A longitudinal study of the physical growth and health of postinstitutionalized Romanian adoptees

Abstract

In the present longitudinal study, the physical growth and health of 36 early-deprived postinstitutionalized Romanian orphans (ROs; 17 boys) adopted by Canadian families are documented. Data were collected for each child at three time points: at 11 months postadoption, at 4.5 years of age and at 10.5 years of age. Data from the RO children were compared with those from children in two matched comparison groups (Canadian-born [CB] nonadopted children and early-adopted [EA] Romanian children without institutional experience).

Results indicated that there was considerable growth retardation in the RO children at times 1 and 2. Nonsignificant differences in height and weight among the RO, CB and EA groups were obtained at time 3, indicating significant growth catch-up. Precocious puberty, a documented feature of some postinstitutionalized children, was not observed in the present study population. At time 3, the RO children did not differ from the CB or EA children on indicators of puberty. The poor physical health of the RO children at time 1 was no longer apparent at time 3. According to parent reports, the RO children were equally as healthy as the CB and EA children.

Following the fall of the dictatorship of Nicolae Ceausescu in Romania in December 1989, the world became aware of the horrendous situations of more than 100,000 children languishing in over 600 state-run ‘orphanages’. The rearing conditions in these institutions were deplorable, and the children were extremely deprived in all respects. Malnourished, they spent the majority of their days alone in cribs lacking in physical, social, auditory and visual stimulation. Knowledge of this situation drove hundreds of Canadians to travel to Romania to adopt children. In 1990/91, approximately 700 children were adopted by Canadians from Romanian orphanages. Scientific study of these children subsequent to their adoption by middle-class families in Canada provides a rare opportunity to assess the effects of early severe deprivation and to determine whether adequate rearing later in life can compensate for early deficits. As the number of adoptions of postinstitutionalized children in Canada remains steady, at just under 2000 per year (1), it is of considerable significance to understand the sequelae of institutionalization and the potential for amelioration.

The Romanian Adoption Project began in 1991 with the purpose of tracking the development of Romanian orphans (ROs) adopted by Canadian families (2). These children and those in two matched comparison groups (described below) were comprehensively assessed at three time points: at 11 months postadoption (phase 1), when the majority were 4.5 years old (phase 2) and when the majority were 10.5 years old (phase 3). In the present paper, we present the longitudinal findings of the physical growth and health of these children.

PHYSICAL GROWTH

Arguably, one of the most immediately apparent deficits of institutionalized children is their profound growth delay. Previous research on postinstitutionalized children has consistently indicated that at the time of adoption, they are very small for their age, often below the fifth percentile for height and weight (3–9).

While the malnutrition of institutionalized children contributes to their growth deficiency, another contributing factor may be the poor quality of interaction and stimulation offered by the low caretaker-to-child ratio in these institutions (4). This type of growth deficit, known as psychosocial dwarfism, can be very serious. However, upon removal from stressful or neglectful conditions, children suffering from psychosocial dwarfism tend to make tremendous gains in both height and weight (10,11). Likewise, after a period of time in their adoptive homes, postinstitutionalized children have been observed to make remarkable gains in both height and weight (3–5,9,12). For example, Rutter et al (9) reported that although at the time of adoption 51% and 34% of his sample of Romanian adoptees fell below the third percentile for weight and height, respectively, approximately two years later only 2% and 1% of the sample remained that low in weight and height, respectively. Similarly, Johnson and Dole (4) reported that the majority of 252 children adopted from Eastern European institutions showed rapid growth catch-up once they were adopted. Nevertheless, at three years postadoption, length of institutionalization was correlated with physical size, and of those children who had spent eight months or more in an orphanage, 31% remained below the 10th percentile in height.

In the present study, we document the height and weight growth trajectories of postinstitutionalized Romanian adoptees from birth to age 10.5 years at approximately eight years postadoption, extending the duration of follow-up considerably beyond that previously reported in the literature.

ONSET OF PUBERTY

One explanation for the failure of some postinstitutionalized children to exhibit full growth catch-up may be the prevalence of precocious puberty in this population. It has been reported that girls, in particular, who are reared in deprived environments in early life followed by a nutritionally rich and stimulating environments are at risk for early onset of puberty (7,8,13,14). Suggested mechanisms underlying this phenomenon relate to hormonal activation as a result of rapid weight gain and dietary change (14). Completion of puberty is associated with completion of the adolescent growth spurt and, hence, early puberty is associated with shorter final height. In the present paper, we assessed pubertal development of the ROs and compared it with that of children in the two comparison groups.

MEDICAL PROBLEMS AND GENERAL HEALTH

Studies have consistently documented the poor health of postinstitutionalized children, which is characterized by high instances of infectious diseases (12,15–18). Hostetter et al (16) reported that 28 of 52 internationally adopted children had an infectious disease, including acute or chronic liver disease ascribable to hepatitis B, cytomegalovirus excretion, tuberculosis, intestinal parasites and scabies. In addition, Hostetter et al (17) reported that 168 of 286 internationally adopted children had clinical or laboratory findings indicative of serious medical illness.

In investigating children adopted specifically from Romanian institutions, Johnson et al (12) found only 10 of 65 adoptees to be judged physically healthy and developmentally normal. In this sample, 53% had hepatitis B and 33% had intestinal parasites. Results from a previous phase of the present longitudinal study were consistent with these findings. Ames et al (15) reported that the medical problems most frequently reported by parents of newly adopted ROs were intestinal parasites (31%), hepatitis B (hepatitis B surface antigen positivity) (28%) and anemia (15%).

Although we expected the previously institutionalized children in the present study to have poorer health at the time of adoption and beyond than their noninstitutionalized peers, there was reason to believe that these children would be less susceptible to certain medical conditions than children in the two comparison groups or Canadian children in general. Specifically, the hygiene hypothesis proposes an explanation for the rising number of children affected by allergic diseases and asthma. It has been demonstrated that an increase in infections during early childhood and number of siblings is inversely associated with several allergic disorders, including asthma (19–21). Further, it is explained that lower hygiene levels during early infancy and exposure to numerous children increases the likelihood of bacterial or viral infections, which in turn may “provide important signals to the newborn’s maturing immune system” (19).

The conditions that children experienced in Romanian institutions were, without a doubt, far less hygienic that any North American standard. Indeed, 18% of parents reported that their children were dirty or soiled when they first met them (22). These institutions were also characterized by many children sleeping in one room with up to 98 children in each room at times. Based on previous findings and the conditions of the Romanian institutions, we hypothesized that the RO children would be less likely to suffer from asthma and allergies than children in comparison groups or in relation to Canadian norms.

SUBJECTS AND METHODS

Participants

Findings reported in the present study were based on data from 36 RO children (17 boys, 19 girls), each of whom had lived in an orphanage for a minimum of nine months (range nine to 53 months, mean = 24 months). The age of these children at adoption and the total time spent in an institution are almost perfectly correlated (r=0.97), indicating that most had been institutionalized since birth. Data were also analyzed for a Canadian-born (CB), nonadopted, noninstitutionalized comparison group (n=42), 35 of whom were individually matched to RO children for sex, age (within three months at assessment) and family sociodemographic characteristics (family income, mother’s education and father’s education), and for an early-adopted (EA) comparison group (n=25) individually matched to the youngest RO and CB children. The CB children were recruited from a research participant bank. For the past two decades, four local hospitals have routinely distributed to new mothers a letter asking those who were interested in research participation to mail back an attached postcard. All children who were the same sex as and born within one month of each RO child were drawn from this bank. This yielded from 12 to 38 potential matches for each RO child. Parents of these children were contacted and asked about family demographic variables. The child who was the best overall match, particularly on mother’s education, was chosen for the CB comparison group. The EA children, also from Romania, were adopted before four months of age and came from hospitals, orphanages or their biological parents’ homes. These children shared similar birth family histories, and pre- and perinatal care with the RO children, and were destined to be raised in orphanages similar to those from which the RO children were adopted. However, because they were adopted early in life, they did not share the extensive institutional experience. Demographic data on the three groups can be found in Table 1.

TABLE 1.

Mean scores and standard deviations for demographic characteristics of Romanian orphan (RO), Canadian-born (CB) and early-adopted (EA) groups in phase 3

	ROa			CBb			EAc
	Mean	SD	n	Mean	SD	n	Mean	SD	n
Time in institution, months	22.67	13.40	36	–	–	–	2.67	1.15	4
Age at adoption, months	23.97	14.86	36	–	–	–	2.42	1.44	23
Age at assessment, months	127.58	12.84	36	127.00	12.63	42	119.84a,b	2.46	25
Mother’s education*	4.26	1.04	35	4.30	0.91	40	4.63	0.82	24
Father’s education*	4.28	1.42	32	4.51	1.07	39	4.67	1.09	24
Mother’s age, years†	44.2c	6.13	36	43.1a,c	4.54	42	47.0	6.44	25
Father’s age, years†	46.3c	6.24	33	44.8a,c	4.89	41	48.6	6.06	25
Income‡	6.03	2.46	35	6.88	2.24	40	6.67	2.62	24

^*1 = elementary school, 2 = some high school, 3 = high school completion, 4 = vocational or some college or university, 5 = college or university graduate, 6 = graduate or professional school;

^†Age at time target child was assessed;

^‡Gross annual income: 1 = less than $20,000, 2 = $21,000 to $30,000, 3 = $31,000 to $40,000, 4 = $41,000 to $50,000, 5 = $51,000 to $60,000, 6 = $61,000 to $70,000, 7 = $71,000 to $80,000, 8 = $81,000 to $90,000, 9 = $91,000 to $100,000, 10 = above $100,000.

^{a, b, c}Indicate means that differ significantly (P<0.05) from one another

In a longitudinal study, one of the major problems is attrition. In phase 1, the sample comprised 46 RO, 46 CB and 29 EA children. Between phases 1 and 2, three RO, one EA and five CB families were lost. At phase 2, three new RO families who had not participated in phase 1 were found and added to the sample. Eight new CB children were added at phase 2, three to be matches for the new RO children and five to replace the lost CB children. In phase 3, an additional 11 RO, five CB and five EA families either declined to participate or could not be found. Five new CB families were added in phase 3. All longitudinal analyses included data only from participants for whom data were available at each time point included in the analysis. For this reason, the number of cases in each analysis was variable.

Procedures and measures

Anthropometric measurements

Weight data were obtained from birth records and parent reports at the time of adoption and in phases 1, 2 and 3. Height data were obtained from parent reports in phases 2 and 3.

Pubertal development

Puberty data were obtained in phase 3 from parent and child self-reports adapted from questionnaires used in the National Longitudinal Study of Children and Youth (NLSCY) and based on the Tanner stages. For girls, indices of puberty were body hair, breast development and menarche. For boys, indices of puberty were voice change, facial hair and body hair. For each index, except menarche, parents and children rated development on a four-point scale: 1 = has not started growing, 2 = barely started, 3 = definitely underway and 4 = seems nearly completed. Menarche was rated as either having or not having occurred.

Medical problems and general health

During comprehensive interviews in phase 1, parents were asked to describe any medical problems their children experienced. In phase 3, parents were asked to complete a questionnaire regarding their children’s health. The questionnaire, developed for the NLSCY, consisted of 51 questions concerning both general and specific health matters. For the present study, only four of those questions were used: In general, how would you say your child’s health is? In the past few months, how would you say your child’s health has been? Has your child ever had asthma that was diagnosed by a health professional? and, Has your child ever had allergies that were diagnosed by a health professional? Parents rated the first two questions on a five-point scale: 1 = excellent, 2 = very good, 3 = good, 4 = fair and 5 = poor. The last two questions were rated as either having or not having the condition in question.

RESULTS

Results are presented in three main sections. First, longitudinal height and weight data are reported, comparing the RO group with the two comparison groups, CB and EA children, as well as with Canadian norms. Second, the pubertal development of the RO children is presented, with comparisons made with the CB and EA groups. Lastly, the health of the RO children is examined over time and in relation to the two comparison groups.

Height and weight

All weight and height data were transformed to percentile scores using North American growth charts (23). Longitudinal height and weight data are presented in Table 2. The mean birth weight percentile for the 20 RO children whose parents had birth weight information was 22nd. At the time of adoption, the RO children, as a group, had dramatically fallen off their growth curves, with the average weight at adoption dropping to the fourth percentile. Specifically, 66.7% (n=20) fell at or below the third percentile for North American norms, 80% (n=24) fell at or below the fifth percentile, 93.3% (n=28) fell at or below the 10th percentile, and 100% (n=30) fell at or below the 15th percentile. For the RO children for whom birth weight data were available, the correlation between birth weight and weight at adoption was virtually zero (r=−0.02, P=0.945).

TABLE 2.

Percentage of Romanian orphan children with height and weight below the third, fifth, 10th and 15th percentiles^* in each phase

	n	Mean percentile	SD	<3rd percentile (%)	<5th percentile (%)	<10th percentile (%)	<15th percentile (%)
Birth weight	20	21.8	19.2	25	25	40	40
Weight at adoption	30	3.97	3.67	66.7	80	93.2	100
Weight in phase 1 (11 months postadoption)	27	7.85	4.69	22.2	37	70.4	85.2
Weight in phase 2 (4.5 years of age)	35	43.7	25.7	8.6	11.4	20	20
Weight in phase 3 (10.5 years of age)	33	59.9	28.06	2.8	2.8	6.1	9.1
Height in phase 2 (4.5 years of age)	30	36.98	31.08	20	23.3	26.7	36.7
Height in phase 3 (10.5 years of age)	33	48.8	32.2	12.1	12.1	15.2	18.2

^*Percentiles based on Canadian norms

Growth catch-up was evident soon after adoption. At 11 months postadoption, although the children were still significantly underweight (mean weight below the eighth percentile), improvements could be seen, with only 22.2% (n=6) falling below the third percentile. By phase 2, when the RO children were, on average, 4.5 years of age, almost complete weight catch-up could be seen, with their mean weight near the 50th percentile and only 8.6% (n=3) below the third percentile. There was, however, still a statistically significant difference between the RO and CB children’s weight (P<0.05; Table 3). In phase 3, when the children were, on average, 10.5 years of age, their weight catch-up continued, with the mean weight nearing the 60th percentile and only 2.8% (n=1) below the fifth percentile. Moreover, there were no longer statistically significant differences among the groups.

TABLE 3.

Mean height and weight percentiles^* for Romanian orphan (RO), Canadian-born (CB) and early-adopted (EA) groups in each phase

	ROa		CBb		EAc
	Mean	SD	Mean	SD	Mean	SD
Birth weight	21.8	19.2
Weight at adoption	3.97	3.67
Weight in phase 1	7.85	4.69
Weight in phase 2	43.7b,c	25.7	61.5	30.25	57.9	28.12
Weight in phase 3	59.9	28.06	60.0	28.05	56.3	35.53
Height in phase 2	36.98b,c	31.08	66.7	26.41	53.3	27.73
Height in phase 3	48.8	32.2	64.3	32.40	57.6	37.15

^*Percentiles based on Canadian norms.

^a,b,cIndicate means that differ significantly (P<0.05) from one another

Height data were not available either at the time of adoption or in phase 1; however, data from both phase 2 and phase 3 indicate that height may have been equally delayed at adoption because weight and growth catch-up began shortly after adoption. In phase 2, the RO children’s mean height was near the 40th percentile, with 20% (n=6) below the third percentile. At this time, statistically significant differences were observed among the groups, with the RO children significantly smaller in stature than the CB children (see Table 3). In phase 3, the RO children’s mean height was at the 49th percentile, with only 12.1% (n=4) below the fifth percentile. Again, by phase 3, there were no longer statistically significant differences among the groups, indicating that eight years after adoption, the RO children’s height and weight had caught up to those of their nondeprived peers.

Within the RO group, weight data from phases 1, 2 and 3 and height data from phases 2 and 3 were examined in relation to length of institutionalization before adoption. Correlation coefficients were nonsignificant, ranging from 0.004 to −0.23, indicating virtually no association between duration of institutional experience and subsequent growth.

Height and weight data were obtained for the EA and CB children only in phases 2 and 3. Table 3 contains longitudinal mean height and weight percentile data for all three groups. As can be seen in Table 3, at phase 2, when the RO children were still below average for height (mean 37th percentile) and weight (mean 44th percentile), the EA children had a mean weight at the 58th percentile and a mean height at the 53rd percentile. Additionally, in phase 2, only 9.5% (n=2) of the EA children were below the 15th percentile on weight (one child each at the first and ninth percentiles) and only 16.7% (n=3) were below the 15th percentile on height (one at the seventh percentile and two at the 12th percentile). In phase 2, indices of growth for the EA group were intermediary to those for the RO and CB groups, not differing significantly from either on height or weight.

Pubertal development

Because previous research has focused on the effects of institutionalization on the pubertal development of girls, analyses in the present paper examine the pubertal development of girls and boys separately. The correlation between girls’ and boys’ self-ratings and parent reports of pubertal development indicated significant agreement between sources (r=0.57, P<0.01 for girls; r=0.31, P<0.05 for boys). Given our belief that children’s self-reports are a more accurate indicator of pubertal development (parents rarely see children naked after the onset of puberty), only self-report data were used in the remaining analyses on pubertal development. The authors acknowledge that a more objective measure of pubertal development would have been a clinical evaluation by a physician, but such data were unavailable.

To determine whether the RO children were experiencing early onset of puberty, their self-ratings on each index of pubertal development were compared with those in the two comparison groups (Tables 4 and 5). For these comparisons, only children who were younger than 125 months of age were included (RO group, n=21; CB group, n=27; EA group, n=24) because older children were beyond the age at which the onset of puberty could be considered early. As can be seen in Tables 4 and 5, no striking differences in pubertal development among the groups were observed.

TABLE 4.

Number and percentage of Romanian orphan (RO), Canadian-born (CB) and early-adopted (EA) girls reporting each degree of development on indices of puberty^*

	RO, n (%)	CB, n (%)	EA, n (%)
Menarche
Yes			1 (7)
No	10 (100)	13 (100)	13 (93)
Breast development
Not started	4 (40)	3 (23)	3 (21)
Barely started	5 (50)	9 (69)	9 (64)
Definitely underway	1 (10)	1 (8)	2 (14)
Nearly complete
Body hair growth
Not started	7 (70)	6 (46)	5 (36)
Barely started	2 (20)	5 (39)	8 (57)
Definitely underway		2 (15)	1 (7)
Nearly complete	1 (10)

^*Data only for children younger than 125 months

TABLE 5.

Number and percentage of Romanian orphan (RO), Canadian-born (CB) and early-adopted (EA) boys reporting each degree of development on indices of puberty^*

	RO, n (%)	CB, n (%)	EA, n (%)
Voice change
Not started	9 (75)	7 (50)	4 (50)
Barely started	1 (8)	4 (29)	3 (37)
Definitely underway	2 (17)	3 (21)	1 (12)
Nearly complete
Facial hair growth
Not started	9 (75)	12 (86)	2 (25)
Barely started	3 (25)	2 (14)	4 (50)
Definitely underway			2 (25)
Nearly complete
Body hair growth
Not started	6 (50)	7 (50)	4 (50)
Barely started	6 (50)	4 (29)	2 (25)
Definitely underway		3 (21)	2 (25)
Nearly complete

^*Data only for children younger than 125 months

Medical problems and general health

As previously reported by Ames et al (15), at the time of adoption, 31% of RO children were reported to have intestinal parasites, 15% had anemia and 28% had hepatitis B. Although we were unable to determine the national incidence of intestinal parasites or childhood anemia, the national incidence of hepatitis B in 1992 (shortly after most RO children were adopted) was slightly more than 10 per 100,000 population or 0.010% (24). In other words, the RO children were more than 300 times more likely to have hepatitis B than the average Canadian. Despite the early health problems of the RO children, in phase 3 there were no statistically significant differences among the groups on parent’s rating of their child’s health ‘in general’ or ‘in the past three months’.

To address the hypothesis that the incidence of allergies and asthma would be lower in the RO group than in the comparison groups, the prevalence of these conditions was examined in each group. Data indicated that fewer children in the RO group suffered from asthma than in either the EA group or the CB group. Specifically, only two (5.7%) RO children were reported to have asthma, while nine (22.5%) of the CB children and four (17.4%) of the EA children were reported to have it. χ² analyses comparing pairs of groups on incidence of asthma indicated significant differences between the RO and CB groups (χ²=4.20, P<0.05). Differences between the RO and EA groups and EA and CB groups were not statistically significant. These incidence figures are interesting to consider in comparison with the national incidence of asthma for Canadian children, which is 12.2% for those under 19 years (25). Another way to look at these data is to point out that the incidence of asthma in the RO group was one-quarter of that found in the CB group and less than one-half of the rate found in the Canadian childhood population. Within the RO group, five (14.3%) children were reported to have allergies in comparison with seven (17.5%) CB children and 10 (43.5%) EA children. χ² analyses comparing pairs of groups indicated that the EA group had a significantly higher rate of allergies than both the RO group (χ²=6.17, P<0.05) and the CB group (χ²=5.00, P<0.05). The RO and CB groups did not significantly differ on incidence of allergies.

DISCUSSION

The results reported here extend previous research by examining the physical growth and health of postinstitutionalized children (4,9). At the time of their adoptions, the physical impact of the extreme deprivation experienced by the RO children was clear, with two-thirds at or below the third percentile for weight and the majority suffering from some medical problem (26). Despite their condition at adoption, their catch-up in growth and medical recovery was spectacular.

In terms of physical growth, the RO children’s mean weight rose from below the fourth percentile at the time of adoption to the eighth percentile after only 11 months in their adoptive homes. By the time these children were 4.5 years of age, and, thus, had spent approximately 2.5 years in their adoptive homes, their mean weight was nearing the 50th percentile. By phase 3, the RO children’s physical growth had reached North American norms, and there were no longer differences between them and the comparison groups on either height or weight. This dramatic catch-up following a major change in their rearing environments provides clear evidence that their initial growth retardation was caused by the profoundly depriving circumstances of the institutions. This is further supported by the lack of an association between birth weight and weight at adoption. While the majority of the RO children were underweight at birth, it was nowhere near the extreme degree of underweight at the time of adoption, indicating that a disruption of their natural growth trajectories had occurred. The findings reported here on the growth catch-up in the RO group are consistent with those of Rutter et al (9), who reported on a similar group of postinstitutionalized Romanian adoptees. Rutter et al (9) reported that at the time of their entry into the United Kingdom, the percentage of Romanian adoptees with weight, height and head circumferences below the third percentile according to United Kingdom standards was 51%, 34% and 38%, respectively. In addition, they also reported a dramatic catch-up in growth for the Romanian adoptees following time in their adoptive homes. By four years of age, only 2% of the Romanian adoptees were below the third percentile on weight, and only 1% on height and 13% on head circumference.

In contrast to Johnson and Dole (4), who found that at three years postadoption, length of institutionalization was correlated with physical size in a sample of Eastern European adoptees, we found no such relationship in the present study. Different ranges in institutional experience in the two samples may explain this difference in findings between the two studies. Johnson and Dole’s (4) sample included children with minimal institutional experience, whereas all children in the present sample had spent at least nine months in an institution before adoption. It may be that beyond a certain threshold, variation in length of institutionalization has no impact on subsequent growth.

While the physical growth of the RO children differed dramatically from that of the CB and EA children, this was not the case in regard to pubertal development. This is inconsistent with previous research documenting that institutionalization followed by rapid growth catch-up leads to precocious or early onset of puberty (7,8,14). One possible reason that we did not observe earlier puberty in the RO children is that they were assessed at 10.5 years of age and, thus, were still too young to demonstrate differences in pubertal onset. Moreover, it has been reported that it is not merely a greater likelihood of early onset of puberty that distinguishes postinstitutionalized children from their peers, but also a faster rate of pubertal completion. Continued follow-up of these children further into adolescence would allow for these issues to be addressed.

Alternatively, it may be that the RO children in the present study will not experience early onset of puberty because of the length of time they were deprived before adoption. In a study of rats, Bourguignon et al (13) discovered that food-restricted pups that were later allowed to resume normal feeding showed earlier maturation than well-fed pups. However, the effect was only observed when unrestricted feeding resumed before weaning, indicating that hypothalamic maturation is sensitive to changes in nutrition only during a critical period. If the rat model extends to humans, it is possible that the RO children were adopted after this critical time period and, therefore, would not experience precocious puberty. It is interesting that of the three groups studied here, it was only the EA group who showed any tendency toward early puberty, and, admittedly, this tendency was not strong. These children were also nutritionally deprived very early in life, but only for a short time. The suggestion of early puberty in this group may indicate that their change in environment (ie, adoption) occurred during the critical period identified by Bourguignon et al (13) in rats. Again, continued study of these children may shed more light on this issue.

With respect to health status, at the time of adoption, the RO children experienced multiple medical problems and malnutrition. Therefore, it is remarkable that after approximately eight years in their adoptive homes, parents of RO children rated their children’s general health on average between ‘excellent’ and ‘very good’. Moreover, it is equally noteworthy that there were no differences between the RO parents’ ratings of their children’s health and parents’ ratings of the EA and CB children. One limitation to this finding, however, is our reliance on parent reports. It is possible that what the parents of RO children think of as ‘excellent’ or ‘very good’ health for their children is different from what the parents of EA and CB children think. This possibility is likely considering the health of the RO children at the time of adoption.

In regard to specific health conditions, almost all of the RO children who had contracted hepatitis B in the institutions were able to develop antibodies and become symptom free. In addition, in phase 3, fewer RO children were reported to have asthma and allergies than were CB children and EA children, respectively. Consistent with the hygiene hypothesis, it is possible that the crowded and unsanitary conditions that characterized Romanian institutions may have played a role in lowering the RO children’s likelihood of developing allergies and asthma.

Similar to other studies assessing the development of postinstitutionalized children, the present longitudinal study has several unavoidable limitations. First, there were no reliable data on the characteristics of the biological parents for most of the children. Therefore, no information on possible genetic influences on the children’s physical growth and health could be taken into consideration. Undoubtedly, genetic factors played a role in the children’s development, but they are not likely to have been responsible for the growth catch-up observed in the RO children or for the infectious diseases that the RO children contracted while institutionalized. In addition, the EA children came from the same population as the RO children and, therefore, can be assumed to originate from similar genetic backgrounds. If genetic factors were responsible for the RO children’s physical growth, the EA children would be expected to demonstrate similar growth and health trajectories.

A second limitation is that we had no systematic measurements of the children’s physical growth at the time of adoption and, thus, had to rely on parents’ retrospective accounts. Because of this, we cannot be certain of the children’s weights at adoption and, hence, the change in their growth trajectories. However, all of the parents had their children weighed by health professionals near the time of adoption and, thus, their reports are likely reliable. We also need to be cautious of any conclusions we draw from the birth weight data because this information was provided to the adoptive parents by the institutions, and the reliability of it is unknown.

CONCLUSIONS

Although it may still be too early to make any conclusive statements, it appears that despite their severe early global deprivation, the RO children have been able to make remarkable gains in the areas of physical growth and health. Continued study of these children into adolescence will continue to expand our knowledge of the importance of early life experiences on physical growth and health and, in particular, the onset and completion of puberty.