Elevated Infant Mortality Rates among Oral Cleft and Isolated Oral Cleft Cases: A Meta-Analysis of Studies from 1943 to 2010

Abstract

Objective

To review available published literature in order to obtain a more comprehensive assessment of the association between infant mortality and oral clefts, overall and isolated.

Design

A wide-ranging search of published studies relating infant mortality rates to oral clefts and congenital malformations was conducted, yielding applicable data sets. Settings of the included studies varied to a limited extent, although all were conducted within high-income countries with superior health indicators. These results were tabulated and meta-analyzed. Random effects odds ratios comparing each data set with its respective population have been used to account for differing settings and specific infant mortality rates.

Results

Nine articles with relatable data were included in these meta-analyses. The calculated odds ratio of infant mortality associated with oral cleft cases, including those with associated malformations, was substantial: 9.466 (95% confidence interval, 6.153 to 14.560). Excluding oral cleft cases with associated malformations, the odds ratio, although lower, was still significant: 2.073 (95% confidence interval, 1.390 to 3.092).

Conclusions

Compared with the overall population, infants with oral clefts have nine times the odds of dying within the first year of life. Furthermore, even without additional malformations, infants with oral clefts are still twice as likely to die before 1 year of age. Current research points to a substantially higher relative risk of infant mortality among oral cleft cases in developing countries. Additional research is essential to determine the sources of these raised infant mortality rates and possible interventions to decrease them.

Keywords

cleft lip infant mortality malformation oral cleft

In recent decades, infant mortality has decreased substantially around the world (Centers for Disease Control and Prevention [CDC]–Office of Minority Health and Health Disparities, 2011). During this time, though, infant mortality due to congenital malformations has proportionally increased (Berry et al., 1987; Rosano et al, 2000). Cleft lip and/or cleft palate, also known as orofacial clefts or oral clefts (OCs), is one of the most common congenital anomalies, affecting approximately 1 in 700 infants (Mossey and Castillia, 2003). The term oral cleft is defined as a range of disorders affecting the lips and oral cavity (Mossey et al., 2009). It is always inclusive of, and in nearly all considered publications has been limited to, cleft lip and/or cleft palate. Thus, OC will be used here synonymously with cleft lip and/or palate.

Over the past 60 years, there have been several studies conducted to evaluate the infant mortality rate (IMR), the standard percentage of infant deaths (at 1 year of age or younger) or the number of deaths of infants under age 1 per 1000 live births, of OC cases. In these studies, additional congenital malformations have been reported to play a major role in mortality rates of infants with OCs. The objective of this paper is to review and compare the individual results of these studies in order to consolidate the available information on the subject and develop a metaanalysis with more robust statistics on the IMRs associated with inclusive OC cases, which include all infants with cleft lip and/or palate, and isolated OC cases, which exclude those with additional malformations.

Method of Literature Search

The literature search and overall evaluation for this metaanalysis has been compiled in line with the guidelines described in the PRISMA guidelines for reviews and meta-analyses (Moher, 2009). Studies published in English and available through MEDLINE, JSTOR, standard Google search, and follow-up exploration under key words cleft, orofacial, infant mortality, death, congenital anomaly, malformation, and craniofacial were considered in this review. This broad search yielded more than 600 publications, only 14 of which were found to have specific data on mortality rates in OC cases. In order to have relatable data among all of the studies included in this meta-analysis, eligibility criteria were formed to limit studies to only those with explicit data specifically regarding infant mortality in cleft cases. Publications that did not have a follow-up period and accompanying data on the entire first year of life were not included. Similarly, studies including deaths occurring after the first year of life were excluded. No publications were excluded due to the date of the study. Furthermore, the eligibility criteria also limited reviewed studies to those of liveborn infants. Studies inclusive of still births were excluded or, when possible, subtracted from the reported figures.

Nine of the 14 studies met our eligibility criteria (see Table 1). Two studies were excluded for presenting data on only perinatal and neonatal mortality in cleft cases, and three others were excluded for including OC case deaths occurring after 1 year of age. As mentioned, no study was excluded for its study period, and our analysis incorporates studies conducted across a 60-year time frame, 1943 to 2005. No appreciable variation in relative IMR of OC cases was observed across this time. Furthermore, odds ratio (OR) statistical analysis has been used in this review to also reflect the significantly decreasing IMRs of the background population during the period of concern. Also to note, although the eligibility criteria did not limit the location of the studies, the localities of those available on the topic are all high-income countries with superior health indicators (the United States, Denmark, and the United Kingdom) (World Bank, 2010). These results, therefore, only reflect the relationship of IMRs and OCs in developed, advanced economic settings comparable to the included study countries.

Table 1

List of Included Articles and Study Characteristics^*

Author, Year	Title	Study Location	Period of Study	Inclusive OC IMR Data Presented	Isolated OC IMR Data Presented	Associated Malformation Rate Data Presented	Control/Reference IMR Data Source	Case Ascertainment and Mortality Data Source
Christensen, 2004	Long Term Follow Up Study of Survival Associated With Cleft Lip and Palate at Birth	Denmark	1943–1998	No	Yes	Yes	Calculated from Danish mortality rate, stratified for age, sex, and time	Birth defect registry linked to civil registration system
Druschel, Hughes, and Olson, 1996	First Year-of-Life Mortality Among Infants With Oral Clefts: New York State, 1983–1990	New York (U.S.A.)	1983–1990	Yes	Yes	Yes	Random sample of four liveborn infants without reported malformations for every case	Birth defect registry linked to birth and death certificates
Emanuel, Culver, Erickson, Guthrie, and Schuldberg, 1973	Further Epidemiological Differentiation of Cleft Lip and Palate: A Population Study of Clefts in King County, Washington, 1856–1965	Washington (U.S.A.)	1956–1965	Yes	Yes	No	General King County, WA, mortality rates for time period	Birth certificates (includes birth defect report), hospital records, and death certificates
Forrester and Merz, 2003	First-Year Mortality Rates for Selected Birth Defects, Hawaii, 1986–1999	Hawaii (U.S.A.)	1986–1999	Yes	Yes	Yes	General U.S. mortality rates for time period, as per UNData^†	Birth defect registry linked to death certificates
Hujoel, Bollen, and Mueller, 1992	First-Year Mortality Among Infants With Facial Clefts	Washington (U.S.A.)	1984–1988	Yes	Yes	Yes	Determined from birth and death certificates of non-OC live births in birth record source	Birth certificates (includes birth defect report) and death certificates
Mackeprang and Hay, 1972	Cleft Lip and Palate Mortality Study	U.S.A. (Multi-site)	1956–1965	Yes	No	No	General U.S. mortality rates for time period	Birth certificates (includes birth defect report) and death certificates via nis
Nembhard, Waller, Sever, and Canfield, 2001	Patterns of First-Year Survival Among Infants With Selected Congenital Anomalies in Texas, 1995–1997	Texas (U.S.A.)	1995–1997	Yes	Yes	Yes	General U.S. mortality rates for time period, as per UNData^†	Birth defect registry linked to birth and death certificates
Richards, 1973	Fetal and Infant Mortality Associated With Congenital Malformations	U.K.	1964–1966	Yes	No	No	General U.K. mortality rates for time period, as per UNData^†	Birth certificates (includes birth defect report), hospital records, and death certificates
Tennant, Pearce, Bythell, and Rankin, 2010	20-Year Survival of Children Born With Congenital Anomalies: A Population-Based Study	U.K. (North England)	1985–2003	No	Yes	No	General U.K. mortality rates for time period, as per WHO life tables	Birth defect registry linked to death certificates and hospital records

List of Excluded Articles and Study Characteristics
Author, Year	Title	Study Location	Period of Study	Reason for Exclusion
Agha, Williams, Marrett, To, and Dodds, 2006	Determinants of Survival in Children With Congenital Abnormalities: A Long-Term Population-Based Cohort Study	Canada	1979–1986	Specific data only available for age 5 mortality; full IMR data not given
Dai, Zhu, Mao, Li, Deng, Wang, Liang, Tang, Wang, Kilfoy, Zheng, and Zhang, 2010	Time Trends in Oral Clefts in Chinese Newborns: Data From the Chinese National Birth Defects Monitoring Network	China	1996–2005	Only perinatal mortality considered; IMR data not given
Dastgiri, Gilmour, and Stone, 2003	Survival of Children Born With Congenital Anomalies	U.K. (Glasgow)	1980–1997	Specific data only available for age 5 mortality; full IMR data not given
Kang, Narayanan, and Kelsall, 2011	Mortality Among Infants Born With Orofacial Clefts in a Single Cleft Network: A Brief Communication	U.K. (East England)	2002–2010	Specific data only available for age 4 mortality; full IMR data not given
Ngai, Martin, Tonks, Wyldes, and Kilby, 2005	Are Isolated Facial Cleft Lip and Palate Associated With Increased Perinatal Mortality? A Cohort Study from the West Midlands Region, 1995–1997	U.K. (West Midlands)	1995–1997	Only perinatal mortality considered; IMR data not given

OC = oral cleft; IMR = infant mortality rate.

†

No control IMR reported by study.

Throughout the included studies, several data collection methods and sources were used to identify study participants, and often multiple sources were used within the same study to corroborate findings, diagnoses, birth and death dates, and related details. The data collection sources are listed in Table 1. Comparison of birth and death certificates alone is traditionally viewed to be the least reliable (Druschel et al., 1996a), but in favor of comprehensiveness, no study has been excluded on the basis of method of information ascertainment.

The data from the nine selected studies, extracted from the tables and text of the publications, were then divided up by OC case type. Seven studies on the infant mortality of inclusive OC cases were included. Five of these included separate data for the etiologically distinct subtypes cleft palate (CP) and cleft lip with or without cleft palate (CL±P) (Mossey et al., 2009), and only one further broke down CL±P data into the subtypes cleft lip (CL) and cleft lip and palate (CLP).

Increased IMRs have been observed in OC infants affected with additional major birth defects, including malformations, syndromes, and sequences. As defined by the CDC, a major birth defect is any structural anomaly that adversely affects health, development, or functional ability. For example, clubfoot would be considered a major malformation, where as a single transverse palmar crease would not (CDC, 2011). For this reason, isolated OC cases, which are those infants without additional major birth defects, have been considered separately. Seven studies were found to include specific data on the IMR of isolated OC cases, five of which also contained inclusive OC data, which has been used in the inclusive IMR analysis. Of these seven, five presented select data for the isolated subtypes CL±P and CP and two provided specific IMR data for CL and CLP cases. Separate analyses of the IMRs associated with classifications OC, CP, and CL±P have been conducted for both inclusive OC cases and isolated OC cases.

Method of Statistical Analysis

After consolidating the data, the collected results were tabulated and aggregated according to OC type and study, with inclusive OC data presented in Table 2 and isolated OC data in Table 3. These tables include all available breakdowns of the data by OC type (i.e., CL/P, CP, CL±P, CLP, and CL). Reported case IMRs were then compared with reference population IMRs. When explicitly stated, the reference population IMR printed in the study was used. Otherwise, the average reference population IMR was calculated according to United Nations (UN) data for the study country during the study time period in consideration (UNData–UN Statistics Division, 2010). Table 1 lists the reference population IMR source used for each study.

Table 2

Study Results: Infant Mortality Associated With Inclusive Oral Cleft Cases^*

Author, Year	Diagnosis	Total Cases	Infant Deaths	Case IMR (%)	Control IMR (%)
Nembhard, Waller, Sever, and Canfield, 2001	CL/P	649	83	12.79	0.70
Druschel, Hughes, and Olson, 1996	CL/P	2786	249	8.94	0.77
Forrester and Merz, 2003	CL/P	502	43	8.57	0.80^†
Hujoel, Bollen, and Mueller, 1992	CL/P	384	40	10.42	0.80
Richards, 1973	CL/P	176	10	5.68	2.05^†
Emanuel et al, 1973	CL/P	378	51	13.49	2.20
Mackeprang and Hay, 1972	CL/P	12,553	1743	13.89	2.57
Total and weighted average	CL/P	17,439	2219	12.73	2.11
Nembhard, Waller, Sever, and Canfield, 2001	CP	246	33	13.41	0.70^†
Druschel, Hughes, and Olson, 1996	CP	1187	115	9.69	0.77
Forrester and Merz, 2003	CP	181	16	9.00	0.80^†
Hujoel, Bollen, and Mueller, 1992	CP	130	20	15.38	0.80
Emanuel et al., 1973	CP	127	23	18.00	2.20
Total and weighted average	CP	1871	207	11.06	0.86
Nembhard, Waller, Sever, and Canfield, 2001	CL±P	403	50	12.41	0.70^†
Druschel, Hughes, and Olson, 1996	CL±P	1599	134	8.38	0.77
Forrester and Merz, 2003	CL±P	321	27	8.00	0.80^†
Hujoel, Bollen, and Mueller, 1992	CL±P	254	20	7.87	0.80
Emanuel et al., 1973	CL±P	251	28	11.20	2.20
Total and weighted average	CL±P	2828	259	9.16	0.89
Emanuel et al., 1973	CLP	162	25	15.50	2.20
Emanuel et al., 1973	CL	89	3	3.40	2.20

IMR = infant mortality rate; CL/P = cleft lip and/or palate; CP = cleft palate; CL±P = cleft lip with or without cleft palate; CLP = cleft lip and palate; CL = cleft lip.

†

No control infant mortality rate reported by study; general control population IMR obtained from United Nations Data for country and time period in consideration.

Table 3

Study Results: Infant Mortality Associated With Isolated Oral Cleft Cases^*

Author, Year	Diagnosis	Total Cases	Infant Deaths	Case IMR (%)	Control IMR (%)
Nembhard, Waller, Sever, and Canfield, 2001	Isolated CL/P	535	14	2.62	0.70^†
Tennant, Pearce, Bythell, and Rankin, 2010	Isolated CL/P	584	13	2.20	0.70
Druschel, Hughes, and Olson, 1996	Isolated CL/P	1802	18	1.00	0.77
Forrester and Merz, 2003	Isolated CL/P	283	0	0.00	0.80^†
Hujoel, Bollen, and Mueller, 1992	Isolated CL/P	307	9	2.93	0.80
Christensen, 2004	Isolated CL/P	5331	175	3.28	2.05
Emanuel et al., 1973	Isolated CL/P	277	7	2.53	2.20
Total and weighted average	Isolated CL/P	9119	236	2.59	1.55
Nembhard, Waller, Sever, and Canfield, 2001	Isolated CP	196	6	3.06	0.70^†
Tennant, Pearce, Bythell, and Rankin, 2010	Isolated CP	217	8	3.70	0.70
Druschel, Hughes, and Olson, 1996	Isolated CP	670	6	0.90	0.77
Hujoel, Bollen, and Mueller, 1993	Isolated CP	87	1	1.15	0.80
Emanuel et al., 1973	Isolated CP	77	2	2.60	2.20
Total and weighted average	Isolated CP	1247	23	1.84	0.84
Nembhard, Waller, Sever, and Canfield, 2001	Isolated CL±P	339	8	2.36	0.70^†
Tennant, Pearce, Bythell, and Rankin, 2010	Isolated CL±P	367	5	1.40	0.70
Druschel, Hughes, and Olson, 1996	Isolated CL±P	1132	12	1.06	0.77
Hujoel, Bollen, and Mueller, 1994	Isolated CL±P	220	8	3.64	0.80
Emanuel et al., 1973	Isolated CL±P	200	5	2.50	2.20
Total and weighted average	Isolated CL±P	2258	38	1.68	0.88
Tennant, Pearce, Bythell, and Rankin, 2010	Isolated CLP	140	3	1.80	0.70
Emanuel et al., 1973	Isolated CLP	122	5	4.10	2.20
Total and weighted average	Isolated CLP	262	8	3.05	1.40
Tennant, Pearce, Bythell, and Rankin, 2010	Isolated CL	227	2	0.70	0.70
Emanuel et al., 1973	Isolated CL	78	0	0.00	2.20
Total and weighted average	Isolated CL	305	2	0.66	1.08

IMR = infant mortality rate; CL/P = cleft lip and/or palate; CP = cleft palate; CL±P = cleft lip with or without cleft palate; CLP = cleft lip and palate.

†

No control infant mortality rate (IMR) reported by study; general control population IMR obtained from United Nations Data for country and time period in consideration.

The individual sets of collected IMR data for each OC type were then meta-analyzed, comparing observed mortalities in OC case populations to the calculated mortality figure of a reference population of 1,000,000, based on the respective standard population IMR. The standardized, large reference population ensures reported effects are due to the cleft case group findings. Although it is also possible that potential bias may exist if characteristics such as socioeconomic status, gender, or education level vary between the cleft population and the reference population used, the most precise reference data available has been used to limit this potential biasing factor. Additionally, only two of the studies reported data from actual control group cohorts. This reference population standardization also helps prevent possible biasing against these studies. Due to the differing base IMRs, total ORs are more reliable figures for analysis than the aggregated case IMRs. Random effects ORs with 95% confidence intervals (CIs) were calculated according to the method described by DerSimonian and Laird (1986). Given the heterogeneity of the listed studies and the varied study settings, the random effects total OR has been selected as the most suitable and complete statistical analysis for this review. Furthermore, a formal I² test for heterogeneity among the meta-analyzed studies also demonstrated a high level of heterogeneity among both the isolated OC and inclusive OC analyses, further supporting the use of random effects total OR (Higgins, 2003). Sensitivity analyses investigating the effects of excluding the highest OR, lowest OR, or both showed unremarkable variance in each case. Internal sensitivity analysis was also performed, examining the effect of excluding studies with specific characteristics (i.e., case ascertainment technique, country of study, study time period, different relative mortality rates). This also revealed no notable variances.

The meta-analyses for inclusive CL/P and isolated CL/P have also been visually represented in forest plots (see Fig. 1). The results of the meta-analyses performed for each subcategory of oral cleft type are listed in Table 4. Studies of IMRs associated with CL and CLP have not been meta-analyzed due to the limited data, but this IMR data is included in Tables 2 and 3.

Figure 1

Meta-analysis: IMR of inclusive oral cleft cases.

Table 4

Total Infant Mortality Odds Ratios for Oral Cleft Cases^*

Diagnosis	Meta-Analysis OR (95% CI)
Inclusive CL/P	9.466 (6.153–14.560)
Inclusive CP	15.247 (11.406–20.383)
Inclusive CL±P	11.142 (7.662–16.201)
Isolated CL/P	2.073 (1.390–3.091)
Isolated CP	2.439 (1.473–3.364)
Isolated CL±P	2.212 (1.274–3.841)

OR = odds ratio; CI = confidence interval; CL/P = cleft lip and/or palate; CP = cleft palate; CL±P = cleft lip with or without cleft palate.

Additionally, due to the significance of associated malformations in the IMRs of oral cleft cases, data on the prevalence of additional malformations in OC cases in the reviewed publications has also been consolidated in Table 5 in order to provide a multifaceted, though incomplete, assessment of this relationship. Statistical analyses were aided by MedCalc for Windows, version 11.3.4 (MedCalc Software, Mariakerke, Belgium).

Table 5

Study Results: Additional Malformations Associated With Oral Cleft Cases^*

Author, Year	Diagnosis	Overall Cases	Cases With Additional Malformation	Cases With Additional Malformation (%)
Christensen, 2004	CL/P	6394	1063	16.62
Druschel, Hughes, and Olson, 1996	CL/P	2786	984	35.32
Forrester and Merz, 2003	CL/P	502	219	43.63
Hujoel, Bollen, and Mueller, 1992	CL/P	384	77	20.05
Nembhard, Waller, Sever, and Canfield, 2001	CL/P	595	91	15.29
Total and Weighted Average	CL/P	10,661	2434	22.83
Druschel, Hughes, and Olson, 1996	CP	1187	517	43.56
Forrester and Merz, 2003	CP	181	101	55.80
Hujoel, Bollen, and Mueller, 1992	CP	130	43	33.08
Nembhard, Waller, Sever, and Canfield, 2001	CP	196	45	22.96
Total and weighted average	CP	1694	706	41.68
Druschel, Hughes, and Olson, 1996	CL±P	1599	467	29.21
Forrester and Merz, 2003	CL±P	321	118	36.76
Hujoel, Bollen, and Mueller, 1992	CL±P	254	34	13.39
Nembhard, Waller, Sever, and Canfield, 2001	CL±P	339	46	13.57
Total and weighted average	CL±P	2513	665	26.46

CL/P = cleft lip and/or palate; CP = cleft palate; CL±P = cleft lip with or without cleft palate.

Analysis of Results

Inclusive OC-IMR Studies

According to the published studies considered in this review, the reported IMRs of inclusive OC cases range from 56 to 139 infant deaths per 1000 live births. The weighted average of these IMRs is 12.72%, compared with a weighted average standard population IMR of 2.11% (see Table 2). The meta-analysis yielded a total OR of 9.466 (95% CI, 6.153–14.560; see Table 4). Given the rare outcome in consideration, infant mortality, the OR approaches the relative risk (McNutt, 2003). Therefore, it indicates that infants with OCs in these developed study settings have approximately nine times the risk of dying before the age of 1, compared with the overall populations. These figures have been largely influenced by the Mackeprang and Hay study (1972), which found a relatively high mortality rate but also reported a relatively higher standard population IMR. Among these studies, there was one substantial outlier in the lower end (Richards, 1973). This study by Richards reported an infant mortality relative risk for OC cases compared with the overall population nearly one half the next lowest finding. As previously mentioned, though, the sensitivity analysis showed little effect from excluding the outlying studies.

The studies that included breakdowns of IMRs by OC subtypes consistently showed a higher IMR in CP cases compared with CL±P cases. This was substantiated by the meta-analyses of the subtypes, which generated total ORs of 15.247 (95% CI, 11.406 to 20.382) and 11.142 (95% CI, 7.662 to 16.201) for CP and CL±P, respectively (see Table 4). Similarly, the studies that included data on the percentages of CP cases and CL±P cases with additional malformations consistently presented considerably higher incidence of additional malformations in CP cases (see Table 5). The overall averages from the studies showed that additional malformations were found in 41.68% of CP cases compared with 26.47% of CL±P cases. Neither Mackeprang and Hay (1972) nor Richards (1973) were included in the CP or CL±P analyses, which explains the raised ORs for both CP and CL±P in these analyses.

Only the Emanuel et al. study (1973) presented data divided further into CL and CLP subtypes. Although the case populations considered were small, this showed a much lower IMR in CL cases (3.40%) when compared with CLP cases (15.5%). Additionally, a higher percentage of CLP cases are affected with an additional malformation than CL cases are (Milerad et al., 1997; Harville et al, 2005).

Isolated OC-IMR Studies

Excluding OC cases with associated syndromes and additional major birth defects or malformations, the IMR of isolated OC cases is significantly lower (26 infant deaths per 1000 live births) averaged across the studies. The studies found relatively similar mortality rates in the case populations, although Forrester and Merz (2003) found no mortalities among isolated OC cases. As noted in Table 4, the calculated total OR of infant mortality in isolated OC cases is 2.073 (95% CI, 1.390 to 3.092; see Fig. 2). This shows a significant relationship between isolated OCs and raised IMRs. According to Cohen's rule-of-thumb, this would be classified as a medium-small effect (Chen et al., 2010). Again, given the relatively small probability of the outcome being considered (i.e., infant mortality), the total OR, 2.073, approximates the relative risk of infant mortality for infants with oral clefts. This means that in the developed-world setting of these studies, infants with OCs but no other malformations are approximately twice as likely to die before 1 year of age compared with the overall population.

Figure 2

Meta-analysis: IMR of isolated oral cleft cases.

The studies varied in their findings for IMRs of CP versus CL±P. Overall, the meta-analysis showed a total OR of 2.439 (95% CI, 1.473 to 3.364) for CP cases, which was slightly higher than the total OR for CL±P cases, 2.212 (95% CI, 1.274 to 3.841). Similar to the studies for inclusive OC cases, few data were divided into CLP and CL. Only the Emanuel et al. (1973) and the Tennant et al. (2010) studies broke the data down into these subtypes. They did, however, both show a much higher occurrence of infant mortality in CLP cases (3.05% on average) than in CL cases (0.66% on average).

Discussion

In regard to the inclusive OC cases, this review's results have provided a more complete epidemiological understanding of the IMRs associated with cleft lip and/or cleft palate. Bringing together the available publications on the topic, this review shows a significant odds ratio for infant mortality in OC cases compared with the general population, 9.393 (95% CI, 6.1 to 14.464). As Druschel et al. (1996b) noted, as have other included studies, cleft lip and/or palate is a marker of an at-risk group for infant mortality, usually related to additional congenital malformations and syndromes. The findings here support this claim.

This review also found the IMR associated with CP to be considerably higher than the IMR associated with CL±P. It is likely that this is closely related to the finding that there is a higher chance of infants with CP having additional malformations than those with CL±P.

Cleft lip and/or cleft palate alone is not traditionally viewed as a source of infant mortality (Druschel et al., 1996a). Whether OC cases without associated malformations exhibit an increased IMR has been debated, but this review indicates that these cases likewise exhibit a significantly raised infant mortality rate, with a calculated total OR compared with the overall population of 2.073 (95% CI, 1.390 to 3.092).

It is interesting that the OR for infant mortality associated with isolated CP was slightly higher than that for isolated CL±P as well. Furthermore, in the limited available data, CL cases, both isolated and inclusive, demonstrated exceedingly lower mortality rates than CLP cases. The premise that CL cases do not exhibit increased mortality rates might explain the higher IMR associated with isolated CP compared with isolated CL±P, because CL±P IMRs would be notably lowered by the inclusion of CL. This could also partially account for why inclusive CP cases exhibit higher IMRs than inclusive CL±P, in addition to the theory that this is due to increased probability of CP cases having additional malformations. From these points, it could be inferred that possessing a CL±P is associated with a raised IMR. More extensive research would need to be performed on the IMRs associated with isolated CLP cases versus isolated CL cases to substantiate this theory.

This leads to the next essential question: Why do infants with cleft lip and/or palate, even without additional malformations, face an increased infant mortality risk? What are the underlying causes of mortality in infants with OCs? In regard to OC cases with associated malformations, Christensen and colleagues’ (2004) study of the long-term survival of OC cases investigated the underlying causes of mortality. They found that OC cases had an increased risk associated with all major causes of death, which therefore would suggest that the reasoning for this amplified mortality risk is multifactorial. The highest reported cause of infant death in OC cases was “cleft lip and palate,” which was a somewhat ambiguous category. Upon following up on these cases, 28% of the mortalities were due to pneumonia and bronchopneumonia. Christensen et al. (2004) asserted that one possible cause of increased mortality in OC cases without associated malformations were undiagnosed structural brain anomalies, which would predispose infants with OCs to seizures, another elevated cause of mortality. Certain studies also found excess risk in OC cases for certain cancers, although mostly adult cancers (Zhu, 2002; Bille, 2004), but Christensen and colleagues’ (2004) study did not support this finding. With the increased risk of infant mortality among OC cases, with and without additional malformations, it will be important to accumulate more information on the underlying causes so they can be anticipated and prevented. Several factors need additional research in order to understand not only how often they are associated with OCs but also what role they might play in increased mortality risk. Understanding the specific causes of mortality in infants with OCs will enable health care professionals and organizations to develop enhanced, guided interventions to prevent mortality and lower the IMR overall.

Confounding Factors and Bias

It is important to recognize that several confounding factors and biases, both within the literature review and within the original studies, might have affected this analysis. Methodology and eligibility criteria designed in line with the PRISMA guidelines were formed to control possible systematic biases of the literature review and meta-analysis processes. Issues common to meta-analyses were observed, though. Because the topic is rather narrow with small case populations, only nine usable studies were available and some of these had relatively small data sets, as few as 78 cases when analyzing CL cases alone (Emanuel et al, 1973). This review found these data sets to effectively correspond with one another, though. For example, CP case IMRs were greater than CL±P case IMRs, and raised rates of infant mortality were found in isolated OC cases in all but one small study. By combining the data sets, we were able to generate stronger, more robust statistics on the subject as well as to determine significance that was not necessarily clear previously.

Another issue in this meta-analysis was that these studies were performed on separate populations, although entirely in developed country settings. Random effects ORs were used to balance this concern. Also, the agreement among these data sets illustrates the congruence and comparability among the developed settings of these studies.

One major advantage of this review is that it easily circumvents publication bias (i.e., the “file drawer problem”), a common bias affecting meta-analyses where insignificant findings remain unpublished and therefore overlooked in the review. Unlike other meta-analyses, many of the studies included in this review originally did deem their findings regarding isolated OC cases as insignificant (e.g., Emanuel et al., 1973, and Druschel et al, 1996a), and the significance has only been recognizable by the compilation of these small data sets. Also, several of the articles were not focused mainly on OC case mortality rates but rather on other specific findings (e.g., Richards, 1973, and Tennant et al., 2010). As a result, the effect of publication bias in this review would be minimal.

Various factors might also have confounded the individual studies reviewed here. One point to take into consideration in respect to the reported IMRs, specifically those inclusive of OC cases with additional malformations, is that these rates have likely been artificially lowered by elective pregnancy terminations. Due to the growing capacity of health care professionals to prediagnose congenital anomalies in utero, the practice of elective pregnancy termination for fetuses with major malformations has expanded in recent years and, therefore, has artificially lowered IMRs among populations with congenital anomalies (Bourke, 2005), including nonisolated OC cases. Estimates of the accuracy of the detection rate of CL±P vary widely, from 18% to 75%, although the accuracy of the detection of CL increases when associated anomalies are present (Demircioglu et al., 2008). Because this analysis looks at studies across several decades, it is important to note that there has been a marked increase in the accuracy of detection since the early 1980s. For example, one study found the detection rate of CL±P to be only 5.3% between 1979 and 1988 compared with 26.5% between 1989 and 1998 (Stoll, 2003). It is also important to note that this analysis is inclusive of several studies that were conducted before or concurrently with the point in time when laws legalizing abortion were implemented, specifically 1973 for Denmark, 1967 for the U.K., and 1973 for the United States. This fact combined with the increasing rate of detection by ultrasound may have had some effect on our study. The rate at which mothers choose to terminate a pregnancy after the detection of a cleft varies widely from study to study. Several studies show that without an associated malformation, mothers do not choose to terminate their pregnancies. One review of termination after an isolated cleft diagnosis showed nine studies with rates varying from 0% to 92%, the latter being a clear outlier with the next closest rate of termination being 27% (Johnson and Sandy, 2003). Decisions to terminate a pregnancy after a cleft diagnosis also vary by culture (Johnson and Sandy, 2003). Given the variety of factors that go into making a decision to terminate a pregnancy, such as the existence of laws legalizing abortion, the sensitivity of ultrasound detection, the presence of additional malformations, and cultural values, this inextricably remains a possible confounding factor in this review. Conversely though, an appreciable variation in OC case IMRs was not noted across the wide time period under review, and therefore the effect of this in these studies was most likely minor.

One other possible confounding factor is low birth weight. Several studies have shown an association between OCs and low birth weight, but the correlation between low birth weight and increased risk of mortality specifically among infants with OCs is presently unclear (Lilius and Nordstrom, 1992; Becker et al, 1998; Wyszynski et al., 2003). More targeted research would be necessary to discern any effect of this factor.

In Wyszynski's review of the issues affecting OC studies, he cited the underreporting of associated anomalies in OC cases as another possible confounding factor in OC research (Wyszynski, 2006). It is possible that certain defects have been overlooked in the examinations of OC mortalities—for example, the structural brain anomalies that Christensen et al. (2004) cited as possible factors in mortalities of isolated OC cases. On the other hand, if underreporting of associated malformations were significant, we would expect the IMR related to isolated OC cases to drop substantially across the 60-year time period in consideration, as malformation diagnostic tools progressed. These tools allow health care professionals to better assess whether there was an associated malformation, decreasing the percentage of overlooked additional malformations. Therefore, fewer cases would be misidentified as “isolated OC cases” in later studies, which had hypothetically raised IMR rates in earlier isolated OC research. One would then expect the observed IMRs in isolated OC cases to fall as the time period of the study advanced. This again, however, has not been observed in this review.

As in most retrospective studies, the methods of case and outcome ascertainment in the individual studies allow possible biasing. As listed in Table 1, the studies used different techniques to establish OC case populations and to determine the number of deaths in each population. As previously mentioned, birth certificate reporting is usually the least reliable source for birth defect statistics, as used in the Hujoel et al. study. As this misreporting is assumedly random, though, it is unlikely that this would bias these results. On the other hand, many of the studies (Emanuel et al., 1973; Richards, 1973; Tennant et al., 2010) used hospital records as a data source for OC cases, which might bias studies by omitting infants born outside of hospitals. The actual effect of this, although very small, cannot be determined. Last, unlike following patients across their lifetime, retrospective follow-up techniques, such as exploring a state's death certificates, are susceptible to losing cases over the study period. Because survival is the assumed outcome, this would result in underreporting infant death outcomes. Only Christensen et al. (2004) followed patients across their lifetimes. This, therefore, would likely bias our results, but toward the null hypothesis that IMRs and OCs are not associated.

One shortcoming of this review to note is that the included studies have been performed in similarly developed, high-income settings only. This is because the available published research relating OCs and IMR has been carried out only in high-income countries. Whereas, it made the data from the selected publications more comparable and therefore better suited for meta-analysis, it also considerably limits the generalizability of the results. Some authors have claimed a clear relationship between OCs, regardless of additional malformations, and IMR in the developing world due to feeding difficulties and a higher susceptibility to infection among infants with cleft lip and/or palate (Hodges et al., 2009). With infection and malnutrition being two of the leading causes of infant mortality around the world (Rosano et al., 2000), this seems obvious. Furthermore, risks associated with congenital malformations are amplified in developing-world settings (Rosano et al., 2000).

One large study on this subject has recently been conducted in China (Dai et al., 2010). This study reported that 6.46% of liveborn infants with OCs without additional malformations die within the first 7 days of life (401 early neonatal mortalities of 6206 liveborn infants with OCs). Additionally, 32.15% of liveborn infants with OCs who have additional malformations die within the first 7 days of life (235 early neonatal mortalities of 731 liveborn infants with OCs). This can be compared with a reported 0.37% standard neonatal mortality rate among the total sample population (17,990 mortalities out of 4,843,080 live births). According to this study's findings, cases with isolated CL and CP, compared with the overall population, have a 17.46 relative risk of early neonatal mortality in China.

There is an outstanding need for further research regarding mortality associated with oral clefts in developing countries. More information on raised IMRs in infants with cleft lip and/or cleft palate, as well an understanding of the causes associated with the amplified risk of these conditions, would enable governments and aid organizations to better develop and implement intervention strategies to diminish this risk and in turn lower IMRs throughout the developing world.

Conclusion

Overall, it is clear from our review that persons with cleft lip and/or palate are an infant mortality at-risk group, largely due to the presence of additional malformations, having nine times the risk of mortality in the first year of life than the overall population. Additionally, although their rate is lower than the inclusive OC infant mortality relative risk, infants with isolated OCs also exhibit a significantly increased risk of infant mortality and are approximately twice as likely to die during the first year of life than the overall population. Currently, the causes of this raised mortality rate in isolated OC cases are not fully understood.

Current thought and Dai's study in China point to an even more highly increased IMR among OC cases in the developing world. Conducting additional research in developing countries is critical to provide a more complete picture of the relationship between infant mortality and OCs throughout the world.

Footnotes

Acknowledgments

The authors would like to acknowledge Dr. Bill Magee, Dr. Randy Sherman, and Dr. Luis Bermudez for their input and advice throughout this review, as well as Ellen Agler, Allison Bradshaw, and Shannon Mintz for their overall guidance and assistance. The authors would also like to offer special thanks to Operation Smile, Inc., for the organization's support during this project.

References

Agha

M.M.

Williams

J.I.

Marrett

Dodds

Determinants of survival in children with congenital abnormalities: a long-term population-based cohort study.

Birth Defects Res A Clin Mol Teratol. 2006; 76: 46–54.

Becker

Svensson

Källén

Birth weight, body length, and cranial circumference in newborns with cleft lip or palate.

Cleft Palate Craniofac J. 1998; 35: 255–261.

Berry

R.J.

Buehler

J.W.

Strauss

L.T.

Hogue

C.J.R.

Smith

J.C.

Birth weight–specific infant mortality due to congenital anomalies, 1960 and 1980.

Public Health Rep. 1987; 102: 171–181.

Bille

Winther

J.F.

Bautz

Murray

J.C.

Olsen

Christensen

Cancer risk in persons with oral cleft—a population-based study of 8093 cases.

Am J Epidemiol. 2005; 161: 1047–1055.

Bourke

Bower

Charles

Knuiman

Effects of terminations of pregnancy for fetal abnormalities on trends in mortality to one year of age in Western Australia.

Paediatr Perinat Epidemiol. 2005; 19: 284–293.

Centers for Disease Control and Prevention Office of Minority Health & Health Disparities. Eliminate disparities in infant mortality. Available at http://www.cdc.gov/omhd/amh/factsheets/infant.htm. Accessed March 1, 2011.

Centers for Disease Control and Prevention. Facts about birth defects. Available at http://www.cdc.gov/ncbddd/birthdefects/facts.html. Accessed June 25, 2011.

Chen

Cohen

Chen

How big is a big odds ratio? Interpreting the magnitudes of odds.

Commun Stat Simulation Comput. 2010; 39: 860–864.

Christensen

Juel

Herskind

A.M.

Murray

J.C.

Long term follow up study of survival associated with cleft lip and palate at birth.

BMJ. 2004; 328(7453): 1405.

10.

Dai

Zhu

Mao

Deng

Wang

Liang

Tang

Wang

Kilfoy

B.A.

. Time trends in oral clefts in Chinese newborns: data from the Chinese National Birth Defects Monitoring Network. Birth Defects Res A Clin Mol Teratol. 2010; 88: 41–47.

11.

Dastgiri

Gilmour

W.H.

Stone

D.H.

Survival of children born with congenital anomalies.

Arch Dis Child. 2003; 88: 391–394.

12.

Demircioglu

Kangesu

Ismail

Lake

Hughes

Wright

Sommerlad

B.C.

Increasing accuracy of antenatal ultrasound diagnosis of cleft lip with or without cleft palate, in cases referred to the North Thames London Region.

Ultrasound Obstet Gynecol. 2008; 31: 647–651.

13.

DerSimonian

Laird

Meta-analysis in clinical trials.

Control Clin Trials. 1986; 7: 177–188.

14.

Druschel

C.M.

Hughes

J.P.

Olsen

C.L.

First year-of-life mortality among infants with oral clefts: New York state, 1983–1990.

Cleft Palate Craniofac J. 1996a; 33: 400–405.

15.

Druschel

Hughes

J.P.

Olsen

Mortality among infants with congenital malformations, New York state, 1983–1988.

Public Health Rep. 1996b; 111: 359–365.

16.

Emanuel

Culver

B.H.

Erickson

J.D.

Guthrie

Schuldberg

Further epidemiological differentiation of cleft lip and palate: a population study of clefts in King County, Washington, 1856–1965.

Teratology. 1973; 7: 271–281.

17.

Forrester

M.B.

Merz

R.D.

First-year mortality rates for selected birth defects, Hawaii, 1986–1999.

Am J Med Genet A. 2003; 119A: 311–318.

18.

Hagberg

Larson

Milerad

Incidence of cleft lip and palate and risks of additional malformations.

Cleft Palate Craniofac J. 1998; 35: 40–45.

19.

Harville

E.W.

Wilcox

A.J.

Lie

R.T.

Vindenes

Åbyholm

Cleft lip and palate versus cleft lip only: are they distinct defects?

Am J Epidemiol. 2005; 162: 448–453.

20.

Hodges

Wilson

Hodges

Plastic and reconstructive surgery in Uganda—10 years experience.

Paediatr Anaesth. 2009; 19: 12–18.

21.

Hujoel

P.P.

Bollen

A.M.

Mueller

B.A.

First-year mortality among infants with facial clefts.

Cleft Palate Craniofac J. 1992; 29: 451–455.

22.

Johnson

Sandy

J.R.

Prenatal diagnosis of cleft lip and palate.

Cleft Palate Craniofac J. 2003; 40: 186–189.

23.

Lilius

G.P.

Nordstrom

R.E.

Birth weight and placental weight in cleft probands.

Scand J Plast Reconstr Surg Hand Surg. 1992; 26: 51–54.

24.

Mackeprang

Hay

Cleft lip and palate mortality study.

Cleft Palate J. 1972; 9: 51–63.

25.

McNutt

L.A.

Xue

Hafner

J.P.

Estimating the relative risk in cohort studies and clinical trials of common outcomes.

Am J Epidemiol. 2003; 157: 940–943.

26.

Milerad

Larson

Hagberg

Idaberg

Associated malformations in infants with cleft lip and palate: a prospective, population-based study.

Pediatrics. 1997; 100: 180–186.

27.

Moher

Liberati

Tetzlaff

Altman

D.G.

, The PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 2009; 6: e1000097. doi: 10.1371/journal.pmedl000097.

28.

Mossey

Castillia

Global Registry and Database on Craniofacial Anomalies. Geneva: World Health Organization; 2003.

29.

Mossey

Little

Munger

R.G.

Dixon

M.J.

Shaw

W.C.

Cleft lip and palate.

Lancet. 2009; 374: 1773–1785.

30.

Nembhard

W.N.

Waller

D.K.

Sever

L.E.

Canfield

M.A.

Patterns of first-year survival among infants with selected congenital anomalies in Texas, 1995–1997.

Teratology. 2001; 64: 267–275.

31.

Ngai

C.W.

Martin

W.L.

Tonks

Wyldes

M.P.

Kilby

M.D.

Are isolated facial cleft lip and palate associated with increased perinatal mortality? A cohort study from the West Midlands Region, 1995–1997.

J Matern Fetal Neonatal Med. 2005; 17: 203–206.

32.

Petrini

Damus

Johnston

Jr . Trends in infant mortality attributable to birth defects—United States, 1980–1995. Morbid Mortal Wkly Rep. 1998; 47: 773–778.

33.

Richards

I.D.G.

Fetal and infant mortality associated with congenital malformations.

Br J Prev Soc Med. 1973; 27: 85–90. Richards IDG, Lowe CR. Incidence of congenital defects in South Wales, 1964–6. Br J Prev Soc Med. 1971;25: 59–64.

34.

Richmond

Atkins

A population-based study of prenatal diagnosis of congenital malformation over 16 years.

BJOG. 2005; 112: 1349–1357.

35.

Rosano

Botto

L.D.

Botting

Mastroiacovo

Infant mortality and congenital anomalies from 1950 to 1994: an international perspective.

J Epidemiol Community Health. 2000; 54: 660–666.

36.

Stoll

Alembik

Dott

Roth

M.P.

Associated malformations in cases with oral clefts.

Cleft Palate Craniofac J. 2000; 37: 41–47.

37.

Tennant

P.W.G.

Pearce

M.S.

Bythell

Rankin

20-year survival of children born with congenital anomalies: a population-based study.

Lancet. 2010; 375: 649–656. Available at http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(09)61922-X/fulltext. Accessed July 10, 2010.

38.

Thong

M.K.

J.J.

Khatijah

N.N.

A population-based study of birth defects in Malaysia.

Ann Hum Biol. 2005; 32: 180–187.

39.

UNData–United Nations Statistics Division. Infant mortality rate per 1000 live births. Available at http://data.un.org/Data.aspx?d=PopDiv&f=variableID:77#PopDiv. Accessed August 10, 2010.

40.

Wehby

G.L.

Castilla

E.E.

Goco

. Description of methodology used in an ongoing pediatric care interventional study of children born with cleft lip and palate in South America. BMC Pediatrics. 2006; 6: 9. doi: 10.1186/1471-2431-6-9.

41.

World Bank. Data: Countries and economies. Available at http://data.worldbank.org/country. Accessed July 21, 2010.

42.

Wyszynski

D.F.

Sárközi

Czeizel

A.E.

Oral clefts with associated anomalies: methodological issues.

Cleft Palate Craniofac J. 2006; 43: 1–6.

43.

Wyszynski

D.F.

Sárközi

Vargha

Czeizel

A.E.

Birth weight and gestational age of newborns with cleft lip with or without cleft palate and with isolated cleft palate.

J Clin Pediatr Dent. 2003; 27: 185–190.

44.

Zarakauskaité

Matulevičiené

Utkus

Kučinskas

Linkevičiené

Orofacial clefts with associated anomalies in Lithuania.

Acta Medico Lituanica. 2007; 14: 17–23.

45.

Zhang

K.F.

What's the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes.

JAMA. 1998; 280: 1690–1691. doi: 10.1001/jama.280.19.1690.

46.

Zhu

J.L.

Basso

Hasle

Winther

J.F.

Olsen

J.H.

Olsen

Do parents of children with congenital malformations have a higher cancer risk? A nationwide study in Denmark.

Br J Cancer. 2002; 87: 524–528.