Abstract
purpose. To investigate maternal, demographic, and pre- and perinatal risk factors for idiopathic congenital/infantile (ICI) cataract.
methods. Based on national registries, a cohort of all children born in Denmark and aged 0 to 17 years during 1977 to 2001 was established, and congenital/infantile cataract cases were identified. Cases of unknown/idiopathic cause were included in the study. Associations between maternal, demographic, and pre- and perinatal factors with the development of cataract were investigated.
results. In a cohort of 2.9 million children, 1027 cases of congenital/infantile cataract were identified. Of the children in those cases, 629 were born in Denmark and had ICI. Bilateral isolated cataract cases were male dominated (62%; 95% confidence interval [CI], 56%–69%) but not unilateral isolated cases (40%; 95% CI, 34%–47%). Older age (≥40 years) of mothers at delivery and caesarean section increased the risk of ICI cataract. Low birth weight (<2000 g) was associated with a 10.6-fold (95% CI, 6.99–16.10) increased risk of bilateral, but not unilateral, ICI cataract. No significant associations were found with birth order, month/place of birth, or cigarette smoking during pregnancy.
conclusions. Variables indicative of environmental influence were not associated with ICI cataract. Low-birth-weight children (<2000 g) had a significantly increased risk of bilateral ICI, whereas no strong risk factors were found for unilateral cataract. Together with the sex difference, this suggests that the etiologies of bilateral and unilateral cataract are different.
Congenital (or infantile) cataract is a rare disease but a frequent cause of low vision or blindness among children in both developed
1 2 3 4 and developing
5 6 7 countries. The cause of most cases of congenital/infantile cataract is unknown.
8 9 10 11 12 13 14
Some of the cases of unknown cause may be genetically based, but infections during pregnancy and other environmental factors influencing fetal life are also suspected as underlying causes.
15 16 However, surprisingly little has been done to investigate the cause of idiopathic congenital/infantile cataract (ICI). There are only two available analytical studies of risk factors for ICI cataract.
13 17 They both showed that children of low birth weight had an increased risk of development of bilateral ICI cataract and one of the studies also showed older maternal age to be a risk factor.
In the present study, the long-existing national registers in Denmark were used to study maternal, demographic, and pre- and perinatal risk factors for ICI cataract based on the Danish population.
The present analysis is part of a broader project on childhood cataract in Denmark. The methods for cohort ascertainment are described in detail elsewhere.
18 Details pertinent to ICI cataract are provided herein.
Since 1968, Danish citizens have been assigned a personal 10-digit identification number by the Civil Registration System (CRS).
19 Attached to this CRS number is continuously updated information on vital status, emigration status, family relation, and place of birth and living. Other national registries also use the CRS number as their personal identification key, thus allowing the different registries to be linked. Based on information from the CRS registry, a cohort of all persons who were born in Denmark and who were between 0 and 17 years of age at some time during the period 1977 to 2001 was established (
n = 2,902,741).
Cases of ICI cataract within the cohort were ascertained from the Danish National Register of Patients (NRP). The NRP was established in 1977 and contains all discharge diagnoses for all patients admitted to public hospitals (inpatients) in Denmark. Since 1995, the registration has also included outpatients. From 1977 to 1993, the International Classification of Diseases version 8 (ICD-8) was used, and beginning in 1994, version 10 (ICD-10) was used. The diagnostic codes for congenital or infantile cataract were 74439, 37400, and 37401 in ICD-8 and Q12.0 and H26.0 in ICD-10. Information from the medical records retrieved from the Danish hospitals together with information from private ophthalmologists by linkage to the Danish Health Security System’s register (HSS) and from the Danish Family Archive for Hereditary Eye Disorders was obtained to classify the ICI cases as having an unknown cause, described in detail elsewhere.
14 Demographic data including information on multiple births were obtained from the CRS.
Information on pre- and perinatal factors of the children in the cohort was obtained from the Danish National Medical Birth Registry (MBR). Since 1968, midwives have been required to report all births in Denmark, including home births, to the MBR. This registry has been computerized since 1973.
20 Information on birth weight, gestational age, fetal length, the use of forceps, vacuum extractor, and caesarean section was available from 1973 to 2001. Other risk factors were registered in only parts of the period 1973 to 2001: Apgar score from 1978 to 2001; maternal smoking during pregnancy from 1991 to 2001; abnormal delivery presentation from 1997 to 2001; and maternal medical disease, toxemia, abruptio placentae, placenta previa, polyhydramnios and disproportion from 1978 to 1990 and from 1997 to 2001. The Apgar score is used to evaluate the newborn child and its need for resuscitative measures. It includes a score for activity (muscle tone), pulse, grimace (reflex irritability), appearance (skin color), and respiration.
21 To generate the variable birth weight percentiles by gestational age, the birth weight of each child in the cohort was used to estimate percentiles according to gestational age and sex.
In the primary study we categorized all types of cataract, and in the present study the focus was on cataract cases of unknown cause. These cases were divided into unilateral and bilateral cases and further into cases involving isolated cataract and those involving additional anomalies (including additional ocular dysmorphology and systemic anomalies), denoted “cataract plus.”
The study was approved by the Scientific-Ethical Committees for Copenhagen and Frederiksberg ([KF] 01-253/00) and permission to receive data from the national registries was obtained from the Danish Data Protection Agency (2000-41-0285). The protocol adhered to the tenets of the Declaration of Helsinki.
Confidence intervals for proportions were determined by log-linear binomial regression. The risk ratio of ICI cataract (overall and subtype specific) according to selected risk factors was estimated by log-linear binomial regression (PROC GENMOD; SAS, Cary, NC). All risk ratios were adjusted for sex and birth cohort (5-year groups). The adjustment for year of birth was primarily made to take into account the increase in the total number of registered cases due to the inclusion of outpatients in the Danish National Register of Patients after 1995. When further adjustment for other variables was made, the variables were categorized as shown in the tables. Not all cataract cases among children born before 1977 had been registered, as computerized registration of cataract was only initiated in 1977. To evaluate how this affects the estimation of risk ratios, we repeated all analyses in birth cohorts from 1977 and later. The estimates from these restricted analyses were similar to the estimates based on all birth cohorts. Only individuals with nonmissing observations of the included variables in each analysis were considered. Risk factor data were missing for no more than 7.5% of the total number of participants, considering the different registration periods of the risk factors. Differences in the case distribution of risk factors according to subtype were evaluated by a log-linear multinomial model (PROC CATMOD in SAS). Test for trend was evaluated by treating the variable as a continuous variable. Difference in risk according to month of birth was evaluated by describing the risk ratios by a sine function where both the amplitude and the peak risk were estimated.
22 Nonsignificant amplitude was taken as an indication of no difference in risk according to month of birth.
A total of 1093 children aged 0 to 17 years during 1977 to 2001 were registered with a diagnosis of congenital/infantile cataract in the Danish National Register of Patients (NRP). As described in detail elsewhere,
14 18 132 cases were excluded because of erroneous diagnosis, and another 46 cases because the children were born abroad (
n = 43) or were registered with cataract on an admittance date before 1977 (
n = 3). Thorough validation of cases recorded under other types of cataract revealed an additional 66 cases of congenital/infantile cataract. With these cases excluded, we were left with 981 cases of congenital/infantile cataract, of which 629 cases (321 boys and 308 girls) were of unknown cause and remained in the study.
The ICI cases were distributed into four groups according to laterality and clinical appearance, as (1) unilateral, isolated cataract; (2) unilateral, cataract plus; (3) bilateral, isolated cataract; and (4) bilateral, cataract plus. There were significantly more boys among bilateral isolated cataract cases 62% (95% CI, 56%–69%) and significantly fewer boys among the unilateral cases (40%; 95% CI, 34%–47%), whereas there was an equal sex distribution among cataract plus cases regardless of laterality.
The distribution of cases by month of birth is depicted in
Figure 1 . There was no statistically significant difference according to month of birth (
P = 0.84). Further analyses among subgroups of ICI cataract revealed a similar lack of association with specific months of birth (data not shown).
Table 1shows the risk ratios of the four groups of ICI cataract according to maternal and demographic factors. Among these risk factors, birth order was the only variable that differed significantly across the four subtypes. The case distribution according to birth order differed significantly for isolated cataract and cataract plus, but in neither group was there a clear or a significant association. Given that there was no clear, significant ordering or association, we have chosen to report all remaining analyses based on all ICI cataracts combined.
There was a significant difference in the risk of ICI cataract according to maternal age (P = 0.03). The difference was primarily due to a higher risk in children of older mothers (≥40 years of age; risk ratio [RR] = 2.27; 95% CI, 1.36–3.77) compared with children of mothers aged 25 to 29 years at delivery. Furthermore, mothers younger than 20 years carried a relatively higher risk, but it did not reach statistical significance (RR = 1.36; 95% CI, 0.95–1.94). The absolute risk of ICI cataract per 100,000 children with a mother aged ≥40 years was 40.6. There was no difference in the risk of ICI cataract in relation to whether a child was born in a rural or urban area (P = 0.96) or had a foreign mother (P = 0.94) or a foreign father (P = 0.96). However, although there was no significant differences in the effects of having foreign parents between the four subtypes, we find it interesting when looking at the subtype analyses that having a foreign mother was associated with an increased risk of bilateral cataract plus and that having a foreign father approached significance, whereas the association with unilateral cataract was in the opposite direction. Finally, there was no overall association between birth order and the risk of ICI cataract (P trend = 0.81).
Table 2presents the associations between the risk of ICI cataract and birth weight, gestational age, birth weight by gestational age, and fetal length. As for the variables in
Table 1 , the distributions of the variables in
Table 2for the four subtypes were initially compared. The distribution of all variables in
Table 2was significantly different in unilateral cases compared with bilateral cases, but not between isolated cataract and cataract plus cases. Thus, the associations are presented according to laterality. With respect to unilateral ICI cataract, we found no association with birth weight, gestational age, or birth weight percentile by gestational age. Fetal length was significantly associated with unilateral cataract. However, there was no clear trend.
In contrast to unilateral cataract, all variables in
Table 2were highly significantly associated with the risk of bilateral ICI cataract. The higher risk was mainly observed in children of low birth weight (<2500 g), small for gestational age (children with birth weights below the 10th percentile for gestational age), preterm (<37 weeks), and children of short fetal length (<48 cm). These factors are correlated by nature, and it is therefore difficult to distinguish their effects. Thus, collinearity excluded the possibility of a simultaneous evaluation of these variables in a multivariate analysis. However, stratified analyses suggested that the higher risk in children with low gestational age could be explained by the higher risk associated with having a low birth weight. For instance, the risk ratio for bilateral ICI cataract in children with a gestational age below 37 weeks was 7.13 (95% CI 3.49–14.6) if the birth weight was below 1500 g; 4.74 (95% CI, 2.24–10.0) if the birth weight was 1500 to 2000 g; and 1.02 (95% CI, 0.32–3.28) if the birth weight was 2000 to 2500 g (reference: term children of gestational ages 39–41 weeks, with a birth weight of 3000–3500 g). Similarly, the effect of birth weight could explain the effect of both fetal length and birth weight percentiles by gestational age. Adjustment for maternal age did not affect the risk ratios in
Table 2 .
Table 3presents the risk ratios of ICI cataract according to other pre- and perinatal variables that were registered only during part of the entire study period or had a rare positive outcome. Smoking during pregnancy was not associated with a higher risk of ICI cataract. Caesarean section was the only variable associated with a higher risk after adjustment for birth weight.
The significant associations from
Tables 1 2 and 3were evaluated in a multivariate analysis presented in
Table 4 . Birth weight represents the highly correlated variables in
Table 2 , as the results of the analyses suggest that the higher risk in small infants were best described by birth weight. In the multivariate analysis, birth weight and caesarean section were found to be significantly associated with bilateral CI cataract, whereas none of the three factors included were associated with unilateral CI cataract.
In a population-based cohort of 2.9 million children there were 629 cases of ICI cataract. We found children of older mothers to have a modestly increased risk of ICI cataract. However, most significant were our findings of differences in the risk factor profile between bilateral and unilateral ICI cataract. In particular, low birth weight was strongly associated with an increased risk of bilateral cataract, but not unilateral cataract. In children born with a birth weight below 2000 g, the risk was 10.6 times higher than in children with a birth weight of 3000 to 3500 g.
All medical records of congenital/infantile cataract cases that occurred in our cohort and were registered in the mandatory national hospital registry were reviewed to validate the diagnosis and to classify them according to presumed etiology. The hospital charts were supplemented with information from private ophthalmologists and from a hereditary eye disease register, and cases for which medical records contained inadequate information were excluded. This procedure limited problems of misclassification and increased the likelihood of detecting true associations.
The information on risk factors was free from recall bias, as it was obtained and recorded independently of and before the diagnosis of cataract. Data on maternal and demographic variables have been shown to be very complete, as they were extracted from the CRS. The information in the CRS is updated daily and used by the official authorities to keep track of Danish citizens. A validation study of the Danish National Medical Birth Registry (MBR) data has reported good agreement of these, compared with data from medical records. An exception is gestational age, which was generally registered as one week greater in the registry.
23 However, such a systematic overestimation of gestational age would be similar in all members of the study cohort, and it would not influence the variable birth weight percentile by gestational age.
Some studies have reported a higher occurrence of children with idiopathic cataract among low-birth-weight and/or preterm children,
24 25 26 27 28 29 30 but in other studies no such association was found.
31 32 33 34 35 36 37 Only two previous studies from the United States have analyzed risk factors for ICI cataract. The study by SanGiovanni et al.
17 was based on 32 isolated cases of cataract out of 73 cases ascertained during the first 7 years of life by ocular examination, though with uncertainty regarding the identification of all familial cataract cases. The study was based on the birth cohort years 1959 to 1965. Compared with 50,000 children without cataract the investigators found low birth weight (<2500 g) to be associated with a higher risk of isolated bilateral cataract. They did not perform risk factor analyses on nonisolated cataracts. The other study by Bhatti et al.
13 was based on 117 idiopathic (up to 1 year of age) out of 199 cataract cases ascertained from a birth-defect register. They did not distinguish between isolated cataract and cataract in association with minor defects/other ocular defects, and laterality could only be determined for 102 cases. Among this reduced group of cases they found very low birth weight (< 1500 g) and maternal age more than 35 years to be risk factors for bilateral cataracts.
Older maternal age is associated with an increased risk of complications during pregnancy and with low-birth-weight children, who are preterm and small-for-gestational age.
38 39 Furthermore, older maternal age increases the likelihood of caesarean section
39 40 and chromosomal abnormalities.
41 However, in the multivariate analysis, older maternal age was no longer associated with an increased risk.
Seasonal variation may be interpreted as a proxy for exposure to infections during pregnancy.
22 Similarly, birth order may be used as a proxy for infectious burden. A child of the low birth order is the lone child in the family and therefore less likely to have been exposed to infections during fetal life than a child of higher birth order.
42 We found no clear association with birth order or with month of birth. Thus, our results do not support that infections during pregnancy play a role in ICI cataract.
16 Furthermore, there was no difference in risk whether a child was born in a rural or an urban area, which further supports the notion that environmental factors do not play a significant role. These findings corroborate our previous report of a stable incidence of idiopathic cataract over the past two decades, when many changes in environmental exposures and infectious disease pressure occurred.
14
No specific ocular defect has been reported to be associated with cigarette smoking,
43 but it is well established that maternal cigarette smoking is associated with low birth weight.
44 45 46 We did not find maternal smoking during pregnancy to be a risk factor for ICI cataract.
Unilateral ICI cataract cases distinguished themselves clearly from bilateral cases. First, they differed by sex, as most unilateral isolated cataract cases were girls and most bilateral isolated cataract cases were boys. The sex difference is puzzling and so far, unexplained. Second, unilateral ICI cataracts also clearly differed regarding significance of birth weight. Children of low birth weight had a higher risk of bilateral, but not of unilateral, cataract.
The pronounced association between the risk of bilateral cataract and low birth weight may indicate that low birth weight per se is associated with a higher risk of bilateral congenital/infantile cataract or, more likely, that the phenomena, low birth weight and cataract, share some common etiology. It is thus plausible that influences acting during early pregnancy may induce drastic changes in the fetal milieu, which not only interfere with the early development of the lenses and result in cataract in both eyes, but also result in failure to thrive in utero. Further studies of this high-risk group are needed to identify the mechanism behind this phenomenon.
In contrast, we found no indication of influences during pregnancy to be associated with unilateral cataract cases. The lack of an association with birth weight may suggest that some unilateral cases will develop later than will bilateral cases. Our finding that caesarean section was associated with an increased risk of cataract suggests that complicated deliveries may pose a risk.
In conclusion, the present study of ICI cataract provided little evidence that environmental factors, such as infections, play a role in causing the disease. With the exception of a possible small influence of a complicated delivery, the most significant finding was a substantial association between low birth weight and risk of bilateral ICI cataract. Finally, the differences between unilateral and bilateral ICI cataract, both according to gender and other risk factors, strongly suggest that they have different etiologies.