Abstract The prevalence of the isolated ultrasonographic finding of polycystic ovaries (PCO) in the Finnish population and among women with a history of gestational diabetes (GDM) and changes in the present carbohydrate metabolism were investigated in the present study. One aim of this study was to investigate the prevalence of the recently discovered variant type LH (v-LH) in PCOS and to compare patient cohorts from Finland, the Netherlands, the United Kingdom and the United States of America. In addition, this study attempted to evaluate the nature of the ovarian streoidogenic response of women with PCOS to exogenously administered human chorionic gonadotrophin (hCG), human menotrophin (hMG) and follicle stimulating hormone (FSH). The effect of metformin on ovarian steroidogenesis was also studied. The prevalence of PCO was significantly higher in younger (≤ 35 years, 21.6%) than among older women (in ≥ 36 years, 7.8%). The overall prevalence of PCO in Finnish women was 14.2%. Women with previous GDM revealed a high prevalence of PCO (39.4%). The carrier frequency of the v-LHb allele in the entire study population was 18.5%. The frequency of the v-LH carrier was significantly lower in obese PCOS subjects in the Netherlands (2.0%) and Finland (4.5%). Women with previous GDM had impaired insulin sensitivity and β-cell function. They also had higher adrenal androgen secretion than the control women. Women with PCO and previous GDM had marked hyperinsulinemia which was not explained by obesity. Obese PCOS women achieved peak peripheral serum T concentrations at 48 hours after a hCG injection, preceded by peak levels of 17-OHP and E2 at 24 hours. In contrast, all steroids measured in the control women reached their maximum serum concentrations at 96 hours. HMG stimulated the production of ovarian androgens more efficiently than a urinary FSH after pituitary suppression with a gonadotrophin releasing hormone agonist (GnRHa). In conclusion, the prevalence of PCO is common in healthy Finnish women and even more common in women with a history of GDM. The ultrasonographic appearance of PCO may be a predictive factor with regards abnormal glucose tolerance during and after pregnancy and, these women should therefore be advised as to possible consequences. The high overall frequency of the v-LH allele in women in general and its low frequency in obese PCOS patients suggests that v-LH plays a role in reproductive functions and may counteract the pathogenesis of PCOS in obese individuals. The differences observed in steroid responses to hCG between normal and PCOS women might be explained by higher theca cell activity or mass in polycystic ovaries. Women with PCOS did not show a distinctly exaggerated steroidogenic response to hMG or FSH administration compared with control women. FSH administration also resulted in increased A and T production.

Gestational diabetes mellitus (GDM) is a carbohydrate metabolism issue during pregnancy that is dangerous for mother and the baby. GDM occurs in 1 out of 3 diabetic women in 16.2% of live births. GDM knowledge and treatment practices among nurses were found inadequate when nurses’ effectiveness in treating a disease they have a shallow knowledge about (GDM) was investigated in the local medical facility. A GDM instructional module was applied and its effectiveness in promoting nurse’s use of GDM education as a treatment strategy tested. The total concept for knowledge and care, empowerment and the social cognitive theories grounded this research. Methodology was Mixed. A population/patient problem-intervention-comparison-outcome-time (PICOT) design was applied in the analysis of data from a sample size {n=40}, whereby the treatment group (TG=20) had an intervention, and control group (CG=20) did not. Data was analyzed descriptively and inferentially with t-test statistic, including the Cohen’s d test for effect size. Evidence showed a significantly high post-intervention gain in scores CG and TG, higher among DNPs than other nurses. Also, the Cohen’s d test indicated high magnitude effect size. Overall confidence in GDM treatment method improved. A comparison of mean test completion time and scores indicated that TG completed the posttest at a shorter time than CG. Knowledge improvement results were TG 27%; CG 2%. GDM education is an effective path to positive social change, beneficial to nurses, the medical facility and the community. Improved GDM treatment means a healthier population and increased productivity for the community. GDM education is non-medicated and more affordable - a huge savings for the community.

A thesis submitted to the Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Doctor of Philosophy December 2017. / Gestational diabetes mellitus (GDM) is characterized by high blood glucose levels that first develop during pregnancy. GDM has been linked with many adverse short and long term health outcomes for the developing foetus as well as for the mother. The Developmental Origins of Health and Disease (DOHaD) concept suggests that in the presence of adverse stimuli, the foetus will adapt, through epigenetic mechanisms, to ensure its immediate survival. For this reason, epigenetic modifications are emerging as mediators linking early environmental exposures during pregnancy with programmed changes in gene expression that alter offspring growth and development. The objective of this research study was to explore the role of altered gene expression and methylation in the development of GDM and determine whether these alterations are inherited by the exposed foetus. Transcriptome sequencing was performed on mRNA extracted from blood samples collected from six women with GDM and from six controls; as well as from exposed (N=6) and unexposed placenta (N=6). Genes that displayed significant (p<0.005) differential expression (log2 fold change >2 and <-2) between cases and controls were identified from the blood (N=60) and placenta (N=56) datasets. Gene ontology and enrichment was performed using DAVID and PANTHER with the aim to narrow down the candidate gene lists. The ten most likely candidate genes for differential gene expression from the blood dataset were G6PD, DCXR, TKT, ALDOA, PGLS, KCNQ1, C14orf80, KCNQ1, SLC25A22 and GSK3A. Gene enrichment revealed that five of these significantly under-expressed genes (G6PD, DCXR, TKT, ALDOA and PGLS) encode enzymes in the pentose phosphate pathway (PPP). In the placental dataset the top ten candidate genes were CXCR1, CXCR2, G6PD, TKT, IGFBP-1, IGFBP-2, IGFBP-6, GGT3P, MMP12 and GLT1D1. The direction and fold change of differential expression of all twenty genes were validated using TaqMan qPCR probes. Of these twenty genes, the five most promising biological candidates (G6PD, TKT, IGFBP-1, IGFBP-2 and IGFBP-6) were identified and the level of promoter region methylation was assessed using EpiTech Methyl II PCR Assays. The level of methylation in the promoter region of G6PD in both blood and placenta tissue was found to be significantly higher (p=1.90 x 10-5 and p=1.2 x 10-11 respectively) in the case groups, correlating with decreased mRNA expression levels. There was a significant negative correlation between G6PD mRNA expression in the blood and placenta with the level of maternal glucose at fasting (p=0.006 and p=0.001, respectively), 1-hr (p=0.016 and p=0.007, respectively) and 2-hr post OG (p=0.045 in placenta). We observed a significant positive correlation between G6PD promoter region methylation in both blood and placental tissues with maternal glucose levels at fasting (p=0.023 and p=0.001, respectively) and at 1-hr post OG (p=0.001 and p=0.004, respectively). IGFBP-1 was found to be significantly under-expressed in exposed placental tissue and hypermethylated (p=1.1 x 10-6) at the promoter region when compared to unexposed samples. There was a significant negative correlation between the expression of IGFBP-1 mRNA in the blood and placenta with foetal birth weight (p=0.005 and p=0.017, respectively). Our results suggest that high glucose levels, an important characteristic of GDM, result in the disturbance of the pentose phosphate pathway, a pathway linked closely to glycolysis, and the IGF-axis, which is important in foetal growth and development. In GDM there is suppression of G6PD mRNA expression in both the blood and placental tissue which influences the pentose phosphate pathway. We hypothesize that this is mediated through an epigenetic mechanism since it is correlated with increased methylation of the G6PD promoter region. Down regulation of G6PD would suppress the PPP and reduce the levels of NADPH production, which may in turn lead to an increase in oxidative stress and an adverse outcome in the mother and foetus. With regard to the IGF-axis, our results demonstrated that IGFBP-1 and IGFBP-2 mRNA expression in the placenta may be inhibited due to the presence of high glucose and insulin levels and this decrease in mRNA expression is likely implicated in the abnormal foetal growth which is often associated with GDM. This study has provided novel insights into gene expression and DNA methylation changes in the blood of women with GDM and the placenta of their female offspring that involve genes in the PPP and the IGF-axis. / LG2018

Gestational diabetes mellitus impacts between 3-10% of pregnancies, and increases the risk of pregnancy complications and lifelong health effects for mother and child (Bellamy, Casas, Hingorani, & Williams, 2009; Ross, 2006; Ryser Rüetschi et al., 2016). About half of cases occur without an evident risk factor (American College of Obstetricians and Gynecologists (ACOG), 1994; Dode & Santos, 2009). The present study was designed to examine possible psychophysiological connections linking psychological stress and stress reactivity, the magnitude of an individual’s response to stress, to blood sugar metabolization during mid-pregnancy between 24-28 weeks gestation. Participants were recruited from Sanford Health in Fargo, where patients underwent routine Oral Glucose Tolerance Testing (OGTT) a diagnostic assessment in which higher results indicate less blood sugar metabolization. They also completed a Virtual Trier Social Stress Task while psychological and physiological markers of stress reactivity were assessed. Additionally, maternal stress and stress reactivity were assessed using psychosocial questionnaires. There was support for proposed psychophysiological connections, including models in which positive associations between OGTT and maternal stress and anxiety were moderated by psychological stress reactivity. Results suggest that both the presence of stress and a women’s responses to that stress are influential over blood glucose metabolization during pregnancy. Continuing research in this area may have implications for improving outcomes of women at higher risk of GDM and other adverse pregnancy and perinatal outcomes.

Pigs exposed to chronic intrauterine hyperthermia (gHS) experience greater fat deposition during life and yield carcasses with greater fat:lean content at slaughter compared to pigs gestated under thermoneutral conditions (gTN). The objectives of this study were to 1) determine whether gHS impacts the lactational performance of affected gilts (F1 generation), and 2) determine whether these effects of gHS are also evident in the next generation (F2 generation). Twenty-four gilts were bred and exposed to thermoneutral or heat stressed conditions for the entirety of gestation, and F1 female offspring were retained. At puberty, gHS and gTN gilts were bred to farrow in either spring (March / April) or summer (July / August). Colostrum and milk samples were collected at farrowing and on d 7, 14, and 21 of lactation. At weaning, four offspring (two male, two female) were retained and grown to market weight in mixed-pens under identical management conditions. Carcass characteristics were analyzed at slaughter. Milk nutrient analysis indicated that gHS gilts produced less lactose, and tended to produce greater protein, than did gTN gilts. There was no difference in the growth rate of F2 offspring, but pigs born of gHS dams did have a tendency for greater backfat thickness. The patterns of altered milk nutrient content observed in F1 gilts reflects a metabolic profile consistent with previous gHS research, and the greater backfat of F2 pigs at slaughter indicates the adipose-promoting effects of gHS may be diluted, but still evident, in the second generation. / Master of Science

<p><strong>Aim:  </strong>To describe health professional's treatment of pregnant overweight or obese women at an antenatal clinic in Nairobi, Kenya. <strong>Method: </strong>A descriptive qualitative method was used including interviews with 11 health professionals at an antenatal clinic in Nairobi, Kenya. The interviews were semi-structured and consisted of five open-ended question. The result was then analysed through a systematic content analyse. <strong>Result: </strong>Four categories were identified. The methods used to state gestational overweight or obesity were weighting and assessing from the previous weight. BMI was also measured. The most common intervention is counselling on diet and lifestyle. Some respondents mentioned referral to nutritionist and also multidisciplinary involvement. Many of the respondents stated however that gestational overweight and obesity is very rare and that they don't have any specific routines or interventions. <strong>Conclusion:</strong> The basic measures used to survey  pregnant women´s weight are  pre- pregnant weight, continuously weighting and measuring BMI are used. There are also basic interventions such as counselling on diet and exercise. However, gestational overweight or obesity is not a prioritised issue at Kenyatta National Hospital and few routines and interventions exist. This study demonstrates the necessity for training and educating the staff to raise awareness about the risks gestational overweight or obesity implies. It also highlights the need of clear guidelines on how manage gestational overweight and obesity.  </p> / <p><strong>Syfte:</strong> Att beskriva hälso- och sjukvårdspersonals behandling av överviktiga eller feta gravida kvinnor. <strong>Metod:</strong> Beskrivande kvalitativ metod där intervjuer med 11 vårdpersonal utfördes på en mödravårdscentral i Nairobi, Kenya. Intervjuerna var semi- strukturerade och bestod av fem öppna frågor. Resultatet analyseras sedan genom en systematisk innehållsanalys. <strong>Resultat: </strong>Fyra kategorier identifierades. De metoder som användes för att fastställa övervikt eller fetma under graviditet var vägning, jämförelse med tidigare vikt samt BMI. Den vanligaste insatsen var rådgivning om kost och livsstil. Några respondenter nämnde att de vid behov remitterar patienter till en dietist och att de arbetar tvärvetenskapligt. Många av respondenterna uppgav dock att övervikt och fetma under graviditet är mycket ovanligt och att inga särskilda rutiner eller interventioner finns. <strong>Slutsats:</strong> Den vanligaste åtgärden för att förebygga graviditetsdiabetes övervikt och fetma är rådgivning om kost och motion. Kvinnornas vikt, kontinuerligt vägning och mätning av BMI är de grundläggande redskapen för att hålla en god kontroll på gravida kvinnors vikt. Övervikt och fetma hos gravida kvinnor är dock inte en prioriterad fråga på Kenyatta National Hospital och mycket få rutiner och åtgärder finns. Studien belyser behovet av att utbilda personal för att öka medvetenheten om vilka risker övervikt och fetma hos gravida kvinnor kan medföra. Det finns även ett behov av tydliga riktlinjer för hur man identifierar, behandlar och följer upp överviktiga eller feta gravida kvinnor.</p>

Aim:  To describe health professional's treatment of pregnant overweight or obese women at an antenatal clinic in Nairobi, Kenya. Method: A descriptive qualitative method was used including interviews with 11 health professionals at an antenatal clinic in Nairobi, Kenya. The interviews were semi-structured and consisted of five open-ended question. The result was then analysed through a systematic content analyse. Result: Four categories were identified. The methods used to state gestational overweight or obesity were weighting and assessing from the previous weight. BMI was also measured. The most common intervention is counselling on diet and lifestyle. Some respondents mentioned referral to nutritionist and also multidisciplinary involvement. Many of the respondents stated however that gestational overweight and obesity is very rare and that they don't have any specific routines or interventions. Conclusion: The basic measures used to survey  pregnant women´s weight are  pre- pregnant weight, continuously weighting and measuring BMI are used. There are also basic interventions such as counselling on diet and exercise. However, gestational overweight or obesity is not a prioritised issue at Kenyatta National Hospital and few routines and interventions exist. This study demonstrates the necessity for training and educating the staff to raise awareness about the risks gestational overweight or obesity implies. It also highlights the need of clear guidelines on how manage gestational overweight and obesity. / Syfte: Att beskriva hälso- och sjukvårdspersonals behandling av överviktiga eller feta gravida kvinnor. Metod: Beskrivande kvalitativ metod där intervjuer med 11 vårdpersonal utfördes på en mödravårdscentral i Nairobi, Kenya. Intervjuerna var semi- strukturerade och bestod av fem öppna frågor. Resultatet analyseras sedan genom en systematisk innehållsanalys. Resultat: Fyra kategorier identifierades. De metoder som användes för att fastställa övervikt eller fetma under graviditet var vägning, jämförelse med tidigare vikt samt BMI. Den vanligaste insatsen var rådgivning om kost och livsstil. Några respondenter nämnde att de vid behov remitterar patienter till en dietist och att de arbetar tvärvetenskapligt. Många av respondenterna uppgav dock att övervikt och fetma under graviditet är mycket ovanligt och att inga särskilda rutiner eller interventioner finns. Slutsats: Den vanligaste åtgärden för att förebygga graviditetsdiabetes övervikt och fetma är rådgivning om kost och motion. Kvinnornas vikt, kontinuerligt vägning och mätning av BMI är de grundläggande redskapen för att hålla en god kontroll på gravida kvinnors vikt. Övervikt och fetma hos gravida kvinnor är dock inte en prioriterad fråga på Kenyatta National Hospital och mycket få rutiner och åtgärder finns. Studien belyser behovet av att utbilda personal för att öka medvetenheten om vilka risker övervikt och fetma hos gravida kvinnor kan medföra. Det finns även ett behov av tydliga riktlinjer för hur man identifierar, behandlar och följer upp överviktiga eller feta gravida kvinnor.

Gestational diabetes mellitus (GDM) is an increasing problem in the U.S. Many comorbidities are associated with GDM: increased risk for type 2 diabetes, neonatal hypoglycemia and fetal malformation. Healthcare organizations develop GDM educational programs to provide women with knowledge and skills to manage GDM and reduce health risks. While there are significant benefits to attending GDM educational programs, attendance rates are low. Little research has been conducted to determine reasons for low attendance in GDM educational programs. The purpose of this study was to explore the experiences of women with GDM and to describe factors influencing GDM educational program attendance. Semi-structured telephone interviews were conducted with GDM program participants at a large hospital in central Kentucky. The sample size was N=21. Results indicated that meal management changes and blood glucose monitoring characterized the GDM experience and many attended the educational program to receive information on these topics. Few participants reported barriers to attendance. The majority was satisfied with information received. Motivators to attendance included flexibility, location, and support of family members.. Participants preferred face-to-face meetings although some expressed a need for online classes and communication via text messaging. Participants expressed the need for GDM information postpartum.

Introdução: O monitoramento do ganho de peso gestacional é de extrema importância nos cuidados pré-natais, pois pode evitar diversos desfechos desfavoráveis tanto para mãe quanto para o bebê. A maioria dos países utiliza algum tipo de referência para o acompanhamento do peso gestacional. Essas referências, muitas vezes, são baseadas em suas próprias populações ou em populações de outros países. Considerando-se que características populacionais variam de acordo com etnia, localização geográfica, hábitos alimentares, medidas antropométricas e condições socioeconômicas, dentre outros fatores, as recomendações baseadas em populações específicas são preferíveis para monitorar o peso gestacional. Várias metodologias são utilizadas para a construção de referências de peso gestacional. A OMS (Organização Mundial da Saúde) recomenda utilizar estudos longitudinais, a partir de populações selecionadas com baixa prevalência de complicações maternas e fetais. No Brasil, as referências utilizadas para peso gestacional são baseadas em duas populações internacionais; essas populações não utilizaram estudos longitudinais para gerar as referências, o que faz com que o Brasil necessite desenvolver sua própria abordagem para o monitoramento do peso gestacional. Objetivo: Construir uma curva de peso gestacional a partir de uma coorte de gestantes brasileiras, utilizando a metodologia estatística recomendada pela OMS para a construção das curvas padrão de crescimento infantil. Método: Dados do Estudo Brasileiro de Diabetes Gestacional (EBDG), estudo multicêntrico que reuniu gestantes de seis capitais brasileiras (Porto Alegre, Rio de Janeiro, São Paulo, Salvador, Manaus e Fortaleza), foram utilizados para a construção da curva. Foram selecionadas 2.103 gestantes eutróficas, de acordo com IOM (Institute of Medicine) (2009), com bons desfechos gestacionais, ou seja, foram excluídas gestantes com diabetes mellitus gestacional, com distúrbios hipertensivos, com gestações múltiplas e com partos prematuros (<37 semanas de gestação); excluíram-se também casos de gestantes com recém-nascidos de baixo peso ao nascer (BPN) ou com recém-nascidos grandes para a idade gestacional (GIG) e recém-nascidos pequenos para a idade gestacional (PIG), bem como casos de macrossomia. Dentre essas gestantes eutróficas, foram sorteadas 918, que irão fazer parte da validação da curva. Para a construção da curva, ficaram 1.179 gestantes eutróficas com bons desfechos gestacionais. Para o ajuste, foi utilizado o método GAMLSS (Modelos Aditivos Generalizados de Localização, Escala e Forma) do software R, que estimou os percentis 3, 5, 10, 25, 75, 90, 95 e 97. Após algumas exclusões, a segunda etapa do trabalho utilizou as 918 sorteadas (gestantes eutróficas com bons desfechos) e mais 901 gestantes eutróficas que tinham pelo menos um dos desfechos gestacionais indesejáveis descritos acima (com exceção de gestações múltiplas e diabetes mellitus), totalizando 1.817 gestantes para o grupo de validação. Com os percentis estimados pela curva de peso gestacional, foram definidos pontos de corte que determinaram os fatores de risco para os desfechos de interesse. A influência dos fatores de risco sobre os desfechos foi medida através do risco relativo (RR) e seus respectivos intervalos, com 95% de confiança, estimados através de regressão de Poisson com variância robusta. Os riscos relativos e seus respectivos intervalos de 95% de confiança foram estimados para a exposição em algum momento da gestação e para a exposição em algum momento dentro de cada trimestre. Os percentis foram avaliados sem ajustar para nenhum possível fator de confusão. Resultados: Após testados vários modelos GAMLSS, o que melhor ajustou os dados foi o que utilizou a família de distribuição BCPE (Box Cox de Potência Exponencial), com suavizador pb (B-splines), utilizando dois parâmetros e . O percentil 25 estimado foi capaz de predizer baixo peso ao nascer, prematuridade e PIG; já o percentil 75 pôde ser utilizado como preditor de distúrbios hipertensivos, macrossomia e GIG. Conclusão: o modelo obtido para a construção da curva de peso gestacional indicou que a relação entre peso gestacional e idade gestacional não é linear. A flexibilidade da metodologia estatística utilizada no estudo é suficiente para que possa ser aplicada utilizando-se o Índice de Massa corporal (IMC) em vez de peso gestacional. Essa metodologia também apresenta uma série de vantagens no que diz respeito às suas opções de modelagem. As curvas de percentis ajustadas foram eficientes em predizer desfechos gestacionais adversos. A metodologia aplicada nesta tese pode ser replicada para todas as categorias de IMC pré-gestacional. / Introduction: Monitoring gestational weight gain is extremely important in prenatal care, as it can avoid a series of unfavorable outcomes both for the mother and for the baby. Most countries use some kind of reference to follow up gestational weight. These references are often based in their own populations or in populations from other countries. Considering that population characteristics vary according to ethnics, geographical location, eating habits, anthropometric measures and socio-economic conditions, among other factors, recommendations based on specific populations are preferable to measure gestational weight. Several methodologies are used in the construction of references of gestational weight. WHO (World Health Organization) recommends using longitudinal studies based on selected populations with low prevalence of maternal and fetal complications. In Brazil the references used for gestational weight are based in two international populations which did not use longitudinal studies to generate the references, which brings to Brazil the need to develop its own approach to monitor gestational weight. Purpose: Build a gestational weight curve based on a Brazilian pregnant women cohort using the statistical methodology recommended by WHO to build standard curves of child growth. Method: Data from the Brazilian Gestational Diabetes Study, multicentric study which gathered women from six Brazilian capital cities (Porto Alegre, Rio de Janeiro, São Paulo, Salvador, Manaus and Fortaleza), was used to build the curve. 2,103 eutrophic pregnant women were selected, according to the IOM (Institute of Medicine) (2009), with good gestational outcomes, that is, there was an exclusion of pregnant women with: gestational diabetes mellitus, hypertensive disorders, multiple pregnancies, preterm deliveries (less than 37 weeks), newborns with low birth weight (LBW), large for gestational age newborns (LGA), small for gestational age newborns (SGA), and macrosomia. From these eutrophic pregnant women, 918 were drawn, who will be part of the validation curve. To build the curve 1,179 eutrophic pregnant women with good gestational outcomes remained. The method GAMLSS (Generalized Additive Models of Location, Scale and Shape) from the software R was used for adjustment, which estimated the percentiles 3, 5, 10, 25, 75, 90, 95 and 97. After some exclusions, the second stage of the work used the 918 drawn eutrophic pregnant women with good outcomes and other 901 eutrophic pregnant women who had at least one unwanted gestational outcomes described above (except for multiple pregnancies and diabetes mellitus), in a total of 1.817 pregnant women for the validation group. With the percentiles estimated by the gestational weight curve, cutoff points were defined which determined the risk factors for the interest outcomes. The influence of risk factors on the outcomes was measured through the relative risk (RR) and its respective intervals with 95% confidence, estimated by Poisson regression with strong variance. The relative risks and their respective intervals of 95% confidence for exhibition at some point during pregnancy and for exhibition at some moment in each trimester. The percentiles were assessed with no adjustment for any possible confounding factor. Results: After testing several GAMLSS methods, the one which best adjusted the data was the one which used the distribution family BCPE (Box Cox of Exponential Power), with pb smoothing (B-splines), using two parameters and . The percentil 25 estimated was able to predict low birth weight, prematurity and SGA, whereas the percentil 75 can be used as a predictor of hypertensive disorders, macrosomia and LGA. Conclusion: the model obtained for the gestational weight curve construction indicated that the relationship between gestational weight and gestational age is not linear. The flexibility of the statistical methodology used in the study is sufficient to be applied using BMI instead of gestational weight. This methodology also presents a series of advantages concerning its modeling options. The adjusted percentile curves were efficient to predict adverse gestational outcomes. The methodology applied in this thesis can be replicated for all pre gestational BMI categories.

Introdução: O monitoramento do ganho de peso gestacional é de extrema importância nos cuidados pré-natais, pois pode evitar diversos desfechos desfavoráveis tanto para mãe quanto para o bebê. A maioria dos países utiliza algum tipo de referência para o acompanhamento do peso gestacional. Essas referências, muitas vezes, são baseadas em suas próprias populações ou em populações de outros países. Considerando-se que características populacionais variam de acordo com etnia, localização geográfica, hábitos alimentares, medidas antropométricas e condições socioeconômicas, dentre outros fatores, as recomendações baseadas em populações específicas são preferíveis para monitorar o peso gestacional. Várias metodologias são utilizadas para a construção de referências de peso gestacional. A OMS (Organização Mundial da Saúde) recomenda utilizar estudos longitudinais, a partir de populações selecionadas com baixa prevalência de complicações maternas e fetais. No Brasil, as referências utilizadas para peso gestacional são baseadas em duas populações internacionais; essas populações não utilizaram estudos longitudinais para gerar as referências, o que faz com que o Brasil necessite desenvolver sua própria abordagem para o monitoramento do peso gestacional. Objetivo: Construir uma curva de peso gestacional a partir de uma coorte de gestantes brasileiras, utilizando a metodologia estatística recomendada pela OMS para a construção das curvas padrão de crescimento infantil. Método: Dados do Estudo Brasileiro de Diabetes Gestacional (EBDG), estudo multicêntrico que reuniu gestantes de seis capitais brasileiras (Porto Alegre, Rio de Janeiro, São Paulo, Salvador, Manaus e Fortaleza), foram utilizados para a construção da curva. Foram selecionadas 2.103 gestantes eutróficas, de acordo com IOM (Institute of Medicine) (2009), com bons desfechos gestacionais, ou seja, foram excluídas gestantes com diabetes mellitus gestacional, com distúrbios hipertensivos, com gestações múltiplas e com partos prematuros (<37 semanas de gestação); excluíram-se também casos de gestantes com recém-nascidos de baixo peso ao nascer (BPN) ou com recém-nascidos grandes para a idade gestacional (GIG) e recém-nascidos pequenos para a idade gestacional (PIG), bem como casos de macrossomia. Dentre essas gestantes eutróficas, foram sorteadas 918, que irão fazer parte da validação da curva. Para a construção da curva, ficaram 1.179 gestantes eutróficas com bons desfechos gestacionais. Para o ajuste, foi utilizado o método GAMLSS (Modelos Aditivos Generalizados de Localização, Escala e Forma) do software R, que estimou os percentis 3, 5, 10, 25, 75, 90, 95 e 97. Após algumas exclusões, a segunda etapa do trabalho utilizou as 918 sorteadas (gestantes eutróficas com bons desfechos) e mais 901 gestantes eutróficas que tinham pelo menos um dos desfechos gestacionais indesejáveis descritos acima (com exceção de gestações múltiplas e diabetes mellitus), totalizando 1.817 gestantes para o grupo de validação. Com os percentis estimados pela curva de peso gestacional, foram definidos pontos de corte que determinaram os fatores de risco para os desfechos de interesse. A influência dos fatores de risco sobre os desfechos foi medida através do risco relativo (RR) e seus respectivos intervalos, com 95% de confiança, estimados através de regressão de Poisson com variância robusta. Os riscos relativos e seus respectivos intervalos de 95% de confiança foram estimados para a exposição em algum momento da gestação e para a exposição em algum momento dentro de cada trimestre. Os percentis foram avaliados sem ajustar para nenhum possível fator de confusão. Resultados: Após testados vários modelos GAMLSS, o que melhor ajustou os dados foi o que utilizou a família de distribuição BCPE (Box Cox de Potência Exponencial), com suavizador pb (B-splines), utilizando dois parâmetros e . O percentil 25 estimado foi capaz de predizer baixo peso ao nascer, prematuridade e PIG; já o percentil 75 pôde ser utilizado como preditor de distúrbios hipertensivos, macrossomia e GIG. Conclusão: o modelo obtido para a construção da curva de peso gestacional indicou que a relação entre peso gestacional e idade gestacional não é linear. A flexibilidade da metodologia estatística utilizada no estudo é suficiente para que possa ser aplicada utilizando-se o Índice de Massa corporal (IMC) em vez de peso gestacional. Essa metodologia também apresenta uma série de vantagens no que diz respeito às suas opções de modelagem. As curvas de percentis ajustadas foram eficientes em predizer desfechos gestacionais adversos. A metodologia aplicada nesta tese pode ser replicada para todas as categorias de IMC pré-gestacional. / Introduction: Monitoring gestational weight gain is extremely important in prenatal care, as it can avoid a series of unfavorable outcomes both for the mother and for the baby. Most countries use some kind of reference to follow up gestational weight. These references are often based in their own populations or in populations from other countries. Considering that population characteristics vary according to ethnics, geographical location, eating habits, anthropometric measures and socio-economic conditions, among other factors, recommendations based on specific populations are preferable to measure gestational weight. Several methodologies are used in the construction of references of gestational weight. WHO (World Health Organization) recommends using longitudinal studies based on selected populations with low prevalence of maternal and fetal complications. In Brazil the references used for gestational weight are based in two international populations which did not use longitudinal studies to generate the references, which brings to Brazil the need to develop its own approach to monitor gestational weight. Purpose: Build a gestational weight curve based on a Brazilian pregnant women cohort using the statistical methodology recommended by WHO to build standard curves of child growth. Method: Data from the Brazilian Gestational Diabetes Study, multicentric study which gathered women from six Brazilian capital cities (Porto Alegre, Rio de Janeiro, São Paulo, Salvador, Manaus and Fortaleza), was used to build the curve. 2,103 eutrophic pregnant women were selected, according to the IOM (Institute of Medicine) (2009), with good gestational outcomes, that is, there was an exclusion of pregnant women with: gestational diabetes mellitus, hypertensive disorders, multiple pregnancies, preterm deliveries (less than 37 weeks), newborns with low birth weight (LBW), large for gestational age newborns (LGA), small for gestational age newborns (SGA), and macrosomia. From these eutrophic pregnant women, 918 were drawn, who will be part of the validation curve. To build the curve 1,179 eutrophic pregnant women with good gestational outcomes remained. The method GAMLSS (Generalized Additive Models of Location, Scale and Shape) from the software R was used for adjustment, which estimated the percentiles 3, 5, 10, 25, 75, 90, 95 and 97. After some exclusions, the second stage of the work used the 918 drawn eutrophic pregnant women with good outcomes and other 901 eutrophic pregnant women who had at least one unwanted gestational outcomes described above (except for multiple pregnancies and diabetes mellitus), in a total of 1.817 pregnant women for the validation group. With the percentiles estimated by the gestational weight curve, cutoff points were defined which determined the risk factors for the interest outcomes. The influence of risk factors on the outcomes was measured through the relative risk (RR) and its respective intervals with 95% confidence, estimated by Poisson regression with strong variance. The relative risks and their respective intervals of 95% confidence for exhibition at some point during pregnancy and for exhibition at some moment in each trimester. The percentiles were assessed with no adjustment for any possible confounding factor. Results: After testing several GAMLSS methods, the one which best adjusted the data was the one which used the distribution family BCPE (Box Cox of Exponential Power), with pb smoothing (B-splines), using two parameters and . The percentil 25 estimated was able to predict low birth weight, prematurity and SGA, whereas the percentil 75 can be used as a predictor of hypertensive disorders, macrosomia and LGA. Conclusion: the model obtained for the gestational weight curve construction indicated that the relationship between gestational weight and gestational age is not linear. The flexibility of the statistical methodology used in the study is sufficient to be applied using BMI instead of gestational weight. This methodology also presents a series of advantages concerning its modeling options. The adjusted percentile curves were efficient to predict adverse gestational outcomes. The methodology applied in this thesis can be replicated for all pre gestational BMI categories.