Accuracy of reporting food energy intake: influence of ethnicity and body weight status in South African women

Similar documents
Health-Related Quality of Life and Symptoms of Depression in Extremely Obese Persons Seeking Bariatric Surgery

XII. HIV/AIDS. Knowledge about HIV Transmission and Misconceptions about HIV

Input from external experts and manufacturer on the 2 nd draft project plan Stool DNA testing for early detection of colorectal cancer

EVALUATION OF DIFFERENT COPPER SOURCES AS A GROWTH PROMOTER IN SWINE FINISHING DIETS 1

Metabolic Syndrome and Health-related Quality of Life in Obese Individuals Seeking Weight Reduction

Clinical Study Report Synopsis Drug Substance Naloxegol Study Code D3820C00018 Edition Number 1 Date 01 February 2013 EudraCT Number

Assessment of Depression in Multiple Sclerosis. Validity of Including Somatic Items on the Beck Depression Inventory II

SYNOPSIS Final Abbreviated Clinical Study Report for Study CA ABBREVIATED REPORT

PNEUMOVAX 23 is recommended by the CDC for all your appropriate adult patients at increased risk for pneumococcal disease 1,2 :

Invasive Pneumococcal Disease Quarterly Report July September 2018

The Effects of Small Sized Rice Bowl on Carbohydrate Intake and Dietary Patterns in Women with Type 2 Diabetes

The Effect of Substituting Sugar with Artificial. Sweeteners on the Texture and Palatability of Pancakes

Study of Stress Distribution in the Tibia During Stance Phase Running Using the Finite Element Method

Prevalence of Prehypertension in a Rural District of Southern India

EFFECT OF DIETARY ENZYME ON PERFORMANCE OF WEANLING PIGS

Effect of orthodontic treatment on oral health related quality of life

Factors influencing help seeking in mentally distressed young adults: a cross-sectional survey

Extraction and Some Functional Properties of Protein Extract from Rice Bran

Optimisation of diets for Atlantic cod (Gadus morhua) broodstock: effect of arachidonic acid on egg & larval quality

obesità nel bambino: epidemiologia e prevenzione

Effects of physical exercise on working memory and prefrontal cortex function in post-stroke patients

EFFECTS OF AN ACUTE ENTERIC DISEASE CHALLENGE ON IGF-1 AND IGFBP-3 GENE EXPRESSION IN PORCINE SKELETAL MUSCLE

Using Paclobutrazol to Suppress Inflorescence Height of Potted Phalaenopsis Orchids

Dietary Sodium Intake in People with Diabetes in Korea: The Korean National Health and Nutrition Examination Survey for 2008 to 2010

Metabolic syndrome (MetS) is defined by a group

Recall Bias in Childhood Atopic Diseases Among Adults in The Odense Adolescence Cohort Study

BMI and Mortality: Results From a National Longitudinal Study of Canadian Adults

ENERGY CONTENT OF BARLEY

The Acute Time Course of Concurrent Activation Potentiation

A FACTORIAL STUDY ON THE EFFECTS OF β CYCLODEXTRIN AND POLOXAMER 407 ON THE SOLUBILITY AND DISSOLUTION RATE OF PIROXICAM

Diabesity & Associated Disorders in Australia

Debra A. Ignaut, R.N., B.S., C.D.E., and Haoda Fu, Ph.D.

Receiving family physician s advice and the stages of change in smoking cessation among Arab minority men in Israel

ARTICLES. The potential role of biotin as dietary risk marker for hypertension in black South African children the THUSA BANA study

Studies on the Predisposing Factors of Protein Energy Malnutrition Among Pregnant Women in a Nigerian Community

Single-Molecule Studies of Unlabelled Full-Length p53 Protein Binding to DNA

Effect of linear and random non-linear programming on environmental pollution caused by broiler production

Overweight and health-related quality of life in adolescents of Florianópolis, Southern Brazil

Community. Profile Yellowstone County. Public Health and Safety Division

Feeding state and age dependent changes in melaninconcentrating hormone expression in the hypothalamus of broiler chickens

THE EVALUATION OF DEHULLED CANOLA MEAL IN THE DIETS OF GROWING AND FINISHING PIGS

Community. Profile Big Horn County. Public Health and Safety Division

Body mass index, waist-to-hip ratio, and metabolic syndrome as predictors of middle-aged men's health

Community. Profile Lewis & Clark County. Public Health and Safety Division

Community. Profile Missoula County. Public Health and Safety Division

Roughage Type & Level & Grain Processing Interactions with Distiller s s Grains Diets. Matt May High Plains Bio Fuels Co-Product Nutrition Conference

Community. Profile Powell County. Public Health and Safety Division

Knowledge, Attitude, and Concussion-Reporting Behaviors Among High School Athletes: A Preliminary Study

Appendix J Environmental Justice Populations

3. DRINKING WATER INTAKE BACKGROUND KEY GENERAL POPULATION STUDIES ON DRINKING WATER INTAKE RELEVANT GENERAL POPULATION

EFFECTS OF INGREDIENT AND WHOLE DIET IRRADIATION ON NURSERY PIG PERFORMANCE

Detecting Undiagnosed Type 2 Diabetes: Family History as a Risk Factor and Screening Tool. Rodolfo Valdez, Ph.D.

A Study of Serological Markers of Hepatitis B and C Viruses in Istanbul, Turkey

Review TEACHING FOR GENERALIZATION & MAINTENANCE

URINARY incontinence is an important and common

Prime Enrollees Consumer Watch NHC Patuxent River FY 2016 Defense Health Cost Assessment & Program Evaluation

Community. Profile Anaconda- Deer Lodge County. Public Health and Safety Division

Dietary Characteristics of Hong Kong Young Children: Implications for Nutrition Education

Effect of supplemental fat from dried distillers grains with solubles or corn oil on cow performance, IGF-1, GH, and NEFA concentrations 1

Abstract. Background. Aim. Patients and Methods. Patients. Study Design

Randomized Controlled Trial to Improve Adiposity, Inflammation, and Insulin Resistance in Obese African-American and Latino Youth

ORIGINAL INVESTIGATION. Exercise and Weight Loss Reduce Blood Pressure in Men and Women With Mild Hypertension

Fat intake in patients newly diagnosed with type 2 diabetes: a 4-year follow-up study in general practice

Geographical influence on digit ratio (2D:4D): a case study of Andoni and Ikwerre ethnic groups in Niger delta, Nigeria.

Ethnic Disparities in Stroke Recognition in Individuals with Prior Stroke

Invasive Pneumococcal Disease Quarterly Report. July September 2017

How adaptations of substrate utilization regulate body composition

CheckMate 153: Randomized Results of Continuous vs 1-Year Fixed-Duration Nivolumab in Patients With Advanced Non-Small Cell Lung Cancer

A novel Mediterranean diet index from Lebanon: comparison with Europe

DXA: Can It Be Used as a Criterion Reference for Body Fat Measurements in Children?

Check your understanding 3

Introduction. In developing countries, children s weight gain commonly falters in relation to reference data

Regression of electrocardiographic left ventricular hypertrophy predicts regression of echocardiographic left ventricular mass: the LIFE study

Diagnostic Accuracy of Mini-Mental Status Examination and Revised Hasegawa Dementia Scale for Alzheimer s Disease

In 2006, the prevalence of bipolar

unit 1.9 Problems with unknown (II) 15 days D Figures within days 1 G 1 1 OA 1 1 OA 1 1 OA 1 1 G 2 1 OA 6 1 OA 4 1 OA 4 1 G 3 1 OA 7

Employment Status and Depressive Symptoms in Koreans: Results From a Baseline Survey of the Korean Longitudinal Study of Aging

Relationship of the Dental Aesthetic Index to the oral health-related quality of life

USE OF SORGHUM-BASED DISTILLERS GRAINS IN DIETS FOR NURSERY AND FINISHING PIGS

SUPPLEMENTARY INFORMATION

Original Research Article. Dement Geriatr Cogn Disord 2007;24: DOI: /

Consumer perceptions of meat quality and shelf-life in commercially raised broilers compared to organic free range broilers

Teacher motivational strategies and student self-determination in physical education

Supplementary Online Content

Utilization of dental services in Southern China. Lo, ECM; Lin, HC; Wang, ZJ; Wong, MCM; Schwarz, E

Bioactive milk components to secure growth and gut development in preterm pigs ESTER ARÉVALO SUREDA PIGUTNET FA1401 STSM

Potassium Intake of the U.S. Population

Comparison of three simple methods for the

Longitudinal Association of Maternal Attempt to Lose Weight During the Postpartum Period and Child Obesity at Age 3 Years

2. Hubs and authorities, a more detailed evaluation of the importance of Web pages using a variant of

Physical activity and risk of cancers of the colon and rectum: an Italian case-control study

Life satisfaction 6 15 years after a traumatic brain injury

Community. Profile Carter County. Public Health and Safety Division

Chronic high-sodium diet intake after weaning lead to neurogenic hypertension in adult Wistar rats

There has been little systematic

Dependency on Smartphone Use and Its Association with Anxiety in Korea

Meat and Food Safety. B.A. Crow, M.E. Dikeman, L.C. Hollis, R.A. Phebus, A.N. Ray, T.A. Houser, and J.P. Grobbel

INFLUENCE OF DIFFERENT STRAINS AND WAYS OF INOCULATION ON THE RABBIT S RESPONSE TO EXPERIMENTAL INFECTION WITH PASTEURELLA MULTOCIDA

Transcription:

Accurcy of reporting food energy intke: influence of ethnicity nd ody weight sttus in South Africn women Mchiz ZJ, PhD Goedecke JH, PhD Lmert EV, PhD Knowledge Systems, Nutrition Unit, Humn Science Reserch Council, Cpe Town, South Afric UCT/MRC Reserch Unit for Exercise Science nd Sports Medicine, Dept of Humn Biology, Fculty of Helth Sciences, University of Cpe Town, South Afric Correspondence to: Dr Zndile J Mchiz, e-mil: zmchiz@hsrc.c.z Keywords: South Africn women; misreporting; energy intke; ody imge; socioeconomic sttus Astrct The current study sought to identify chrcteristics tht my e ssocited with the misreporting of food energy intke (EI) in urn South Africn women. A totl of 198 women (61 lck, 76 of mixed ncestry, 61 white) completed quntified food frequency questionnire, from which dily energy nd mcronutrient intke were clculted. Body composition (ody mss index [BMI], percentge of ody ft), ody imge (Feel-Idel Difference index nd Body Shpe questions) nd socio-economic sttus (SES) (household density nd sset index) were lso mesured. Food EI in reltion to estimted sl metolic rte rtio tht ws less thn 1.05 represented under-reporting, wheres rtio greter thn 2.28 represented over-reporting. Results suggested tht 26% of the prticipnts under-reported, 64% dequtely reported nd 10% over-reported. Prticipnts who under-reported hd higher BMI (p < 0.01) nd higher percentge of ody ft (p < 0.05) thn those who dequtely nd over-reported. The mjority of under-reporters were lck (38%) versus 21% under-reporters of mixed ncestry nd 20% white under-reporters (p < 0.01). Eighty-three per cent of lck under-reporters were oese. On the other hnd, mjority (63%) of overweight women of mixed ncestry nd mjority (50%) of white norml-weight women under-reported their food EI. Under-reporters reported lower intke of dietry ft (p < 0.01) nd higher intke of dietry protein (p < 0.01) thn dequte or over-reporters. Food EI reporting ws not influenced y SES or ody imge. In conclusion, results suggest tht food EI reporting is influenced y ody size, nd my e ethnic-specific in South Africn women. Peer reviewed. (Sumitted: 2009-09-24, Accepted 2010-04-25). SAJCN : 84-89 Introduction Oesity is glol prolem. 1 In South Afric, s in mny other developing countries, dult women re the most vulnerle group, with mrkedly higher prevlence of oesity thn men. 1,2 There is sustntited evidence suggesting oesity s result of chronic positive energy lnce in the form of food energy intke (EI) tht is higher thn physicl ctivity energy expenditure (PAEE). 3 However, food EI nd PAEE (whether reported or mesured) my e ised (under-reported, in prticulr) 4,5 to greter extent in women thn in men. 6 Wht is of concern is tht studies suggest tht if food EI-reporting is is not evenly distriuted in the popultion nd ffects certin nutrients, the interprettion of the reltionship etween diet nd diseses in tht popultion my e ltered. 5,7 As such, identifying the group or groups likely to under-report their food EI nd resons for misreporting my help to understnd the reltionship etween diet nd oesity in South Africn women. In ddition to gender, 6 interntionl reserchers hve lso highlighted fctors such s ge, 8 socioeconomic sttus (SES) nd eduction level, 6,9,10 ody composition, 6,11 ethnicity or culture 10,11 nd socil desirility 9 s influences of food EI misreporting. Bis in food EI reporting cn e mesured y the rtio etween reported food EI nd energy expenditure (EE). 12 EE my e estimted using sl metolic rte (BMR est ), clculted using Schofield et l s 13 equtions, or mesured (BMR mes ), using indirect clorimetry 14 or douly lelled wter (DLW). 15 However, the mesurement of EE using indirect clorimetry or DLW in lrge group of individuls is costly. Vrious consulttive groups nd reserchers hve proposed guidelines for determining the EI : BMR rtio. For exmple, in 1985, the Joint FAO/WHO/UNU Expert Consulttion on Energy nd Protein Requirements proposed tht n EI of 1.55 x BMR ws dequte for dults to sustin resonle helth nd light ctivity. Moreover, Golderg et l nd others hve recommended rnge of EIs from 1.35 to 1.67 x BMR s plusile rtio for dults in ffluent societies to mintin helth nd lower levels of ctivity. 12,15 Susequently, Blck 16 reviewed the evidence supporting vrious proposed cut-off points in terms of the method used for mesuring dietry intke, whether or not resting EE ws mesured or estimted, nd the mens y which physicl ctivity levels were determined. In her nlysis, she determined the specificity nd sensitivity of vrious 84

cut-off points in identifying under- nd over-reporting. As such, she proposed strtegy tht employed the 95% confidence intervls (CIs) round the EI : BMR rtio under vrious conditions to reflect under- nd over-reporting. The selection of criteri for clssifying misreporting of food EI is therefore informed y one s ility to estimte or mesure resting nd PAEE nd the method used for collecting dietry intke dt. Only one study y McIntyre et l 17 previously identified food EI misreporting in 43% of South Africns from multi-ethnic smple of 178 men nd women. However, the extent of misreporting nd the fctors ssocited with over- nd under-reporting hve not yet een studied. Therefore, the im of the current study ws to identify the extent nd determinnts of fctors ssocited with the misreporting of food EI in smple of South Africn women of mixed ethnic origin. Methods nd procedures Study popultion This study ws undertken s prt of lrger project in which diet, physicl ctivity, knowledge, ttitudes, eliefs nd helth ehviours in South Africn women nd their dughters were evluted, nd pulished in prt elsewhere, in which more detiled methodology is included. 18 In rief, 15 primry schools in the Cpe Town Metropole re were rndomly selected nd smpled on the sis of divergent SES. All girls (ges 9 12 yers, grdes 4 5) nd their mothers were invited to prticipte in the study. Only the mothers were included in the current nlysis nd of the mothers who responded, ll women who were involved in ny weight chnge intervention strtegy t the time of the study were excluded. As result, the finl smple of 198 women consisting of 31% lck, 38% mixed-ncestry nd 31% white women were included in this nlysis. Ethics pprovl to undertke this study ws otined from the Western Cpe Deprtment of Eduction, s well s the Reserch Ethics Committee of the Fculty of Helth Sciences t the University of Cpe Town. All prticipnts completed informed consent forms prior to prticiption in the study. Body composition, socioeconomic nd psychosocil vriles Body composition ws ssessed in the form of ody mss index (BMI), clculted s weight (in kilogrms) divided y the squre of height (in metres) nd percentge of ody ft. Body weight ws ssessed in light clothing, without shoes, nd recorded to the nerest 0.5 kg using clirted electronic scle (TANITA HD-309, Tnit Corportion of Americ Inc, USA). Height ws mesured without shoes to the nerest 0.5 cm using clirted odometer. Furthermore, to mesure ody ft stores, triceps, iceps, suscpulr nd suprilic skinfold thicknesses were mesured using clirted Hrpenden cllipers, nd recorded to the nerest 0.1 mm. Therefter the ody ft percentge mesurements were clculted using stndrd equtions y Durnin nd Womersely. 19 Age, ethnic group, eduction level nd SES (determined s sset index [which is the totl numer of pplinces in one household out of list of nine pplinces] nd household density [which is the totl numer of people residing in the sme household for five or more dys week]) were lso investigted. 20 Dietry intke In the current study, quntified food frequency questionnire (QFFQ) y McIntyre et l 17 ws employed. The QFFQ is more pproprite, vlid method tht is more likely to give true representtive picture of the usul dietry intke of individuls over period of time. It is lso more pproprite method used y reserchers to link dietry intke to diseses. The QFFQ used in this regrd ws comprised of 122 food items otined from locl pulished studies on dietry intke y different ethnic groups of South Afric. The QFFQ is presented such tht prticipnts re le to choose the food items tht they regulrly consume, long with the quntity nd frequency of consumption of these items, within the preceding month. Among other methods to vlidte the QFFQ, McIntyre et l 17 compred it to the seven-dy weighed food records (in 74 multi-ethnic dult South Africns etween the ges of 15 nd 65 yers). Some of the results otined y McIntyre et l 17 were Spermn s rnk correltions of 0.21 for ft, 0.35 for met, 0.38 for fruit, 0.41 for vegetles s well s 0.56 for mize mel nd dded sugr. The QFFQ ws dministered y trined registered dieticins. The field worker resource mnul used for trining the dieticins ws prepred y the principl investigtor for the current study. To id prticipnts in estimting their food portion sizes nd food preprtion methods, food portion photogrph ook (FPPB) 21 ws used. The food EI generted y the QFFQ ws nlysed using the South Africn Medicl Reserch Council (MRC) Foodfinder 3 softwre progrm (WAMTechnology nd MRC RISD, 2001). The totl energy (in kj) nd mcronutrient intke (in the form of the percentge of totl energy derived from dietry ft, dded sugr nd dietry protein) generted from the nlysis ws then clculted to express verge intke per dy. Energy reporting sttus (EI : BMR est ) The reported EI in reltion to BMR est (EI : BMR est rtio) ws clculted for ech individul to determine the reporting sttus of the women. The BMR ws estimted using Schofield stndrd equtions. 13 In the current study, n estimted lnketed physicl ctivity level (for light ctivity) of 1.55 derived from the FAO/WHO/UNU in 1985 ws used. This ws sed on the evidence from South Africn study tht suggested tht urn women engge in light occuptionl ctivity. 22 In this regrd, the 95% CI were employed for the Golderg et l 12 nd Blck 16 cut-off points with the sensitivity of 0.52 nd specificity of 0.99 for under-reporting food EI. As such, in the current study, ny EI : BMR est vlue elow cut-off point of 1.05 represented underreporting. 16 In ddition, ny EI : BMR est vlue ove 2.28 represented over-reporting. 16 All other prticipnts were considered dequte reporters. 16 85

Golderg et l 12 formulted these cut-off points y compring food EI to n individul s resting energy expenditure (mesured using n ojective mesure such s DLW). Blck 16 lso vlidted the food EI results reported y Golderg et l 12 ginst 24-hour urinry nitrogen excretion nd EE mesured y DLW in British middle-ged women. Body imge Body shpe nd size cceptnce ws ssessed using the Body Shpe Questionnire (BSQ) y Cooper et l 23 nd the Feel-Idel Difference (FID) index y Mciz et l. 18 The BSQ is 34-item questionnire tht mesures ody shpe concerns. A BSQ score < 1.23 indicte lower ody shpe concerns, wheres BSQ score 123 indicted higher ody shpe concerns. Body size disstisfction ws ssessed using the FID index creted y determining the difference in the numer of silhouettes selected tht est represented the prticipnts current ppernce (determined s Feel ), nd the one the prticipnts thought ws their Idel (the silhouettes they wnt to look like). A higher FID index score represents greter ody size disstisfction, wheres FID index score tht pproches zero represents less ody size disstisfction. The silhouettes used were derived from set of nine silhouettes y Stunkrd et l, 24 rnging from very thin to very hevy ody imge. Sttisticl nlysis All dt were nlysed using STATISTICA version 7.0 (SttSoft Inc, Tuls, OK, USA). Vlues re presented s mens ± stndrd devition. Chi-squred nlysis ws used to ssess the frequency of food EI-reporting sttus ccording to ethnicity nd BMI groups nd presented s percentge. In ddition, GrphPd Instt t m Dos Progrmme (Copyright 1990 1994, Lmert M, University of Cpe Town) ws used to clculte Chi-squred test for trend of the dequte reporters food EI ccording to ethnic group nd BMI ctegories. Two-wy nlysis of vrince (ANOVA) ws used to compre ody composition, mcronutrient intke, ody imge nd SES etween the ethnic nd food EI-reporting groups. Furthermore, dvnced generl liner model nlysis ws used to djust for confounding fctors of ge nd BMI on food EI-reporting sttus etween different ethnic groups. Results Prticipnts chrcteristics Detiled chrcteristics of the min 204 smple re presented y Mciz et l. 18 In rief, the verge ges of the 198 women who prticipted in the current study were 42 ± 5, 38 ± 5 nd 40 ± 11 yers for the white, mixed-ncestry nd lck women respectively (p = 0.02). Furthermore, lck women hd significntly higher BMIs (33.6 ± 7.8 vs 26.5 ± 4.7 nd 25.4 ± 4.1 kg/m 2, p < 0.05) nd ody ft percentge (34.5 ± 6.3 vs 32.2 ± 4.8 nd 31.8 ± 4.9 %, p < 0.05) thn the mixed-ncestry nd white women respectively. They lso hd significntly lower levels of eduction nd lso presented with significntly lower SES on the sis of sset index nd household density thn the other groups of women (ll p vlues < 0.001). In ddition, when djusting for ge nd BMI, white women scored significntly higher on the BSQ, indicting higher ody shpe concerns, thn the mixed-ncestry nd lck women (86.7 ± 24.1 vs 81.7 ± 28.1 nd 80.8 ± 35.0, p < 0.05, respectively). On the other hnd, mixed-ncestry women scored significntly higher on the ody size disstisfction (presented s FID index scores) thn white nd lck women (1.7 ± 1.1 vs 1.5 ± 1.3 nd 1.2 ± 2.2, p < 0.05, respectively). Misreporting of food energy intke Using the cut-off points of 1.05 2.28 for food EI reporting, it ws oserved tht overll, 26% of the women in this nlysis underreported, 64% dequtely reported nd 10% over-reported their food EI (p < 0.05). When compring the food EI-reporting sttus of the three ethnic groups of women, lck women under-reported food EI to greter extent thn mixed-ncestry nd white women (45% vs 31% nd 24%, p < 0.01, respectively). Food EI-reporting sttus ccording to BMI ctegory nd ethnicity is presented in figures 1A nd 1B. Figure 1A shows tht, of the 45% (n = 23) lck women who under-reported food EI, only 4% (n = 1) were within the norml rnge of BMI, 13% (n = 3) were overweight nd 83% (n = 19) were oese. Of the 31% (n = 16) mixed-ncestry women who under-reported food EI, 13% (n = 2) were within the norml rnge of BMI, 63% (n = 10) were overweight nd 24% Figure 1: Frequency of (A) under-reporting nd (B) dequte reporting of food EI ccording to BMI ctegory nd ethnicity in South Africn women. Mtching superscripts represent groups tht re different from ech other, (c2 = 8.782, p < 0.01); (c2 = 10.667, p < 0.01 nd c2 = 18.910, p < = 0.001, for lck, mixed-ncestry nd white, respectively) A 100 BMI < 25 B 100 BMI < 25 75 BMI = 25-30 BMI > 30 75 BMI = 25-30 BMI > 30 Totl prticipnts (%) 50 25 Totl prticipnts (%) 50 25 0 Blck (n=23) Mixed ncestry (n=16) White (n=12) 0 Blck (n=31) Mixed ncestry (n=54) White (n=42) 86

Tle I: Chrcteristics of South Africn women ccording to ethnicity nd food EI-reporting sttus Physicl chrcteristics Blck 31% (n = 61) Mixed ncestry 38% (n = 76) White 31% (n = 61) UR (n = 23) AR (n = 31) OR (n = 7) UR (n = 16) AR (n = 54) OR (n = 6) UR (n = 12) AR (n = 42) OR (n = 7) Age (yrs) 40.6 ± 11.6* 39.2 ± 10.9 36.6 ± 7.6 37.1 ± 4.7* 38.0 ± 4.5 40.8 ± 3.4 42.7 ± 5.8* 41.0 ± 4.1 43.3 ± 5.9 Weight (kg) 92.2 ± 20.5 * 79.5 ± 22.5* 78.7 ± 21.2 70.0 ± 10.1 65.0 ± 11.8 71.6 ± 19.3 67.5 ± 15.4 68.5 ± 9.5 68.4 ± 16.4 BMI (kg/m 2 ) 36.6 ± 7.1 # 31.9 ± 7.8 # 31.3 ± 8.5 28.1 ± 4.3 25.8 ± 4.6 28.5 ± 5.7 26.2 ± 5.8 25.2 ± 3.8 24.9 ± 3.1 Body ft (%) 36.4 ± 6.2* 34.1 ± 5.6* 29.8 ± 9.7 33.9 ± 4.0 31.2 ± 4.9 36.1 ± 4.9 33.4 ± 5.3 31.5 ± 4.9 30.9 ± 4.9 Socioeconomic chrcteristics Eduction sttus 3.2 ± 1.2 3.3 ± 1.1 3.6 ± 0.8 4.1 ± 0.3 4.4 ± 0.6 4.2 ± 0.8 4.4 ± 0.7 4.7 ± 0.5 4.9 ± 0.4 Asset index 5.5 ± 2.4 5.3 ± 2.0 6.1 ± 1.8 8.6 ± 0.6 8.4 ± 0.9 8.5 ± 0.8 8.8 ± 0.4 8.7 ± 0.6 9.0 ± 0.0 Household density 3.1 ± 1.4 3.2 ± 1.8 3.7 ± 2.4 1.5 ± 0.6 1.6 ± 0.5 1.4 ± 0.4 1.4 ± 0.0 1.3 ± 0.4 1.1 ± 0.2 Body imge BSQ 79.6 ± 39.2 78.5 ± 34.9 85.0 ± 17.3 91.8 ± 27.7 80.5 ± 28.4 64.8 ± 18.9 92.3 ± 31.3 83.8 ± 22.5 83.9 ± 22.0 FID Index 1.5 ± 1.9 1.3 ± 2.4 0.3 ± 2.8 1.7 ± 0.8 1.0 ± 1.4 1.5 ± 0.5 1.5 ± 1.5 1.1 ± 1.0 0.6 ± 0.5 Vlues re men ± stndrd devition: UR = under-reporters (EI : RMR est < 1.05); AR = dequte reporters (EI : RMR est = 1.05 2.28); OR = over-reporters (EI : RMR est > 2.28). Asset index: totl numer of pplinces in one household out of nine; Household density: totl numer of people sleeping in the sme household for five or more dys week; Eduction sttus: highest grde ctegory pssed. Mtching superscripts represent groups tht re significntly different from ech other: *p < 0.05, p < 0.001, # p < 0.01. After djusting for BMI nd ge: p < 0.05 Tle II: Reported mcronutrient intke (in the form of percentge of totl energy derived from dietry ft, protein nd crohydrtes, s well s dded sugr in grms) of South Africn women ccording to ethnicity nd food EI-reporting sttus Blck 31% (n = 61) Mixed ncestry 38% (n = 76) White 31% (n = 61) UR (n = 23) AR (n = 31) OR (n = 7) UR (n = 16) AR (n = 54) OR (n = 6) UR (n = 12) AR (n = 42) OR (n = 7) Crohydrtes (% TE) 64.4 ± 9.2 59.3 ± 9.5 57.9 ± 6.5 52.5 ± 9.6 55.2 ± 7.5 53.9 ± 9.0 55.9 ± 10.4 53.6 ± 7.6 53.4 ± 13.0 Ft (% TE) 22.9 ± 8.6 # 28.4 ± 8.3 # 29.3 ± 7.7 # 30.9 ± 7.6 # 31.7 ± 6.9 # 34.7 ± 9.6 # 29.7 ± 8.5 # 32.8 ± 6.6 # 35.0 ± 12.2 # Protein (% TE) 12.9 ± 3.1 12.2 ± 3.2 12.8 ± 3.0 16.2 ± 4.5 # 12.9 ± 2.4 # 11.3 ± 2.6 # 13.6 ± 3.4 13.1 ± 3.0 11.2 ± 2.6 Added sugr (g) 15.5 ± 6.9 17.5 ± 6.7 19.7 ± 5.5 19.2 ± 8.3 16.0 ± 7.2 11.5 ± 8.5 20.6 ± 12.8 16.4 ± 6.6 12.5 ± 7.4 Vlues presented s Men in grmsrived from ody ft percentge insted of percentge ody ft. More over, we ment Vlues re men ± stndrd devition: % TE = percentge of totl energy; g = grms UR = under-reporters (EI : RMR est < 1.05); AR = dequte reporters (EI : RMR est = 1.05 2.28); OR = over-reporters (EI : RMR est > 2.28). Mtching superscripts represent groups tht re significntly different from ech other: p < 0.001: Blck women reported higher crohydrte intke nd less ft intke thn mixed-ncestry nd white women. # p < 0.01: UR reported less ft intke thn AR nd OR in ll ethnic groups. Furthermore, UR reported higher protein intke thn AR nd OR for the mixed-ncestry women. p < 0.05: UR reported higher protein intke in lck nd white women only fter djusting for BMI nd ge of women. p < 0.05: Blck UR reported less dded sugr intke thn AR nd OR, wheres mixed-ncestry nd white UR reported higher sugr intke thn OR. (n = 4) were oese. Of the 24% (n = 12) of white women who under-reported food EI, 50% (n = 6) were within the norml rnge of BMI, 25% (n = 3) were overweight nd 25% (n = 3) were oese. A greter-thn-expected proportion of mixed-ncestry women dequtely reported food EI, compred to their white nd lck counterprts (71% [n = 54] vs 69% [n = 42] nd 51% [n = 31], p < 0.05, respectively) (Figure 1B). The frequency of dequte reporters incresed linerly with incresing BMI ctegory in the lck women (Chi-squre [χ 2 ] = 8.782 nd p < 0.01), wheres the frequency of dequte reporters decresed linerly with incresing BMI ctegory in the mixed-ncestry (χ 2 = 10.667 nd p < 0.01) nd white women (χ 2 = 18.910 nd p < 0.001). There were no significnt differences in over-reporting etween ethnic groups cross BMI ctegories. The chrcteristics of the women ccording to ethnicity nd food EI-reporting sttus re presented in Tle I. Those women who under-reported food EI were significntly hevier (p < 0.05), hd higher BMI (p < 0.01) nd percentge of ody ft (p < 0.05) thn those tht dequtely reported food EI. Although lck women were significntly hevier in terms of weight (p < 0.001) nd hd higher BMIs (p < 0.001) thn the mixed-ncestry nd white women, no significnt ethnic differences in food EI reporting were oserved (p = 0.20). Despite the fct tht lck women presented with lower eduction levels nd lower SES, these chrcteristics did not influence food EI-reporting sttus. Similrly, food EI-reporting sttus ws not influenced y ody imge (chrcterised y FID index nd BSQ scores). 87

Reported mcronutrient intke ccording to ethnicity nd food EI-reporting sttus is presented in Tle II. Irrespective of food EI-reporting sttus, lck women hd higher crohydrte intke nd less dietry ft intke thn the other groups of women (ll p vlues < 0.001). These results were independent of ge nd BMI. Signific nt ethnic differences in reported protein intke were oserved only fter djusting for the ge nd BMI of women. Furthermore, underreporters in ll ethnic groups reported less dietry ft nd higher dietry protein intke compred to the dequte nd over-reporters (oth p vlues < 0.01). There ws significnt interction effect for ethnicity nd food EI-reporting sttus for reported dded sugr intke (p < 0.05). In this regrd, lck women reported consistently higher dded sugr intke for oth dequte nd over-reporters thn the other groups of women. Blck under-reporters lso reported less dded sugr intke thn lck dequte nd over-reporters. White nd mixed-ncestry under-reporters, on the other hnd, reported significntly higher dded sugr intke compred to the white nd mixed-ncestry over-reporters. These significnt differences disppered fter djusting for the ge nd BMI of women. Discussion The min im of this study ws to identify chrcteristics in urn South Africn women tht my e ssocited with the misreporting of food EI, including ethnicity, SES, ody composition (BMI nd ody ft percentge), ody imge nd mcronutrient intke. The min findings of this study were tht 26% of the women under-reported their food EI, with greter proportion of under-reporters eing lck nd oese. In contrst, more overweight women of mixed ncestry nd more norml weight white women under-reported food EI. Under-reporters lso reported less dietry ft intke nd higher dietry protein intke thn the dequte nd over-reporters. However, EI : BMR est ws not ssocited with the SES or the ody imge of women. These findings highlight significnt prolem in terms of estimting food EI, s 26% of the women reported implusile food EI. This percentge is higher thn tht reported in similr study conducted in nother developing country. 11 For exmple, only 10% of the Egyptin women in Hrrison s study under-reported their food EI. However, the proportion of under-reporters in the current study is closer to tht found in one-third of Americn women. 11 As such, the role of dietry fctors in the etiology of oesity in South Africn studies should e interpreted with cution, ering in mind the influence of this food EI-reporting is. However, otining more knowledge of the fctors tht influence food EI reporting in South Afric will help with the design of etter dietry ssessment instruments nd possily of studies evluting diet disese reltionships, s proposed y Johnsson et l. 5 In the current study, ethnic ckground ppered to ply role in the is of food EI reporting, in tht lck women were more likely to under-report their food EI thn mixed-ncestry nd white women. An explntion for these findings my e tht the mjority of lck South Africn women in the current study were oese compred to other ethnic groups of women. There is well-estlished interntionl evidence suggesting tht oese women re more likely to underreport their food EI thn lighter women. 9 As such, it my seem s if under-reporting is not ethnic- or culture-ound, ut only reltes to ody composition. However, the ethnic differences in food EI reporting hve lso een oserved in other similr interntionl studies. 10,11 The results of the current study regrding ethnic differences in food EI reporting, however, contrst with those of Kimm et l, 11 who found tht white dolescent Americn girls under-reported their food EI to greter extent thn their lck counterprts. Differences in the findings of Kimm et l s 11 study nd the findings of the current study my relte to the ge difference etween the two study groups. However, oth the current study nd Kimm et l s 11 study highlight tht BMI is one of the most consistent fctors in predicting food EI under-reporting in women of different ethnic origins. For exmple, in the current study, only 4% of lck women in the lowest BMI ctegory under-reported food EI compred to 13% of mixed-ncestry nd 50% of white women. Similrly, Kimm et l 11 found tht lck girls in the highest tertile of BMI under-reported food EI to greter extent thn white girls, while within the lowest BMI tertile, lck girls under-reported food EI somewht less thn white girls. The differences in the BMI level t which lck, mixed-ncestry nd white dult women in the current study under-reported their food EI might e explined, in prt, y ethnic differences in ody imge discussed y Mciz et l. 18 In this study, results suggested tht lck dult women experienced disstisfction out their ody size sttus t higher level of BMI (BMI > 30 kg/m 2 ) thn the mixedncestry women, who experienced ody size disstisfction even if they were somewht overweight (BMI > 25 kg/m 2 ) nd the white women, who experienced ody size disstisfction even if they were not overweight t ll (BMI < 25 kg/m 2 ). To our surprise, in the current study, ody imge prmeters such s ody size disstisfction (presented s greter FID index scores) nd ody shpe concerns (presented s greter BSQ scores) were not specificlly ssocited with food EI-reporting sttus. Previous studies hve lso suggested tht socil clss is n importnt risk fctor for under-reporting. 5,11 In the current study, the mjority of lck women were of lower socil clss (sed on eductionl level, household density nd sset index scores) thn the mjority of mixed-ncestry nd white women. However, in the current study, eductionl level nd SES did not influence food EI reporting. Similrly, Hrrison et l 10 oserved no reltionship etween food EIreporting sttus nd forml eduction in Egyptin women. Misreporting in the current study did not only influence totl food EI, ut lso ised the reporting of mcronutrient intke in tht under-reporters reported less dietry ft nd higher dietry protein intke thn dequte nd over-reporters. Similr results hve een reported in other interntionl studies. 4,5 In some of these studies, 88

guilt ssocited with the consumption of food items emphsised in dietry modifiction interventions s unhelthy ft in prticulr hve lso een regrded s the drivers of under-reporting this food item in women with higher BMIs. From the current dt, reserchers were not le to scertin whether guilt influenced mcronutrient reporting. However, ll prticipnts who prticipted in dietry modifiction interventions directed t losing weight were excluded from the current nlysis, reducing the likelihood of this fctor confounding the results. However, future reserch should explore whether guilt influences mcronutrient reporting in the South Africn context, nd whether there re ny culturl differences tht my lso e driven y socil norms regrding this spect. In conclusion, the current study identified significnt group of women who misreported their food EI, sed on the cut-off rnge of 1.05 to 2.28 EI : BMR est. Food EI under-reporting in these women ws influenced y ody size sttus nd differed ccording to ethnicity. Furthermore, food EI reporting influenced mcronutrient reporting. As such, studies designed to explore the reltionship etween dietry intke nd oesity might e confounded y the is in food EI nd mcronutrient reporting, compromising interventions imed t preventing nd mnging oesity in South Africn women. Disclosure This reserch ws funded y the Nestlé Foundtion. We lso give specil thnks to the MRC of South Afric, Unit of Chronic Diseses of Lifestyle, the South Africn Deprtment of Science nd Technology nd the Ntionl Reserch Foundtion for the primry investigtor s scholrship. Acknowledgements Grtitude is extended to ll the principls, prents nd lerners in the Cpe Town Metropole primry schools in which the study ws conducted. Moreover, the ssistnce of Nsreen Jffer, Luren Hill, Mdlne Crsten nd Alici Hess in the process of dt collection nd dt entry is pprecited. Lst ut not lest, thnks to the MRC/ UCT Unit for Exercise Science nd Sports Medicine, Deprtment of Humn Biology, t the University of Cpe Town, for supplying conducive environment for this reserch. References 1. Ono T, Guthold R, Strong K. WHO Glol Comprle Estimtes. 2005 Accessed [2010 My 27]. Aville from http://www.who.int/infose IBRRef:199999 2. Puone T, Steyn K, Brdshw D, et l. Oesity in South Afric: The South Africn demogrphic nd helth survey. Oes Res. 2002;10(10):1038 1048. 3. Grier JL, Mnznres JM, Brny M, Contrers J, Amigó P, Sls-Slvdó J. Physicl ctivity, energy lnce nd oesity. Pulic Helth Nutr. 2007;10(10A):1194 1199. 4. Cook A, Pryer J, Shetty P. The prolem of ccurcy in dietry surveys: Anlysis of the over 65 UK Ntionl Diet nd Nutrition Survey. J Epidemiol Community Helth 2000;54(8):611 616. 5. Johnsson L, Solvoll K, Bjorneoe GE, Drevon CA. Under- nd overreporting of energy intke relted to weight sttus nd lifestyle in ntionwide smple. Am J Clin Nutr. 1998;68(2):266 274. 6. Amirklli B, Njfi M, Atie-Jfri A, Hosseini S, Heshmt R. Under- nd overreporting of energy in group of cndidtes for CABG surgery nd its ssocition with some nthropometric nd sociodemogrphic fctors, Tehrn, Irn. Vsculr Helth nd Risk Mngement 2008;4(5):1115 1120. 7. Beton GH. Approches to nlysis of dietry dt: Reltionship etween plnned nlyses nd choice methodology. Am J Clin Nutr. 1994;59(suppl):253S 261S. 8. Bndini LG, Cyr H, Must A, Dietz WH. Vlidity of reported energy intke in pre-dolescent girls. Am J Clin Nutr. 1997;65:1138S 1141S. 9. Horner NK, Ptterson RE, Neuhouser ML, Lmpe JW, Beresford SA, Prentice RL. Prticipnt chrcteristics ssocited with errors in self-reported energy intke from the Women s Helth Inititive food-frequency questionnire. Am J Clin Nutr. 2002;6(4):766 773. 10. Hrrison GG, Gll OM, Irhim N, et l. Underreporting of food intke y dietry recll is not universl: A comprison of dt from Egyptin nd Americn women. J Nutr. 2000;130:2049 2054. 11. Kimm SY, Glynn NW, Orznek E, Aston CE, Dniels SR. Rcil differences in correltes of misreporting of energy intke in dolescent femles. Oesity (Silver Spring) 2006;14(1):156 164. 12. Golderg GR, Blck AE, Je SA, et l. Criticl evlution of energy intke dt using fundmentl principles of energy physiology. 1: Derivtion of cut-off vlues to identify under-recording. Eur J Clin Nutr. 1991;45:569 581. 13. Schofield WN, Schofield C, Jmes WPT. Bsl metolic rte. Humn Nutrition: Clinicl Nutrition 1985;39C(Suppl.1):1 96. 14. Schoeller DA. Recent dvnces from ppliction of douly lelled wter to mesurement of humn energy expenditure. J Nutr. 1999;129:1765 1768. 15. Cowrd WA, Cole, TJ. Determintion of optimum dosing rtios. In: Prentice AM, editor. The doulylelled wter method for mesuring energy expenditure. Vienn: IAEA, 1990; p. 294 301. 16. Blck AE, Welch AA, Binghm SA. Vlidtion of dietry intkes mesured y diet history ginst 24 h urinry nitrogen excretion nd energy expenditure mesured y the douly-lelled wter method in middle-ged women. Br J Nutr. 2000;83(4):341 354. 17. McIntyre UE, Venter CS, Vorster HH. A culture-sensitive quntittive food frequency questionnire used in n Africn popultion. 1: Development nd reproduciility. Pulic Helth Nutr. 2001;4(1):53 62. 18. Mciz Z, Goedecke JH, Steyn NP, et l. Development nd vlidtion of instruments mesuring ody imge nd ody weight disstisfction in South Africn mothers nd their dughters. Pulic Helth Nutr. 2005;8(5):509 519. 19. Durnin JV, Womersley J. Totl ody ft, clculted from ody density nd its reltionship to skinfold thickness in 571 people ged 12 72 yers. Proc Nutr Soc 1973;32(1):45A. 20. Moser C, Felton A. The construction of n sset index: Mesuring sset index ccumultion in Ecudor. Glol economy nd development. Working Pper 87. Mss: Chronic Poverty Reserch Centre, The Brooking Institution; July 2007. 21. Venter CS, McIntyre UE, Vorster HH. The development nd testing of food portion photogrph ook fuse in n Africn popultion. Journl of Humn Nutrition & Dietetics 2000;13(3):205 218. 22. Kruger HS, Venter C, Vorster HH, Mrgetts BM. Physicl inctivity is the mjor determinnt of oesity in lck women in the North West province, South Afric: The THUSA Study. Nutrition 2002;18:422 427. 23. Cooper PJ, Tylor MJ, Cooper Z, Firurn CG. The development nd vlidtion of the Body Shpe Questionnire. Interntionl Journl of Eting Disorders 1987;6(4):485 494. 24. Stunkrd AJ, Sorensen T, Schulsinger F. Use of the Dnish Adoption Register for the study of oesity nd thinness. Res Pul Assoc Res Nerv Ment Dis. 1983;60:115 20. 89