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

Transcription

1 nture publishing group Longitudinl Assocition of Mternl Attempt to Lose Weight During the Postprtum Period nd Child Obesity t Age 3 Yers Kendrin R. Sonneville 1,2, Sheryl L. Rifs-Shimn 3, Emily Oken 3, Kren E. Peterson 2,4,5, Steven L. Gortmker 5, Mtthew W. Gillmn 2,3 nd Elsie M. Tvers 3,6 The effect of mternl ttempt to lose weight during the postprtum period on lter child weight hs not been explored. Among 1,044 mother infnt pirs in Project Viv, we estimted longitudinl ssocitions of mternl ttempt to lose weight during the postprtum period with child weight nd diposity t ge 3 yers nd exmined differences in ssocitions by type of weight loss strtegy used. Using covrite-djusted liner nd logistic regression models, we estimted ssocitions before nd fter djusting for mternl weight-relted vribles including prepregnncy BMI. At 6 months postprtum, 53% mothers were trying to lose weight. At ge 3 yers, men (s.d.) child BMI z-score ws 0.44 (1.01) nd 8.9% of children were obese. Children whose mothers were trying to lose weight t 6 months postprtum hd higher BMI z-scores (0.30 (95% confidence intervl (CI) 0.18, 0.42)) nd were more likely to be obese (3.0 (95% CI 1.6, 5.8)) t 3 yers of ge. Addition of mternl prepregnncy BMI to the models ttenuted but did not eliminte the ssocitions seen for BMI z-score (0.24 (95% CI 0.12, 0.36) nd obesity (2.4 (95% CI 1.2, 4.7)). Attempting to lose weight by exercising lone ws the only weight loss strtegy tht consistently predicted higher child BMI z-score (0.36 (95% CI 0.14, 0.58)) nd odds of obesity (6.0 (95% CI 2.2, 16.5)) t ge 3 yers. In conclusion, we observed n ssocition between mternl ttempt to lose weight t 6 months postprtum, prticulrly through exercise lone, mesured using single item nd child diposity t ge 3 yers. This ssocition should be thoroughly exmined in future studies. Obesity (2011) 19, doi: /oby Introduction Gesttionl weight gin is necessry for fetl growth nd development, lthough excessive gesttionl weight gin is common (1,2). To reduce complictions in subsequent pregnncies (3) nd to prevent long-term consequences of excess body weight nd obesity, women should lose retined postprtum weight. There is currently no consensus regrding wht weight loss strtegy postprtum women should be encourged to dopt. Accordingly, mny women my not receive weight loss dvice from helth-cre providers during the postprtum period (4). Severl studies hve exmined the effectiveness of prticulr postprtum weight loss interventions (5 8). However, it is uncler whether mother s ttempt to lose weight during the postprtum period is ssocited with longer term effects on the child. Limited reserch suggests tht mternl dieting my be ssocited with offspring overeting. In one longitudinl study of children, mternl restrint nd drive for thinness predicted the emergence of overeting during the first 5 yers of life (9). In retrospective study of young dults, overeting ws more common in femles who remembered their mother dieting (10). The objective of the present study ws to exmine the longitudinl ssocition of mternl ttempt to lose weight t 6 months postprtum with mesures of child diposity t ge 3 yers, while controlling for mternl weight-relted vribles tht could confound the reltionship. Furthermore, we sought to ssess the extent to which the ssocition differed by type of weight loss strtegy used. Methods nd Procedures Study design nd prticipnts Study subjects were prticipnts in Project Viv, prospective cohort study of pre- nd perintl fctors, pregnncy outcomes, nd offspring helth. Detils of study design nd recruitment re reported elsewhere (11). Prticipnts were recruited t their initil prentl visit from eight urbn nd suburbn obstetric offices in Msschusetts between Deprtment of Medicine, Division of Adolescent Medicine, Children s Hospitl Boston, Boston, Msschusetts, USA; 2 Deprtment of Nutrition, Hrvrd School of Public Helth, Boston, Msschusetts, USA; 3 Obesity Prevention Progrm, Deprtment of Popultion Medicine, Hrvrd Medicl School nd Hrvrd Pilgrim Helth Cre Institute, Boston, Msschusetts, USA; 4 Humn Nutrition Progrm, Deprtment of Environmentl Helth Sciences, University of Michign School of Public Helth, Ann Arbor, Michign, USA; 5 Deprtment of Society, Humn Development, nd Helth, Hrvrd School of Public Helth, Boston, Msschusetts, USA; 6 Deprtment of Medicine, Division of Generl Peditrics, Children s Hospitl Boston, Boston, Msschusetts, USA. Correspondence: Kendrin R. Sonneville (kendrin.sonneville@childrens.hrvrd.edu) Received 10 September 2010; ccepted 13 Jnury 2011; published online 24 Februry doi: /oby VOLUME 19 NUMBER 10 OCTOBER

2 nd Of the 2,128 women who delivered live singleton infnt, 1,579 were eligible for 3-yer follow-up by virtue of hving completed prentl nutritionl ssessment nd consenting for their children to be followed up. From this group, we excluded 321 prticipnts who hd missing mesured child height nd weight t ge 3 yers nd 214 who hd missing dt for mternl ttempt to lose weight ssessed t 6 months postprtum. Thus, our smple size ws 1,044 mother infnt pirs. When compred with the 535 who were eligible for but not included in this nlysis, mothers in the present study were more likely to be white (77% vs. 53%), hve t lest college eduction (75% vs. 53%), nd hve n nnul household income $ (68% vs. 49%); prticipnts did not differ in birth weight or men mternl prepregnncy BMI. We conducted in-person study visits with both mothers nd children t enrollment, t 6 months postprtum, nd t 3 yers postprtum. We obtined sociodemogrphic dt nd other prticipnt chrcteristics from study visit interviews, self-dministered questionnires completed t study visits, nd through miled yerly self-dministered questionnires. We mesured length nd weight t the 6-month study visit nd mesured height nd weight t the 3-yer visit. The study ws pproved by the institutionl review bords of Hrvrd Pilgrim Helth Cre, Brighm nd Women s Hospitl, nd Beth Isrel Deconess Medicl Center. Mesures Min exposure mternl ttempt to lose weight t 6 months postprtum. We ssessed mternl ttempt to lose weight bsed on yes/no response to the question re you ctively trying to lose weight? sked vi questionnire t 6 months postprtum. Women who provided yes response to this question were sked to select the types of weight loss strtegies they were using from the following options: chnging eting hbits, exercising, brestfeeding, using diet pills, bottle feeding, purging, smoking, nd/or other. We lso strtified this vrible into ctegories, which represented the most common types of weight loss strtegies used in the smple: (i) trying to lose weight by exercising only, (ii) trying to lose weight by chnging eting hbits only, nd (iii) trying to lose weight by both exercising nd chnging eting hbits. Outcome mesures child diposity t ge 3 yers. We mesured height nd weight of children using clibrted stdiometer (Shorr Productions, Olney, MD) nd scle (Sec model 881; Sec, Hnover, MD). We clculted ge- nd sex-specific weight-for-length z-scores t 6 months nd BMI percentiles nd z-scores t 3 yers using US reference growth dt (12). We defined obesity s BMI for ge nd sex 95th percentile nd overweight s BMI for ge nd sex between the 85th nd 95th percentiles for ge nd sex (12), nd used BMI <85th percentile s the comprison. Reserch ssistnts performing ll mesurements followed stndrdized techniques (13), nd prticipted in binnul in-service trining to ensure mesurement vlidity (Shorr Productions). Inter- nd intrrter mesurement error were well within published reference rnges for ll mesurements (14). Other mesures. Mothers reported informtion bout mternl ge, height, prepregnncy weight, rce/ethnicity, eduction (some college or less, college grdute or more), nnul household income (<$70,000, $70,000, missing), durtion of brestfeeding, nd pternl height nd weight. We clculted totl gesttionl weight gin by subtrcting prepregnncy weight from the lst prentl weight. We report gesttionl weight gin s continuous vrible nd s ctegoricl vrible clssified s indequte, dequte, or excessive weight gin bsed on 2009 Institute of Medicine (IOM) recommendtions (15). IOM gesttionl weight gin recommendtions re kg for underweight women (BMI <18.5), kg for norml weight women (BMI ), kg for overweight women (BMI ), nd 5 9 kg for obese women (BMI 30.0). Indequte weight gin nd excessive weight gin re defined s gesttionl weight gin tht flls below or bove recommended rnges, respectively. Sttisticl nlysis We exmined the bivrite ssocitions between mternl ttempt to lose weight t 6 months postprtum with our min outcomes (BMI z-score, overweight, nd obesity) nd other covrites. We used liner nd logistic regression to exmine the independent ssocition of mternl ttempt to lose weight with child BMI z-score, overweight, nd obesity t 3 yers of ge (model 1). Adjusted multivrite models included child chrcteristics (ge, sex, birthweight, brestfeeding durtion, weight-for-length z-score t 6 months) nd prentl nd household chrcteristics (mternl ge, rce/ethnicity, eduction; pternl BMI; household income) (model 2). Mternl obesity before (16) nd during (17) pregnncy is ssocited with offspring obesity in erly childhood nd excessive gesttionl weight gin is lso ssocited with higher child obesity t ge 3 yers (18). Accordingly, we djusted our models for mternl weight-relted vribles which my confound the ssocition between mternl ttempt to lose weight during the postprtum period nd child weight t 3 yers of ge. We individully dded mternl weight-relted vribles to the djusted models: prepregnncy BMI (model 3), postprtum weight retention t 6 months (model 4), IOM gesttionl weight gin ctegory (model 5), mternl BMI t 6 months postprtum (model 6), nd chnge in mternl BMI from 6 months to 3 yers postprtum (model 7). To exmine whether ssocitions differed by type of weight loss strtegy, we rn nlogous models using 4-ctegory exposure vrible representing mternl weight loss strtegy used. We report regression estimtes (β) or odds rtios nd 95% confidence intervls (CIs) for the min outcomes. We conducted ll of the nlyses using SAS version 9.1 (SAS Institute, Cry, NC). Results The men ge of women ws 32.7 (4.7) yers. Men gesttionl weight gin ws 15.6 (5.3) kg, with most women (59%) gining n mount of weight tht is considered excessive ccording to IOM recommendtions. Among the children t ge 3 yers, the men (s.d.) BMI z-score ws 0.44 (1.02). At 3 yers, 275 (26%) of children were either overweight or obese. Other prticipnt chrcteristics re shown in Tble 1. Of the 1,044 women comprising the smple, 557 (53%) indicted they were ttempting to lose weight t 6 months postprtum. Weight loss strtegies used most frequently by these women included chnging eting hbits (83%) nd exercising (77%). Of the women who were ttempting to lose weight, 83 (15%) were trying to lose weight through exercise only, 116 (21%) were trying to lose weight by chnging eting hbits only, 346 (62%) were trying to lose weight using combined pproch tht involved both exercising nd chnging eting hbits. The remining 12 (2%) women were ttempting to lose weight using strtegy other thn exercising or chnging eting hbits. Becuse of the smll number of women who were ttempting to lose weight using strtegy other thn exercising or chnging eting hbits, these women were not included nlyses strtified by type of weight loss strtegy used. Prticipnt chrcteristics strtified by whether or not mothers were ttempting to lose weight t 6 months postprtum re shown in Tble 1. Women who were trying to lose weight t 6 months postprtum hd higher prepregnncy BMIs, greter gesttionl weight gin, were more likely to hve excessive gesttionl weight gin, nd were hevier t 6 months postprtum thn those who were not. Chnge in BMI from 6 months to 3 yers postprtum ws similr in women who were trying obesity VOLUME 19 NUMBER 10 OCTOBER

3 Tble 1 Descriptive sttistics nd bivrite ssocition of mternl ttempt to lose weight with prentl/household nd child chrcteristics (n = 1,044) Prentl/household chrcteristics Entire smple Mternl ttempt to lose weight t 6 months postprtum Yes (n = 557, 53%) No (n = 487, 47%) P vlue Mternl ge (yers) 32.7 (4.7) 32.7 (4.7) 32.8 (4.7) 0.62 Prepregnncy BMI (kg/m 2 ) 24.4 (5.0) 25.6 (5.1) 23.0 (4.6) < Gesttionl weight gin (kg) 15.6 (5.3) 16.4 (5.6) 14.8 (4.8) < Postprtum weight retention (kg) t 6 months 3.3 (5.1) 4.0 (5.8) 2.4 (4.1) < BMI (kg/m 2 ) t 6 months postprtum 26.0 (5.4) 27.7 (5.2) 24.0 (4.9) < Chnge in BMI (kg/m 2 ) from 6 months to 3 yers postprtum 0.04 (2.46) 0.17 (2.84) 0.10 (1.94) 0.10 Pternl BMI (kg/m 2 ) 26.4 (3.8) 26.8 (3.9) 26.0 (3.7) <0.001 Mternl rce/ethnicity 0.20 White 803 (77%) 433 (78%) 370 (76%) Blck 96 (9%) 46 (8%) 50 (10%) Hispnic 53 (5%) 34 (6%) 19 (4%) Other 91 (9%) 44 (8%) 47 (10%) Mternl eduction <0.01 Some college 261 (25%) 159 (29%) 102 (21%) College grdute 782 (75%) 398 (71%) 384 (79%) Household income 0.20 <$70, (30%) 165 (30%) 151 (31%) $70, (65%) 358 (64%) 318 (65%) Don t know/missing 52 (5%) 34 (6%) 18 (4%) Gesttionl weight gin ctegory < Indequte 129 (13%) 44 (8%) 85 (18%) Adequte 295 (29%) 125 (23%) 170 (35%) Excessive 604 (59%) 378 (69%) 226 (47%) Child chrcteristics Age t 3 yer visit (yers) 3.2 (0.3) 3.2 (0.3) 3.3 (0.3) 0.20 Birthweight (kg) 3.5 (0.6) 3.5 (0.6) 3.5 (0.5) 0.57 Brestfeeding durtion (months) 6.5 (4.5) 5.9 (4.4) 7.3 (4.5) < WFL z-score t 6 months 0.71 (0.96) 0.78 (0.99) 0.64 (0.93) 0.04 Sex Mle 521 (50%) 232 (45%) 289 (55%) 0.17 BMI percentile t 3 yers < <5th 24 (2%) 10 (2%) 14 (3%) 5th <85th 745 (71%) 374 (67%) 371 (76%) 85th <95th (overweight) 182 (17%) 104 (19%) 78 (16%) 95th percentile (obese) 93 (9%) 69 (12%) 24 (5%) WFL, weight-for-length. P vlues from t-test or χ 2. to lose weight t 6 months postprtum (men (s.d.)) ( 0.17 (2.84) kg) nd those who were not (0.10 (1.94) kg). As shown in Tble 2, women who were trying to lose weight t 6 months postprtum by chnging eting hbits only hd greter reduction in BMI from 6 months to 3 yers postprtum ( 0.6 (2.9) kg) thn women who were not trying to lose weight by exercising or chnging eting hbits (0.1 (1.9) kg). BMI from 6 months to 3 yers did not differ pprecibly between women 2048 VOLUME 19 NUMBER 10 OCTOBER

4 Tble 2 Bivrite ssocition of mternl weight loss strtegies used t 6 months postprtum with prentl/household nd child chrcteristics (n = 1,044) Prentl/household chrcteristics None b (n = 499, 48%) Mternl weight loss strtegies Exercise c (n = 83, 8%) Et d (n = 116, 11%) Exercise + Et e (n = 346, 33%) P vlue Prepregnncy BMI (kg/m 2 ) 23.1 (4.6) 23.7 (4.5) 26.0 (4.4) 26.0 (5.4) < Gesttionl weight gin (kg) 14.9 (4.8) 15.8 (4.3) 16.5 (5.4) 16.4 (6.0) Postprtum weight retention (kg) t 6 months 2.3 (4.1) 3.2 (4.4) 5.2 (5.5) 4.0 (6.1) < BMI (kg/m 2 ) t 6 months postprtum 24.0 (4.8) 25.6 (4.5) 29.0 (5.0) 27.9 (5.3) < Chnge in BMI (kg/m 2 ) from 6 months to 3 yers postprtum 0.1 (1.9) 0.2 (1.8) 0.6 (2.9) 0.1 (3.0) <0.05 Pternl BMI (kg/m 2 ) 26.0 (3.7) 26.0 (4.0) 26.9 (4.0) 27.0 (3.9) <0.001 Mternl rce/ethnicity <0.01 White 380 (76%) 63 (76%) 78 (67%) 282 (82%) Blck 50 (10%) 4 (4%) 11 (9%) 31 (9%) Hispnic 19 (4%) 8 (10%) 11 (9%) 15 (4%) Other 49 (10%) 8 (10%) 16 (14%) 18 (5%) Mternl eduction <0.01 Some college 105 (21%) 19 (23%) 42 (36%) 95 (27%) College grdute 393 (79%) 64 (77%) 74 (64%) 251 (73%) Household income 0.09 <$70, (31%) 16 (19%) 44 (38%) 102 (29%) $70, (65%) 62 (75%) 66 (57%) 222 (64%) Don t know/missing 19 (4%) 5 (6%) 6 (5%) 22 (6%) Gesttionl weight gin ctegory < Indequte 86 (17%) 9 (11%) 9 (8%) 25 (7%) Adequte 173 (35%) 22 (27%) 20 (18%) 80 (24%) Excessive 234 (47%) 51 (62%) 84 (74%) 235 (69%) Child chrcteristics Birthweight (kg) 7.3 (4.5) 7.2 (4.4) 5.9 (4.4) 5.5 (4.3) < WFL z-score t 6 months 0.27 (0.95) 0.57 (1.17) 0.56 (1.13) 0.61 (0.99) < BMI percentile t 3 yers <0.01 <5th 14 (3%) 3 (4%) 2 (2%) 5 (1%) 5th <85th 377 (76%) 51 (61%) 79 (68%) 238 (69%) 85th <95th (overweight) 82 (16%) 17 (20%) 19 (16%) 64 (19%) 95th percentile (obese) 26 (5%) 12 (14%) 16 (14%) 39 (11%) WFL, weight-for-length. P vlues from t-test or χ 2. b Not trying to lose weight by chnging eting hbits or exercising. c Trying to lose weight by exercising only. d Trying to lose weight by chnging eting hbits only. e Trying to lose weight by exercising nd chnging eting hbits. who were trying to lose weight by exercising (0.2 (1.8) kg) nd those who were not trying to lose weight by exercising or chnging eting hbits ( 0.1 (3.0) kg). As shown in Tble 3, mternl ttempt to lose weight t 6 months postprtum ws ssocited with higher BMI z-score t child ge 3 yers (β = 0.30 (95% CI 0.18, 0.42)) nd with elevted odds of obesity (3.0 (95% CI 1.6, 5.8)) in covritedjusted models. Becuse both mternl ttempt to lose weight nd child weight sttus re relted to mother s own weight sttus, we rn dditionl models tht included severl mternl weight-relted vribles. Adding prepregnncy BMI (model 3) ttenuted the covrite-djusted ssocitions of mternl ttempt to lose weight with BMI z-score (β = 0.24 (95% CI 0.12, 0.36)) by ~10% nd with odds of obesity (2.4 (95% CI 1.2, 4.7)) obesity VOLUME 19 NUMBER 10 OCTOBER

5 Tble 3 Associtions of mternl ttempt to lose weight t 6 months postprtum with child diposity (n = 1,044) Model BMI z-score β (95% CI) Child diposity outcomes t ge 3 yers Overweight (BMI 85th 94th percentile) OR (95% CI) Obesity (BMI 95th percentile) Model 1. Undjusted 0.35 (0.22, 0.47) 1.3 (1.0, 1.9) 2.9 (1.8, 4.7) Model 2. Adjusted 0.30 (0.18, 0.42) 1.5 (1.0, 2.3) 3.0 (1.6, 5.8) Model 3. M2 + prepregnncy BMI 0.24 (0.12, 0.36) 1.5 (1.0, 2.3) 2.4 (1.2, 4.7) Model 4. M2 + postprtum weight retention t 6 months 0.31 (0.19, 0.43) 1.6 (1.1, 2.5) 3.1 (1.6, 6.0) Model 5. M2 + IOM gesttionl weight gin ctegory 0.30 (0.18, 0.42) 1.5 (1.0, 2.4) 3.1 (1.5, 6.0) Model 6. M2 + mternl BMI t 6 months postprtum 0.22 (0.09, 0.35) 1.5 (1.0, 2.4) 2.3 (1.2, 4.6) Model 7. M2 + chnge in mternl BMI from 6 months to 3 yers postprtum 0.32 (0.20, 0.44) 1.6 (1.1, 2.5) 3.1 (1.6, 6.1) CI, confidence intervl; IOM, Institute of Medicine; OR, odds rtio. Adjusted for child chrcteristics (ge, sex, birthweight, brestfeeding durtion, WFL z-score t 6 months) nd prentl/household chrcteristics (mternl ge, rce/ ethnicity, eduction; pternl BMI; household income). Tble 4 Associtions of type of mternl weight loss strtegy used t 6 month postprtum with child diposity (n = 1,032) Model Model 1 Model 2 Model 3 Undjusted Adjusted prepregnncy BMI M2 + BMI z-score (β (95% CI)) None b (ref.) Exercise c 0.32 (0.08, 0.55) 0.39 (0.16, 0.61) 0.36 (0.14, 0.58) Et d 0.30 (0.10, 0.50) 0.27 (0.06, 0.47) 0.19 ( 0.02, 0.39) Exercise + et e 0.36 (0.22, 0.50) 0.27 (0.14, 0.41) 0.21 (0.08, 0.35) Overweight (OR (95% CI)) None (ref.) Exercise 1.6 (0.9, 2.8) 1.9 (0.9, 4.4) 1.9 (0.8, 4.5) Et 1.2 (0.7, 2.0) 1.5 (0.8, 3.1) 1.5 (0.7, 3.0) Exercise + et 1.3 (0.9, 1.9) 1.4 (0.9, 2.2) 1.4 (0.8, 2.2) Obesity (OR (95% CI)) None (ref.) Exercise 3.6 (1.7, 7.5) 6.5 (2.4, 17.5) 6.0 (2.2, 16.5) Et 3.2 (1.6, 6.2) 2.2 (0.8, 6.0) 1.5 (0.5, 4.4) Exercise + et 2.6 (1.5, 4.4) 2.5 (1.2, 5.1) 1.9 (0.9, 4.0) CI, confidence intervl; OR, odds rtio; WFL, weight-for-length. Adjusted for child chrcteristics (ge, sex, birthweight, brestfeeding durtion, WFL z-score t 6 months) nd prentl/household chrcteristics (mternl ge, rce/ethnicity, eduction; pternl BMI; household income). b Not trying to lose weight by chnging eting hbits or exercising. c Trying to lose weight by exercising only. d Trying to lose weight by chnging eting hbits only. e Trying to lose weight by exercising nd chnging eting hbits. by 20%. Similrly, dding mternl BMI t 6 months (model 6) ttenuted the covrite-djusted ssocition between mternl ttempt to lose weight nd BMI z-score (β = 0.22 (95% CI 0.09, 0.35)) nd odds of obesity (2.3 (95% CI 1.2, 4.6)). Similrities in the ssocitions seen in models 3 nd 6 were expected to given the high correltion between mternl BMI before pregnncy nd t 6 months postprtum (r = 0.93). Adding postprtum weight retention (model 4), IOM gesttionl weight gin ctegory (model 5), or chnge in mternl BMI from 6 months to 3 yers postprtum (model 7) s covrite did not substntilly ttenute the ssocition between mternl ttempt to lose weight nd ny of the mesures of diposity t child ge 3 yers. The ssocition between the type of weight loss strtegy used by mothers t 6 months postprtum nd mesures of child diposity t 3 yers of ge in undjusted models (model 1), covrite-djusted models (model 2), nd covrite-djusted models including mternl prepregnncy BMI (model 3) re shown in Tble 4. In model 3, women who were trying to lose weight by exercising only, compred with women who were not trying to lose weight, hd children with higher BMI z-scores (0.36 (95% CI 0.14, 0.58)) nd odds of obesity (6.0 (95% CI 2.2, 16.5)) t ge 3 yers. In model 3, mternl ttempt to lose weight by chnging eting hbits only ws not significntly ssocited with BMI z-score (0.19 (95% CI 0.02, 0.39)), odds of overweight or odds of obesity (1.5 (95% CI 0.5, 4.4)) in children t ge 3 yers. Also in model 3, mternl ttempt to lose weight by both exercising nd chnging eting hbits predicted higher BMI z-score (0.21 (95% CI 0.80, 0.35)), but not higher odds of obesity (1.9 (95% CI 0.9, 4.0)). Models djusted for postprtum weight retention t 6 months (model 4), IOM gesttionl weight gin ctegory (model 5), mternl BMI t 6 months postprtum (model 6), nd chnge in mternl BMI from 6 months to 3 yers postprtum (model 7) produced similr trends such tht only ttempting to lose weight through exercise lone consistently predicted higher BMI z-score nd odds of overweight nd obesity in children t ge 3 yers. Discussion In this prospective cohort, mternl ttempt to lose weight t 6 months postprtum ws ssocited with higher child BMI z-score, nd elevted odds of overweight nd obesity t child ge 3 yers. Although hevier women hd hevier children nd 2050 VOLUME 19 NUMBER 10 OCTOBER

6 were lso more likely to be trying to lose weight t 6 months postprtum, confounding by mother s weight sttus does not entirely explin the observed reltionship. Adjusting for mternl prepregnncy BMI or BMI t 6 months postprtum ttenuted the ssocitions by 10 20%, suggesting tht some, but not ll, of the observed ssocitions re due to mother s own weight sttus. Women with excessive gesttionl weight gin nd those with more postprtum weight retention were more likely to ttempt to lose weight t 6 months postprtum. However, controlling for these vribles did not substntilly ttenute the observed ssocitions between mternl ttempt to lose weight t 6 months postprtum nd child diposity t ge 3 yers. We observed higher offspring diposity t ge 3 yers mong women who exercised to lose weight t 6 months postprtum. Children of mothers in the exercise only nd diet plus exercise groups hd higher BMI z-score nd greter odds of overweight t ge 3 yers. Mternl weight loss strtegies tht involved chnging diet lone did not predict higher child diposity. To our knowledge, this is the first study to investigte nd report n ssocition between mternl ttempt to lose weight during the postprtum period nd offspring weight outcomes. The pthwy through which mother s ttempt to lose weight during the postprtum period could influence her child s weight remins to be estblished. Although previous studies hve observed n ssocition between mternl dieting nd child overeting (9,10), we did not find tht mternl ttempt to lose weight by chnging eting hbits ws ssocited with elevted child weight. Rther mternl ttempts tht included exercise were most predictive of child diposity, mking lterntive explntions more likely. Ongoing reserch is needed to explore mechnisms which could explin this ssocition. Future studies should exmine how the home food environment, dietry intke, eting styles, nd feeding hbits of women differ bsed on the type of weight loss strtegy they choose to dopt. Furthermore, dditionl reserch should exmine how exercise type, durtion, nd frequency impcts energy blnce mong postprtum women. Becuse of the known dverse consequences of postprtum weight retention, dditionl reserch is needed to determine the best weight loss strtegies to recommend. Pst reserch indictes tht dvice should be fmily centered (19), encourge brestfeeding (20,21), promote chnge in food intke (20), exercise frequency (20), or both (22), or should ddress mternl self-efficcy (23). Our findings suggest tht postprtum weight loss strtegies tht include chnging eting hbits my be most beneficil to both mother nd offspring. We observed tht women who reported trying to lose weight t 6 months postprtum by chnging eting hbits were most successful in reducing weight t 3 yers postprtum nd tht women who reported trying to lose weight by exercising only did not reduce their BMI from 6 months to 3 yers postprtum. This observtion supports recent findings, which suggest tht, in the bsence of cloric restriction, exercise is not n effective weight loss strtegy, prticulrly mong overweight women (24). Experimentl studies re need to exmine the impct of vrious types of postprtum weight loss strtegies on both mternl nd child weight. Our study hs severl strengths, including reltively lrge smple size, use of longitudinl dt, nd bility to djust for lrge number of covrites including mternl weight-relted vribles which could confound the reltionship between mternl ttempt to lose weight nd child diposity. Our study is subject to limittions. First, unmesured confounding my bis our findings. We were not ble to control for pst eting nd exercise hbits or long-term weight loss ttempts by the mother, lthough we expect tht mother s weight control behviors t 6 months postprtum re highly correlted with these behviors both prepregnncy nd throughout the offspring s first 3 yers of life. The use of mostly self-reported mesures, including prepregnncy weight, could lso hve introduced bis. Furthermore, we ssessed mternl ttempt to lose weight t 6 months postprtum using single item. Multi-item scles of dieting my be more predictive of behviors thn generl single-item questions (25). We sked women to indicte the generl types of weight loss strtegies they were using, but did not obtin detiled ssessment of weight control behviors used by the women in our smple. It is likely tht substntil individul vribility in behviors exists within ech weight loss strtegy ctegory nd could vry by mternl chrcteristics such s socioeconomic bckground nd weight sttus (25). Accordingly, future studies should collect detiled informtion bout the weight loss strtegies nd behviors used by postprtum mothers. Future studies could lso explore pthwys through which mternl weight loss strtegies influence behviors nd weight sttus of children. Acknowledgments The bstrct of this mnuscript ws published s prt of the proceedings of the 2009 Annul Meeting of The Obesity Society. This study ws supported in prt by grnts from the Centers for Disese Control nd Prevention, U.S. Ntionl Institutes of Helth (HD 34568, HL 64925, HL 68041). Disclosure The uthors declred no conflict of interest The Obesity Society REFERENCES 1. Stuebe AM, Oken E, Gillmn MW. Associtions of diet nd physicl ctivity during pregnncy with risk for excessive gesttionl weight gin. Am J Obstet Gynecol 2009;201:58.e1 58.e8. 2. Mmun AA, Kinrivl M, O Cllghn MJ et l. Associtions of excess weight gin during pregnncy with long-term mternl overweight nd obesity: evidence from 21 y postprtum follow-up. Am J Clin Nutr 2010;91: Villmor E, Cnttingius S. Interpregnncy weight chnge nd risk of dverse pregnncy outcomes: popultion-bsed study. Lncet 2006;368: Ferrri R, Sieg-Riz A, Evenson K, Moos M, Melvin C, Herring A. Provider dvice bout weight loss nd physicl ctivity in the postprtum period. J Womens Helth 2010;19: O Toole ML, Swicki MA, Artl R. Structured diet nd physicl ctivity prevent postprtum weight retention. J Womens Helth (Lrchmt) 2003;12: McCrory MA, Nommsen-Rivers LA, Molé PA, Lönnerdl B, Dewey KG. Rndomized tril of the short-term effects of dieting compred with dieting plus erobic exercise on lcttion performnce. Am J Clin Nutr 1999;69: obesity VOLUME 19 NUMBER 10 OCTOBER

7 7. Loveldy CA, Grner KE, Moreno KL, Willims JP. The effect of weight loss in overweight, lctting women on the growth of their infnts. N Engl J Med 2000;342: Dewey KG, Loveldy CA, Nommsen-Rivers LA, McCrory MA, Lönnerdl B. A rndomized study of the effects of erobic exercise by lctting women on brest-milk volume nd composition. N Engl J Med 1994;330: Stice E, Agrs WS, Hmmer LD. Risk fctors for the emergence of childhood eting disturbnces: five-yer prospective study. Int J Et Disord 1999;25: Brunstrom JM, Mitchell GL, Bguley TS. Potentil erly-life predictors of dietry behviour in dulthood: retrospective study. Int J Obes (Lond) 2005;29: Gillmn MW, Rich-Edwrds JW, Rifs-Shimn SL et l. Mternl ge nd other predictors of newborn blood pressure. J Peditr 2004;144: Ntionl Center for Helth Sttistics. CDC Growth Chrts, United Sttes. < (2000). 13. Shorr IJ. How to Weigh nd Mesure Children. U.N.: New York, Mueller WH, Mrtorell R. Relibility nd ccurcy of mesurement. In: Lohmn TG, Roche AF, Mrtorell R (eds). Anthropometric Stndrdiztion Reference Mnul. Humn Kinetics Books: Chmpign, IL, Rsmussen KM, Yktine AL. Committee to Reexmine IOM Pregnncy Weight Guidelines. Institute of Medicine, Ntionl Reserch Council. The Ntionl Acdemies Press: Wshington, D.C Slsberry PJ, Regn PB. Dynmics of erly childhood overweight. Peditrics 2005;116: Whitker RC. Predicting preschooler obesity t birth: the role of mternl obesity in erly pregnncy. Peditrics 2004;114:e29 e Oken E, Tvers EM, Kleinmn KP, Rich-Edwrds JW, Gillmn MW. Gesttionl weight gin nd child diposity t ge 3 yers. Am J Obstet Gynecol 2007;196:322.e1 322.e Setse R, Grogn R, Cooper LA et l. Weight loss progrms for urbn-bsed, postprtum Africn-Americn women: perceived brriers nd preferred components. Mtern Child Helth J 2008;12: Olson CM, Strwdermn MS, Dennison BA. Mternl weight gin during pregnncy nd child weight t ge 3 yers. Mtern Child Helth J 2009;13: Bker JL, Gmborg M, Heitmnn BL et l. Brestfeeding reduces postprtum weight retention. Am J Clin Nutr 2008;88: Amorim AR, Linne YM, Lourenco PM. Diet or exercise, or both, for weight reduction in women fter childbirth. Cochrne Dtbse Syst Rev 2007;CD Hinton PS, Olson CM. Postprtum exercise nd food intke: the importnce of behvior-specific self-efficcy. J Am Diet Assoc 2001;101: Lee IM, Djoussé L, Sesso HD, Wng L, Buring JE. Physicl ctivity nd weight gin prevention. JAMA 2010;303: Neumrk-Sztiner D, Jeffery RW, French SA. Self-reported dieting: how should we sk? Wht does it men? Associtions between dieting nd reported energy intke. Int J Et Disord 1997;22: VOLUME 19 NUMBER 10 OCTOBER