A comprehensive monocentric ophthalmic study with Gaucher disease type 3 patients: vitreoretinal lesions, retinal atrophy and characterization of abnormal saccades

Gaucher disease is the most common sphingolipidosis and is defined by the autosomal recessively inherited deficiency of lysosomal acid β-glucocerebrosidase. The gold standard for diagnosing Gaucher disease is the confirmation of deficient enzyme activity. Mutation detection in the affected GBA-gene can enhance the diagnostic accuracy [1]. However, there is a strong genetic heterogeneity of patients with similar symptoms and vice versa [2, 3]. Few genotypes are known to have a prognostic impact. The homozygote L444P allele is associated with Gaucher disease type 3 (GD3), which is the rare chronic neuronopathic form [4‐7]. The diagnosis of Gaucher disease type 3 (GD3) is challenging as specific neurologic symptoms develop at various ages [8]. Reclassification of type 1 (visceral) to type 3 (chronic neuronopathic) in later stages occurs [9]. The most common early neurologic symptoms are oculomotor disturbances [10], such as slowed and hypometric saccades [8, 10] and saccadic initiation errors [8], and esotropia [8, 11] and/or abduction deficits [8, 12‐14]. Horizontal saccades decelerate earlier, while vertical saccades follow [15‐17]. Vertical downwards saccades are known to be impaired earlier than the less affected upwards saccades [8, 15, 16, 18]. As the clinical examination of oculomotor abnormalities has a limited reproducibility, depending on patient’s cooperation as well as the investigator’s evaluation, an objective oculomotor testing has the potential to result in a more objective assessment and a better quality of data for follow-up investigation [8].

Beside neuro-ophthalmologic abnormalities, Gaucher type 3 may present posterior segment abnormalities of the eye. Specific vitreous opacities are better known from Gaucher type 1, with a prevalence of 3% [19]. The prevalence of posterior segment abnormalities in Gaucher type 3 is not yet established. Recent case reports describe preretinal accumulations of matter [20‐24]. In fact, they even may promote severe tractive retinal detachment in Gaucher type 3 [25, 26].

The purpose of this study is to propose recommendations for a set of useful ophthalmologic investigations for diagnosis and follow up as well as saccadometry parameters of Gaucher disease type 3 patients enabling a correlation to disease severity.

Inclusion criteria for participants were a biochemically and genetically diagnosed Gaucher disease type 3 and the cognitive ability to perform the investigations. Children under 6 years were excluded, because cooperation was regarded not to be sufficient for the examinations at that age. A visual acuity of better than 1.3 logMAR (at far distance) was required.

Sixteen patients were included in this study. A cohort of 100 healthy individuals at a stratified age distribution served as control.

The patients underwent an ophthalmologic and neuro-ophthalmologic examination according to a standardized investigation form. Ophthalmologic examination included: refraction, best corrected visual acuity, anterior and posterior segment examination, Spectral Domain (SD)-OCT (Spectralis OCT, Heidelberg Engineering GmbH, Heidelberg, Germany) of the optic nerve head and the macula. Only OCT scans with a quality index above 20 were analyzed. Neuro-ophthalmologic examination included: orthoptic status (stereopsis, cover test, motility, optokinetic nystagmus and clinical saccade testing). A technical saccadometry was performed using an infrared-videooculography device (see below). Saccadometry was done by a specifically trained investigator. Orthoptic examination was carried out by an experienced orthoptist in 12 cases.

The neuropathic Gaucher symptoms were assessed quantitatively by the disease specific “modified severity scoring tool” (mSST) using 12 categories [27]. Additionally, the “scale for the assessment and rating of ataxia” (SARA) including 8 items was assessed. Herein, the items ataxia and dysarthria [28] are characteristic in GD3 [29‐31] and are also part of the mSST [27]. The IQ, one of the categories of the mSST, represents the cognitive function and in Gaucher patients serves to evaluate the neurologic involvement of the disease. The wide range of cognitive impairment in Gaucher patients has been reported by several authors [11, 15, 32].

Saccades were recorded using the infrared video-oculography (VOG) device EyeSeeCam HIT (Eyeseetec, Fürstenfeldbruck), sampling at 220 Hz (every 3.6 ms) which was linked to a MacBook Pro 13″ (OS X Version 10.9.5, Apple Inc.) equipped with the EyeSeeCam software (EyeSeeCam VOG HIT System Reversion r3429 and r3444, EyeSeeTec Fürstenfeldbruck, Germany). The study participants were seated on a height-adjustable chair, facing the center of a computer monitor (Dell Technologies Inc., 19″ screen BQR-1908FPb, 1280 × 1024 resolution, 60 Hz refresh rate, 300 CD/m² display luminance, and 5 ms images building time) at a distance of 60 cm from the glabella to the monitor. The linear visual range was at 5°/15°/30° leftwards and rightwards and 5°/10°/20° upwards and downwards. Each target was a smiley icon. Prior to the measurement, a qualitative calibration was performed. Movements of the left eye were measured. To avoid alternating fixation in heterotropia, the right eye was covered in these cases.

Saccades were counted and calculated at a threshold of 100°/s. The beginning and ending of a saccade was determined at a threshold of 5°/s. Further criteria were an occurrence 0.5 s after stimulus at latest and an amplitude of more than 0.5 of the relevant stimulus. In case that standard calculation did not detect enough saccades due to pathologically decelerated saccades, a “special saccades calculation” with more sensible criteria was used (threshold of 5°/s, beginning and end at 2°/s, latest detection 0.7 s after stimulus and detection of saccades even if they measure less than 0.5 of the stimulus amplitude).

The evaluation by “special saccades calculation” was necessary in 13/14 patients horizontally and in 8/14 patients horizontally to allow the detection of a sufficient number of saccades. Thereby a minimum of 4 saccades could be included in 82.4% of the 4 direction and 3 degree (12 different eccentricities) in the 14 patients.

With 16 patients and 100 controls, it was possible to establish a difference of 0.67 standard deviation between groups at 5% level with a power of 80% using a two sample t-test. This required that the parameter of interest followed a normal distribution with common variance. This study did not need an age matching as the main measurement parameters have been shown not to be age dependent in the targeted population (as described in a parallel paper) [33].

To assess correlations between the saccade parameters (peak velocity, amplitude, latency) and the mSST, SARA and IQ, Spearman correlation was performed. Spearman’s rank correlation coefficient of ≥0.8 was considered a strong correlation; ≥ 0.5 was considered a moderate correlation; ≥ 0.2 a weak correlation; and < 0.2 no correlation.

Bilateral abduction deficits were found in 8/16 patients, while 6 patients additionally presented downgaze or/and upgaze restriction. While patients with mild phenotype show free motility until the middle of the 3rd decade, restricted motility already occurs in the 2nd decade in patients with a severe phenotype (Table 2). The motility deficit was of marked severity in two of the patients precluding saccadometric measurements due to poor quality. Strabismus was present in 8 of 16 patients (most common strabismus convergens, followed by strabismus convergens alternans and decompensated esophoria). Half of the patients had full stereopsis in Titmus ring test, however stereo resolution was significantly worse than in the control group (mean degree of disparity in patients 130 (± 149) arc seconds vs. 54 (± 45) arc seconds in the controls, p = 0.041). Limited triggerable optokinetic nystagmus was seen in 14/16 patients (stronger impairment horizontally in 8/16).

Binocular clinical assessment revealed horizontal saccades to be more often decelerated (15/16) than vertical ones (12/16) (Table 3). The patient presenting normal saccades horizontally suffered epilepsy and was genetically diagnosed as type 3 as both mutations N188S and L395P are known in type 3 GD. No type 1 mutation/ no neuroprotective mutation was present.

Rightwards and leftwards saccades were comparable in all patients, however in the above mentioned two patients not assessable due to marked motility restriction. In three patients horizontal saccades were triggerable exclusively by the use of blinking. Two of them also presented head thrust.

A minimum of 4 recorded saccades (as response to 6–8 stimuli per direction and target displacement) in each of the 16 patients was detected in 82.4% of the 12 eccentricities in 16 patients (98.5% in the 100 control subjects). Missing values were few. The saccade parameters had an approximately normal distribution except for latency.

Saccade parameters, especially peak velocity, showed to be impaired in a characteristic pattern of damage: horizontally (rightwards/leftwards) followed by downwards and last upwards (Fig. 7).

Peak velocity was significantly reduced (exemplarily peak velocity 15° saccades horizontally: 69°/s rightwards and 83°/s leftwards in GD3 patients vs. 339°/s rightwards and 364° leftwards in the controls, p ≤ 0.001), while main sequence was preserved. The difference to the control group was more evident horizontally and downwards (p ≤ 0.001), followed by upward saccades (p = 0.001–0.003). Moreover, paired t-test showed downwards saccades in GD3 patient to be slower (144°/s ± 87°/s) compared to upwards saccades (192°/s ± 94°/s).

Further saccadometric data are provided in the Additional file 2 displayed as boxplots [see Additional file 2]. Gain or saccadic accuracy as defined per amplitude divided by target eccentricity proved to have a larger range (in terms of interquartile distance) in GD3 patients. The saccades were of similar accuracy vertically, while being less accurate horizontally. Both study groups showed hypomentric saccades more often in horizontal than in vertical direction (see Additional file 2A).

The latency of Gaucher patients was not normally distributed. The non-parametric Wilcoxon rank sum test revealed that latency was significantly longer in Gaucher patients compared to controls except for upward direction and large target displacements (20° vertically and 30° horizontally) (Table 4) (see Additional file 2B).

Clinical scores correlated with subjective and objective saccadic peak velocity and with saccade hypometria. Phenotype severity showed mild correlation with longer latency. High mSST (adjusted for saccades), high SARA, long duration of disease and low IQ correlate with slow and hypometric saccades. The correlation was found to be greater in vertical than in horizontal saccades. Phenotype severity correlated with longer latency, while its correlation with other saccade parameters was not universal. The underlying analyses data are provided in the Additional file 3, and the neurologic items are listed in the Additional file 4 [see Additional files 3 and 4].

Diffuse corneal opacities have been infrequently reported in GD 3 patients [24]. They occurred in D409H homozygous patients [35‐38] in association with cardiac valve calcification, while we found corneal haze in a heterozygous patient as did Inui et al. in toddler Gaucher type 2 [39]. Our heterozygous patient with corneal opacity was significantly older (44 years) than other patients. If the corneal opacity is due to the presence of the D409H mutation (albeit in the heterozygous compound state), it could be hypothesized that the absence of this finding in the other carriers of the same mutation (L444P/D409H) might be due to their still young age. In another case of Gaucher type 1 disease (F216Y/L444P), corneal abnormality preceded the diagnosis almost 15 years [40].

First descriptions on posterior segment abnormalities in Gaucher patients, before enzyme replacement therapy was established, focused on vitreous opacities [41, 42]. Recent case reports confirmed vitreous opacities, condensations, preretinal hyperreflective dots and posterior vitreous detachment by OCT [21, 23] [24]. Sheck et al. [23] located a hyperreflective preretinal accumulation between the posterior hyaloid interface and the nerve fiber layer of the retina in the mid peripheral and perimacular area (as also shown in Fig. 3). An additional sign in this 14-year-old girl was partial posterior vitreous detachment [23]. One year later in 2013, Coussa showed comparable OCT recordings in a 13-year-old girl – preretinal deposits - with slight progression after a follow up period of 5 years [21]. These are in accordance with our experiences in 16 GD3 patients, in which we found epiretinal particles in the macula, epiretinal and vitreous opacities in the mid periphery, and partial detachment of the posterior vitreous. However, in the present study, we did not only see vitreoretinal lesions (at the interface), but also subretinal lesions. Their occurrence seems more typical with severe phenotype and longer disease duration. The OCT characteristics seem similar to drusen but unlike drusen, the color was whitish instead of yeallowish, similar to the pre-retinal presumed Gaucher cell bodies. One could speculate that the drusen-like deposits are Gaucher cell collections. These lesions located at the peripapillary region, reaching the macula were accompanied by retinal atrophy per loss of photoreceptors and retinal pigment epithelium. We were unable to find comparable findings in a computerized literature search. While chorioretinal atrophy is described in other storage diseases [43], their appearance in these diseases is considerably less marked and without deposits.

Likewise, tortuosity of the retinal vessels is known from other lysosomal storage diseases such as Fabry disease and alpha-mannosidosis. Unlike the isolated finding of tortuosity in our cohort, a previously reported case of GD showed severe tortuosity with extensive opacities in the vitreous body accompanied by visual impairment [44]. Red cherry spot is no specific sign for Gaucher disease [24], which we could confirm. In our cohort, we could not detect any nerve fiber layer deposits as described by Sawicka-Gutaj et al. [22].

At the time being, we cannot answer the crucial question whether the occurence of subretinal (or vitreoretinal) affection indicates a neuronopathic (or visceral) manifestation.

Watanabe [25] and Zhao [26] described tractional retinal detachment due to strong vitreoretinal adhesions and massive vitreous opacities in GD3. In teenagers, the underlying mechanism might include early development of liquefied cavity that exercises traction by the vitreous [25]. If retinal detachment develops, surgical treatment will be essential to preserve vision [25], otherwise permanent vision loss might occur at a young age [26].

Therefore, from the ophthalmologist’s point of view, Gaucher patients whether they are diagnosed type 1 or 3, should undergo the following diagnostic procedures:

GD3 patients develop a progressive horizontal supranuclear gaze palsy [15, 45]. Motility restriction, namely bilateral abduction deficits, indicate abducens motoneuron/nucleus affection [12], esotropia being the likely result of this N. abducens involvement. We found esotropia or decompensated esophoria in 8/18 patients (44%), which is in a comparable range as described earlier [11], and considerably more frequent than its prevalence in healthy subjects (2–3.5%). We could additionally demonstrate motility restriction involving all directions of gaze in 6/16 patients (38%). Oculomotor disturbances follow a typical pattern in Gaucher type 3, in accordance to the topo-anatomical areas (PPRF (paramedian pontine reticular formation), riMLF (rostral interstitial nuclei of the medial longitudinal fasciculus), motoneurons of the abducens nucleus, flocculus/cerebellum and vestibular system [12, 18]. Besides slowed saccadic velocity, the initiation of saccades is delayed and [10] and saccadic gain (accuracy) is reduced, horizontally more severely than vertically [15]. Horizontal gaze may be affected so severely that technical measurement is impossible. Therefore, being a more sensitive measurement, the investigation of the less impaired vertical saccades has gained much more attention in saccadometric studies or possible effects of medical treatment [15, 17]. 2/16 of our patients exhibited saccades that were not reliably measurable due to motility restriction.

Peak velocity: The relation between increasing peak velocity by increasing amplitude, termed main sequence in healthy subjects [46], is preserved in GD3 patients despite abnormal saccades. The overall saccadometric results of decelerated horizontal and secondary vertical saccades confirms previous data on slowed saccades in GD3 patients [8, 10, 18].

Latency: As in previous reports, latency was found to be prolonged [10, 12, 15, 18], except for upwards saccades, which impair lastly. The results show enlarged interquartile distances in the boxplots, which in our view rather reflects the susceptibility for confounders such as age than for disease severity.

Gain: Saccadic accuracy (gain) received relatively little attention in studies on oculomotor involvement in GD3 patients. A study on Norbottnian GD3 has recently reported normal accuracy [18], while a previous report described it to be reduced [15]. In our cohort, horizontal saccades were less precise and rather hypometric. The same held true for vertical ones, although to a lesser extent. One of the reasons for low gain values in general is motility restriction, especially regarding the maximal target eccentricity. Anticonvulsive medication might be a confounder resulting in slowed and hypometric saccades [45]. We figured out the patients without anticonvulsive medication (12/16) to have slowed saccades (just like the patients with anticonvulsive medication (4/16)) with one exception corresponding to a mild genotype (N188/−).

Saccadometry was feasible in all Gaucher type 3 patients (14/16 performed saccadometry with sufficient data quality). We also recommend performing OCT as it is a widely used, expeditious, and informative investigation method feasible in less than 10 min in Gaucher patients, while saccadometry rather takes 10 to 30 min depending on the protocol used and the necessity to retake a test. The additional, typical signs of oculomotor abnormalities such as synkinetic blinking [8] are assessable by clinical evaluation. Another technique is an upwards curvature to horizontal saccades likely a Bell’s phenomenon (but keeping the lids open). Both are felt to inhibit the omnipause region allowing the saccade to begin. The latter phenomenon is noticeable during saccadometry and may disturb measurements by producing artefacts. Clinical assessment is difficult and less sensitive: While clinical assessment could not reveal differences between upwards and downward saccades, saccadometry was able to measure significant differences regarding this measurement. This is why we recommend to screen Gaucher type 1 patients and to monitor Gaucher type 3 patients for their neurologic manifestation by meticulous motility evaluation, ideally by video-oculography. We suggest that GD1 patients should be assessed for individual follow-up and that a normative GD1 cohort data, which does not exist to the best of our knowledge, should be established.

Infrared video-oculography is non-invasive, not stressful and easy to understand. It is well tolerated by children, however case-by-case-decision is surely advocated. The measurements are time consuming and editing of the raw data is needed to ensure optimum data acquisition. The quality of measurements depends primarily on the patient’s cooperation of the investigator’s experience.

Correlation between peak velocity and neurologic status (adjusted SARA, disease duration, IQ and to a slightly less extend the mSST) was greatest vertically. We were unable to reproduce the strongest correlation for downwards saccades as reported by Bremova (2018) [12]. However, the better preserved vertical spectrum of velocity allows stronger correlations. This fact has been explained by the horizontal floor effect [15] or ceiling effect [12]. Vertical saccadic peak velocity therefore functions as indicator for neurologic manifestation [12, 15, 17]. Saccadometry has the potential to qualify for those diagnostic procedures allowing quantification of neurologic manifestation such as the mSST. It might replace the less objective clinical saccade assessment and optimize the scoring system. Although correlation between clinical assessment of saccadic impairment and adjusted mSST, SARA and IQ were demonstrated, saccadometry is definitely more objective.

Despite the low prevalence of all GD-phenotypes of 0.7 to 1.75 per 100.000 inhabitants [47] and an amount of 5% GD3 [48] within these, we were able to investigate a group of 16 GD3 patients. Most studies reporting on oculomotor disorders in GD3 are based on smaller samples [8, 10, 14, 15, 18] or display a multicenter study design [12]. Schiffmann et al. performed saccadometry on 30 GD3 patients in two study centers as part of a randomized controlled study on miglustat [17]. One older report exists on clinical oculomotor and ophthalmological findings in 22 Norbottnian GD3 patients [11]. Statistical analysis of the limited samples encountered in rare diseases is challenging. In this explorative study, the significance level has not been adjusted for multiple testing, which may result in an excess of false positive results. Confirmatory studies are required to verify the results.

Another problem is a possible selection bias as the included patients mostly presented mild or intermediate phenotype severity. For patients with severely affected oculomotor disturbances, a favorable atmosphere for the measurement was created and the eyelid was fixed to minimize blinking artefacts.

A potential flaw of our investigation is the fact that in presbyopic study participants, no near correction was used. Quite interestingly, up to now there is no data on how near correction influences saccadometric precision [49]. Myopes and emmetropes show similar saccadic eye movements [50].

Regarding the pre-set program of the video oculography system we used, it would be useful to provide a shorter protocol to maintain attention and thus to reduce artefacts during saccadometry in patients with abnormal eye movements. As patients with restricted motility are not able to reach large target eccentricities, smaller maximal target displacements should be considered in those cases. Longitudinal saccadometry might be appropriate for monitoring and seems to be more sensitive than the clinical testing used in the scoring system.