Discussion
This study confirms the presence of classical risk factors for cardiovascular disease in adult patients with PKU, including an increased blood pressure, obesity and an atherogenic lipoprotein profile. In addition, we found an increase in resting heart rate, elevated biochemical indicators of inflammation and oxidative stress, marked endothelial dysfunction and increased vascular stiffness. Risk factors and biochemical markers (LDLc/HDLc, triglycerides, non-HDLc, HDL2c) were in part modified by dietary adherence as indicated by the current plasma Phe concentrations. These data further define the cardiovascular phenotype of adult PKU patients and suggest an important pathophysiological role for dietary adherence.
The patients in this study had higher
Phe plasma concentrations than recommended in the European guidelines [
28] and also compared to other studies [
4,
6]. Indeed, poor adherence to the strict PKU diet and consequently high Phe concentrations are a well-known problem in patients with PKU [
28]. In a recent multicenter survey, most patients achieved Phe blood concentrations < 1200 μmol/l, although there was limited consensus regarding target Phe levels for adult PKU patients [
29]. The majority of surveyed adult patients in another multicenter study admitted poor compliance with the Phe-restricted diet, which seems to reflect worsening of metabolic control with age reported in several studies [
30]. Thus, the rather high Phe plasma concentrations in our study patients most likely indicate poor compliance with dietary advice. However, Phe plasma concentrations in our study were not associated with cardiovascular measurements, including systolic or diastolic blood pressure, resting heart rate, endothelial function and PWV.
Our patient’s characteristics were comparable with other studies of adult PKU patients, not only including increased Phe plasma concentrations but also concerning dietary data and a high
BMI [
4,
29,
31]. In our patients, the total protein intake, intake of synthetic protein and the BMI (i.e. 27.6 kg/m
2) were almost identical to adult PKU patients included in a recent study of body composition of PKU patients in the US [
31] . In another large study of 236 British adult PKU patients, a similar mean BMI (26 kg/m
2) was reported, which increased with age and was significantly correlated with the Phe level [
4]. In our study, the BMI was associated with Phe plasma concentrations, confirming previous studies in adult PKU patients, while such association was not found in children with PKU [
6]. Furthermore, the BMI was associated with inflammatory markers (CRP, SAA) and non-HDLc levels; it correlated with endothelial dysfunction, but not with other cardiovascular measurements (blood pressure, resting heart rate and PWV). This is in contrast to the well-known association of BMI with systolic and diastolic blood pressure in the general population [
32].
Our study confirms several previous reports of a high prevalence of
obesity in the adult PKU population [
4,
11,
33], with 52% of the patients preobese or obese. In contrast, Rocha et al. found a much lower mean BMI (20.1 kg/m
2) and no evidence of obesity in their study of body composition of 89 patients [
34]; however, the majority of their patients was much younger (mean age 14.4 years) and had milder forms of PKU.
We found significantly higher systolic and diastolic blood pressure as well as a significantly increased resting heart rate. While some studies in children and young adults with PKU found normal or even lower
blood pressure compared to controls [
11,
35], a significantly higher blood pressure was found in obese PKU children [
7] and in a study of renal function in adolescents and young adults [
5]. To our knowledge an increased resting heart rate, a predictor of cardiovascular mortality in the general population [
36], has not been described before in PKU patients.
The
lipoprotein profile in PKU patients was significantly different from the control group. Patients had higher levels of total cholesterol and non-HDLc and a higher LDLc/HDLc ratio, while HDLc levels were lower. This pattern is widely used for the definition of dyslipidemia and generally indicates a higher risk for cardiovascular disease [
37]. Importantly, the (elevated) non-HDLc lipid fraction, but not the (decreased) HDLc, was positively associated with the BMI, which is at variance with the most common finding in the general population, which typically shows strong correlations between obesity and (elevated) triglyceride levels and (decreased) HDLc [
38]. Instead, the low HDLc levels were inversely associated with inflammation parameters and with Phe plasma levels in our patients; this seems to indicate complex influences on the lipoprotein profile in adult PKU patients.
Previous studies of
total cholesterol levels in PKU patients have shown inconsistent results. While in some studies lower cholesterol levels were measured [
9,
39], we and others found higher cholesterol levels in PKU patients [
40]. Lower HDLc levels have been described before in children and adults with PKU [
6,
41].
The
PKU diet in adults is similar to a vegetarian diet, avoiding protein mainly of animal origin, and supplemented by a synthetic Phe-free amino acid mixture [
39]. Patients have a low natural protein intake, a low fat intake and a high carbohydrate intake [
39]. A vegetarian diet is beneficial for cardiovascular health [
42] and associated with decreased levels of total cholesterol, LDLc and HDLc [
43]. Therefore, the low HDLc and HDL2c levels in our patients could be attributed at least in part to the PKU diet and in this context, the low HDLc levels and the elevated LDLc/HDLc ratio may not be interpreted as a classical cardiovascular risk factor [
39]. However, the negative associations of HDLc and HDL2c with Phe plasma levels and MDA levels indicate that HDLc lipid parameters were independently influenced not only by dietary non-adherence (higher Phe plasma concentrations) but also by higher oxidative stress.
HDL is a lipoprotein which has been shown to be inversely correlated to the risk of cardiovascular disease [
44]. Its antiatherogenic capacity derives mainly from its function to promote cholesterol efflux from cells (i.e. lipid laden macrophages) [
45].
HDL ameliorates endothelial functions and has anti-inflammatory properties. It can be classified into subtypes differing in density [
46]. HDL2c has been suggested a more adequate risk indicator for cardiovascular disease in the general population than HDLc and HDL3c [
24], although other studies could not confirm an association of CVD with HDL subclasses [
47].
HDL subtypes have not been studied previously in PKU patients. We found that the decrease in HDLc was due to low cholesterol levels in the HDL2, but not the HDL3 fraction. Studies analyzing the HDL subtypes in vegetarian populations also found a decrease in the HDL2 fraction, probably due to lower cholesterol intake and a higher polyunsaturated-to-saturated fat ratio [
48]. Thus, low HDLc and HDL2c levels in our patients were negatively correlated with Phe plasma levels, suggesting that a low adherence to the PKU diet with high Phe plasma concentrations lowered cholesterol content of the HDL2 fraction. We found an inverse correlation between MDA and HDLc as well as HDL2c in patients with PKU. MDA has been supposed to alter the HDL mediated cholesterol efflux by modifying Apo A1 and can thereby lead to dysfunctional HDL [
49].
Oxidative stress as well as high SAA can alter the antiatherogenic properties of HDL [
45,
50]. Importantly, HDL may acquire pro-inflammatory functions by enrichment with SAA [
50]. We observed a strong positive correlation between HDLc and SAA suggesting an enrichment of SAA in HDL, which could possibly indicate an altered function of HDL in PKU patients.
Beside the traditional cardiovascular risk markers, we further analyzed
oxidative stress and inflammation in PKU patients as possible mediators of endothelial dysfunction and vascular stiffness. Biomarkers included MDA, a marker of lipid peroxidation, the peroxidase MPO, and 3-NT, reflecting tyrosine oxidation by reactive nitrogen species. We found that MDA-levels, but not MPO or 3-NT, were increased in patients with PKU. This supports the results of Ercal et al. [
51], who found increased MDA levels in a PKU animal model. Furthermore Wilke et al. [
27] found higher MDA levels in children with PKU, which were reversible after selenium substitution. Selenium, as part of the antioxidatively acting glutathione peroxidase, was found to be reduced in PKU patients [
26]. However, our patients had normal selenium levels, indicating the presence of increased oxidative stress independent of the serum selenium status. These data support previous studies demonstrating oxidative stress in patients with PKU [
8,
41]. The underlying causes of increased oxidative stress in PKU are still a matter of debate, but increased Phe plasma concentrations have been found associated with several markers of oxidative stress in other studies [
52,
53]. Possible mechanisms proposed for oxidative stress in patients with PKU include the high blood concentrations of phenylalanine, which may directly induce oxidative damage but also a decreased antioxidative defense as result of the strict diet leading to a lack of macro- and micronutrients with antioxidative functions [
8,
54].
While oxidative stress is well documented in PKU patients, suggesting a high pro-inflammatory potential, markers of
inflammation have been rarely studied thus far. Deon et al., could demonstrate increased serum levels of IL-1b, IL-6 and IL-10 in a study of 7 well-controlled adolescent patients with PKU compared to controls [
10]. In our study, CRP and SAA in serum were elevated and associated with the BMI, but not with Phe levels. It is very well established that SAA plays a relevant role for HDL-functionality, endothelial dysfunction and progression of atherosclerosis [
55]. In addition, there was a significant correlation between a higher total protein intake and lower CRP, suggesting a suppressive effect of dietary adherence on inflammation. Altogether these data indicate an obesity-induced inflammatory status in the patients included in our study.
We further analyzed the vascular status and found alterations in its function and structure.
To our knowledge,
endothelial function has not been studied previously in patients with PKU. We here show a significant reduction of post-ischemic blood flow in these patients as measured by venous occlusion plethysmography, a long established and validated method [
56]. The post-ischemic flow reserve (PIFR) was reduced by 34%, indicating striking endothelial dysfunction, which was associated with the BMI, as shown in the multivariate analysis. Obesity is a well-known cause of endothelial dysfunction, mainly mediated by inflammation and oxidative stress [
57].
A decline of the elastic properties of the aorta is a result of vascular ageing and leads to an acceleration of the pulse wave. Thus, a high PWV can be interpreted as a sign of relevant arterial damage. The development of vascular stiffness in the general population is strongly promoted by hypertension, inflammation and oxidative stress [
58], but it is unknown whether PKU itself or dietary treatment of PKU may affect arterial properties by specific metabolic mechanisms. In a small study of vegetarian men, carotid intima-media thickness and distensibility and vascular stiffness measured by PWV were reduced by the vegetarian diet [
59]. In our study, patients with PKU had a significant increase in vascular stiffness, which was independent of blood pressure or BMI but associated with the serum CRP. Hermida-Ameijeiras et al. found a similar increase in PWV (classical PKU only) in a study of 41 PKU patients (age 6–50 years; mean age 23; 61% overweight, 39% obese) who had no significant changes in blood pressure, heart rate and blood lipid levels [
11]. In their study, PWV was associated with several other variables including age, BMI, central diastolic blood pressure, and median Phe plasma concentrations; however, significant predictors of PWV in a multivariate analysis were age and central blood pressure. In another study Htun et al. found an increase in carotid intima media thickness and local vascular stiffness [
60], which is in accordance with our results.
Taken together, these data indicate endothelial dysfunction and premature vascular ageing in PKU patients. Increased oxidative stress and a decrease in antioxidative activity has been demonstrated in PKU patients [
53], which along with the presence of inflammation might promote vascular damage. Although we could not find a significant association of PWV with oxidative stress, it should be noted that our observations were not targeted to comprehensively analyze oxidative stress or inflammation but were limited to few selected makers. Further limitations include the small number of patients and the lack of additional studies of validated surrogate markers for cardiovascular disease, such as echocardiography or measurement of intima-media thickness, which could not be performed in this study for logistical reasons.
In summary, this study provides evidence for an increased cardiovascular risk in PKU patients. An accumulation of traditional cardiovascular risk factors, high inflammatory and oxidative stress markers, endothelial dysfunction and vascular stiffness characterize the cardiovascular phenotype of adult PKU patients. These findings indicate the need for cardiovascular monitoring and early preventive measures against cardiovascular disease in patients with PKU.
Recent guidelines by the American Heart Association provide detailed recommendations for risk calculation and treatment of atherosclerotic cardiovascular disease (ASCVD) for the general population [
61]. For adult PKU patients, we therefore suggest to consider calculating the individual risk by ASCVD-calculator and to follow the primary prevention concept of the 2018 AHA recommendation. A statin therapy should be considered in patients with PKU with an estimated ASCVD risk higher than 5%.