- Email: [email protected]
Weight gain and triceps skinfolds fat mass after gastrostomy placement in children with developmental disabilities
RESEARCH
Weight gain and triceps skinfolds fat mass after gastrostomy placement inchildren with developmental disabilities JANET SUGARMANISAACS, PhD, RD; KEITHE. GEORGESON, MD; HARRIET H. CLOUD, MS, RD; NANCY WOODALL, RN
eeding difficulties in children with severe disabilities may AIBSTRCT Objective To determine appropriate outcome indicators of nutritional status that are measurable over time after gastrostomy placement in children with severe neurologic impairments. Design Twenty-two nonambulatory children met the selection criteria: feeding by gastrostomy of at least 50% of total energy, age between 1 and 12 years, diagnosis of neurologic impairments, and presurgical recommendation for weight gain. Each child served as his or her own control; three assessments were made after gastrostomy placement. Setting Children were seen in specialty outpatient clinics. Statistical analyses z Scores and Pearson product moment correlations Results Outcomes of gastrostomy placement were (a) increase in actual weight, (b) increase in weight-age equivalent, (c) rate of weight accretion as expected by National Center for Health Statistics growth charts and improved z scores for half of the children, and (d) improvement in triceps skinfolds percentiles for nearly half (n=10) of the children. The results reflect the heterogeneity of children with severe disabilities. Pearson correlations showed a significant relationship between chronologic age and weight-age equivalent (r=.96), but not for weight for age and weight-age equivalent, or triceps skinfolds fat mass and weight-age equivalent. Conclusions/applications Weight and triceps skinfolds fat mass were appropriate outcome indicators of nutritional status measurable over time. Weight-age equivalent and z scores were more helpful than standard growth plots for interpreting weight gain over time. Our data also support findings that undernutrition limits growth before gastrostomy placement in patients with disabilities. Nutritionists are encouraged to track improvement in nutritional status after gastrostomy placement with measurements of triceps skinfolds fat mass and to use the information to support families facing decisions about the need for this surgery. JAm Diet Assoc. 1994; 94:849-854.
require families and health care providers to consider gastrostomy placement (1-3). Stuart et al (4) critically
reviewed feeding gastrostomies and reported the increasing use of percutaneous endoscopic gastrostomy; they recommended careful patient selection. Taylor et al (5) studied a communitybased cohort and outcomes in adults and noted the lack of practice guidelines for placement of gastrostomies. Significant controversy exists in regard to the appropriateness, type, and timing of a protective antireflux procedure and gastrostomy
placement for neurologically impaired children. McGrath et al (6) documented unexpectedly high survival of children with gastrostomies compared with studies of survival of similar patients who had severe cerebral palsy and no gastrostomy. Their work and other longitudinal studies were based on retrospective
chart reviews rather than prospective experimental designs. The purpose of our study was to track longitudinally two outcome indicators of nutritional status after gastrostomy placement: body weight and triceps skinfolds fat mass. The following were our main research questions: Can growth be documented
over time as an outcome of gastrostomy placement in severely impaired children? Are weight gain and triceps skinfolds fat mass clinical indicators of improved nutritional status for this group of patients? Evidence of improved nutritional outcomes would help families and health care providers make difficult decisions about the appropriateness and timeliness of such surgery.
Feeding difficulties are a threat to overall health; conversely, the ability to self-feed is a strong predictor of life expectancy for
persons with Down syndrome (7). Bax (8) pointed out that assistance with feeding and eating problems is not provided in
routine health care for children with special health care needs, in part, because underweight and short stature are so common.
J. S. Isaacs is an assistantprofessor and H. H. Cloud is a professor emeritus with the Departmentof Nutrition Sciences, CivitanInternationalResearch Center, University ofAlabama at Birmingham. K E. Georgeson is a professor of surgery and the director of the Division of PediatricSurgery, University of Alabama at Birmningham, The Children'sHospital ofAlabama, Birmingham. N. Woodall is with Children'sRehabilitationServices in Birmingham. Address correspondence to: J. S. Isaacs, PhD,RD, Civitan InternationalResearch Center, University of Alabama at Birmingham, SC 313, Birmingham, AL 35294. JOURNAL OF THE AMERICAN DIETETIC ASSOCIATION / 849
RERAWC
Table 1 Descriptive characteristics of neurologically impaired children with gastrostomies (n = 22)
Characteristic
No.
Primary diagnosis Cerebral palsy/seizures Congenital anomaly or genetic syndrome
15 7
was placed, received less than 50% of energy via the gastrostomy, or were infants. The children were assessed at two community settings by the same nutrition staff, who used the same procedures and type of equipment. Training in measurement techniques was provided to all nutrition staff and rechecked over time by the researchers. Both sites provided medical teams, social service coordination, and assistance in procuring gastrostomy-related supplies and formula.
Feeding regimen No intake orally Intake orally and by gastrostomy
9 13
Gastrostomy placement or revision Within 1 yr of study 1 yr or longer before study
17 5
Age at gastrostomy placement Between 6 mo and 3 yr Between 3 and 5 yr Older than 5 yr
10 3 9
Demographics Girls Boys White Black
10 12 12 10
Inschool or program Attending Not attending
15 7
Rempel et al (9) and Shapiro et al (10) used retrospective records to determine weight gain after gastrostomy placement. Other studies have looked at outcomes such as surgical complications. Wheatley et al (11) reported on 29 children with recurrent gastroesophageal reflux after gastrostomy placement and fundoplication surgery. Another study (12) that reported on 79 children retrospectively used weight and height for weight to categorize children into groups that were then compared for surgical complications. Methods of gastrostomy placement, related surgeries, and long-term complications are still being evaluated (4,5,12-16).
METHODS The study design consisted of repeated measures over time, with each child serving as his or her own control. This method resolved the difficulty of finding an adequate control group, as orally fed children with the same diagnoses differ in degree of impairments. Subjects were selected from a prospective longitudinal cohort of all children with gastrostomies seen by the nutrition staff of the Civitan International Research Center Sparks Clinics. Twentytwo patients met all the selection criteria: (a) feeding of 50% of total energy or more by gastrostomy; (b) nonambulatory status, age between 1 and 12 years, and a diagnosis of neurologic impairment; (c) presurgical growth data indicating that weight gain was recommended; and (d) completion of at least three nutrition assessments after gastrostomy placement. Although dietary intake was not the focus of this research, 24-hour recalls were used to determine whether subjects met the first inclusion criterion. Presurgical nutrition and physician recommendations were reviewed to confirm that weight gain was an initial treatment goal. Subjects were measured over 10 months ormore for children between 1 and 2 years, and over 1 year or more for children between 3 and 12 years. Some families decreased feeding by gastrostomy as oral feeding progressed, but the children were not dropped from the study. Excluded children were at or above the 50th percentile according to National Center for Health Statistics (NCHS) growth charts for weight for age when the gastrostomy 850 / AUGIJST 1994 VOLUME 94 NUMBER 8
Anthropometric Measurements Nutrition assessments consisted of anthropometric measurements, growth plots, interviews with a parent or caretaker, and review of medical records. All subjects were weighed without shoes or orthopedic devices on a beam balance chair scale with a supportive insert. Recumbent lengths were measured on a recumbent-length board with an embedded measure. Arm circumferences were taken at the measured midpoint between the acromial process of the shoulder and the tip of the elbow, with the unclothed arm flexed. Measures of skinfolds fat mass were made according to standard methods using Lange calipers (Cambridge Scientific Industries, Cambridge, Md), which were checked for calibration. When more than three assessments were completed, the initial and final measurements and one interim measurement closest to the midpoint in time were selected. Although recumbent length was measured, its validity was so questionable that this measure was not considered reliable for individual orgrouped data. All the children were nonambulatory and had or were at risk for developing scoliosis, contracture, and hip dislocations. Data Analysis Anthropometric data were plotted and expressed as a percentile on the NCHS growth chart for the appropriate age and sex (17,18) and on the Frisancho standards (19,20). The convention termed weight-age equivalent was used to characterize weight for chronologic age in children below the 5th percentile on the NCHS growth charts. This convention assigns an age not based on the date of birth but on the 50th percentile of the NCHS weight-forage chart that matches the child's weight. The method expresses the observation that an underweight 6-year-old child can weigh the same as a typical 3-year-old child. Total weight gain over time was calculated for each child and expressed as a percentage of weight gained at the 50th percentile or 5th percentile on the NCHS growth charts for the appropriate age and sex. For children who initially measured well below the 50th percentile, catch-up growth was defined as weight gained above the rate expected for children of the same chronologic age at the 50th percentile. Pearson product moment correlations were calculated using the Statistical Analysis System (version 6.03, 1988, SAS Institute, Cary, NC) for three data points in time for weight and for each growth parameter for each child. The Pearson correlation tests whether one variable influences another by measuring the degree of association; independent variables are not correlated. Percentile scores for each baseline, intermediate, and final assessment were compared for weight for age vs chronologic age, and weight for age vs triceps skinfolds fat mass percentile. The z scores were calculated using the Centers for Disease Control and Prevention Antho software (version 1.01, 1990, USDepartment of Health and Human Services, Atlanta, Ga) for weight for age, height for age, and weight for height for each assessment. The z score is the difference between a measured parameter and the median of the standard reference population divided by the standard deviation. It describes the relationship of a given subject to the 50th percentile of the NCHS growth standards. Two standard deviations from the mean (50th percentile) are equivalent to the 2nd and 98th percentiles.
THE JONAL TIILKS WITH ... Janet Sugarman Isaacs, PhD, RD, who began her career in ruminant nutrition at a college of agriculture, recalls that her first exposure to gastrostomy, and the concept that an artificial hole in the stomach could be used to study digestion, involved cows. Isaacs' current research on gastrostomy feeding evolved from her work with a surgeon who asked for help documenting the effectiveness of this procedure. By tracking changes in nutritional status indicators of patients, Isaacs not only showed that gastrostomy feeding can improve the health of patients, she also highlighted the extent to which undernutrition and malnutrition complicate the medical condition of children with severe neurologic impairments. Journal: Is gastrostomy a cormnon procedure for treating feeding problems in children with disabilities? Isaacs: Gastrostomy is relatively rare in children, except for those with severe disabilities. It is more often used for nutrition support with adults who have cancers or catastrophic illnesses, but advances in technology and surgical procedures, and an increase in the number of children surviving neonatal intensive care, have made gastrostomy an increasingly common treatment option for children with severe disabilities. Journal: Outside of pediatric or major medical centers, is it likely most dietitians will become involved in this treatment? Isaacs: Many dietitians probably encounter these children and their families without realizing it. All of us who work with children have had parents tell us about children who have difficulty eating. A mother who might not feel comfortable talking to a physician often is quite willing to tell a dietitian at the local WIC clinic about problems she has feeding her child. We need to take advantage of the fact that mothers and other caregivers talk to us about their children. I think sometimes dietitians hear about these problems but think they can't help, especially when there are serious physical problems or complicating medical conditions. It's especially important that we support parents of a child who is severely disabled because they face a series of difficult decisions regarding treatment for their child. We can explain to these parents how changing the texture or consistency of foods might make it easier for their child to eat by mouth. We can also help parents decide whether or when it is unsafe for their child to continue eating orally and then explain other options, such as gastrostomy. As gastrostomy for children becomes more common, it's increasingly likely that dietitians who work in schools, public health departments, or other community settings will encounter children who have already had gastrostomies placed. These children live at home. They go to school. They are part of the community. They simply are not fed by mouth. Journal: What roles do dietitians play in treatment? Isaacs: As part of the medical team, the dietitian's main role is deciding at what point referral to more specialized treatment is necessary. This would be after the dietitian has ensured that every effort has been made to allow a child to eat by mouth. After the gastrostomy has been placed, dietitians help parents determine the best method for feeding their child. For example, it's common to use an intermittent pump that
delivers the liquid nutrition while a child sleeps. Many children with gastrostomies continue to eat some food by mouth, and the dietitian should continually adjust the diet to maintain a proper balance between food eaten by mouth and liquid nutrition delivered via the gastrostomy. Dietitians work with the family to determine whether a child can return to oral feeding. Not all children with gastrostomies have to keep them, especially children who receive gastrostomies because of trauma. In many cases, children are unable to eat because their muscles are extremely weak. Once they've been fed by gastrostomy for 4 to 6 months, children may regain enough strength to benefit from physical or occupational therapies, and they can then learn to eat by mouth. Gastrostomy necessarily becomes permanent only when coordination problems prevent a child from swallowing safely. One of the most important ways a dietitian can support parents of children with disabilities is to understand that complex emotional and psychological dynamics affect how a family copes with a child's feeding problem and illness. Many parents equate their ability to get their child to eat with their ability to be a good parent. They are devastated when their child won't or can't eat. Some pediatricians give parents tlhe message that a feeding problem exists because they are not feeding the child correctly. This is unfortunate because once a family's emotional security has been undermined, it's difficult for them to feel anything but loss and failure if their child later needs a gastrostomy. That's why the way a dietitian talks to parents from the time a child is first identified as being difficult to feed is so critical. Those early messages c an set in motion dynamics that help parents either feel good about their ability to care for their child or feel failure and despair. I always tell parents that I know they are good parents and are trying their best. I try to help them understand that a feeding problem doesn't reflect whether a parent, and child love each another. I explain that feeding problems are related to a child's underlying diagnoses or brain damage. It's usually helpful to tell parents that any child with the same medical problems would be difficult to feed. I also try to nirnunlze concern over the gastrostomy by telling parents to hink of it as a "second mouth" that can be used to bypass some if the difficulties the child was having. I explain that we can still feed the child some foods by mouth and that they can still have meals together as a family. Journal: WIhy do some parents, and medical professionals, resist the use of gastrostomy in children? Isaacs: Choosing gastrostomy is a difficult decision for many families because they might have been told that their child is severely disabled. They are therefore not; surprised when their child gets frequent infections or doesn't grow. As a dietitian, I believe we need to consider all the options for care and to realize that the child has not only a disability but also undernutrition or malnutrition superimposed on that disability. If the child were relieved of the difficulties involved in eating by mouth, we might see that the child is capable of smiling at the family, attending school, or performing tasks previously thought impossible. Gastrostomy for children is not only about whether a child will continue to live, it may also determine how much a child can interact with his or her family or benefit from other therapies.
JO( )RNkAL
OF THE AMERICAN DIETETIC ASSOC(IATION / 851
RESEARCH
RESULTS
Table 2 Neurologically impaired children with gastrostomies (n = 22): anthropometric percentiles at baseline and final assessments at least 12 months later
Characteristic No.
Baseline Frequency (%)
No.
Final Frequency (%)
Weight for age Less than 5th percentile Between 5th and 50th Between 50th and 95th Over 95th percentile
14 7 1 0
64 32 5 0
14 3 5 0
64 14 23 0
11 6 0 0 5
50 27 0 0 23
7 2 0 0 13
32 9 0 0 59
9 8 0 0 5
41 36 0 0 23
4 4 0 0 14
18 18 0 0 64
16 4 2
73 18 9
16 4 2
73 18 9
12 6 4
50 27 18
8 6 8
36 27 36
7 9 4
32 41 18
5 9 7
23 41 32
Recumbent length Less than 5th percentile Between 5th and 50th Between 50th and 95th Over 95th percentile Interfering condition
Length-for-weight relationship Less than 5th percentile Between 5th and 50th Between 50th and 95th Over 95th percentile Interfering conditiona
Head circumference Microcephaly 2nd and 98th percentile Macrocephaly
Triceps skinfolds fat mass Under 5th percentile Below 50th percentile 50th percentile or greater
Arm circumference Under 5th percentile Below 50th percentile 50th percentile or greater
aCondition interfering with length measurement such as scoliosis, kyphosis, contracture, hip dislocation.
852 / AUGUST 1994 VOLUME 94 NUMBER 8
Characteristics of the Subjects Elapsed time between baseline and final measurements ranged from 10 months to 4 years. Five children were observed for 3 or more years; 11 children were observed between 1 and 2 years. Table 1 characterizes the subjects. All of the chil(lren were nonverbal, nonambulatory, and dependent for care. The initial dietary intake after gastrostomy placement for all patients determined that the inclusion criteria were met: substantial energy (more than 50%) was consistently being fed through the gastrostomy, and energy was sufficient to expect weight gain, based on calculated basal metabolic rate (21). Ten children were fed both orally and by gastrostomy at the initial evaluation. Mean total energy intake for the 22 children was 68 kcal/kg body weight (standard deviation=26 kcal), which is consistent with their age and inactivity. The most common diagnosis was spastic quadriplegia in 15 children. Its etiology was prernaturity-related sequela or birth trauma for seven of the subjects. Seven others had congenital anomalies or genetic syndromes. Chronologic age ranged from 15 months to 12 years at baseline assessments.
Growth Over Time Table 2 shows the anthropometric data collected at the first and final postsurgical nutrition assessments, based on NCHS growth charts for the appropriate age and sex and the Frisancho standards (17-20). The children were shorter, lighter, and more inicrocephalic than predicted using standard growth charts. Changes between initial and final assessments were seen in percentiles for triceps skinfolds fat mass and weight for age. This suggests that tracking changes bymeasurement oftriceps skinfolds fat mass would show improved nutritional status more easily than other standard measurements. Incremental weight gain in pounds or kilograms was an outcome of gastrostomy placement for 21 of 22 children. One child was diagnosed with a failed fiundoplication and had bolus feedings reduced in volume after the family refused surgical intervention. Only nine children had incremental increases in recumbent length, with no change in percentiles. Measuring changes in recumbent length overtime was ineffective because a majority of the children (n= 13) had diagnoses that interfered with length meas iuremelnts by the final assessment. Initial weight-age equivalents raged from 3 months to 9 years 6 months. Only two children had chronologic ages within 1 year of their initial weight-age equivalent. Weight-age equivalent. increased at least 1 year between initial and final assessments for 20 of the 22 children. Pearson correlations showed a significant (r=.96) relationship between chronologic age and weight-age equivalent. No significant correlation was found between weight; for age and weight-age equivalent (r=.61) or triceps skinfolds fat mass and weight-age equivalent (=.39). This suggests that weight-age equivalent, which is an extrapolated measure, does not change as percentiles (doover time. Evidence of a growth response to gastrostorny placement is shown in Table 3 according to two methods: rate of weight accretion and catch-up growth. Both NCHS growth charts and z scores showed the same results about weight gain, but z scores quantified poor growth. Small children who were below the fifth percentile on the NCHS weight-for-age chart before gastrostomy placement remained below the fifth percentile, although they gained weight. The z score calculated a negative number greater than 2 for such a child. After weight gain the z score became more positive, even if it was still a negative number. This advantage of the z score is shown in Table 3 for the weight-for-age category. Of
the 14 children who had initial and final measurements below the fifth percentile in weight for age, z scores increased for 9 children. The z scores were not reported for height for age or weight for height because of the interfering diagnoses that affected recumbent length measurements over time. Table 3 provides evidence of a growth response to gastrostomy placement even at the 50th percentile of the NCHS growth charts for weight for age. Catch-up growth was defined as weight gain at a faster than anticipated rate for a child of the same age and sex and was not found for most children. All the children had weight gain as a goal at the time of gastrostomy placement, so catch-up growth was anticipated. However, half of the children studied added weight at the same rate or faster than children without these severe developmental disabilities. The comparison of rate of weight gain for these subjects at the fifth percentile of the NCHS growth charts shows a more positive response -13 (59%) children showed normal or catch-up growth. Evidence of a growth response was also seen with triceps skinfolds fat mass and arm circumference (Table 3) in that children's percentiles increased.
Table 3
Evidence of growth response to gastrostomy placement over time: weight accretion and catch-up growth Growth response
50th percentile
5th percentile
Rate of weight accretion' Low: weight accretion 0%-70% Normal: weight accretion 70%-130% Catch-up growth: weight accretion exceeded 130%
12 7
9 8
3
5
Change in percentile
Increase in z scoreb
3 10 7
11
Catch-up growth Weight for age Triceps skinfolds Midarm circumference
aBased on growth charts of the National Center for Health Statistics for children of the same age and sex. bThe z score is 0 at the mean.
DISCUSSION For the group of children studied, growth as an outcome of gastrostomy placement was shown as increase in actual weight, increase in weight-age equivalent, rate of weight accretion as expected by NCHS growth charts for 11 of the children, improvement in triceps skinfolds percentiles for 10 of the children, and z score increases for 11 of the children subjects (moving closer to the median). These results reflect the heterogeneity of a group of children with severe disabilities. Growth in these subjects must be interpreted with care. Growth rates in children with specific neurologic disorders are not known. An underlying assumption is that growth and activity would continue to be limited by the children's neurologic impairments to the same degree after gastrostorny placement as before. Feeding by gastrostomy inchildren with severe disabilities is affected by the type of surgical procedures, primary and secondary diagnoses, family coping styles, and social and financial support. Also, the low activity of the nonambulatory children is an important limitation in interpreting growth using the NCHS growth charts. Weight gain was clinically significant, particularly when compared at the fifth percentile. Comparison with this lower percentile may be more appropriate in clinical applications because the group as a whole was so small in head size and weight at the first assessment. The lack of statistical significance of the Pearson product moment correlations between changes in weight for age and triceps skinfolds fat mass may be explained by assumption of a linear relationship in the statistic. Linear growth would mean crossing percentiles as seen in catch-up growth. The Pearson correlation was selected because all children had weight gain as an initial goal at the time of gastrostomy placement. Change in percentile over time, however, is not usual; normally growing children stay at the same percentile as weight increases. This study showed that growth as an outcome of gastrostomy placement in severely impaired children cannot be consistently shown by usual methods. Improvement in weight-for-age percentiles and z scores only described half of the children. Interfering conditions (scoliosis and contractures) limited the validity of length measurements over time. Consequently, estimation of energy needs based on kilocalories per centimeter of height or other body mass index calculations are not helpful in patients with such orthopedic complications (22). Although low weight was a consistent finding, this may not be interpreted as poor growth by some families. Families who lift children often prefer weight maintenance rather than growth. Even though the need for weight gain was noted at the time of
gastrostomy placement, children were observed long enough for this goal to be reached and modified, particularly for older children. This treatment goal would violate a basic assumption of the Pearson product moment statistic; no linear change over time would be anticipated. Clinical outcome indicators of growth for the children studied are suggested by our results. The most sensitive indicators of growth changes with feeding by gastrostomy over time were weight-age equivalent and triceps skinfolds fat mass. The usefulness of tracking triceps skinfolds fat mass in this group of patients was shown. The skinfolds measurement includes skin and fat only (23,24). Because it does not include muscle mass, it would be less likely to reflect the underlying neuromuscular diagnosis in these children. Triceps skinfolds thickness remains stable in age groupings of well-nourished children but changes in marasmic children compared with obese children (25). Triceps fat is related to effective energy intake and to body energy reserves (25). This is a more sensitive clinical indicator of changes in nutritional status over time than weight for age. There are many measurements of skinfolds among various populations, which show that the triceps site has been most completely analyzed (23). Measurements of triceps skinfolds fat mass have been used for patients with a variety of chronic illnesses, such as those undergoinghemodialysis (26) and young children with ascaris (25), and for pregnant women (27). Midarm circumference measurement would include muscle and bone as well as skin and fat (23). Small muscle size is a common finding in patients with cerebralpalsy and microcephaly, so arm circumference may reflect the underlying diagnosis more than nutritional intake. We found similar weight gain response to feeding bygastrostomy as reported by two retrospective studies (9,10). A prospective study of surgical complications noted weight gain as an outcome (28). Our study had a narrower age range than the patients studied by Rempel (9), but fewer subjects younger than 2 years of age. Rempel found that children younger than 2 had a more positive response than older children (9). Our study agreed with the study of Shapiro et al (10) which did not find age at gastrostomy placement to be a factor in growth response. Other studies involving children with a similar degree of neurologic impairment and age range have reported that the children tolerate percutaneous gastrostomy and gain weight. Such results contribute to wider application of newer techniques (5,12,28). Microcephaly has not been reported previously in evaluating response to feeding by gastrostomy, but it is an important predic,JOt [RNAL OF THE AMERICAN DIETETIC ASSOCIATION / 853
tor of neurological impairment. It may be a predictor of small body size also and maybe useful in setting realistic goals for weight gain. Several limitations of our study are inherent in working with children with severe disabilities: lack of control of medications, frequent acute infections, and complex family reactions to the stress of caring for a child with disabilities. Other parameters of nutritional status were not measured, such as iron status and other sites of skinfolds fatmass. Regardless ofhowanthropometric changes are interpreted in these children, our results support gastrostomy placement as a mechanism to improve the nutritional status indicators of weight and fat mass. Normal growth expectations are not justified in children with severe neurologic impairments, but improvement in growth was found for all children overtime. Even though limited to 23% of the children, catchup growth was documented; that is, these children gained weight above the rate expected for children of the same chronologic age at the 50th percentile. The results of this subgroup support the initial recommendation for weight gain before gastrostomy placement. Feeding by gastrostomy corrected undernutrition as a limiting factor for the children in this subgroup.
For children with severe neurologic impairments, use of gastrostomy seems to be an effective method for correcting undernutrition and achieving weight gain or catch-up growth
APPLICATIONS This study supports dietetics practitioners in their efforts to achieve weight gain and even catch-up growth in children with severe neurologic impairments. It also supports the assumption that undernutrition limits growth of the child before gastrostomy placement. Such information may be helpful to families facing a decision about gastrostomy placement surgery. Nevertheless, health care practitioners should not overstate the growth expectations for families considering a gastrostomy. For children with gastrostomies, triceps skinfolds fat mass may be a more sensitive clinical indicator of change after placement than weight and height measurements. Triceps skinfolds fat mass and weight accretion are appropriate outcome measures of improved nutritional status after gastrostomy placement. · This studyfunded, in part, by MCH ProjectMCJ-019143 University Affiliated TrainingGrant to the Civitan InternationalResearch Center University of Alabama at Birmingham References 1. Gisel E, Patrick J. Identification of children with cerebral palsy unable to maintain a normal nutritional state. Lancet. 1988; 1: 283-286. 2. Griggs CA, Jones PM, Lee RE. Videofluoroscopic investigation of feeding disorders of children with multiple handicap. Dev Med Child Neurol. 1989; 31:303-308. 3. Jones PM. Feeding disorders in children with multiple handicaps. Dev Med Child Neurol. 1989; 31:398-406. 4. Stuart SP, Tiley EH, Boland JP. Feeding gastrostomy: a critical 854 / AUGUST 1994 VOLUME 94 NUMBER 8
review of it indications and mortality rate. South Med J 1993; 86:169172. 5. Taylor CA, Larson DE, Ballard DJ, Bergstrom LR, Silverstein MD, Zinsmeister AR, Dimagno EP. Predictors of outcome after percutaneous endoscopic gastrostomy: a community-based study. Mayo Clin Proc. 1992; 67:1042-1049. 6. McGrath SJ, Splaingard ML, Alba HM, Kaufman BH, Glicklick M. Survival and functional outcome of children with severe cerebral palsy following gastrostomy. Arch Phys Med Rehabil. 1992; 73:133-137. 7. Eyman RK, Call TL, White JF. Life expectancy of persons with Down syndrome. Am J Ment Retard. 1991; 95:603-612. 8. Bax M. Eating is important. Dev Med Child Neurol. 1989; 31:285286. 9. Rempel GR, Colwell SO, Nelson RP. Growthin children with cerebral palsy fed via gastrostomy. Pediatrics. 1988; 82:857-862. 10. Shapiro BK, Green P, Krick J, Allen D, Capute AJ. Growth of severely impaired children: neurological versus nutritional factors. Dere Med Child Neurol. 1986; 28:729-733. 11. Wheatley MJ, Coran AG, Wesley R, Oldham KT, Turnage RH. Redo fundoplication in infants and children with recurrent gastroesophageal reflux. JPediatrSurg. 1991; 26:758-761. 12. Albanese CT, Towbin RB, Ulman I, Lewis J, Smith SD. Percutaneous gastrojejunostomy versus Nissan fundoplication for enteral feeding of the neurologically impaired child with gastroesophageal reflux. J Pediatr. 1993; 123:371-375. 13. Martinez D, Ginn-Peasem ME, Caniano DA. Sequelae of antireflux surgery in profoundly disabled children. JPediatrSurg. 1992; 27:267273. 14. Hassett JM, Sunby C, Flint LM. No elimination of aspiration pneumonia in neurologically disabled patients with feeding gastrostomy. Surg Gynecol Obstet. 1988; 167:383-388. 15. Lazares B, Murphy JB, Culpepper L. Aspiration associated with long-term gastric versus jejunal feeding: a critical analysis of the literature. Arch Phys Med Rehabil. 1990; 71:46-53. 16. Wilkinson JD, Dudgeon DL, Sondheimer JM. A comparison of medical and surgical treatment of gastroesophageal reflux in severely retarded children. JPediatr.1981; 99:202-205. 17. National Center for Health Statistics. Height and Weight of Youths 12 17. Washington, DC: Mental Health Administration, Public Health Service; 1973. Vital and Health Statistics. Series 11. No 124. 18. National Center for Health Statistics. NCHS Growth Curves for Children 0-18 Years Washington, DC: Mental Resources Administration, Public Health Service, 1977. Vital and Health Statistics. Series 11. No. 165. 19. Frisancho AR. Triceps skinfold and upper arm muscle size norms for assessment of nutritional status. Am J Clin Nutr. 1974; 27:10521057. 20. Frisancho AR. New norms of upper limb fat, and muscle areas for assessment of nutritional status. Am J Clin Nutr. 1981; 34:2540-2545. 21. Hendricks KM, Walker WA. Manual of PediatricNutrition. 2nd ed. Philadelphia, Pa: B. C. Decker, 1990:59-69. 22. Krick J, Murphy PE, Markham JFB, Shapiro BK. A proposed formula for calculating energy needs of children with cerebral palsy. Dev Med Child Neurol. 1992; 34:481-487. 23. Eveleth PB, TannerJM. Worldwide Vaiation in Human Growth. 2nd ed. New York, NY: Cambridge University Press, 1990. 24. Gregory JW, Greene SA, Scringeour CM, Rennie MJ. Body water measurements in growth disorders: a comparison of bioelectrical mlpedance and skinfold thickness techniques with isotope dilution. Arch Dis Child. 1991; 66:220-222. 25. Rammohan M, Aplasca EC. Caliper method vs bioelectrical impedance analysis for determining body fat in patients undergoing chronic dialysis and in health individuals. JAm DietAssoc. 1992; 92:1395-1396. 26. Stephenson LS, Crompton DWT, Latham MC, Schulpen TWJ, Nesheim MC, Janesen AAJ. Relationships between ascaris infection and growth of malnourished preschool children in Kenya. Am J Clin Nutr. 1980; 33:1165-1172. 27. Neggers Y, Goldenberg RL, Cliner SP, Hoffman HJ, Cutter GR. Usefulness of various maternal skinfold measurements for predicting newborn birth weight. JAm Diet Assoc. 1992; 92:1393-1394. 28. King SJ, Chait PG, Daneman A, Pereira J. Retrograde percutaneous gastrostomy: a prospective study in 57 children. PediatrRadiol. 1993; 23:23-25.
Weight gain and triceps skinfolds fat mass after gastrostomy placement inchildren with developmental disabilities JANET SUGARMANISAACS, PhD, RD; KEITHE. GEORGESON, MD; HARRIET H. CLOUD, MS, RD; NANCY WOODALL, RN
eeding difficulties in children with severe disabilities may AIBSTRCT Objective To determine appropriate outcome indicators of nutritional status that are measurable over time after gastrostomy placement in children with severe neurologic impairments. Design Twenty-two nonambulatory children met the selection criteria: feeding by gastrostomy of at least 50% of total energy, age between 1 and 12 years, diagnosis of neurologic impairments, and presurgical recommendation for weight gain. Each child served as his or her own control; three assessments were made after gastrostomy placement. Setting Children were seen in specialty outpatient clinics. Statistical analyses z Scores and Pearson product moment correlations Results Outcomes of gastrostomy placement were (a) increase in actual weight, (b) increase in weight-age equivalent, (c) rate of weight accretion as expected by National Center for Health Statistics growth charts and improved z scores for half of the children, and (d) improvement in triceps skinfolds percentiles for nearly half (n=10) of the children. The results reflect the heterogeneity of children with severe disabilities. Pearson correlations showed a significant relationship between chronologic age and weight-age equivalent (r=.96), but not for weight for age and weight-age equivalent, or triceps skinfolds fat mass and weight-age equivalent. Conclusions/applications Weight and triceps skinfolds fat mass were appropriate outcome indicators of nutritional status measurable over time. Weight-age equivalent and z scores were more helpful than standard growth plots for interpreting weight gain over time. Our data also support findings that undernutrition limits growth before gastrostomy placement in patients with disabilities. Nutritionists are encouraged to track improvement in nutritional status after gastrostomy placement with measurements of triceps skinfolds fat mass and to use the information to support families facing decisions about the need for this surgery. JAm Diet Assoc. 1994; 94:849-854.
require families and health care providers to consider gastrostomy placement (1-3). Stuart et al (4) critically
reviewed feeding gastrostomies and reported the increasing use of percutaneous endoscopic gastrostomy; they recommended careful patient selection. Taylor et al (5) studied a communitybased cohort and outcomes in adults and noted the lack of practice guidelines for placement of gastrostomies. Significant controversy exists in regard to the appropriateness, type, and timing of a protective antireflux procedure and gastrostomy
placement for neurologically impaired children. McGrath et al (6) documented unexpectedly high survival of children with gastrostomies compared with studies of survival of similar patients who had severe cerebral palsy and no gastrostomy. Their work and other longitudinal studies were based on retrospective
chart reviews rather than prospective experimental designs. The purpose of our study was to track longitudinally two outcome indicators of nutritional status after gastrostomy placement: body weight and triceps skinfolds fat mass. The following were our main research questions: Can growth be documented
over time as an outcome of gastrostomy placement in severely impaired children? Are weight gain and triceps skinfolds fat mass clinical indicators of improved nutritional status for this group of patients? Evidence of improved nutritional outcomes would help families and health care providers make difficult decisions about the appropriateness and timeliness of such surgery.
Feeding difficulties are a threat to overall health; conversely, the ability to self-feed is a strong predictor of life expectancy for
persons with Down syndrome (7). Bax (8) pointed out that assistance with feeding and eating problems is not provided in
routine health care for children with special health care needs, in part, because underweight and short stature are so common.
J. S. Isaacs is an assistantprofessor and H. H. Cloud is a professor emeritus with the Departmentof Nutrition Sciences, CivitanInternationalResearch Center, University ofAlabama at Birmingham. K E. Georgeson is a professor of surgery and the director of the Division of PediatricSurgery, University of Alabama at Birmningham, The Children'sHospital ofAlabama, Birmingham. N. Woodall is with Children'sRehabilitationServices in Birmingham. Address correspondence to: J. S. Isaacs, PhD,RD, Civitan InternationalResearch Center, University of Alabama at Birmingham, SC 313, Birmingham, AL 35294. JOURNAL OF THE AMERICAN DIETETIC ASSOCIATION / 849
RERAWC
Table 1 Descriptive characteristics of neurologically impaired children with gastrostomies (n = 22)
Characteristic
No.
Primary diagnosis Cerebral palsy/seizures Congenital anomaly or genetic syndrome
15 7
was placed, received less than 50% of energy via the gastrostomy, or were infants. The children were assessed at two community settings by the same nutrition staff, who used the same procedures and type of equipment. Training in measurement techniques was provided to all nutrition staff and rechecked over time by the researchers. Both sites provided medical teams, social service coordination, and assistance in procuring gastrostomy-related supplies and formula.
Feeding regimen No intake orally Intake orally and by gastrostomy
9 13
Gastrostomy placement or revision Within 1 yr of study 1 yr or longer before study
17 5
Age at gastrostomy placement Between 6 mo and 3 yr Between 3 and 5 yr Older than 5 yr
10 3 9
Demographics Girls Boys White Black
10 12 12 10
Inschool or program Attending Not attending
15 7
Rempel et al (9) and Shapiro et al (10) used retrospective records to determine weight gain after gastrostomy placement. Other studies have looked at outcomes such as surgical complications. Wheatley et al (11) reported on 29 children with recurrent gastroesophageal reflux after gastrostomy placement and fundoplication surgery. Another study (12) that reported on 79 children retrospectively used weight and height for weight to categorize children into groups that were then compared for surgical complications. Methods of gastrostomy placement, related surgeries, and long-term complications are still being evaluated (4,5,12-16).
METHODS The study design consisted of repeated measures over time, with each child serving as his or her own control. This method resolved the difficulty of finding an adequate control group, as orally fed children with the same diagnoses differ in degree of impairments. Subjects were selected from a prospective longitudinal cohort of all children with gastrostomies seen by the nutrition staff of the Civitan International Research Center Sparks Clinics. Twentytwo patients met all the selection criteria: (a) feeding of 50% of total energy or more by gastrostomy; (b) nonambulatory status, age between 1 and 12 years, and a diagnosis of neurologic impairment; (c) presurgical growth data indicating that weight gain was recommended; and (d) completion of at least three nutrition assessments after gastrostomy placement. Although dietary intake was not the focus of this research, 24-hour recalls were used to determine whether subjects met the first inclusion criterion. Presurgical nutrition and physician recommendations were reviewed to confirm that weight gain was an initial treatment goal. Subjects were measured over 10 months ormore for children between 1 and 2 years, and over 1 year or more for children between 3 and 12 years. Some families decreased feeding by gastrostomy as oral feeding progressed, but the children were not dropped from the study. Excluded children were at or above the 50th percentile according to National Center for Health Statistics (NCHS) growth charts for weight for age when the gastrostomy 850 / AUGIJST 1994 VOLUME 94 NUMBER 8
Anthropometric Measurements Nutrition assessments consisted of anthropometric measurements, growth plots, interviews with a parent or caretaker, and review of medical records. All subjects were weighed without shoes or orthopedic devices on a beam balance chair scale with a supportive insert. Recumbent lengths were measured on a recumbent-length board with an embedded measure. Arm circumferences were taken at the measured midpoint between the acromial process of the shoulder and the tip of the elbow, with the unclothed arm flexed. Measures of skinfolds fat mass were made according to standard methods using Lange calipers (Cambridge Scientific Industries, Cambridge, Md), which were checked for calibration. When more than three assessments were completed, the initial and final measurements and one interim measurement closest to the midpoint in time were selected. Although recumbent length was measured, its validity was so questionable that this measure was not considered reliable for individual orgrouped data. All the children were nonambulatory and had or were at risk for developing scoliosis, contracture, and hip dislocations. Data Analysis Anthropometric data were plotted and expressed as a percentile on the NCHS growth chart for the appropriate age and sex (17,18) and on the Frisancho standards (19,20). The convention termed weight-age equivalent was used to characterize weight for chronologic age in children below the 5th percentile on the NCHS growth charts. This convention assigns an age not based on the date of birth but on the 50th percentile of the NCHS weight-forage chart that matches the child's weight. The method expresses the observation that an underweight 6-year-old child can weigh the same as a typical 3-year-old child. Total weight gain over time was calculated for each child and expressed as a percentage of weight gained at the 50th percentile or 5th percentile on the NCHS growth charts for the appropriate age and sex. For children who initially measured well below the 50th percentile, catch-up growth was defined as weight gained above the rate expected for children of the same chronologic age at the 50th percentile. Pearson product moment correlations were calculated using the Statistical Analysis System (version 6.03, 1988, SAS Institute, Cary, NC) for three data points in time for weight and for each growth parameter for each child. The Pearson correlation tests whether one variable influences another by measuring the degree of association; independent variables are not correlated. Percentile scores for each baseline, intermediate, and final assessment were compared for weight for age vs chronologic age, and weight for age vs triceps skinfolds fat mass percentile. The z scores were calculated using the Centers for Disease Control and Prevention Antho software (version 1.01, 1990, USDepartment of Health and Human Services, Atlanta, Ga) for weight for age, height for age, and weight for height for each assessment. The z score is the difference between a measured parameter and the median of the standard reference population divided by the standard deviation. It describes the relationship of a given subject to the 50th percentile of the NCHS growth standards. Two standard deviations from the mean (50th percentile) are equivalent to the 2nd and 98th percentiles.
THE JONAL TIILKS WITH ... Janet Sugarman Isaacs, PhD, RD, who began her career in ruminant nutrition at a college of agriculture, recalls that her first exposure to gastrostomy, and the concept that an artificial hole in the stomach could be used to study digestion, involved cows. Isaacs' current research on gastrostomy feeding evolved from her work with a surgeon who asked for help documenting the effectiveness of this procedure. By tracking changes in nutritional status indicators of patients, Isaacs not only showed that gastrostomy feeding can improve the health of patients, she also highlighted the extent to which undernutrition and malnutrition complicate the medical condition of children with severe neurologic impairments. Journal: Is gastrostomy a cormnon procedure for treating feeding problems in children with disabilities? Isaacs: Gastrostomy is relatively rare in children, except for those with severe disabilities. It is more often used for nutrition support with adults who have cancers or catastrophic illnesses, but advances in technology and surgical procedures, and an increase in the number of children surviving neonatal intensive care, have made gastrostomy an increasingly common treatment option for children with severe disabilities. Journal: Outside of pediatric or major medical centers, is it likely most dietitians will become involved in this treatment? Isaacs: Many dietitians probably encounter these children and their families without realizing it. All of us who work with children have had parents tell us about children who have difficulty eating. A mother who might not feel comfortable talking to a physician often is quite willing to tell a dietitian at the local WIC clinic about problems she has feeding her child. We need to take advantage of the fact that mothers and other caregivers talk to us about their children. I think sometimes dietitians hear about these problems but think they can't help, especially when there are serious physical problems or complicating medical conditions. It's especially important that we support parents of a child who is severely disabled because they face a series of difficult decisions regarding treatment for their child. We can explain to these parents how changing the texture or consistency of foods might make it easier for their child to eat by mouth. We can also help parents decide whether or when it is unsafe for their child to continue eating orally and then explain other options, such as gastrostomy. As gastrostomy for children becomes more common, it's increasingly likely that dietitians who work in schools, public health departments, or other community settings will encounter children who have already had gastrostomies placed. These children live at home. They go to school. They are part of the community. They simply are not fed by mouth. Journal: What roles do dietitians play in treatment? Isaacs: As part of the medical team, the dietitian's main role is deciding at what point referral to more specialized treatment is necessary. This would be after the dietitian has ensured that every effort has been made to allow a child to eat by mouth. After the gastrostomy has been placed, dietitians help parents determine the best method for feeding their child. For example, it's common to use an intermittent pump that
delivers the liquid nutrition while a child sleeps. Many children with gastrostomies continue to eat some food by mouth, and the dietitian should continually adjust the diet to maintain a proper balance between food eaten by mouth and liquid nutrition delivered via the gastrostomy. Dietitians work with the family to determine whether a child can return to oral feeding. Not all children with gastrostomies have to keep them, especially children who receive gastrostomies because of trauma. In many cases, children are unable to eat because their muscles are extremely weak. Once they've been fed by gastrostomy for 4 to 6 months, children may regain enough strength to benefit from physical or occupational therapies, and they can then learn to eat by mouth. Gastrostomy necessarily becomes permanent only when coordination problems prevent a child from swallowing safely. One of the most important ways a dietitian can support parents of children with disabilities is to understand that complex emotional and psychological dynamics affect how a family copes with a child's feeding problem and illness. Many parents equate their ability to get their child to eat with their ability to be a good parent. They are devastated when their child won't or can't eat. Some pediatricians give parents tlhe message that a feeding problem exists because they are not feeding the child correctly. This is unfortunate because once a family's emotional security has been undermined, it's difficult for them to feel anything but loss and failure if their child later needs a gastrostomy. That's why the way a dietitian talks to parents from the time a child is first identified as being difficult to feed is so critical. Those early messages c an set in motion dynamics that help parents either feel good about their ability to care for their child or feel failure and despair. I always tell parents that I know they are good parents and are trying their best. I try to help them understand that a feeding problem doesn't reflect whether a parent, and child love each another. I explain that feeding problems are related to a child's underlying diagnoses or brain damage. It's usually helpful to tell parents that any child with the same medical problems would be difficult to feed. I also try to nirnunlze concern over the gastrostomy by telling parents to hink of it as a "second mouth" that can be used to bypass some if the difficulties the child was having. I explain that we can still feed the child some foods by mouth and that they can still have meals together as a family. Journal: WIhy do some parents, and medical professionals, resist the use of gastrostomy in children? Isaacs: Choosing gastrostomy is a difficult decision for many families because they might have been told that their child is severely disabled. They are therefore not; surprised when their child gets frequent infections or doesn't grow. As a dietitian, I believe we need to consider all the options for care and to realize that the child has not only a disability but also undernutrition or malnutrition superimposed on that disability. If the child were relieved of the difficulties involved in eating by mouth, we might see that the child is capable of smiling at the family, attending school, or performing tasks previously thought impossible. Gastrostomy for children is not only about whether a child will continue to live, it may also determine how much a child can interact with his or her family or benefit from other therapies.
JO( )RNkAL
OF THE AMERICAN DIETETIC ASSOC(IATION / 851
RESEARCH
RESULTS
Table 2 Neurologically impaired children with gastrostomies (n = 22): anthropometric percentiles at baseline and final assessments at least 12 months later
Characteristic No.
Baseline Frequency (%)
No.
Final Frequency (%)
Weight for age Less than 5th percentile Between 5th and 50th Between 50th and 95th Over 95th percentile
14 7 1 0
64 32 5 0
14 3 5 0
64 14 23 0
11 6 0 0 5
50 27 0 0 23
7 2 0 0 13
32 9 0 0 59
9 8 0 0 5
41 36 0 0 23
4 4 0 0 14
18 18 0 0 64
16 4 2
73 18 9
16 4 2
73 18 9
12 6 4
50 27 18
8 6 8
36 27 36
7 9 4
32 41 18
5 9 7
23 41 32
Recumbent length Less than 5th percentile Between 5th and 50th Between 50th and 95th Over 95th percentile Interfering condition
Length-for-weight relationship Less than 5th percentile Between 5th and 50th Between 50th and 95th Over 95th percentile Interfering conditiona
Head circumference Microcephaly 2nd and 98th percentile Macrocephaly
Triceps skinfolds fat mass Under 5th percentile Below 50th percentile 50th percentile or greater
Arm circumference Under 5th percentile Below 50th percentile 50th percentile or greater
aCondition interfering with length measurement such as scoliosis, kyphosis, contracture, hip dislocation.
852 / AUGUST 1994 VOLUME 94 NUMBER 8
Characteristics of the Subjects Elapsed time between baseline and final measurements ranged from 10 months to 4 years. Five children were observed for 3 or more years; 11 children were observed between 1 and 2 years. Table 1 characterizes the subjects. All of the chil(lren were nonverbal, nonambulatory, and dependent for care. The initial dietary intake after gastrostomy placement for all patients determined that the inclusion criteria were met: substantial energy (more than 50%) was consistently being fed through the gastrostomy, and energy was sufficient to expect weight gain, based on calculated basal metabolic rate (21). Ten children were fed both orally and by gastrostomy at the initial evaluation. Mean total energy intake for the 22 children was 68 kcal/kg body weight (standard deviation=26 kcal), which is consistent with their age and inactivity. The most common diagnosis was spastic quadriplegia in 15 children. Its etiology was prernaturity-related sequela or birth trauma for seven of the subjects. Seven others had congenital anomalies or genetic syndromes. Chronologic age ranged from 15 months to 12 years at baseline assessments.
Growth Over Time Table 2 shows the anthropometric data collected at the first and final postsurgical nutrition assessments, based on NCHS growth charts for the appropriate age and sex and the Frisancho standards (17-20). The children were shorter, lighter, and more inicrocephalic than predicted using standard growth charts. Changes between initial and final assessments were seen in percentiles for triceps skinfolds fat mass and weight for age. This suggests that tracking changes bymeasurement oftriceps skinfolds fat mass would show improved nutritional status more easily than other standard measurements. Incremental weight gain in pounds or kilograms was an outcome of gastrostomy placement for 21 of 22 children. One child was diagnosed with a failed fiundoplication and had bolus feedings reduced in volume after the family refused surgical intervention. Only nine children had incremental increases in recumbent length, with no change in percentiles. Measuring changes in recumbent length overtime was ineffective because a majority of the children (n= 13) had diagnoses that interfered with length meas iuremelnts by the final assessment. Initial weight-age equivalents raged from 3 months to 9 years 6 months. Only two children had chronologic ages within 1 year of their initial weight-age equivalent. Weight-age equivalent. increased at least 1 year between initial and final assessments for 20 of the 22 children. Pearson correlations showed a significant (r=.96) relationship between chronologic age and weight-age equivalent. No significant correlation was found between weight; for age and weight-age equivalent (r=.61) or triceps skinfolds fat mass and weight-age equivalent (=.39). This suggests that weight-age equivalent, which is an extrapolated measure, does not change as percentiles (doover time. Evidence of a growth response to gastrostorny placement is shown in Table 3 according to two methods: rate of weight accretion and catch-up growth. Both NCHS growth charts and z scores showed the same results about weight gain, but z scores quantified poor growth. Small children who were below the fifth percentile on the NCHS weight-for-age chart before gastrostomy placement remained below the fifth percentile, although they gained weight. The z score calculated a negative number greater than 2 for such a child. After weight gain the z score became more positive, even if it was still a negative number. This advantage of the z score is shown in Table 3 for the weight-for-age category. Of
the 14 children who had initial and final measurements below the fifth percentile in weight for age, z scores increased for 9 children. The z scores were not reported for height for age or weight for height because of the interfering diagnoses that affected recumbent length measurements over time. Table 3 provides evidence of a growth response to gastrostomy placement even at the 50th percentile of the NCHS growth charts for weight for age. Catch-up growth was defined as weight gain at a faster than anticipated rate for a child of the same age and sex and was not found for most children. All the children had weight gain as a goal at the time of gastrostomy placement, so catch-up growth was anticipated. However, half of the children studied added weight at the same rate or faster than children without these severe developmental disabilities. The comparison of rate of weight gain for these subjects at the fifth percentile of the NCHS growth charts shows a more positive response -13 (59%) children showed normal or catch-up growth. Evidence of a growth response was also seen with triceps skinfolds fat mass and arm circumference (Table 3) in that children's percentiles increased.
Table 3
Evidence of growth response to gastrostomy placement over time: weight accretion and catch-up growth Growth response
50th percentile
5th percentile
Rate of weight accretion' Low: weight accretion 0%-70% Normal: weight accretion 70%-130% Catch-up growth: weight accretion exceeded 130%
12 7
9 8
3
5
Change in percentile
Increase in z scoreb
3 10 7
11
Catch-up growth Weight for age Triceps skinfolds Midarm circumference
aBased on growth charts of the National Center for Health Statistics for children of the same age and sex. bThe z score is 0 at the mean.
DISCUSSION For the group of children studied, growth as an outcome of gastrostomy placement was shown as increase in actual weight, increase in weight-age equivalent, rate of weight accretion as expected by NCHS growth charts for 11 of the children, improvement in triceps skinfolds percentiles for 10 of the children, and z score increases for 11 of the children subjects (moving closer to the median). These results reflect the heterogeneity of a group of children with severe disabilities. Growth in these subjects must be interpreted with care. Growth rates in children with specific neurologic disorders are not known. An underlying assumption is that growth and activity would continue to be limited by the children's neurologic impairments to the same degree after gastrostorny placement as before. Feeding by gastrostomy inchildren with severe disabilities is affected by the type of surgical procedures, primary and secondary diagnoses, family coping styles, and social and financial support. Also, the low activity of the nonambulatory children is an important limitation in interpreting growth using the NCHS growth charts. Weight gain was clinically significant, particularly when compared at the fifth percentile. Comparison with this lower percentile may be more appropriate in clinical applications because the group as a whole was so small in head size and weight at the first assessment. The lack of statistical significance of the Pearson product moment correlations between changes in weight for age and triceps skinfolds fat mass may be explained by assumption of a linear relationship in the statistic. Linear growth would mean crossing percentiles as seen in catch-up growth. The Pearson correlation was selected because all children had weight gain as an initial goal at the time of gastrostomy placement. Change in percentile over time, however, is not usual; normally growing children stay at the same percentile as weight increases. This study showed that growth as an outcome of gastrostomy placement in severely impaired children cannot be consistently shown by usual methods. Improvement in weight-for-age percentiles and z scores only described half of the children. Interfering conditions (scoliosis and contractures) limited the validity of length measurements over time. Consequently, estimation of energy needs based on kilocalories per centimeter of height or other body mass index calculations are not helpful in patients with such orthopedic complications (22). Although low weight was a consistent finding, this may not be interpreted as poor growth by some families. Families who lift children often prefer weight maintenance rather than growth. Even though the need for weight gain was noted at the time of
gastrostomy placement, children were observed long enough for this goal to be reached and modified, particularly for older children. This treatment goal would violate a basic assumption of the Pearson product moment statistic; no linear change over time would be anticipated. Clinical outcome indicators of growth for the children studied are suggested by our results. The most sensitive indicators of growth changes with feeding by gastrostomy over time were weight-age equivalent and triceps skinfolds fat mass. The usefulness of tracking triceps skinfolds fat mass in this group of patients was shown. The skinfolds measurement includes skin and fat only (23,24). Because it does not include muscle mass, it would be less likely to reflect the underlying neuromuscular diagnosis in these children. Triceps skinfolds thickness remains stable in age groupings of well-nourished children but changes in marasmic children compared with obese children (25). Triceps fat is related to effective energy intake and to body energy reserves (25). This is a more sensitive clinical indicator of changes in nutritional status over time than weight for age. There are many measurements of skinfolds among various populations, which show that the triceps site has been most completely analyzed (23). Measurements of triceps skinfolds fat mass have been used for patients with a variety of chronic illnesses, such as those undergoinghemodialysis (26) and young children with ascaris (25), and for pregnant women (27). Midarm circumference measurement would include muscle and bone as well as skin and fat (23). Small muscle size is a common finding in patients with cerebralpalsy and microcephaly, so arm circumference may reflect the underlying diagnosis more than nutritional intake. We found similar weight gain response to feeding bygastrostomy as reported by two retrospective studies (9,10). A prospective study of surgical complications noted weight gain as an outcome (28). Our study had a narrower age range than the patients studied by Rempel (9), but fewer subjects younger than 2 years of age. Rempel found that children younger than 2 had a more positive response than older children (9). Our study agreed with the study of Shapiro et al (10) which did not find age at gastrostomy placement to be a factor in growth response. Other studies involving children with a similar degree of neurologic impairment and age range have reported that the children tolerate percutaneous gastrostomy and gain weight. Such results contribute to wider application of newer techniques (5,12,28). Microcephaly has not been reported previously in evaluating response to feeding by gastrostomy, but it is an important predic,JOt [RNAL OF THE AMERICAN DIETETIC ASSOCIATION / 853
tor of neurological impairment. It may be a predictor of small body size also and maybe useful in setting realistic goals for weight gain. Several limitations of our study are inherent in working with children with severe disabilities: lack of control of medications, frequent acute infections, and complex family reactions to the stress of caring for a child with disabilities. Other parameters of nutritional status were not measured, such as iron status and other sites of skinfolds fatmass. Regardless ofhowanthropometric changes are interpreted in these children, our results support gastrostomy placement as a mechanism to improve the nutritional status indicators of weight and fat mass. Normal growth expectations are not justified in children with severe neurologic impairments, but improvement in growth was found for all children overtime. Even though limited to 23% of the children, catchup growth was documented; that is, these children gained weight above the rate expected for children of the same chronologic age at the 50th percentile. The results of this subgroup support the initial recommendation for weight gain before gastrostomy placement. Feeding by gastrostomy corrected undernutrition as a limiting factor for the children in this subgroup.
For children with severe neurologic impairments, use of gastrostomy seems to be an effective method for correcting undernutrition and achieving weight gain or catch-up growth
APPLICATIONS This study supports dietetics practitioners in their efforts to achieve weight gain and even catch-up growth in children with severe neurologic impairments. It also supports the assumption that undernutrition limits growth of the child before gastrostomy placement. Such information may be helpful to families facing a decision about gastrostomy placement surgery. Nevertheless, health care practitioners should not overstate the growth expectations for families considering a gastrostomy. For children with gastrostomies, triceps skinfolds fat mass may be a more sensitive clinical indicator of change after placement than weight and height measurements. Triceps skinfolds fat mass and weight accretion are appropriate outcome measures of improved nutritional status after gastrostomy placement. · This studyfunded, in part, by MCH ProjectMCJ-019143 University Affiliated TrainingGrant to the Civitan InternationalResearch Center University of Alabama at Birmingham References 1. Gisel E, Patrick J. Identification of children with cerebral palsy unable to maintain a normal nutritional state. Lancet. 1988; 1: 283-286. 2. Griggs CA, Jones PM, Lee RE. Videofluoroscopic investigation of feeding disorders of children with multiple handicap. Dev Med Child Neurol. 1989; 31:303-308. 3. Jones PM. Feeding disorders in children with multiple handicaps. Dev Med Child Neurol. 1989; 31:398-406. 4. Stuart SP, Tiley EH, Boland JP. Feeding gastrostomy: a critical 854 / AUGUST 1994 VOLUME 94 NUMBER 8
review of it indications and mortality rate. South Med J 1993; 86:169172. 5. Taylor CA, Larson DE, Ballard DJ, Bergstrom LR, Silverstein MD, Zinsmeister AR, Dimagno EP. Predictors of outcome after percutaneous endoscopic gastrostomy: a community-based study. Mayo Clin Proc. 1992; 67:1042-1049. 6. McGrath SJ, Splaingard ML, Alba HM, Kaufman BH, Glicklick M. Survival and functional outcome of children with severe cerebral palsy following gastrostomy. Arch Phys Med Rehabil. 1992; 73:133-137. 7. Eyman RK, Call TL, White JF. Life expectancy of persons with Down syndrome. Am J Ment Retard. 1991; 95:603-612. 8. Bax M. Eating is important. Dev Med Child Neurol. 1989; 31:285286. 9. Rempel GR, Colwell SO, Nelson RP. Growthin children with cerebral palsy fed via gastrostomy. Pediatrics. 1988; 82:857-862. 10. Shapiro BK, Green P, Krick J, Allen D, Capute AJ. Growth of severely impaired children: neurological versus nutritional factors. Dere Med Child Neurol. 1986; 28:729-733. 11. Wheatley MJ, Coran AG, Wesley R, Oldham KT, Turnage RH. Redo fundoplication in infants and children with recurrent gastroesophageal reflux. JPediatrSurg. 1991; 26:758-761. 12. Albanese CT, Towbin RB, Ulman I, Lewis J, Smith SD. Percutaneous gastrojejunostomy versus Nissan fundoplication for enteral feeding of the neurologically impaired child with gastroesophageal reflux. J Pediatr. 1993; 123:371-375. 13. Martinez D, Ginn-Peasem ME, Caniano DA. Sequelae of antireflux surgery in profoundly disabled children. JPediatrSurg. 1992; 27:267273. 14. Hassett JM, Sunby C, Flint LM. No elimination of aspiration pneumonia in neurologically disabled patients with feeding gastrostomy. Surg Gynecol Obstet. 1988; 167:383-388. 15. Lazares B, Murphy JB, Culpepper L. Aspiration associated with long-term gastric versus jejunal feeding: a critical analysis of the literature. Arch Phys Med Rehabil. 1990; 71:46-53. 16. Wilkinson JD, Dudgeon DL, Sondheimer JM. A comparison of medical and surgical treatment of gastroesophageal reflux in severely retarded children. JPediatr.1981; 99:202-205. 17. National Center for Health Statistics. Height and Weight of Youths 12 17. Washington, DC: Mental Health Administration, Public Health Service; 1973. Vital and Health Statistics. Series 11. No 124. 18. National Center for Health Statistics. NCHS Growth Curves for Children 0-18 Years Washington, DC: Mental Resources Administration, Public Health Service, 1977. Vital and Health Statistics. Series 11. No. 165. 19. Frisancho AR. Triceps skinfold and upper arm muscle size norms for assessment of nutritional status. Am J Clin Nutr. 1974; 27:10521057. 20. Frisancho AR. New norms of upper limb fat, and muscle areas for assessment of nutritional status. Am J Clin Nutr. 1981; 34:2540-2545. 21. Hendricks KM, Walker WA. Manual of PediatricNutrition. 2nd ed. Philadelphia, Pa: B. C. Decker, 1990:59-69. 22. Krick J, Murphy PE, Markham JFB, Shapiro BK. A proposed formula for calculating energy needs of children with cerebral palsy. Dev Med Child Neurol. 1992; 34:481-487. 23. Eveleth PB, TannerJM. Worldwide Vaiation in Human Growth. 2nd ed. New York, NY: Cambridge University Press, 1990. 24. Gregory JW, Greene SA, Scringeour CM, Rennie MJ. Body water measurements in growth disorders: a comparison of bioelectrical mlpedance and skinfold thickness techniques with isotope dilution. Arch Dis Child. 1991; 66:220-222. 25. Rammohan M, Aplasca EC. Caliper method vs bioelectrical impedance analysis for determining body fat in patients undergoing chronic dialysis and in health individuals. JAm DietAssoc. 1992; 92:1395-1396. 26. Stephenson LS, Crompton DWT, Latham MC, Schulpen TWJ, Nesheim MC, Janesen AAJ. Relationships between ascaris infection and growth of malnourished preschool children in Kenya. Am J Clin Nutr. 1980; 33:1165-1172. 27. Neggers Y, Goldenberg RL, Cliner SP, Hoffman HJ, Cutter GR. Usefulness of various maternal skinfold measurements for predicting newborn birth weight. JAm Diet Assoc. 1992; 92:1393-1394. 28. King SJ, Chait PG, Daneman A, Pereira J. Retrograde percutaneous gastrostomy: a prospective study in 57 children. PediatrRadiol. 1993; 23:23-25.