Letters

(2); the latter was the main bladder abnor- Objective Evidence continued, and her weight had returned mality in our patient, presumably reflect- to baseline. The aberrations in nerve func- ing the brainstem atrophy. Previously, for the Reversibility tion had normalized; most dramatic was bowel dysfunction has only rarely been of Nerve Injury in the improvement in nerve axon function, documented in WFS (3,4). In our case, Diabetic Neuropathic represented by a doubling of amplitude loss of SPRCs and sphincter weakness potentials. For example, in the sural nerve were the main bowel abnormalities. The the conduction velocity improved by SPRCs are likely to reflect the intrinsic 30% (from 36 to 47 m/s) and the sensory neuronal activities of the pacemaker cells amplitude potentials doubled (from 3.4 iabetic neuropathic cachexia is an in the myenteric plexus, which can be to 7.5 ␮V) from baseline. Similar changes acute complication of diabetes damaged in peripheral neuropathies that were seen in the median and peroneal marked by such extreme pain and involve small fibers. Sphincter tone is D nerves for these parameters and also for that, although exceptionally maintained by the extrinsic somatic nerve F-wave latencies and vibration perception for the external sphincter and sympa- rare, it imparts major challenges in man- thresholds. thetic nerve for the internal sphincter, re- agement and diagnosis. Little is known of The profound weight loss, the sym- spectively, which can also be damaged in the fundamental pathophysiologic fea- metrical sensorimotor polyneuropathy peripheral neuropathies. These bowel tures of the peripheral nervous system in associated with dramatic painful pares- dysfunctions need specific management this condition; for example, the symp- thesias devoid of weakness, the temporal to maximize the quality of life in patients tomatic resolution that is classically ob- relation with insulin therapy, and the with WFS. served may arise from either complete chronic course are in complete accor- destruction of pain-transmitting nerve fi- 1 dance with the diagnosis of diabetic neu- ZHI LIU, MD bers or from their repair. To reconcile this 1 ropathic cachexia (1). This report RYUJI SAKAKIBARA, MD issue, we report the first case to our 1 emphasizes the need for vigilant symp- TOMOYUKI UCHIYAMA, MD knowledge that the nerve dysfunction is 1 tomatic therapy of diabetic neuropathic TATSUYA YAMAMOTO, MD reversible. A 36-year-old woman pre- 1 cachexia while expediting investigations TAKASHI ITO, MD 1 sented with subacute hyperglycemic for eliminating alternate causes. Unique SHOICHI ITO, MD 2 symptoms. Soon after initiation of insulin to this report, however, is the objective YUSUKE AWA, MD 3 therapy and the decline of HbA1c from finding of nerve dysfunction that was dra- TAKEO ODAKA, MD 3 14.9 to 5.5%, she developed severe lanci- matically reversed after symptomatic re- TAKETO YAMAGUCHI, MD 1 nating pain and profound weight loss as- covery; it supports the proposed TAKAMICHI HATTORI, MD sociated with , amenorrhea, hypothesis that paradoxical hypoxic in- From the 1Department of Neurology, Chiba Univer- insomnia, and dehydration. On examina- jury occurs at the initiation of insulin sity, Chiba, Japan; the 2Department of Urology, tion, allodynia was so pronounced that a therapy (2,3). This case fundamentally 3 Chiba University, Chiba, Japan; and the Depart- light touch to her shoulder would cause suggests that the reversal of painful symp- ment of Gastroenterology, Chiba University, Chiba, her to weep. Profound loss of subcutane- Japan. toms is associated with repair of func- Address correspondence to Ryuji Sakakibara, ous adipose tissue and loss of muscle bulk tional aberrations in nerve fibers induced MD, Neurology Department, Chiba University, was evident, such that her weight had de- by hypoxia rather than their irreversible 1-8-1 Inohana Chuo-ku, Chiba 260-8670, Japan. creased from a baseline of 58.3 to 41.8 kg ischemic destruction. E-mail: [email protected]. (corresponding to a decrease in BMI from © 2006 by the American Diabetes Association. 2 1 21 to 15.7 kg/m ). Pain, temperature, and JASPREET GREWAL, MD 1 ●●●●●●●●●●●●●●●●●●●●●●● light touch sensation were abnormal in VERA BRIL, MD 2 References the hands and feet. GARY LEWIS, MD 2 1. Sakakibara R, Odaka T, Uchiyama T, Asa- Blood count, chemistries, and thyroid BRUCE A. PERKINS, MD, MPH hina M, Yamaguchi K, Yamaguchi T, Ya- and cortisol levels were normal. Further manishi T, Hattori T: Colonic transit time, tests for malignancy and malabsorption, From the 1Division of Neurology, Department of sphincter EMG and rectoanal videoma- including serum immunoelectrophoresis, Medicine, University of Toronto, Toronto, Canada; and the 2Division of Endocrinology and Metabo- nometry in multiple system atrophy. Mov computed tomography, bone scan, and Disord 19:924–929, 2004 lism, Department of Medicine, University of To- endoscopy with biopsies, were normal. ronto, Toronto, Canada. 2. Simsek E, Simsek T, Tekgul S, Hosal S, Address correspondence to Bruce A. Perkins, Seyrantepe V, Aktan G: Wolfram (DID- Titers for antinuclear antibodies and viral etiologies were negative. Nerve conduc- MD, MPH, FRCP(C), Endocrinology and Metabo- MOAD) syndrome: a multidisciplinary lism, Assistant Professor, University of Toronto, clinical study in nine Turkish patients and tion studies demonstrated impaired con- Staff Physician, University Health Network, Toronto review of the literature. Acta Paediatrica duction velocity (indicative of impaired General Hospital, 200 Elizabeth St., Room EN-12- 92:55–61, 2003 myelin sheath function) and impaired 217, Toronto, Ontario, Canada M5G 2C4. E-mail: 3. Barrett TG, Bundey SE, Macleod AF: Neu- amplitude potentials (indicative of im- [email protected]. rodegeneration and diabetes: UK nation- paired nerve axon function) of both sen- © 2006 by the American Diabetes Association. wide study of Wolfram (DIDMOAD) sory and motor nerves. ●●●●●●●●●●●●●●●●●●●●●●● syndrome. Lancet 346:1458–1463, 1995 Hydration, oral nutritional support, 4. Medlej R, Wasson J, Baz P, Azar S, Salti I, References Loiselet J, Permutt A, Halaby G: Diabetes and opiate therapy were provided during 1. Weintrob N, Josefberg Z, Galazer A, Vardi mellitus and optic atrophy: a study of an 8-week hospitalization. She subse- P, Karp M: Acute painful neuropathic ca- Wolfram syndrome in the Lebanese pop- quently received symptomatic therapy chexia in a young type I diabetic woman: ulation. J Clin Endocrinol Metab 89:1656– with amitriptyline and gabapentin; a year a case report. Diabetes Care 20:290–291, 1661, 2004 later these analgesic therapies were dis- 1997

DIABETES CARE, VOLUME 29, NUMBER 2, FEBRUARY 2006 473 Letters

2. Yuen KC, Day JL, Flannagan DW, Ray- may not have been determined correctly, Venter C, Xiaomei Wu: Determination of man G: Diabetic neuropathic cachexia the composition or manufacturing proce- the glycaemic index of foods: interlabora- and acute bilateral cataract formation fol- dures of individual products may change tory study. Eur J Clin Nutr 57:475–482, lowing rapid glycaemic control in a newly over time, and shelf life and preparatory 2003 diagnosed type 1 diabetic patient. Diabet methods may also affect glycemic index. 3. Brouns F, Bjorck I, Frayn KN, Gibbs AL, Med 18:854–857, 2001 Lang V, Slama G, Wolever TMS: Glycae- 3. Kihara M, Zollman PJ, Smithson IL, La- Such concerns are not unique to studies mic index methodology. Nutr Res Rev 18: gerlund TD, Low PA: Hypoxic effect of of glycemic index. One cannot assume, 145–171, 2005 exogenous insulin on normal and diabetic for example, that a published value for 4. Velangi A, Fernandes G, Wolever TM: peripheral nerve. Am J Physiol 266:E980– vitamin C content of Valencia orange will Evaluation of a glucose meter for deter- E985, 1994 apply to every piece of fruit, at all times of mining the glycemic response of foods. year, from any location. Clin Chim Acta 356:191–198, 2005 Major categories of food differ in gly- 5. Wolever TM, Bolognesi C: Prediction of COMMENTS AND cemic index with reasonable consistency; glucose and insulin responses of normal most fruits, legumes, minimally pro- subjects after consuming mixed meals RESPONSES cessed grain products, and pasta prepared varying in energy, protein, fat, carbohy- from hard wheat have low– to moderate– drate and glycemic index. J Nutr 126: 2807–2812, 1996 glycemic index, whereas highly processed 6. Chew I, Brand JC, Thorburn AW, Trus- Influence of Glycemic grains products and pasta previously pre- well AS: Application of glycemic index to Index/Load on pared and canned have a high–glycemic mixed meals. Am J Clin Nutr 47:53–56, 1988 index. Most of the foods used by Alfenas 7. Bornet FR, Costagliola D, Rizkalla SW, Glycemic Response, and Mattes for the low–glycemic index diet Blayo A, Fontvieille AM, Haardt MJ, Leta- , and Food included highly processed grain products noux M, Tchobroutsky G, Slama G: Insu- (quick pizza, quiche, pita, bagel, etc.). linemic and glycemic indexes of six Intake in Healthy There are many studies demonstrat- starch-rich foods taken alone and in Humans ing that the glycemic index of individual mixed meal by type 2 diabetics. Am J Clin foods predicts a response to mixed meals Nutr 45:588–595, 1987 8. Ludwig DS, Majzoub JA, Al-Zahrani A, Response to Alfenas and Mattes when appropriate methodology is uti- Dallal G, Blanco I, Roberts SB: High gly- lized (5–7). With regard to the authors’ cemic index foods, overeating, and obe- description of our study, two of the test sity. Pediatrics 103: e26, 1999 n their recent article, Alfenas and meals did have identical macronutrient Mattes (1) conclude that the glycemic composition and solid food components, I index values of individual foods do not and the measured glycemic response cor- Influence of Glycemic predict glycemic response to mixed responded closely with prediction (8). meals, nor influence measures of hunger. Clearly, research into the relationship Index/Load on Because the observed glycemic response between glycemic index and glycemic re- Glycemic Response, did not differ between diets, the lack of sponse merits study. To advance the dia- Appetite, and Food effect on appetite is not surprising. Thus, logue, adequately powered studies the potentially important aspect of the employing accepted methodology will be Intake in Healthy study pertains to the prediction of glyce- needed. A more fundamental question is Humans mic index in mixed meals. whether diets comprised of low–glycemic The authors’ approach was to validate index foods improve important clinical Response to Alfenas and Mattes published glycemic index values in a pre- end points related to , diabetes, test, selecting 48 of 79 foods with consis- heart disease, and cancer. tent glycemic responses. However, their ecently, Alfenas and Mattes (1) con- methods do not conform to standard pro- DAVID S. LUDWIG, MD, PHD cluded that the differential glyce- cedures (2–4). Only 3 subjects were used SUSAN B. ROBERTS, PHD R mic responses of foods tested in for each food instead of the recommended isolation are not preserved under condi- minimum number of 10 (3). Blood glu- From Children’s Hospital Boston and Tufts Univer- tions of chronic ad libitum consumption sity, Boston, Massachusetts. cose was measured by glucometer, a de- Address correspondence to David S. Ludwig, of mixed meals (1). This conclusion is un- vice that is not sufficiently accurate in the MD, PhD, Division of Endocrinology, 300 Long- warranted because of serious method- normal range for research purposes (4). wood Ave., Boston, MA 02115. E-mail: ological problems that undermine the With such a small subject number, CIs [email protected]. validity of their results. around the mean would likely overlap for © 2006 by the American Diabetes Association. Foods were classified as low– or high–glycemic index by the investigators; most foods on both diets. From a statisti- ●●●●●●●●●●●●●●●●●●●●●●● cal perspective, the selection of foods with the glycemic index of each food was de- an underpowered pretest using inaccu- References termined in three subjects by measuring rate methods would produce regression 1. Alfenas RCG, Mattes RD: Influence of gly- glucose four times with a glucose meter cemic inex/load on glycemic response, to the mean. appetite, and food intake in healthy hu- and discarding means with inconsistent It is important to emphasize that pub- mans. Diabetes Care 28:2123–2129, 2005 values. Since white bread was used as the lished values for specific foods cannot be 2. Wolever TM, Vorster HH, Bjorck I, reference, all glycemic index values dis- used for a study such as this without care- Brand-Miller J, Brighenti F, Mann JI, Ram- cussed here are adjusted accordingly. We ful validation because published values dath DD, Granfeldt Y, Holt S, Perry TL, commend the authors for wanting to mea-

474 DIABETES CARE, VOLUME 29, NUMBER 2, FEBRUARY 2006