen-I producing cells with a stellate or epitheloid appearance (47). In addition, factor XIIIa-positive cells as well as CD68-positive multinuclear macrophages can be observed. The immunohistopathological profile leads to the hypothesis that the cells are circulating fibrocytes, a cell population that originates from the bone marrow and seems to play a role in wound-healing processes. Skin biopsy results are not specific and therefore have to be correlated with the clinical presentation. Thus, clinical information given to the dermatopathologist should include renal failure and gadolinium exposure before the start of the disease (48). There are also characteristic histopathological changes in the affected muscle. These are particularly visible in areas that are directly below affected skin segments. In regions within clinically healthy skin, no changes are generally found, so that the muscle participation is closely linked to skin involvement. Myopathic changes are frequently diagnosed in the patient cohort; however, they seem more likely to represent secondary signs of uraemic neuropathy (8,49).
So far, imaging diagnostics have no clear value in making the diagnosis and in assessing the progression of NSF, because the findings are nonspecific (2,8).
The differential diagnosis of NSF includes a series of other fibrosing skin disorders that can be clinically and histologically differentiated. Scleromyxoedema shows a different distribution pattern (inclusion of the face and neck as well as, frequently, paraproteinemia). Systemic sclerosis might resemble NSF in affecting the extremities and the trunk, but it often also involves the face and might be discriminated from NSF serologically (by the presence of autoantibodies) and histopathologically. Other differential diagnoses such as pretibial myxoedema in Basedow's Disease or eosinophilic fasciitis are clearly delimitable clinically and/or by immunohistology (50).
Overall, there are currently no sufficient data available for recommendation of an effective therapy. Restoration of sufficient renal function is the primary goal when NSF is identified (51). This goal might be reached by recovery from AKI or by renal transplantation. Renal transplantation is not a well-assessed therapeutic option in NSF. Although a recent publication reported clinical improvement of patients with NSF after renal transplantation, there are only single case reports available and not all patients benefited from renal transplantation after NSF was diagnosed (8,32,51–56). Thus, the value of renal transplantation has to be further elucidated; there is not enough evidence for a clear conclusion or recommendation at present (32,51–55). In addition, physical therapy and pain management particularly for prophylaxis and therapy of joint contractures is the mainstay of therapy in NSF. Further therapy attempts were made with systemic corticosteroids, thalidomide, pentoxyphyllin, intravenous immunoglobulin, plasmapheresis and cyclophosphamide, and the chelating agent sodium thiosulfate. Overall, improvements were only anecdotally successful, so that no clear recommendations can be given (8,33). Extracorporeal photopheresis has been studied in 7 small studies with 1 to 5 patients reporting improvements of skin symptoms, implying that this relatively safe therapy might be a good initial treatment option (8,52,57).
Considerations for the use of gadolinium-containing contrast agents in patients with impaired renal function
The FDA has reported that exposure to gadolinium-containing contrast agents increases the risk for NSF in patients with severe acute or chronic renal failure with a GFR below <30 ml/min/1.73 m2and in patients with hepatorenal-syndrome or in the perioperative liver transplantation period. The FDA ordered Black Box warnings in 2007 suggesting that all gadolinium contrast media available in the U.S. might be linked to NSF and that gadolinium-based magnetic resonance contrast media should be avoided in the presence of advanced renal failure with estimated GFR below 30 ml/min/1.73 m2, unless the diagnostic information is essential and not available with other imaging techniques (Table 2)(19). European regulatory recommendations are slightly different: the Danish Medicines Agencies and the United Kingdom Medicines and Healthcare products Regulatory Agency focus on gadodiamide, because this agent is associated with the highest number of NSF patients. In view of the Danish Medicines Agencies and the United Kingdom Medicines and Healthcare products Regulatory Agency, gadodiamide is contraindicated in patients with a GFR <30 ml/min/1.73 m2(Table 2) (58,59). The American College of Radiology recommends using gadolinium-based contrast agents in patients with a GFR <30 ml/min/1.73 m2only if it is absolutely essential; the lowest possible dose has to be administered (Table 2) (60,61). The American College of Radiology also stressed that no gadolinium-based contrast agents can be regarded as safe (61).
In Germany, the manufacturer of gadodiamide recommends that this product should not be used in patients with an impaired GFR of <30 ml/min/1.73 m2(62). Meanwhile, such recommendation has also been issued for gadopentetane dimeglumin (Magnevist, Bayer Schering Pharma AG) (62). Therefore, renal function should be determined by history and/or laboratory tests before administering gadolinium-containing MR contrast agents. The GFR should be assessed by measuring creatinine clearance or by means of the abbreviated Modification of Diet in Renal Disease formula. However, in case of acute renal failure, renal function cannot be determined by the Modification of Diet in Renal Disease formula, because creatinine is not in steady state. (16).
Alternatives to gadolinium-based MRI studies
For all patients at risk, it must be determined whether the diagnostic information can be obtained without the administration of contrast. There are several alternatives to contrast-enhanced 3-dimensional MRA for vascular application. Time of flight and phase contrast MRA introduced more than 20 years ago have almost completely been replaced by contrast-enhanced 3-dimensional MRA but still allow visualization of vascular pathologies (e.g., carotid artery stenoses). Recently, new techniques have been introduced for non-contrast MRA of different vascular territories. Steady state free precession sequences can be applied for the assessment of the aorta (63,64), the renal arteries (65), as well as the coronary arteries (66). Additionally, new techniques have been introduced to cover a large field of view required for peripheral MRA (67). In some applications, ultrasound might be an alternative in selected patients with renal failure. However, ultrasound contrast microbubbles have many limitations in patients with different types of cardiac disease, and they are contraindicated in pregnancy and in pediatrics (68). Contrast enhanced computed tomography could be a viable alternative in patients with renal failure. The nephrotoxicity of iodine contrast media has to be taken into account (69,70).
Strategies for prevention of NSF in high-risk patients with gadolinium exposure
If gadolinium-containing contrast agents are to be used in patients with stage 5 CKD or AKI, one of the most controversial issues is the need for hemodialysis after gadolinium exposition. Currently there are no data supporting any of the following recommendations. However, if hemodialysis after gadolinium exposure has a role, it seems logical to remove as much gadolinium as soon as possible after exposure. The FDA and the American College of Radiology recommend considering “prompt hemodialysis” and hemodialysis “no later than 2 hours” after gadolinium exposure (Table 2) (19,60). Two hemodialysis sessions that are performed on the following days could remove approximately 93% of the injected gadolinium (23). Because elimination of the gadolinium via peritoneal dialysis is considerably lower than via hemodialysis (21), it is reasonable to perform hemodialysis procedures after gadolinium exposure in peritoneal dialysis patients (71). Because exogenous iron therapy and application of ESA has been suggested to contribute to the development of NSF (33,37), it might be reasonable to suspend medication with these agents before gadolinium application.
In patients with stage 4 CKD, the discussion about the use of gadolinium-containing contrast agents is currently underway and far from being completed, due to insufficient data (7,71,72). For the time being, gadolinium-containing contrast agent should only be used in patients suffering from stage 4 CKD with a strict indication and after obtaining written informed consent. Dialysis after gadolinium exposure in these patients does not seem to be indicated, due to the high morbidity associated with placement of a large lumen dialysis catheter (73). The applied quantity of gadolinium-containing contrast agent should be kept as low as possible in all patients with renal failure, and repeated examinations should be avoided. In planning the MR examination, several organs and body regions should be examined with a single application of the contrast agent.
Consequences for Clinical Work and Practice
• NSF is a potentially lethal systemic disorder that frequently manifests itself on the skin in its initial stages.
• Nephrogenic systemic fibrosis is observed in patients suffering from renal failure after exposure with gadolinium-containing MR contrast agents.
• Renal function should be determined by history and/or laboratory tests before administering gadolinium-containing MR contrast agents. Strict indication for MRI is necessary in the presence of advanced renal failure.
• In patients with stage 4 or 5 CKD, contrast-free MRI has to be used when possible.
• Skin biopsies can be used for confirmation of the diagnosis.
• There is no established therapy for NSF.
The subject of this article is independently presented, and the depiction of contents is presented with product neutrality. Dr. Kribben has received lecture fees from Bayer Schering AG and GE Healthcare Buchler (Amersham). Dr. Barkhausen has received speaker honorarium from Bayer, Covidien, and Guerbet, and grant support from Bayer and Guerbet.
- Abbreviations and Acronyms
- acute kidney injury
- chronic kidney disease
- Food and Drug Administration
- glomerular filtration rate
- magnetic resonance angiography
- magnetic resonance imaging
- nephrogenic systemic fibrosis
- Received August 11, 2008.
- Revision received December 3, 2008.
- Accepted December 8, 2008.
- American College of Cardiology Foundation
- The International Center for Nephrogenic Fibrosing Dermopathy Research (ICNFDR)
- Kribben A.
- Grobner T.
- Marckmann P.,
- Skov L.,
- Rossen K.,
- et al.
- Deo A.,
- Fogel M.,
- Cowper S.E.
- Reilly R.F.
- ↵Food and Drug Administration Public Alert. Information for health care professionals: gadolinium-based contrast agents for magnetic resonance imaging (marketed as magnevist, Multihance, Omniscan, OptiMARK, ProHance). http://www.fda.gov/cder/drug/InfoSheets/HCP/gcca_200705.htm. Accessed August 5, 2007.
- European Agency for the Evaluation of Medicinal Products (EMEA)
- Agarwal R.,
- Brunelli S.M.,
- Williams K.,
- et al.
- Morcos S.K.
- Othersen J.B.,
- Maize J.C.,
- Woolson R.F.,
- et al.
- Swaminathan S.,
- Shah S.V.
- Heeschen C.,
- Aicher A.,
- Lehmann R.,
- et al.
- Artunc F.,
- Schanz S.,
- Metze D.,
- et al.
- Leung N.,
- Shaikh A.,
- Cosio F.G.,
- et al.
- Pesek G.D.,
- Tyler L.,
- Theus J.,
- et al.
- ↵Medicines and Healthcare products Regulatory Agency (online 7 February 2007) Public assessment report: increased risk of nephrogenic fibrosing dermopathy/nephrogenic systemic fibrosis and gadolinium-containing MRI contrast agents. http://www.mhra.gov.uk/home/idcplg?IdcService=SS_GET_PAGE&ssDocName=CON2030229&ssSourceNodeId=221&ssTargetNodeId=221. Accessed October 17, 2008.
- Segal A.,
- Ellis J.,
- Baumgartner B.,
- et al.
- Kane G.C.,
- Stanson A.W.,
- Kalnicka D.,
- et al.
- Thomsen H.S.,
- Marckmann P.,
- Logager V.B.