Study population
This study included 19 patients aged 6 to 46 years with mild, or classical, phenylketonuria or tetrahydrobiopterin-deficient hyperphenylalaninemia, 15 patients aged 6 to 44 years with Gaucher disease type 3, and 93 controls aged 6 to 75 years. The patients’ inclusion criteria were genetical and/or biochemical prove of their disease. Participants under 6 years of age were excluded. Controls were included if they had no ocular disease and age-appropriate visual function, as well as no relevant systemic disease (e.g., neurodegeneration). Controls were recruited stratifying for age (seven subgroups with at least 10 participants were built for an even age distribution).
The study was approved by the Medical Ethical Committee of the State Chamber of Medicine of Rhineland Palatinate in Mainz, Germany (reference number 837.373.14). All persons or their parents/guardians gave their written informed consent prior to inclusion in the study. The research adhered to the tenets of the Declaration of Helsinki.
Ophthalmic examination procedure
The examination included non-cycloplegic auto-refraction measurements (NIDEK AR-360A, Nidek Co., Japan), best-corrected visual acuity testing, slit lamp biomicroscopy, and fundus examination, as well as orthoptic examination, which were published elsewhere [21,22,23]. Spherical equivalents defined as the sum of the spherical power and half of the cylindric power were used in the statistic models.
Imaging of the optic nerve head and the macula was carried out using spectral-domain (SD) optical coherence tomography (OCT) (Spectralis, Heidelberg Engineering GmbH, Heidelberg, Germany) with automatic real-time function for image averaging. We acquired a peripapillary OCT and a macular OCT. The peripapillary retinal nerve fiber layer (pRNFL) was imaged with a diameter of 12° (corresponding to 3.47 mm in the standard eye), and a standard corneal curvature of 7.7 mm. For the macular OCT, 49 horizontal single scans were acquired. After semi-automated segmentation of the retinal layers as provided by the OCT software (Heidelberg Eye Explorer version 1.10.2.0, viewing module 6.9.5.0; HEYEX, Heidelberg, Germany), all scans were assessed regarding their quality by a board-certified ophthalmologist (SH). Those with segmentation errors were corrected, or excluded in cases of poor image quality. In cases of poor data in only one sector, this sector was excluded prior to analysis; if more sectors were affected, then the complete OCT dataset of this eye was excluded.
Mean retinal thickness of 9 macular layers was used for the analysis: total retinal layer thickness of the macula, nerve fiber layer (NFL), ganglion cell layer (GCL), inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer (OPL), outer nuclear layer (ONL), outer retinal layer (ORL) (being limited by the external limiting membrane and Bruch’s membrane, this layer corresponds to the photoreceptors), and retinal pigment epithelium (RPE).
The 6-mm macular scan measurements were classified according to the ETDRS segments (“Early Treatment Diabetic Retinopathy Study” subfields). Central zone, inner ring, and outer ring with diameters of 1, 3, and 6 mm, respectively, were included in the analysis. The average of all points within the central zone (1 mm diameter) was defined as foveal thickness, the inner ring (1 to 3 mm) as parafoveal thickness, and the outer ring (3 to 6 mm) as perifoveal thickness.
Clinical data
Typical variables described for disease stage in PKU and GD3 were determined and obtained from the patient’s records. For PKU, we used first, current phenylalanine serum concentrations of each individual (mean 693 µmol/l ± 384 µmol/l); second, current tyrosine serum concentrations of each individual (mean 105 µmol/l ± 60 µmol/l); third, disease treatment (16 early-treated vs. three late-treated individuals with PKU); and fourth, disease type (mild phenylketonuria with untreated blood phenylalanine concentrations of less than 1000–1200 µmol/l) [24], classical phenylketonuria, and tetrahydrobiopterin-deficient hyperphenylalaninemia). For GD3, we used first, the modified severity scoring tool (mSST), which is based on twelve domains including horizontal gaze palsy, cranial nerve palsy, seizures and age at first seizures, cognitive ability, ataxia, tremor, spasticity, rigidity, dysphagia, dysarthria, and spinal alignment [25]. Second, we considered the phenotype severity (mild, intermediate, severe), of which intermediate phenotype was associated with homozygous L444P mutation [22]. Third/fourth, we included horizontal/vertical peak velocity of reflexive saccades in GD3 (69°/s ± 58°/s and 192°/s ± 92°/s, respectively).
Statistical analysis
Medians, interquartile ranges, minimums, and maximums were calculated for all continuous variables. For variables distributed normally, means and standard deviations were computed. For dichotomous variables, absolute and relative frequencies were computed.
To analyze the differences of retinal thickness with respect to PKU and GD3, we used linear mixed models to control for the inclusion of one and two eyes of a study participant (as random effect). A further adjustment for age, sex, and spherical equivalent was included in the statistical analysis. Full thickness measurements are more susceptible to these parameters, than single layer measurements, which is why we focused on single layer correlation analysis as follows. Spearman’s rank correlation was conducted to correlate thinned or thickened layers (e.g., GCL or OPL, respectively) with disease-specific variables of PKU (current phenylalanine and tyrosine serum concentration, early- vs. late-treated PKU, and disease type) and GD3 (mSST, phenotype severity, horizontal and vertical eye movements). The other retinal layers, which were not different to controls, or which were underrepresented in the specific subfield (fovea), were not further analyzed. Correlation coefficient rho of ≥ 0.5 was regarded a moderate correlation, correlation coefficient of ≥ 0.3 was considered a weak correlation, and < 0.3 was considered a no correlation. Statistical analysis was performed using R version 4.0.4. All p-values should be regarded as continuous parameters that reflect the level of evidence from our explorative analysis and are therefore reported exactly.