Nephrogenic systemic fibrosis (NSF) was first described in 1997 in patients with end-stage renal disease (ESRD) [1]. It is characterised by scleroderma-like skin changes that mainly affect the limbs and trunk. The induration of the skin can progress to cause flexion contracture of the joints. The fibrotic changes may also affect other organs such as muscles, heart, liver and lungs [1–7]. The disease can be aggressive in some patients, leading to serious physical disability or even death [1]. Recently, it has been suggested that extracellular gadolinium based contrast agents (Gd-CA) might have a causal relation with NSF [4–7]. According to a recent editorial there are more than 150 patients who have developed NSF following administration of Gd-CA. Strikingly, the overwhelming majority (,90%) followed administration of the non-ionic agent gadodiamide (only 15% of all Gd-CA injections worldwide are gadodiamide) [7]. This was completely unexpected considering the good safety records of all Gd-CA, including gadodiamide [8]. According to the manufacturer (GE Healthcare, Waukesha, WI) gadodiamide has been administered to about 30 million patients since its introduction for clinical use in 1993 without high incidence of important adverse effects [6]. Hence, this recently observed strong association between NSF and gadodiamide requires an explanation. Currently, there are seven extracellular Gd-CA available for clinical use (Table 1). They are all chelates containing the Gd ion (Gd). The configuration of the molecules is either linear or cyclic. They are available as ionic or non-ionic preparations (Table 1). The important difference between these agents that could be of relevance to patients with advanced renal impairment (GFR ,30 ml min) and a factor in the development of NSF is the stability of the chelate molecule [6]. Gd-CA are eliminated from the body through the kidneys and biological half-life in patients with normal renal function is 1.5 h. In patients with advanced renal impairment, elimination half-life can be prolonged to 30 h or more [9]. Patients on haemodialysis would require three consecutive dialysis sessions over 6 days to remove 97% of the administered dose of Gd-CA from the body. Continuous ambulatory peritoneal dialysis for 20 days eliminates 69% of the injected dose of Gd-CA [10]. Transmetallation is likely to occur when the Gd-chelate remains in the body for a long period, as is the case in patients with ESRD, including those on dialysis. Transmetallation of Gd-CA leads to release of free gadolinium through replacement of the Gd within the chelate molecule by body cations such as zinc or copper [11]. Free gadolinium is highly toxic and animal studies showed that it can cause splenic degeneration, central lobular necrosis of the liver and a variety of haematological abnormalities [9, 12]. Therefore, it is crucially important that Gd should be strongly attached to a chelate to avoid its toxic effects. Understanding the synthesis of metal chelates is somewhat difficult, especially for those of us who have no deep knowledge in chemistry. However, the author of the article attempted to present some of the chemical principles involved in the production of Gd-chelate in a simplified manner, hopefully without important compromise of scientific accuracy. The gadolinium ion has Address correspondence to: Professor Sameh K Morcos, Consultant Radiologist, Department of Diagnostic Imaging, Northern General Hospital, Herries Road, Sheffield S5 7AU, UK. E-mail: sameh.morcos@sth.nhs.uk The British Journal of Radiology, 80 (2007), 73–76