Lifestyle Intervention in Preparation for Pregnancy (LIPP)

Specific Aim 1: To investigate the physiological significance of lifestyle intervention in preparation for pregnancy (LIPP) on maternal and neonatal metabolism and adiposity. Introduction/Rationale: Our preliminary data demonstrate that supervised lifestyle intervention leads to significant weight loss, improved insulin sensitivity, glucose tolerance and incretin secretion, together with healthier cardiovascular and body composition outcomes in overweight and obese adults. We expect that timing and implementation of the proposed lifestyle intervention will produce similar health benefits in overweight/obese women planning a second pregnancy, and lead to greater insulin sensitivity, reduced insulin secretion, and less inflammation. These improvements will result in the prevention of excess nutrient availability (glucose and lipids) from contributing to excess fetal growth/adiposity. The working hypothesis for this Aim is that in contrast to GWG, the decreased pre-pregnancy insulin sensitivity in obese mothers accounts for the greatest clinical variance in fat accretion in the infant. Although clinically we anticipate a decrease in weight and BMI in the LIPP group, the improvement in insulin sensitivity and metabolic profile are the key physiological measures related to the primary outcome of decreased neonatal adiposity, and not the weight loss per se. The rationale is that the optimal time to implement lifestyle intervention that effectively improves maternal health at the physiological, cellular and molecular level, and results in optimal adiposity in the baby, is prior to pregnancy. Women who lose weight postpartum, experience a decrease in neonatal birth weight (primarily adipose tissue) in subsequent pregnancy, whereas women who gain weight, experience an increase in neonatal birth weight and adiposity. We hypothesize that maternal pre-pregnancy metabolic condition determines the obesogenic in-utero environment, which in turn affects placental programming of mitochondrial and lipid pathways (Specific Aim 2), and body composition of the baby. An additional rationale is that there is a need to understand how improved pregravid metabolic condition affects maternal physiological and molecular function. We anticipate that obese women who complete the LIPP program will enter pregnancy with improved insulin regulated metabolism and reduced insulin resistance, thus facilitating a lower neonatal birth weight and adiposity. We will recruit mothers who delivered their first baby at MHMC. We recognize that these mothers represent a demographic that has limited access to exercise facilities or family support systems that would facilitate free time for exercise. In order to reduce barriers to participation, we will conduct the exercise sessions in local Community Recreation Centers. The Centers have child care facilities and we will support the cost so participants may bring their babies to the LIPP sessions. To further increase participation and maximize retention, we will provide transportation to and from the Recreation Centers. Support for transportation will be requested from the Cleveland Mt. Sinai Foundation. Lifestyle Intervention Weight Loss Phase: The LIPP program is designed to promote weight loss that is 5-10% of body weight. The 4-month weight loss phase consists of aerobic exercise training with diet and behavioral counseling to induce weight loss as was successfully achieved in prior studies. Initially, supervised exercise will be prescribed at 55-60% of HRmax and gradually increased so that after 1-2 weeks, the subjects are exercising at 75-85% of HRmax (\~65-70% VO2max). Supervised exercise will consist of walking/jogging on a treadmill and stationary cycling, 3 days/week, 60 min/session (i.e., 500 kcal/session). The women will wear heart rate (HR) monitors (Polar Electro, Woodbury, NY) during each exercise session so that they have visual feedback of their personalized target heart rate goal. Participants will be advised to reduce caloric intake by \~500 kcal/d in order to support their weight loss goals. The recommended diet will provide \~55% of calories as carbohydrate, 25% as fat, and 20% as protein. Participants will be instructed to consume complex carbohydrates and to avoid simple sugars. Specific caloric needs will be estimated by indirect calorimetry and a sedentary (x1.3) physical activity correction factor. Energy intake will be estimated using the food photo diary app, Meal Snap. Digital photography provides an excellent estimate of energy intake (67). Participants who do not own a smart phone will have one provided through support from the Cleveland Foundation. Records covering a 72-hour diet period will be shared with the research team for determination of calorie and nutrient intake. Meal Snap has a food database of over 350,000 items. However, some meals will not be in this database, therefore our Lifestyle Coaches will enter all foods eaten into our diet database (NDSR, Minneapolis, MN) to facilitate analysis of calorie, and macro/micro nutrient intake. Subjects will generate photos from before and after the meal in order to estimate the amount of food that was eaten. Data will be obtained at baseline, and at 2-week intervals during the initial 16-week supervised weight loss period. Lifestyle Intervention - Weight Management/Maintenance: The pre-pregnancy weight management program (phase 2A, 2B and 2C) is designed to facilitate personalized weight loss goals using lifestyle behaviors that include, exercise, diet and behavioral modification, and is based in part on the Look AHEAD trial. The intervention includes a toolbox concept to help meet weight loss goals. The Lifestyle Coaches will provide personalized instruction on physical activity/exercise - 10,000 steps/day, and the participants will use FitBit Flex (Fitbit.com), to track step count and exercise time. During the first phase of weight maintenance (2A) the women will attend 2 supervised group sessions/week. These sessions include structured exercise (e.g., Zumba, jazzercise, stroller walking), review of dietary photo records, where meals have been eaten (home or away; recorded on smart phone app), and behavioral counseling (with their Lifestyle Coach). Participants will be encouraged to eat 1,200-1,500 kcal/d (\~55% carbohydrate, 25% fat, 20% protein) if below 113 kg, or 1,500-1,800 kcal/d if \>113 kg. Dietary data will be analyzed at 4-week intervals during the weight maintenance period. Importantly, to minimize subject burden and to maximize retention and data acquisition, the Lifestyle Coach will track calorie intake using the participant's photo diary. Behavioral strategies to motivate healthy lifestyle decisions will include: self-monitoring (food, activity, and weight, using FitBit online resources), goal setting (steps/day, weight loss), stimulus control (i.e., social eating, fast food, sitting vs. standing), problem solving (have snacks available, exercise at home), and relapse prevention (i.e., holidays, alcohol, fast food, sweets, problem foods, compulsive eating). After 4 months and the desired weight loss, subjects move to Phase 2B. During this phase subjects maintain an exercise goal of 10,000 steps/day, but will be required to attend only 1 supervised session/week. If a subject fails to maintain weight loss, defined as weight gain of 3% of current body weight, participants will return to Phase 2A for more supervised weight management. Alternatively, if weight loss is maintained after 3 months, subjects will progress to Phase 2C until the subsequent pregnancy. Phase 2C consists of no supervised exercise sessions. However, the subject and Lifestyle Coach will converse weekly by phone to review progress, including Fitbit exercise data and diet. Data suggest that weight management programs delivered by phone are comparable to clinically delivered programs. During the calls participants will be counseled to continue to exercise at the intensity and duration prescribed during Phase 2B. They will be provided with language-specific food and exercise log books. These will be used to guide calorie intake and will provide another record of compliance. Participants randomized to the Control group will receive information on post-pregnancy diet/weight loss from the CRU nutritionist as distinct from the LIPP nutritionist (HB) to decrease cross contamination between groups. Weight Management during Pregnancy: All LIPP and usual care/control groups will be followed by their primary Obstetrical provider. The OB/GYN department at MHMC has recently revised its clinical guidelines for the management of the overweight/obese women based on the December 2015 ACOG practice bulletin (2). All overweight/obese women will be offered nutrition counseling early in pregnancy by a registered dietician from the MHMC Nutrition Department with follow-up visits as needed to support GWG within the IOM guidelines. Nutritional therapy will consider maternal pregravid BMI, ethnic, cultural and social factors in individualizing healthy eating. The electronic health record (EPIC) includes a graphic GWG nomogram and so GWG will be monitored at each visit. All subjects will be encouraged to increase physical activity for at least 30 minutes/day (primarily walking). Clinical management such as, ultrasounds to estimate fetal growth, and fetal surveillance will be based on the ACOG recommendations. The Lifestyle Coach will continue to follow-up only with the LIPP subjects as described in the maintenance phase of the research design. Metabolic Evaluations: Metabolic evaluations will be performed at baseline (3 months ? 2 weeks) postpartum. Follow-up evaluations will be performed after 4 (+/-2 weeks) and 12 months (+/-2 weeks), and then every 6 months (+/-2 weeks) prior to pregnancy up to a maximum of 24 months. Once a subject has pregnancy dating and viability confirmed by ultrasound, metabolic evaluations will continue at 12-16 and 32-36 weeks of gestation. Maternal Body Composition: Anthropometry measurements will include height, weight, and hip and waist circumferences. Total body fat will be measured by whole body plethysmography (Bod Pod; Cosmed, Rome, Italy). We will use a hydration constant of 76% for fat free mass during late pregnancy. Resting Energy Expenditure: Resting Metabolic Rate (RMR) will be determined after an overnight fast using the Cosmed OMNIA metabolic cart (Cosmed, Rome Italy) with a canopy system. We will control for diet by providing a standardized energy balanced meal the evening prior to the test from the CRU. Participants will relax in a quiet low-light metabolic room for 30 min prior to obtaining a 30 min measure of exhaled breath. Oxidative and non-oxidative glucose metabolism will be estimated and urine samples will be obtained before and after the measure in order to calculate non-protein RQ (NPRQ). These data will be used in conjunction with Specific Aim 2, and will be correlated with change in mitochondrial function during pregnancy. Exercise Capacity: An incremental graded treadmill test will be performed at baseline, 4, and 6 month time points in both groups. Oxygen consumption (Jaeger OxyCon Pro/Delta System, Hoechberg, Germany), heart rate, and ratings of perceived exertion will be performed as previously described. Insulin Sensitivity and ?-Cell Function: A 75 g oral glucose tolerance test (OGTT) will be used to assess postprandial glycemia, and insulin sensitivity and secretion. After an overnight fast, blood samples will be drawn at 10 min intervals for the first 60 min, and at 20 min intervals thereafter. C-peptide data will be analyzed using a combined model approach to provide pre-hepatic insulin secretion rates, insulin sensitivity, disposition index, and relate insulin and Cpeptide kinetics (73). Plasma glucose will be measured using the glucose oxidase method (YSI; Yellow Springs, OH). Insulin will be assayed by RIA (Millipore, Billerica, MA). Diagnosis of GDM will be made using criteria recommended by the ACOG (74). Enteroinsular Axis Responses: Plasma samples (with appropriate additives) will be obtained to measure incretin hormones (glucagon-like peptide-1 (GLP-1), and glucose-dependent insulinotropic polypeptide (GIP), and satietyrelated gut peptides (cholecystokinin (CCK), and peptide YY (PYY). Measurements will be made under static (fasting) and dynamic (glucose-stimulated) conditions (10 min intervals up to 1 hour). Metabolic and Inflammatory Biomarkers: Fasting blood samples will be obtained to measure CBC, TSH, HbA1c, lipid panel, and total free fatty acids (FFA). Adipocytokines (adiponectin, leptin interleukin-6, interleukin- 8, TNF-? and hsCRP) will be measured using ELISA (R\&D Systems, Minneapolis, MN). All samples from each subject will be stored at -80oC and run in the same assay at completion to decrease variability. Quality of Life Questionnaire: The SF-36 health survey will be used at baseline, 4 and 12 months, and then at 6 month intervals until pregnancy to assess the subject's health-related quality of life. These data will provide a generic measure of physical and mental health through assessment of physical functioning, bodily pain, limitations due to physical, personal, or emotional problems and well-being, energy/fatigue, and general health perceptions. During pregnancy the questionnaire will be administered at 12-16 and 32-36 weeks. Measuring Fat Mass in the Infant: We have extensive experience in estimating body composition in neonates, and were one of the first centers to procure a Pea Pod (pediatric air densitometry), which is housed in the CRU next to Labor \& Delivery and the postpartum unit. Insulin Resistance at Birth: At birth we will obtain umbilical cord blood for insulin and glucose to estimate insulin resistance using HOMA. Full lipid profile, CRP, and the adipokines IL-6 and leptin (an excellent marker of neonatal fat mass) will be measured in cord blood as described above. Anticipated Results, Challenges, and Alternative Approaches: The primary outcome measure of this proposal is lower neonatal adiposity at birth in the LIPP group relative to: 1) the Control group and 2) with the subject's firstborn. As secondary outcomes we anticipate that prior to a subsequent pregnancy, LIPP will produce significant improvement (absolute and percent changes) in maternal weight and body composition, and more importantly improved insulin sensitivity, beta-cell function, incretin response to glucose, lipid and inflammatory biomarkers, compared to the Control group. We also expect reduced insulin resistance, cord lipids and inflammatory profiles in babies born to LIPP compared to Controls. Recruitment is a well-recognized problem for the successful implementation and completion of lifestyle intervention trials. However, given the novel strategy for recruitment detailed above, our unique access to the patient population, and our extensive experience in metabolic research in pregnancy, we do not anticipate recruitment to present an insurmountable challenge. We will recruit 200 subjects in the first 4 years and complete all mother/baby evaluations according to the proposed timeline. If necessary we will recruit subjects from Cleveland Clinic and University Hospitals, both affiliated with Case Western Reserve University (CWRU). Retention strategies include free transportation to the exercise sessions, free childcare service during the lifestyle sessions and consults with the Lifestyle Coach, cell phone apps to reduce subject burden with data entry, regular telephone contacts, and financial incentives, including an infant car seat at delivery. Our team has had outstanding success in retaining pregnant women in our previous metabolic research and lifestyle intervention studies in overweight and obese populations (48,64). In the current proposal, additional strategies are also in place to maximize retention and minimize dropout. These include phone calls and emails from the Lifestyle Coach to review and reinforce problem solving and self-monitoring skills as described for participants in the Diabetes Prevention Program. We will also establish a buddy system where each participant becomes a buddy to another participant, the two buddies may prepare meals together, exercise together, share problems, etc. The buddy system fosters the sense that staying in the study is important not just for the individual's health but also for the buddy's as well. If subjects in the LIPP program do not meet their target weight loss goal, we will implement a meal replacement strategy using the resources of the CRU/CTSC metabolic kitchen, with attention to the caloric and nutrient needs of the mother depending on breast feeding. Not all participants will conceive. Since the primary outcome is neonatal adiposity, subjects who do not become pregnant will not be included in the primary analysis but will be included in secondary analyses relating to pregravid metabolic improvement. By recruiting women with a previous pregnancy the risk of infertility is significantly reduced. We anticipate that approximately 20% of women will experience a spontaneous abortion, but they will be allowed to continue and resume their previous participation in either the LIPP or Control group. We anticipate that approximately 25% of subjects will drop out before becoming pregnant, and another 15% may drop out during pregnancy. We will enroll 100 subjects in each arm to account for the unlikely event of 40% dropout. Adopting an even more stringent strategy, we report statistical power is available for as low as 50 subjects per group as a worst-case scenario. If retention lags behind projections we will recruit additional subjects at MHMC and CWRU affiliated hospitals. The CRU staff will assist in collecting cord blood and placenta at delivery and performing the Pea Pod body measures. If there is a Pea Pod equipment malfunction we will estimate body composition using neonatal anthropometrics. Although LIPP subjects will be encouraged to delay a second pregnancy until they are in the maintenance phase of the program, we will exclude LIPP subjects only if they conceive during the first 4 month weight loss phase. For the Control subjects, exclusion will occur if a subject becomes pregnant before the 3 month postpartum CRU randomization visit. Use of birth control in this phase is an inclusion criterion. Not all LIPP subjects will conceive at similar times after the 4 month weight loss intervention. We will not adjust for the time between pregnancies between the LIPP and Control subjects. We will use the LIPP and control subject's metabolic status (body composition, insulin sensitivity and response, etc.) at the last metabolic evaluation in the CRU before becoming pregnant as their pregravid or baseline status for the subsequent pregnancy. Because this is a pregnancy study only women with be recruited. However, we will assess the effect of LIPP on males and female offspring, together and independently, based on sex. Statistical Approach: The primary analyses of Specific Aim 1a will be intent-to-treat comparisons of the LIPP and Control groups in regard to changes in maternal insulin sensitivity, BMI, and fat mass. The comparisons will be performed first with two-sample t-tests at p=0.05. Should any imbalance of confounding factors in the groups be recognized, linear regression models that include significant differences (e.g. GDM) will be used to perform covariate adjustments. Based on our 1 year postpartum follow-up studies (62,63), we will have 90% power to detect an absolute or covariate-adjusted improvement in insulin sensitivity of 30% and an 80% power to detect an improvement as small as 25% in the LIPP vs. Control group. Corresponding 95% confidence intervals (95% CI) for the absolute or covariate-adjusted differences or percent improvements in insulin sensitivity between groups will be reported. We estimate the standard deviation (SD) of change in BMI from randomization until subsequent pregnancy to be 5.1 kg/m2. We will have 90% power to detect an absolute or covariate-adjusted difference in BMI of 2.6 kg/m2 and 80% power to detect a difference of 2.26 kg/m2 and 90% power to detect an absolute or covariate-adjusted difference in fat mass of 5.9 kg and 80% power to detect a difference of 5.1 kg between groups prior to the second pregnancy. The primary analysis for Specific Aim 1b is an intent-to-treat comparison of LIPP versus Control neonates with respect to fat mass at birth. The comparison will be performed using a two-sample t-test at p=0.05. Linear regression will be performed, which will include weight (body composition measures) of the subject's first child as a covariate. Should any imbalance of confounding factors (for example gestational age) in the groups be recognized, linear regression models will be used to perform a covariate adjustment. Based on our preliminary data, we estimate the SD of neonatal fat mass between the LIPP and Control groups to be no more than 225g. With at least 50 women in each group, (assuming a 50% dropout), the t-test or linear regression will have 90% power to detect an absolute or covariate-adjusted difference of 146 g fat mass between groups. We have 80% power to detect an absolute or covariate-adjusted difference as small as 126 g fat mass between groups. Corresponding 95% CI for the absolute or covariate-adjusted difference in neonatal fat mass between groups will be reported. For secondary analyses, we will be using the same statistical approach. Based on our preliminary data, we estimate the SD of birth weight to be 700 g; with 50 neonates in each group we will have 90% power to detect an absolute or covariate-adjusted difference of 455 g in birth weight and 80% power to detect a 393 g difference between groups. Additional secondary analyses will include umbilical cord cytokines. Comparisons will be performed using a two-sample t-test at a significance level of p=0.05; however, Mann- Whitney U tests or log transformations will be employed if data are not normally distributed. Linear regression models, including confounding factors, will be used to perform covariate adjustments. Based on our published data (80), we estimate the standard deviations of umbilical cord IL-6 and CRP to be 3.4 pg/ml and 7,900 ng/ml, respectively. With 50 women in each group, we will have 90% power to detect an improvement in IL-6 and CRP levels of 50% and 42%, and 80% power to detect an improvement of 42% and 36%, respectively. Specific Aim 2: To determine the molecular effects whereby lifestyle intervention initiated before pregnancy can improve placental mitochondrial lipid oxidation and accumulation. Introduction/Rationale: Our data suggest that in obese women, a mitochondrial defect in placental tissue is present early in pregnancy, inhibiting the placental capacity for fatty acid oxidation and shunting fatty acids to esterification pathways and lipid accumulation, potentially leading to increased nutrient availability to the fetus and higher adiposity at term. Our group has shown that other mitochondrial processes, such as cholesterol transport and steroidogenesis are impaired in placentas of obese, insulin resistant women at term. Placental mitochondrial content (assessed by mtDNA and citrate synthase activity) is not affected by maternal obesity and insulin resistance at term, suggesting that the observed defects in function are due to changes in mitochondrial activity, rather than number. Previous dietary interventions initiated during pregnancy were unable to alter placental ?-oxidation or fetal fat deposition, potentially because the intervention was begun after placental mitochondrial function was impaired. We anticipate that LIPP will improve placental mitochondrial fatty acid oxidation, which will be measurable at term and associated with lower fatty acid esterification and accumulation, and neonatal fat mass. The hypothesis is that the decreased insulin sensitivity and increased inflammatory environment in obese mothers impairs mitochondrial ?-oxidation in the developing placenta. It is the changes in placental metabolism beginning early in pregnancy that lead to altered nutrient delivery and increased fetal fat deposition. Our rationale is based on the need to understand how changes in placental lipid metabolism mediate the effects of improved pregravid metabolism on neonatal adiposity. We expect that placentas of obese women in the LIPP program will show improved fatty acid oxidation and decreased lipid esterification and accumulation at term. Furthermore, we expect that these changes will correlate with lower maternal inflammation and insulin resistance in early pregnancy and lower neonatal adiposity at term. Anticipated Results and Endpoints: We anticipate that compared to the Control group, placentas of women in LIPP will have: 1) increased ?-oxidation, 2) decreased fatty acid esterification, 3) lower lipid content, 4) increased activity of mitochondrial CPT1B, the rate-limiting enzyme in ?-oxidation, and higher phosphor ACC, which, when phosphorylated, produces less malonyl CoA, the main inhibitory regulator of CPT1B, and 5) no difference in mitochondrial content (as measured by mtDNA and citrate synthase activity). We also anticipate that mitochondrial ?-oxidation and CPT1B activity will negatively correlate with early pregnancy maternal serum inflammatory cytokine markers and insulin resistance and neonatal adiposity. Experimental Design: We will accomplish the objectives of Specific Aim 2 by measuring changes in placental mitochondrial enzyme activity and lipid metabolism in women enrolled in the Control or LIPP groups described in Specific Aim 1. Placental tissue will be collected at delivery from all study participants and embedded in paraffin, or snap frozen in liquid nitrogen, and stored at -80oC for molecular analysis. In a subset of women delivering by scheduled Cesarean delivery (we estimate \~30% of our participants or N=15-18/group), we will also collect fresh placental tissue for lipid metabolism activity assays. Placental lipid metabolism: These assays are well established in the O'Tierney-Ginn lab. Mitochondrial fatty acid oxidation (FAO) and esterification into total lipid assays will be performed in placental explants as described previously, with some modifications. Freshly isolated placental explants will be incubated in the presence of 100 ?M cold palmitate and 3H-palmitate (Moravek Radiochemicals) for 18 hrs. At the end of the incubation period, media will be collected to quantify the FAO rate by detection of 3H2O using the vapor phase equilibration method of Hughes. Esterificationinto total lipids will be determined by homogenizing the treated explants in HPLC-grade acetone and incubating with agitation at room temperature overnight. An aliquot of the acetone-extract lipid suspension will be used to determine the radioactive content by liquid scintillation counting. Oxidation and esterification rates will be defined as nmol palmitate/mg tissue/hr. Assessment of placental mitochondria: Mitochondria will be isolated from frozen placental tissue as previously described. Markers of lipid oxidation (CPT1B) and synthesis/esterification (phospho-ACC) activity will be measured using commercially available kits (Cell Signaling, Abcam) in isolated mitochondria from all samples. Markers of mitochondrial content (mtDNA and citrate synthase activity) will be measured in whole placental tissue in all samples as previously described. Placental Lipid Accumulation: Total placental lipid content will be measured using the Folch method. Anticipated Challenges and Alternatives: 1) We will use only placentas delivered by scheduled cesarean delivery for in vitro oxidation and esterification assays, to avoid changes pertinent to labor. At our hospital, the rate of cesarean for obese women is \~40%. Our conservative estimate of 40-50% drop-out followed by 30% of subjects delivering by cesarean, results in N=15-18/group. Based on our preliminary data, this will leave us adequately powered to detect differences in placental lipid metabolism due to LIPP. We will collect placental samples from all study participants for mitochondrial enzyme activity assays, providing us with an additional assessment of metabolic activity in a larger number of participants. 2) Placental mitochondrial fatty acid oxidation capacity may be affected by mitochondrial number, oxidative phosphorylation activity, and energetic efficiency (coupling). Assessment of mitochondrial oxidative phosphorylation capacity or energetic efficiency requires freshly isolated mitochondria and/or living cells which, for our proposal, would be overly ambitious and costly considering the unpredictable nature of delivery and the large number of participants. Alternatively, we will measure markers of mitochondrial content and key enzymes in mitochondrial lipid metabolism in all placental samples to assess some potential mechanisms underlying changes in placental fatty acid oxidation. Additionally, frozen samples collected from all placentas can be used to measure enzymes involved in electron transport (e.g., ATP synthase) as a marker of mitochondrial oxidative phosphorylation. Statistical Analysis: The primary goal of Aim 2 is to determine the effect of lifestyle intervention prior to pregnancy on placental mitochondrial fatty acid oxidation at term. We hypothesize that placental ?-oxidation will be higher in the LIPP group. We will conduct an intent-to-treat analysis using two sample t-test or non-parametric Wilcoxon rank-sum test to assess differences between groups. Regression analysis will be used to assess the association of placental ?-oxidation and enzyme activity with maternal inflammatory cytokine levels and insulin resistance in early pregnancy, along with neonatal fat mass with adjustment for gestational age and gender. Descriptive statistics, such as mean, median and range will be calculated for all variables. Power and sample size analysis based on our preliminary mitochondrial ?-oxidation data in obese women (38?14 nmol/mg/hr) showed that a sample size of N=18/group achieves 80% power to detect a difference of 25% between groups using a two-sample t-test with a significance level of 0.05.