© 2002 American Thoracic Society
Familial Predisposition and Cosegregation Analysis of Adult Obstructive Sleep Apnea and the Sudden Infant Death SyndromeDepartment of Pulmonary Medicine, Vífilsstadir, Gardabaer; Departments of Pathology and Pediatrics, University Hospitals; and deCODE Genetics, Reykjavík, Iceland Correspondence and requests for reprints should be addressed to Thorarinn Gislason, Vifilsstadir, Department of Pulmonary Medicine, 210 Gardabaer, Iceland. E-mail: thorarig{at}rsp.is
Previous studies suggest a familial link between adult obstructive sleep apnea syndrome (OSAS) and sudden infant death syndrome (SIDS). However, most of these studies were hampered by the availability of too few cases of SIDS to draw conclusions. To examine the familial nature of this association in Iceland, hospital-based lists of all patients who were diagnosed with OSAS (n = 2,350) and SIDS (n = 58) from 1979 to 1998 were used to separately determine the familial occurrence of OSAS and SIDS and to search for evidence of cosegregation of these conditions in Icelandic families, using a nationwide genealogy database. The risk ratio for a first-degree relative of a patient with OSAS was 2.0 (1.7–2.8, 95% confidence interval). The risk ratio of the more severely affected patients with OSAS was slightly higher (2.3). Likewise, the kinship coefficient (KC) for the OSAS patient group, which determines the relatedness of the patients, was significantly larger than the mean KC of 1,000 matched control groups. Estimation of the KC for the SIDS group showed a trend toward significance when compared with control groups, but after excluding one of the half-siblings in the SIDS group from the analysis, the difference did not show any trend toward significance. Although the results of the analysis of the relatedness between all patients with OSAS and infants who died of SIDS were not significant, a trend toward significance was evident when the data were separately analyzed for the more severely affected patients with OSAS. Collectively, these results demonstrate a strong familial component in OSAS and suggest that infants who died of SIDS may have shared some of the same susceptibility factors with OSAS.
Key Words: sleep apnea • sudden infant death • genetics • epidemiology
Breathing disturbances during sleep have been widely recognized during the last two decades, the most common being the obstructive sleep apnea syndrome (OSAS) (1). The main characteristics of OSAS are repeated cessations of breathing (apneas and hypopneas) during sleep, associated with decreased oxygen saturation and sleep disturbances. OSAS is found in all age groups, but the majority of patients with OSAS are middle-aged, overweight males, often with an anatomic narrowing of the upper airway (1). There have been several reports of a familial aggregation of patients with OSAS (2–10). In a large-scale study, Redline and coworkers (6) reported that among first-degree relatives of 47 subjects with laboratory-confirmed sleep-disordered breathing, the prevalence of the disorder was almost twice as that among control subjects and that the probability of sleep-disordered breathing increased progressively with increasing numbers of affected relatives. In a recent report some of the same authors describe an oligogenic inheritance of OSAS in Whites, partly due to a major gene, with a small contribution from modifying genes. The genetics of sleep apnea have recently been described extensively in a review article (10). Sudden infant death syndrome (SIDS) is a major cause of infant mortality. It is defined as the death of a baby less than 1 year of age that remains unexplained after a complete postmortem examination (11). Kahn and coworkers (12) have reported 30 SIDS cases out of 21,000 infants who had been included in a sleep study. The infants who had died of SIDS had shown more obstructive apneas during sleep than the other infants had. In a recent follow-up study on 40 infants who subsequently died of SIDS and 607 healthy infants, it was observed that the infants who had succumbed to SIDS had more frequent apneas and that they were more likely to be males, as compared with the healthy infants (13). A familial occurrence of SIDS has also been suggested (14), although much disputed (15). From the early recognition of OSAS, a resemblance to SIDS has been suggested (16, 17). Strong familial association between SIDS and OSAS was reported from Cleveland, OH, with predisposing factors to both being multiple OSAS cases, allergy, narrow upper airway, and decreased hypoxic ventilatory response (17). In another survey from Edinburgh, strong cosegregation between OSAS and SIDS was reported with a common X-ray pattern of a narrow upper airway (18). Also, by comparing the facial bone structure of 15 consecutive SIDS victims with that of 15 control infants, the same group reported that the SIDS victims have a retroposition of the maxilla more often than control infants do, and this characteristic might predispose to retropalatal upper airway narrowing (20). It is therefore possible that the two disorders share a pathophysiologic abnormality based on anatomic characteristics and/or an abnormal ventilatory control (16–20). None of the studies cited previously has been population-based, and they have usually included only a few SIDS cases and sometimes cases of different ethnicities. The consequence of the near-isolation of the Icelandic population for 1,100 years, in contrast to other European populations, is that the nation is genetically relatively homogeneous (21). This genetic and cultural homogeneity is one reason why it is convenient to work on human genetics in Iceland. Another reason is the easily accessible genealogical information. At deCODE Genetics, a computerized database has been set up containing genealogic information on the entire Icelandic population several centuries back in time (21). This facilitates determining whether a particular disease is familial and whether there is cosegregation of two or more diseases within the same families. The purpose of this study was to determine whether in Iceland (1) adult OSAS is a familial disease, (2) SIDS is a familial disease, and (3) OSAS cases cosegregate with SIDS cases.
Patients with OSAS This study is population-based and includes all Icelanders diagnosed with OSAS at the Vifilsstadir Pulmonary Department, University Hospital, from August 1987 to July 1999. With the exception of several dozen night studies performed elsewhere in Iceland during that period, individuals suspected of sleep-related breathing disturbances have been referred to Vífilsstadir. A few cases of severe OSAS diagnosed elsewhere were also referred to Vifilsstadir for treatment with continuous positive airway pressure (CPAP) and are included in our study population. Most patients were referred because of habitual loud snoring and/or daytime sleepiness. All the OSAS diagnoses are based on whole-night studies. At the beginning of the study period (1987–1990), the diagnoses were based only on oximetry and thermistor signals registered on a paper recorder, in addition to occasional full polysomnographic studies, also recorded on paper. Then came a period (1990–1995) with different types of 4-channel digitized units for respiratory recordings, but since 1995, all recordings have been based on the same 12-channel system (EMBLA system; Flaga Medical Devices, Reykjavik, Iceland) and more (but variable) parameters have been used. The OSAS diagnosis has been based on a minimum of at least 30 oxygen desaturations of more than 3% during the study night. The minimum estimated sleep time for a negative study has been 5 hours. Each patient who fulfilled the aforementioned criteria and also reported daytime impairment (usually sleepiness) was given a specific diagnostic number for OSAS and was registered as such in the hospital computerized data system. Mild and moderate cases were often referred for palatal surgery, especially at the beginning of the study period. A weight reduction program and mandibular advancement devices have also been available. Altogether 2,350 patients with OSAS were found in the hospital files and were included in the study population: 1,727 men and 623 women (Figure 1A) . The women were slightly older (52.4 ± 13.4 versus 49.4 ± 13.7 years, p < 0.001). In the case of more severe OSAS (usually more than 100 oxygen desaturations per night), the patients were hospitalized again for CPAP therapy and thereafter yearly. Among these 2,350 patients with OSAS, 886 had been hospitalized repeatedly because of severe OSAS and had been or are currently being treated with CPAP.
SIDS Cases The SIDS group consists of all cases of SIDS registered in Iceland from 1979 to 1998. In Iceland, all cases of unexpected child death are investigated. Autopsy reports and death certificates for the entire country were reviewed to ensure proper diagnosis. Cases not confirmed by autopsy were excluded. The number of live births in Iceland during 1978–1998 totaled 87,734, so the 58 SIDS cases constitute on average 0.67 cases per 1,000 births. Autopsies were performed in the same institution in all cases and by the same pathologist (J.H.J.) in 54 (93%) cases. There were 32 boys and 26 girls in the SIDS group, and their mean age at death was 4 months and ranged from 1 to 8 months (Figure 1B).
Consent
Background Information Meiotic distance is defined as the distance: one, from a child to a parent; two, between siblings; two, from child to its grandparents, etc. The meiotic distance can be used as a measure of familial relatedness.
Statistical Analysis Control groups. The control groups were matched (one to one) according to sex, year of birth, and sum of ancestors in the genealogy database five generations back. The p values reported are one-sided and empiric. If the value of the patient list exceeds all control values, a p value of less than 0.001 is reported. Risk ratio calculations. The risk ratio for relatives of affected patients was defined as the risk of OSAS in the relatives divided by the risk in the general population; this ratio is directly related to the power to identify or map susceptibility genes (26). Obtaining valid estimates of the risk ratio is not straightforward because many sampling schemes lead to biased or inflated estimates (27). The use of a population-based group of patients eliminates much of the potential sampling bias. A near-complete genealogy database facilitates identification of patients who are related to other patients. The fact that OSAS was not diagnosed in Iceland until the late 1980s and that the majority of patients are middle aged leads to relatively few very old and young individuals on the patient list (Figure 1A). In calculating the estimated risk of OSAS in relatives, we restricted our analyses to patients born during the time periods 1920–1950 or 1925–1955 and included more than 90% of the patients in the calculations (Figure 1A). Relative risk assessments were made separately for the entire OSAS group and the CPAP-treated group. The SIDS group was too small for estimation of risk ratios. As an example, the risk ratio for offspring of patients with OSAS was calculated by first estimating the risk to offspring. The patients with OSAS had 5,475 offspring in the 1925–1955 time period. Of these, 164 were affected (they were also patients with OSAS). The estimated risk to offspring is then 164/5,475 = 0.030. There were 111,447 persons born in Iceland during this time period, and 1,797 of these were patients with OSAS. Thus, the estimate of the population risk for this time period is 1,797/111,447 = 0.016. The estimate for the risk ratio is simply the ratio of these two fractions, 0.030/0.016 = 1.9 (for two significant digits). Confidence intervals are provided for the risk ratios to indicate the range of values supported by the data. They are determined by the distribution of the risk ratios calculated for the matched controls sets (25). Kinship coefficient. The kinship coefficient (KC) is a measure of the relationship between two relatives. It is defined as the probability that a randomly selected allele from each pair of individuals is inherited from a common ancestor (28). For example, with no consanguinity in previous generations, the KC is 0.25 for siblings and other first-degree relatives, 0.125 for second-degree relatives, 0.0625 for third-degree relatives, and so on, each value being half the expected fraction of the genome shared by these relatives. The reported KC is the average KC for all the pairs of patients being studied. The KC was estimated for both the entire patient group (OSAS) and the 886 CPAP-treated OSAS cases. Furthermore, the KC was estimated for both groups with the contributions of first-degree relative pairs discarded. A possible relatedness within the SIDS group was estimated by comparing their KCs with the values for 1,000 sex- and age-matched control groups. This large number of control groups was used to have as exact a distribution of the controls as possible. Because the pedigrees were extensive, the overall average KC could not be calculated exactly. We used Monte Carlo simulations to approximate the average KC for each group (patients or control subjects) and ensured that the Monte Carlo errors had a negligible effect on the reported results. Cross-risk ratio. The cross-risk ratio was used to estimate the possible relatedness between the OSAS and SIDS cases. A list of the parents of the SIDS cases was used for the calculations. The cross-risk ratio is therefore an estimation of the risk for a relative of a patient with OSAS of having a child who might succumb to SIDS. These are computed in a manner similar to the risk ratio calculations but by using the patients with OSAS as the proband (or reference relative) and computing the risk to relatives and the population risk of being a parent of a SIDS case. We could identify 113 parents for 58 SIDS cases; altogether 57 mothers and 56 fathers, as one mother had two children who had succumbed to SIDS and two fathers were missing on the list. The cross-risk ratio was estimated for first-, second-, and third-degree relatives of patients with OSAS. Confidence intervals for the cross-risk ratios are provided.
Relatedness among Patients with OSAS Relatedness was estimated among the OSAS group using the computerized genealogy database and the minimum meiotic distance found between patients (the meiotic distance has been defined previously). A cluster is a family tree in which a patient with OSAS is related to at least one other patient with OSAS in the tree. When we include relatives of patients as distant as second cousins (a meiotic distance of six), then 90% of the patients form a single multilineal cluster. To more carefully establish whether such relatedness is significant or simply due to the background relatedness of Icelanders, we computed KCs and risk ratios and tested their significance. We note that the SIDS group contained one pair of half-siblings, and four additional SIDS cases were related to each other at seven meiotic events.
Risk Ratios for Relatives of Patients with OSAS
Kinship Coefficient The KC for the patients with OSAS was 7.3 x 10-4, which is significantly larger than the mean KC of the 1,000 control groups (p < 0.001) (Table 2) . Even after excluding first-degree relations, the KC of the patients with OSAS remained significantly larger than the mean KC of the 1,000 control groups (p = 0.002). Similarly, the KC of the 886 CPAP-treated patients with OSAS was 8.1 x 10-4, which is significantly larger than the mean KC of the 1,000 control groups (p < 0.001) (Table 2).
Relatedness among SIDS Cases The KC for the SIDS cases was 9.4 x 10-4. The estimated value of the mean KC of the 1,000 control groups was 3.7 x 10-4 (p = 0.055) (Table 3) . This difference is not significant. The KC for the SIDS group remained 9.4 x 10-4 after excluding first-degree relatives because the closest relationship within the list of patients who had succumbed to SIDS was for the half-sib pair. However, the estimated mean KC for the control groups then changed to 3.2 x 10-4, compared with the earlier value of 3.7 x 10-4 because, by chance, there had been first-degree relatives in the control groups; the KC was reduced when these relations were removed, resulting in a lower KC estimate (Table 3). The difference in the KC between the SIDS cases and control groups was then significant (p = 0.022). However, when one of the half-siblings was excluded from the calculations, there was no longer a significant difference in the KC between the SIDS cases and the control groups (p = 0.601).
Familial Relationship between SIDS and OSAS The cross-risk was estimated between first-, second-, and third-degree relatives for both total number of patients with OSAS and CPAP-treated patients with OSAS against parents of SIDS victims (Table 4) . There was a significantly increased risk for relatedness between nieces/nephews of CPAP-treated patients with OSAS and SIDS victims, but other cross-risk ratios were not statistically significant (Table 4). The number of SIDS cases is small, and this affects the power of the study to detect changes. The 95% confidence interval for the risk ratio for offspring of patients with OSAS to be parents of SIDS cases indicates that the ratio could be, for example, as low as 0.8 or as high as 5.5. Values outside that range can be excluded from these data. Even in the case of strongest significance, the risk ratio for nieces/nephews of CPAP-treated patients to be parents of SIDS cases, the data do not rule out the ratio being as small as 1.1 or as large as 3.5.
DISCUSSION The results of this study indicate that the Icelanders we have diagnosed so far with OSAS show a strong familial relationship. There is a marginal evidence for an increased relatedness within the SIDS group compared with control groups and a trend toward significance for nieces/nephews of patients more severely affected with OSAS and SIDS cases to be related to each other than accounted for by chance. The small sample size of "only" 58 SIDS cases, however, makes the results inconclusive for SIDS. Although genetic factors are most likely responsible for the observed aggregation of patients with OSAS in Icelandic families, the latter issue warrants further discussion. Because the general knowledge about OSAS is probably higher in families where an OSAS case has already been diagnosed and treated, a selection bias could occur. This might cause a higher referral rate among relatives of patients with OSAS than among others. This might have occurred especially at the beginning of the sleep studies in Iceland (1987) when OSAS was less well known, both in the society in general and among referring physicians. We have, however, already diagnosed OSAS among a high proportion of the total population. In some age groups, such as in males 35- to 65-years old, the number we have diagnosed is comparable to reported OSAS prevalence in large epidemiologic studies (1). We have currently diagnosed 1,500 patients per total population of 51,500 in this age group, which is 2.9% (Figure 1A). Therefore, it is unlikely that the familial awareness of OSAS accounts for more than a fraction of the familial aggregation that we found. Whereas the SIDS group is well defined and includes all cases diagnosed in Iceland over a period of 20 years, the OSAS group is based on those who seek help themselves or are referred by their doctors and is clearly different from a theoretic group identified through a population-based screening. Among the randomly chosen control subjects, there are likely to be undiagnosed patients with OSAS. Our waiting list for a sleep study of almost 400 individuals adds support to this assertion. This would, however, lead to an underestimate of the family association and cannot explain our findings. Our data confirm what has been described in previous studies on other populations (10). On the basis of our findings, we draw the conclusion that patients with OSAS should be asked about OSAS among relatives and that the level of suspicion should be high. In addition, further studies in this field are warranted, first of all to determine whether or not patients with OSAS share alleles of particular genes. Second, such studies could result in more specific treatment possibilities, e.g., through early life intervention of the upper airway by dental appliances or later in life by pharmacologic treatment. The patients with OSAS in this study may have genetically determined anatomic factors in common that are known to predispose to OSAS such as overweight, micrognathia, and narrow upper airways (5, 6). These factors could possibly explain some of the familial aggregation of OSAS, although they might be basically more important in the pathogenesis of OSAS unless occurring together with other abnormalities like impaired ventilatory response or other still unknown mechanisms within the central nervous system. Recently, an impaired respiratory response to resistive loading during sleep was described in healthy offspring of patients with OSAS (29). In another study, OSAS was related to upper airway dysfunction due to pharyngeal neuropathy (30). In a small-scale Japanese study, the human leukocyte antigen A2 was more common among patients with OSAS than among control subjects (31). The possible role of the HLA system in the pathogenesis of OSAS remains to be studied. The finding that SIDS cases are somewhat more related to each other than expected by chance is based on a pair of half-siblings sharing the same mother. Whether this observation warrants taking special measures is open to discussion as other reports have not found an increased mortality rate among siblings of SIDS victims as shown in a recent article (32). Previous studies (16–18) that have examined the familial relatedness between OSAS and SIDS cases are only based on a handful of near-miss sudden infant death children and/or SIDS cases. These studies have reported the high prevalence of OSAS among their close relatives. Recently, it has also been reported that infants of families with multiple histories of SIDS, apparent life-threatening events, and OSAS are more likely to have OSAS than infants of families with only one case of SIDS or apparent life-threatening events (33). Although the results from the association between OSAS and SIDS in our study concur with these observations and are of borderline significance for at least one test, the number of cases is too small to allow us to draw conclusions. The importance of the cosegregation of SIDS and OSAS is yet to be shown in the apparently multifactorial syndrome we call SIDS. Previous studies have indicated impaired ventilatory responses to hypoxia (17) or narrow upper airway as possible common denominators (16, 20). Thus, OSAS and SIDS might be having the same predisposing factors. Another highly speculative mechanism might be that OSAS induces changes such as gastroesophageal reflux (34), and the acid in the esophagus could lead to apnea (34, 35). Sudden infant death might ensue when the apnea is not terminated, perhaps due to insufficient arousal response. Although genetic factors are clearly important, other factors may also contribute to the familial clustering observed for OSAS and SIDS. However, little is known about the relative contributions of lifestyle-related risk factors such as eating habits, diet, physical inactivity, sleeping position, and exposure to environmental factors such as smoking, and more research is needed to better understand the mechanisms by which environmental factors modulate the expressions of OSAS susceptibility genes. Although it has been relatively well accepted that the familial aggregation of body weight may, at least in part, contribute to the familial aggregation of the sleep apnea, in a recent study of 1,560 male–male twin pairs with self-reported symptoms of snoring and daytime sleepiness, the genetic factors identified were largely independent of the genetic variance for obesity (36). A recent study from Cleveland also lends support to the notion that the genetic basis of obstructive sleep apnea is largely independent of factors that relate to body mass index (37). In contrast, little information is available on the familial recurrence and genetic basis of SIDS. Whereas many studies suggest that some SIDS cases may be attributed to undiagnosed infanticide or that infant care practices such as prone sleep position may play an important role in the susceptibility to SIDS, others have identified significant familial occurrence of SIDS cases (14). Our study demonstrates a strong familial component in OSAS and a trend toward significance between the relatedness of patients with OSAS and infants who succumbed to SIDS, suggesting that genetic factors are involved. These results are in keeping with the notion that infants who died of SIDS may have shared some of the susceptibility factors with OSAS. Received in original form July 24, 2001; accepted in final form June 14, 2002
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
bjartsson H, Stefánsson K. Protection of privacy by third-party encryption in genetic research in Iceland. Eur J Hum Genet 2000;8:739–742.
