Objective. To assess the effectiveness and safety of Kangaroo Mother Care (KMC) for infants of low birth weight.
Methods. An open, randomized, controlled trial of a Colombian social security referral hospital was conducted. A total of 1084 consecutive infants who were born at ≤2000 g were followed, and 746 newborns were randomized when eligible for minimal care, with 382 to KMC and 364 to “traditional” care. Information on vital status was available for 693 infants (93%) at 12 months of corrected age. KMC consisted of skin-to-skin contact on the mother's chest 24 hours/day, nearly exclusive breastfeeding, and early discharge, with close ambulatory monitoring. Control infants remained in incubators until the usual discharge criteria were met. Both groups were followed at term and at 3, 6, 9, and 12 months of corrected age. The main outcomes measured were morbidity, mortality, growth, development, breastfeeding, hospital stay, and sequelae.
Results. Baseline variables were evenly distributed, except for weight at recruitment (KMC: 1678 g; control participants: 1713 g). The risk for death was lower among infants who were given KMC, although the difference was not significant (KMC: 11 [3.1%] of 339; control participants: 19 [5.5%] of 324; relative risk: 0.57; 95% confidence interval: 0.17–1.18). The growth index of head circumference was statistically significantly greater in the group given KMC, but the developmental indices of the 2 groups were similar. Infants who weighed ≤1500 g at birth and were given KMC spent less time in the hospital than those who were given standard care. The number of infections was similar in the 2 groups, but the severity was less among infants who received KMC. More of these infants were breastfed until 3 months of corrected age.
Conclusion. These results support earlier findings of the beneficial effects of KMC on mortality and growth. Use of this technique would humanize the practice of neonatology, promote breastfeeding, and shorten the neonatal hospital stay without compromising survival, growth, or development.
Low birth weight is a major public health problem worldwide, but the burden represented by this condition is considerably higher in developing countries. In Colombia, it is estimated that 11% of all live newborn infants weigh ≤2500 g at birth (M. Ronderos, Ministry of Health of Colombia, personal communication, March 1998), and at the main social security facility in the country (where this study was conducted) 25% of all liveborn infants had a birth weight of ≤2500 g and 10% had a birth weight of ≤2000 g. Proper care of these infants requires access to expensive, sophisticated techniques and to highly qualified health care professionals, although it has been estimated that 95% of all low-birth-weight infants are delivered in low-income countries (N. Bergman, personal communication, December 1998). All such infants should have access to humane, efficacious, and efficient care, despite the enormous existing inequalities.
Kangaroo Mother Care (KMC) is an alternative to standard hospital care for low-birth-weight infants. Early, continuous skin-to-skin contact between the mother and her infant is the cornerstone of the treatment, with breastfeeding and careful ambulatory clinical follow-up.1,,2 Three uses of KMC have been identified2:
Where appropriate neonatal care facilities do not exist, KMC is proposed as the alternative to incubators in health facilities.
Where all levels of neonatal care are readily accessible, early mother–infant skin-to-skin contact in health facilities can enhance the quality of mother–infant bonding and promote successful breastfeeding.
Where technical and human resources are of a high standard but are insufficient to cope with the demand, as in our institute in Colombia, KMC is provided mainly as outpatient care after a short in-hospital adaptation period, provided the infant has overcome major neonatal problems and is eligible for minimal care.
Between September 1993 and December 1996, we conducted a randomized clinical trial to compare KMC with “traditional” care. The methods and early (40–41 weeks of gestational age) outcomes have been described fully elsewhere,3,,4 and the results with regard to the quality of mother–infant bonding at 12 months of corrected age will be reported in the near future. The early results showed that at 40 weeks of postconceptional age, the mortality and growth indices and the total number of episodes of infection were similar, but the KMC infants had less severe infections. Furthermore, they had a shorter average hospital stay (1.1 days less), with an even larger difference for infants with lower birth weights. There was also a significant difference in early breastfeeding patterns. In addition, we found that mothers who provided skin-to-skin contact gained confidence in the care of their fragile infants, and the effect was stronger in stressful situations such as a prolonged hospital stay as a result of disease or prematurity.4 This report complements the early observations with data on longer-term outcomes with regard to health status, including somatic and neurologic development, during the first 12 months of corrected age.
The study population consisted of all liveborn infants who weighed ≤2000 g at the Clı́nica San Pedro Claver and were eligible for KMC and whose mother or a relative was able to understand and was willing to follow the general instructions for taking care of a premature infant and complying with a 1-year follow-up schedule. Infants were eligible for KMC regardless of their actual weight or gestational age, as soon as they had overcome major problems of adaptation to extrauterine life, any infection or concomitant condition had been properly treated, the infant was sucking and swallowing properly, and the daily weight changes were appropriate for gestational and chronological age. Major problems of adaptation to extrauterine life were overcome at a median postpartum age of 4 days (95% confidence interval [CI]: 3–5), and 6% of the infants were eligible before they were 1 week old. Infants who weighed ≤1200 g at birth were eligible at a median postnatal age of 33 days, whereas those who weighed 1200 to 1499 g, 1500 to 1799 g, and 1800 to 2000 g were eligible at median postnatal ages of 19 days, 7 days, and 1 day, respectively.
The exclusion criteria were referral to another institution, plans to leave Bogotá in the near future, lethal or major malformations, early major conditions arising from perinatal problems (eg, severe hypoxic ischemic encephalopathy, pulmonary hypertension), and parental or family refusal to comply with the follow-up program or, in those assigned to KMC, with the intervention. This design had the potential for introducing imbalance between study groups, but none of the mothers who were assigned to KMC refused to participate.
Between September 20, 1993, and September 20, 1994, 13 560 live infants were delivered at the Clı́nica San Pedro Claver. Of these, 3350 (24.7%) had a low birth weight (≤2500 g) and 1084 (8%) infants who weighed ≤2000 g were assessed and followed up to determine their eligibility. Before randomization, 307 infants (28%) were declared ineligible because they had died before evaluation (160 infants [15%]), had major malformations and dysmorphic syndromes (6 infants [0.005%]), were referred to other institutions (131 infants [12%]), and other reasons (10 infants [0.01%]). The remaining 777 (72%) were randomized to 1 of the 2 interventions; when 31 infants were subsequently withdrawn for conditions not detectable at the time of randomization that precluded admission (early detection of severe sequelae of neonatal conditions as severe encephalopathy or bronchopulmonary dysplasia [31 infants]), 746 remained in the study. When severe neurologic problems were detected or when serologic examinations indicated intrauterine infection (eg, congenital rubella, toxoplasmosis) during the first 2 weeks after randomization, the infant was excluded by 1 of the authors (J.G.R.), who was not involved in clinical care or in direct data collection and who was unaware of the treatment status of the patient, in accordance with recommended procedures for randomized, controlled trials with prolonged follow-up.5 The withdrawals affected 14 of the infants who were receiving KMC and 17 in the control group. Complete follow-up data were obtained for 679 (91%) infants at 40 to 41 weeks of postconceptional age and for 630 (84.5%) infants at 12 months of corrected age; the vital status of 63 other infants at 12 months was assessed by telephone. At 12 months of corrected age, therefore, vital status was known for 693 infants (93% of the sample).
According to our previous study,6 the death rate among eligible patients who were under traditional care at the Clı́nica San Pedro Claver was approximately 7%. To detect a twofold increase in the risk for death, a sample of 215 patients per group would be needed (<0.05, 2-tailed test, 80% power). If it is assumed that the death rate among control participants is as high as 10%, then at least 328 patients per group would be needed to detect a 2-fold increment in the risk for dying. As we expected that only 70% of the newborn infants who weighed ≤2000 g would be eligible, we planned to evaluate at least 940 infants who weighed ≤2000 g to identify enough eligible patients.
Recruitment and Allocation
An observer reviewed the delivery and admission records daily to identify all infants who were born at or referred to the Clı́nica San Pedro Claver with a birth weight or a weight at admission of ≤2000 g. Infants who were admitted to the neonatal care unit and those who were placed in the mother's hospital room were assigned a consecutive identification number. Eligibility for KMC was assessed by an attending physician for those who were admitted to the neonatal care unit and before discharge of the mother and infant for those who were not admitted to the unit. Infants who were identified as eligible were assigned randomly to KMC or the control intervention. Four stratified block randomization lists were prepared from a list of random numbers (permutations of the number 16) according to birth weight: ≤1200 g, 1201 to 1500 g, 1501 to 1800 g, and 1801 to 2000 g, and blocks of 2 infants assigned to KMC and 2 to control were prepared.
The parents of infants who were assigned to KMC were interviewed by one of the study nurses and were asked whether they wished to participate. If they agreed, then the infant was discharged and the mother and infant were transferred to the Kangaroo Ambulatory Clinic in another hospital (Clinica del Niño), where the parents began providing KMC and the adaptation of the mother and infant to the kangaroo position and feeding were monitored during the daytime. At night, mothers and infants went to their homes. Daily attendance to the Kangaroo Ambulatory Clinic was continued for as long as was necessary to ensure that the adaptation was appropriate. Infants who were assigned to the control treatment remained in the Clı́nica San Pedro Claver neonatal care unit until the treating neonatologist decided on discharge.
A modified version of the original Rey-Martı́nez KMC method1 was used, in which eligible infants who weighed <2000 g at birth are discharged and their mothers are used as “incubators” and as the main source of food and stimulation. The infants were kept in an upright position, in skin-to-skin contact, firmly attached to the mother's chest for 24 hours/day. Their temperature thus was maintained within the normal range by the mother's body heat. The infants were breastfed regularly, and premature formula supplements were used to guarantee adequate weight gain if necessary; they were examined daily until they gained at least 20 g/d. They remained in the kangaroo position until they no longer accepted it.
In the control (traditional) intervention, infants were kept in incubators until they could regulate their temperature and showed appropriate weight gain. The infants were discharged according to the practice at the Clı́nica San Pedro Claver, usually when they weighed approximately 1700 g. At the time of the study, the practice of the neonatal care unit was severely to restrict the parents' access to their infants.
Data Collection and Follow-Up
After discharge, the 2 groups had similar access to ambulatory care and follow-up. For minimizing contamination bias, follow-up visits to KMC infants took place in the mornings and to control infants in the afternoons. The ambulatory care included administration of metoclopramide7 and supplemental vitamins (A, D, E, and C) to all infants and prophylactic aminophylline to those who were <34 weeks' corrected age at entry to the study. Iron supplementation was started at 2 months of corrected age, and metoclopramide was stopped at 6 months.
The follow-up visits were scheduled at 1, 3, 6, 9, and 12 months of corrected age, when information was collected on health status, including the results of a neurologic examination and a neurologic screening test. Additional evaluations of psychomotor development and of the mother–infant relationship were conducted at term and at 6, 12, and 15 months of corrected age.
The primary outcomes were mortality, development, and growth (weight, height, and head circumference). Mortality was measured as the proportion of children whose vital status was known at 12 months of corrected age and who died during follow-up (cumulative mortality) and as the mortality rate (cases per person-time). Development was assessed on the Griffith scale of mental development, validated for Colombia. Growth was measured as changes in weight, length, and head circumference and expressed as both absolute values and as a proportion of expected values for corrected age, according to the National Center for Health Statistics standards.
The secondary outcomes were length of hospital stay for survivors; the overall incidence of infection, measured as all infectious episodes requiring antibiotic treatment, in the hospital or elsewhere; the incidence of severe nosocomial infections requiring systemic antibiotics or severe infections detected after discharge and requiring hospital admission; and the proportion of infants entirely or partially breastfed, at each follow-up time.
Baseline and follow-up data were collected prospectively by interviews with the mothers and review of clinical charts at the neonatal unit and at scheduled visits and were entered into a computerized database; manual and computerized data cleaning was then conducted. As the baseline data concerned variables and events that occurred before randomization, the observers who collected the information were unaware of the treatment status of the infant. The clinicians who assessed the infants at follow-up could not be blinded, however, because the KMC mothers were holding their infants in the kangaroo position.
Comparisons between study groups for discrete variables were performed with the χ2 or Fisher's exact test. Continuous variables were compared by means of Student's ttest or nonparametric tests, when appropriate. Adjustments for potential confounders and tests for interactions were conducted with analysis of variance (ANOVA) and/or multiple linear regression for continuous variables; robust regression ANOVA–analysis of covariance (ANCOVA) and multiple robust regression8 were used for continuous variables with nonnormal distributions and logistic regression for dichotomous outcome variables. As the analysis was performed from the “intention to treat” perspective (effectiveness as opposed to efficacy), all patients were analyzed according to the group to which they were allocated, regardless of compliance with treatment or contamination of the intervention.
Baseline Comparison of Treatment Groups
The 2 study groups did not differ with regard to sociodemographic, pregnancy, or labor characteristics. The high prevalences of pathologic conditions such as preeclampsia (38%) during pregnancy and the high rate of cesarean sections (68%) reflect that our study population was the high-risk segment of the deliveries conducted at the clinic. The distribution of these variables was similar, except that there were 22 more multiple deliveries in the KMC group.
All relevant characteristics at birth were distributed evenly (Table 1). Of the total eligible population, 64% were preterm infants who were adequate for gestational age, 22% were preterm infants who were small for gestational age, and only 14% were term infants who were small for gestational age. The prevalence of asphyxia at birth according to the Apgar score at 1 minute was similar in the 2 groups. The most frequent diagnoses at primary discharge were transient tachypnea (24%), jaundice (all causes, 23%), hyaline membrane disease (15%), neonatal sepsis (11%), necrotizing enterocolitis (5%), and suspected and/or confirmed nosocomial infection before eligibility (10%). These diagnoses were evenly distributed between the study groups.
At the time of eligibility, the groups showed a few slight imbalances, such as a lower average weight among KMC infants (although the distribution according to weight was similar) and a higher proportion of infants never admitted to the neonatal unit in the control group. The proportion of infants who required intensive care was similar in the 2 groups, as was the duration of mechanical ventilation.
Compliance with follow-up visits was 78% to 91% per visit, and 80% attended the last visit at 12 months of corrected age. Although the proportions of deceased, compliant, and noncompliant infants were similar at each scheduled visit in the 2 groups, multiple logistic regression showed that after control for weight at eligibility and educational level of the mother, KMC infants were 1.61 times more likely than control participants to attend scheduled visits at 3 months (P < .001), despite identical efforts to obtain good compliance. A comparison of the baseline characteristics of the infants who attended each scheduled visit showed that no additional imbalance in baseline risk was introduced by failure to attend (data not shown).
Cumulative mortality during follow-up to 12 months of corrected age was 4.3%, with 11 deaths (3.0%) in the KMC group and 19 (5.5%) in the control group (crude relative risk: 0.57; 95% CI: 0.17–1.18). Cumulative mortality at each follow-up visit was always lower in the KMC group, and the measurements at 3 and 9 months almost reached statistical significance (Table 2). Cumulative mortality at 12 months depended on 4 groups of factors: biological baseline conditions (ante- and perinatal and clinical events), socioeconomic and educational levels (availability of resources, social support, and health-seeking behavior), type of intervention, and random factors leading to lethal events.
The fourth group of factors was controlled by random allocation of infants to 1 of the 2 interventions. After control for imbalances in the distribution of the first 2 sets of factors (which might have behaved as effect modifiers) by multiple logistic regression, only 2 factors were statistically significantly associated with mortality: level of education of the father and weight at eligibility. After control for these factors, the adjusted odds ratio for mortality was 1.97 (95% CI: 0.91–4.24). Nevertheless, we found a protective interaction between having a father with a primary or lower education and the KMC intervention, with a lower risk for death than in the control group.
In a Kaplan-Meyer analysis, the survival function was lower in the control group, although the difference in cumulative survival was not statistically significant. A Cox proportional hazards multivariate regression analysis to control for potential confounders showed an adjusted death hazard ratio between the control and KMC groups of 1.95 (95% CI: 0.93–4.10).
General Morbidity: Length of Stay and Ambulatory Visits
We evaluated use of health resources in the ambulatory setting and the number and duration of hospital admissions during the first year of corrected age. The average number of ambulatory visits between eligibility and 12 months of corrected age was 17 for the KMC group and 12.8 for the control group (P < .0001), and the density of visits (number of visits per 100 days of child follow-up) was 4.5 for the KMC group and 3.4 for the control group (P < .001). The somewhat higher values for the KMC group reflect visits initiated by parents and all visits to ensure satisfactory daily weight gain, whereas the 2 visits per day to control infants, who remained in hospital, were not counted. The adjusted number of visits per 100 infant-days did not change after controlling for potential confounders such as gestational age and weight at eligibility, per-capita income, and gender (robust regression ANOVA–ANCOVA).
The number of readmissions after primary discharge was similar in the 2 groups, but the total length of hospital stay from eligibility until the end of the follow-up was shorter for the KMC infants and particularly for those who weighed <1500 g at birth (Table 3). We used a robust regression ANOVA to confirm that the interaction between KMC and weight at birth is significant, and the savings in hospital stay persist up to 12 months of corrected age.
The overall frequency of infections, expressed as the number of infectious episodes and incidence density, was similar during follow-up. Although the frequency of nosocomial infections was statistically significantly higher in the control group (6.8%; KMC: 3.4%; rate ratio: 2.01; 95% CI: 1.04–3.87), there was no difference in the cumulative frequency of infections grouped by severity: mild to moderate (ambulatory use of antibiotics), 1.02 episodes per KMC infant and 0.9 episodes per control infant; severe (requiring hospitalization) between eligibility and 12 months: 0.31 admissions per KMC infant and 0.33 admissions per control infant.
The incidence density of ambulatory and hospital treatment of infections was computed as the number of episodes in each study group by patient-time of follow-up. Length of hospital stay for an infection divided by the total time of follow-up for each patient was computed as a proxy for burden of illness. The density of infectious episodes was 0.25 ambulatory antibiotic courses per 100 KMC infant-days of follow-up and 0.21 per 100 in control infants (P = .053), and 0.09 hospital admissions for severe infection per 100 KMC infant-days and 0.24 for the control group (P = .14). The length of hospital stay for infection between eligibility and 12 months of corrected age standardized by time of follow-up was 1.1 days per 100 KMC infant-days and 2.1 days per 100 control infant-days (P = .12).
Given that there were moderate imbalances in some of the baseline characteristics at eligibility and we wanted to explore further the relationships among risk factors for severe infection and the interventions, we used a series of multivariable models of various aspects of infectious morbidity, including length of stay for severe infection standardized by time of follow-up. None of the models converged.
The growth indices were similar, except for height for corrected age at 9 and 12 months and head circumference from 3 to 12 months of corrected age (Table 4). To identify factors that might have influenced growth, we fitted the data to a robust regression model (F7568 = 32.46,P > F = 0.00 001) to estimate height as a proportion of expected height at 12 months of corrected age. After adjustment for gender, educational level of the mother, weight and gestational age at eligibility, and proportion of expected length for gestational age at birth, the crude differences disappeared. The growth of head circumference, expressed as a proportion of expected circumference at 12 months of corrected age, was associated with the KMC intervention (P = .014); these infants had a larger circumference than control infants. Surprising is that in a comparison of exclusively breastfed infants up to term or to 3 months of corrected age with infants who received either partial breastfeeding or bottle feeding, breastfeeding was not associated with the growth indices.
Psychomotor development was evaluated at 6 and 12 months of corrected age. The average overall and subscale Griffith quotients (100 × observed/expected score for corrected age) were for the most part satisfactory and were similar in the 2 groups (Table 5), even after adjustment for imbalances in gender and weight at eligibility by ANOVA–ANCOVA.
Both exclusive and partial breastfeeding were highly prevalent in both groups (46% exclusive and 98% partial) at 40 to 41 weeks of postconceptional age, but the proportions of KMC mothers who breastfed up to 3 months (exclusively or partially) were statistically significantly higher (Table 6). After the 3-month visit, the proportion dropped dramatically and decreased steadily to approximately 20% at 12 months of corrected age, and from 6 to 12 months, there was no difference between the 2 groups. The proportions of exclusively breastfed infants did not differ in the 2 groups at any time during follow-up.
Sequelae at 12 Months of Corrected Age
No difference was found between the 2 groups in the proportions of infants with cerebral palsy, psychomotor delay, or visual or hearing impairment (Table 7). The only factor associated with an increased risk for cerebral palsy was the total number of days spent in a neonatal intensive care unit, which reflects the severity of the initial condition of the infant.
Our results demonstrate that KMC is not associated with an additional risk for death. On the contrary, after control for weight at eligibility (baseline biological events) and level of education of the father (also representing differences in socioeconomic status), the point estimate for death for KMC infants was lower than that for control infants. This finding is not explained by differences in length of follow-up or time of occurrence of death, as shown by life-table analysis. Furthermore, almost identical results were obtained with Cox logistic regression analysis.
The sample size and previous observational data indicated expected mortality rates in the 2 study groups of 7% to 10% and 14% to 20%,6 whereas improvements in the general health status of the population and in the obstetric and neonatal services at the institution resulted in a dramatic reduction in overall mortality of the study patients. Our sample size computation therefore fell short, and we had insufficient power to detect a much smaller difference than expected. Nevertheless, the reduction in the mortality rate of KMC infants was almost statistically significant. As most of the deaths in both groups occurred during the first 3 months of corrected age, any enhancement of the intervention aimed at lowering mortality should address this seemingly critical period.
KMC reduced the number of total days spent in the hospital. The reduction was most marked for newborns with particularly low birth weights and was not seen for those who weighed >1800 g, as traditional care in the study institution suggests discharge when the infant weighs 1700 to 1800 g. In institutions where infants are discharged only when heavier, the savings in the hospital stay could extend to infants who weigh >1800 g. The savings in days spent in the hospital persisted up to the age of 12 months.
Use of ambulatory care, measured as the density of visits, was somewhat higher in the KMC group. Although KMC infants received many visits between eligibility and term, the control infants were in the hospital for most of this period, and visits from the attending pediatrician were not counted. Furthermore, the parents of KMC infants are more sensitive to their health needs and therefore may demand more ambulatory health care. We consider that ambulatory visits are preferable to prolonged hospitalization if the outcome is equivalent.
One of the potential criticisms of early discharge is that the savings in hospital stay might be lost as a result of an increase in the number of infections that require readmission. The number and length of hospitalizations as a result of infection were similar in the 2 groups. The control infants had statistically significantly more nosocomial infections and a greater need for inpatient care because of infections. This effect disappeared during prolonged follow-up because the risk for severe infection is high before 3 months of corrected age and decreases markedly thereafter.
As the use of much-needed neonatal hospital beds was reduced by KMC, it thus can alleviate the pressure on the already strained neonatal care facilities in less developed countries. Ambulatory care of infants who are discharged early with KMC requires close monitoring by qualified health personnel (eg, neonatologists, registered nurses) and efficient follow-up, including home visits backed by the possibility of readmitting sick infants to a tertiary care hospital.
No reduction in early physical growth was seen with KMC. This finding differs from that of our previous study6 and that of Dı́az-Rosello et al,9 perhaps because in our revised KMC program breast milk was supplemented with special formula for premature infants with daily weight gains of <20 g, independently of the assigned intervention. Having completed the data collection, our goal for KMC infants is that they gain at least 15 g/kg daily in weight and 0.7 cm in length per week.
We documented a statistically significant larger head circumference in KMC infants from 3 months of corrected age onward. Although the absolute difference is not impressive (0.5 cm on average at 12 months), it cannot be ignored. Prolonged follow-up of premature infants suggests an association between head size during the first year of life and cognitive development up to the age of 12 years.10–13After controlling for confounding factors, the head circumference of girls was closer to that expected for age and gender than that of boys, regardless of the assigned intervention. An interaction was also found between gender and type of intervention: girls who were given KMC had larger head circumferences. We have no explanation for this finding.
Psychomotor development was similar in the 2 groups. The only 2 factors that affected the Griffith scores were having been admitted to a neonatal intensive care unit (a marker of the severity of neonatal health status before eligibility) and gender; girls had higher average quotients. A modest but significant positive correlation was seen between head circumference and Griffith quotient (Spearman'sr = 0.1553; P = .0003), suggesting either a higher mortality rate among “intelligent” boys before eligibility or greater “intelligence” among girls, provocative interpretations that might add fire to the battle of the genders.
A full assessment of the potential role of KMC in the promotion and maintenance of breastfeeding was limited by 2 factors. First, Colombian law allows at best only 3 months of maternity leave. Although both study groups were exposed to an intensive campaign to promote breastfeeding, more KMC mothers breastfed their infants for the full 3 months. After that age, however, most of the full-time working mothers in both groups (more than half of the study population) returned to their jobs and found it difficult to maintain breastfeeding. Second, data were collected in categories that were different from those used by other investigators, making it difficult to compare the results. In the future, we plan to use the set of breastfeeding definitions developed for consistent use worldwide under the auspices of the Interagency Group for Action on Breast-feeding Definitions.14
As some otherwise healthy preterm infants did not gain weight properly when receiving exclusive breastfeeding early in their lives,6 in this study, we reluctantly supplemented their nutrition with preterm formula and in many cases were able to achieve appropriate weight-gain rates. We are now refining and enhancing the KMC intervention, with focus on the feeding policy. In the 669 of our 746 study infants who were available for evaluation, the average growth indices with the current feeding policy at 40 weeks of postconceptional age were 2809 g (±525 g) weight, 47 cm (±2.4 cm) height, and 34.6 cm (±1.6 cm) head circumference.3 We consider these results satisfactory, taking into account that the socioeconomic level of our study population is low and that Bogotá is 2640 m above sea level. In addition, the quoted figures include infants with preexisting intrauterine growth retardation, who represented 36% of the sample.
The sequelae identified seem to have originated from processes that occurred before eligibility, and it would be unrealistic to expect that KMC could modify these outcomes. There were no differences in sequelae between the groups. Because infants with major malformations, early neurologic impairment, and intrauterine infections were not included in the study, the prevalence of sequelae in the population is lower than that expected for a cohort of low-birth-weight infants.
Other studies of early use of skin-to-skin contact, 1 of the components of KMC, have been reported in the literature. Such contact can be started very early, almost immediately after birth and even in intubated infants, and seems to confer benefits not only in the maternal “healing” process2,,15 but also in stabilizing vital parameters,16 promoting breastfeeding, and improving weight gain. The outcomes of our low birth weight infants and the use of resources could be improved by earlier use of skin-to-skin contact, and we plan to evaluate the effects of this intervention in the neonatal intensive care unit.
The findings on low birth weight infants reported here and previously indicate that KMC can be recommended for populations and health facilities similar to that in the present study. We are aware that our findings of savings in the length of hospitalization and a decreased frequency of nosocomial infections should be supplemented with a formal economic analysis to justify the investment of resources necessary to start a KMC ambulatory clinic. Additional desirable components of the intervention are that it empowers families by giving them the responsibility and the tools for providing health care to their infants, and it might encourage breastfeeding and even the quality of family-to-infant bonding.4 All health professionals involved in caring for newborn infants therefore should consider using part or all of KMC to complement usual neonatal care. In particularly, KMC should be recognized as a means of humanizing such care.
All care during pregnancy, the neonatal period, and infancy is directed to promoting a healthy, well-adjusted adult with a high probability of successful integration into society. Although having a healthy and caring environment early in life does not guarantee the achievement of those goals, reaching a minimum physical health status, a given level of psychological and physical development, and a place within a well-adjusted and caring family certainly increase the likelihood of becoming a happy, healthy adult.
This study was funded jointly by the Instituto de Seguros Sociales de Colombia, the World Laboratory (NGO, Lausanne, Switzerland, Project number MCD13), and Colciencias (Colombian Government).
We thank all members of the Kangaroo Research Team: A. Mondragón, MD; R. Gómez, MD; M. Cristo, Psc; E. Vélez, Psc; M. Girón, SW; R. Martı́nez, RN; F.A. Gómez, RN; and M. V. Jiménez. Without their commitment and devotion, this study could not have been conducted.
- Received February 25, 2001.
- Accepted May 2, 2001.
Reprint requests to (N.C.) Carrera 7a #46-20, Apto 2001, Santa Fe de Bogotá DC, Colombia. E-mail:
- KMC =
- Kangaroo Mother Care •
- CI =
- confidence interval •
- ANOVA =
- analysis of variance •
- ANCOVA =
- analysis of covariance
- ↵Rey E, Martı́nez H. Manejo Racional del Niño Prematuro. Bogotá, Colombia: Universidad Nacional; 1983
- Charpak N,
- Ruiz-Pelaez JG,
- Figueroa de CZ,
- Charpak Y
- Charpak N,
- Ruiz-Peláez JG,
- Charpak Y
- ↵Voyer M. Prematurité I, II, III. Encyclopedie Medico-Chirurgical de Pediatrie. Actualization pour 1996. Paris, France: Elsevier; 1996
- ↵Hamilton LC. Regression With Graphics: A Second Course in Applied Statistics. Belmont, CA: Duxbury Press; 1992
- ↵Diaz-Rosello JL, Lozano P, Tenzer S. Impaired Growth of Low Birth Weight Infants in an Early Discharge Program. New York, NY: UNICEF; 1992
- Hack M,
- Breslau N
- Copyright © 2001 American Academy of Pediatrics