Diabetic proximal neuropathy goes by a bewildering variety of names: diabetic femoral neuropathy, amyotrophy, neuropathic cachexia, anterior neuronopathy, polyradiculopathy, myopathy, mononeuritis multiplex or lumbar plexopathy. We are all familiar with the typical syndrome in elderly in Type 2 diabetics who develop severe unilateral anterior thigh pain followed by wasting and weakness of the quadriceps muscle with loss of the knee jerk [1]. This soon becomes bilateral in roughly 50 per cent of patients. The underlying pathogenic mechanisms have not yet been identified. It is presumed that peripheral nerve infarcts are responsible for those cases which develop acutely over a few days. However, an earlier report of small peripheral nerve infarcts in an autopsy study [2] has been tempered by subsequent recognition that such lesions can be difficult to distinguish from Renaut bodies, a non-specific histopathological finding [3]. Furthermore, any theory that proximal diabetic neuropathy represents nerve infarction due to diabetic microvascular disease would have to account for the rarity with which it is accompanied by diabetic retinopathy or nephropathy, [4] the excellent recovery achieved by many patients, and the marked tendency for abnormalities to develop symmetrically on the two sides within a few weeks of one another [4, 5]. The other end of the clinical spectrum is less well recognized: symmetrical proximal muscle weakness of slow onset, which is relatively painless and is presumed to be metabolic in origin [6-8]. A paper in this issue of the journal by Coppack and Watkins [5] addresses two important issues concerning the treatment and prognosis of diabetic proximal neuropathy. First, does hypoglycaemic therapy really aid recovery of the leg weakness? Second, to what extent do these patients ultimately recover from disability which may be seriously incapacitating for many months? There is no doubt that many patients with diabetic proximal neuropathy do recover. Few physicians can resist the temptation to improve glycaemic control in patients who develop diabetic proximal neuropathy and most feel this action is responsible for recovery when it occurs. But is hypoglycaemic therapy itself really responsible for recovery, or is recovery merely part of the natural history of diabetic proximal neuropathy? Garland [9] stated that 'the only treatment necessary is full diabetic control. . . the use of insulins is almost invariably essential' a view supported by Casey and Harrison [10]. However it is clear that diabetic proximal neuropathy can develop in patients with rather mild glucose intolerance, and that recovery can coincide merely with the institution of a low calorie diet, without additional hypoglycaemic medication [7, 10].