Insulin Misuse by Women with Type 1 Diabetes Mellitus Complicated by Eating Disorders does not Favorably Change Body Weight, Body Composition, or Body Fat Distribution
1998; Elsevier BV; Volume: 98; Issue: 6 Linguagem: Inglês
10.1016/s0002-8223(98)00155-2
ISSN1878-3570
AutoresSandra G. Affenito, Nancy R. Rodriguez, Jeffrey R. Backstrand, Garry Welch, Cynthia H. Adams,
Tópico(s)Obesity and Health Practices
ResumoReports indicate that adolescent and young adult females with type 1 diabetes are at high risk for development of the classic eating disorders anorexia nervosa and bulimia nervosa ((1)Rodin G.M. Daneman D. Eating disorders and IDDM.Diabetes Care. 1992; 10: 1402-1412Google Scholar, (2)Steel J.M. Lloyd G.G. Young R.J. MacIntyre C.C.A. Changes in eating attitudes during the first year of treatment for diabetes.J Psychosom Res. 1990; 34: 313-318Google Scholar, (3)Hudson J.I. Wentworth S.M. Hudson M.S. Pope H.G. Prevalence of anorexia nervosa and bulimia among young diabetic women.J Clin Psychol. 1985; 46: 88-89Google Scholar). Studies suggest that 15% to 39% of women with type 1 diabetes have omitted or reduced insulin doses in an effort to control weight ((4)Polonsky W.P. Anderson B.J. Aponte J.A. Lohrer P.A. Jacobson A.M. Cole C.F. Insulin omission in females with IDDM.Diabetes Care. 1994; 17: 1178-1185Google Scholar, (5)Biggs M.M. Basco M.R. Patterson G. Raskin P. Insulin withholding for weight control in women with diabetes.Diabetes Care. 1994; 17: 1186-1189Google Scholar, (6)Peveler R.C. Fairburn C.G. Boller I. Eating disorders in adolescents with IDDM a controlled study.Diabetes Care. 1992; 15: 1356-1360Google Scholar, (7)Fairburn C.G. Peveler R.C. Davies B. Mann J.I. Mayou R.A. Eating disorders in young adults with insulin-dependent diabetes mellitus a controlled study.BMJ. 1991; 303: 17-20Google Scholar). As a result, the 4th edition of the Diagnostic and Statistical Manual of Mental Disorders has incorporated the misuse of insulin (medications) as one of the diagnostic criterion for bulimia nervosa ((8)Diagnostic and Statistical Manual of Mental Disorders. 4th ed. American Psychiatric Association, Washington, DC1994Google Scholar). Improper and inconsistent use of insulin notably affects metabolic control and may result in altered body weight, body composition, and fluid balance ((9)Man S.O. Water, electrolyte and acid-base balance.in: Shils M.E. Olson J.A. Shike M. Modern Nutrition in Health and Disease. 8th ed. Lea & Febiger, Philadelphia, Pa1994: 112-143Google Scholar).This study characterizes the relationship between insulin misuse and eating disorders with regard to body weight, body composition, body fat distribution, and other variables associated with body composition analyses in women with type 1 diabetes.MethodsSubjectsAfter this study was approved by the Institutional Review Board for the Use of Human Subjects by the University of Connecticut and all participating clinics, nonpregnant, nonlactating women (18 to 46 years old) who had type 1 diabetes for at least 1 year, and who were otherwise healthy, were recruited (n=90). Written informed consent was obtained from all subjects.Assessment of Eating Disordered Behaviors and Insulin MisuseDiagnosis of eating disorders was based on criteria of the 3rd edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R) ((10)Diagnostic and Statistical Manual of Mental Disorders. 3rd ed. American Psychiatric Association, Washington, DC1987Google Scholar) and confirmed by clinical interview using the validated Eating Disorder Examination, which assesses eating habits and attitudes toward shape and weight ((11)Cooper Z. Fairburn C.G. The Eating Disorder Examination a semi-structured interview for the assessment of the specific psychopathology of eating disorders.Int J Eating Disord. 1987; 6: 1-8Google Scholar). Subjects were separated into 1 of 3 groups based on eating disorder severity: clinical (n=14:4 women with anorexia nervosa and 10 with bulimia nervosa), subclinical (n=13), and control (n=63). Subjects were classified as clinical if all DSM-III-R criteria were met, and as subclinical if the criteria were partially fulfilled. Control subjects were free of eating disorders. The Bulimia Test Revised was also administered to each subject to assess the severity and frequency of bulimic behaviors ((12)Brelford T.N. Hummel R.M. Barrios B.A. The bulimia test revised a psychometric investigation.Psychol Asses. 1992; 4: 399-401Google Scholar). Attitudes and behaviors regarding misuse of insulin were determined by clinical interview according to the procedure of Polonsky et al ((4)Polonsky W.P. Anderson B.J. Aponte J.A. Lohrer P.A. Jacobson A.M. Cole C.F. Insulin omission in females with IDDM.Diabetes Care. 1994; 17: 1178-1185Google Scholar). Insulin misuse was defined as the intentional reduction in insulin dosage or omission of insulin injection throughout the past 30 days for weight-control purposes.Anthropometric Measurements and Body Composition AnalysesWeight and height were measured and body mass index (BMI) was calculated as kg/m2. Skinfold measurements were taken at 3 sites: triceps, suprailiac crest, and subscapular areas. Circumference measurements were obtained at the mid-upper arm area, waist, and hips for determination of mid-arm muscle circumference and waist-to-hip ratio (WHR), respectively. Bioelectrical impedance analysis (Model BIA, 101, RJL systems, Detroit, Mich) was also performed ((13)Kushner R.F. Bioelectrical impedance analysis a review of principles and applications.J Am Coll Nutr. 1992; 11: 199-209Google Scholar).Biochemical AssessmentBlood was collected from a large antecubital vein into plastic tubes containing ethylenediaminetetraacetic acid. Whole blood was analyzed for glycated hemoglobin (HbAlc) by affinity chromatography ((14)Fiechtner M. Ramp J. England B. Knudson M.A. Little R.R. England J.D. Goldstein D.E. Wynn A. Affinity binding assays of glycohemoglobin by two-dimensional centrifugation referenced to hemoglobin A1C.Clin Chem. 1992; 38: 2372-2379Google Scholar). Serum was analyzed to determine glucose concentration and establish a health profile using a standard chemistry-26 panel.Statistical AnalysesData were analyzed using the Statistical Analysis System (version 6.07, 1993, SAS Institute, Cary, NC). One-way analysis of variance was used to compare groups based on DSM-III-R categorization. Group differences based on insulin misuse were determined by Student's t tests. Differences on dichotomous variables were determined by χ2 procedure. No differences were found when groups were analyzed by DSM-III-R categorization or insulin misuse; therefore, results are presented for women with type 1 diabetes and eating disorders vs women with type 1 diabetes only. Analyses did not include 4 subjects with anorexia nervosa because by diagnosis, anthropometric and body composition measurements would be significantly decreased.Results and discussionWeight-controlling behaviors used by subjects with either a subclinical or clinical eating disorder (n=27) were as follows: laxative abuse (1 of 27, 3%), diuretic misuse (3 of 27, 11%), self-induced vomiting (5 of 27, 18%), dietary restraint (10 of 27, 37%), exercise (11 of 27, 41%), and insulin misuse (12 of 27, 44%). Women with diabetes who did not have eating disorders were more educated, held more professional occupations, and were more likely to be married compared with women with diabetes who had eating disorders (Table 1). The sample was predominantly white.Table 1Demographic, behavioral, and glycemic characteristics of women with type 1 diabetes with and without eating disordersVariableWomen with type 1 diabetes and eating disorders (n=23)Women with type 1 without eating disorders (n=63)P value←mean±standard deviation→Age (y)29.4±7.0aNS=no significance.29.5±6.0NSaNS=no significance.Duration diabetes (y)14.7±8.714.9±9.2NSGlucose level (mmol/L)bTo convert mmol/L glucose to mg/dL, multiply mmol/L by 18.0. To convert mg/dL glucose to mmol/L, multiply mg/dL by 0.0555. Glucose of 6.0mmol/L=108mg/dL.10.7±3.58.2±3.0<.05HbA1c (%)cHbA1c=glycated homoglobin; %=average blood glucose concentration over past 120 days: reference range=4.0% to 6.0% without diabetes; >8% indicates poor control (20).9.9±1.88.3±1.6<.0002Total dose insulin prescribed (units/d)47.6±21.5dUnequal variances were calculated using the folded form of the F statistic (F′) (P<.05).41.7±14.6NSBULIT-ReBULIT-R=Bulimia Test Revised (range of scores=28–140; diagnosis of bulimia nervosa ≥98 cutoff point).84.1±27.143.6±27.1<.0001No. of daily injections2.5±0.782.8±0.73NSn%n%College education4172540NSMarital statusSingle18782540<.05Married4173759Divorced1.05Separated1.01BULIT-R ≥ 98fNumber and percent of subjects who were diagnosed with bulimia nervosa.156500Intentional insulin misusegIntentional insulin misuse: current=any omission or reduction of prescribed insulin dosage during the past 30 days for the purpose of weight control; past=any omission or reduction of prescribed insulin dosage in the subject's lifetime for the purpose of weight control.Current104400<.0001hFisher's exact test.Past18781118<.0001hFisher's exact test.a NS=no significance.b To convert mmol/L glucose to mg/dL, multiply mmol/L by 18.0. To convert mg/dL glucose to mmol/L, multiply mg/dL by 0.0555. Glucose of 6.0mmol/L=108mg/dL.c HbA1c=glycated homoglobin; %=average blood glucose concentration over past 120 days: reference range=4.0% to 6.0% without diabetes; >8% indicates poor control (20)American Diabetes Association.Clinical Practice Recommendations.Diabetes Care. 1997; 20 (14–S17)Google Scholar.d Unequal variances were calculated using the folded form of the F statistic (F′) (P<.05).e BULIT-R=Bulimia Test Revised (range of scores=28–140; diagnosis of bulimia nervosa ≥98 cutoff point).f Number and percent of subjects who were diagnosed with bulimia nervosa.g Intentional insulin misuse: current=any omission or reduction of prescribed insulin dosage during the past 30 days for the purpose of weight control; past=any omission or reduction of prescribed insulin dosage in the subject's lifetime for the purpose of weight control.h Fisher's exact test. Open table in a new tab Table 1 also illustrates the association of eating disorders with glycemic variables. Women with eating disorders misused insulin to a greater extent for weight-control purposes (P<.05). Mean insulin doses were similar between groups, but women with eating disorders had a greater range of insulin doses (P<.05).Anthropometric and body composition measures are given in Table 2. With the exception of WHR, no significant differences were found between groups for any of the variables. However in the group with eating disorders, significantly larger variances among women were noted in body weight, BMI, triceps skinfold measurement, mid-arm muscle circumference, and percentage total body water and intracellular water. Nearly a third of the women with eating disorders had triceps skinfold measurements below the 10th percentile.Table 2Anthropometric and body composition measurements in women with type 1 diabetes and eating disordersAnthropometric measurementWomen with type 1 diabetes and eating disorders (n=23)Women with type 1 without eating disorders (n=23)P value←mean±standard deviation→Height (cm)164.6±5.7163.9±5.5NSaNS=no significance.Weight (kg)70.4±19.3bUnequal variances were calculated using the folded form of the F statistic (F′) (P<.0001).65.6±10.0NSBody mass index26.0±7.1bUnequal variances were calculated using the folded form of the F statistic (F′) (P<.0001).24.3±3.4NSTriceps skinfold (mm)24.5±13.925.3±7.0NSn%n%≤ 10th percentilecPercentiles from Frisancho (21); Fisher's exact test.83436<.001≥ 90th percentilecPercentiles from Frisancho (21); Fisher's exact test.9391219<.10Suprailiac skinfold (mm)21.2±14.119.9±7.2NSSubscapular skinfold (mm)15.3±6.316.0±5.1NSMAMC (cm)dMAMC=mid-arm muscle circumference.20.3±3.320.5±3.0NSn%n%≤ 10th percentilecPercentiles from Frisancho (21); Fisher's exact test.10432337NS≥ 90th percentilecPercentiles from Frisancho (21); Fisher's exact test.3131118NSWaist-to-hip ratio0.78±0.080.74±0.06<.05Body fat (%)eEstimated from Durnin and Womersley (22).28.9±7.830.5±3.6NSBIAfBIA=bioelectrical impedance analysis.Body fat (%)29.8±11.830.4±10.7NSLean body mass (%)70.1±11.869.2±10.9NSTotal body water (%)52.5±8.9gUnequal variances were calculated using the folded form of the F statistic (F′) (P<.005).52.0±5.6NSIntracellular water (%)63.2±6.6gUnequal variances were calculated using the folded form of the F statistic (F′) (P<.005).61.8±3.9NSExtracellular water (%)36.8±6.537.6±5.6NSa NS=no significance.b Unequal variances were calculated using the folded form of the F statistic (F′) (P<.0001).c Percentiles from Frisancho (21)Frisancho A.R. New norms of upper limb fat and muscle areas for assessment of nutritional status.Am J Clin Nutr. 1981; 34: 2540-2545Google Scholar; Fisher's exact test.d MAMC=mid-arm muscle circumference.e Estimated from Durnin and Womersley (22)Durnin J.V.G.A. Wormersley J. Body fat assessed from total body density and its estimation from skinfold thickness measurements on 481 men and women aged 16 to 72 years.Br J Nutr. 1974; 32: 77-97Google Scholar.f BIA=bioelectrical impedance analysis.g Unequal variances were calculated using the folded form of the F statistic (F′) (P<.005). Open table in a new tab In our study, higher WHRs were associated with eating disorders. In contrast, others have reported lower WHRs for women who have eating disorders but do not have diabetes ((15)Radke-Sharpe N. Whitney-Saltiel D. Rodin J. Fat distribution as a risk factor for weight and eating concerns.Int J Eating Disord. 1990; 9: 927-936Google Scholar). Our findings may differ because higher WHRs are associated with increased rates of diabetes ((16)Hartz A.J. Rupley D.C. Rim A.A. The association of firth measurements with disease in 32,856 women.Am J Physiol. 1984; 119: 71-80Google Scholar).ApplicationsThis study characterized eating disordered behaviors in women with type 1 diabetes and confirmed the misuse of insulin as a method of purging. Nearly half of women with eating disorders reported current misuse of insulin, which was motivated by the desire to lose weight. On average, the women with eating disorders were more variable in weight, BMI, triceps and suprailiac skinfold measurements, and percentage total body water and intracellular water compared to women without eating disorders. This variability in the group with eating disorders may reflect achievement of weight loss via insulin misuse, insulin misuse combined with dietary restriction, or dietary restriction alone. Because insulin misuse results in poor metabolic control ((17)Affenito S.G. Backstrand J.R. Welch G.W. Lammi-Keefe C.J. Rodriguez N.R. Adams C.H. Subclinical and clinical eating disorders in insulin-dependent diabetes mellitus (IDDM) negatively affect metabolic control.Diabetes Care. 1997; 20: 182-184Google Scholar) and diabetes-related complications ((18)Rydall A.C. Rodin G.M. Olmsted M.P. Devenyl R.G. Daneman D. Disordered eating behavior and microvascular complications in young women with insulin-dependent diabetes mellitus.N Engl J Med. 1997; 336: 1849-1854Google Scholar), this behavior as a method of weight control should be identified in initial screening protocols ((19)Nutrition Practice Guidelines for Type 1 and Type 2 Diabetes Mellitus. American Dietetic Association, Chicago, Ill1998Google Scholar, (20)American Diabetes Association.Clinical Practice Recommendations.Diabetes Care. 1997; 20 (14–S17)Google Scholar). ■This report was supported, in part, by the University of Connecticut Research Foundation, Storrs Agricultural Experiment Station, National Institutes of Health General Clinical Research Grant (M01RR06192), University of Connecticut Health Center. Reports indicate that adolescent and young adult females with type 1 diabetes are at high risk for development of the classic eating disorders anorexia nervosa and bulimia nervosa ((1)Rodin G.M. Daneman D. Eating disorders and IDDM.Diabetes Care. 1992; 10: 1402-1412Google Scholar, (2)Steel J.M. Lloyd G.G. Young R.J. MacIntyre C.C.A. Changes in eating attitudes during the first year of treatment for diabetes.J Psychosom Res. 1990; 34: 313-318Google Scholar, (3)Hudson J.I. Wentworth S.M. Hudson M.S. Pope H.G. Prevalence of anorexia nervosa and bulimia among young diabetic women.J Clin Psychol. 1985; 46: 88-89Google Scholar). Studies suggest that 15% to 39% of women with type 1 diabetes have omitted or reduced insulin doses in an effort to control weight ((4)Polonsky W.P. Anderson B.J. Aponte J.A. Lohrer P.A. Jacobson A.M. Cole C.F. Insulin omission in females with IDDM.Diabetes Care. 1994; 17: 1178-1185Google Scholar, (5)Biggs M.M. Basco M.R. Patterson G. Raskin P. Insulin withholding for weight control in women with diabetes.Diabetes Care. 1994; 17: 1186-1189Google Scholar, (6)Peveler R.C. Fairburn C.G. Boller I. Eating disorders in adolescents with IDDM a controlled study.Diabetes Care. 1992; 15: 1356-1360Google Scholar, (7)Fairburn C.G. Peveler R.C. Davies B. Mann J.I. Mayou R.A. Eating disorders in young adults with insulin-dependent diabetes mellitus a controlled study.BMJ. 1991; 303: 17-20Google Scholar). As a result, the 4th edition of the Diagnostic and Statistical Manual of Mental Disorders has incorporated the misuse of insulin (medications) as one of the diagnostic criterion for bulimia nervosa ((8)Diagnostic and Statistical Manual of Mental Disorders. 4th ed. American Psychiatric Association, Washington, DC1994Google Scholar). Improper and inconsistent use of insulin notably affects metabolic control and may result in altered body weight, body composition, and fluid balance ((9)Man S.O. Water, electrolyte and acid-base balance.in: Shils M.E. Olson J.A. Shike M. Modern Nutrition in Health and Disease. 8th ed. Lea & Febiger, Philadelphia, Pa1994: 112-143Google Scholar). This study characterizes the relationship between insulin misuse and eating disorders with regard to body weight, body composition, body fat distribution, and other variables associated with body composition analyses in women with type 1 diabetes. MethodsSubjectsAfter this study was approved by the Institutional Review Board for the Use of Human Subjects by the University of Connecticut and all participating clinics, nonpregnant, nonlactating women (18 to 46 years old) who had type 1 diabetes for at least 1 year, and who were otherwise healthy, were recruited (n=90). Written informed consent was obtained from all subjects.Assessment of Eating Disordered Behaviors and Insulin MisuseDiagnosis of eating disorders was based on criteria of the 3rd edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R) ((10)Diagnostic and Statistical Manual of Mental Disorders. 3rd ed. American Psychiatric Association, Washington, DC1987Google Scholar) and confirmed by clinical interview using the validated Eating Disorder Examination, which assesses eating habits and attitudes toward shape and weight ((11)Cooper Z. Fairburn C.G. The Eating Disorder Examination a semi-structured interview for the assessment of the specific psychopathology of eating disorders.Int J Eating Disord. 1987; 6: 1-8Google Scholar). Subjects were separated into 1 of 3 groups based on eating disorder severity: clinical (n=14:4 women with anorexia nervosa and 10 with bulimia nervosa), subclinical (n=13), and control (n=63). Subjects were classified as clinical if all DSM-III-R criteria were met, and as subclinical if the criteria were partially fulfilled. Control subjects were free of eating disorders. The Bulimia Test Revised was also administered to each subject to assess the severity and frequency of bulimic behaviors ((12)Brelford T.N. Hummel R.M. Barrios B.A. The bulimia test revised a psychometric investigation.Psychol Asses. 1992; 4: 399-401Google Scholar). Attitudes and behaviors regarding misuse of insulin were determined by clinical interview according to the procedure of Polonsky et al ((4)Polonsky W.P. Anderson B.J. Aponte J.A. Lohrer P.A. Jacobson A.M. Cole C.F. Insulin omission in females with IDDM.Diabetes Care. 1994; 17: 1178-1185Google Scholar). Insulin misuse was defined as the intentional reduction in insulin dosage or omission of insulin injection throughout the past 30 days for weight-control purposes.Anthropometric Measurements and Body Composition AnalysesWeight and height were measured and body mass index (BMI) was calculated as kg/m2. Skinfold measurements were taken at 3 sites: triceps, suprailiac crest, and subscapular areas. Circumference measurements were obtained at the mid-upper arm area, waist, and hips for determination of mid-arm muscle circumference and waist-to-hip ratio (WHR), respectively. Bioelectrical impedance analysis (Model BIA, 101, RJL systems, Detroit, Mich) was also performed ((13)Kushner R.F. Bioelectrical impedance analysis a review of principles and applications.J Am Coll Nutr. 1992; 11: 199-209Google Scholar).Biochemical AssessmentBlood was collected from a large antecubital vein into plastic tubes containing ethylenediaminetetraacetic acid. Whole blood was analyzed for glycated hemoglobin (HbAlc) by affinity chromatography ((14)Fiechtner M. Ramp J. England B. Knudson M.A. Little R.R. England J.D. Goldstein D.E. Wynn A. Affinity binding assays of glycohemoglobin by two-dimensional centrifugation referenced to hemoglobin A1C.Clin Chem. 1992; 38: 2372-2379Google Scholar). Serum was analyzed to determine glucose concentration and establish a health profile using a standard chemistry-26 panel.Statistical AnalysesData were analyzed using the Statistical Analysis System (version 6.07, 1993, SAS Institute, Cary, NC). One-way analysis of variance was used to compare groups based on DSM-III-R categorization. Group differences based on insulin misuse were determined by Student's t tests. Differences on dichotomous variables were determined by χ2 procedure. No differences were found when groups were analyzed by DSM-III-R categorization or insulin misuse; therefore, results are presented for women with type 1 diabetes and eating disorders vs women with type 1 diabetes only. Analyses did not include 4 subjects with anorexia nervosa because by diagnosis, anthropometric and body composition measurements would be significantly decreased. SubjectsAfter this study was approved by the Institutional Review Board for the Use of Human Subjects by the University of Connecticut and all participating clinics, nonpregnant, nonlactating women (18 to 46 years old) who had type 1 diabetes for at least 1 year, and who were otherwise healthy, were recruited (n=90). Written informed consent was obtained from all subjects. After this study was approved by the Institutional Review Board for the Use of Human Subjects by the University of Connecticut and all participating clinics, nonpregnant, nonlactating women (18 to 46 years old) who had type 1 diabetes for at least 1 year, and who were otherwise healthy, were recruited (n=90). Written informed consent was obtained from all subjects. Assessment of Eating Disordered Behaviors and Insulin MisuseDiagnosis of eating disorders was based on criteria of the 3rd edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R) ((10)Diagnostic and Statistical Manual of Mental Disorders. 3rd ed. American Psychiatric Association, Washington, DC1987Google Scholar) and confirmed by clinical interview using the validated Eating Disorder Examination, which assesses eating habits and attitudes toward shape and weight ((11)Cooper Z. Fairburn C.G. The Eating Disorder Examination a semi-structured interview for the assessment of the specific psychopathology of eating disorders.Int J Eating Disord. 1987; 6: 1-8Google Scholar). Subjects were separated into 1 of 3 groups based on eating disorder severity: clinical (n=14:4 women with anorexia nervosa and 10 with bulimia nervosa), subclinical (n=13), and control (n=63). Subjects were classified as clinical if all DSM-III-R criteria were met, and as subclinical if the criteria were partially fulfilled. Control subjects were free of eating disorders. The Bulimia Test Revised was also administered to each subject to assess the severity and frequency of bulimic behaviors ((12)Brelford T.N. Hummel R.M. Barrios B.A. The bulimia test revised a psychometric investigation.Psychol Asses. 1992; 4: 399-401Google Scholar). Attitudes and behaviors regarding misuse of insulin were determined by clinical interview according to the procedure of Polonsky et al ((4)Polonsky W.P. Anderson B.J. Aponte J.A. Lohrer P.A. Jacobson A.M. Cole C.F. Insulin omission in females with IDDM.Diabetes Care. 1994; 17: 1178-1185Google Scholar). Insulin misuse was defined as the intentional reduction in insulin dosage or omission of insulin injection throughout the past 30 days for weight-control purposes. Diagnosis of eating disorders was based on criteria of the 3rd edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R) ((10)Diagnostic and Statistical Manual of Mental Disorders. 3rd ed. American Psychiatric Association, Washington, DC1987Google Scholar) and confirmed by clinical interview using the validated Eating Disorder Examination, which assesses eating habits and attitudes toward shape and weight ((11)Cooper Z. Fairburn C.G. The Eating Disorder Examination a semi-structured interview for the assessment of the specific psychopathology of eating disorders.Int J Eating Disord. 1987; 6: 1-8Google Scholar). Subjects were separated into 1 of 3 groups based on eating disorder severity: clinical (n=14:4 women with anorexia nervosa and 10 with bulimia nervosa), subclinical (n=13), and control (n=63). Subjects were classified as clinical if all DSM-III-R criteria were met, and as subclinical if the criteria were partially fulfilled. Control subjects were free of eating disorders. The Bulimia Test Revised was also administered to each subject to assess the severity and frequency of bulimic behaviors ((12)Brelford T.N. Hummel R.M. Barrios B.A. The bulimia test revised a psychometric investigation.Psychol Asses. 1992; 4: 399-401Google Scholar). Attitudes and behaviors regarding misuse of insulin were determined by clinical interview according to the procedure of Polonsky et al ((4)Polonsky W.P. Anderson B.J. Aponte J.A. Lohrer P.A. Jacobson A.M. Cole C.F. Insulin omission in females with IDDM.Diabetes Care. 1994; 17: 1178-1185Google Scholar). Insulin misuse was defined as the intentional reduction in insulin dosage or omission of insulin injection throughout the past 30 days for weight-control purposes. Anthropometric Measurements and Body Composition AnalysesWeight and height were measured and body mass index (BMI) was calculated as kg/m2. Skinfold measurements were taken at 3 sites: triceps, suprailiac crest, and subscapular areas. Circumference measurements were obtained at the mid-upper arm area, waist, and hips for determination of mid-arm muscle circumference and waist-to-hip ratio (WHR), respectively. Bioelectrical impedance analysis (Model BIA, 101, RJL systems, Detroit, Mich) was also performed ((13)Kushner R.F. Bioelectrical impedance analysis a review of principles and applications.J Am Coll Nutr. 1992; 11: 199-209Google Scholar). Weight and height were measured and body mass index (BMI) was calculated as kg/m2. Skinfold measurements were taken at 3 sites: triceps, suprailiac crest, and subscapular areas. Circumference measurements were obtained at the mid-upper arm area, waist, and hips for determination of mid-arm muscle circumference and waist-to-hip ratio (WHR), respectively. Bioelectrical impedance analysis (Model BIA, 101, RJL systems, Detroit, Mich) was also performed ((13)Kushner R.F. Bioelectrical impedance analysis a review of principles and applications.J Am Coll Nutr. 1992; 11: 199-209Google Scholar). Biochemical AssessmentBlood was collected from a large antecubital vein into plastic tubes containing ethylenediaminetetraacetic acid. Whole blood was analyzed for glycated hemoglobin (HbAlc) by affinity chromatography ((14)Fiechtner M. Ramp J. England B. Knudson M.A. Little R.R. England J.D. Goldstein D.E. Wynn A. Affinity binding assays of glycohemoglobin by two-dimensional centrifugation referenced to hemoglobin A1C.Clin Chem. 1992; 38: 2372-2379Google Scholar). Serum was analyzed to determine glucose concentration and establish a health profile using a standard chemistry-26 panel. Blood was collected from a large antecubital vein into plastic tubes containing ethylenediaminetetraacetic acid. Whole blood was analyzed for glycated hemoglobin (HbAlc) by affinity chromatography ((14)Fiechtner M. Ramp J. England B. Knudson M.A. Little R.R. England J.D. Goldstein D.E. Wynn A. Affinity binding assays of glycohemoglobin by two-dimensional centrifugation referenced to hemoglobin A1C.Clin Chem. 1992; 38: 2372-2379Google Scholar). Serum was analyzed to determine glucose concentration and establish a health profile using a standard chemistry-26 panel. Statistical AnalysesData were analyzed using the Statistical Analysis System (version 6.07, 1993, SAS Institute, Cary, NC). One-way analysis of variance was used to compare groups based on DSM-III-R categorization. Group differences based on insulin misuse were determined by Student's t tests. Differences on dichotomous variables were determined by χ2 procedure. No differences were found when groups were analyzed by DSM-III-R categorization or insulin misuse; therefore, results are presented for women with type 1 diabetes and eating disorders vs women with type 1 diabetes only. Analyses did not include 4 subjects with anorexia nervosa because by diagnosis, anthropometric and body composition measurements would be significantly decreased. Data were analyzed using the Statistical Analysis System (version 6.07, 1993, SAS Institute, Cary, NC). One-way analysis of variance was used to compare groups based on DSM-III-R categorization. Group differences based on insulin misuse were determined by Student's t tests. Differences on dichotomous variables were determined by χ2 procedure. No differences were found when groups were analyzed by DSM-III-R categorization or insulin misuse; therefore, results are presented for women with type 1 diabetes and eating disorders vs women with type 1 diabetes only. Analyses did not include 4 subjects with anorexia nervosa because by diagnosis, anthropometric and body composition measurements would be significantly decreased. Results and discussionWeight-controlling behaviors used by subjects with either a subclinical or clinical eating disorder (n=27) were as follows: laxative abuse (1 of 27, 3%), diuretic misuse (3 of 27, 11%), self-induced vomiting (5 of 27, 18%), dietary restraint (10 of 27, 37%), exercise (11 of 27, 41%), and insulin misuse (12 of 27, 44%). Women with diabetes who did not have eating disorders were more educated, held more professional occupations, and were more likely to be married compared with women with diabetes who had eating disorders (Table 1). The sample was predominantly white.Table 1Demographic, behavioral, and glycemic characteristics of women with type 1 diabetes with and without eating disordersVariableWomen with type 1 diabetes and eating disorders (n=23)Women with type 1 without eating disorders (n=63)P value←mean±standard deviation→Age (y)29.4±7.0aNS=no significance.29.5±6.0NSaNS=no significance.Duration diabetes (y)14.7±8.714.9±9.2NSGlucose level (mmol/L)bTo convert mmol/L glucose to mg/dL, multiply mmol/L by 18.0. To convert mg/dL glucose to mmol/L, multiply mg/dL by 0.0555. Glucose of 6.0mmol/L=108mg/dL.10.7±3.58.2±3.0<.05HbA1c (%)cHbA1c=glycated homoglobin; %=average blood glucose concentration over past 120 days: reference range=4.0% to 6.0% without diabetes; >8% indicates poor control (20).9.9±1.88.3±1.6<.0002Total dose insulin prescribed (units/d)47.6±21.5dUnequal variances were calculated using the folded form of the F statistic (F′) (P<.05).41.7±14.6NSBULIT-ReBULIT-R=Bulimia Test Revised (range of scores=28–140; diagnosis of bulimia nervosa ≥98 cutoff point).84.1±27.143.6±27.1<.0001No. of daily injections2.5±0.782.8±0.73NSn%n%College education4172540NSMarital statusSingle18782540<.05Married4173759Divorced1.05Separated1.01BULIT-R ≥ 98fNumber and percent of subjects who were diagnosed with bulimia nervosa.156500Intentional insulin misusegIntentional insulin misuse: current=any omission or reduction of prescribed insulin dosage during the past 30 days for the purpose of weight control; past=any omission or reduction of prescribed insulin dosage in the subject's lifetime for the purpose of weight control.Current104400<.0001hFisher's exact test.Past18781118<.0001hFisher's exact test.a NS=no significance.b To convert mmol/L glucose to mg/dL, multiply mmol/L by 18.0. To convert mg/dL glucose to mmol/L, multiply mg/dL by 0.0555. Glucose of 6.0mmol/L=108mg/dL.c HbA1c=glycated homoglobin; %=average blood glucose concentration over past 120 days: reference range=4.0% to 6.0% without diabetes; >8% indicates poor control (20)American Diabetes Association.Clinical Practice Recommendations.Diabetes Care. 1997; 20 (14–S17)Google Scholar.d Unequal variances were calculated using the folded form of the F statistic (F′) (P<.05).e BULIT-R=Bulimia Test Revised (range of scores=28–140; diagnosis of bulimia nervosa ≥98 cutoff point).f Number and percent of subjects who were diagnosed with bulimia nervosa.g Intentional insulin misuse: current=any omission or reduction of prescribed insulin dosage during the past 30 days for the purpose of weight control; past=any omission or reduction of prescribed insulin dosage in the subject's lifetime for the purpose of weight control.h Fisher's exact test. Open table in a new tab Table 1 also illustrates the association of eating disorders with glycemic variables. Women with eating disorders misused insulin to a greater extent for weight-control purposes (P<.05). Mean insulin doses were similar between groups, but women with eating disorders had a greater range of insulin doses (P<.05).Anthropometric and body composition measures are given in Table 2. With the exception of WHR, no significant differences were found between groups for any of the variables. However in the group with eating disorders, significantly larger variances among women were noted in body weight, BMI, triceps skinfold measurement, mid-arm muscle circumference, and percentage total body water and intracellular water. Nearly a third of the women with eating disorders had triceps skinfold measurements below the 10th percentile.Table 2Anthropometric and body composition measurements in women with type 1 diabetes and eating disordersAnthropometric measurementWomen with type 1 diabetes and eating disorders (n=23)Women with type 1 without eating disorders (n=23)P value←mean±standard deviation→Height (cm)164.6±5.7163.9±5.5NSaNS=no significance.Weight (kg)70.4±19.3bUnequal variances were calculated using the folded form of the F statistic (F′) (P<.0001).65.6±10.0NSBody mass index26.0±7.1bUnequal variances were calculated using the folded form of the F statistic (F′) (P<.0001).24.3±3.4NSTriceps skinfold (mm)24.5±13.925.3±7.0NSn%n%≤ 10th percentilecPercentiles from Frisancho (21); Fisher's exact test.83436<.001≥ 90th percentilecPercentiles from Frisancho (21); Fisher's exact test.9391219<.10Suprailiac skinfold (mm)21.2±14.119.9±7.2NSSubscapular skinfold (mm)15.3±6.316.0±5.1NSMAMC (cm)dMAMC=mid-arm muscle circumference.20.3±3.320.5±3.0NSn%n%≤ 10th percentilecPercentiles from Frisancho (21); Fisher's exact test.10432337NS≥ 90th percentilecPercentiles from Frisancho (21); Fisher's exact test.3131118NSWaist-to-hip ratio0.78±0.080.74±0.06<.05Body fat (%)eEstimated from Durnin and Womersley (22).28.9±7.830.5±3.6NSBIAfBIA=bioelectrical impedance analysis.Body fat (%)29.8±11.830.4±10.7NSLean body mass (%)70.1±11.869.2±10.9NSTotal body water (%)52.5±8.9gUnequal variances were calculated using the folded form of the F statistic (F′) (P<.005).52.0±5.6NSIntracellular water (%)63.2±6.6gUnequal variances were calculated using the folded form of the F statistic (F′) (P<.005).61.8±3.9NSExtracellular water (%)36.8±6.537.6±5.6NSa NS=no significance.b Unequal variances were calculated using the folded form of the F statistic (F′) (P<.0001).c Percentiles from Frisancho (21)Frisancho A.R. New norms of upper limb fat and muscle areas for assessment of nutritional status.Am J Clin Nutr. 1981; 34: 2540-2545Google Scholar; Fisher's exact test.d MAMC=mid-arm muscle circumference.e Estimated from Durnin and Womersley (22)Durnin J.V.G.A. Wormersley J. Body fat assessed from total body density and its estimation from skinfold thickness measurements on 481 men and women aged 16 to 72 years.Br J Nutr. 1974; 32: 77-97Google Scholar.f BIA=bioelectrical impedance analysis.g Unequal variances were calculated using the folded form of the F statistic (F′) (P<.005). Open table in a new tab In our study, higher WHRs were associated with eating disorders. In contrast, others have reported lower WHRs for women who have eating disorders but do not have diabetes ((15)Radke-Sharpe N. Whitney-Saltiel D. Rodin J. Fat distribution as a risk factor for weight and eating concerns.Int J Eating Disord. 1990; 9: 927-936Google Scholar). Our findings may differ because higher WHRs are associated with increased rates of diabetes ((16)Hartz A.J. Rupley D.C. Rim A.A. The association of firth measurements with disease in 32,856 women.Am J Physiol. 1984; 119: 71-80Google Scholar). Weight-controlling behaviors used by subjects with either a subclinical or clinical eating disorder (n=27) were as follows: laxative abuse (1 of 27, 3%), diuretic misuse (3 of 27, 11%), self-induced vomiting (5 of 27, 18%), dietary restraint (10 of 27, 37%), exercise (11 of 27, 41%), and insulin misuse (12 of 27, 44%). Women with diabetes who did not have eating disorders were more educated, held more professional occupations, and were more likely to be married compared with women with diabetes who had eating disorders (Table 1). The sample was predominantly white. Table 1 also illustrates the association of eating disorders with glycemic variables. Women with eating disorders misused insulin to a greater extent for weight-control purposes (P<.05). Mean insulin doses were similar between groups, but women with eating disorders had a greater range of insulin doses (P<.05). Anthropometric and body composition measures are given in Table 2. With the exception of WHR, no significant differences were found between groups for any of the variables. However in the group with eating disorders, significantly larger variances among women were noted in body weight, BMI, triceps skinfold measurement, mid-arm muscle circumference, and percentage total body water and intracellular water. Nearly a third of the women with eating disorders had triceps skinfold measurements below the 10th percentile. In our study, higher WHRs were associated with eating disorders. In contrast, others have reported lower WHRs for women who have eating disorders but do not have diabetes ((15)Radke-Sharpe N. Whitney-Saltiel D. Rodin J. Fat distribution as a risk factor for weight and eating concerns.Int J Eating Disord. 1990; 9: 927-936Google Scholar). Our findings may differ because higher WHRs are associated with increased rates of diabetes ((16)Hartz A.J. Rupley D.C. Rim A.A. The association of firth measurements with disease in 32,856 women.Am J Physiol. 1984; 119: 71-80Google Scholar). ApplicationsThis study characterized eating disordered behaviors in women with type 1 diabetes and confirmed the misuse of insulin as a method of purging. Nearly half of women with eating disorders reported current misuse of insulin, which was motivated by the desire to lose weight. On average, the women with eating disorders were more variable in weight, BMI, triceps and suprailiac skinfold measurements, and percentage total body water and intracellular water compared to women without eating disorders. This variability in the group with eating disorders may reflect achievement of weight loss via insulin misuse, insulin misuse combined with dietary restriction, or dietary restriction alone. Because insulin misuse results in poor metabolic control ((17)Affenito S.G. Backstrand J.R. Welch G.W. Lammi-Keefe C.J. Rodriguez N.R. Adams C.H. Subclinical and clinical eating disorders in insulin-dependent diabetes mellitus (IDDM) negatively affect metabolic control.Diabetes Care. 1997; 20: 182-184Google Scholar) and diabetes-related complications ((18)Rydall A.C. Rodin G.M. Olmsted M.P. Devenyl R.G. Daneman D. Disordered eating behavior and microvascular complications in young women with insulin-dependent diabetes mellitus.N Engl J Med. 1997; 336: 1849-1854Google Scholar), this behavior as a method of weight control should be identified in initial screening protocols ((19)Nutrition Practice Guidelines for Type 1 and Type 2 Diabetes Mellitus. American Dietetic Association, Chicago, Ill1998Google Scholar, (20)American Diabetes Association.Clinical Practice Recommendations.Diabetes Care. 1997; 20 (14–S17)Google Scholar). ■This report was supported, in part, by the University of Connecticut Research Foundation, Storrs Agricultural Experiment Station, National Institutes of Health General Clinical Research Grant (M01RR06192), University of Connecticut Health Center. This study characterized eating disordered behaviors in women with type 1 diabetes and confirmed the misuse of insulin as a method of purging. Nearly half of women with eating disorders reported current misuse of insulin, which was motivated by the desire to lose weight. On average, the women with eating disorders were more variable in weight, BMI, triceps and suprailiac skinfold measurements, and percentage total body water and intracellular water compared to women without eating disorders. This variability in the group with eating disorders may reflect achievement of weight loss via insulin misuse, insulin misuse combined with dietary restriction, or dietary restriction alone. Because insulin misuse results in poor metabolic control ((17)Affenito S.G. Backstrand J.R. Welch G.W. Lammi-Keefe C.J. Rodriguez N.R. Adams C.H. Subclinical and clinical eating disorders in insulin-dependent diabetes mellitus (IDDM) negatively affect metabolic control.Diabetes Care. 1997; 20: 182-184Google Scholar) and diabetes-related complications ((18)Rydall A.C. Rodin G.M. Olmsted M.P. Devenyl R.G. Daneman D. Disordered eating behavior and microvascular complications in young women with insulin-dependent diabetes mellitus.N Engl J Med. 1997; 336: 1849-1854Google Scholar), this behavior as a method of weight control should be identified in initial screening protocols ((19)Nutrition Practice Guidelines for Type 1 and Type 2 Diabetes Mellitus. American Dietetic Association, Chicago, Ill1998Google Scholar, (20)American Diabetes Association.Clinical Practice Recommendations.Diabetes Care. 1997; 20 (14–S17)Google Scholar). ■ This report was supported, in part, by the University of Connecticut Research Foundation, Storrs Agricultural Experiment Station, National Institutes of Health General Clinical Research Grant (M01RR06192), University of Connecticut Health Center.
Referência(s)