Patterns of postnatal weight changes in very-low-birth-weight infants.

Factors that affect initial weight loss and gain in infants < 1.501 grams are numerous. The purpose of this retrospective study was to examine short-term, postnatal weight changes in the very-low-birth-weight (VLBW) and extremely-low-birth-weight (ELBW) infants at a university hospital. Sixty-two charts of infants admitted to the University of Utah Hospital Neonatal Intensive Care Unit from July 1990 through November 1992 were reviewed. Data were collected on a three-part data collection tool and included demographic and treatment variables. Infants < 1,001 grams were categorized as ELBW, and infants 1,001 to 1,500 grams were categorized as VLBW. Each group was comprised of 31 infants. Fifty percent of the sample was male and 50% were female. Eighty-five percent of the samples were Anglo-American, and 15% were non-Anglo-American. There was no significant difference in the time to return to birth weight between the two groups. However, there was a significant difference in the maximum percent weight lost between the two groups, with the ELBW group losing a mean of 14.77% of birth weight and the VLBW group losing a mean of 11.35% of birth weight (t = 2.45, p < .05). Factors associated with postnatal weight changes were intraventricular hemorrhage, use of diuretics and steroids, day of maximum weight lost, and maximum percent of weight lost. Significant multicollinearity exists among many of independent variables (p < .01 to p < .001). Number of days diuretics was given before return to birth weight correlated significantly with time to return to birth weight (r = .77, F = 26.66, p < .0001). Prospective research examining the variables that may affect weight gain and loss patterns in infants < 1,501 grams in needed; thus, intervention studies could be developed to minimize postnatal weight loss and to promote growth in these infants.

PATTERNS OF POSTNATAL WEIGHT CHANGES IN VERY-LOW-BIRTH-WEIGHT INFANTS by Sandra Lee Wilson A thesis submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Master of Science College of Nursing The University of Utah August 1993 Copyright C Sandra Lee Wilson 1993 All Rights Reserved THE UNIVERSITY OF UTAH GRADUATE SCHOOL SUPERVISORY COMMITTEE APPROVAL of a thesis submitted by Sandra Lee Wilson This thesis has been read by each member of the following supervisory committee and by majority vote has been found to be satisfactory. Chair: Karin T. Kirchhoff Susan J. Squire THE UNIVERSITY OF UTAH GRADUATE SCHOOL FINAL READING APPROVAL To the Graduate Council of the University of Utah: I have read the thesis of Sandra Lee Wilson in its final form and have found that (1) its format, citations, and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables, and charts are in place; (3) the final manuscript is satisfactory to the supervisory committee and is ready for submission to The Graduate School. Date Approved for the Major Department :-£i1{2~ Linda K. Amos Chair/Dean Approved for the Graduate Council dtnvkJ.~ Ann W. Hart Dean of The Graduate School ABSTRACT Factors that affect initial weight loss and gain in infants < 1,501 grams are numerous. The purpose of this retrospective study was to examine short-term, postnatal weight changes in the very-low-birth-weight (VLBW) and extremely-low-birth-weight (ELBW) infants at a university hospital. Sixty-two charts of infants admitted to the university of Utah Hospital Neonatal Intensive Care unit from July 1990 through November 1992 were reviewed. Data were collected on a three-part data collection tool and included demographic and treatment variables. Infants < 1,001 grams were categorized as ELBW, and infants 1,001 to 1,500 grams were categorized as VLBW. Each group was comprised of 31 infants. Fifty percent of the sample were male and 50% were female. Eighty-five percent of the sample were Anglo-American, and 15% were non-Anglo­American. There was no significant difference in the time to return to birth weight between the two groups. However, there was a significant difference in the maximum percent weight lost between the two groups, with the ELBW group losing a mean of 14.77% of birth weight and the VLBW group losing a mean of 11.35% of birth weight (t = 2.45, ~ < .05). Factors associated with postnatal weight changes were intraventricular hemorrhage, use of diuretics and steroids, day of maximum weight lost, and maximum percent of weight lost. Significant multicollinearity exists among many of the independent variables (R < .01 to R < .001). Number of days diuretics was given before return to birth weight correlated significantly with time to return to birth weight (~= .77, E = 26.66, R < .0001). Prospective research examining the variables that may affect weight gain and loss patterns in infants < 1,501 grams is needed; thus, intervention studies could be developed to minimize postnatal weight loss and to promote growth in these infants. v TABLE OF CONTENTS ABSTRACT • LIST OF TABLES • ACKNOWLEDGMENTS Chapter I. INTRODUCTION Research Questions II. LITERATURE REVIEW. Conceptual Framework III. METHOD Design Sample Instrument Coding Rules • Procedures Operational Definitions Data Analysis . IV. RESULTS • V. DISCUSSION APPENDIX: DATA COLLECTION INSTRUMENT--WEIGHT CHANGES REFERENCES . Page iv vii viii 1 2 4 8 10 10 10 12 13 15 16 18 20 30 36 41 Table 1. 2. 3. 4. 5. 6. LIST OF TABLES Mean, Standard Deviation, and Range of Characteristics of Extremely-Low-Birth-Weight and Very-Low-Birth-Weight Infants Who Returned to Birth Weight • • • . • • • • • comparison of Mean Time to Return to Birth Weight between Extremely-Low-Birth-Weight and Very-Low-Birth-Weight Infants • • • • • Mean, Standard Deviation, and Standard Error of the Mean of Maximum Percent of Weight Lost between Extremely-Low-Birth-Weight and Very-Low-Birth-Weight Infants • • • • • • • Correlation of Demographic and Treatment Variables with Time to Return to Birth Weight in Fifty-Nine Infants • • • • • • Correlation Matrix of Twenty-Two variables in Fifty-Nine Infants Who Returned to Birth Weight Multiple Correlation Coefficient and Beta Weights for Predictors and Time to Return to Birth Weight • . • • . • • . • • • • • Page 21 24 24 25 27 29 ACKNOWLEDGMENTS I wish to thank my supervisory committee, Gary Chan and Sue Squire, for their time and support. I would especially like to thank the chair of my committee, Karin Kirchhoff, for her guidance, encouragement, and availability. Her support was limitless, and she encouraged me to go beyond my level of knowledge, to learn more, and to explore new areas of inquiry. Finally, thanks go to my husband, Jim Smith, who encouraged and supported me during my graduate education. CHAPTER I INTRODUCTION Infants weighing 500 to 1,500 grams and born at 22 to 25 weeks gestation are being successfully resuscitated and are surviving in greater numbers due to the advances of technology in Neonatal Intensive Care units (NICUs). For example, between 1975 and 1988, survival of very-low-birth-weight (VLBW) infants weighing 1,001 to 1,500 grams at birth increased from 56% to 77%, whereas survival of extremely-low-birth-weight (ELBW) infants weighing < 1,001 grams also increased from 24% to 59% in two New England hospitals (Pursley & Cloherty, 1991). Studies of postnatal weight change, defined as initial weight loss and regain to birth weight, are limited in VLBW and ELBW infants. Many of these studies were reported before the increased survival of ELBW infants. Therefore, understanding weight change becomes a significant issue with regard to the care and length of hospital stay for infants < 1,501 grams. In particular, weight loss, followed by poor or delayed weight gain, contributes to the morbidity of these infants and may, in fact, prolong hospitalization and increase costs. The length of time required by the VLBW and ELBW 2 infant to regain birth weight is variable. Numerous studies describe early postnatal weight loss and infant body composition; however, studies describing the patterns of weight gain after the initial postnatal weight loss are limited. In addition to the limited number of studies, most were published 6 to 9 years ago. Due to the advances of technology and increased knowledge regarding care of infants < 1,501 grams, the time to regain birth weight in these infants merits additional study. Research Questions The purpose of this descriptive, retrospective study was to examine postnatal weight changes in VLBW and ELBW infants at the University of Utah Hospital. The research questions were as follows: 1. What is the mean length of time required for the VLBW and ELBW infants to regain birth weight. 1a. Is there a difference between the means of these two groups? 2. What is the mean percent of body weight that is lost by the VLBW and ELBW groups? 3. Is there a relationship between day of life the infant reached 80 parenteral kilocalories/ kilogram/day (Kcal/kg/day) (maintenance parenteral Kcal/kg/day) and time to regain birth weight? 4. Is there a relationship between day of life the infant is started on total parenteral nutrition (amino acid and intralipid) and time to regain birth weight? 5. Is there a relationship between day of life the infant reached 120 enteral Kcal/kg/day (maintenance enteral Kcal/kg/day) and time to regain birth weight? 6. Is there a relationship between day of life patent ductus arteriosus ligation surgery occurred and length of time to regain birth weight? 7. Is there a relationship between the day of life an infant is moved to an incubator and time to regain birth weight? 3 8. Which of the independent variables contributes significantly to the dependent variable (time to regain birth weight)? CHAPTER II LITERATURE REVIEW Early postnatal weight loss in neonates is attributed to loss of total body water (Bauer & Versmold, 1989; Brosius, Ritter, & Kenny, 1984) and ranges from 7.9% to 14.6% of birth weight (Brosius et al., 1984; Shaffer, Quimiro, Anderson, & Hall, 1987). The major route of water loss in the premature infant is via the skin (transepidermal water loss) and respiratory system (Bell & Oh, 1983; Fanaroff, Rand, Wald, Gruber, & Klaus, 1972; Wu & Hodgman, 1974). Transepidermal water loss increases as gestational age decreases, and it was found to be 15 times greater in infants born at 25 weeks gestation than in full-term infants (Sedin, Hammarlund, & Stroemberg, 1983). Birth weight was regained in infants 620 to 900 grams by day of life (DOL) 22 and in infants between 930 to 1,100 grams by DOL 15 (Brosius et al., 1984). Shaffer et ale (1987) found that similar weight gain patterns occurred in infants 500 to 1,500 grams, with birth weight being regained between 15 to 25 DOL. V.d.Wagen, Okken, Zweens, and Zijlstra (1985) studied small for gestational age (SGA) infants born at 32 to 39 weeks and found that these infants regained birth weight by 8 to 15 DOL. Decsi and Fekete (1990) found that weight gain steadily increased after 2 weeks of postnatal life but was less than intrauterine fetal controls in 20 VLBW infants. 5 Although Dancis, O'Connell, and Holt (1948) developed weight curves 4 decades ago, they are still used in NICUs for infants weighing < 2,501 grams at birth based on weight curves of 100 infants with birth weights of 1,000 to 2,500 grams. A small group of infants weighing 750 grams had a weight curve plotted with birth weight regained by DOL 32. Jaworski (1974) developed a weight chart for premature infants based on mean weights of 235 premature infants with birth weights of 907 to 2,268 grams. This curve also is used in the University of Utah Hospital NICU. A third growth chart that is used in the University of Utah Hospital NICU was created by Babson (1970). This growth chart includes weight, length, and head circumference and was based on data collected on 36 infants weighing 950 to 2,000 grams at birth. Infants in all three studies were fed via the enteral route, and none of the infants was given total parenteral nutrition (Babson, 1970; Dancis et al., 1948; Jaworski, 1974). In addition to postnatal growth charts, intrauterine growth charts are available and used. Lubchenco, Hansman, and Boyd (1966) and Usher and McLean (1969) developed intrauterine fetal growth curves including weight, head circumference, and length. Lubchenco et al. extrapolated the weights for the 24~-week and 25-week gestation infant since their sample did not include infants < 26 weeks gestation. Usher and McLean included 25-week gestation infants in their sample. The growth curves are similar; however, the weight curves in the Lubchenco et al. (1966) study showed infants to be smaller. Both studies noted that the altitude in which these studies were conducted could have been a confounding variable. Usher and McLean studied infants 100 feet above sea level, whereas Lubchenco et al. studied infants 5,000 feet above sea level. 6 Feeding regimens and caloric intake were deemed to be major variables in weight gain of infants < 2,500 grams (Brosius et al., 1984; Dancis et al., 1948; Shaffer et al., 1987; V.d.Wagen et al., 1985). Several studies included infants fed only enterally with formula or breast milk (Babson, 1970; Brosius et al., 1984; Dancis et al., 1948; Decsi & Fekete, 1990; V.d.Wagen et al., 1985). Shaffer et al. (1987) included infants in their study on parenteral and enteral nutrition regimens. other variables that were not addressed specifically that may affect weight gain include the premature infant's environment, respiratory support needed by the infant, sepsis, ligation of a patent ductus arteriosus, severity of illness, gestational age, sex, and race. Other variables, beyond the scope of this study, that also may affect neonatal weight gain in these infants include maternal factors such as smoking, age, drug use, and disease. 7 Bauer et al. (1991) studied postnatal weight changes in infants < 1,500 grams, finding that initial weight loss was due to loss of total body water but that loss of lean body tissue did not occur. The infants in this study were given parenteral amino acids on admission in the amount of 1 gram/kilogram/day, with the concentration of amino acids increased to 2.4 grams/kg/day by the 3rd day. Energy intake for this group of infants was 26±7 Kcal/kg/day on DOL 1 and was increased up to DOL 5. Energy calories (carbohydrates + lipids) for these infants was 86±15 Kcal/kg/day during the time infants were gaining weight. In addition, premature infants 25 to 33 weeks gestation and 4 to 55 days postnatal age who were given 80 Kcal/kg/day of parenteral carbohydrates and lipid plus amino acids gained more weight than infants on 52 Kcal/kg/day of parenteral nutrition (Zlotkin, Bryan, & Anderson, 1981). Technology in the care of premature infants has grown tremendously over the past 20 years. Infants as small as 500 grams and 22 weeks gestation are being resuscitated successfully and are surviving. The growth charts of Babson (1970), Dancis et al. (1948), and Jaworski (1974) do not include these tiny infants. Babson (1970) 8 extrapolated growth curves for all groups of infants in his study, whereas Dancis et ale (1948) extrapolated his growth curve for infants weighing 750 grams based on "a very few infants of this weight" (p. 572). This lack of growth curve for infants ~ 1,000 grams may be due to the fact that the VLBW and ELBW infants were not part of the population during the time when these studies were conducted, or they were included in limited numbers. In addition, the studies previously conducted included healthy preterm infants. Babies who are 22 to 25 weeks gestation require more sophisticated care and technology during their hospital stays, including surgery for patent ductus arteriosus, respiratory support, antibiotics, prolonged intravenous nutrition, and special environmental settings. Conceptual Framework Infant growth requires energy conservation, along with intake of an energy source (nutrients). Energy conservation in the premature neonate requires a multifactorial approach from the caretakers. Levine's Conservation Model (Schaeffer & Pond, 1991) is the conceptual framework for this study. This framework is based on the premature infant's need to conserve energy and the infant's dependence on caretakers to provide an environment conducive to energy conservation, growth, and adaptation to the intensive care environment. 9 The four conservation principles of Levine's Conservation Model (Schaeffer & Pond, 1991) are "conservation of energy, conservation of structural integrity, conservation of personal integrity, and conservation of social integrity" (Fawcett, 1991, p. 26). Preterm infants are completely dependent on their caregivers to provide an external environment conducive to maintaining these four principles. Nursing interventions that promote energy conservation in the neonate include supplying nutrients, maintaining a quiet thermoneutral environment, and positioning the infant in a flexed position. The nurse assists the infant in maintaining structural integrity by using an aseptic technique when providing hands-on care, in addition to the previously mentioned activities. Personal and social integrity are promoted by the nurse by involving the family in the infant's care and by educating the family on the principles of energy conservation and structural integrity. CHAPTER III METHOD Design A retrospective descriptive design was used to explore and describe the phenomenon of postnatal weight loss and regain to birth weight in ELBW and VLBW infants. The dependent variable was time to return to birth weight. Twenty-four independent variables were identified by the investigator. Chart review of discharged infants was the data collection method. sample A purposive sample of 62 infants was selected based on the following parameters: (a) appropriate for gestational age (AGA) infants (defined by maternal ultrasound, Dubowitz score, and maternal dates) and (b) born at the University of Utah Hospital with a birth weight of < 1,501 grams. Infants with congenital anomalies, multiple gestations, inborn errors of metabolism, SGA, and large for gestational age (LGA) were excluded. These infants may have been born to mothers with disease states that affect fetal growth and development, which was beyond the scope of this study. The sample size of 31 infants in each group (VLBW and 11 ELBW) for Research Question 1 was determined by power analysis, with the alpha set at .05 (one-tailed test) and the power set at .75, with an effect size of .6 (Cohen, 1988). A sample size of 59 infants was determined by power analysis for the remaining research questions. The alpha was set at .05 (one-tailed test); the power was set at .75; and an effect size (~) was set at .3 (Cohen, 1988). Since the total number of subjects required for Research Question 1 was 3 greater than those needed for the remaining research questions, 62 infants were included in the sample. All infants received routine nursing care per NICU policy. Infants in the ELBW group experienced three different environments during hospitalization: (a) swamp, (b) heat shield, and (c) incubator. These infants were admitted initially to an open radiant warmer bed with a bed scale. A four-sided, 0.5 cm thick Plexiglas~ shield with internal dimensions of 52.5 cm long X 29.5 cm wide X 19 cm high was placed around the infant and then covered with plastic wrap. Humidified air (80% to 90% moisture) that was warmed to 36°C to 37°C was funneled into the covered Plexiglas~ shield, providing a "swamp" (warm and humid) atmosphere. Infants were kept in the swamp for 72 hours. After 72 hours, the humidity usually was discontinued. The Plexiglas~ shield with the plastic wrap cover (heat shield) remained over the infant. The infant 12 stayed in the heat shield as long as the radiant warmer was used. The heat shield was discontinued when the infant was moved into a double-walled incubator. Infants in the VLBW group experienced the heat shield and incubator environments. Infants were weighed by staff nurses on two different scales. Weights were taken using bed scales in the open warmer bed or in the incubator. Despite the environments utilized to provide neutrothermal stability for these infants, nursing care and interventions were disruptive and caused a loss of the environmental integrity that may have led to, cold stress and increased energy expenditure by the infant. Duxbury, Henly, Broz, Armstrong, and Wachdorf (1984) noted that nursing interventions accounted for 79% of all infant disruptions, with the average number of disruptions at two per hour, for an average of 14 minutes per hour of disruptions. These disruptions may have affected the infants' energy requirements, resulting in weight changes. Instrument The investigator-completed data collection tool consisted of three sections. The first section provided a framework to collect demographic variables such as date of birth, sex, gestational age in weeks, and race. In addition to the demographic variables, the DOL that various treatments and/or interventions occurred were noted. The second section was a chart for data relevant 13 to infant intake in Rcal/kg/day. The final section of the data collection tool was a chart for comments and the following confounding variables: (a) sepsis, (b) rule out necrotizing enterocolitis, and (c) need for increased Rcal > 120 Rcals/kg/day (see the Appendix). Coding rules established by the investigator were used to facilitate consistent use of the instrument. Coding Rules An average weight loss of > 15% total body weight, with an increase in weight of > 15% in a 24-hour period, was not accepted as accurate. These large weight changes were considered erroneous and likely due to scale error or scale use error. The day enteral feedings were begun and the day full enteral feedings were accomplished were coded as two variables. First, enteral feedings are frequently started early in neonatal life in quantities as small as 1 milliliter every 12 hours. Although this amount of enteral feeding does not provide significant Rcal/kg/day, it was counted as the DOL enteral feedings were begun. Second, the DOL full enteral feedings were accomplished was the day the infant received enteral feedings without IV fluid supplementation. VLBW and ELBW infants were given intravenous fluids upon admission to the University of Utah Hospital NICU. The fluid used on admission of the VLBW infant was 10% 14 dextrose water and for the ELBW infant was 5% dextrose water. Amino acids and intravenous lipid emulsions are started at 0.5 grams/kilogram/day and are increased by 0.5 grams/kilogram/day to a maximum of 3 grams/kilogram/day. Parenteral and enteral Kcal/kg/day were determined by the investigator, using the weight of the infant from the previous night shift flow sheet and the following equations. For parenteral intake, ml/kg/day were calculated, based on the total intake for each 24-hour period (0700 to 0700); then, this figure was multiplied by the Kcal/ml of the dextrose solution that the infant received, based on an energy density of 3.4 kilocalories per gram for intravenous dextrose, to achieve Kcal/kg/day (Crouch & Rubin, 1991). Lipid emulsions given intravenously have energy densities of 1.1 Kcal/ml for 10% and 2 Kcal/ml for 20% emulsions (Crouch & Rubin, 1991). The calculation used to determine total intravenous fat calories was ml/kg/day of lipid solution X Kcal/ml (1.1 or 2.0, depending on the lipid solution used) = total Kcal of lipid/kg/day. The totals of the carbohydrate and fat Kcal/kg/day were added, resulting in the total parenteral Kcals/kg/day. Enteral Kcal/kg/day were calculated, using the following formula: (Formula Kcal/ounce + additive Kcal/ounce) X ml/kg/day of formula X 1 ounce/30 ml = total Kcal/kg/day (Crouch & Rubin, 1991). Additives include 15 medium chain triglyceride oil, polycose, microlipid, and human milk fortifier. Premature babies require 80 Kcal/kg/day for growth if nutrition was provided via the parenteral route and 120 Kcal/kg/day if nutrition was provided enterally (Neu, Valentine, & Meetze, 1990). Therefore, 80 Kcal/kg/day for parenterally fed infants and 120/Kcal/kg/day for enterally fed babies was noted the day the infant reached these goals. Confounding variables that may have affected weight changes included documented sepsis/infection, need for increased caloric intake> 120 Kcal/kg/day, and suspected or confirmed necrotizing enterocolitis. Documented sepsis/infection was a positive blood, urine, skin, or tracheal aspirate culture and treatment with antibiotics for at least 7 days. Need for increased caloric intake > 120 Kcal/kg/day was documented poor weight gain on 120 Kcal/kg/day, and the infant was placed on Kcal in excess of 120 Kcal/kg/day. Suspected or confirmed necrotizing enterocolitis included infants who had enteral feedings discontinued for suspected necrotizing enterocolitis, as documented by the physician or neonatal nurse practitioner. Procedures Data were collected from patient medical records, including demographic infant characteristics, growth variables, and other confounding variables. Operational 16 definitions of the dependent variable, VLBW, and ELBW infant groups were established prior to data collection. Medical records were reviewed by the principal investigator. Data were obtained from various parts of the record including, but not limited to (a) nurses' notes, (b) neonatal nurse practitioner progress notes, (c) physician progress notes, (d) laboratory report slips, and (e) medication sheets. Privacy of the subjects was protected by assignment of an identification number to each infant. Assigned numbers were cross-referenced with the infant's medical record number and were kept in a separate file. operational Definitions The dependent variable was the length of time to regain birth weight and was defined as the day the infant reached birth weight, followed by a minimum average weight gain of 10 to 15 grams/kg/day (Crouch & Rubin, 1991; D'Harlingue & Byrne, 1991). Some infants return to birth weight due to fluid retention, not anabolic growth. Therefore, the infant must have demonstrated anabolic growth after attaining birth weight in order to be considered returned to birth weight. Parameters that demonstrate anabolic growth were defined as (a) serum albumin of > 2.0 grams/dl within 1 week of attaining birth weight (Reading, Ellis, & Fleetwood, 1990), (b) serum sodium of 130 to 145 mEq/l (Thomas & Reichelderfer, 1968) 17 within 1 week of attaining birth weight, (c) average weight gain of 10 to 15 grams/kg/day for 7 days after attaining birth weight, and (d) no edema documented in the neonatal nurse practitioner or physician progress notes within 1 week of attaining birth weight. The VLBW and ELBW infants differ based on birth weight and percent of total body water. The smaller the infant and the earlier the gestation, the higher the percentage of body weight is due to fluid. widdowson and Spray (1951) studied the chemical composition of 19 stillborn infants weighing 225 grams to 3,994 grams and 17 to 40 weeks gestation. Their analysis revealed that the smallest infants « 1,000 grams) were composed of 87% to 96% water, and, as the fetuses increased in size, the proportion of water decreased. Infants weighing < 1,501 grams but> 1,000 grams were composed of 80% to 85% water; 2,500-gram infants were composed of 77% water; and full­term infants were composed of 68.8% water (Widdowson & Spray, 1951). The ELBW AGA infant weighs < 1,001 grams. These infants are usually 27 to 28 weeks gestation or less, based on the intrauterine growth curves of Lubchenco et ale (1966) and Usher and McLean (1969). The VLBW AGA infant weighs < 1,501 grams but> 1,000 grams. These infants are usually 31 to 32 weeks gestation, but not less than 27-28 weeks gestation 18 (Lubchenco et al., 1966; Usher & McLean, 1969). Data Analysis According to VLBW and ELBW, infants were divided into two groups for Research Questions 1 and 2. The following statistics were calculated for each group: (a) mean length of time that it takes infants to regain birth weight, (b) standard deviation, and (c) range of time. In order to answer the first and second nondirectional research questions, a two-tailed student's t test was used to test the difference between the group means of the VLBW and ELBW groups. Alpha was set at .05 to evaluate the difference between the means of the VLBW and ELBW groups. Pearson product-moment correlation coefficients were used to analyze Research Questions 3, 4, 5, 6, and 7. The alpha was set at .05 (one-tailed test). The meaningfulness of ~ was calculated by squaring r. The square of the coefficient multiplied by 100 gave the percent of variance shared between the independent and dependent variables (Munro, Visintainer, & Page, 1986). Following are the independent variables for Research Questions 3, 4, 5, 6, and 7: (a) the DOL the infant reached maintenance parenteral Kcal/kg/day; (b) the DOL amino acid was begun; (c) the DOL intralipid was begun; (d) the DOL the infant reached maintenance enteral Kcal/kg/day; (e) the DOL patent ductus arteriosus ligation was done; and (f) the DOL the infant was moved to an incubator. The dependent variable was time to regain birth weight. 19 Multiple regression analysis was used to test Research Question 8. A correlation matrix of 25 variables was generated prior to the multiple regression analysis to assess the direction and strength of relationships among the variables. The independent variables were entered into the multiple regression equation using the stepwise method of entry. These variables included the number of days diuretics were given prior to return to birth weight, maximum percent of weight lost, and the DOL maximum weight loss occurred. These variables were included based on the relationship they had with the dependent variable in the correlation matrix (R < .01). In addition, the number of subjects affected by a particular variable was a consideration. For example, because only 5 infants received steroids before return to birth weight out of the entire sample (N = 59), the independent variables associated with steroid therapy prior to return to birth weight were not entered into the regression analysis equation. CHAPTER IV RESULTS Sixty-two charts of infants admitted to the University of Utah Hospital NICU from July 1990 through November 1992 were reviewed. Infants were placed in the ELBW group if their birth weight was < 1,001 grams and in the VLBW group if their birth weight was 1,001 to 1,500 grams. Each group had 31 infants. Two ELBW infants and 1 VLBW infant were transferred to another facility prior to regaining birth weight. Fifty percent of the sample were male and 50% were female, with 85% of the sample Anglo­American and 15% non-Anglo-American. Demographic and treatment variable means and standard deviations for each group and total mean, standard deviation, and range are displayed in Table 1. There are significant differences between the ELBW and VLBW groups for the following variables: (a) birth weight, (b) gestational age, (c) number of days of mechanical ventilation, (d) DOL infants were moved to incubators, (e) IVH, (f) DOL enteral feedings were begun, and (f) DOL the infant reached maintenance enteral Kcals (n < .001). In addition, significant differences between the two groups (n < .01) are seen in DOL amino acids and Table 1 Mean, Standard Deviation, and Range of Characteristics of Extremely-Low-Birth-Weight and Very-Low-Birth-Weight Infants Who Returned to Birth Weight ELBW variable m SD DOL returned to birth 15.55 4.68 weight Birth weight (grams) 812.10 141.95 Gestational age (weeks) 26.32 1.70 Number of days 35.61 20.15 mechanically ventilated DOL patent ductus arteriosus ligated DOL moved to incubator Intraventricular hemorrhage (grades 1 to 4) Number of days of diuretics prior to return to birth weight First DOL steroids given prior to return to birth weight Number of days steroids given prior to return to birth weight DOL enteral feedings begun 5.26 11.94 2.29 1.62 1.55 0.69 8.23 6.04 7.01 1.56 2.01 4.00 2.07 4.51 VLBW m 15.27 1262.90 29.52 6.61 2.58 4.42 1.16 1.33 0.43 0.17 4.52 SD 5.04 151.78 1.31 8.41 5.88 3.91 0.58 3.82 2.37 0.91 2.66 Total m SD 15.41 4.85 1037.50 269.80 27.92 2.01 21.11 21.17 3.92 8.18 1.73 1.48 1.00 0.42 6.37 6.06 6.83 1.31 3.04 3.29 1.60 4.12 Total range 7-29 n 59 510-1500 62 23-32 62 0-67 62 0-25 1-30 0-05 0-15 0-14 0-10 1-25 62 62 62 59 59 59 62 t -12.10*** - 8.27*** 7.40*** 5.16*** 3.74*** 3.94*** l\) ~ Table 1 (Continued) ELBW VLBW Total Variable Total n .t m SD m SD m SD ran2e DOL full feedings 17.48 6.71 12.42 6.13 15.00 6.87 5-45 62 3.10** accomplished DOL reached maintenance 3.71 7.02 1.87 4.76 2.80 6.02 0-28 62 total parenteral nutrition kilocalories DOL reached maintenance 21.71 10.58 13.03 9.22 17.37 10.77 0-41 62 3.44*** enteral kilocalories DOL amino acids begun 4.16 1.19 2.87 1.95 3.52 1.73 0-07 62 3.15** DOL intralipids begun 4.55 1.52 3.00 2.10 3.77 1.98 0-09 62 3.33** DOL maximum weight loss 7.16 2.72 5.94 2.07 6.55 2.47 3-13 62 2.00* occurred Maximum percent weight 14.77 6.48 11.35 4.30 13.07 5.72 3-33 62 2.45* lost Note. ELBW = extremely-low-birth-weight infants, VLBW = very-low-birth-weight infants, and DOL = day of life. *R < .05 for t tests between groups. **R < .01 for t tests between groups. ***R < .001 for t tests between groups. tro.J tro.J 23 intralipids are begun. Finally, significant differences are seen between the two groups in DOL maximum weight loss occurred and maximum percent of weight lost (~< .05). However, because multiple t tests were used to test the differences between the ELBW and VLBW groups on multiple variables, the high number of variables achieving the .05 level of significance is tentative and should be viewed with caution. There is no statistically significant difference in the mean time to return to birth weight between the ELBW and VLBW groups. The mean time to return to birth weight for the ELBW group is 15.55 days and for the VLBW group is 15.27 days (see Table 2). The mean maximum percent of weight lost in the ELBW group is 14.77% and in the VLBW group is 11.35%. The difference between these two means is statistically significant (R < .05), as shown in Table 3. The relationship between various demographic and treatment variables is shown in Table 4. Three infants were not included because they were transferred to another institution before returning to birth weight. There are no statistically significant relationships between time to return to birth weight and the following independent variables: (a) the DOL the infant reached 80 parenteral Kcals/kg/day, (b) the DOL total parenteral nutrition was started and time to return to birth weight, (c) the DOL 24 Table 2 Comparison of Mean Time to Return to Birth Weight between Extremely-Low-Birth-Weight and Very-Low-Birth-Weight Infants Group ELBW Cn = 29) VLBW Cn = 30) Total Time to return to birth weight 15.55 4.68 15.27 5.04 15.41 4.83 0.87 NS 0.92 0.63 Note. ELBW = extremely-low-birth-weight infants and VLBW = very-low-birth-weight infants. Table 3 Mean, Standard Deviation, and Standard Error of the Mean of Maximum Percent of Weight Lost between Extremely-Low­Birth- Weight and Very-Low-Birth-Weight Infants Maximum percent of weight lost Group m SD SEM :t. ELBW Cn = 31) 14.77 6.48 1.16 2.45* VLBW (n = 31) 11.35 4.30 0.77 Total 13.07 5.72 0.73 Note. ELBW = extremely-low-birth-weight infants and VLBW = very-low-birth-weight infants. *R < .05. Table 4 Correlation of Demographic and Treatment Variables with Time to Return to Birth Weight in Fifty-Nine Infants variable r Sex -0.19 Birth weight (grams) 0.02 Gestational age (weeks) -0.12 Number of days mechanically ventilated 0.20 Surgical vs. indocin ligation (grades 0 to 3) 0.21 DOL patent ductus arteriosus ligated 0.25 Intraventricular hemorrhage (grades 1 to 4) 0.28* DOL moved to incubator 0.19 Days diuretics given before return to birth 0.61** weight First DOL steroids given before return to 0.35** birth weight Number of days steroids given before return to 0.41** birth weight DOL enteral feedings begun DOL full feedings established DOL reached maintenance total parenteral nutrition kilocalories DOL reached maintenance enteral kilocalories DOL amino acids begun DOL intralipids begun DOL maximum weight loss occurred Maximum percent weight lost Note. DOL = day of life. *R < .05. **R < .01. 0.21 0.23 0.02 -0.17 0.02 0.14 0.38** 0.49* 25 26 the infant reached 120 enteral Kcals/kg/day and time to return to birth weight, (d) the DOL patent ductus arteriosus ligation occurred and time to return to birth weight, and (e) the DOL infant was moved to an incubator and time to return to birth weight. statistically significant relationships are seen between number of days diuretics were given and time to return to birth weight, first DOL steroids were given prior to return to birth weight and time, to return to birth weight, number of days steroids were given prior to return to birth weight and time to return to birth weight, DOL maximum weight loss occurred and time to return to birth weight, maximum percent of weight lost and time to return to birth weight, and intraventricular hemorrhage and time to return to birth weight. In effect, treatment variables appear to prolong return to birth weight, and nutritional variables do not appear relevant to shortening the time to return to birth weight. A correlation matrix of the demographic and treatment variables was produced prior to the multiple regression analysis (see Table 5). Multicollinearity exists among many of the variables. Variables that have a statistically significant relationship with time to return to birth weight are number of days diuretics were given prior to return to birth weight (~ = .61, R < .001, with 42% of the infants receiving diuretics prior to return to Table 5 Correlation Matrix of Twenty-Two Variables in Fifty-Nine Infants Who Returned to Birth Weight Variable 1 DOL RTBW 2 Birth weight (grams) 3 Gestational age (weeks) 4 Days of mechanical ventilation 5 Chest tubes (yes/no) 6 PDA ligation (yes/no) 7 Surgical vs. indocin (grades 0 to 3) 8 DOL PDA ligated 9 Intraventricular hemorrhage (grades 1 to 4) 10 DOL moved to incubator 11 Days of diuretics prior to RTBW 12 First DOL steroids given prior to RTBW 13 Days of steroid therapy prior to RTBW 14 DOL enteral feedings begun 15 DOL full feedings accomplished 16 DOL reached maintenance TPN kilocalories 17 DOL reached maintenance enteral kilocalories 18 DOL amino acids begun 19 DOL intralipids begun 20 Use of the swamp (yes/no) 3 .86** 4 5 6 7 -.75** .49** -.33* -.81** .52** .42** -.45** - .57** -.45** .47** .74** - .87** - .90** 8 9 10 -.44** -.53** -.31* -.46** - .60** .48** .72** - .57** -.51** -.49** -.71** - .37* -.55** .75** .37* .49** .31* .47** 11 .61** -.41** .39* .37* .35* - .32* .31* .43** .43** .40** (I,J -...J Table 5 (continued) Variable 1 DOL RTBW 2 Birth weight (grams) 3 GestationaL age (weeks) 4 Days of mechanical ventilation 5 chest tubes (yes/no) 6 PDA Ligation (yes/no) 7 SurgicaL vs. indocin (grades 1 to 4) 8 DOL PDA ligated 9 IntraventricuLar hemorrhage (grades 1 to 4) 10 DOL moved to incubator 11 Days of diuretics prior to RTBW 12 First DOL steroids given prior to RTBW 13 Days of steroid therapy prior to RTBW 14 DOL enteral feedings begun 15 DOL fuLL feedings accompLished 16 DOL reached maintenance TPN kilocaLories 17 DOL reached maintenance enteral kiLocalories 18 DOL amino acids begun 19 DOL intralipids begun 13 14 15 16 .41* -.41** - .38* - .48** - .50** .31* .56** .56** -.59** - .62** - .34* - .67** - .62** .61** .68** .35* .31* .48** .65** .37* .44** .55** .36* .50** .76** .66** .38* .31* .32* .87** .51** .40** .35* .64** .56** .51** .75** .61** .59** 17 18 19 20 21 22 -.38* .49** -.42** - .34* -.34* .79** -.34* - .40** -.41** .71** -.38* -.42** .35* .44** .47** - .69** .41** .44** -.39* .49** -.31* -.39* .35* .37* - .37* .30* .46** - .36* .41** .40** .30* .33** -.53** .39* .41** .33* .39* - .48** .33* .31* .34* - .34* .41** -.51** .91** Note. DOL = day of life, RTBW = return to birth weight, PDA = patent ductus arteriosus, and TPN = total parenteral nutrition. variable 21 = day of life maximum weight loss occurred and Variable 22 = maximum percent weight lost. *.Q < .01. **12. < .001. ~ (X) birth weight); first DOL steroids were given prior to return to birth weight (~ = .35, R < .01, with 8% of 29 infants receiving steroids prior to return to birth weight and the remaining infants not receiving steroids prior to return to birth weight); number of days steroids were given prior to return to birth weight (r = .41, R < .01); DOL maximum weight loss occurred (~= -.38, R < .01); and maximum percent of weight lost (~= .49, R < .001). stepwise multiple regression with the criterion variable of time to return to birth weight produced a multiple correlation coefficient of .77 (R < .0001) and beta weights for three predictors of time to return to birth weight (see Table 6). Table 6 Multiple Correlation Coefficient and Beta Weights for Predictors and Time to Return to Birth Weight Criterion variable Time to return to birth weight Multiple correlation coefficient .77* Note. DOL = day of life. n = 59 infants. I: = 26.66. *R < .0001. Beta weights for predictors Number of days Maximum DOL diuretics given percent maximum before return weight weight to birth weight lost lost 0.61 0.41 0.25 CHAPTER V DISCUSS lOB The results of this study do not support prolonged time to return to birth weight for infants < 1,SOl grams. The mean time to return to birth weight for infants in this study is lS.41 days, with a range of 7 to 29 days. Maximum weight loss in this study occurs at a mean of 6.SS days, with a range of 3 to 13 days, which does not concur with past studies. Growth charts of Babson (1970), Dancis et ale (1948), and Jaworski (1974), used in the University of Utah Hospital NICU, have various weight loss and gain patterns. The DOL that maximum weight loss occurred, based on the growth curve of Jaworski, was DOL 3, with a return to birth weight on DOL 9 for the 1,SOO-gram infant. The 1,200-gram infant had a maximum weight loss on DOL 6, with a return to birth weight on DOL 17. Babson showed no weight loss on his charts because he used intrauterine growth curves. Finally, Dancis et ale graphed maximum weight loss at DOL 6 for the 1,SOO-gram infant, DOL 8 for the 1,000-gram infant, and DOL 13 for the 7S0-gram infant. Return to birth weight occurred on DOL 14 for the 1,SOO­gram infant, DOL 17 for the 1,000-gram infant, and DOL 32 31 for the 750-gram infant. significant differences between the ELBW and VLBW groups among the variables (a) birth weight, (b) gestational age, (c) number of days of mechanical ventilation, (d) DOL moved to incubator, (e) IVH, (f) DOL enteral feedings begun, (g) DOL full feedings accomplished, (h) DOL reached maintenance enteral Kcals, and (i) DOL amino acids and intralipids were begun are not surprising (see Table 1). The ELBW infant is expected to be smaller and require more technology in a longer period of time than the VLBW infant, given the degree of immaturity of the ELBW infant. It is logical that birth weight and gestational age are different based on the definition of the ELBW infant and VLBW infant (Lubchenco et al., 1966; Usher & McLean, 1969; widdowson & Spray, 1951). Maximum percent of weight lost by the ELBW group is greater than the VLBW group (R < .05), which is expected when the body composition is considered. The ELBW infant is 87% to 96% water, and the VLBW infant is 80% to 85% water (Widdowson & Spray, 1951). The greater the water content, the more weight will be lost. This result was expected; however, this result must be interpreted with caution at the .05 level of significance since the t test was used to test each IV between the ELBW and VLBW groups. The day that maximum weight loss occurred is later in 32 the ELBW group than in the VLBW group (~ < .05); this may be due to environment. ELBW infants were in the high humidity environment for the first 720 of life and VLBW infants were not; however, this was not specifically tested for and is speculation. Given that the ELBW infant's body composition is 7% to 11% more water than the VLBW infant (Widdowson & Spray, 1951), it would seem more logical for the ELBW group to reach their nadir earlier than the VLBW group. Maximum percent of weight lost has a moderate effect on time to return to birth weight, and the DOL that maximum weight is lost has a small effect on time to return to birth weight. The amount of weight lost and the DOL maximum weight loss occurs are expected to be predictors affecting time to return to birth weight. However, treatment variables such as the DOL total parenteral nutrition is started would seem to have had more effect on the criterion variable than weight loss variables and diuretic therapy. The fact that treatment variables such as nutrition and environmental control did not have a moderate or strong effect on the criterion variable is surprising. Predictors of weight gain and time to return to birth weight in the VLBW and ELBW infant are obfuscated by the treatment variable diuretic therapy, thus making it difficult to determine which predictors enhance growth and 33 return to birth weight. Based on this retrospective analysis, diuretic therapy has an impact on time to return to birth weight. However, because of the large effect of diuretic therapy, other treatment variables are overshadowed and the effect the other variables may have on time to return to birth weight is difficult to evaluate. The amount of multicollinearity among the independent variables may have affected the coefficients in the regression analysis. The effect of this multicollinearity may have been less precision in the analysis. The relationships of the independent variables with the dependent variable are surprising. It was expected that environmental and nutritional interventions would have had a stronger, more positive correlation with time to return to birth weight. In fact, neither of these categories of variables has a significant effect on return to birth weight. An unexpected result of the study is the correlation of intraventricular hemorrhage with time to return to birth weight (R < .05). Diuretic therapy has a moderate effect on return to birth weight (R < .01). steroid therapy has a weak to moderate effect on return to birth weight (R < .01); however, the number of infants who received steroids was so small that this result may have been skewed by the small number of infants. All of these results are unexpected and surprising. 34 Multicollinearity presented a challenge in this study. As seen in Table 5, strong correlations exist among the independent variables. For example, birth weight and gestational age are strongly correlated with the number of days of mechanical ventilation (~< .001). The DOL full enteral feedings were accomplished is moderately to strongly correlated with (a) the DOL enteral feedings were begun, (b) surgical versus indocin ligation, and (c) the DOL the infant was moved to an incubator (~ < .001). Implications for future research include prospectively studying predictors that may affect time to return to birth weight. Then, controls can be implemented that are lacking in this study, such as use of a consistent scale and consistent scale use technique. Relationships among variables, as seen in Table 5, can be explored further in an effort to analyze variables that may affect weight-change patterns in these infants. Multicollinearity beyond inspection of the correlation matrix can be diagnosed (Schroeder, 1990). Tolerance of the independent variables and variance inflation factors can be evaluated in an effort to discard the independent variables that are significantly correlated. Eventually, interventions may be developed that can promote growth and return to birth weight in the VLBW and ELBW infant. Studies examining the differences in treatment 35 variables between ELBW and VLBW infants, as seen in Table I, also can be conducted to determine why these differences occur and if these differences are necessary. There are significant differences in the DOL the infants were moved into an incubator, the DOL enteral feeds were begun, and the DOL infants reached maintenance enteral Kcal, which all can be studied to determine if these differences are necessary and if they affect the infant's weight loss and gain. APPENDIX DATA COLLECTION INSTRUMENT--WEIGHT CHANGES 37 Date of Birth: ID#: Column # 1-2 Value ColUill ., Variable 0 1 2 3 4 5 6 7 8 9 10 1. Sex M F 3 2. Bi rth weight (grams) 4-7 3. Gestational age 22 23 24 25 26 27 28 29 30 31 32 8-9 (weeks) 4. Race CA B1 AI As Hi To Sa Ot 10 5. Number of days of 11-12 mechanical ventilation 6. Chest tubes Y N 13 7. Patent ductus Y N 14 arteriosus ligation 8. Surgical vs. indocin NA I S BO 15 9. DOL patent ductus 16-17 arteriosus ligated 10. IVH UK N Y1 Y2 Y3 Y4 18 11. DOL moved to 19-20 incubator 12. DOL returned to 21-22 birth weight 13. Number of days 23-24 diuretics given prior to return to birth weight 14. First DOl Iteroids 25-26 given prior to return to birth weight 15. Number of days of 27-28 steroids prior to return to birth weight 16. DOL enteral 29-30 feedings begun 17. HM vs. fornIJla BO H F 31 18. DOL full enteral 32-33 feedings accomplished 19. DOL reached 34-35 maintenance total parenteral nutrition Kcals 20. DOL reached 36-37 maintenance enteral Kcals 21. DOL AA begl.ll 38-39 38 VaLue CoLLiln # Variable 0 1 2 3 4 5 6 7 8 9 10 22. DOL IL begun 40-41 23. Swan., Y N 42 24. DOL maximum weight 43-44 lost 25. Maximum percent 45-46 weight lost 39 ID#: DOL Weight KcaL/kg/day Total Lab test Serlin Serlin Ma+ Edema percent (yes/no) albunfn body weight lost 10#: possible confounding variables: 1. Sepsis (documented problem, i.e., + blood, urine, skin cultures +/or tracheal asp with X ray changes and treatment with antibiotics for a full 7-day course. 2. Rule out necrotizing enterocolitis (R/O NEe). 3. Need for increased caloric intake> 120 Rcal/kg/day for work of breathing, poor weight gain on maintenance, Kcal. DOL Sepsis RIO NEe Increased Keals Comments 40 REFERENCES Babson, S. G. infants. (1970). Growth of low-birth-weight The Journal of Pediatrics, 77, 11-18. 42 Bauer, K., Bovermann, G., Roithmaier, A., Goetz, M., Proelss, A., & Versmold, H. T. (1991). Body composition, nutrition, and fluid balance during the first two weeks of life in preterm neonates weighing less than 1,500 grams. The Journal of Pediatrics, 118, 615-620. Bauer, K., & Versmold, H. (1989). Postnatal weight loss in preterm neonates < 1,500 g is due to isotonic dehydration of the extracellular volume. Acta Paediatrica Scandinavica Supplement, 360, 37-42. Bell, E. F., & Oh, W. (1983). water requirement of premature newborn infants. Acta Paediatrica Scandinavica Supplement, 305, 21-26. Brosius, K. K., Ritter, D. A., & Kenny, J. D. (1984). Postnatal growth curve of the infant with extremely low birth weight who was fed enterally. Pediatrics, 74, 778-782. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates. Crouch, J. B., & Rubin, L. P. (1991). Nutrition. In J. P. Cloherty & A. R. Stark (Eds.), Manual of neonatal care (pp. 526-558). Boston: Little, Brown. Dancis, J., O'Connell, J. R., & Holt, L. E. (1948). A grid for recording the weight of premature infants. The Journal of Pediatrics, 33, 570-572. Decsi, T., & Fekete, M. (1990). Postnatal weight gain and serum total protein and albumin levels in very low birth weight (1,500 g) preterm infants. Acta Paediatrica Hungarica, 30, 373-382. D'Harlingue, A. E., & Byrne, W. J. (1991). Nutrition in the newborn. In H. W. Taeusch, R. A. Ballard, & M. E. Avery (Eds.), Schaffer and Avery's diseases of the newborn (6th ed.) (pp. 709-727). Philadelphia: W. B. Saunders. Duxbury, M. L., Henly, S. J., Broz, L. J., Armstrong, G. D., & Wachdorf, C. M. (1984). Caregiver disruptions and sleep of high-risk infants. Heart & Lung, 13, 141-147. Fanaroff, A. A., Rand, M. B., Wald, M., Gruber, H. S., & Klaus, M. H. (1972). Insensible water loss in low birth weight infants. Pediatrics, 50, 236-245. 43 Fawcett, J. (1991). Analysis and conservation model. In K. M. (Eds.), Levine's conservation nursing practice (pp. 13-43). Davis. evaluation of Levine's Schaefer & J. B. Pond model: A framework for Philadelphia: F. A. Jaworski, A. A. (1974). New premature weight chart for hospital use. Clinical Pediatrics, 13, 513-516. Lubchenco, L. C., Hansman, C., & Boyd, E. (1966). Intrauterine growth in length and head circumference as estimated from live births at gestational ages from 26 to 42 weeks. Pediatrics, 37, 403-408. Munro, B. H., Visintainer, M. A., & Page, E. B. (1986). statistical methods for health care research. Philadelphia: J. B. Lippincott. Neu, J., Valentine, C., & Meetze, W. (1990). Scientifically-based strategies for nutrition of the high-risk low birth weight infant. European Journal of Pediatrics, 150, 2-13. Pursley, D. M., & Cloherty, J. P. (1991). Identifying the high-risk newborn and evaluating gestational age, prematurity, postmaturity, large-for-gestational-age, and small-for-gestational-age infants. In J. P. Cloherty & A. R. Stark (Eds.), Manual of neonatal care (pp. 85-103). Boston: Little, Brown. Reading, R. F., Ellis, R., & Fleetwood, A. (1990). Plasma albumin and total protein in preterm babies from birth to eight weeks. Early Human Development, 22, 81-87. Schaeffer, K. M., & Pond, J. B. (1991). Levine's conservation model: A framework for nursing practice. Philadelphia: F. A. Davis. Schroeder, M. A. (1990). Diagnosing and dealing with multicollinearity. Western Journal of Nursing Research, 12, 175-187. Sedin, G., Hammarlund, K., & Stroemberg, B. (1983). Transepidermal water loss in full-term and pre-term infants. Acta Paediatrica Scandinavica Supplement, 2, 27-31. 44 Shaffer, S. G., Quimiro, C. L., Anderson, J. V., & Hall, R. T. (1987). Postnatal weight changes in low birth weight infants. Pediatrics, 79, 702-705. Thomas, J. L., & Reichelderfer, T. E. (1968). Premature infants: Analysis of serum during the first seven weeks. Clinical Chemistry, 14, 272-280. Usher, R., & McLean, F. (1969). Intrauterine growth of live born Caucasian infants born between 25 and 44 weeks of gestation. Journal of Pediatrics, 74, 901-910. V.d.Wagen, A., Okken, A., Zweens, J., & zijlstra, W. G. (1985). Composition of postnatal weight loss and subsequent weight gain in small for dates newborn infants. Acta Paediatrica Scandinavica, 74, 57-61. Widdowson, E. M., & Spray, C. M. (1951). Chemical development in utero. Archives of Disease in Childhood, 26, 205-214. Wu, P. Y. K., & Hodgman, J. E. (1974). Insensible water loss in preterm infants: Changes with postnatal development and non-ionizing radiant energy. Pediatrics, 54, 704-712. Zlotkin, S. H., Bryan, M. H., & Anderson, G. H. (1981). Intravenous nitrogen and energy intakes required to duplicate in utero nitrogen accretion in prematurely born human infants. The Journal of Pediatrics, 99, 115-120.