This national study is convening a sample of 232 youths with SB to complete the following aims.

A multisite, prospective cross-sectional study of youth with the most severe form of SB, myelomeningocele, is in progress. The project is being conducted collaboratively through 4 regional pediatric SB programs.

The SB programs from across the country were strategically chosen based on (1) the ethnic and geographic representation of clinic patients, (2) broad research experience with this vulnerable population, (3) participation in the Centers for Disease Control and Prevention National Spina Bifida Patient Registry (NSBPR), and (4) the ability to recruit the needed sample size. After the COVID-19 pandemic, during the initial implementation phase of the protocol, one of the original sites withdrew after data collection challenges and was replaced by an alternate site. In total, 3 of the 4 sites use their associated clinical and translational science institutes’ research unit for study visits, and the fourth uses their designated clinical research setting. Each site has a predetermined recruitment goal based on its patient enrollment. The NSBPR investigators at each site have established routine communication patterns, annual meetings, and a history of collaboration to determine, gather, and evaluate data collected through the registry. Each site recruits and collects data on its predetermined number of participants and has coinvestigators and staff to support their roles.

On the basis of the data from the NSBPR [7], we estimated that 1399 youth are seen yearly in the 4 participating SB programs. Among these, we expect approximately 81%, or 1133, to be diagnosed with myelomeningocele. Of those, 725 (64%) are expected to be aged between 5 and 18 years and eligible for recruitment. On the basis of the NSBPR data [7], we expect our sample to be approximately 52% female.

This study has a targeted enrollment of 232 youths (aged 5-18 years) diagnosed with myelomeningocele-type SB. Purposive sampling is being used to achieve a minimum of 60 youths who identify as being Hispanic or Latino and 116 who primarily use a wheelchair for mobility. Participants are stratified by age (5-11 years and 12-18 years) and mobility status (ambulates independently vs primarily uses a wheelchair for mobility) to increase the usefulness of the study findings to both mobility groups (Table 1). Once the collection is complete for a subgroup, sites will be notified to stop recruiting for that specific mobility status, age, or ethnicity.

The chosen age range (5-18 years) allows for representation of 2 developmental stages that include prepubertal and pubertal status; children in these groups were successful at following the protocol during the pilot study. In addition, the initial age of 5 years was chosen based on the likelihood of being able to cooperate with the study protocol. The end cutoff point of 18 years was chosen because many youth transfer to an adult SB program at that time.

The specific inclusion of minority groups provides a more accurate representation of the population of youth with SB and makes the findings more useful as obesity prevalence has been documented as varying by ethnic descent [3]. To accommodate participants who do not speak English, interpreters are sufficiently budgeted for, and all study materials are professionally translated into Spanish.

The inclusion criteria are (1) youth aged between 5 and 18 years diagnosed with myelomeningocele and (2) English or Spanish speakers (interpreters will be provided).

Participants are excluded from the study if they meet one or more of the following criteria that either interfere with their ability to complete the protocol or compromise the study results: (1) a medical restriction to a 6-hour fast, (2) pregnancy, (3) expectation to participate in abnormally vigorous and out-of-the-ordinary activity during the 7-day test period, (4) living or traveling >200 miles from the study site during the week before or during the protocol, (5) use of supplemental oxygen, (6) a pacemaker, (7) blood transfusion or intravenous infusion of >500 mL of intravenous fluids the week before the test date or expectation to do so during the 7-day test period, (8) claustrophobia or predisposition to the sensation of claustrophobia (for sites including a resting metabolic cart assessment), and (9) an active viral or bacterial illness during the initiation of the 7-day protocol (if study participants are ill during their in-person study visit, they are rescheduled).

Recruitment is conducted by the participating pediatric SB programs, their surrounding communities, and national organizations (eg, Spina Bifida Association) through newsletters, social media, and referrals from family members and health care providers. It is reiterated during and within the consent and assent process that participation is completely voluntary and that there will be no consequences if individuals do not take part or complete the study.

Recruitment processes used in the previous pilot studies were effective in achieving targeted participation and are being replicated. Recruiting strategies include institutional review board (IRB)–approved advertisements that are posted and distributed within clinics, locations where families congregate, and special events that focus on families with disabilities. In addition, the advertisements can be mailed to patients enrolled in the participating SB programs, shared through social media sites such as X (formerly known as Twitter) and Facebook, and posted on the websites of relevant professional organizations. Within the clinic, a Permission to Contact Form is used, which allows a family (if interested) to indicate where, how, and when they would like to be contacted to discuss the study further. Recruitment materials include a pictorial orientation study manual (in print and web-based), modified from that of our initial pilot study, explaining the study protocol and measurements so that families can make an informed decision to participate. By using SB programs as our recruitment sites, families that have an ongoing relationship with the program can participate in this study in a familiar environment.

To reimburse the families for their time and cost of participation in the study, each family is offered a US $100 gift card for completing the first-day (<5 hours) appointment and a US $150 gift card when the day-7 urine specimens and accelerometers are returned to the study site. The second gift card is sent only after the day-7 samples are received. All supplies and prepaid shipping mailers are provided to the family before their departure on day 1. Early in the implementation process, unanticipated transportation and childcare issues were encountered by a subset of sites. Due to this concern, an amendment was submitted to allow for an additional voucher of up to US $125 per the site’s discretion to accommodate families who expressed interest in participating in the study but acknowledged barriers (eg, lack of transportation, distance to the clinic requiring a hotel stay for an early-morning visit necessary due to fasting for the child, or childcare), along with local parking vouchers if relevant. The request and reason for use are documented by each site. Study dates are on weekdays and include flexible appointment times to meet the needs of the families. Morning appointments are recommended based on the need for a minimum 6-hour fasting status before testing.

Retention of study participants (to complete the full 7-day protocol) is enhanced by replicating strategies successfully used in our pilot studies, including the provision of (1) a comprehensive explanation of what the study entails (aided by the pictorial recruitment manual) so that the family can make an informed choice to participate, (2) flexible scheduling options, (3) clear instructions provided verbally and in print when they leave on day 1, and (4) all materials needed for the collection and the return of the day-7 urine samples and accelerometers with prepaid mailing labels and supplies. To enhance retention and completion of data collection, a call or SMS text message is made or sent to the family on day 5 or 6 to remind them of the final samples needed on day 7 and review instructions and answer questions. The final retention aid is that the second gift card is not sent to the family until the day-7 sample collections are received. The process and requirements for receipt of the second gift card are discussed during their day-1 visit, which we believe may support study retention.

Table 2 outlines vital signs and anthropometrics, Textbox 1 provides a listing of body composition and energy expenditure measures, and Textbox 2 provides study instruments. All instruments and study materials are available in English and Spanish. Specific training, orientation, and validation checks are completed for data collectors along with the provision of a study manual with detailed instructions supported with pictures. For all measures, a standardized process is followed to obtain the measure, and an open comment box is available for the data collector to provide additional details related to any challenges encountered (eg, orthopedic complications) to be considered during analysis. In total, 2 of the measures, DXA and resting metabolic rate through a resting metabolic cart, are only being collected at the lead principal investigator’s site to provide a subset of participants to be examined as a secondary validation assessment for the DLW. The decision to use this site for these additional measures was made as the measures were successfully collected during the pilot work at this site.

After agreeing to participate (see the Recruitment section), consenting and assenting is completed, and a 5-hour morning appointment on day 1 is scheduled. A reminder call is placed to the family before the visit to review the study protocol and fasting requirements and answer any questions. On the day of the visit, the fasting status of the child is confirmed, and standard-of-care vital signs are obtained, including a baseline heart rate, blood pressure, and respiratory rate. The youth’s weight (wheelchair or standing based on mobility status) is obtained and used to verify the accurate dosing of DLW.

A baseline urine or saliva sample (participant’s choice) is obtained. The method of urine collection depends on the youth’s home bladder regimen. If they typically perform clean intermittent catheterization, catheters are provided for them (or their parent) to collect urine; if they typically void, a clean catch in a clean urine hat is completed. If they prefer to provide a saliva sample, they are asked to collect 2 to 3 mL of saliva in a tube.

After baseline urine or saliva sample collection and measurement of weight, DLW is dosed based on the youth’s weight and provided to the youth (1-3 oz depending on the youth’s weight) in a sterilized bottle provided by the laboratory and requested to be consumed using a straw. After drinking the DLW, 50 mL of drinking water are added to the empty DLW bottle and swished around, and the youth is asked to drink the added water to ensure that the DLW was completely consumed. Subsequent urine and saliva collection are conducted at hours 1 (only urine), 3, and 4 after consumption of DLW. The hour-1 sample is collected but discarded as the urine is not isotopically equilibrated with body water at this time and no saliva is collected at this time. The hour-3 and hour-4 samples are collected, timed, and saved for analysis. After hour 1 or between hours 1 and 3, the youth is offered a small snack that is consistent with the caloric requirements (≤250 kcal) of the DLW protocol, such as a granola bar, prepackaged snack, or small can of PediaSure supplement, and up to 8 oz of water.

Between timed urine or saliva samples, up to 5 measures of height (stadiometer [if able to stand], arm span [adjusted by level of lesion], recumbent length, knee height, and ulnar length) and 3 additional measures of alternative body composition beyond DLW (waist circumference, 4-site skinfold, and BIA) are collected. In addition, for participants at site 1, the child has a resting metabolic cart and a DXA scan performed in the hospital’s radiology department if no contraindications are present (eg, metal implants in the body). The resting metabolic cart measurement is completed as the resting metabolic rate is the largest contributor to TDEE. The DXA scan is used to address potential age-related variation in bone mass and hydration and provides coefficients between the DXA and TBW, eliminating concern of variance [69]. All sites conduct a ramped protocol while the study participant wears an accelerometer on their nondominant wrist and a heart rate monitor. The ramped protocol includes walking or wheeling in their wheelchair (based on ambulatory status) for set periods at different levels of intensity (eg, sedentary, light, and moderate to vigorous). Activities include 5 minutes each for sitting quietly and playing with Legos for arm movement, walking or wheeling slowly, and walking or wheeling at a normal pace. The final activity includes walking or wheeling at a fast pace for 6 minutes. After each activity, the child is asked to rate their perceived level of exertion using a pictorial scale [68].

Parents complete the demographic form, including a report of the youth’s Tanner stage, a post–day 1 survey that includes questions on the measurements performed and weight management in children with SB, the Patient Reported Outcomes Measurement Information System parent-proxy peer relationships questionnaire to assess their child’s peer relationships, and the Pediatric Evaluation of Disability Inventory–Computerized Adaptive Test survey related to their child’s function. Youths, with parental assistance as needed, complete the nutrition screener and physical activity screener. The physical activity screener was developed for youths aged 8 to 18 years, so youths aged between 5 and 7 years will not be assessed using this instrument.

Before the end of the day-1 appointment, the family receives a “participant kit” containing all materials needed for urine or saliva collection at home: sterile urine cups and urine hats or saliva collection tubes, labels with participant ID number and space for time and date of collection to be completed, and appropriate-sized catheter if the participant uses it. In addition, the participants are asked to continue wearing the accelerometer provided on their nondominant wrist for the following 7 days. They are provided an accelerometer wear and sample collection log to document when and why the accelerometer was removed and reminders for day-7 urine or saliva sample collection. At the end of the day-1 visit, the family receives a US $100 gift card in return for their time and travel, a US $25 hospital cafeteria voucher, and any additional site-specific gift cards (eg, parking or travel).

On day 7, the participants are asked to collect 2 urine or saliva samples 1 hour apart as close to their original appointment time of day as possible. Families keep urine or saliva samples refrigerated until they are able to mail them back along with their accelerometer to their original collection site using prepaid and addressed mailers and packing materials that are provided. After the study team receives the final samples, the family is mailed their second and final gift card for US $150. The gift card at the end of the study is emailed or mailed based on the participants’ preference. If mailed, the cards are sent through the US Postal Service and require a signed signature for receipt of mail. Table 3 provides a summary of the protocol.

Descriptive statistics, including mean, SD, median, and range, will be calculated for each variable or measurement. When analyzing arm span, knee height, ulnar length, waist circumference, and skinfolds, additional analysis or equations will be used based on the literature [30,38-41,43-48,50-52,78,79]. For BMI (kg/m²) and bioelectrical impedance measures, the average of each surrogate height measure will be used. Arm span can be used as a direct measure of height if there is no leg muscle mass loss (low-level lesions), or it can be modified if there are higher-level lesions and variations in leg muscle mass. If there is partial leg muscle loss (midlevel lesions), arm span × 0.95 can be used; with high-level lesions and complete leg muscle loss (thoracic-level lesion), arm span × 0.90 will be used. We will consider use of the measured arm span with and without the adjustment based on the level of lesion for our analysis. For knee height, 3 prediction equations developed for use with individuals with mobility impairment (cerebral palsy and Duchenne muscular dystrophy) [43-45] will be used. For ulnar length, a prediction equation using the ulnar length and age of the youth will be used [44].

For aim 1, Bland-Altman plot analysis and concordance correlation will evaluate the agreement between DLW body fat percentage and modified BMI, waist circumference, BIA, and skinfold measures. This marginal association analysis will also provide information for the relationship between different measures (linear, quadratic, and nonlinear association). To construct an algorithm that can predict body fat percentage, we will first choose the measures of skinfold thickness at 4 sites in our base model using the equations by Jackson and Pollock [79], Slaughter et al [51], and Gurka et al [78] that have been successfully applied to several adolescent populations [80]. To further improve accuracy, we will construct linear gender-specific regression models by considering the modified BMI (eg, weight or arm span and weight or ulnar length), FFM, waist circumference, mobility status, race, age, or pubertal status as potential predictor variables. Stepwise linear regression and all-possible-subsets regression procedures will be used to develop prediction equations. Potential interactions between age, race, and other variables will be examined. The Mallow Cp statistic and the Schwarz-Bayesian criterion will be used to select the appropriate models. The resulting predicted body fat percentage is a weighted sum of a subset of those potential predictors, where the weights are the parameter estimates associated with each predictor in the regression model. Predictive accuracy will be evaluated using the mean square error through a 10-fold cross-validation; all the samples will be randomly split into 10 parts, 9 folds of the samples will be used to build the regression model, and the remaining hold-out samples will be used to test the predictive accuracy between the predicted and DLW body fat percentage. This procedure will be repeated until all the samples are tested. Upon completion of our body fat percentage algorithm, we can use the distribution of this metric to categorize weight status (obesity, overweight, and normal weight) by selecting appropriate cutoff points. The determination of cutoff points to use for the categorization of the weight status was thoughtful and deliberate. While there are no accepted cutoff points for body fat percentage to define obesity or other weight categories for youth, we will use cutoff points based on body fat percentages correlated to BMI percentiles presented by McCarthy et al [81] that are age specific to account for the growth and development of the child. We will use the agreed-upon criteria for categorizing weight status: body fat percentage between 85% and 94% for overweight, ≥95% for obesity, and 5% to 84% for normal weight [19].

For aim 2, we will predict TDEE by estimating BMR and physical activity level and using these as our primary predictors. Several BMR equations from cross-sectional studies with good model performance (ie, R²=0.7-0.8) have been reported [82] that mainly used age, FFM, FM, and weight as key predictors. Thus, we will use those as our base model predictors and further consider height, race, gender, assistive devices used in ambulation, pubertal status, and mobility as potential predictors. Furthermore, the subset of 42 participants who will undergo a resting metabolic cart to measure resting energy expenditure will provide additional data to construct an equation for predicting TDEE in SB. The model selection procedure and predictive accuracy will be evaluated as described in the statistical methods for aim 1. Once the best BMR equation is identified, TDEE will be estimated as a weighted sum of estimated BMR and physical activity level. To handle the missing data issue, regression analysis will be conducted under multiple imputation under the missing-at-random (MAR) assumption [83]. Sensitivity analysis will be conducted to check for potential violation of the MAR assumption [84]. Simulation under MAR will be conducted to quantify the efficiency loss, defined as 1 – E (Ŷ [X*] – Y)2/E(Ŷ [X] – Y)², where X* represents the observed data and X is the data with missingness. The exploratory aim based on the Block food and physical activity screeners will be analyzed through NutritionQuest. Descriptive analysis on frequency and duration of physical activities, sedentary activity, and youths’ dietary intake by food groups and servings will be presented. Raw accelerometer data are collected at 30 Hz and converted to counts through ActiGraph’s proprietary software (ActiLife; ActiGraph LLC). Analysis will include calibration of individualized activity counts aided by 20-minute ramped protocol data based on the perceived exertion of the youth and measured heart rate. The Block Kids Physical Activity Screener will complement the objective accelerometer data.

The power calculation was based on the increased coefficient of determination (R²) from a baseline simple linear model to a multivariate model that regresses body fat percentage on potential predictors (ie, alternate height, BMI, race, and age). Pilot studies from the literature suggest that the correlation between the body fat percentage measured using DLW and alternate height (arm span and recumbent length) is in the range of approximately 0.6 to 0.7 (R²=0.36-0.49). As we will use 10-fold cross-validation to construct our predictive models for 2 gender groups, a sample size of 94 (232/2 × 0.9 × 0.9) achieves 90% power to detect an increase of R²≥0.066 (0.078) attributed to 1 (2) independent variable(s) using an F test with a significance level of .05. The variables tested are adjusted for 1 independent variable in the simple linear model with an R² of 0.36. We assume that 10% of the participants will not complete all measures based on pilot data. Because it is likely that we will include >2 additional variables in the final model, the proposed sample size will have sufficient power to detect a clinically meaningful increase in R².

Data are collected and stored using REDCap (Research Electronic Data Capture; Vanderbilt University) [85]. Through REDCap, individual levels of privilege and access can be granted to study team members based on need. This database provides menu-driven options and built-in checks and generates reports. All data are checked for accuracy through biweekly reports run by the lead study coordinator. Signed consent or assent forms are kept in a locked file cabinet in the office of each site’s lead investigator and are separate from deidentified study data. REDCap projects rely on a thorough study-specific data dictionary defined in an iterative self-documenting process by team members, with planning assistance from the administrator. The development and testing process results in a well-planned data collection strategy. The REDCap server is housed in a local data center of the lead study coordinator, and all web-based information transmission is encrypted with a 256-bit Secure Sockets Layer certificate. Data backups are performed nightly and stored in a separate location.

This human participant study was reviewed and approved by the Western IRB–Copernicus Group (20190134). Informed consent from a parent or legal guardian for all participants and assent for participating children aged ≥7 years is obtained by a trained staff member who has experience working with children. All documents are IRB approved and provided in English and Spanish based on the needs of the family. During the consent and assent process, it is reiterated that participation is voluntary and that there will be no consequences if they do not participate or complete the study. Signed consent and assent forms are kept in a locked file cabinet in each lead site investigator’s locked office, and a copy is provided to each family. The confidentiality and privacy of the participants’ information is protected by deidentifying the data and through the use of a secure, encrypted, and password-protected database for storage. Compensation for participation is provided using gift cards at 2 time points of the protocol. The monetary value reflects the participants’ time and effort required to complete the study protocol. To reduce family burden, an optional gift card of up to US $125 is available if needed to support travel costs to the study site or associated childcare costs. The amount is determined by each site based on their geographical region, distance to the study site, and associated travel costs.

The aims of this study are to develop 2 algorithms for use in youth with SB in a clinical setting, one to model body fat and one to predict TDEE. In addition, the physical activity and dietary intake of the sample will be described. To increase the generalizability of the findings, the goal is to recruit a diverse sample of youths aged between 5 and 18 years who are diagnosed with myelomeningocele, the most severe type of SB. This population has an increased prevalence of overweight and obesity that influences their physical health, mobility, and level of independence.

Data cleaning and preliminary analysis have documented that there are limited missing data and a high retention of study participants who start the protocol. Current enrollment of 61.6% (143/232) of the proposed sample is on target to reach the goal of including 25.9% (60/232) of participants who self-identify as Hispanic.

Challenges encountered have primarily centered on COVID-19. Since this research project was funded, the COVID-19 pandemic drastically altered clinical programs and, thus, recruitment and data collection. COVID-19 affected the number of health care staff in all settings, resulting in overloads, reduced clinical activities, and unexpected vacancies in key clinic management and study personnel, thus hampering the implementation of the project in some sites. Although each of the clinical sites’ research compliance mechanisms differed, research at all sites was completely suspended for multiple months, with variability by site for resumption of research activities (recruitment and data collection). Each program or organization limited the number of research personnel allowed in clinical settings and the scope of recruitment activities upon resumption, which varied by site. In addition, telehealth drastically reduced clinic attendance in some settings. Finally, travel to sites for training and validation of data collection was challenged by the pandemic.

Although a 6-month extension of year 2 was obtained from the funding agency (National Institutes of Health/NICHD), recruitment remained variable and reduced at each of the study sites. Concurrent to the disruption of clinical processes for conducting research, families experienced challenges with school programs altering attendance and having multiple children at home, and more clinic participants lived a substantial distance from the clinic than was originally anticipated. Families voiced concerns about (1) childcare as clinical and research settings were allowing only 1 adult and no siblings to attend data collection, (2) the risk of public mass transit, (3) the cost of alternative transportation, and (4) difficulty traveling to the data collection site for early-morning appointments necessitated by the 6-hour preappointment fast. These concerns prompted amendments to the study protocol related to participant reimbursement and discontinuing data collection at 1 site and adding a new site.

Despite the unforeseen challenges, the strengths of this study are notable. We are the first to convene a sample of youth with SB to develop an algorithm that incorporates clinically available, cost-effective, but not yet validated measures of height or body composition to accurately model body fat percentage in this population and estimate TDEE. In addition, the purposive recruitment from clinics located in different regions allows for stratification by age and mobility status and provides representation of ethnicity while providing an adequately powered sample to test the aims. The use of DLW, a validated noninvasive measure, as the criterion for body fat percentage and TDEE increases the study’s rigor. The use of a DXA scan in a subgroup of the sample allows for a second comparison of body fat percentage and FFM between DLW and DXA, decreasing concerns of depending on a 2-compartment body water method in this unique population. Limitations include the lower-than-expected enrollment at the time of this publication. The team continues to strategize on methods to further recruitment and enrollment within the timeline of the grant. The findings will enhance the screening, prevention, and treatment of an abnormal weight by facilitating the accurate identification of the youths’ weight status and allow for recommendations of daily caloric needs to be provided for this population at a higher risk of having obesity. With the ability to accurately estimate body fat percentage, we can accurately categorize weight status and monitor weight trends longitudinally. Furthermore, our findings in youth with SB have the potential to impact outcomes for youth diagnosed with disabilities other than SB who experience similar challenges related to alterations in body composition or fat distribution or measurement challenges secondary to mobility issues or musculoskeletal problems. The NICHD research agenda that prioritized weight management for youth with disabilities was not limited to SB but included youth with a variety of intellectual and physical disabilities [28]. The successful execution of this study, methodological approach, and dissemination of the findings will shift current research and clinical practice paradigms as well as raise awareness for the need to conduct future studies in this area. Furthermore, the measures used in the proposed research will serve as a framework for use with children in other diagnostic categories and, thus, will make a major contribution to science.

Dissemination of the study findings will be a priority. Outcomes will be shared with national SB organizations (ie, Spina Bifida Association); the Centers for Disease Control and Prevention NSBPR national and international conferences focused on SB, weight management, and developmental disabilities; manuscripts; and individual and family organizations. Future directions for this work will be to test the application of these algorithms in youth with SB and expand this work to adults with SB. The outcomes of this study will also be implemented in weight management interventions.

This innovative study will provide 2 clinically accessible and feasible algorithms related to body fat and TDEE. The lack of an accurate clinically available and cost-effective method to measure body fat and the inability to provide daily caloric recommendations obstruct the advancement of the science of weight management of youth with SB. These foundational deficiencies significantly impede successful prevention and treatment of obesity for this vulnerable population. The outcomes of this study will advance the science of weight management in individuals with disabilities, address national research priorities, and shift clinical practice paradigms.