Birth Defect Co-Occurrence Patterns Among Infants With Cleft Lip and/or Palate

Abstract

Objective:

To investigate 2- to 5-way patterns of defects co-occurring with orofacial clefts using data from a population-based registry.

Design:

We used data from the Texas Birth Defects Registry for deliveries between 1999 and 2014 to Texas residents, including 1884 cases with cleft palate (CP) and 5289 cases with cleft lip with or without cleft palate (CL±P) without a known syndrome. We identified patterns of defects co-occurring with CP and with CL±P observed more frequently than would be expected if these defects occurred independently. We calculated adjusted observed-to-expected (O/E) ratios to account for the known tendency of birth defects to cluster nonspecifically.

Results:

Among infants without a syndrome, 23% with CP and 21% with CL±P had at least 1 additional congenital anomaly. Several combinations of defects were observed much more often than expected. For example, the combination of CL±P, congenital hydrocephaly, anophthalmia, and other nose anomalies had an O/E ratio of 605. For both CP and CL±P, co-occurrence patterns with the highest O/E ratios involved craniofacial and brain abnormalities, and many included the skeletal, cardiovascular, and renal systems.

Conclusions:

The patterns of defects we observed co-occurring with clefts more often than expected may help improve our understanding of the relationships between multiple defects. Further work to better understand some of the top defect combinations could reveal new phenotypic subgroups and increase our knowledge of the developmental mechanisms that underlie the respective defects.

Background and Objectives

Orofacial clefts, which include cleft palate alone (CP) and cleft lip with or without cleft palate (CL±P), are among the most common birth defects (Calzolari et al., 2004 2007; Dixon et al., 2011; Mossey et al., 2012; Seto-Salvia et al., 2014; Pereira et al., 2018). Cleft palate alone occurs in 1 in 1500 to 2000 births, and CL±P occurs in 1 in 700 to 1000 births, with prevalence varying by sex, geographic location, and ethnic group (Derijcke et al., 1996; Calzolari et al., 2004, 2007). Although it is possible to correct clefts with surgery, our understanding of the causes and optimal management of nonsyndromic clefts is still evolving. Developmental and psychological consequences are often present even after surgical correction, such as feeding difficulty, abnormal speech and/or hearing, dental abnormalities, and social stigmatization (Seto-Salvia et al., 2014). Additionally, first-year mortality is estimated to be 5% to 14% among infants with CP and 6% to 12% among infants with CL±P (Forrester & Merz, 2003). For both morbidity and mortality, the number and severity of birth defects present in addition to the orofacial cleft likely play a role (Milerad et al., 1997; Carlson et al., 2013).

In fact, orofacial clefts are frequently accompanied by other birth defects (Milerad et al., 1997; Stoll et al., 2000; Calzolari et al., 2004; Forrester et al., 2006; Wyszynski et al., 2006; Calzolari et al., 2007; IPDTOC, 2009; Sekhon et al., 2011; Mossey et al., 2012; Seto-Salvia et al., 2014; Pereira et al., 2018). Approximately 27% of cases with CP and 7% to 11% of cases with CL±P have a known syndrome diagnosis (eg, 22q11.2 deletion; van der Woude syndrome), and even infants without an identified syndrome often have additional co-occurring defects (25% of cases with CP and 17%-20% of cases with CL±P; Calzolari et al., 2004, 2007; IPDTOC, 2009; Mossey et al., 2012). An unknown proportion of the infants who are reported as nonsyndromic may actually have a known syndrome that is undiagnosed or even a syndrome that is not yet described in the literature (Agopian et al., 2018). Additionally, there are likely truly nonsyndromic patterns of co-occurring birth defects that share overlapping genetic and/or developmental mechanisms with orofacial clefts (Agopian et al., 2018). It is not clear how frequently this may occur and for which birth defect combinations, but identifying these nonrandom associations could lead to a better understanding of the etiology of clefts.

Several analytic methods have been proposed for evaluating co-occurrence patterns of multiple birth defects, including comparisons of the observed-to-expected (O/E) ratio for a given combination of defects, clustering techniques, multiple regression analysis, and log-linear analysis (Khoury et al., 1990; Beaty et al., 1991; Kallen et al., 1999; Agopian et al., 2018). These methods have been applied to a few specific known birth defect patterns (eg, Vertebral anomalies, Anal atresia, Cardiac defects, TracheoEsophageal anomalies, Renal abnormalities, Limb anomalies [VACTERL] and Omphalocele, Exstrophy, Imperforate anus, Spinal defects [OEIS]) (Mastroiacovo, 1991; Khoury et al., 1993; Botto et al., 1997; Kallen et al., 1999, 2000). However, to date, our understanding of patterns of birth defect co-occurrence among infants with CL±P is relatively limited. While previous studies have described the most common types of associated defects, most have focused on simply assessing the proportion of affected cases with select, single co-occurring defects of interest (eg, the percent with microphthalmia) (Shprintzen et al., 1985; Lilius, 1992; Milerad et al., 1997; Stoll et al., 2000; Shaw et al., 2004; Rittler et al., 2008; Beriaghi et al., 2009; Genisca et al., 2009; Sekhon et al., 2011; Hadadi et al., 2017; Pereira et al., 2018). The goal of this study is, therefore, to investigate the complex patterns (ie, 2- to 5-way combinations) of birth defects that co-occur with orofacial clefts, using data from a population-based registry.

Materials and Methods

Study Patients

This analysis used data from the Texas Birth Defects Registry (TBDR) for deliveries between 1999 and 2014 to Texas residents. The TBDR is an active, population-based surveillance system within the Texas Department of State Health Services. Trained staff identify cases with birth defects (including live births, fetal deaths, and induced terminations) by reviewing birth logs, discharge lists, and other medical records from health care facilities throughout Texas. Diagnosis codes are assigned based on a modified version of the British Pediatric Association (BPA) Classification of Diseases, used by the Centers for Disease Control and Prevention (National Center on Birth Defects and Developmental Disabilities, 2002). To allow for the analysis of distinct categories of co-occurring major birth defects, we grouped defects into 146 categories created from the first 4 digits of the 6-digit BPA code, as previously described (Benjamin et al., 2019).

We analyzed data on infants with a diagnosis of CP or CL±P within the first year of life. Cases were selected based on the inclusion of BPA codes 749.00–749.08 for CP and 749.10–749.22 for CL±P. We included only singletons in the analysis. Infants from multiple gestation pregnancies represent a unique group, since they tend to be particularly at risk for certain birth defect patterns (Schinzel et al., 1979; Mastroiacovo et al., 1999; Dawson et al., 2016). We excluded cases with diagnoses of syndromes, associations (eg, VACTERL), or chromosomal abnormalities identified by the TBDR within the first year of life. Such cases were identified based on BPA codes as well as through manual review of text descriptions. Sex-specific birth defects (such as hypospadias) and minor defects, as classified by the National Birth Defects Prevention Study (Rasmussen et al., 2003) and/or clinicians at TBDR, were not included in our analyses. The TBDR data were linked to vital records from the Texas Vital Statistics Section at the Texas Department of State Health Services to obtain information on maternal and infant characteristics (eg, maternal age and race/ethnicity).

Analysis

We conducted all analyses separately for infants with CP and with CL±P. The counts and proportions of cases across different categories of infant and maternal characteristics (eg, race/ethnicity) were tabulated separately for isolated and nonisolated (ie, additional major co-occurring birth defects present) cases with CP and CL±P.

To identify the individual birth defects most frequently observed among infants with orofacial clefts in our data, and for comparison to prior literature, we tabulated the count and proportion of infants with CP or CL±P that had each specific co-occurring defect (eg, the proportion of infants with CP with microphthalmia; Table 2).

Table 2.

Top 30 Most Frequent Birth Defects Co-Occurring With Nonsyndromic Cleft Palate Alone (CP) and With Nonsyndromic Cleft Lip With or Without Cleft Palate (CL ± P), Texas Birth Defects Registry, 1999–2014 (N = 1884 CP; N = 5289 CL ± P).

Description of co-occurring defecta	Number of cases with the defect combination (%)
With CP
Ostium secundum–type atrial septal defect	93 (4.9)
Ventricular septal defect	70 (3.7)
Anomalies of ear causing impairment of hearing	40 (2.1)
Other deformities of feet	35 (1.9)
Reduction deformities of brain	29 (1.5)
Other unspecified anomalies of face and neck	29 (1.5)
Other anomalies of aorta	27 (1.4)
Microcephalus	25 (1.3)
Congenital hydrocephalus	24 (1.3)
Polydactyly	24 (1.3)
Syndactyly	20 (1.1)
Obstructive defects of renal pelvis and ureter	20 (1.1)
Anomalies of spine	18 (1.0)
Other specified anomalies of the heart	17 (0.9)
Other specified anomalies of unspecified limb	17 (0.9)
Varus (inward) deformities of feet	16 (0.8)
Other anomalies of ribs and sternum	16 (0.8)
Other anomalies of lower limb including pelvic girdle	15 (0.8)
Anencephalus	15 (0.8)
Other anomalies of upper limb including shoulder girdle	14 (0.7)
Microphthalmos	14 (0.7)
Anomalies of pulmonary artery	14 (0.7)
Renal agenesis and dysgenesis	14 (0.7)
Atresia and stenosis of large intestine rectum and anal canal	12 (0.6)
Reduction defects of upper limb	12 (0.6)
Other specified anomalies of ear	12 (0.6)
Other specified anomalies of brain	11 (0.6)
Tricuspid atresia and stenosis	11 (0.6)
Anomalies of pulmonary valve	11 (0.6)
Other anomalies of intestine	11 (0.6)
With CL ± P
Ostium secundum type atrial septal defect	234 (4.4)
Ventricular septal defect	171 (3.2)
Reduction deformities of brain	133 (2.5)
Microcephalus	96 (1.8)
Other deformities of feet	84 (1.6)
Anomalies of ear causing impairment of hearing	80 (1.5)
Other anomalies of nose	76 (1.4)
Anencephalus	66 (1.2)
Polydactyly	60 (1.1)
Obstructive defects of renal pelvis and ureter	60 (1.1)
Congenital hydrocephalus	59 (1.1)
Syndactyly	47 (0.9)
Other unspecified anomalies of face and neck	39 (0.7)
Anomalies of spine	38 (0.7)
Other anomalies of aorta	37 (0.7)
Reduction defects of upper limb	33 (0.6)
Other anomalies of ribs and sternum	32 (0.6)
Anomalies of great veins	32 (0.6)
Microphthalmos	32 (0.6)
Other specified anomalies of the heart	31 (0.6)
Renal agenesis and dysgenesis	31 (0.6)
Anomalies of pulmonary artery	28 (0.5)
Other anomalies of lower limb including pelvic girdle	28 (0.5)
Other specified anomalies of ureter	26 (0.5)
Anomalies of pulmonary valve	25 (0.5)
Other specified anomalies of ear	25 (0.5)
Varus (inward) deformities of feet	24 (0.5)
Other specified anomalies of brain	23 (0.4)
Congenital anomalies of posterior segment	23 (0.4)
Coloboma and other anomalies of anterior segments	23 (0.4)

^aBPA codes for defects included in the analysis specified in Supplementary material.

We used the software platform Co-Occurring Defect Analysis (CODA) (Benjamin et al., 2019) to analyze complex co-occurrence patterns of CP and CL±P with other major birth defects in our data. The CODA examines patterns of birth defects and estimates the adjusted observed-to-expected prevalence ratio (O/E_adj ratio) for each observed birth defect combination, using an approach originally proposed by Khoury et al. (1990). Briefly, this method compares the observed proportion of infants with a given combination of birth defects to the proportion that would be expected if the defects occurred independently. An O/E ratio greater than 1.0, therefore, indicates that a combination occurs more frequently than expected. The CODA also implements an adjustment to the O/E ratios to account for the nonspecific, generalized tendency of birth defects to co-occur with other major defects more often than predicted by chance alone, which is known to lead to inflated O/E ratios (Khoury et al., 1990). In other words, without this adjustment, nearly all birth defect combinations are expected to have O/E ratios greater than 1.0 (Khoury et al., 1990).

We separately calculated the O/E_adj ratio for all combinations of 1, 2, 3, or 4 co-occurring birth defects among cases with CP and CL±P with at least 3 observed cases. We ranked all of the patterns by O/E_adj ratio and focused our interpretation on the 30 combinations with the highest O/E_adj ratios (Tables 3 and 4). If cases included in a combination were identical to those in a higher O/E_adj ratio combination, and if the defects in the combination were a subset of the defects in the higher O/E_adj ratio combination (eg, same 3 cases with CP and microcephaly vs CP, microcephaly, and coloboma), the redundant, lower order combination was excluded from our top patterns. In other words, we focused on the more specific defect combination. To address potential ascertainment differences between live births and non–live births, we also repeated the calculation of O/E_adj ratios, specifically among live births.

Table 3.

Birth Defect Combinations Co-Occurring With Nonsyndromic Cleft Palate (CP) With the Top 30 Ratios of Observed: Expected Prevalence, Texas Birth Defects Registry, 1999–2014 (N = 1884).

Description of co-occurring defectsa	Number of cases with the defect combination	O/Eadj ratiob
Microphthalmos, ventricular septal defect, ostium secundum–type atrial septal defect, tricuspid atresia and stenosis	3	305.2
Anomalies of ear causing impairment of hearing, choanal atresia	3	81.7
Other anomalies of tongue	6	75.0
Reduction deformities of brain, other specified anomalies of brain, other anomalies of aorta	3	50.9
Unspecified anomalies of genital organs	3	49.4
Other anomalies of upper limb including shoulder girdle, other anomalies of ribs and sternum	5	46.3
Transposition of great vessels, tricuspid atresia and stenosis, other anomalies of aorta, anomalies of pulmonary artery	3	33.8
Microphthalmos, other unspecified anomalies of face and neck	4	31.6
Other unspecified anomalies of face and neck, other anomalies of upper limb including shoulder girdle	3	28.1
Renal agenesis and dysgenesis, other anomalies of upper limb including shoulder girdle	3	28.0
Reduction deformities of brain, other anomalies of nose	3	27.8
Ventricular septal defect, ostium secundum–type atrial septal defect, other specified anomalies of unspecified limb	3	27.7
Congenital hydrocephalus, other specified anomalies of unspecified limb	4	24.7
Anomalies of ear causing impairment of hearing, other anomalies of ribs and sternum	3	24.2
Other specified anomalies of brain, congenital anomalies of posterior segment	3	23.9
Other unspecified anomalies of face and neck, other specified anomalies of unspecified limb	3	23.2
Cystic hygroma, other deformities of feet	3	21.9
Reduction defects of upper limb, other anomalies of ribs and sternum	3	20.8
Renal agenesis and dysgenesis, syndactyly	3	19.5
Other unspecified anomalies of face and neck, other anomalies of ribs and sternum	3	19.4
Varus (inward) deformities of feet, other anomalies of upper limb including shoulder girdle	3	18.9
Microphthalmos, ventricular septal defect	5	18.9
Other specified anomalies of brain, anomalies of ear causing impairment of hearing	3	18.9
Reduction deformities of brain, anomalies of peripheral vascular system	3	18.8
Encephalocele, reduction deformities of brain	3	18.6
Other unspecified anomalies of face and neck, other anomalies of lower limb including pelvic girdle	3	17.5
Other anomalies of lower limb including pelvic girdle, other anomalies of ribs and sternum	3	16.1
Hypothyroidism congenital, ostium secundum–type atrial septal defect	4	15.9
Reduction defects of upper limb, other anomalies of upper limb including shoulder girdle	4	15.3
Other unspecified anomalies of face and neck, polydactyly	3	15.3

^aBPA codes for defects included in the analysis specified in Supplementary material.

^bAdjusted observed-to-expected (O/E_adj) prevalence ratio.

Table 4.

Birth Defect Combinations Co-Occurring With Nonsyndromic Cleft Lip With or Without Cleft Palate (CL ± P) With the Top 30 Ratios of Observed-to-Expected Prevalence, Texas Birth Defects Registry, 1999–2014 (N = 5289).

Description of co-occurring defectsa	Number of cases with the defect combination	O/Eadj ratiob
Congenital hydrocephalus, anophthalmos, other anomalies of nose	3	604.7
Reduction deformities of brain, anophthalmos, obstructive defects of renal pelvis and ureter	3	459.6
Reduction deformities of brain, anophthalmos, microphthalmos	3	452.8
Reduction deformities of brain, anophthalmos, other anomalies of nose	4	276.8
Anomalies of ear causing impairment of hearing, renal agenesis and dysgenesis, anomalies of spine	3	269.3
Microphthalmos, congenital anomalies of eyelids lacrimal system and orbit, other unspecified anomalies of face and neck	4	263.8
Coloboma and other anomalies of anterior segments, congenital anomalies of eyelids lacrimal system and orbit, other unspecified anomalies of face and neck	3	263.4
Anophthalmos, other anomalies of nose	8	259.7
Microcephalus, reduction deformities of brain, other anomalies of nose	18	236.3
Microcephalus, reduction deformities of brain, ventricular septal defect, other anomalies of nose	5	220.4
Microphthalmos, coloboma and other anomalies of anterior segments, congenital anomalies of eyelids lacrimal system and orbit	3	196.6
Anomalies of great veins, renal agenesis and dysgenesis, anomalies of spine	3	157.7
Microphthalmos, coloboma and other anomalies of anterior segments, other unspecified anomalies of face and neck	3	156.2
Anomalies of great veins, atresia and stenosis of large intestine rectum and anal canal, renal agenesis and dysgenesis	3	155.2
Anophthalmos, other unspecified anomalies of face and neck	4	150.7
Reduction deformities of brain, other anomalies of nose	32	147.2
Reduction deformities of brain, congenital anomalies of posterior segment, other anomalies of nose	5	141.9
Congenital hydrocephalus, tricuspid atresia and stenosis, congenital mitral stenosis	3	134.9
Reduction deformities of brain, anophthalmos	9	131.7
Anomalies of ear causing impairment of hearing, renal agenesis and dysgenesis, other anomalies of ribs and sternum	3	131.2
Anophthalmos, microphthalmos	3	129.4
Reduction deformities of brain, microphthalmos, other anomalies of ribs and sternum	3	123.4
Varus (inward) deformities of feet, anomalies of adrenal gland	3	122.8
Microcephalus, reduction deformities of brain, ostium secundum–type atrial septal defect, other anomalies of nose	3	119.0
Congenital hydrocephalus, anophthalmos	4	114.0
Microcephalus, ventricular septal defect, other anomalies of nose	5	113.9
Ventricular septal defect, anomalies of great veins, anomalies of spine	3	110.3
Microcephalus, other anomalies of nose	21	110.0
Anophthalmos, obstructive defects of renal pelvis and ureter	3	105.1
Reduction deformities of brain, ventricular septal defect, other anomalies of nose	6	100.8

^aBPA codes for defects included in the analysis specified in Supplementary material.

^bAdjusted observed-to-expected (O/E_adj) prevalence ratio.

To generate a visual representation of the overlap in defects featured in the 30 combinations with the highest O/E_adj ratios for CP and for CL±P, we further grouped the defects into broader domains by organ system and created Venn diagrams showing the overlap across combinations. We limited these diagrams to 5 domains that were commonly represented among the top combinations: skeletal (eg, limb defects), cardiovascular (eg, atrial septal defect, anomalies of pulmonary artery), eye (eg, microphthalmia), central nervous system (CNS), and ear, nose, or throat (ENT) defects. The Venn diagrams were constructed such that each cell indicated the number of top combinations involving the corresponding overlapping domains (eg, the number of top combinations that included CP, 1 or more skeletal defects, and 1 or more eye defects).

Results

From 1999 to 2014, there were 6 131 631 live births in Texas. The TDBR identified 195 536 infants with a birth defect delivered during this time, out of which 30 108 were diagnosed with known (eg, syndromic) patterns and were excluded from analysis. Of the nonsyndromic cases, 1884 infants were diagnosed with CP (3.1/10 000 live births) and 5289, with CL±P (8.6/10 000 live births). We tabulated the count and proportion of maternal and infant characteristics for infants with isolated versus nonisolated CP and CL±P (Table 1). Our main analyses focused on cases with nonsyndromic, nonisolated CP and CL±P.

Table 1.

Characteristics of Mothers of Infants With Nonsyndromic Cleft Palate Alone (CP) and Cleft Lip With or Without Cleft Palate (CL ± P) in the Texas Birth Defects Registry (1999–2014).

	CP		CL ± P
Demographic characteristic	Nonisolated cases, a N = 434 (%)	Isolated cases, N = 1450 (%)	Nonisolated cases, a N = 1106 (%)	Isolated cases, N = 4183 (%)
Race/ethnicity
White, non-Hispanic	150 (34.6)	594 (40.9)	335 (30.3)	1624 (30.8)
Black, non-Hispanic	30 (6.9)	149 (10.2)	114 (10.3)	262 (6.2)
Hispanic	235 (54.3)	615 (42.4)	615 (55.7)	2097 (50.1)
Other	18 (4.2)	92 (6.3)	40 (3.6)	198 (4.7)
Age (years)
<20	50 (11.5)	157 (10.8)	169 (15.2)	608 (14.5)
20–24	111 (25.5)	395 (27.2)	325 (29.3)	1206 (28.8)
25–29	117 (26.9)	403 (27.8)	283 (25.5)	1074 (25.6)
30–34	96 (22.1)	300 (20.7)	191 (17.2)	834 (19.9)
35–39	41 (9.4)	150 (10.3)	104 (9.4)	365 (8.7)
40+	19 (4.3)	44 (3)	34 (3)	95 (2.2)
Education
Less than high school	133 (31.9)	414 (29)	334 (32.4)	1288 (31.3)
High school	129 (31)	425 (29.8)	332 (32.2)	1242 (30.2)
Greater than high school	154 (37)	585 (41)	365 (35.4)	1575 (38.3)
Sex
Male	187 (43.1)	588 (40.5)	589 (53.4)	2653 (63.4)
Female	242 (55.8)	860 (59.3)	509 (46.1)	1529 (36.5)
Pregnancy outcome
Live birth	395 (91)	1433 (98.8)	969 (87.6)	4105 (98.1)
Other	39 (8.9)	17 (1.1)	137 (12.3)	78 (1.8)
Birth year
1999–2004	159 (36.6)	523 (36)	389 (35.1)	1576 (37.6)
2005–2009	150 (34.5)	512 (35.3)	365 (33)	1363 (32.5)
2010–2014	125 (28.8)	415 (28.6)	352 (31.8)	1244 (29.7)
Number of co-occurring birth defects
1–2	298 (68.6)	–	822 (74.3)	–
3–5	105 (24.1)	–	221 (19.9)	–
6 or more	31 (7.1)	–	63 (5.7)	–

^aMore than 1 major birth defect present as grouped by BPA4.

Approximately 23% of nonsyndromic infants with CP and 21% with CL±P had at least 1 additional major birth defect. For comparison with prior literature, we identified the single birth defects that most frequently co-occurred with nonsyndromic clefts (Table 2). For both types of orofacial clefts, the defects most commonly associated were atrial septal defect (4.9% of cases with CP and 4.4% with CL±P) and ventricular septal defect (3.7% and 3.2%, respectively). Other defects that commonly co-occurred with both types of clefts included defects involving the limbs and digits, such as clubfoot and syndactyly, and eye defects, such as microphthalmia.

Tables 3 and 4 show the 30 combinations, with the highest O/E_adj for CP and for CL±P (ie, the patterns of multiple defects that co-occurred with orofacial clefts more often than would be expected if the defects occurred independently). For CP, the top combination included microphthalmia, ventricular septal defects, ostium secundum–type atrial septal defects, and tricuspid atresia and stenosis (O/E_adj ratio: 305.2). The other top CP combinations had O/E_adj ratios ranging from 15.3 to 81.7.

For CL±P, the top combination included congenital hydrocephalus, anophthalmia, and other anomalies of nose (eg, absent nose; O/E_adj ratio: 604.7). Other top combinations for CL±P included reduction deformities of the brain (eg, absent corpus collosum) and anophthalmia with obstructive defects of renal pelvis and ureter (O/E_adj ratio: 459.6) or with microphthalmia (O/E_adj ratio: 452.8). The other top CL±P combinations had O/E_adj ratios ranging from 92.9 to 269.3. Some of these combinations were observed in a relatively high number of infants, for example, the combination of microcephaly, reduction deformities of brain, and other anomalies of nose (N = 18).

To address potential ascertainment differences between live births and non–live births, we repeated the above main analyses among live births. The results from this analysis were similar to the main results (data not shown).

We generated Venn diagrams to visualize the overlap in domains (skeletal, cardiovascular, eye, CNS, and ENT) of birth defects among the top 30 combinations for CP and for CL±P (Figures 1 and 2). For CP, 18 (60.0%) of the top combinations involved 2 or more of these domains. Ten (33.3%) of the top combinations involved ENT defects in combination with skeletal, CNS, or eye defects. For CL±P, 26 (86.7%) of the top combinations involved 2 or more of the 5 domains: 16 (61.5%) of these involved CNS defects, 12 (46.2%) involved ENT defects, and 13 (50.0%) involved eye defects.

Figure 1.

Number of the top 30 birth defect combinations by organ systems involved, among infants with cleft palate (CP). Each count represents 1 birth defect combination with CP (eg, CP + reduction deformities of the brain + other anomalies of nose would be shown in the cell with CNS and ENT overlap). Numbers do not sum to 30 due to combinations not represented in any of these 5 organ systems (eg, CP + unspecified anomalies of genital organs). CNS indicates central nervous system; ENT, ear, nose, and throat.

Figure 2.

Number of the top 30 birth defect combinations by organ systems involved, among infants with cleft lip with or without cleft palate (CL±P). Each count represents 1 birth defect combination with CL±P (eg, CL±P + congenital hydrocephaly + anophthalmia + other anomalies of nose would be shown in the cell with CNS, eyes, and ENT overlap). CNS indicates central nervous system; ENT, ear, nose, and throat.

Discussion

Our study used data from a large, population-based registry to identify birth defect patterns involving orofacial clefts. Our estimates of the proportion of affected infants with co-occurring defects (23% for CP and 21% for CL±P) are consistent with previous population-based studies of infants without a syndrome diagnosis (Calzolari et al., 2004, 2007; IPDTOC, 2009). The head and neck regions (including ear, nose, and throat), and the skeletal, cardiovascular, and CNS have been identified in the literature as the most common sites of co-occurring malformations associated with orofacial clefts (Shprintzen et al., 1985; Lilius, 1992; Milerad et al., 1997; Stoll et al., 2000; Shafi et al., 2003; Shaw et al., 2004; Forrester et al., 2006; Rawashdeh & Abu-Hawas, 2008; Rittler et al., 2008; Beriaghi et al., 2009; Genisca et al., 2009; Rittler et al., 2011; Sekhon et al., 2011; Mossey et al., 2012; Hadadi et al., 2017; Pereira et al., 2018). This was also observed in our analysis. Previous studies, however, have generally focused on single defects associated with orofacial clefts (Shprintzen et al., 1985; Lilius, 1992; Milerad et al., 1997; Stoll et al., 2000; Shaw et al., 2004; Rittler et al., 2008; Beriaghi et al., 2009; Genisca et al., 2009; Sekhon et al., 2011; Hadadi et al., 2017; Pereira et al., 2018) rather than complex patterns. Patterns involving 2 or more additional defects co-occurring with orofacial clefts have been identified in a few studies (Kallen et al., 1996, 1999), but they have not been extensively described. We identified patterns of up to 4 defects co-occurring with CP and CL±P and reported on the combinations that co-occurred more often than would be expected if these defects occurred independently of one another.

For infants with CP, the defect pattern with the highest O/E ratio included eye and cardiovascular defects (CP, microphthalmia, atrial septal defect, ventricular septal defect, and tricuspid atresia and stenosis). Although heart defects are among the most commonly occurring defects in the general population, our O/E calculations accounted for their high prevalence. The association of clefts with anophthalmia/microphthalmia has been documented in studies of pairwise combinations (Kallen et al., 1996; Shaw et al., 2004; Rittler et al., 2008; Anchlia et al., 2011). Many of our other top CP patterns included skeletal anomalies, such as anomalies of the upper limb (eg, hand, fingers, or wrist) and of the ribs and sternum (eg, absent or extra ribs). These skeletal anomalies were often present in combinations that included facial defects, in particular, other anomalies of face and neck (eg, macro/microstomia or macro/microcheilia). Anomalies of the ear were also present among our top CP patterns, for example, in combinations that included other specified anomalies of brain (eg, cerebral dysgenesis) or choanal atresia. Ear defects and hearing loss in infants with CP have been associated with brain abnormalities in the superior temporal plane (involved in auditory processing and language), as well as with eustachian tube dysfunction caused by alterations in the cranial and muscular anatomy (Sharma & Nanda., 2009).

Infants with CL±P had some of the same co-occurring defects as infants with CP, but they appeared in different combinations and with generally higher O/E_adj ratios (604.7–100.8 for CL±P vs 305.2–15.3 for CP). The larger O/E_adj ratios may suggest a stronger tendency for the defects in these top combinations to co-occur with CL±P, as compared to those in the combinations to co-occur with CP. However, the higher ratios for CL±P could also be a reflection of there being more observed eligible unique combinations of birth defects co-occurring with CL±P (555 combinations) than with CP (255 combinations) in our analysis, which may be related to the larger overall number of cases with CL±P. The patterns with the highest O/E_adj among infants with CL±P included combinations of anomalies in the eyes, face, and CNS. For example, CL±P co-occurred much more often than expected with congenital hydrocephaly, anophthalmia, and other anomalies of the nose (eg, absent nose). Based on prior literature, for both types of orofacial clefts, we expected patterns including the brain, face, and eyes to have high O/E_adj ratios (Kallen & Winberg, 1969). Since both the face and the brain develop from the prechordal region, and the eyes develop as an extension of the forebrain, it is not surprising that anomalies in these regions co-occur (van der Plas et al., 2010; Anchlia et al., 2011). There is some evidence suggesting that certain combinations of eye anomalies, brain defects, and clefts could have shared genetic etiologies. Slavotinek et al. (2012) identified VAX1 as a mutation associated with microphthalmia, corpus callosum agenesis, and orofacial clefting in humans. Indeed, one of the top combinations in our analysis included CL±P, reduction deformities of the brain (ie, corpus callosum), and microphthalmia (O/E_adj ratio: 452.8; N = 3). These combinations may also be consistent with patterns seen among patients with Sonic Hedgehog (Shh) mutations (Dworkin et al., 2016). We also observed cardiovascular defects co-occurring with CL±P, CNS anomalies, and facial defects much more often than expected (eg, CP, microcephaly, reduction deformities of the brain, ventricular septal defect, and other anomalies of the nose). There is speculation that co-occurrence patterns including clefts, cardiovascular defects, and facial defects may be related to mechanisms of early embryonic development, during which the aortic arches of the primitive heart surround the pharyngeal arches that eventually give form to the face (Genisca et al., 2009). Other combinations with high O/E_adj in infants with CL±P included renal defects (eg, obstructive defects of renal pelvis and ureter, and renal agenesis and dysgenesis). This is consistent with developmental field theory, which proposes that any disruption in the midline has the potential to influence the development of other midline structures and result in defects such as oral clefts, CNS, cardiac, abdominal wall, and genitourinary defects (Forrester et al., 2006).

Overall, our findings support clinical practices for screening infants with orofacial clefts for common co-occurring defects, and they may be helpful to better understand differences between CP and CL±P patterns. For example, our results emphasize the importance of continuing screening for hearing loss, particularly for infants with CP (Chen et al., 2008). On the other hand, we observed CNS defects co-occurring with CL±P much more often than expected, which is consistent with studies that have found brain structural abnormalities in individuals with seemingly isolated clefts during brain imaging (Nopoulos et al., 2001; van der Plas et al., 2010; Adamson et al., 2014). In fact, the potential biological association between clefts and CNS anomalies (discussed above) could play a role in the reported higher incidence of neurodevelopmental problems in infants with clefts (including language and intellectual impairment, learning disabilities, and attention deficits). Our findings emphasize the importance of considering brain imaging for affected infants, particularly those with developmental problems (Nopoulos et al., 2001; van der Plas et al., 2010; Richman et al., 2012; Adamson et al., 2014).

The main strengths of this study are its large overall sample size and the application of an agnostic, analytic approach for evaluating complex combinations of birth defects and adjusting for their nonspecific clustering among infants with orofacial clefts. To our knowledge, this is the first comprehensive, epidemiologic evaluation of complex patterns of defect co-occurrence in this population. We also conducted subanalyses among live births, and the results from these analyses suggested that our main findings were not strongly influenced by potential bias related to ascertainment differences between live births and non–live births.

There are some limitations to this study that should be considered, particularly when comparing our findings to previous studies on CL±P. To improve interpretability and facilitate the computational analysis of high-order combinations of birth defects, we collapsed some individual defects into broader anomaly groups (eg, hand anomalies and wrist anomalies were grouped into “other unspecified anomalies of the upper limb”). This may have resulted in some loss of specificity. In addition, while we excluded all infants with documented syndromes from our analysis, some remaining cases may have had syndromes that were not diagnosed in the first year of life (eg, Coloboma, Heart defects, Atresia of the choanae, Retardation of growth and development, Genital and urinary anomalies, and Ear anomalies [CHARGE] syndrome) in the hospitals and other facilities accessed by the TBDR. Some infants with CL±P among the top combinations may have had undiagnosed oculo-auriculo-vertebral dysplasia. This condition is characterized by facial asymmetry, ear and/or eye malformations, and vertebral anomalies and may include CL±P and cardiac defects (Kallen et al., 2004; Heike et al., 2009).

Furthermore, our results may not be generalizable to all populations, given population differences with respect to birth defect co-occurrence tendencies. Our method also did not account for additional infant characteristics (eg, race/ethnicity, sex, or environmental risk factors), though exploring these variables would be an interesting direction for future work. Lastly, despite our use of an extremely large sample, we were limited by small numbers for several of the top birth defect combinations. Our top combinations observed in a larger number of infants may be more informative with respect to population burden and may represent more robust results.

In summary, we identified patterns of defects that include orofacial clefts that were observed much more frequently than would be expected if these defects were independent of each other. For both CP and CL±P, many of these combinations involved craniofacial structures. The co-occurrence of these defects may suggest shared etiologic elements or, in some cases, even a craniofacial sequence (eg, developmental field defect). We also identified patterns that involve defects in other organ systems (eg, renal, skeletal, and heart). Independent confirmation of the observed patterns, as well as an improved characterization, refinement, and understanding of them, may lead to the identification of birth defect combinations that are candidates for further research, and potentially to the identification of new syndromic patterns. Evaluating the nonrandom associations between CP and CL±P and other defects could also improve our understanding of the genetic etiologies and developmental mechanisms that underlie the respective birth defects. Lastly, we hope that increased awareness of the co-occurrence patterns could inform clinical practices in diagnosis, screening, and treatment of infants with orofacial clefts.