- Top of page
Objective To investigate myo-inositol, glucose and zinc status in mothers and their infants on cleft lip with or without cleft palate risk (CLP).
Design Case–control study.
Setting University Medical Centre Nijmegen, the Netherlands.
Population Eighty-four mothers and their CLP child and 102 mothers and their healthy child.
Methods Venous blood samples were obtained to determine serum myo-inositol and glucose and red blood cell zinc concentrations in mothers and children. Geometric means were calculated and compared between the groups. The blood parameters were dichotomised with cutoff points based on control values, <P10 for myo-inositol and zinc concentrations and >P90 for glucose concentrations.
Main outcome measures Geometric means (P5–P95) and odds ratios (95% confidence intervals).
Results The CLP children (P= 0.003) and their mothers (P= 0.02) had significantly lower red blood cell zinc concentrations than controls. A low maternal serum myo-inositol concentration (<13.5 μmol/L) and a low red blood cell zinc concentration (<189 μmol/L) increased CLP risk [odds ratio 3.0 (95% CI 1.2–7.4) and 2.0 (95% CI 0.8–4.8), respectively]. Children with low myo-inositol (<21.5 μmol/L ) or low red blood cell zinc concentrations (<118 μmol/L) were more likely to have CLP [odds ratio 3.4 (95% CI 1.3–8.6) and 3.3 (95% CI 1.3–8.0), respectively]. Glucose was not a risk factor for CLP in mothers and children. Maternal and child myo-inositol as well as zinc concentrations were slightly, albeit significantly correlated, rPearson= 0.33 (P= 0.0006) and rPearson= 0.23 (P= 0.01), respectively.
Conclusion This study demonstrates for the first time that zinc and myo-inositol are important in the aetiology of CLP.
- Top of page
Non-syndromic cleft lip with or without palate (CLP) are common birth defects that arise between the 7th and 14th week of pregnancy and are thought to be partially caused by interactions between genetic and environmental factors.1 Nutrition is considered to be important in the aetiology of these midline defects, as a preventive effect of multivitamins containing folic acid has been suggested.2 We recently showed a preventive effect against CLP of increasing food folate intake in the periconceptional period.3 Evidence is cumulating that deficiencies of myo-inositol and zinc are involved in reproductive failures, in particular, in midline defects as neural tube defects.4–7 Non-syndromic CLP and neural tube defects both originate from neural crest cells. Therefore, we hypothesised that myo-inositol, glucose and zinc status may be involved in the development of CLP as well.
Myo-inositol is the dominant form of inositol, a hexa-hydroxycyclohexane sugar alcohol isomer. Humans acquire myo-inositol from a variety of dietary sources of which vegetables, fruits and grains are the main contributors. In the body, myo-inositol is available as free forms and as inositol 1,4,5 triphosphate that is endogenously synthesised from glucose by myo-inositol 1-phosphate synthase. Free myo-inositol is involved in cellular osmoregulation. Its phosphorylated forms (i.e. the phospholipids) are important for the anchoring of proteins to the cellular membrane, modulation of enzyme activity, hormone secretion and intracellular signal transduction processes. Besides, phospholipids are a source of stearic and arachidonic acids for eicosanoid synthesis.8,9 The biological functions of myo-inositol underline its essential role for human growth, fertility, fetal development and pregnancy outcome.4
The studies of Grove et al.10,11 performed in the 1980s first suggest a potential role of myo-inositol in the pathogenesis of orofacial clefts. They demonstrate that glucocorticoids inhibit the proliferation of cells by altering the myo-inositol metabolism in fibroblastic cells derived from a human embryonic palate.
Diabetic pregnant women are at increased risk for having offspring with neural tube defects and orofacial clefts.12 Besides the induction of oxidative stress and its influence on gene expression, hyperglycaemia could possibly induce teratogenic effects through deficient myo-inositol tissue concentrations as glucose interferes with myo-inositol cellular uptake.13–15 This is supported by findings of Greene and Copp6 showing that dietary supplementation of myo-inositol reduces the occurrence of neural tube defects in the offspring of diabetic Curly tail mice.
An inadequate zinc status has been postulated to be teratogenic in both animal and human studies.16–18 Zinc is a cofactor for several metalloenzymes and a constituent of proteins, hormones and neuropeptides. It is crucial for embryonic development as cellular multiplication, differentiation and apoptosis as well as for the integrity of cellular membranes.18 Zinc finger proteins are important in their control of genes involved in embryonic development. Mice mutant in specific zinc finger proteins exhibit several congenital malformations including cleft palates.19 In addition, zinc is of interest because of its role in the absorption of natural folate (e.g. polyglutamates), its involvement in the conversion of 5-methyltetrahydrofolate into tetrahydrofolate by the zinc-dependent methionine synthase enzyme.20 Lower serum zinc concentrations have been reported in mothers of neural tube affected offspring.16,17 The data of Shaw et al.2 point to a possible preventive effect of periconceptional multivitamin use containing zinc on CLP risk. Zinc status is often compromised in diabetics and could be another explanation for the higher prevalence of CLP in these patients.
From this background, we hypothesise that a compromised supply to embryonic tissues of myo-inositol and zinc and excessive exposure to glucose could play a role in the pathogenesis of non-syndromic CLP. Therefore, the aim of this study is to investigate associations between the blood concentrations of myo-inositol, glucose and zinc in mothers and their children with and without a non-syndromic CLP.
- Top of page
We performed a case–control study of Dutch Caucasians in the Netherlands in the period 1998–2001. The study protocol has been extensively described previously.2 We recruited 177 children at around 14 months of age with CLP and their mothers from the registration of the Dutch Association for Cleft Palate and Craniofacial Anomalies in collaboration with the nine largest Cleft Palate Teams in the Netherlands. In each centre, a clinical geneticist and/or paediatrician diagnosed the malformation of the children and recorded the clinical features on a standard registration form.21 Two hundred and nine unrelated children of similar age, without major congenital abnormalities and without a familiar relationship with the cases, and their mothers were recruited from the domain population in collaboration with nurseries (25.5%) and by the case parents (74.5%). We selected all case and control pairs of which blood samples were available (n= 84 CLP and n= 102 control pairs). After exclusion of pregnant, lactating and non-fasting mothers at the time of blood sampling, mothers who reported a change in diet compared with the preconceptional period, and one control mother with gestational diabetes, 68 CLP mothers, 82 control mothers, 84 CLP children and 102 control children remained for analysis. The standardised collection of questionnaire data and blood sampling was performed 24 months after the conception of the index pregnancy. All mothers filled out a general questionnaire providing information about the index pregnancy, current and periconceptional lifestyle factors and demographics. Blood samples were taken for the determination of myo-inositol, glucose and zinc concentrations in the children and in their mothers.
The preconceptional intake of dietary folate was evaluated as confounder. Nutritional intake was assessed using the validated food frequency questionnaire (FFQ) developed for the Dutch cohorts of the European Prospective Investigation into Cancer and Nutrition Study (EPIC).3 The FFQ accounted for at least 90% of the population mean intake of food groups and nutrients of interest.22,23 Average daily folate intake was estimated by multiplying the frequency of consumption of the food items by the portion size and the nutrient content per gram. Total energy intake was calculated by using the 1996 Dutch food composition table; folate intake was calculated using that of 2001.24,25 The energy-independent residuals of this analysis were standardised to the predicted nutrients intake at the average energy intake (8926 kJ/day) in our population.26
Our considerations to choose 24 months after conception of the index pregnancy as study moment were that the hormonal and metabolic state of the mother had returned to the non-pregnant state.3 Moreover, in general nutritional habits are rather constant and are only influenced by episodes of illnesses, dieting and increased needs due to pregnancy and breastfeeding. By taking these determinants into account, we assume that the myo-inositol, glucose and zinc status preconceptionally to the index pregnancy approaches the status 14 months after the delivery of the child. The assumption of rather constant nutritional habits in women around pregnancy has been supported by others. Devine et al.27 interviewed 36 women on dietary habits from pregnancy through the postpartum period and concluded that in general no changes occurred from the beginning of pregnancy and one year postpartum. Leck et al.28 demonstrated that red blood cell folate concentrations in early pregnancy were well correlated with those one year after delivery. Quirk and Bleasdale29 substantiated this hypothesis for myo-inositol as these concentrations were comparable in pregnant and non-pregnant women. An additional argument was that at the age of 14 months in most cases the diagnosis of the child with non-syndromic CLP is completed.
The data obtained for confounder adjustment were current use of nutritional supplements, the use of medication, oral contraceptives, cigarettes, alcohol and the presence of illnesses. The current use of vitamins was defined by the use of any nutritional supplements. Periconceptional use of folic acid supplements was defined as daily intake from four weeks before through eight weeks after conception. Incidental users and women who started to use folic acid supplements later than four weeks before conception were categorised as ‘non-users’. Current illnesses were considered to be present when women reported acute or chronic illnesses at the time of blood sampling. Women were considered to be smokers and to consume alcohol when any smoking or alcohol consumption was reported. Educational level was categorised into low education (primary/lower vocational/intermediate secondary/intermediate vocational education) and high education (higher secondary/higher vocational or university education).
The Medical Ethical Committees of all participating hospitals approved the study protocol and written consent was obtained from all participants.
All mothers and children provided a venous blood sample at around 14 months after the delivery of the index pregnancy to measure the concentrations of serum myo-inositol and glucose and red blood cell zinc concentrations in whole blood. The blood samples of the mothers were fastened. The mothers were asked not to add sugar to the meals of their children prior to blood sampling.
Myo-inositol and glucose concentrations were determined by Tri-sil-TBT-esterification and gas chromatography.30 A volume of 500 μL of serum was pipetted into an Eppendorf tube and 100 nmol of mannoheptulose and 100 nmol of trehalose were added as internal standards. Then, 25 μL of 8 M perchloric acid was added. After the tube was put on ice for 10 minutes, the samples were centrifuged for 5 minutes at maximum speed in an Eppendorf centrifuge (model 5414). The supernatants were transferred into new Eppendorf tubes and 50 μL of 4 M K2HPO4 was added. After 10 minutes on ice, the samples were again centrifuged for 5 minutes at maximum speed. The supernatants were transferred into glass tubes and frozen until dry overnight (Lyovac GT II). To the residues 500 μL of Tri-sil-TBT was added and the mixtures were heated at 100°C for 30 minutes (heating block by Marius Instrumenten, Utrecht, The Netherlands). One millilitre of aquadest and 1 mL of hexane were added to the solutions. The mixtures were then shaken for 5 minutes (Gyrotory Shaker Model G2) and centrifuged for 5 minutes at 1000 ×g. The supernatants were transferred to clean glass tubes and 1 mL of 0.1 M HCl was added, again shaken for 5 minutes and centrifuged at 1000 ×g for 10 minutes. The supernatants were transferred to new tubes and 200 mg of Na2SO4 was added. Samples were left at room temperature for 15 minutes and centrifuged at 1000 ×g for 10 minutes. The supernatants were decanted into new tubes and one drop of bis(trimethylsilyl)trifluoracetamide was added, mixed and centrifuged at 1000 ×g for 10 minutes. The supernatants were then decanted into clean glass tubes and evaporated under a stream of nitrogen. Finally, the residues were dissolved in 100 μL of hexane and analysed on a gas chromatograph (Hewlett Packard 5890 series II). The detection limit of myo-inositol was ±0.5 μmol/L. The intra-assay coefficient of variation for myo-inositol and glucose was 10.4% and 4.3%, respectively.
Red blood cell zinc concentrations were determined by flame atomic absorption spectrophotometry (AAS) (Perkin Elmer 4100; Norwalk, Connecticut, USA). To 1 mL of venous blood 9 mL of a physiological salt solution was added and centrifuged for 10 minutes at 400 ×g at 4°C in a Hettich Rotanta/RP. The supernatant was removed, the cell pellet resuspended into 10 mL of saline and centrifuged for 10 minutes at 400 ×g at 4°C. After the supernatant was removed, the red blood cells were lysed by adding 2 mL of MilliQ. The lysed red blood cell samples were stored at −20°C until assayed. The intra- and inter-assay coefficient of variation for zinc were 3.4% and 3.5%, respectively.
No biochemical data were available for three mothers: two cases and one control mother. Due to parental refusal to take a venous sample from their child or as a consequence of a limited volume of the blood sample, from 46 blood samples we missed the myo-inositol concentrations (10 CLP and 36 controls), of 46 blood samples the glucose concentrations (10 CLP and 36 controls) and of 45 blood samples zinc concentrations (15 CLP and 30 controls) were lacking. Due to haemolysis, we excluded the myo-inositol and glucose concentrations derived of five control mothers and children and two CLP children from analyses. Thus, the study group comprised 66 CLP and 81 control mothers and 76 CLP and 77 control children.
Comparability of case mothers and controls on educational level, current illnesses, the use of vitamins, oral contraceptives, medication, smoking and alcohol consumption and periconceptional folic acid supplement use was performed using χ2 tests. When the cells had expected counts less than 5, the two-tailed Fisher's exact test was used. Maternal body mass index was positively skewed for which we performed a logarithmic transformation and presented the results as geometric means (P5–P95). Comparability of maternal and child age at blood sampling and maternal body mass index was tested by a Student's t test.
After logarithmic transformation, the biochemical values of myo-inositol, glucose and zinc in mothers and children were presented as geometric means (P5–P95) and compared between cases and controls. Differences in serum myo-inositol, glucose and red blood cell zinc concentrations between mothers of CLP children and controls were evaluated using Student's t test. Because of the multifactorial aetiology of CLP, we assume that only subsets of mothers and children are at risk for CLP. Therefore, serum concentrations of myo-inositol, glucose and red blood cell zinc in mothers and infants were dichotomised with cutoff points of the lowest 10% (P10) based on the control values for myo-inositol and zinc and the highest 10% (P90) for glucose concentrations. Odds ratios (ORs) with 95% confidence intervals (95% CI) were calculated to estimate the risk of CLP offspring. Logistic regression models were fitted to the continuous and dichotomised data, respectively, to adjust for potential confounders. Furthermore, as periconceptional vitamin supplement use and especially folic acid is suggested to affect the occurrence of CLP, we performed a stratified analysis for maternal periconceptional vitamin use containing folic acid. Pearson correlation coefficients within mother and child biochemical values were calculated to estimate the dependency of the log-transformed blood concentrations. Significance was defined as a P value equal to or less than 0.05. All analyses were performed using SAS Statistical Analysis System version 6.12 (SAS Institute, Cary, North Carolina).
- Top of page
The characteristics of the mothers and children at the moment of investigation, 24 months after conception of the index pregnancy, are presented in Table 1. Educational level and body mass index were significantly lower and significantly higher, respectively, in case mothers compared with controls. Alcohol consumption was more frequent in control mothers (P < 0.05). CLP mothers used significantly more folic acid supplements periconceptionally compared with controls. CLP children were more frequently males and older than controls (P < 0.05). Approximately 50% of the children in each group used vitamins A and D; only one CLP child used multivitamins.
Table 1. Characteristics of mothers of orofacial cleft children and controls and their children.
| ||CLP group||Control group|
|Mothers||n= 66||n= 81|
|Age at blood sampling (years), mean (SD)||32.2 (4.2)||32.9 (3.5)|
|Educational level, (% low)*||54.6||32.1**|
|BMI (kg/m2), geometric mean (P5–P95)||24.7 (20.2–34.3)||23.4 (19.1–30.0)***|
| Vitamin use†||13.6||17.3|
| Containing zinc||7.6||1.2|
| Oral contraceptive use||54.6||46.9|
| Medication use||27.3||22.2|
|Periconceptional folic acid supplement use†† (%)||29.7||46.2***|
|Children||n= 76||n= 77|
|Age at time blood sampling (months) mean (SD)||14.9 (1.9)||13.5 (2.9)***|
|Gender (% boy)||68.4||45.5**|
| Vitamin use†||50.0||55.8|
| Medication use||26.3||18.2|
The geometric mean concentrations of myo-inositol and glucose were not significantly different in case mothers compared with control mothers (Table 2). This is in contrast to the significantly lower geometrical mean zinc concentrations in case mothers compared with controls (P= 0.02). The serum myo-inositol and glucose concentrations in CLP children were comparable to the control values. Red blood cell zinc concentrations, however, were significantly lower in CLP children compared with controls (P= 0.003). Mothers had significantly lower myo-inositol, higher glucose and higher zinc concentrations compared with their child (P= 0.0001).
Table 2. Concentrations of myo-inositol, glucose and zinc in case and control mothers and their CLP children and in control mothers and children.
| ||n||CLP group, geometric mean (P5–P95)||n||Control group, geometric mean (P5–P95)||P|
|Myo-inositol (μmol/L)||66||16.0 (11.4–23.6)||76||16.4 (11.9–23.3)||0.50|
|Glucose (mmol/L)||66||4.1 (3.7–4.4)||76||4.1 (3.7–4.7)||0.79|
|Zinc (μmol/L)||66||207 (174–247)||81||218 (169–262)||0.02|
|Myo-inositol (μmol/L)||72||25.7 (18.2–37.8)||61||27.1 (20.4–35.0)||0.20|
|Glucose (mmol/L)||72||3.8 (3.1–4.7)||61||3.7 (3.1–4.4)||0.77|
|Zinc (μmol/L)||69||132 (97–173)||72||145 (114–206)||0.003|
Due to the multifactorial aetiology of CLP, we hypothesised that a subset of CLP mothers and children have a compromised myo-inositol, glucose and zinc status. Therefore, we defined subsets of mothers with a high risk for CLP offspring by the lowest 10% cutoff based on the controls for myo-inositol and zinc and by the highest 90% values for glucose. Table 3 revealed that a maternal serum myo-inositol concentration <13.5 μmol/L and a red blood cell zinc concentration <189 μmol/L are associated with an increased risk for having an CLP child [OR (95% CI): 3.0 (1.2–7.4) and 3.4 (1.3–8.6), respectively]. The separate associations of myo-inositol and zinc with CLP did not change after adjusting for each other and for glucose, suggesting an independent effect of the two blood parameters. Adjustment for maternal education and body mass index altered the results only marginally [OR (95% CI): low myo-inositol 2.7 (1.1–7.0), high glucose 0.3 (0.06–1.2) and low zinc 1.7 (0.6–4.4)]. Separate adjustment for maternal age, current use of vitamins or oral contraceptives, medication, alcohol consumption and smoking at the moment of investigation did not significantly alter the associations, nor did adjustment for periconceptional folic acid supplement use (data not shown). Adjustment for energy-adjusted dietary folate intake slightly lowered the risk estimate for low myo-inositol and CLP risk [OR 2.4 (95% CI 0.9–6.1)]. No mothers had low myo-inositol, high glucose and low zinc concentrations simultaneously and the presence of combinations of low myo-inositol/high glucose (one CLP mother), low zinc/high glucose (one CLP mother), low myo-inositol/low zinc (two CLP mothers and one control mother) did not differ between CLP mothers and controls.
Table 3. Risk for CLP associated with low myo-inositol, high glucose or low zinc concentrations in mothers and children (mother: zinc cases = 66, controls = 81; children: zinc cases = 69, controls = 72).
| ||Cutoff value*||No. of cases||No. of controls||OR||95% CI|
|Mothers|| ||66||76|| || |
|Myo-inositol, serum||<13.5 μmol/L||17||8||3.0||1.2–7.4|
|Glucose, serum||>4.5 mmol/L||3||7||0.5||0.1–1.9|
|Zinc, red blood cells||<189 μmol/L||13||9||2.0||0.8–4.9|
|Children|| ||72||61|| || |
|Myo-inositol, serum||<21.5 μmol/L||22||7||3.4||1.3–8.6|
|Glucose, serum||>4.2 mmol/L||11||6||1.7||0.6–4.8|
|Zinc, red blood cells||<118 μmol/L||20||8||3.3||1.3–8.0|
Children with a serum myo-inositol and red blood cell zinc concentration below 21.5 and 118 μmol/L, respectively, were more likely to have CLP, low myo-inositol [OR 3.4 (95% CI 1.3–8.6)] and low zinc [OR 3.3 (95% CI 1.3–8.0)]. A glucose concentration above 4.2 mmol/L in children was not a risk factor for CLP [OR 1.7 (95% CI 0.6–4.8)]. Adjustment for the age of the child at the time of study did not alter the conclusions. No children had low myo-inositol, high glucose and low zinc concentrations simultaneously, one control child has high glucose concentrations and low zinc concentrations, three CLP children and one control had high glucose concentrations and low myo-inositol concentrations and only the simultaneous presence of both low myo-inositol/low zinc (seven CLP children) differed significantly between CLP children and controls (Fisher's exact test, two-tailed P= 0.006).
The chosen cutoff point for glucose might be too low considering our data. Therefore, we repeated the analyses with a cutoff point based on the highest 98% of the control values. A glucose concentration above 5.1 mmol/L in mothers did not reveal an increased risk for CLP [OR 1.2 (95% CI 0.07–18.8)]. Children with a high glucose (≥4.9 mmol/L) concentration were more likely to have a CLP although the risk estimate was not significant [OR 2.6 (95% CI 0.3–25.8)]. Adjustment for current vitamin use by the child did not alter the risk estimates [low myo-inositol OR 3.5 (95% CI 1.4–8.9), high glucose OR 1.8 (95% CI 0.6–5.1) and low zinc OR 3.4 (95% CI 1.4–8.5)].
The aetiology of non-syndromic CLP is multifactorial. Therefore, we evaluated whether a low myo-inositol or a low zinc status in both mother and child increased CLP risk even further. There was no significant interaction between low maternal and child myo-inositol concentrations on CLP risk. The same conclusions could be drawn for a low maternal and child zinc status.
As periconceptional folic acid supplement use was significantly more frequent in controls, we performed a stratified analyses for periconceptional folic acid supplement use. The risk estimates for low myo-inositol and zinc concentrations and high glucose concentrations and CLP were not significantly different according to folic acid supplement use periconceptionally. Nevertheless, a maternal glucose concentration above 4.5 mmol/L revealed an odds ratio of 1.8 (95% CI 0.2–20.3) in non-supplement users compared with 0.3 (95% CI 0.03–2.3) in supplement users (Table 4).
Table 4. Risk for CLP associated with low myo-inositol, high glucose or low zinc concentrations in mothers by periconceptional supplement use.
| ||CLP/controls (n/n)||Supplement users (n= 55)||CLP/controls (n/n)||Non-supplement users (n= 87)|
|Myo-inositol, serum (<13.5 μmol/L)||3/2||3.0 (0.5–19.8)||14/6||2.5 (0.8–7.3)|
|Glucose, serum (>4.5 mmol/L)||1/6||0.3 (0.03–2.3)||2/1||1.8 (0.2–20.3)|
|Zinc, red blood cells (<189 μmol/L)||4/4||2.1 (0.5–9.7)||8/5||1.6 (0.5–5.3)|
Correlation analyses of the maternal and child values revealed significant correlations between maternal and child myo-inositol (rPearson= 0.33, P= 0.0006) and maternal and child zinc concentrations (rPearson= 0.23, P= 0.01).
- Top of page
This study reveals for the first time that mothers of a CLP child and their CLP children had significantly lower red blood cell zinc concentrations compared with controls. These results could not be explained by higher intake of zinc containing vitamin supplements by the control group. A compromised zinc status possibly due to altered metabolism or increased clearance in mothers and their CLP child is suggested. This is in line with reports on neural tube defects, but has not yet been demonstrated before for CLP.
To our knowledge, our study is the first report of a trend towards lower myo-inositol concentrations in CLP mothers and their children compared with controls, which increased the risk for CLP offspring and is confirming research in vitro.10,11 An increased glucose concentration was not associated with CLP risk in this study, not supporting previous findings that maternal glucose levels are involved in the pathogenesis of CLP.12 This is consistent with our lack of evidence for an inhibitory effect of glucose on cellular myo-inositol uptake and no interaction was observed between maternal serum myo-inositol and glucose concentrations. Such an effect might however occur intracellularly and would be undetectable in our serum sample. It should be mentioned though that none of the participants were diabetics or had hyperglycaemic values for glucose, which questions the power of this study to provide information regarding the role of glucose in the formation of orofacial facial clefts.
Mothers had significantly lower myo-inositol, higher glucose and zinc concentrations compared with their children (P= 0.0001). Quirke and Bleasdale29 found higher serum myo-inositol concentrations in newborns compared with maternal serum values, which would originate from an increased endogenous synthesis from glucose. Zinc is not endogenously synthesised by the body and humans acquire zinc only through their diet. The consumption of zinc (mg/kJ) and glucose (mono- and disacharides or carbohydrates, g/kJ) was comparable in children and mothers according to the Dutch food consumption survey of 1997–1998.31 The higher myo-inositol, lower glucose and zinc status could thus be a reflection of increased requirements of the child compared with adults.
Of interest is the correlation between myo-inositol and zinc concentrations, both in mothers and children, possibly indicating a common genetic trait. Although the results of this study on mothers are rather consistent with those in children, the latter are less precise due to a rather large amount of missing values for the children, either due to parental refusal or to the limited sample size.
The lower blood concentrations of myo-inositol and zinc could be a reflection of a poorer dietary intake of myo-inositol and zinc sources such as vegetables, fruits and meats. In our study population, case mothers had a lower educational level compared with the controls and it has been reported that individuals with a lower level of education are more likely to consume less healthy diets compared with their counterparts.32 Nevertheless, adjustment for maternal education did not significantly alter the risk estimates. Other lifestyle or genetic factors might also be involved.
We collected our data 24 months after the conception of the index pregnancy with the assumption that the nutritional status at that moment is the best reflection of the preconceptional period of the index pregnancy. Previous research confirmed this hypotheses and was substantiated by the fact that adjustment for maternal and child illnesses at the time of investigation did not alter our results.27,28 Furthermore, in our study, zinc status was determined in red blood cell, known to reflect the zinc status over a longer period and to be less sensitive to external influence.
Due to the lack of adequate clinical references for the blood concentration of myo-inositol and zinc, we dichotomised the values for myo-inositol and zinc with cutoff points of the lowest 10% (P10) based on the control values. This cutoff point was chosen as it demonstrated to be the optimal cutoff point compared with analyses at higher cutoff points. The choice of a lower cutoff point was rejected due to the small number of cases and controls having such low blood concentrations and due to the explorative feature of our study.
We conclude that maternal zinc and myo-inositol are associated with the risk for having a child with CLP, although the exact mechanisms remain unclear. Larger studies need to test our hypothesis of myo-inositol and zinc in the pathogenesis of CLP before our results are clinically applicable.
- Top of page
The authors would like to thank the members of the participating Cleft Palate teams and their coordinators (Rijnstate Hospital Arnhem, Dr W. Brussel; Free University Hospital Amsterdam and Sophia Children's Hospital/Erasmus University Medical Centre Rotterdam, Professor Dr B. Prahl-Andersen; University Hospital Groningen, Professor Dr S. M. Goorhuis-Brouwer; Medical Centre Leeuwarden, Dr J. J. van der Biezen; University Medical Centre Nijmegen, Professor Dr A. M. Kuijpers-Jagtman; St Elisabeth Hospital Tilburg, Dr J. G. Daggers; University Medical Centre Utrecht, TFJMC Specken, Dr A. B. Mink van der Molen and Sophia Hospital Zwolle, P. Houpt). The authors acknowledge the nurseries and infant welfare centres for their participation in the recruitment of the subjects and are very grateful for the assistance of M. van der Doelen and Dr C. van Oostrom, chief nurse officer and head of the outpatient paediatric clinic of the University Medical Centre Nijmegen, respectively, for their assistance, The authors thank Professor Dr Chr. Vermeij-Keers, Department of Plastic and Reconstructive Surgery, Erasmus University Medical Centre Rotterdam for the opportunity to use the registration system of the Dutch Association for Cleft Palate and Craniofacial Anomalies. Furthermore, the authors thank Mrs F. Janssen for her work on the determination of the myo-inositol concentrations, Mrs A. Pellegrino, Mrs M. van Drongelen, Mr P. Groenen and Mr L. Lemmens for data entry, P. Groenen and N. Peer for their help with the statistical analyses and Mr W. Lemmens for data management.
This work was funded by the Royal Netherlands Academy of Arts and Sciences (KNAW), Amsterdam, The Netherlands, 1997–2003.