Main

CAD is a major cause of morbidity and premature mortality in Singapore and some Western cultures. The precedent event of CAD is coronary atherosclerosis and atherogenesis is closely associated with lipoprotein metabolism. Clinical studies have demonstrated a close association between lipoprotein abnormalities and susceptibility to CAD. A high level of Lp(a) has been recognized to be an independent risk factor for atherogenesis(1–4). Lp(a) is present in all individuals but in varying degrees. The level of Lp(a) is under strict genetic control, and its plasma concentration is determined significantly by inheritance(5–7). The level of plasma Lp(a) is largely determined by alleles at the apo(a) locus which accounts for greater than 90% of its variation(7). Apo(a) was found to be a significant predictor of CAD, independent of total cholesterol and HDL-cholesterol but not independent of Lp(a) levels(8).

Ethnic dependence of Lp(a) levels has been found in different populations(9). Relevant to the Singapore scene are the findings in adults of high levels of Lp(a) in ethnic Asian Indians and low levels among the ethnic Chinese of Singapore(8, 9). Of 7 ethnic groups studied by Sandholzer et al.(8, 9), the Chinese and Indians of Singapore were found to have Lp(a) levels at the two extremes of the spectrum in distribution. It is important to note that the Indians of Singapore are Asian Indians.

The prevalence rates of CAD among the three major ethnic groups of multiracial Singapore differ, with highest mortality seen in ethnic Indians and lowest in ethnic Chinese(10, 11). Interestingly, their relative coronary risk levels are concordant with the relative pattern of distribution of plasma Lp(a) in these population groups(12).

As the concentration of Lp(a) in plasma is genetically determined through the apo(a) gene, it has become in recent years an interesting parameter to study in the newborn. Studies in adult Caucasian populations have shown that the distribution of Lp(a) concentrations is highly skewed and that the concentrations are highly variable among individuals. At birth it is present in very low levels. The concentration and distribution pattern of Lp(a) at birth and the subsequent evolution and relationship of cord Lp(a) levels to the parent's profile have recently been reported(13, 14).

The Lp(a) profile of newborns of the three ethnic groups of Singapore is therefore a point of great interest in our investigation of the variation of this highly genetically charged biochemical marker of CAD within populations which show a diversity of CAD mortality among the different ethnic groups.

It has been recognized that there is ethnic variation in the heritability of Lp(a) levels. Variability at the apo(a) locus was observed to account for 77, 70, and 33% of variation of Lp(a) levels in the ethnic Malays, Chinese, and Indians of Singapore(9). In this study, we report the results of the Lp(a) profile of 542 male and 468 female newborns from these three ethnic groups of Singapore. Their Lp(a) levels in the cord plasma are then related to the coronary risk levels of their respective adult populations.

METHODS

Clinical data collection. Normal newborns of pure Chinese, Malay, and Indian heritage born at the National University Hospital, Singapore, were recruited into the study. Participating babies had no history of mixed heritage in the preceding three generations. There were 1059 samples collected over a period of 8 mo from August 1993 to March 1994. Information regarding any antenatal events or complications and maternal health problems was extracted from maternal obsteric records. The birth order, gestational age, mode of delivery, birth weight and length, sex, race, the 1- and 5-min Apgar scores of the babies, and any intrapartum and immediate postnatal problems were noted. Mothers along with fathers were interviewed on the first or second postdelivery day in the postnatal wards for family history of CAD, hypertension, and diabetes mellitus.

Exclusion criteria. Babies of mixed heritage and those with major congenital malformations were excluded from the study. As perinatal stress is found to have an influence on cord blood lipids, three perinatal deaths and all babies with 5-min Apgar scores of <7 were also excluded from the study. Babies of families who were unable to give a clear history of CAD in either of the four grandparents of the infants were excluded from analysis of the parameters with relation to family history. A total of 1010 newborns comprising 542 male infants (158 Chinese, 128 Malays, and 256 Indians) and 468 female infants (174 Chinese, 107 Malays, and 187 Indians) were available for final analysis.

Cord blood collection. After delivery of the infant but before delivery of the placenta, the umbilical cord was divided between clamps. Five milliliters of free flowing umbilical cord blood were collected in heparinized tubes from the placental end of the cord.

Lp(a) estimation. Plasma Lp(a) quantification was performed by a sandwich-ELISA using commercially available kits (Biopool, Umea, Sweden). The results were read on an automated Bio-Rad 3550 reader. The coefficient of variance between assays was 10%, and the intraassay coefficient of variance was 2.8%.

Statistical analysis. Statistical analysis was performed using the Version 6 of the Statistical Package for the Social Sciences (SPSS) for Windows. The significance of differences for the means of plasma Lp(a) between the races was determined by the t test. ANOVA was performed to determine the effect of race, gestational age, and birth weight. The percentages of sample variance (R2 × 100) were calculated from sums of squares. The significance of the sample variance was tested by F and p values. As the frequency distributions of Lp(a) were highly skewed, parametric statistical calculations (t test and ANOVA) were performed on transformed data by taking the natural logarithm (ln) of their raw values. Because the values of Lp(a) in milligrams/dL in cord blood specimens were small with values below 1, and as the natural logarithm of <1 is negative, logarithmic transformations for these parameters were done after addition of 1 to the raw values. Nonparametric tests were also applied to the skewed distributions. The Mann-Whitney U-Wilcoxon rank sum test was also used to test the equality of medians of Lp(a), and the Kruskal-Wallis one-way ANOVA test was used to analyze variance by ranks of observations for Lp(a).

The quartiles in the distribution of the cord plasma levels of Lp(a) were determined. To determine whether there was a higher proportion of newborns with a family history of CAD with Lp(a) in the top quartile, χ2 analysis was done to compare the two proportions in the top quartile and that in the lower three quartiles.

RESULTS

Frequency Distributions of Cord Plasma Lp(a) Levels

The frequency distributions of plasma Lp(a) levels in cord blood in male and female infants was highly positively skewed (Fig. 1,left). The distribution pattern of plasma Lp(a) is characteristic of the Chinese and Malay adult populations of Singapore and several other Caucasian populations(9) and of Belgian newborns(13). Before comparison of the means between male and female infants by t test, the raw data were first transformed by natural logarithm (Fig. 1,right).

Figure 1
figure 1

(Left) Frequency distribution (%) of plasma Lp(a) levels (mg/dL) in cord blood. Very positively skewed distribution for both sexes (males 2.653, females 1.689). (Right) Frequency distribution after natural logarithmic transformation. Skewness: in male infants, 0.597, and in female infants, 0.388.

Full size image

Plasma Lp(a) Levels in Cord Blood

Table 1 shows the means and SD of Lp(a) for male and female infants of the three ethnic groups. Female infants as a group had higher levels of Lp(a) compared with male infants at birth (1.7 ± 1.4 mg/dL versus 1.5 ± 1.4 mg/dL), although this difference is not statistically significant. These levels are more akin to that found in Belgian babies(13). Both the present study and that of Van Biervliet estimated plasma Lp(a) levels in cord specimens by ELISA, and the results are therefore more comparable. The figures on Austrian newborns of 3.1 ± 2.5 mg/dL measured by electroimmunoassays were much higher than what is found here(15). The different method of estimation and the small sample size in Strobl et al.'s(15) study could have resulted in the difference in levels.

Table 1 The cord blood Lp (a) levels (mg/dL) in male and female newborns of the three different ethnic groups
Full size table

ANOVA by Race, Gestational Age, and Birth Weight of Cord Plasma Lp(a) Levels

The percentage of variance (R2 × 100) in ln [Lp(a)] levels attributed to race as a variable is shown in Table 2. The phenotypic variance of ln[Lp(a)] levels contributed by race in male and female infants was 1.5 and 4.3% (p < 0.02 and < 0.0001), respectively. The adjusted means showed that Indians had the highest levels, Chinese the lowest, and Malays had levels in between. This is the pattern for both male and female newborns. Gestational age and birth weight as covariates did not make any significant contribution to variance of Lp(a) levels for both sexes. This is in agreement with the finding by Wang et al.(16) that the levels of apo(a) which serve as a marker for quantitation of Lp(a) are independent of birth weight. As the frequency distribution of Lp(a) after natural logarithmic transformation was still skewed, ANOVA of Lp(a) levels was also done using nonparametric tests. The Kruskal-Wallis one-way ANOVA test revealed the same findings as in ANOVA after logarithmic transformation of Lp(a) (Table 3). The mean ranks of Lp(a) levels were highest in Indians followed by Malays and then Chinese for both sexes, and the differences in ranks were at the same level of significance as tested by ANOVA on ln[Lp(a)] values (Table 2).

Table 2 Percentage of sample variance (R2× 100) in ln [Lp (a)] explained by race, gestational age and birth weight in males and females of the three ethnic groups (ANOVA)
Full size table
Table 3 Rank correlation of Lp(a) levels by race in male and female newborns of the three ethnic groups (Kruskal-Wallis one-way ANOVA)
Full size table

Plasma Lp(a) Levels in Cord Blood in the Three Ethnic Groups

Comparison of means between the two sexes. Between the sexes, the means of Lp(a) after logarithmic transformation were higher in female than in male newborns for the whole cohort (0.9 ± 0.5 versus 0.8± 0.4) as well as for ethnic Malays and Indians(Table 1). The gender differences in means were, however, not statistically significant. This is the experience of a previous study of apo(a) in Caucasian babies(16). There has been no reports on cord blood levels of Lp(a) in Asian populations. There is, however, a significant difference in Lp(a) levels between male and female newborns and in adult Chinese populations of Singapore(17).

Comparison of the means in male and female newborns of different ethnic groups. The means and SD of cord Lp(a) levels among the male and female newborns of the three ethnic groups are graphically presented inFigure 2. Indian male and female newborns had the highest Lp(a) levels compared with their Malay and Chinese counterparts, whereas Chinese male and female newborns had the lowest Lp(a) levels at birth. The difference between Indian and Chinese male newborns and Indian and Chinese female newborns was statistically significant by t test for means of ln[Lp(a)] levels (p < 0.04 and < 0.002 for male and female newborns, respectively), as well as by the nonparametric Mann-WhitneyU- Wilcoxon rank sum test for equality of means of Lp(a) (p< 0.006 and < 0.0001 for male and female infants, respectively)(Tables 1 and 4).

Figure 2
figure 2

(Left) Means and SD of Lp(a) mg/dL in cord plasma of ethnic Chinese, Malay, and Indian babies. Significance of difference, Chinese vs Indian, male newborns. p < 0.04; female, p < 0.002. (Right) The age-standardized relative coronary risk of the Chinese, Malay, and Indian adult populations of Singapore [from Hughes et al.(11)].

Full size image
Table 4 Differences in medians of Lp(a) levels in cord blood among male and female newborns of the three ethnic groups (Mann-Whitney U-Wilcoxon rank sum test)
Full size table

In this study, family history of CAD in grandparents was found to have no significant association with plasma Lp(a) levels at birth.Table 5 shows the data on the comparison by χ2 analysis of proportions (%) with family history of CAD according to the quartile of cord plasma Lp(a) levels. The significantly higher proportion of Indian newborns with a family history of CAD in the lower 75th centile of distribution of plasma Lp(a) levels was contrary to what it should have been and was probably a chance finding.

Table 5 Comparison by χ2 analysis of proportions (%) with family history of CAD according to quartile of cord plasma Lp(a) levels in Chinese, Malay, and Indian newborns
Full size table

DISCUSSION

Lp(a) levels at birth are not affected by birth weight and sex of the babies. At birth, gestational age affected plasma Lp(a) to a very small degree. These results are consistent with what Wang et al.(16) had found for apo(a) concentrations.

Lp(a) levels rank highest in Indians, followed by Malays, and then Chinese. This trend was present in both male and female groups. The difference, however, was statistically significant only between Chinese and Indians for both sexes. The same difference has been observed in the healthy adult Indians and Chinese of Singapore(8). Elevated concentrations of Lp(a) in plasma are associated with premature CAD(4). The higher Lp(a) levels at birth in ethnic Asian Indians would predispose them to a higher risk for CAD. The cord plasma Lp(a) profiles of the three ethnic groups shown in Figure 2 are concordant with the age-standardized relative coronary risks of their respective adult populations(11).

The fact that the level of Lp(a) is under strict genetic control and that its plasma concentration is determined almost entirely by inheritance indicates that the ethnic characteristics of Lp(a) is inherent. The newborn Lp(a) profile of Indian and Chinese babies corresponded to their adult profile. This finding supports an earlier report that apo(a) levels in infants track closely and are predictive of parental levels(14). Certain physiologic and disease states including diabetes mellitus, liver disease, and end stage renal disease affect Lp(a) levels. Lp(a) levels are relatively stable to most dietary and drug interventions except to high doses of niacin and to estrogen administration and would be an invaluable and reliable marker to track from the newborn period. The expression of the ethnic Lp(a) profile at birth also supports the notion that the apo(a) gene for the essential protein component of Lp(a) is expressed at birth. It has been reported earlier in a Caucasian population that the gene for the regulation of apo(a) is fully expressed before the age of 1 y(14).

The results of this study showed that the cord plasma levels of Lp(a) are reflective of their adult Lp(a) levels according to ethnicity and coronary risk levels. The expression of their ethnic Lp(a) profile at birth would make this a very useful marker of later coronary risk from very early in life.