The latter two polymorphisms and their haplotypes modify the down-regulation of the β2-adrenergic receptor upon agonist stimulation. In a previous study [14], the Gly16/Gln27 haplotype underwent enhanced down-regulation compared with the Arg16/Gln27 haplotype, while the Arg16/Glu27 haplotype showed almost no down-regulation. Many population-based studies have been performed to investigate the modifying potency of these polymorphisms on cardiovascular risk factors in vivo, but the results have been inconsistent or contradictory.

Previously, a significant negative influence of the severity of OSA on BP, prevalence of arterial hypertension, heart rate, and serum levels of high-density lipoprotein cholesterol (HDL-C) was shown in the parent study cohort of the patients [4, 7] included in this report. In the present study, we scrutinised the modifying effects of β2-adrenergic receptor polymorphisms on these risk factors, which are mediated by elevated sympathetic activity. We further investigated whether the three polymorphisms have an effect on the comparatively high prevalence of coronary heart disease (CHD) and survived myocardial infarction (MI) in our patients with OSA.

Allele frequencies of the different polymorphisms were: Arg-47Cys = 0.53/0.47, Arg16Gly = 0.37/0.63 and Gln27Glu = 0.49/0.51. The β2-adrenergic receptor polymorphisms Arg-47Cys and Gln27Glu were found to be strongly concordant, generating the haplotypes -47Cys/27Glu (83.5% of -47Cys homozygotes are 27Glu homozygotes). The five most common haplotypes accounted for 73.2% of all observed haplotypes.

Use of antihypertensive or lipid-lowering medication was distributed equally in the genotype subgroups and consisted of angiotensin-converting enzyme inhibitors (n = 128; 29.8%), β-receptor blockers (n = 87; 20.3%), diuretics (n = 70; 16.3%), calcium-channel blockers (n = 86; 20.0%), nitrates (n = 41; 9.6%) and α-receptor blockers (n = 1; 0.2%). Of the 40 patients on lipid-lowering medication, 37 were using statins and 3 fibrates. Total cholesterol and triglycerides measurement was available for 360 patients, and lipoprotein determination for 278 patients.

The mean ± SD AHI in our study group was 29.1 ± 23.1 events per hour, and the mean minimal nocturnal oxygen saturation was 81.0 ± 10.9%. Linear and independent relationships between the AHI and daytime systolic and diastolic BP, daytime heart rate, the prevalence of hypertension, and HDL-C serum levels have been shown in this study group previously [4]. Because DNA tests could not be performed for all members of the parent study cohort (n = 540), we performed a new, adjusted regression for our subgroup of 429 subjects [7]. Multivariate linear or logistic regressions revealed a significant influence of the severity of OSA on daytime systolic BP (β = 0.129, P = 0.008), diastolic BP (0.129, P = 0.012), heart rate (β = 0.197. P < 0.001) prevalence of hypertension (β = 0.026, P = < 0.001), and HDL-C serum levels (n = 78, β = -0.146, P = 0.003)[7]. This influence was independent of diabetes, BMI, age, sex, or antidiabetic, lipid-lowering, or antihypertensive treatments.

As shown in Table 2, the β2-adrenergic receptor polymorphisms did not modulate these relationships (Table 2; hypertension is shown in Table 1).

Of the 429 patients with OSA, 27 (6.3%) had survived a MI. Patients heterozygous for the Arg-47Cys and Gln27Glu polymorphisms showed a lower rate of survived MI than those with the homozygous forms. There was a trend towards a lower rate of survived MI in patients heterozygous for the Arg16Gly polymorphism (Table 1).

We analysed β2-adrenergic receptor polymorphisms in a study group of untreated patients with OSA in whom a linear and independent relationship between the frequency of nocturnal apnea/hypopnea and daytime systolic and diastolic BP, daytime heart rate, prevalence of arterial hypertension, and HDL serum levels were shown previously [4, 7]. The elevated sympathetic nerve activity in patients with OSA [1–3] is suggested to mediate these effects, especially in relation to the cardiovascular system.

The underlying mechanisms between the severity of OSA and modulation of lipid serum level have been discussed in detail elsewhere [7]. The fact that we could not find any influence despite the sympathetic pathway supports the notion that the effect of these polymorphisms in vivo is minor, and confirms previous negative results [15]. Kim et al investigated 195 Korean subjects and found no difference between β2-adrenergic receptor genotypes (Arg16Gly and Gln27Glu) in relation to plasma glucose, fasting serum insulin, LDL, HDL, or triglyceride levels. However, in another study of 19 healthy individuals with propanolol-induced dyslipidemia, a significantly decreased HDL serum level was found in subjects homozyous for Gln27 homozygotes, and raised triglyceride levels in subjects homozygous for Glu27 [16].

The genetic role of the β2-adrenergic receptor gene in the development of hypertension has been the subject of many population-based studies. The distal segment of chromosome 5 (5q31.1-qtr), which harbors the β2-adrenergic receptor gene, has been linked both to systolic and to post-exercise diastolic BP in genome-wide scans performed in large study populations [17, 18]. In a large sample of 638 individuals from 212 Polish pedigrees, maximum linkage to essential hypertension and adjusted systolic and diastolic BP was found for a marker in close proximity to the β2-adrenergic receptor gene. However, no significant association of the Arg16Gly and Gln27Glu polymorphisms was detected with these parameters in the same study [19].

The published studies that have shown an influence of these polymorphisms on BP and hypertension are confusing. On the one hand, the Gly16 and Glu27 genotypes were significantly associated with the prevalence of hypertension and systolic BP in two large populations of Chinese individuals [20, 21]. On the other hand, a more than twofold increased relative risk for hypertension and higher BP was detected in homozygous Arg16 carriers with diabetes [22]. One could argue that this represents a linkage phenomenon in two different populations, but in view of the functional relevance of these polymorphisms, such contradictory results remain difficult to explain. Finally, in the most extensive and prospective population based study, the Cardiovascular Health Study, with over 5000 participants, no effect of the Arg16Gly and Gln27Glu polymorphisms on BP control, new-onset hypertension, ankle-arm index, carotide intima-media thickness, or brachial flow-mediated dilation was detected [23]. Interestingly, in this study cohort, Glu27 carriers (heterozygous and homozygous) had a significantly lower risk for coronary events, and there was a similar trend for Gly16 carriers [24]. Recently, in the same study population, it was shown that Glu27 carriers also have a significantly lower risk for sudden cardiac death [25].

In contrast to the aforementioned studies, we investigated a cohort of patients in which an OSA-induced, elevated sympathetic activity exacerbated BP, heart rate, and HDL levels. Thus, the known functional relevance of these polymorphisms in vitro would be likely to be detected in vivo in such patients. The fact that we could not detect any significant result underlines the minor effect of β2-adrenergic receptor polymorphisms in the development of these cardiovascular risk factors, and supports the results of the Cardiovascular Health Study.

Sympathetic activation also influences the risk of MI [26]. Thus, patients with OSA show a higher incidence of fatal and non-fatal cardiovascular events [9]. In our study group, with a mean age of 56 years, we found a high percentage of CHD (12.8%) and survived MI (6.3%). Although no effect was seen on the prevalence of CHD, we found a clear and independent protective effect of the heterozygous β2-adrenergic receptor genotypes on survived MI. This result again supports the hypothesis that was deducted form the Cardiovascular Health Study; namely, that variations of the β2-adrenergic receptor gene influence vulnerability to coronary events.

In the statistical model that was chosen in the Cardiovascular Health Study, carriers of the Gly variation were pooled into one group. It is not apparent whether heterozygosity alone is protective, but a protective effect of heterozygosity in the β2-adrenergic receptor has already been shown in pulmonary research. Joos et al examined 587 smokers from the Lung Health Study. Those subjects who were heterozygous at position 27 of the β2-adrenergic receptor gene had a significantly lower decline in lung function at 1-year follow-up compared with the homozygous genotypes (P < 0.001) [27]. Our results support those findings; by pooling homozygous and heterozygous individuals into two different groups, we could show a protective effect of heterozygosity on MI, independent of other risk factors such as diabetes, arterial hypertension, dyslipidemia and BMI in a logistic regression model. Interestingly, patients heterozygous at position 27 were slightly but significantly older in spite of the lower rate of MI (Table 1). This may be an additional indicator for the protective effect of this polymorphism.

As discussed previously by Loos et al, [27] a reason for the protective effect of heterozygosity may be a phenomenon called heterozygous advantage. For example, a study on the role of the secretor blood group in psoriasis showed that heterozygosity had a strong protective effect (odds ratio = 0.17), even though the allele frequencies did not differ between patients and controls [28]. A possible mechanism for heterozygote advantage is interaction between the wild-type and mutant protein isoforms in heterozygous cells. This is a plausible mechanism for receptor molecules, especially in cases where they form dimers. Heterodimeric receptors consisting of a wild-type and a mutant protein isoform may have different agonist-binding, signal-transduction or agonist-promoted desensitation properties than wild-type or mutant homodimers. Several lines of evidence indicate that β2-adrenergic receptors form dimers like other G-protein-coupled receptors. The formation of β2-adrenergic receptor dimers was shown to have functional effects on receptor-stimulated adenylate cyclase activity [29]. Furthermore, the results of a study of a mutant β2-adrenergic receptor (Cys341Gly) suggested that the dimer is the active form of the receptor [30].

We conclude that the investigated β2-adrenergic receptor polymorphisms do not modulate cardiovascular risk factors in patients with OSA-induced elevated sympathetic activity. However, heterozygosity of these polymorphisms may be protective in reducing the vulnerability to survived MI, especially if heterozygosity is present for all polymorphisms. Because the statistical haplotype with heterozygosity in all three polymorphisms was not included in previous studies, we believe that this potentially protective genetic combination should be focussed on in future.