BIRTH 41:2 June 2014
Charts for Weight Loss to Detect Hypernatremic Dehydration and Prevent Formula Supplementing Paula van Dommelen, PhD, Suzanne Boer, MSc, Sevim Unal, PhD, and Jacobus P. van Wouwe, PhD ABSTRACT: Background: Most breast-fed newborns get the milk they need. However, very rarely milk intake is insufﬁcient mostly as a result of poor breastfeeding techniques. Dramatic weight loss and hypernatremic dehydration may occur. Our aim was to construct charts for weight loss. Methods: A case–control study was performed. Charts with standard deviation score (SDS) lines for weight loss in the ﬁrst month were constructed for 2,359 healthy breast-fed term newborns and 271 cases with breastfeeding-associated hypernatremic dehydration with serum sodium level > 149 mEq/L. Day 0 was deﬁned as the day of birth. Results: Many cases with (or who will develop) hypernatremic dehydration (84%; +1 SDS line) fell below the 1 SDS line at day 3, the 2 SDS line at day 4, and the 2.5 SDS line at day 5 in the chart of the healthy breast-fed newborns. Weight loss of cases with permanent residual symptoms was far below the 2.5 SDS. Conclusions: Already at an early age, weight loss differs between healthy breast-fed newborns and those with hypernatremic dehydration. Charts for weight loss are, therefore, useful tools to detect early, or prevent newborns from developing, breastfeeding-associated hypernatremic dehydration, and also to prevent unnecessary formula supplementing. (BIRTH 41:2 June 2014)
Key words: breastfeeding, dehydration, growth
Breastfeeding is important to achieve optimal growth, development, and health (1). Worldwide programs are available focusing on improvement of support and encouragement of breastfeeding, resulting in an increasing prevalence of mothers initiating breastfeeding (2,3). With this positive trend and the increasing attention for breastfeeding, we need to be aware of inadequate breastfeeding and subsequent neonatal risks (4). Inadequate breastfeeding is mostly a result of poor breastfeeding techniques (incorrect positioning and latching), short or infrequent feedings, breastfeeding on scheduled times, food suppliers, or maternal illness (5,6). If the milk supply is insufﬁcient for several days,
dramatic weight loss and dehydration with increased serum sodium concentration may occur (7,8). Hypernatremic dehydration is a serious condition if not recognized early and treated properly (9–11). Complications that have been reported include signiﬁcant indirect hyperbilirubinemia, prerenal azotemia (9), renal failure, elevated liver enzymes (10), disseminated intravascular coagulation, intracranial hemorrhage, cavernous sinus thrombosis (10), seizure, brain edema (9), and death (11). The risk of developing hypernatremic dehydration is higher in newborns born by a cesarean section than for those born vaginally (12). Other risk factors are primi-
Paula van Dommelen is a Statistician in the Department of Life Style at TNO, Leiden, The Netherlands; Suzanne Boer is a Statistician in the Department of Life Style at TNO, Leiden, The Netherlands; Sevim Unal is a Pediatrician in the Neonatal Intensive Care Unit at the Ankara Children’s Hematology and Oncology Research Hospital, Ankara, Turkey; Jacobus P. van Wouwe is a Pediatrician in the Department of Child Health at TNO, Leiden, The Netherlands.
Address correspondence to Paula van Dommelen, Department of Life Style, TNO, P.O. Box 2215, 2301 CE Leiden, The Netherlands. Accepted January 14, 2014 © 2014 Wiley Periodicals, Inc.
BIRTH 41:2 June 2014
154 parity, delay of initiation of ﬁrst breastfeeding, newborn’s hunger cues are slight or nonexistent, use of the nipple shield, and infrequent feeding (7,13). Initial symptoms of cases with hypernatremic dehydration are jaundice, decreased bowel movements (7,13), decreased urination, dry mucous membranes, sunken fontanel, poor skin turgor, lethargy (7), thirst fever (13,14), and irritability (9,15). An important and obvious symptom of hypernatremic dehydration is a dramatic weight loss (16–19). It is suggested that weight loss exceeding 10 percent or failure to regain birthweight by 10 or 14 days are conventional cutoffs for seeking a pediatric opinion (17,20,21). To obtain the diagnostic values of such cutoffs, weight loss of both healthy newborns and dehydrated cases should be investigated. In countries with an early weighing policy and appropriate lactation support, the risk of hypernatremic dehydration is small (15,20). However, out of concern for hypernatremic dehydration, supplementing with formula is sometimes offered to newborns with weight losses that are still within normal ranges (22). Supplementing formula regularly hampers a successful breastfeeding, and milk supply will diminish (23). To prevent unnecessary supplementing, one needs to know the difference between a normal and an abnormal weight loss. Currently, no growth charts for weight loss in the ﬁrst month of life based on large sample sizes are available to deﬁne a normal weight loss. In addition, no growth chart exists for breastfeeding-associated hypernatremic dehydration to deﬁne an abnormal weight loss. In this study, we construct charts for weight loss in healthy breast-fed newborns and cases with breastfeeding-associated hypernatremic dehydration. The charts will be compared to investigate when and to what degree weight loss of healthy breast-fed newborns differs from those with hypernatremic dehydration. In addition, we will investigate the association between weight loss and serum sodium concentration and general neonatal and maternal characteristics.
Patients and Methods
practices from different parts of The Netherlands. Each midwife practice can have its own weighing policy. In The Netherlands, newborns are weighed at home by a midwife with a calibrated electronic scale. A midwife either assists the delivery at home or in an outpatient clinic, or is involved in the follow-up after hospital delivery by an obstetrician. Longitudinal weights after the neonatal period were retrospectively obtained from 21 participating child health clinics from different parts of The Netherlands. In The Netherlands, young children are weighed with a baby scale calibrated in decagrams at the average ages of 1 and 2 months in the child health clinics by trained health care professionals. Figure 1 shows the exclusion criteria. We excluded newborns with a condition. In our study, those conditions were Down syndrome, pyloric stenosis, intestinal abnormality, cleft lip and palate, arrhythmia, and left heart hypoplasia syndrome. We randomly selected measurements if the number of visits exceeded the regular number of visits according to the weighing policy of the primary care midwife practice. In addition, outliers of weight loss were removed. After exclusion, 2,359 newborns with birthweights and 4,475 additional weight measurements were available for analyses.
Breastfeeding-associated Hypernatremic Dehydration We obtained a sample of cases with hypernatremic dehydration from two studies (10,26). The ﬁrst study consisted of 169 cases that were predominantly breast-
INCLUDED 2,475 BREAST-FED NEONATES 4,823 weight measurements
EXCLUDED NEONATES WITH A CONDITION 22 neonates with 33 weight measurements
2,453 NEONATES 4,790 weight measurements
2,382 NEONATES 4,652 weight measurements
LOW BIRTHWEIGHT (149 mEq/L on the day of presentation. After this selection, a total of 271 cases with hypernatremic dehydration with weight measurements up to 22 days were available for analyses.
Relative Weight Change Relative weight change (RWC) (= - weight loss) was calculated as the difference in weight at day of visit (w(t)) and birthweight (w(t0)) divided by birthweight in percentage, or expressed in the following formula: 100% *(w(t) w(t0))/w(t0)). Day 0 was deﬁned as the day of birth. For example, a 5-day-old boy weighs 3,500 g and his birthweight is 3,700 g, then his RWC is 100% *(3,500–3,700)/3,700 = 5.4 percent and weight loss is +5.4 percent. This boy lost 5.4 percent of his birthweight in 5 days.
Charts for Weight Loss Centiles or standard deviation score (SDS) lines of the charts for RWC for the healthy breast-fed newborns between birth and 30 days and cases with hypernatremic dehydration between birth and 22 days were constructed using the GAMLSS package in R version 4.2. (27). Data between 30 and 60 days were used to obtain stable ﬁt results at the age of 30 days. The smoothing parameters were chosen by creating worm plots: local detrended quantile-quantile plots of the SDS of the reference samples across several age groups (28). The SDS expresses the measurement relative to a reference
population in units of standard deviations above or below the median (29). The 0 SDS line represents the median. At every age, 50 percent of newborns in the reference population fall above this line and 50 percent below it. Most newborns (95.4%) fall between the 2 and +2 SDS lines and almost all newborns (98.8%) fall between 2.5 and +2.5 SDS (30,31).
Associations with Weight Loss Associations between weight loss and serum sodium concentration and background characteristics were estimated by regression analyses. Multilevel linear regression analyses in the healthy breast-fed newborns were performed with RWC SDS as dependent variable and neonatal age (in days), neonatal age2, maternal age (in years), parity (ﬁrst vs second or more), mode of delivery (vaginal vs cesarean section), gender (male vs female), and birthweight (in kg) as independent variables. We performed multilevel analyses as measurements of weight loss are nested within newborns. Linear regression analysis was performed in the case sample with weight loss as dependent variable and serum sodium concentration, neonatal age, and the interaction between serum sodium concentration and neonatal age as independent variables. The descriptive statistics were performed in SPSS Version 20.0 for Windows (SPSS Inc., Chicago, IL, USA). The growth charts and the regression analyses were analyzed in R Version 3.0.1; p < 0.05 (two-sided) were considered statistically signiﬁcant.
Results Table 1 shows the general characteristics of the studies available for analyses. The ﬁrst two studies were healthy breast-fed newborns weighed in the ﬁrst 8 days after birth. Birthweights and weight measurements after 3 weeks were available in the third study. Mean maternal age was approximately 4 years and mean birthweight 200 g higher in these studies of healthy newborns compared with the study of those with hypernatremic dehydration. First parity (75 vs 40%) and vaginal delivery (90 vs 80%) were also higher. With the three pooled healthy studies and the two pooled hypernatremic dehydration studies, respectively, the LMS method and the Box–Cox t distribution in the GAMLSS package were applied, as these methods showed optimal worm plots. In the LMS method, a log transformation was applied to the age axis. The curves were ﬁtted as cubic splines. Figure 2 shows the RWC of the healthy breast-fed newborns together with the SDS lines of their chart
BIRTH 41:2 June 2014
Relative weight change (%) 0 10 20 30
birthweight. Many cases with hypernatremic dehydration (84%; +1 SDS line) fell below the 1 SDS line at day 3, the 2 SDS line at day 4, and the 2.5 SDS line at day 5 in the chart of the healthy breast-fed newborns. Permanent residual symptoms were reported in 11 cases between days 6 and 14, with weight losses ranging from 14.0 percent to 32.3 percent. Mortality was reported in ﬁve cases between days 5 and 15 with weight loss between 16.9 percent and 31.0 percent. In
(the graph is also available at: https://www.tno.nl/ downloads/GraphBreastfedInfants.pdf) (33). The lowest weight (i.e., the nadir) occurred 2 days after birth. Median weight loss was 6.2 percent at day 2, 5.1 percent at day 3, and decreased with approximately 1 percent of birthweight each following day. In total, 4.1 percent of the newborns lost 10 percent or more of their birthweight at day 2, and 0.9 percent at day 4. Only 0.6 percent still had a weight loss of 10 percent or more at day 7. Approximately 50 percent of the healthy breast-fed newborns had regained weight to equal their birthweight by 7 days; 88 percent by 14 days, and 95 percent by 19 days of age. Table 2 shows the association between RWC SDS and general neonatal and maternal characteristics in the healthy breast-fed sample. The results of the analyses with and without adjusting for all characteristics were almost similar. All models showed that more weight loss was observed in newborns with a high maternal age, ﬁrst parity, cesarean section, females, and a high birthweight (R2 = 8.9% by the ﬁxed factors without neonatal age and neonatal age2). Figure 3 shows the RWC of the cases with hypernatremic dehydration, together with the SDS lines of their chart. Each point represents one case. The literature cases (white points) had a lower RWC compared with the cases from the Turkish study, which is mainly caused by a late age of detection or admission. Median weight loss was 11 percent 2 days after birth and increased with 2 percent of birthweight each following day. At day 2, 65 percent of the cases with (or who will develop) hypernatremic dehydration lost 10 percent or more of their birthweight. At day 4, 87 percent of the cases weighed at least 10 percent less than their
3 4 5 7 9 Neonatal age (in days)
Fig. 2. Standard deviation score lines of the reference chart for relative weight change in healthy breast-fed newborns (logarithmic-scale axes). The dots presents the weight change of the newborns.
Table 1. General Neonatal and Maternal Characteristics of the Samples Available for Analyses
Healthy Study 1* Number of newborns Number of measurements Neonatal age range in days (min, max) Maternal age in years (mean, SD) Gestational age in weeks (mean, SD) First parity (%) Vaginal delivery (%) Females (%) Birthweight in g (mean, SD)
Hypernatremic dehydration Study 3†
183 565 0–8
1,470 2,839 0–8
706 1,071 0–60
169 169 0–22
102 102 0–20
36¶ 54 3,555 (390.9)
45 92 50 3,520 (448.6)
41 96 52 3,555 (449.2)
75 76 52 3,351 (453.4)
*Data retrospectively obtained from Dutch primary care midwife practices; †Data retrospectively obtained from Dutch child health clinics, ‡Data retrospectively obtained from the neonatal intensive care unit at the Turkey Ankara Children’s Hematology and Oncology Research Hospital; § cases obtained by a literature search. ¶N = 84; #N = 78.
BIRTH 41:2 June 2014
160 165 170 175 180 Serum sodium concentration (mEq/L)
This study showed that already at an early age weight loss differs between healthy newborns and those with hypernatremic dehydration, and this difference increases with age. Many cases with (or who will develop) hypernatremic dehydration (84%; +1 SDS) already fell below the 2 SDS line at day 4 and the 2.5 SDS line at day 5 in the chart of the healthy breast-fed newborns. Although it is easier to distinguish between weight loss of healthy and hypernatremic newborns at an older age, early detection and treatment leads to better outcomes compared with treatment at older ages. We also have to keep in mind that other
symptoms of hypernatremic dehydration (i.e., jaundice, decreased bowel movements (7,13), decreased urination, dry mucous membranes, sunken fontanel, poor skin turgor, lethargy (7), thirst fever (13,14), and irritability (9,15)) are helpful to early detect hypernatremic newborns with weight losses within the normal ranges of the growth chart.
Relative weight change (%) -35 -30 -25 -20 -15
all 16 cases with reported mortality or permanent residual symptoms, weight losses were far below the 2.5 SDS in the chart for healthy breast-fed newborns. The dashed line in Fig. 3 shows the association between RWC and serum sodium concentration, neonatal age, and the interaction between RWC and neonatal age in the hypernatremic dehydration sample. The adjusted explained variance of the model was 55.1 percent. The dashed line can be interpreted as the increase in serum sodium concentration by day when median weight loss according to the chart of weight loss of the cases occurs. Mean serum sodium concentration increased by 7.6 mEq/L in the ﬁrst week and 27 mEq/L after the ﬁrst week when weight loss increased by 10 percent. Mean serum sodium concentration of the cases was 155 mEq/L at 4 days of age, 160 mEq/L at 7 days of age, and 170 mEq/L at 12 days of age when median weight loss occurred.
5 6 7 8 10 Neonatal age (in days)
Fig. 3. Standard deviation score lines of the reference chart for relative weight change (left y-axis) of newborns with breastfeeding-associated hypernatremic dehydration (logarithmic-scale x-axis). The dashed line shows the increase in serum sodium concentration (right y-axis) by day when median weight loss occurs. The dots present the weight loss of the cases.
Table 2. Association between Standardized Relative Weight Change and General Neonatal and Maternal Characteristics in Healthy Breast-fed Newborns: Results of Multilevel Linear Regression Analyses (N VARIES BETWEEN 2,175 AND 2,359 NEWBORNS)
Characteristic Maternal age (per 10 yr) Parity First Second or more Delivery Vaginal Cesarean section Gender Male Female Birthweight (per 1 kg)
B (95% CI)†
Adj. B (95% CI)‡
0.18 ( 0.25,
0.17 ( 0.26,
0.08 ( 0.14, 1
0.22 ( 0.28, 1
1 0.25 ( 0.42, 1 0.14 ( 0.18, 0.54 ( 0.59,
1 0.20 ( 0.35,
1 0.22 ( 0.28, 0.62 ( 0.66,
† Multivariate model, adjusted for neonatal age and neonatal age2; ‡Multivariate model, adjusted for neonatal age, neonatal age2, maternal age, parity, mode of delivery, gender and birthweight (R2 by the ﬁxed factors without neonatal age and neonatal age2 is 8.9%). *p < 0.05, **p < 0.01, ***p < 0.001.
BIRTH 41:2 June 2014
158 Weight loss exceeding 10 percent or failure to regain birthweight by 10 or 14 days are conventional cutoffs (17,20,21). We found that a high proportion (12%) of the healthy newborns had not regained birthweight within 14 days. If the newborn is well, breastfeeding is adequate, not complicated, and the newborn has stopped losing weight, we suggest using a cutoff of 19 days to regain birthweight to reduce the number of unnecessary referrals to 5 percent. Furthermore, the 10 percent weight loss cutoff leads to unnecessary referrals, 4.1 percent at day 2, and 0.9 percent at day 4, and respectively, 65 percent and 87 percent correct referrals. This shows that the number of correct and unnecessary referrals for the 10 percent cutoff highly depends on age. Centiles of the chart can be used as cutoff values to ﬁx the level of unnecessary referrals at a certain age. However, more measurements usually imply more referrals (30). Further research is needed to evaluate the optimal cutoff levels of the chart in combination with the frequency and timing of the weighing moments to provide an evidence-based weighing policy for the early detection of hypernatremic dehydration. More weight loss was found in healthy newborns with a high maternal age, ﬁrst parity, cesarean section, females, and a high birthweight. Although all these characteristics achieved statistical signiﬁcance, birthweight was the only factor that showed a clinically relevant association. Other studies also found that cesarean section was associated with more weight loss compared with vaginal delivery (21,33). First parity and high birthweight were also shown to be risk factors for more weight loss (21). Other reported risk factors that we did not study were a hot season, jaundice not requiring phototherapy (21), low parental education (33), and epidural pain relief (34). Moreover, in this study, differences were found between the healthy and the hypernatremic dehydration study from Turkey. Maternal age, birthweight, the proportion with ﬁrst parity and vaginal delivery were higher in the healthy newborns compared with the cases with hypernatremic dehydration. As the cases were obtained from Turkey and the healthy sample from The Netherlands, the differences can be attributed to cultural and disease-speciﬁc differences. Other studies showed that cesarean section (12), ﬁrst parity, delay of initiation of ﬁrst breastfeeding, failure of the newborn to demand, use of the nipple shield, and infrequent feeding are risk factors for hypernatremic dehydration (7,13). In this study, we found that every additional kilogram of birthweight is associated with an extra weight loss of 0.62 SDS, whereas high birthweight was not reported as a risk factor for hypernatremic dehydration. Therefore, we suggest being careful in interpreting weight loss on the chart in newborns with very high birthweights (e.g. > 2 SDS).
One of the strengths of this study is that we used large sample sizes of healthy breast-fed newborns, which made it possible to develop a chart with high precisions of the centiles. The other strength is that we were able to convert weight loss into SDS and relate them to several risk factors. The effect of the risk factors was, therefore, better isolated from the effect of age on weight loss. A limitation is that the number of cases to estimate the upper and lower centiles of the chart was relatively small, but still large given the low prevalence and the sample sizes used in other studies. Another limitation is that worldwide scientists and clinicians use different deﬁnitions of day of birth. Examples that are used for day of birth are day one or zero, or only zero when the child was born before midday or before 7 P.M. This complicates a fair comparison between studies focusing on the ﬁrst week of life. We suggest reporting the number of hours after birth or providing the deﬁnition of day of birth in the methods section. In conclusion, the charts that we constructed give insight into a normal and an abnormal weight loss. The charts for weight loss are useful for two reasons: ﬁrst, to detect early or prevent breastfeeding-associated hypernatremic dehydration; second, to prevent unnecessary supplementing with formula with its risk of diminishing milk supply.
Acknowledgments We are grateful to Mieke Huisenga, Wendelien Roepke, Marleen Snabilie, Ilona van Putten, Jacqueline Breuning-Boers, and the midwife practices for obtaining data. We thank Jack Bennebroek Gravenhorst for his suggestions for improvement.
References 1. WHO/UNICEF. Global Strategy for Infant and Young Child Feeding. Geneva: World Health Organization, 2003. http://www. who.int/child_adolescent_health/documents/9241562218/en/index. htmlhttp://webitpreview.who.int/entity/nutrition/publications/gs_ infant_feeding_text_eng.pdf. 2. Britton C, McCormick FM, Renfrew MJ, et al. Support for breastfeeding mothers. Cochrane Database Syst Rev 2007;(1): CD001141. 3. Baby-Friendly Hospital Initiative. Revised, Updated and Expanded for Integrated Care. Geneva: World Health Organization, 2009. 4. Laing I, Wong C. Hypernatraemia in the ﬁrst few days: Is the incidence rising? Arch Dis Child Fetal Neonatal Ed 2002;87(3): F158–F162. 5. Neifert MR. Prevention of breastfeeding tragedies. Pediatr Clin North Am 2001;48:273–297. 6. Giugliani ER. Common problems during lactation and their management. J Pediatr (Rio J) 2004;80(5 Suppl):S147–S154.
BIRTH 41:2 June 2014
7. Livingstone VH, Willis CE, Abdel-Wareth LO, et al. Neonatal hypernatremic dehydration associated with breast-feeding malnutrition: a retrospective survey. CMAJ 2000;162:647–652. 8. Moritz ML, Manole MD, Bogen DL, Ayus JC. Breastfeedingassociated hypernatremia: Are we missing the diagnosis? Pediatrics 2005;116:343–347. 9. Bilgin LK, Akcay F, Altinkaynak K, Altindag H. Hypernatremia in breast-fed newborns: A review of 149 cases. J Trop Pediatr 2012;58:332–334. 10. Unal S, Arhan E, Kara N, et al. Breast-feeding-associated hypernatremia: Retrospective analysis of 169 term newborns. Pediatr Int 2008;50:29–34. 11. Kaplan JA, Siegler RW, Schmunk GA. Fatal hypernatraemic dehydration in exclusively breast-fed newborn infants due to maternal lactation failure. Am J Forensic Med Pathol 1998;19:19–22. 12. Konetzny G, Bucher HU, Arlettaz R. Prevention of hypernatraemic dehydration in breast-fed newborn infants by daily weighing. Eur J Pediatr 2009;168:815–818. 13. Boskabadi H, Maamouri G, Ebrahimi M, et al. Neonatal hypernatremia and dehydration in infants receiving inadequate breastfeeding. Asia Pac J Clin Nutr 2010;19:301–307. 14. Ng PC, Chan HB, Fok TF, et al. Early onset of hypernatraemic dehydration and fever in exclusively breast-fed infants. J Paediatr Child Health 1999;35:585–587. 15. Oddie SJ, Craven V, Deakin K, et al. Severe neonatal hypernatraemia: A population-based study. Arch Dis Child Fetal Neonatal Ed 2013;98:F384–F387. 16. van Dommelen P, van Wouwe JP, Breuning-Boers JM, et al. Reference chart for relative weight change to detect hypernatraemic dehydration. Arch Dis Child 2007;92:490–494. 17. Macdonald PD, Ross SR, Grant L, Young D. Neonatal weight loss in breast and formula fed infants. Arch Dis Child Fetal Neonatal Ed 2003;88:F472–F476. 18. Crossland DS, Richmond S, Hudson M, et al. Weight change in the term baby in the ﬁrst 2 weeks of life. Acta Paediatr 2008;97:425–429. 19. Noel-Weiss J, Courant G, Woodend AK. Physiological weight loss in the breastfed neonate: A systematic review. Open Med 2008;2:e99–e110. 20. Pelleboer RA, Bontemps ST, Verkerk PH, et al. A nationwide study on hospital admissions due to dehydration in exclusively breast-fed infants in the Netherlands: Its incidence, clinical characteristics, treatment and outcome. Acta Paediatr 2009;98: 807–811. 21. Davanzo R, Cannioto Z, Ronfani L, et al. Breastfeeding and neonatal weight loss in healthy term infants. J Hum Lact 2013;29:45– 53.
159 22. van Dommelen P, Verkerk PH, van Buuren S, van Wouwe JP. Weight loss in breast-fed neonates supplemented with formula. 40th Annual Scientiﬁc Meeting of the European Society for Clinical Investigation, Prague, Czech Republic, March, 2006, 15–18. 23. Chantry CJ, Dewey KG, Peerson JM, et al. In-Hospital Formula Use Increases Early Breastfeeding Cessation Among First-Time Mothers Intending to Exclusively Breastfeed. J Pediatr 2014 Feb 1. [epub ahead of print]. 24. van Putten IKF, van Dommelen P, van Wouwe JP, de Wolf BS. Voorspelt de mictiefrequentie het gewichtsverlies van borstgevoede kinderen. Tijdschrift voor verloskundigen 2007;6:25– 29. 25. Herngreen WP, van Buuren S, van Wieringen JC, et al. Growth in length and weight from birth to 2 years of a representative sample of Netherlands children (born in 1988–89) related to socioeconomic status and other background characteristics. Ann Hum Biol 1994;21:449–463. 26. Breuning-Boers JM, van Dommelen P, van Wouwe JP, Verkerk PH. Gewichtsverlies, serumnatriumconcentratie en restverschijnselen bij pati€entjes met hypertone dehydratie door onvoldoende borstvoeding. Ned Tijdschr Geneeskd 2006;150:904–908. 27. Stasinopoulos DM, Rigby RA. Generalized additive models for location scale and shape (GAMLSS) in R. J Stat Softw 2007; 23:1–46. 28. van Buuren S, Fredriks M. Worm plot: A simple diagnostic device for modelling growth reference curves. Stat Med 2001; 20:1259–1277. 29. Cole TJ, Green PJ. Smoothing reference centile curves: The LMS method and penalized likelihood. Stat Med 1992;11:1305– 1319. 30. van Dommelen P, van Buuren S. Methods to obtain referral criteria in growth monitoring. Stat Methods Med Res 2013;Feb 1. [Epub ahead of print] DOI: 10.1177/0962280212473301. 31. Grote FK, van Dommelen P, Oostdijk W, et al. Developing evidence-based guidelines for referral for short stature. Arch Dis Child 2008;93:212–217. 32. TNO Innovation for Life. Weight loss in breast-fed babies [graph of relative weight change]. Accessed January 9, 2014. Available at: https://www.tno.nl/downloads/GraphBreastfedInfants.pdf. 33. Manganaro R, Mamı C, Marrone T, et al. Incidence of dehydration and hypernatremia in exclusively breast-fed infants. J Pediatr 2001;139:673–675. 34. Martens PJ, Romphf L. Factors associated with newborn in-hospital weight loss: Comparisons by feeding method, demographics and birthing procedures. J Hum Lact 2007;23:233–241.