American Journal of Epidemiology Copyright 2005 by the Johns Hopkins Bloomberg School of Public Health All rights reserved

Vol. 162, No. 5 Printed in U.S.A. DOI: 10.1093/aje/kwi214

Cleft Lip and Palate versus Cleft Lip Only: Are They Distinct Defects?

byholm6 Emily W. Harville1, Allen J. Wilcox2, Rolv Terje Lie3,4, Hallvard Vindenes5, and Frank A
1 2

Department of Epidemiology, University of North Carolina-Chapel Hill, Chapel Hill, NC. Epidemiology Branch, National Institute of Environmental Health Sciences, Durham, NC. 3 Department of Public Health and Primary Health Care, University of Bergen, Bergen, Norway. 4 Medical Birth Registry of Norway, Norwegian Institute of Public Health, Bergen, Norway. 5 Department of Plastic Surgery, Haukeland University Hospital, Bergen, Norway. 6 Department of Plastic Surgery, The National Hospital, Oslo, Norway. Received for publication October 16, 2004; accepted for publication April 7, 2005.
Downloaded from http://aje.oxfordjournals.org/ by guest on July 21, 2013

Cleft lip defects are usually regarded as a single entity, with the assumption that an accompanying cleft palate represents the more severe form. The authors linked data from the Medical Birth Registry of Norway with medical records from two centralized centers to provide a population-based data set. They assessed the distribution of cleft lip only and cleft lip with cleft palate by covariate. Among 1.8 million Norwegian livebirths between 1967 and 1998, there were 1,572 cases of cleft lip with cleft palate and 1,122 cases with cleft lip only. Seventeen percent of those with cleft lip and palate had another defect compared with 9% of those with cleft lip only. For boys, the risk was greater for cleft lip and palate than for cleft lip only (odds ratio 2.4 vs. 1.8, p < 0.001 for difference). The risk of cleft lip only, but not of cleft lip and palate, was increased for twins (odds ratio 1.6 vs. 1.1, p 0.11) and infants whose parents were rst cousins (odds ratio 2.7 vs. 0.7, p 0.07). Although cleft lip with cleft palate may simply represent a more severe form of the defect, epidemiologic assessments of cleft lip should, when possible, include separate analyses of these two groups. abnormalities; cleft lip; cleft palate; Norway

Abbreviations: CLO, cleft lip only; CLP, cleft lip and palate; ICD-8, International Classication of Diseases, Eighth Revision.

Cleft palate in the absence of a cleft lip is commonly regarded as etiologically distinct from cleft lip with or without cleft palate on the basis of epidemiologic data, separate patterns of the risks of recurrence (1), and understanding of embryonic palate and lip formation (2). However, researchers have usually considered cleft lip and palate (CLP) and cleft lip only (CLO) as variants of the same defect, differing only in severity (3). Embryologically, a severe defect of the lip can lead to a cleft of the hard palate, which argues for grouping the two as degrees of the same defect. However, there are also reasons why CLO might be distinct from CLP (2). Cleft lip has its origin as a malformation of the primary palate only, while CLP involves both the primary and secondary palates (4). Occasionally, a cleft lip can be found with a separated cleft of the soft but not the hard palate, which suggests two

different defects (5). Treatment classications, geared toward establishing a course of surgery or orthodontics, have always distinguished the two (68), and postnatal craniofacial growth differs in the two groups (4). In the studies that have distinguished CLP from CLO, CLP is usually about twice as common (9), although some African studies have reported an overall higher prevalence of CLO (2). The CLP:CLO ratio is higher in regions where the overall prevalence of cleft is high (2). Generally, other noncleft malformations are more common with CLP than with CLO (10, 11). Because most studies have not separated the two (1214), it is not clear to what degree the epidemiologic features of CLP differ from those of CLO. While some groups report a male excess of CLP compared with CLO (2, 9, 15), we know of no systematic comparison of the epidemiologic

Correspondence to Emily W. Harville, Department of Epidemiology, University of North Carolina, CB #7435, Chapel Hill, NC 27599-7435 (e-mail: ewh@unc.edu).

448

Am J Epidemiol 2005;162:448453

Cleft Lip Only versus Cleft Lip and Palate

449

features of the two types of defects. The objective of this paper is to compare risk factors for CLP with risk factors for CLO in a large, well-dened population.

TABLE 1. Cases of cleft lip and palate versus cleft lip only, based on two sources, Norway, 19681997
Registry classication Hospital classication Cleft lip and palate Cleft lip only Total Neither

MATERIALS AND METHODS

Cleft lip and palate Cleft lip only Neither Total

1,239 146 164 1,549

76 684 121 881

93 171 1,845,357

1,408 1,001

Norwegians are assigned a unique identication number at birth. Since 1967, all births in Norway have been required to be reported in the Medical Birth Registry. Care for oral clefts is centralized at two hospitals, Haukeland in Bergen and Rikshospitalet in Oslo; virtually all cases in the country over the last 50 years have been sent to one of these two hospitals for treatment. The registry captures the rst week of birth, while hospitalization for the surgery captures all children that require surgery. Theoretically, this process should include 100 percent of the surviving children who stay in the country. Records were linked by their identication numbers, and analyses were restricted to livebirths. In the Medical Birth Registry, cases of CLO were dened as those corresponding to International Classication of Diseases, Eighth Revision (ICD-8), code 7590, with no concurrent recording of isolated cleft palate (ICD-8, code 7592). Cases of CLP were dened as those corresponding to ICD-8, code 7591. Four spaces are available for recording birth defects; if any of the spaces contained one of these codes, the infant was counted as a case. A CLO case was dened as an infant who had some degree of cleft lip but no cleft palate, while a CLP case was dened as an infant with any degree of both. In 358 instances (13 percent), the birth registry and the hospital records agreed that a cleft was present but disagreed on the type. For these cases, the hospitals classication was used. If the registry reported a cleft lip but the hospital reported cleft palate only (n 29), the infants were not counted as cases. The other cases for whom the registry reported either CLP or CLO, or both, and the hospital reported neither (n 285) reected a mix of infant deaths (n 120), clefts too minor to require surgery, clefts operated on at another hospital in Norway, children who left the country, and false positives. Since we had no way of distinguishing false positives from the other categories, all of these children were included in the data set as cases. Concurrent birth defects identied from the birth registry were dened as a birth defect according to ICD-8, codes 74007489 and 74937599. Cases who had cleft lip with cleft palate were not considered to have a concurrent defect. Concurrent birth defects identied from the hospital records were dened as a birth defect noted in the eld labeled Other defects or syndrome diagnosis. For the purposes of this analysis, a defect recorded by either source was considered an accompanying defect. Ascertainment was assessed by using capture-recapture methods. These methods use the amount of overlap between cases from the two data sources to estimate the proportion of cases missed by both sources (16). This method assumes that the sources are independent, which was not applicable in this analysis: a case noted at birth would be more likely to be referred for treatment. When the registry and hospital disagreed on whether a case was CLO or CLP, we included
Am J Epidemiol 2005;162:448453

data on the case in analyses of both defects for the purpose of capture-recapture analysis only. We carried out stratied capture-recapture analyses to determine whether ascertainment varied by characteristics of the mother, child, or birthplace. For the remaining analyses, we combined the two sources of information on clefts to dene the case group. We were interested in determining whether facial clefts might be part of a larger picture of subtle fetal growth and development problems, so we examined birth weight and gestational length. The time-trend data were smoothed by using the PROC LOESS procedure in SAS software (SAS Institute, Inc., Cary, North Carolina), with a window of 0.4. All infants had a record in the Medical Birth Registry and so had data on covariates, even if a cleft had not been recorded at birth. Prevalence at birth was stratied by available covariates, including age of the mother and father, sex of the infant, hospital characteristics, region, and mothers marital status. Data were examined by stratied analysis using chi-square tests to test for statistically signicant differences. Logistic regression was used to adjust for the effects of several variables.
RESULTS

Downloaded from http://aje.oxfordjournals.org/ by guest on July 21, 2013

Overall, 1,572 cases of CLP and 1,122 cases of CLO were reported. A total of 1,239 cases of CLP and 684 cases of CLO were reported by both the hospital and the registry (table 1), while information on the remaining cases was either absent from the other source or reported in the other source as a different defect. When we combined both sources of data, we found that the overall prevalence of cleft lip with or without cleft palate in Norway over the last 30 years was 1.46 per 1,000 livebirths (0.85/1,000 livebirths for CLP and 0.61/1,000 livebirths for CLO). Both the hospital and the registry recorded an accompanying defect in about twice as high a proportion of CLP infants as in those with CLO: 13 percent versus 6 percent for cases recorded by the hospital, and 10 percent versus 5 percent for cases recorded by the registry. When the two sources were combined, 17 percent (272/1,572) of CLP infants and 9 percent (103/1,122) of CLO infants had at least one other defect. Overall prevalence of the two defects was fairly constant over time (gure 1). Within this overall at pattern, in the last decade there has been both an increase in the number of infants with CLP and a decline in those with CLO.

450 Harville et al.

Downloaded from http://aje.oxfordjournals.org/ by guest on July 21, 2013

FIGURE 1. Smoothed prevalence of cleft lip with or without cleft palate and no other defect, Norway, 19671998. Dotted lines, unsmoothed data. Patterns for cases with accompanying defects were similar.

Ascertainment for CLP was estimated at 88 percent by the hospital and 79 percent by the registry. For CLO, the ascertainment was somewhat lower: 76 percent by the hospital and 63 percent by the registry. In stratied analysis, ascertainment did not differ by maternal age, marital status, hospital size, or hospital type. Infants with cleft lip and cleft palate together had a poorer outcome in other respects (table 2). They were more likely to be born preterm and had a lower mean birth weight at term. They also had a higher infant mortality rate (25/1,000 for CLP and 15/1,000 for CLO, p for difference 0.09). Stratied analysis showed other differences between infants with CLP and CLO (table 3). Boys were at higher risk of both types of defect, but especially CLP (p < 0.001 for difference in pattern between CLP and CLO). In contrast, twins were at increased risk of CLO but not CLP (p 0.11

for difference in pattern). Risk of CLO was also raised for those whose parents were rst cousins (p 0.07 for difference in pattern). Patterns were generally similar for those cases with and without accompanying defects. The exceptions were that the sex ratio was higher for CLP cases without other defects and for CLO cases with other defects, and that the risk for cases with other defects tended to rise with maternal age. Logistic regression did not reveal any substantially different patterns from those seen in stratied analysis (e.g., results for cases without accompanying defects, table 4).
DISCUSSION

Of the 2,694 infants with a cleft lip, 58 percent (n 1,572) had CLP and 42 percent (n 1,122) had CLO. This

TABLE 2. Preterm birth and birth weight (mean in grams (standard deviation)) among infants with and without cleft lip, cleft palate, and other defects, Norway, 19681997
Cleft lip and palate, no accompanying defect No. Very preterm (1034 weeks)y Preterm (>3437 weeks) Full term % Birth weight* Cleft lip and palate, accompanying defect No. % Birth weight* No. No cleft lip % Birth weight* Cleft lip only, no accompanying defect No. % Birth weight* Cleft lip only, accompanying defect No. % Birth weight*

33

2.7 1,790 (859)

21

8.3 1,594 (594)

29,984

1.7 1,889 (842)

20

2.1 1,839 (938)

4 15 71

4.4 1,620 (983) 16.7 2,068 (635) 78.9 3,430 (668)

72 5.9 2,624 (488) 29 11.4 2,333 (645) 68,899 4.0 2,774 (600) 43 4.5 2,771 (656) 1,109 91.4 3,532 (522) 204 80.3 3,239 (648) 1,631,486 94.3 3,567 (503) 894 93.4 3,559 (533)

* p 0.001 for difference in mean birth weight, cleft lip and palate vs. cleft lip only; p < 0.001 for difference in mean birth weight, cleft lip and palate, cleft lip only, no cleft. y For 115,196 infants, data on gestational age were missing.

Am J Epidemiol 2005;162:448453

Cleft Lip Only versus Cleft Lip and Palate

451

TABLE 3. Prevalence of cleft lip, with and without cleft palate, by covariates, among Norwegian livebirths, 19671998
Cleft lip and palate, no accompanying defect No. Mothers age (years) <20 2024 2529 3034 3539 40 Fathers age (years) <20 2024 2529 3034 3539 40 Mothers marital status Married Cohabiting Single Widowed/divorced Relationship between parents None First cousins Second cousins Plurality Singleton Twin Infants sex Male Female Parity 0 1 2 3 4 Region East Middle/West North 715 437 148 Reference 0.84 0.87 0.74, 0.94 0.73, 1.04 150 88 34 Reference 0.81 0.95 0.62, 1.05 0.65, 1.38 552 360 105 Reference 0.90 0.80 0.78, 1.02 0.65, 0.98 59 31 13 Reference 0.72 0.92 0.47, 1.11 0.51, 1.68 518 475 189 67 38 Reference 1.10 0.93 1.06 1.08 0.97, 1.24 0.79, 1.10 0.82, 1.37 0.78, 1.50 119 97 34 15 6 Reference 0.98 0.73 1.03 0.84 0.75, 1.28 0.50, 1.07 0.60, 1.76 0.37, 1.90 417 369 152 52 28 Reference 1.06 0.93 1.02 1.03 0.92, 1.22 0.78, 1.12 0.76, 1.36 0.70, 1.50 44 35 19 2 3 Reference 0.95 1.11 0.37 1.16 0.61, 1.49 0.65, 1.89 0.09, 1.53 0.36, 3.73 930 370 2.38 Reference 2.11, 2.68 167 105 1.50 Reference 1.18, 1.92 650 369 1.67 Reference 1.47, 1.89 74 29 2.41 Reference 1.57, 3.71 1,270 29 Reference 1.03 0.71, 1.48 266 6 Reference 1.01 0.45, 2.88 985 34 Reference 1.55 1.10, 2.18 98 5 Reference 2.29 0.93, 5.63 1,271 4 6 Reference 0.91 1.33 0.34, 2.43 0.60, 2.97 266 0 1 Reference 0.00 1.06 0.15, 7.56 996 9 5 Reference 2.61 1.42 1.36, 5.03 0.59, 3.41 100 0 0 Reference 0.00 0.00 949 202 131 13 Reference 1.20 1.12 1.02 1.03, 1.40 0.93, 1.34 0.59, 1.76 202 44 22 2 Reference 1.23 0.88 0.73 0.89, 1.70 0.57, 1.37 0.18, 2.96 768 138 100 12 Reference 1.01 1.05 1.16 0.85, 1.22 0.86, 1.30 0.66, 2.05 70 15 14 3 Reference 1.21 1.62 3.18 0.69, 2.11 0.78, 3.36 1.00, 10.10 12 183 416 321 161 112 1.63 1.01 Reference 0.91 0.91 1.17 0.79, 1.05 0.76, 1.09 0.95, 1.44 0.92, 2.90 0.85, 1.20 1 44 89 50 35 33 0.64 1.13 Reference 0.66 0.92 1.61 0.47, 0.94 0.62, 1.37 1.08, 2.40 0.09, 4.57 0.79, 1.63 3 130 334 272 135 72 0.51 0.89 Reference 0.96 0.95 0.93 0.82, 1.12 0.78, 1.16 0.72, 1.21 0.16, 1.59 0.73, 1.09 1 13 35 20 18 7 1.62 0.85 Reference 0.67 1.21 0.87 0.39, 1.17 0.68, 2.13 0.38, 1.95 0.22, 11.82 0.45, 1.61 119 375 428 262 98 18 1.40 1.02 Reference 1.06 1.11 1.03 0.91, 1.24 0.89, 1.39 0.65, 1.66 1.14, 1.72 0.89, 1.17 16 88 78 55 28 7 1.03 1.31 Reference 1.22 1.75 2.21 0.87, 1.73 1.13, 2.69 1.02, 4.79 0.60, 1.77 0.97, 1.78 66 300 366 198 79 10 0.91 0.95 Reference 0.94 1.05 0.67 0.79, 1.12 0.82, 1.34 0.36, 1.26 0.70, 1.18 0.82, 1.11 5 33 33 20 9 3 0.76 1.16 Reference 1.05 1.33 2.24 0.60, 1.83 0.63, 2.77 0.69, 7.29 0.30, 1.95 0.72, 1.88 RR* 95% CI* No. Cleft lip and palate, accompanying defect RR 95% CI No. Cleft lip only, no accompanying defect RR 95% CI No. Cleft lip only, accompanying defect RR 95% CI

Downloaded from http://aje.oxfordjournals.org/ by guest on July 21, 2013

* RR, relative risk; CI, condence interval.

proportion is consistent with other studies that have distinguished the two types of defects (2). In our data, the prevalence of cleft palate only was 0.8 per 1,000 livebirths, the prevalence of CLO was 0.6 per 1,000, and the prevalence of CLP was 0.9 per 1,000, which underscores the fact that CLP is not just the chance co-occurrence of the two defects. Some of our results support the idea that CLP is simply a more severe version of CLO. For instance, 17 percent of our CLP infants had accompanying defects as opposed to 9 percent of CLO infants. This pattern is similar to that in other studies, although the absolute prevalence in our study was lower (10, 11). It is also possible that the underlying
Am J Epidemiol 2005;162:448453

etiology producing the accompanying noncleft defect also increases susceptibility to more severe CLP (17, 18). Preterm delivery was more common among infants with CLP. Unlike Jensen et al. (9), we found that CLP infants were lighter at birth than either CLO or noncleft infants. Similarly, while cleft lip infants had higher infant mortality rates overall than other infants did, mortality was especially high among those with an accompanying cleft palate (25/1,000 compared with 15/1,000). However, other factors suggest a qualitative difference between the two categories of cleft lip (CLP and CLO). We saw a stronger male predominance among CLP than CLO infants, as has been reported in other studies (2, 9,

452 Harville et al.

TABLE 4. Logistic models of cleft lip with and without cleft palate, accompanying defects excluded, Norway, 19671998*
Cleft lip and palate OR 95% CIy OR Cleft lip only 95% CI

Mothers age (years) <20 2024 2529 3034 3539 40 Fathers age (years) <20 2024 2529 3034 3539 40 Infant sex male vs. female Marital status Cohabiting Single Widowed/divorced Married Parents being rst cousins Twins Region Middle/West North East Parity 0 1 2 3 4 1 1.21 1.09 1.26 1.19 1.05, 1.39 0.89, 1.32 0.95, 1.68 0.81, 1.76 1 1.08 0.95 1.03 1.19 0.92, 1.26 0.77, 1.18 0.74, 1.43 0.78, 1.82 1.17 0.95 1 0.96, 1.41 0.77, 1.16 1.36 1.24 1 1.07, 1.72 0.97, 1.58 1.22 1.01 0.85 1 0.70 1.10 0.23, 2.18 0.75, 1.60 1.04, 1.44 0.75, 1.36 0.38, 1.91 1.05 1.08 0.91 1 2.72 1.63 1.41, 5.27 1.15, 2.31 0.87, 1.27 0.77, 1.51 0.38, 2.19 1.50 1 1.05 0.91 0.88 1.12 2.38 0.86, 1.28 0.72, 1.15 0.66, 1.16 0.81, 1.55 2.08, 2.70 0.82, 2.74 0.61 1 1.07 1.01 0.98 0.99 1.75 0.85, 1.34 0.78, 1.32 0.72, 1.35 0.68, 1.43 1.54, 2.00 0.19, 1.94 1.38 1 1.01 1.09 1.11 0.99 0.86, 1.20 0.88, 1.35 0.82, 1.50 0.58, 1.70 1.05, 1.81 0.91 1 1.06 1.01 1.11 0.77 0.88, 1.27 0.79, 1.28 0.80, 1.56 0.38, 1.54 0.64, 1.29

Statistical evidence (capture-recapture estimates) shows that ascertainment of CLP is better than for CLO. While cleft lip may be underascertained in Norway, the overall rate of cleft lip in Norway is still among the highest in Europe (2). The underlying rates of the defect in Norway may indeed be high, or underascertainment may be even greater in other countries. Some cases were no doubt missed by both sources of information, especially in the earliest years of the registry before surgery was centralized, and the two sources are not independent. However, ascertainment across strata of covariates was similar, suggesting that missed cases did not necessarily produce bias. Our examination of covariates was limited to those variables that were recorded in the birth registry; this limitation precluded examining other variables of interest, such as maternal smoking, mothers medication use, exposure to chemicals, medical conditions, and nutrition during pregnancy. We excluded stillbirths from our results. Stillbirths are unlikely to be selective with regard to isolated CLO and CLP, which are not lethal defects. If an infant with these defects dies before birth, it is almost certainly from other causes, most likely other defects. Therefore, this is probably not a source of bias in this analysis, which was limited to cases without accompanying defects. Perhaps CLP and CLO have distinct etiologies, at least in some cases. Unless future studies separate the two, it will not be possible to discover the different causal pathways. Whenever feasible, future studies should analyze the two separately to explore further the possibility that some factors may affect the risk of one but not the other.

Downloaded from http://aje.oxfordjournals.org/ by guest on July 21, 2013

ACKNOWLEDGMENTS

Conict of interest: none declared.

REFERENCES 1. Fogh-Andersen P. Inheritance of harelip and cleft palate. Copenhagen, Denmark: Nyt Nordisk Forlag, Arnold Busck, 1942. 2. Mossey PA, Little J. Epidemiology of oral clefts: an international perspective. In: Wyszynski DF, ed. Cleft lip and palate: from origin to treatment. New York, NY: Oxford University Press, 2002:12758. 3. Mitchell LE, Beaty TH, Lidral AC, et al. Guidelines for the design and analysis of studies on nonsyndromic cleft lip and cleft palate in humans: summary report from a Workshop of the International Consortium for Oral Clefts Genetics. Cleft Palate Craniofac J 2002;39:93100. 4. Kreiborg S, Hermann NV. Craniofacial morphology and growth in infants and young children with cleft lip and palate. In: Wyszynski DF, ed. Cleft lip and palate: from origin to treatment. New York, NY: Oxford University Press, 2002: 8797. 5. Saal HM. Classication and description of nonsyndromic clefts. In: Wyszynski DF, ed. Cleft lip and palate: from origin to treatment. New York, NY: Oxford University Press, 2002: 4752. Am J Epidemiol 2005;162:448453

* All odds ratios (ORs) were mutually adjusted for the other variables in the table. y CI, condence interval.

15). Twins and infants whose parents were rst cousins had a stronger risk of CLO than CLP. The association between CLO and rst-cousin parents may indicate a genetic mutation or a marker for membership in an ethnic group that practices cousin marriage, or it may be a chance association because the numbers are quite small. The fact that twinning and consanguinity appear to be risk factors for CLO and not for CLP suggests that disease severity is not the sole explanation for the difference between the two phenotypes. However, even for the CLP or CLO cases, multiple causes of the defect are likely.

Cleft Lip Only versus Cleft Lip and Palate

453

6. Friedman HI, Sayetta RB, Coston GN, et al. Symbolic representation of cleft lip and palate. Cleft Palate Craniofac J 1991;28:2529; discussion 25960. 7. Smith AW, Khoo AK, Jackson IT. A modication of the Kernahan Y classication in cleft lip and palate deformities. Plast Reconstr Surg 1998;102:18427. 8. Ortiz-Posadas MR, Vega-Alvarado L, Maya-Behar J. A new approach to classify cleft lip and palate. Cleft Palate Craniofac J 2001;38:54550. 9. Jensen BL, Kreiborg S, Dahl E, et al. Cleft lip and palate in Denmark, 19761981: epidemiology, variability, and early somatic development. Cleft Palate J 1988;25:25869. 10. Stoll C, Alembik Y, Dott B, et al. Associated malformations in cases with oral clefts. Cleft Palate Craniofac J 2000;37: 417. 11. Tolarova MM, Cervenka J. Classication and birth prevalence of orofacial clefts. Am J Med Genet 1998;75:12637. 12. Christensen K. The 20th century Danish facial cleft populationepidemiological and genetic-epidemiological studies. Cleft Palate Craniofac J 1999;36:96104.

13. Cooper ME, Stone RA, Liu Y, et al. Descriptive epidemiology of nonsyndromic cleft lip with or without cleft palate in Shanghai, China, from 1980 to 1989. Cleft Palate Craniofac J 2000;37:27480. 14. Farrall M, Holder S. Familial recurrence-pattern analysis of cleft lip with or without cleft palate. Am J Hum Genet 1992; 50:2707. 15. Shapira Y, Lubit E, Kuftinec MM, et al. The distribution of clefts of the primary and secondary palates by sex, type, and location. Angle Orthod 1999;69:5238. 16. Lie RT, Heuch I, Irgens LM. Maximum likelihood estimation of the proportion of congenital malformations using double registration systems. Biometrics 1994;50:43344. byholm FE. Cleft lip and palate in a Norwegian population. 17. A II. A numerical study of 1555 CLP-patients admitted for surgical treatment 195475. Scand J Plast Reconstr Surg 1978; 12:3543. 18. Hagberg C, Larson O, Milerad J. Incidence of cleft lip and palate and risks of additional malformations. Cleft Palate Craniofac J 1998;35:405.
Downloaded from http://aje.oxfordjournals.org/ by guest on July 21, 2013

Am J Epidemiol 2005;162:448453