Early Human Development, 20 (1989) 67-16 Elsevier Scientific Publishers Ireland Ltd.
Effect of early human malnutrition and sleep organization Patricia Peirano,
Igino Fagioli*, Brish Bhanu Singh and Pier0 Salzarulo
INSERM U3, H6pital de la Salp.&ri&e, Paris (France) Accepted for publication 13 March 1989
Summary In order to evaluate the effect of early human malnutrition (EHM) on waking and sleep organization, a group of 12 infants malnourished (M) from birth were investigated in their first year of life through 24-h polygraphic recordings. The same infants were recorded again after nutritional rehabilitation (R). A group of 12 agepaired normal infants served as controls (C). Both young (< 4 months) and older (24 months) M infants showed a distribution of classes of waking duration different from that of C while the trend with age of the longest sustained waking (LSW) is similar in M and in C infants. The durations distribution of sleep episodes does not differ between M and C infants, while the longest sustained sleep (LSS) is significantly shorter in M than in C infants. The LSW is less clearly located in the day-time and the LSS is less clearly located in the night-time in M infants as compared to C infants. The quiet sleep-paradoxical sleep cycles are significantly shorter in older M infants because of the reduction of quiet sleep. Nutritional rehabilitation tends to improve waking and sleep organization. early malnutrition;
waking; sleep; infant; nutritional rehabilitation.
Introduction Early human malnutrition (EHM) is known to affect several aspects of brain development both anatomical and functional [5,6,10,11,16,24,25]. It has already been Correspondence to: P. Peirano, INSEBM U3, Hbpital de la Salpttriere, 47, Bd. de l’H&pital,75651 Paris Ctdex 13, France. *Present address: Laboratory of Psychology, University of Trento, Italy. 0378-3782/89/$03.50 0 1989 Elsevier Scientific Publishers Ireland Ltd. Published and Printed in Ireland
shown  that infants malnourished (M) from birth have diminished amount of quiet sleep on the 24-h period and shorter phases plus an increased amount of ambiguous sleep; these data indicate that M infants have difficulties in coordinating several physiological functions in a stable state. The aim of this work is to investigate whether EHM affects first the consolidation of waking and sleep [2,14,26] and their distribution in the 24-h, and secondly the basic structure of sleep: the cycle made by a sequence of quiet sleep and paradoxical sleep ]4,91, which has been proved to be useful in infant research [3,4,8,12,13,15,23]. The effect of nutritional rehabilitation will be evaluated on the variables affected by malnutrition. Method
Twelve M infants (8 females, 4 males) full-term, between 6 weeks and 14 months of age, of normal birth weight and free from neurological disturbances, were recorded once. Malnutrition began soon after birth and was due to severe diarrhoea in 9 cases and following a surgical resection of the small bowel in 3 cases. The infants weight at the time of recording was - 42.8 f 5.0% of the expected weight for chronological age; the values for length, not being below - 15% of the expected length for chronological age, were less deviant than the values for weight. The weight for length was - 26.1 + 9.0% of normal values. For grading the nutritional state we used the French classification established by Sempt et al. . No water or electrolyte disorders were present at the moment of recording. The same 12 infants were studied again after nutritional rehabilitation (R) performed in a nutritional intensive care unit (Pr. C. Ricour, Hopital des Enfants Malades, Paris) by parenteral nutrition through intracaval catheter [ 171and/or constant rate enteral nutrition U81. Nutritional rehabilitation led to clinical improvement and weight increase; the weight was - 37.0 f 6.3% of the expected weight for chronological age (the difference between malnutrition and rehabilitation is statistically significant: t = 4.8, P< 0.001, Student’s t-test for paired data). All M infants during the rehabilitation period received antibiotics; none received other kinds of drugs, in particular no psychotropic drugs. Twelve normal full-term infants, between 1 and 12 months, age-matched with the M, served as controls (C). Each polygraphic recording lasted for a period of 24-h. Techniques and data analysis, which have already been used by our group [7,8,19-211 were as follows: electroencephalogram (EEG) (l-3 channels), electro-oculogram, electromyogram, respiratory rhythm, electrocardiogram. The behaviour of the child was observed and noted on the EEG paper. The records were analyzed in 20-s epochs. Four states were defined by combining behavioural and electrophysiological data; (i) Quiet sleep (QS): eyes closed, no eye movements, diminished body movements, EEG with slow waves (and spindles) or “trace alternant”, regular respiration; (ii) Paradoxical sleep (PS): eyes closed (or alternatively half open and closed), eye movements and in addition body or limb movements, low voltage EEG, muscular atonia, irregular respiration; (iii) Ambiguous sleep (Amb.S) includes the characteristics of both QS and PS;
(iv) Waking: eyes open, eye movements and in addition body movements, irregular respiration. The minimum length of each state was 2 min, shorter changes being included in the preceding state. Longest sustained waking (LSW) was defined as the longest waking period without being interrupted by a sleep episode lasting for more than 15 min. Longest sustained sleep (LSS) was defined as the longest sleep period without being interrupted by a waking episode lasting for more than 15 min . A sleep cycle (a phase of QS plus a phase of PS) was defined as the period starting from the beginning of the QS to that of the next QS (or the spontaneous awakening); cycles including Amb.S at the QS-PS and/or PS-QS transition were retained; whereas those containing AmbS between two phases of QS or PS were excluded [4,9]; cycles being finished by provoked awakenings were excluded as well. Results were statistically tested using: (a) Chi-square test for the durations distribution of waking and sleep episodes; (b) Student’s t-test: (i) for independent samples for comparisons between M and C and between R and C infants; and (ii) for related samples for comparisons between M and R infants; (c) 2 way-ANOVAs: (i) for two independent factors (nutritional status and age) for comparisons between M and C and between R and C infants; and (ii) for one independent (age) and one related factor (nutritional status) for comparisons between M and R infants. Results
Waking (W) consolidation We separated W episodes in four duration classes according to Wolff’s criteria  (Fig. 1). The durations distribution is different between M and C infants both in young (< 4 months; x2 = 10.222, df = 3, P < 0.02) and older (24 months; x2 = 8.800, df = 3, P< 0.05) subgroups: long duration episodes are scarce in M infants, while in the C infants, mainly in the older subgroup, there is a consistent frequency of long duration episodes (> 60 min). The LSW is significantly longer in older infants as compared to young infants (F (1, 20) = 5.2216, P < 0.05) irrespective of the nutritional status (main effect F (1, 20) = 2.4047, n.s.; interaction F (1, 20) = 0.04133, n.s.) as it is shown by a 2-way ANOVA for independent factors. The LSW is clearly located in the day-time in C infants, while it is less evident in M infants (Fig. 2). Sleep consolidation Sleep episodes were separated according to the same four duration classes as were W episodes (Fig. 3). The durations distribution does not differ between C and M infants in both young and older subgroups (x2 = 0.444, df = 3, n.s. and x2 = 5.714, df = 3, n.s. respectively). However, the short sleep episodes (<20 min) are still prevalent in the older M group. The LSS is significantly shorter in M than in C infants (F (1, 20) = 4.6965, P < 0.05) without differences between young and older subgroups (F (1, 20) = 0.002784, n.s.); the difference between M and C infants is more consistent in the older subgroup, even if the interaction between the two factors (nutrition and age) fails to reach statistical significance (F (1, 20) = 1.43866,
% 80 50 40
C 30 20 10 I
4 M 3 2 10
R 30 20 10 !
ILL Ill4L -
Fig. 1. Distribution of uninterrupted (C = controls; M = malnourished; 41-60 min; D = 361 min).
periods of waking in the young (3’) and older (0) subgroups. R = rehabilitated; A = 2-20 min; B = 21-40 min; C =
n.s.) (2-way ANOVA for independent factors). LSS in C infants is always located in the night-time, while it is less evident in M infants. QS-PS cycles duration The duration of QS-PS C infants 0, = 2.216x + infants 0, = - 1.052.x + dent factors shows that
cycles within LSS shows a tendency to increase with age in 38.649; r = 0.5701, n.s.) while it tends to decrease in M 50.087; r = 0.3937, n.s.). A 2-way ANOVA for indepenQS-PS cycles (Table I) are shorter in older M infants
Fig. 2.24-h distribution of longest sustained waking (A) and of longest sustained sleep (0) in control (C) and malnourished (M) infants. In each group, infants are placed according to the age from the youngest (up) to the oldest (down).
(malnutrition, main effect F (1, 20) = 1.079, n.s.; age, main effect F (1, 20) = 0.828, n.s.; interaction F (1, 20) = 8.054, P < 0.02). The shorter duration of cycles in older M infants is accounted for by a significant diminution of QS (t = 2.062, df = 10, P < 0.05; Student’s t-test for unpaired data); the duration of both PS and Amb.S did not differ between M and C groups (t = - 0.163, df = 10, n.s. and t = 1.242, df = 10, n.s. respectively; Student’s t-test for unpaired data; Table II). Effect of nutritional rehabilitation
We looked at the effects of nutritional rehabilitation on the variables which were affected by malnutrition. The durations distribution of W episodes in R infants is quite different from that of M (x2 = 10.108, df = 3, P < 0.01) in the young subgroup, while in the older subgroup it does not differ from that of M (x2 = 5.974, df
C 30. 20IO-
40 M 30 20 1 I
Fig. 3. Distribution of uninterrupted periods of sleep in the young (Y) and older (0) subgroups. For legend see Fig. 1.
= 3, n.s.) and of C infants (x2 = 7.072, df = 3, n.s.). No significant difference was found for LSS duration after nutritional rehabilitation (a 2-way ANOVA for one independent factor (age) and one related factor (nutritional status) F (1, 10) = 0.852, n.s.) either for young or older subgroups (main effect F(l, 10) = 0.303, n.s.; interaction F (1, 10) = 1.708, n.s.). The 24-h distribution of both LSW and LSS was TABLE I Cycles duration (min) within LSS.
Controls Malnourished Rehabilitated
40.9 + 11.8 48.2 + 11.5 51.6 f 13.8
55.3 + 9.4 41.7 f 5.9 50.5 f 12.7
73 TABLE II Duration of sleep states (rnin) within the cycles in the LSS in older infants.
Controls Malnourished Rehabilitated
slightly infants The infants
35.6 f 10.4 25.3 f 6.4 30.3 f 8.3
12.4 + 3.6 12.8 f 4.1 11.5 f 3.8
7.0 + 3.5 4.6 f 1.8 1.7 f 3.2
modified after nutritional rehabilitation and tends to approach that of C (Figs. 2 and 4). increase of QS-PS cycles duration after nutritional rehabilitation in older fails to reach statistical significance (F (1, 10) = 3.410, n.s.) without differ-
Fig. 4. 24-h distribution of longest sustained waking (A) and longest sustained sleep (0) in controls (C) and rehabilitated (R) infants. In each group, infants are placed according to the age from the youngest (up) to the oldest (down).
ences between young and older subgroups (F(1, 10) = 0.618, n.s.; interaction F(l, 10) = 0.834, n.s., Table I). After nutritional rehabilitation, QS within cycles duration in older infants is not statistically different between M and R (t = - 1.150, df = 5, n.s., Student’s t-test for paired data) and C and R infants (t = 0.981, df = 10, n.s., Student’s t-test for unpaired data, Table II). Discussion
Waking consolidation is consistently disturbed by malnutrition, as it is shown by the percentage distribution of classes of durations (rare episodes of waking longer than 60 min). Data coming from the analysis of the LSW, which show no differences between M and C infants, are apparently contradictory with the previous ones. In fact, this discrepancy could be explained taking into account our definition of LSW which admits the possibility that sleep episodes shorter than 15 min are included, while sleep episodes longer than 2 min lead to an interruption of waking episodes in the classes distribution. Also it is important to stress the fact that the effect of EHM on waking is observed in both young and older subgroups. We may put forward the hypothesis that in EHM energy diminution plus a modification of the social environment puts the infant in difficult condition for initiating goal directed actions, which is the prerequisite, according to Wolff , to prolong waking by the end of 2 months on. Sleep consolidation (see LSS) is also affected by EHM; moreover short sleep episodes are still numerous in the older infants, a finding which recalls that reported by Chavez and Martinez [l] through behavioural observations. Disturbances of waking and sleep consolidation in M infants are often accompanied by a poor location in the 24-h period of the longest episode (LSW and LSS): on the whole these results show an important and global disturbance of the development of the sleep-waking rhythm. The ultradian organization of sleep is disturbed in the older M infants, as it is shown by a shorter duration of the QS-PS cycles. The shortening of the cycles is entirely due to the shortening of the QS. This result together with the shortening of the QS phases on the whole sleep time previously observed , underscores the vulnerability of QS by EHM, whatever the sleep structure. The diminution of QS (together with the disturbance of sleep and waking consolidation) could be considered as an expression of the disruption of homeostatic processes in the M infants (i.e. the ability to maintain a stable biological condition). It is also important to emphasize possible consequences for growth of the shortening of QS-PS cycles. Since relationships between sleep cycles and anabolic processes have been shown , M infants thus may have an additional factor which limits their growth. After nutritional rehabilitation the distribution of the durations of waking episodes in the young subgroup is similar to that of C infants, while that of the older subgroup is intermediate between C and M infants, suggesting that recovery takes more time in long malnutrition periods. Nutritional rehabilitation, which has been shown to be fully effective on the
recovery of the quantity of QS over a 24-h period and on the mean duration of the phases , has a relative effect on the sleep organization recovery. LSS duration, QS-PS cycles duration and QS within cycles duration values are intermediate between C and M values. We suggest that the recovery of QS phases is the first step towards a recovery of the whole sleep organization. Acknowledgements We gratefully acknowledge Marie-The&se Rigoard for her skillful cooperation and Dr. W.R. Webster for revising the English. P. Peirano received grants from Foundation Singer-Polignac and Foundation Mustela. References 1
2 3 4 5 6 I 8 9 10 11 12 13
14 15 16 17 18
Chavez, A. and Martinez, C. (1984): Behavioral measurements of activity in children and their relation to food intake in a poor community. In: Energy Intake and Activity, pp. 303-321. Editors: E. Pollitt and P. Amante. Alan R. Liss, New York. Coons, S. and Guilleminault, C. (1984): Development of consolidated sleep and wakeful periods in relation to the day-night cycle in infancy. Dev. Med. Child Neurol., 26, 169-176. Curzi-Dascalova, L., Peirano, P. and Morel-Kahn, F. (1988): Development of sleep states in normal premature and full-term newborns. Dev. Psychobiol., 21,431-444. Dittrichova, J. (1966): Development of sleep in infancy. J. Appl. Physiol., 21, 1243-1246. Dobbing, J. and Sands, J. (1973): Quantitative growth and development of the human brain. Arch. Dis. Child., 48,757-767. Engel, R. (1956): Abnormal brain wave patterns in kwashiorkor. Electroenceph. Clin. Neurophysiol., 8,489-500. Fagioli, I., Ricour, C., Salomon, F. and Salzarulo, P. (1981): Weight changes and sleep organization in infants. Early Hum. Dev., 5, 395-399. Fagioli, I. and Salzarulo, P. (1982): Sleep states development in the first year of life assessed through 24-h recordings. Early Hum. Dev., 6.215-228. Fagioli, I. and Salzarulo, P. (1982): Organisation temporelle dans les 24 heures des cycles de sommeil chez le nourrisson. Rev. Electroencephalograph. Neurophysiol. Clin., 12,344-348. Fagioli, I., Salzarulo, P., Salomon, F. and Ricour, C. (1983): Sinus pauses in early malnutrition during waking and sleeping. Neuropediatrics, 14,4346. Fagioli, I., Pedroni, E. and Salzarulo, P. (1985): QT interval in early human malnutrition and after nutritional rehabilitation. Pediatr. Cardiol., 6,52-53. Kraemer, H.C., Hole, W.T. and Anders, T.F. (1984): The detection of behavioral state cycles and classification of temporal structure in behavioral states. Sleep, 7,3-17. Monod, N. and Pajot, N. (1965): Le sommeil du nouveau-n& et du premature. Analyse des etudes polygraphiques (mouvements oculaires, respiration et EEG) chez le nouveau-& B terme. Biol. Neonate, 8,281-307. Parmelee, A.H., Wenner, W.H. and Schulz, H.P. (1964): Infants sleep patterns: from birth to 16 weeks of age. J. Pediatr., 65,576-582. Prechtl, H.F.R., Weinmann, H. and Akiyama, Y. (1969): Organization of physiological parameters in normal and neurologically abnormal infants. Neuropadiatrie, 1, 101-129. Prescott, J.W., Read, M.S. and Coursin, D.B. (Editors) (1978): Brain Function and Malnutrition: Neuropsychological Methods of Assessment, 449 pp. John Wiley, New York. Ricour, C. and Nihoul-Fekete, C. (1973): Nutrition parent&ale prolong&e chez l’enfant. Arch. Fr. Ptdiatr., 30,469-490. Ricour, C., Duhamel, J.F. and Nihoul-Fekete, C. (1977): Nutrition enterale a debit constant chez l’enfant. Arch. Fr. Pediatr., 34, 154-170.
76 19 20
22 23 24
Salzarulo, P., Fagioli, I., Salomon, F., Duhamel, J.F. and Ricour, C. (1979): Alimentation continue et rythme veille-sommeil chez I’enfant. Arch. Fr. Pediatr. (suppl.), 36,26-32. Salzarulo, P., Fagioli, I., Salomon, F., Ricour, C., Raimbault, C., Ambrosi, S., Cicchi, O., Duhamel, J.F. and Rigoard, M.T. (1980): Sleep patterns in infants under continuous feeding from birth. Electroenceph. Clin. Neurophysiol., 49,330-336. Salzarulo, P., Fagioli, I., Salomon, F. and Ricour, F. (1982): Developmental trend of quiet sleep is altered by early human malnutrition and recovered by nutritional rehabilitation. Early Hum. Dev., 7,257-264. Sempe, M., Pedron, G. and Roy-Pernot, M.P. (1979): Auxologie, Mtthode et Sequences, p. 205. Thtraplix, Paris. Stern, E., Parmelee, A., Akiyama, Y., Schulz, M. and Wenner, W. (1969): Sleep cycle characteristics in infants. Pediatrics, 43,65--70. Valenzuela, R.H., Hernandez-Peniche, J. and Macias, R. (1959): Clinical, electroencephalographic and psychological aspects of recuperation of the undernourished child. Gac. Med. Mex., 89, 651665. Winick, M. and Rosso, P. (1975): Malnutrition and central nervous system development. In: Brain Function and Malnutrition: Neuropsychological Methods of Assessment, pp. 41-51. Editors: J.W. Prescott, M.S. Read and D.B. Coursin. John Wiley, New York. Wolff, P.H. (1984): Discontinuous changes in human wakefulness around the end of the second month of life: a developmental perspective. In: Continuity of Neural Functions from Prenatal to Postnatal Life, pp. 144-158. Editor: H.F.R. Prechtl. Blackwell, Oxford.