Contents - Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Hector Balcazar and José A. Cobas
This study investigates the relationship between maternal biological, nutritional, and socio-demographic variables as they relate to intra-uterine growth retardation (IUGR) in 162 women who delivered in a maternal and child centre in Mexico City. Data were obtained through a questionnaire administered to each woman after her infant was born. The mothers' height, arm and head circumferences, and skinfolds were measured after delivery. Infants were defined as having IUGR if their gestational age was 37 weeks or over and their birth weight was at or below the 10th percentile of a sex-specific birth weight-gestational age distribution of a reference population. Infants whose gestational age was 37 weeks or over and whose birth weight was above the 10th percentile of the reference population distribution were defined as having normal birth weight. Overall, maternal risk factors associated with past and current nutritional status were associated with IUGR, but social factors were not. Logistic regression analysis showed that height, positive dietary changes during pregnancy, and past spontaneous abortions during the second trimester were significantly related to IUGR.
Low birth weight constitutes an important public health problem, particularly in developing countries . Studies have documented that low-birth-weight infants are heterogeneous with regard to their growth and development characteristics at birth [2: 3]. failing into two major groups: those who are premature or preterm, and those with intra-uterine growth retardation (IUGR) . Our concern is with the latter group.
In developing countries a large percentage of low-birth-weight infants who are born at term suffer some type of foetal growth retardation . They tend to have higher mortality and morbidity during the perinatal period than do full-term infants with adequate weight for gestational age [5-7].
Intervention strategies to decrease the prevalence of IUGR in developing countries require knowledge of the aetiological risk factors associated with this condition. For example, in chronically malnourished populations. maternal indicators of past nutritional status are associated with IUGR . Therefore, identification in clinical settings of maternal risk factors associated with IUGR is useful for designing timely interventions in a given population group. We assessed IUGR in newborns from an urban poor population of Mexico City, and investigated the relationship of maternal biological, nutritional, and sociodemographic variables with IUGR.
Materials and methods
The sample consisted of women who delivered in a maternal and child centre in Mexico City during the spring of 1984. The centre was established to provide health services for poor pregnant women who were not covered by any social security plan, and their infants. A total of 162 women participated in the study.
Data were obtained through a questionnaire administered to each woman after her infant was born in the centre. After delivery, the mothers' height, arm circumference, head circumference, and skinfolds were measured by a nutrition student trained by the principal investigator to use standard procedures. Pre-pregnancy weight and dietary information was collected by maternal recall. A locally developed food frequency questionnaire was used to explore dietary adequacy by identifying positive dietary changes during pregnancy; if the mother complemented her diet by consuming extra foods from any of the basic four food groups, it was coded as a positive dietary change.
Shortly after delivery, birth weights were determined by registered nurses using infant weighing scales that were checked and calibrated periodically. Each infant's weight was recorded to the nearest 10 g. Gestational age was calculated from the mothers' menstrual histories by the attending physicians. Computed gestational age was expressed in completed weeks on the basis of the last normal menses and the date of delivery. Unfortunately. information as to the reliability of the gestational age calculation was not available. The variables collected and their codings are shown in table 1.
TABLE 1. Independent variables and their coding
|Biological profile and maternal anthropometry|
|Maternal age||Completed years at delivery|
|Parity||Parity order of present birth|
|Age at menarche||years|
|Factors associated with current pregnancy|
|Systolic blood pressure||mm Hg|
|Diastolic blood pressure||mm Hg|
|Dietary changes during pregnancy||(0.1)a|
|Illiteracy of head of household||(0. 1)a|
|Illiteracy of mother||(0.1)a|
|Schooling of head of household||years completed|
|Schooling of mother||years completed|
|Income factors and marital status|
|Employment status of head of household||(0.1) (1 = permanent)|
|Household food expenditures||pesos/month (1984 rate)|
|Total expenditures of household||pesos/month|
|Income of head of household||pesos/month|
|Wages of head of household||(0.1) (1 indicates <= minimum wage)|
|Marital status||(0.1) (1 = no partner present in household)|
|Housing living. and sanitary conditions|
|Use of kitchen as bedroom||(0.1)a|
|Presence of animals in house||(0.1)a|
|Water facilities in house||(0.1)a|
|Bathroom facilities in house||(0.1)a|
|Number of persons per bedroom||(0.1) (1 indicates >2)|
|Electricity in house||(0.1)a|
|Access to public garbage collection||(0.1)a|
|Use of public services||(0.1)a|
|Regular use of health services||(0. 1)a|
|Factors associated with earlier pregnancies|
|Spontaneous abortions in 1st trimester||(0.1)a|
|Spontaneous abortions in 2nd trimester||(0.1)a|
|Birth of weight <2.5(N) g||(0.1)a|
|Birth at <37 weeks||(0.1)a|
a. 1 = noted.
b. 1 = less cxpcnsive material.
The infants in the study were divided into two categories. They were assigned to the IUGR group if their gestational age was 37 weeks or more and their birth weight was at or below the 10th percentile of a sex-specific birth weight-gestational age distribution of a reference population . Infants were defined as being of normal birth weight (NBW) if their gestational age was 37 weeks or more and their birth weight was above the 10th percentile of the reference population distribution. Of the 162 infants, l 10 were classified as NBW, 27 as IUGR, and 25 as preterm. Infants identified as preterm (gestational age less than 37 weeks) were not included in the present analysis.
Means and standard deviations were calculated for selected independent variables in the NBW and IUGR groups. Group mean differences were tested using t-tests. Relative risk and 95% confidence intervals were calculated for selected independent variables. The Mantel-Haenszel chi-square statistic was used to test the significance of the risk ratios. Finally, logistic regression analysis was performed to assess the relative contribution of the variables found to be significantly associated with IUGR. The data were analysed using the SAS statistical package on an IBM mainframe computer.
The means and standard deviations of birth weight for the NBW group (>=37 weeks, >10th percentile) were 3,199 and 423 g respectively (N= 110). The IUGR group (>=37 weeks, <=10th percentile) means and standard deviations were 2,382 and 225 g respectively (N = 27). Both groups had the same mean gestational age of 39.4 weeks. The NBW group showed a mean value of crown-heel length of 50.5 cm, compared to 47.3 cm for the IUGR group.
The t-tests were performed to identify differences between the groups with respect to factors posited to affect intra-uterine growth retardation. Table 2 identifies among selected variables those that reached statistical significance at the .05 level or less. Mothers of IUGR infants scored lower on height, milk consumption, and positive dietary changes than did mothers of the NBW group. In addition, the former had a greater proportion of previous spontaneous abortions in the second trimester of pregnancy than the latter.
TABLE 2. Means and standard deviations of selected independent ar ables for the groups with normal birth weight (NBW) and intra-uterir growth retardation (IUGR)
|NBW (N = 110)||IUGR (N = 27)|
|Arm circumference (cm)||25.2||2.2||24.9||2.4|
|Head circumference (cm)||51.9||1.4||51.7||2.4|
|Subscapular skinfold (mm)||19.4||5.1||18.6||4.2|
|Pre-pregnancy weight (kg)||52.9||7.6||51.2||9.7|
|Milk consumption (days/wk)*||4.7||2.6||3.1||2.5|
|Cheese consumption (days/wk)||1.9||1.9||2.5||2.0|
|Mother's schooling (years completed)||6.4||3.6||7.4||3.6|
|Dietary changes during pregnancy (1 = noted)*||0.76||0.4||0.40||0.5|
|Use of kitchen as bedroom (1 = noted)||0.80||0.3||0.62||0.4|
|Number of persons per bedroom (1 = >2)||0.35||0.4||0.55||0.5|
|Access to public garbage collection (1 = noted)||0.94||0.2||0.85||0.3|
|House ownership ( 1 = noted)||0.15||0.3||0.03||0. I|
|Spontaneous abortions in 2nd trimester(1 = noted)*||0.02||0.1||0.14||0.3|
|Previous birth of weight >2,500 g (1 = noted)||0.11||0.3||0.25||0.4|
*Group differences significant at P <=.05.
Table 3 presents the relative risk ratios and 95% confidence intervals for the independent variables that were significant in the univariate analysis. To dichotomize the continuous variables, the 25th percentile value of each variable in the sample was used as the cut-off point. The NBW group was used as the reference in the two-by-two risk analysis tables. The exposed group for each variable comprises those below that level. Short stature (reflection of chronically deficient maternal nutritional status), no positive dietary changes during pregnancy, and previous spontaneous abortions during the second trimester were significantly associated with IUGR.
TABLE 3. Relative risk ratios and 95% confidence intervals for selected variables for the IUGR group
|Maternal height < 146.8 cma||2.20 *||1.10-4.37|
|Milk consumption < 3 days/wka||1.33||0.65-2.69|
|No dietary changes during pregnancyb||3.29**||1,71-6.32|
|Spontaneous abortions in 2nd trimesterb||3.22**||1.30-7.93|
a. Cut-off points for the 25th percentile of the total sample
b. Dichotomized (positive risk associated with presence of the characteristic).
*P < .05 for Mantel-Haeuszel chi-square statistic.
**P < .01 for Mantel-Haeuszel chi-square statistic.
To assess whether these differences persisted in a multivariate setting, logistic regression analysis was used (table 4). In the logistic model that was fitted. the dependent variable represents assignment to the IUGR group as opposed to the NBW group. The independent variables represent the three factors in table 3 with respect to which the two groups showed significant differences. As the mothers' positive dietary changes during pregnancy increased, the probability of their infants' being classified as IUGR decreased. The same held true for increased maternal height. The probability of IUGR classification increased by the occurrence of spontaneous abortions in the second trimester.
TABLE 4. Logistic regression analysis for selected maternal characteristics for the IUGR group
|Dietary changes during pregnancy||- 1.8**||0.50|
|Maternal height||- 0.1**||0.04|
|Spontaneous abortions in 2nd trimester||1.6*||0.84|
R-squared equivalent for model = 15.
*P < .05.
**P < .01.
Maternal risk factors associated with past and current nutritional status were associated with intrauterine growth retardation. The presence of risk factors associated with earlier pregnancies were also related to IUGR. Social factors, however. were not. The multivariate analysis confirmed the association between maternal indicators of nutritional status and past obstetric performance with IUGR.
Several limitations of the study should be mentioned before discussing the implications of these results. The sample was small (162 infants), and therefore the number of infants classified as IUGR was few. Since gestational age was calculated from the last menstrual period based on the mothers' recall, some misclassification of IUGR infants must have occurred as a result of errors. Errors in determining birth weight are also likely to have occurred. Finally, caution should be exercised when defining IUGR using local or international reference growth standards. Differences in the cut-off points of the percentiles of birth weight by gestational age may lead to the selection of infants whose risk might not be the same.
Despite these caveats, our results demonstrated that mothers of IUGR infants have a risk profile characterized by the presence of factors that are associated with nutritional and health conditions of pregnant women. Similar studies conducted in Guatemala support these results [8; 10]. Simple dietary information collected as a proxy for current nutritional status was significantly associated with IUGR. This indicates the need to evaluate simple indicators of dietary assessment as proxies for current nutritional status, especially when information on weight gain or other indicators is not available.
The socio-economic variables studied showed no predictive power when trying to distinguish IUGR from NBW infants. Studies conducted in rural Mexico have demonstrated the usefulness of socio-economic indicators for discriminating pre-school children with a high risk of being undernourished from those who are not [11; 12]. We found, however. that some of the social variables had limited variability. This could be a reflection of the homogeneity in the sample of women with regard to their socio-economic status. If such is the case, it is necessary to identify more meaningful social indicators as markers of perinatal risk.
The identification of IUGR beyond the criterion of birth weight under 2,500 g made it possible to characterize the maternal risk factors for infants most likely to have suffered some type of foetal growth retardation even if their birth weights were greater than 2.500 g.
In summary, there appears to be justification for using the combined information of birth weight and gestational age, together with data associated with maternal risk factors, to characterize better the problem of IUGR in developing countries. Data such as those presented here should be evaluated in developing countries so that the problem of low birth weight can be better defined, together with the combination of maternal and social risk associated with it.
Hernán Delgado, Patricia Palma, and Magda Fischer
Height censuses - involving the measurement of all children attending first grade - have been used in Central America to detect growth retardation to screen high-risk groups and to target social interventions. Data on age sex residence. and socio-economic characteristics are obtained from the teachers. Studies carried out by INCAP hare confirmed the simplicity and low cost of analyses of height-census data as well as their reliability and validity. In the Central American region data on height retardation reflect the variation found among countries when other socioeconomic and biological indicators such as illiteracy infant and child mortality and social services are considered.
The growth of pre-school children is commonly used as an indirect indicator of their nutrition status. In addition, since growth in underprivileged populations is mainly determined by environmental factors such as food intake and illnesses, growth retardation could he used as an indicator of inadequate quality of life . For this reason, the height retardation of children who enter the first grade of elementary school has been proposed as a useful indicator to identify highrisk population groups with poor health. undernutrition, and low socio-economic status.
This paper discusses the results when an entire region, the Central American isthmus, used height censuses to detect growth retardation, to screen highrisk groups, and to target social interventions. This methodology has also been used to evaluate the impact of the last. Thus. the height census may he recommended for three of the four purposes identified for food and nutrition surveillance systems: planning, evaluation. and advocacy .
Materials and methods
The height census consists of measuring all children who attend first grade of primary school. In addition. information on age, sex, residence, and, in some cases, socio-economic characteristics are obtained by the first-grade teachers. The teachers are trained in a standardized method of obtaining data, using simplified guidelines describing the measurement techniques. They also are provided with cardboard measuring charts and a drawing triangle. A more detailed description of the methodology is presented elsewhere . The quality of height information collected by teachers has been shown by INCAP to he highly reliable .
Data processing and flow of information have used two approaches: In some countries all the information collected is sent to the central level for recording and analysis. Elsewhere, the teachers summarize the information so that the central level receives a summary from each school describing the magnitude of height retardation by age group and sex. The latter approach simplifies processing at the central level, thus reducing logistic and other costs; however, information that could be used for other analyses at the individual level is lost. INCAP has participated in the quality control in both methods, demonstrating that the information processed and summarized at the field level is as reliable as that processed at the central level.
The final analysis and interpretation of the information generally are the responsibility of the nutrition departments of the ministries of health and education or of specialized data-processing centres. The primary analysis includes descriptive statistics, such as calculations of mean height and proportions of children with height retardation by age group, sex, and place of residence. In the calculation of height retardation, the height of a child at a given age is compared with the median height for children of the same age in the WHO reference population, derived from the NCHS reference height-for-age standard, and the corresponding Z-score value is calculated. Height retardation in a child has been defined as being shorter than 2 SD below the median height of the reference population. The expected proportion of children in that category in a well-nourished population is 2.5% .
Results of the analyses have been disseminated through reports for each country presented at regional technical meetings [6; 7]. These present descriptive information that allows the identification of highrisk geographic areas, based on the aggregate of schools with high proportions of height retardation.
Table 1 presents information on the countries' demographic characteristics as well as on the height censuses. As can be seen, there is considerable variation in the countries' illiteracy rates: Costa Rica has the lowest rate, and Guatemala the highest. All the countries of the region have carried out one or more censuses during the last decade. Costa Rica began this activity in 1979 and completed its fifth census in August 1989. The table also shows that there is considerable heterogeneity in the magnitude of height retardation among the countries; it is greatest in Honduras and Guatemala and lowest in Costa Rica. Secular changes can be identified in the countries with more than one census, Costa Rica and Panama. In Costa Rica, height retardation decreased by 45% from 1979 to 1985, whereas in Panama there was a decrease of approximately 22% from 1982 to 1985, followed by an increase for the period 1985-1988. Figure 1 (see FIG. 1. Mean heights of first-grade schoolchildren in four countries compared to the World Health Organization reference standard shows the mean heights for Costa Rica, Panama, Honduras, and Guatemala, as well as that of the reference population.
TABLE 1. Height census data from Central America and Panama
|Year||Illiteracy rate (1980)||No. of primary schools||No. of children||Height retardation (%)|
Besides showing the differences between countries, the data collected in the height censuses allow the identification of geographic areas within each country where height retardation is the greatest. Figure 2 (see FIG 2. Geographic distribution of height retardation in first-grade schoolchildren in Central America and Panama. The darkest shaded areas represent the zones with the greatest retardation.) identifies the areas in the highest quartile of height retardation for each country in the Central American isthmus. The information shown for Guatemala in figure 3 (see FIG. 3. Height retardation in first-grade schoolchildren in Guatemala, by departments (1986). Low risk = 24.1%-25.9% retardation; moderate = 26.0%-35.9%; high = 36.0%-50.9%; very high = 51.0%-64.6%) provides a good example of the data aggregated at the departmental level. In addition, census data are useful for identifying smaller geographic areas such as counties and districts within departments and provinces. This helps to detect variation between geographic areas where the average proportion of retardation is extreme.
The data on height retardation reflect the variation found between the countries in other socioeconomic and biological indicators, such as illiteracy, infant and child mortality, and social-services coverage. Furthermore. the high-risk geographic areas identified within countries coincide with those identified by other indicators that are more complex and expensive to collect, such as specific surveys and population censuses.
The changes that have occurred in the prevalence of height retardation in Costa Rica are a valid indicator of improvements in quality of life and development in that country. On the other hand, the recent increase in the proportion of children with height retardation in Panama could be a reflection of the present socio-political crisis and the internal rural-to-urban migration of the population that has been taking place in recent years.
Because of the simplicity and low cost of the collection and analysis of data from height censuses of first-grade schoolchildren and their reliability and validity, as confirmed by the studies carried out by INCAP, it is recommended that the use of such censuses for planning activities, monitoring and evaluating interventions, and advocacy be extended to other countries and regions.
Contents - Previous - Next