L

Reversal of corticosteroid-induced mellitis with supplemental chromium

A. Ravina*, L. Slezak*, N. Mirsky*, N. A. Bryden² and R. A. Anderson²

Abstract

*Department of Diabetes, The Linn Clinic, Haifa, Aims To determine if the stress of corticosteroid treatment increases chromium Department of, Biology, Oranim University of (Cr) losses and if corticosteroid-induced diabetes ( diabetes) can be reversed Haifa, Israel by supplemental chromium. ² Nutrient Requirements and Functions Laboratory, Beltsville Human Nutrition Research Methods The effects of corticosteroid treatment on chromium losses of 13 patients Center, Beltsville, MD, USA 2 days prior to steroid administration and the ®rst 3 days following treatment were determined. Since steroid-induced diabetes was associated with increased Received 20 May 1998; revised 3 August 1998; accepted 18 August 1998 chromium losses and insuf®cient dietary chromium is associated with glucose intolerance and diabetes, we treated three patients with steroid-induced diabetes with 600 mg per day of chromium as chromium picolinate.

Results Urinary chromium losses following corticosteroid treatment increased from 155 6 28 ng/d before corticosteroid treatment to 244 6 33 ng/d in the ®rst 3 days following treatment. Chromium supplementation of patients with steroid- induced diabetes resulted in decreases in fasting blood glucose values from greater than 13.9 mmol/l (250 mg/dl) to less than 8.3 mmol/l (150 mg/dl). Hypoglycaemic drugs were also reduced 50% in all patients when given supplemental chromium.

Conclusions These data demonstrate that corticosteroid treatment increases chromium losses and that steroid-induced diabetes can be reversed by chromium supplementation. Follow-up, double-blind studies are needed to con®rm these observations.

Diabet. Med. 16, 164±167 (1999)

Keywords chromium, corticosteroids, diabetes mellitis, glucose, , steroid- induced diabetes mellitis, trace elements

Abbreviations DM, diabetes mellitus; RfD, reference dose

Introduction observed throughout the world [3]. Dietary and other forms of stress increase chromium losses [4] which may exacerbate the Insuf®cient dietary chromium (Cr) leads to impaired glucose consequences of consumption of a low chromium diet. and lipid metabolism [1]. Dietary chromium intake is usually Thiazide diuretics that alter response to intravenous chro- 60% or less of the United States of America (USA) estimated mium have also been postulated to increase chromium losses safe and adequate daily dietary intake of 50 mg. Chromium but con®rmatory studies have not been completed [5]. intake of women in the USA eating normally was 28 6 1 per Corticosteroid treatment often leads to impaired glucose day and 33 6 1 mg for males [2]. Similar results have been tolerance and diabetes [6]. When steroid-induced diabetes continues for more than 1 week after the withdrawal of the steroid administration, it may be concluded that diabetes is Correspondence to: Dr Richard A. Anderson, USDA, ARS, BHNRC, NRFL, Bldg. 307, Rm. 224, BARC-East, Beltsville, MD 20705±2350, USA. established and needs on-going treatment [7]. Steroid diabetes e-mail: [email protected] is characterized by insulin resistance, usually in the absence of

164 ã 1999 British Diabetic Association. Diabetic Medicine, 16, 164±167 R Case reports 165 ketosis and acidosis. Similarly, the syndrome of marginal acute macular neuritis on her left eye and was given 60 mg of chromium intakes are characterized by insulin resistance in the prednisone daily. Within 2 days of treatment, her postprandial absence of ketosis and acidosis [1]. blood glucose increased to 15.5 mmol/l. Patient was given The purpose of this research was to determine if corticos- 200 mg of chromium as chromium picolinate three times daily teroid treatment causes increased chromium losses and if with meals. The Glibenclamide was maintained at 15 mg supplemental chromium counteracts the negative effects of daily. Within 1 week her postprandial blood sugar level was corticosteroid-induced diabetes. We measured 24-h urine 7.6 mmol/l and supplemental chromium was reduced to chromium in 13 patients before and after starting steroid 200 mg daily. The Glibenclamide dosage was gradually therapy and describe three case studies which test the theory reduced over the next few days and her diabetes was controlled that chromium supplementation counteracts steroid-induced with 5 mg of Glibenclamide and 200 mg of chromium diabetes. picolinate daily. The prednisone treatment was stopped 4 months later. Patients and methods Our second patient was a 44-year-old man who developed terminal renal failure due to chronic glomerulonephritis Thirteen patients scheduled to be treated with corticosteroids following a kidney transplant. After his ®rst operation, he collected 24-h urine samples the day prior to treatment and was given prednisone (60 mg/d) as part of the immuno- over the following 3 days. The subjects were instructed on suppressive therapy. His blood glucose started to rise and proper methods to collect urine samples, to avoid chromium immunosuppression was stopped within a week because the contamination. All patients were informed of the purposes of transplanted kidney was rejected. He continued haemo- the study and gave signed informed consent to participate in dialysis, and blood glucose levels became normal. The second the study. The study was approved by the review board of the renal transplantation was performed 35 months later. His Diabetes Department of the Linn Clinic, Haifa, Israel. blood glucose started to rise as soon as he started immuno- suppressive therapy (which included 10 mg of prednisone daily). Insulin was given in increasing doses with the Chromium analyses expectation that it would be decreased with the reduction of Chromium determinations were performed using a computer- the corticosteroid treatment. He was released from the controlled atomic absorption spectrometer (Perkin- hospital 10 days after the operation on 10 mg of prednisone Elmer, Model HGA 500, Norwalk, CT) as described [2]. daily. He was readmitted to the Diabetes Clinic with blood An in-house control urine sample, whose chromium concen- glucose levels of 19.4 6 2.8 mmol/l, in spite of a daily dosage of tration had been veri®ed by two independent methods, was 100 units of mixed (twice daily). He was then given assayed at least twice daily as an internal check on the accuracy 200 mg of chromium as chromium picolinate three times daily of the results. A value of 0.20 6 0.02 mg/l was obtained by with meals. Daily insulin dosage was reduced to 40 units of graphite furnace atomic absorption and 0.21 6 0.04 mg/l by intermediate acting insulin. Within 24 h his fasting blood gas-chromatography mass spectroscopy [8]. Differences in the glucose was below 8.3 mmol/l and postprandial below urinary chromium concentrations of the control and those 10 mmol/l. The daily supplemental chromium was gradually following corticosteroid treatment were analysed using paired reduced to 200 mg per day and the insulin dosage to 32 units Student's t-tests. daily. He continued to take 10 mg of prednisone daily and within 4 months blood sugar levels started to rise. Supple- Results mental chromium was increased to 400 mg per day. Fasting and postprandial blood sugar levels returned to below Urinary chromium losses of 13 patients increased following 8.3 mmol/l and 10 mmol/l, respectively, following the in- initiation of corticosteroid treatment from 155 6 28 to creased level of supplemental chromium. 244 6 33 ng/d in the ®rst 3 days (P < 0.01). Chromium losses Our third patient was an obese 54-year-old female with for all except two patients increased following prednisone Type 2 DM for 1 year. Her condition was well controlled treatment (Fig. 1). with metformin, 850 mg at bedtime (09.00 pm) and 200 mg of chromium picolinate. She was hospitalized because of right parietal glioblastoma with a severe Case reports peri-tumoral oedema. She was treated with 24 mg of Following this observation, a pilot study involving supple- daily. Following 200 mg of chromium mentation of three patients was completed. The ®rst patient three times daily, there was a symptomatic improvement was a 62-year-old woman with mellitus (DM) and her cerebral oedema decreased. There was no rise in for 12 years. She was on constant haemodialysis for end-stage blood glucose despite the high dosage of dexamethasone. diabetic nephropathy. Her blood glucose levels were well Craniotomy was performed 3 weeks later to remove the controlled with 15 mg of Glibenclamide daily. Her right eye tumour. She survived the operation but died of a was blind because of chronic keratitis, and she developed an pulmonary embolism 10 days later.

ã 1999 British Diabetic Association. Diabetic Medicine, 16, 164±167 L 166 Chromium alleviates steroid DM · A Ravina et al.

Figure 1 Urinary chromium excretion of 13 patients before and the ®rst 3 days following corticosteroid treatment. Patients served as their own controls. There were signi®cant increases in chromium losses due to corticosteroid treatment (P < 0.01).

Discussion four days of supplemental chromium, 250 mg per day, as chromic chloride, fractional glucose clearance was normalized Deterioration of glucose tolerance following corticosteroid and the patient markedly improved. administration or steroid diabetes may occur in patients There are several double-blind, placebo-controlled studies initially free of diabetes or in patients with established DM demonstrating bene®cial effects of supplemental chromium on [6,7]. Our observations, and those published in medical people with impaired glucose tolerance and diabetes (re- textbooks [7], indicate that steroid-induced diabetes occurs de viewed in reference 1). We recently completed a double-blind, novo, treatment is required and that there is rarely a placebo controlled study involving 180 people with Type spontaneous recovery. 2 DM. One-third of the subjects received placebo, one-third Stresses that alter chromium status are glucose loading, 200 mg of chromium as chromium picolinate per day and one- diets high in simple sugars, pregnancy, lactation, infec- third 1000 mg of chromium as chromium picolinate per day. tion, acute and chronic exercise and physical trauma [4]. Supplemental chromium led to improvements in fasting and

Urinary chromium losses can be used as a measure of postprandial glucose, insulin and haemoglobin A1c in > 90% mobilized chromium since mobilized chromium it is not of the patients, with larger improvements in subjects receiving reabsorbed by the kidney but is excreted in the urine [4]. the higher dose of chromium [12]. Chromium supplementa- The degree of stress is roughly proportional to the tion (200 mg/d) to 114 patients with Type 2 DM and 48 amount of chromium lost in the urine. For example, patients with Type 1 DM also led to a decrease in the insulin, mild aerobic exercise at 50% of maximal oxygen sulphonylurea or metformin requirements in > 70% of the consumption does not signi®cantly increase chromium patients [13]. loss in the urine while exercise at 90% of maximal Following the demonstration that chromium losses in- oxygen consumption does. There is a direct correlation creased after initiation of corticosteroid treatment and that between the increases in cortisol in response to stress and glucose metabolism improved in our three patients given the amount of chromium lost [10]. In the present study, supplemental chromium, we have administered chromium to urinary chromium losses were increased following corti- 41 patients with steroid-induced diabetes. Steroid-induced costeroid administration. Since chromium intake is nor- diabetes in 38 of 41 patients was ameliorated following mally marginal, this may lead to marginal chromium supplementation of 200 mg of chromium three times per day, de®ciency [1±5]. despite a 50% decrease in the amount of medication Associations of chromium de®ciency include not only administered (unpublished data). A daily maintenance dose impaired glucose and lipid metabolism but in extreme cases of 200 mg was suf®cient to maintain blood glucose in the impaired nerve function [1]. Glucose intolerance and neuro- normal range thereafter. Double-blind, placebo controlled pathy of a prednisone-treated patient on total parenteral studies are needed to con®rm these observations. nutrition improved following addition of supplemental chro- The estimated safe and adequate daily dietary intake range mium to the total parenteral ¯uids [11], although the patient for chromium is 50±200 mg [14]. There have been several was on other medications including metronidazole. Within chromium supplementation studies using 1000 mg/d or more

ã 1999 British Diabetic Association. Diabetic Medicine, 16, 164±167 R Case reports 167 and there have not been any documented cases of any adverse 3 Cauwenbergh RV, Hendrix P, Robberecht H, Deelstra HA. Daily effects. Trivalent chromium, the form found in foods and dietary chromium intake in Belgium, using duplicate portion nutrient supplements, is considered one of the least toxic sampling. Z Lebenson Unters Forsch 1996; 203: 203±206. 4 Anderson RA. Stress effects on chromium nutrition of humans and nutrients. The reference dose (RfD) established by the United farm animals. In: Biotechnology in the Feed Industry, Proceedings of States Environmental Protection Agency for chromium is 350 Alltech's Tenth Symposium. Nottingham, UK: University Press, times the upper limit of the estimated safe and adequate daily 1994, 267±274. dietary intake (ESADDI) of 200 mg per day. The RfD is de®ned 5 Mossop RT. Diabetogenic effect of thiazides and the relation to as `an estimate (with uncertainty spanning perhaps an order of chromium. A preliminary report. Cent Afr J Med 1985; 7: 129±131. magnitude) of a daily exposure to the human population, 6 Merck Manual, 14th edn. Rahway, NJ: Merck Sharp and Dohme including sensitive subgroups, that is likely to be without an Research Laboratories, 1982, 2385. 7 Taylor R. In: JC Pickup and G Williams, eds. Textbook of Diabetes. appreciable risk of deleterious effects over a lifetime' [15]. This Oxford, UK: Blackwell Science Publications, 1991, 803. conservative estimate of safe intake has a much larger safety 8 Veillon C, Wolf WR, Guthrie BE. Determination of chromium in factor for trivalent chromium than almost any other nutrient. biological materials by stable isotope dilution. Anal Chem 1979; 51: Anderson et al. [16] demonstrated a lack of toxicity of 1022±1024. chromium chloride and chromium picolinate in rats at levels 9 Ravina A, Minuhin O. Bed-time administration of Metformin-HC1 several thousand times the upper limit of the estimated safe (glucophage retard) may reduce insulin requirement (Abstract). and adequate daily dietary intake for humans (based on body MGSD Meeting. November 1990, Nice, France, p. 38. 10 Anderson RA, Bryden NA, Polansky MM, Thorp JW. Effect of weight). carbohydrate loading and underwater exercise on circulating In summary, our data show that corticosteroid treatment cortisol, insulin and urinary losses of chromium and zinc. Eur J increases urinary chromium losses. Our case reports show a Appl Physiol 1991; 63: 146±150. marked bene®t of chromium supplementation of patients with 11 Verhage AH, Cheong WK, Jeejeebhoy KN. Neurologic symptoms steroid-induced diabetes. Double-blind, placebo controlled due to possible chromium de®ciency in long-term parenteral studies are needed to con®rm these observations. nutrition that closely mimic metronidazole±induced syndrome. J Parent Enteral Nutr 1996; 20: 123±127. 12 Anderson RA, Cheng N, Bryden NA, Polansky MM, Chi J, Feng J. Acknowledgements Elevated intakes of supplemental chromium improve glucose and insulin variables of people with type II diabetes. Diabetes 1997; 46: This work was funded in part by a grant from the Diabetes 1786±1791. Action Foundation. 13 Ravina A, Slezak L, Rubal A, Mirsky N. Clinical use of the trace element chromium (III) in the treatment of diabetes mellitus. J Trace Elem Exptl Med 1995; 8: 183±190. References 14 National Research Council. Recommended Dietary Allowances, 1 Anderson RA. Recent advances in the clinical and biochemical 10th edn. Washington, DC: National Academy Press, 1989. manifestation of chromium de®ciency in human and animal 15 Mertz W, Abernathy CO, Olin SS. Risk Assessment of Essential nutrition. J Trace Elem Exptl Med 1998; 11: 241±250. Elements. Washington, DC: ILSI Press, 1994, xix±xxviii. 2 Anderson RA, Kozlovsky AS. Chromium intake, absorption and 16 Anderson RA, Bryden NA, Polansky MM. Lack of toxicity of excretion of subjects consuming self-selected diets. Am J Clin Nutr chromium chloride and chromium picolinate in rats. J Am Coll Nutr 1985; 41: 1177±1183. 1997; 16: 273±279.

ã 1999 British Diabetic Association. Diabetic Medicine, 16, 164±167