Computer-Assisted Visual Training in Children and Adolescents with Developmental Visual Disorders: A Systematic Review

A systematic review

Background: In this systematic review, we address the question whether children and adolescents with developmental visual disorders benefit from computer-assisted visual training.

Methods: Systematic literature searches were carried out in three bibliographic databases (initial search in October 2021) and trial registries. Included were randomized controlled trials that evaluated the efficacy of computer-assisted visual training in children and adolescents with developmental visual disorders in comparison to no training, sham training, or conservative treatment.

Results: The inclusion criteria were met by 17 trials (with a total of 1323 children and adolescents) focusing on binocular or monocular computer-assisted visual training for the treatment of amblyopia. In these trials, visual training was carried out for 2 to 24 weeks, either as “stand alone” therapy or in addition to occlusion therapy. Six trials showed a statistically significant difference in favor of the visual training for the outcome “best corrected visual acuity of the amblyopic eye.” However, this difference was small and mostly below the threshold of clinical relevance of –0.05 logMAR (equivalent to an improvement of 0.5 lines on the eye chart, or 2.5 letters per line). Only few data were available for the outcomes “binocular vision” and “adverse events”; the differences between the groups were similarly small.

Conclusion: The currently available data do not permit any firm conclusions regarding the efficacy of visual training in children and adolescents with amblyopia. Moreover, treatment adherence was often insufficient and the treatment durations in the trials was relatively short. No results from randomized trials have yet been published with respect to other developmental visual disorders (refractive errors, strabismus).

Many children and adolescents suffer from developmental functional vision disorders, including amblyopia (reduced vision in one eye), eye misalignment (strabismus, squint), and refractive errors (for example, myopia and hyperopia).

Amblyopia is due to an inadequate development of the visual pathways during early childhood and is caused by uncorrected unilateral refractive errors (anisometropia) in over 60% of those affected (1). In the presence of such uncorrected refractive errors, only blurred image contours are projected onto the retina, which prevents normal development of visual acuity. Apart from refractive errors, eye misalignment can also lead to amblyopia. In order to avoid double vision from eye misalignment, the visual input from the misaligned eye is suppressed, and this in turn can lead to amblyopia.

A Dutch population-based study estimated the cumulative lifetime risk of bilateral visual impairment among individuals with unilateral amblyopia to be 18%, and 10% in the absence of amblyopia (2). Overall, the risk for bilateral visual impairment in those with unilateral amblyopia is two to three times greater than in persons without amblyopia (3). Furthermore, the worldwide prevalence of amblyopia in 2018 was estimated to be 1.8% (95% confidence interval [1.6; 1.9%]) (1). A noticeably high prevalence of 3.7% [2.9; 4.5] was found in Europe.

Because the plasticity of the nervous system deteriorates with increasing age, early treatment becomes all the more important (4). Conventional treatment of amblyopia involves occlusion of the fellow non-amblyopic eye with a special patch. Occlusion therapy, however, has its disadvantages (3, 5):

• On the one hand, it can have a negative impact on personality development.
• On the other, covering the non-amblyopic eye suppresses eye coordination, which can result in further visual impairments.

It is therefore understandable that research is underway into treatment approaches—based, for example, on the methods of perceptual learning (5).

Financing of these new (mostly digital) treatment methods varies: Some are only used in scientific studies, while others are available commercially (6). Digital binocular (dichoptic) vision training is one of the new methods and presents each eye with a different image component (usually during a video game): the non-amblyopic eye with contrast-reduced (out of focus) and the amblyopic eye with contrast-enhanced (focused) image components (7). The image separation required for binocular training is achieved with the use of a special pair of glasses, for example, anaglyph glasses (use of colors to separate the image), shutter glasses (coated eyeglass lenses which can alternate between transparent and opaque), or headsets (virtual reality glasses). In order to be successful at binocular training (using a video game), both eyes must perceive their respective image component and fuse the image parts.

The contrast of the image component seen by the non-amblyopic eye is increased during treatment—ideally, until fusion capability is achieved with equal contrast of both image components. There are also monocular training measures which, for example, use background stimulation (a moving sinusoidal grating) and are intended to promote neural stimulation (8, 9).

In Germany, visual training by the company Caterna Vision, based on such a procedure, was used by more than 350 ophthalmic practices and hospitals in the year 2020 (8).

In view of the large number of new treatment methods, the aim of the present systematic review, which was conducted as part of a Health Technology Assessment (HTA) report (10), was to evaluate the benefits and harms of digital vision training compared with treatment without vision training in children and adolescents with developmental vision disorders.

Methods

The protocol of this systematic review has been published in the PROSPERO register (CRD42021289044). Reporting was based on the Transparent Reporting of Systematic Reviews and Meta-Analyses (PRISMA) statement (11).

The review included randomized controlled trials in which digital vision training was used to treat children and adolescents with developmental vision disorders.

The best corrected visual acuity of the amblyopic eye was specified using the unit “logMAR”. The MAR is the minimum angle of resolution and represents the reciprocal of the visual acuity (visual acuity = 1/MAR). If the visual resolution is averaged, the logarithm is used (logMAR = log[1/visual acuity]). LogMAR values are the opposite of visual acuity values (logMAR = −log[visual acuity]).

Detailed information on the methodology and the search strategy can be found in the eMethods Section and in eTable 1. A meta-analysis was not performed due to the small number of studies per comparison, heterogeneous patient populations, differences in treatment procedures, as well as due to statistical heterogeneity.

eTable 1

Search strategy in the Medline database

Results

Search

The Figure presents the selection process of the studies. The full texts from 49 of 1303 references were reviewed, of which 17 trials (19 references [12–30]) that exclusively assessed digital vision training for amblyopia met the inclusion criteria.

Figure

Study selection according to the PRISMA flow diagram

Study characteristics

The key characteristics of the 17 randomized trials are included in eTable 2, stratified by intervention (binocular or monocular training) and comparative treatment. Binocular training was assessed in 11 trials (N = 1138 participants) (12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22) and monocular training in six trials (N = 185 participants) (23, 24, 25, 26, 27, 28). Occlusion therapy was used in six trials as the only comparative treatment (12, 13, 14, 15, 16, 17). In the other trials, vision training was compared with no training (16, 18, 19, 20, 23, 24) or with sham training (21, 22, 25, 26, 27, 28).

eTable 2

Study characteristics

The trials were conducted in Europa, North America, Asia, and Australia. Treatment duration ranged between two and 24 weeks, and the average age of the study population varied between 4.3 and 14.3 years. In four trials, only children with an initial diagnosis were treated. In the other trials, the proportion of cases which had undergone previous treatment was between 21 and 97%. Children with amblyopia due to anisometropia alone were examined in four trials. Training was conducted in three trials on an outpatient basis and in 14 trials at home.

The best corrected visual acuity of the amblyopic eye

Data for the endpoint “best corrected visual acuity of the amblyopic eye” are presented in the Table, stratified according to intervention and comparative treatment.

Table

Best corrected visual acuity of the amblyopic eye

Vision training versus no training

When binocular vision training is compared with no training, then two out of four trials treating three- to ten-year-old children demonstrated effects in favor of the intervention: mean difference [MD] in logMAR: −0.10; [−0.16; −0.04] and −0.11; [−0.17; −0.05], respectively. The effect estimators of the two other trials which treated three- to thirteen-year-olds pointed in the same direction (in favor of the intervention), although the group difference was not statistically significant. The two trials on monocular training also demonstrated a statistically significant superiority of digital training: MD in logMAR: –0.11; 95% CI not calculable; [0.0001; −0.04] and –0.13; [-0.20; –0.03], respectively. It is not possible to deduce from the available data whether additional occlusion therapy enhanced the effect of the digital intervention.

Vision training versus sham training

The two trials which compared binocular training with sham training did not find any statistically significant group differences: MD in logMAR: –0.01; [0.08; −0.06] and 0.06; [-0.02; 0.14], respectively. One of four trials reported a statistically significant effect in favor of vision training in four- to eight-year-olds for monocular training. Due to an absence of numerical values, however, it is not possible to estimate in which area this improvement lies. One of the trials without a significant effect also failed to report any numerical values. The MD in logMAR of the two others was −0.02 and 0.03, with a 95% confidence interval of [−0.20; 0.16] and [−0.11; 0.18], respectively.

Vision training versus occlusion therapy

One of six trials demonstrated a significant difference in favor of binocular training in four- to ten-year-olds: MD in logMAR: –0.08; [-0.13; –0.03]. The effect estimators were in favor of the occlusion therapy in the other five trials which covered a wide age range (three- to sixteen-year-olds).

Endpoint binocular vision

eTable 3 contains data on the endpoint “binocular vision”.

eTable 3

Binocular vision

Vision training versus no training

One of three trials demonstrated a significant effect in favor of binocular training in three- to thirteen-year-olds: MD in log angle-seconds: +0.38; [+0.06; +0.70]. One of the two trials without a significant difference did not publish any explicit data. The effect estimator of the other trial was +0.10; [–0.19; +0.34].

Vision training versus sham training

Altogether, two trials (without providing numerical values) reported that no group differences were found for binocular vision.

Vision training versus occlusion therapy

Group differences were found in none of the five trials. One study stated an effect estimator: MD in log angle-seconds: +0.20; [–0.10; +0.50].

Adverse effects

Adverse outcomes, such as double vision, asthenopia (eyestrain) and headache, dizziness, increased blinking, or nightmares were reported in seven studies (eTable 4). However, these events only occurred occasionally, and it is not to be expected that digital training is associated with an increased risk of adverse events.

eTable 4

Adverse events

Risk of bias

The risk of bias was rated as low in one trial (21) and high in 16 trials. Studies in which vision training was compared with no training or occlusion generally do not allow blinding of treatment. Therefore, it cannot be excluded that, to an uncertain degree, the study results are due to the awareness of the study population and the study staff of the respective intervention. In addition, there is uncertainty about adherence to the training plan (compliance). For example, deviations are reported in up to 87% of cases. Furthermore, it remains uncertain whether the children and adolescents wore the glasses, which are required for image separation, all the time and how often their eyes wandered from the screen. Details of the risk-of-bias assessment are shown in eTable 5.

Discussion

Whereas conventional amblyopia therapy is usually integrated as part of everyday life, the actual technique used for vision training often determines whether the procedures are used at home or a visit to the outpatient clinic is needed. Vision training often requires many sessions, and the organizational effort can be quite demanding if it is conducted on an outpatient basis. A transfer to the home setting is therefore of advantage. However, such a transfer can only be successful if the available equipment allows it and if the children are supported when using the devices and games. Yet the present review shows that compliance issues often develop (eTable 5). Younger children in particular experienced concentration difficulties. With adolescents, on the other hand, it was often a question of boredom because the video games offered are inferior to the innovative computer game genres (10). Apart from the lack of compliance, there is also a heightened risk awareness in the parents with regard to digital treatment methods. It is therefore understandable that the use of digitally supported forms of treatment for children and adolescents is approached critically (31). The decision-making process should therefore involve the following elements:

eTable 5

Risk of bias

• Treatment benefits: Although individual trials have shown a statistically significant effect in favor of vision training, it cannot be assumed that the measured differences are of any clinical relevance. For example, the upper and lower limits (95% CI) of the measured MDs for the best corrected visual acuity of the amblyopic eye were −0.03 logMAR (improvement of about +0.3 lines on the eye chart) and −0.20 logMAR (+2.0 lines), respectively. If the published value of −0.05 logMAR (+0.5 lines) is used as the non-inferiority limit for visual acuity (12), then only one study managed to exceed this threshold value (24): The lower and upper limits of the measured MDs after nine weeks were –0.06 logMAR (+0.6 lines) and –0.24 logMAR (+2.4 lines), respectively. A lack of any effect when comparing vision training with sham training could be due to the comparative intervention: For example, in binocular training, different contrast elements were not incorporated into the sham training, but the children still wore image-separating glasses. In monocular training, the background stimulation used for neural stimulation was dispensed with in the sham intervention. It therefore cannot be ruled out that the sham training has the same effect as the intervention which the studies examined. If vision training is compared with occlusion therapy, then—with the exception of one trial (15)—no statistically significant effects were identified in favor of vision training. On the contrary, the effect estimators were in favor of occlusion therapy. Little study data is available for binocular vision, and only one trial (16) demonstrated any benefit in favor of digital treatment. When interpreting the results, it must also be taken into account that, although the stereo tests used (for example, the Titmus test) are suitable for recognizing deficits, they are not good at quantifying thresholds (32). From the heterogeneous study pool available, it was also not possible to deduce whether the age of the children or the form of amblyopia could possibly have an impact on the treatment effect.
• Treatment costs: The average total cost of conventional amblyopia treatment per patient for the first year of treatment is between 606 and 646 euros (10). Except for a personal contribution of about 92 euros, these costs are reimbursed by the statutory health insurance system. Since digital training methods have so far only been used as a supplement to conventional treatment in Germany, these costs are incurred with each amblyopia treatment. If the vision training is provided by Caterna Vision as concomitant therapy, then the overall costs are increased by 380 euros per person. The costs for Caterna vision training, which is limited to three months, are only covered by certain health insurance funds within the scope of selective contracts. Apart from Caterna vision training, the insurance funds do not otherwise reimburse vision training (10).

Ethical aspects: The vulnerability of the population also plays a role in the decision-making process (10). A too-early exposure to digital media has its risks and can lead to increased media consumption. Furthermore, health games fall under the category “Serious computer games” because video games are used as both recreation and a therapeutic measure. “Gaming disorder” as defined in the ICD-11 is also recognized by the WHO as an official disease.

Apart from the present article, other systematic reviews (33, 34, 35) have also been published in recent years. The conclusions of these articles are comparable with the present results, and the authors point out that, at the moment, replacement of conventional therapy by digital therapies cannot be recommended. Furthermore, another randomized study was published in 2022 which assessed Caterna vision training (36). This study included 37 children from the Russian Federation, some with bilateral amblyopia or pathological alterations of the fundus. This patient population did not fulfill the inclusion criteria of the present review, so the trial was not included in the results section. Although the study authors concluded that the children of the intervention group demonstrated a significantly better monocular visual acuity after ten days of treatment as compared with controls, it does not appear that the differences found in this trial are of any clinical relevance either. Other randomized studies published since the last literature search also suggest no clinically relevant effects of digital vision training (for example, 37, 38).

Conclusions

Overall, the available studies do not allow a conclusive statement on the benefits and harms of vision training in children and adolescents with amblyopia. Apart from the marginal effect and often absent compliance, it must be considered that the duration of treatment in the trials was set at a few weeks and does not really reflect the results of amblyopia treatment, which often lasts for some years. It was also not possible to deduce from the heterogeneous study pool available whether children with amblyopia due to anisometropia achieve better results than children with amblyopia caused by strabismus or whether the age of the children had an impact on the effect of treatment. Such results would be particularly important for estimating whether certain patient populations would possibly benefit from such treatment.

Acknowledgments

We would like to thank Ms Carolin Wolf (orthoptist at the Freiburg Eye Clinic) for her specialist expertise.

Funding

This systematic review is part of a Health Technology Assessment (HTA) and was financed by the Institute for Quality and Institute for Quality and Economic Efficiency in Health Care (HTA-21–03).

Conflict of interest statement
The authors confirm that there are no conflicts of interest.

Corresponding author:
PD Dr. sc. hum. Christine Schmucker

Institute for Evidence in Medicine, University Hospital of Freiburg

Faculty of Medicine, Albert Ludwig University of Freiburg

Breisacher Str. 86, 1st Floor, 79110 Freiburg

Cite this as:
Schmucker C, Thörel E, Flatscher-Thöni M, Sow D, Göhner A, Stühlinger V, Mühlberger N, Lagréze WA, Meerpohl J: Computer-assisted visual training in children and adolescents with developmental visual disorders—a systematic review. Dtsch Arztebl Int 2023; 120: 747–53. DOI: 10.3238/arztebl.m2023.0191

►Supplementary material
eReferences, eMethods, eTables, eFigures, eBox:
www.aerzteblatt-international.de/m2023.0191

References

1. Hashemi H, Pakzad R, Yekta A, et al.: Global and regional estimates of prevalence of amblyopia: a systematic review and meta-analysis. Strabismus 2018; 26: 168–83

2. van Leeuwen R, Eijkemans MJC, Vingerling JR, Hofman A, de Jong PTVM, Simonsz HJ: Risk of bilateral visual impairment in individuals with amblyopia: the Rotterdam study. Br J Ophthalmol 2007; 91: 1450–1

3. Pieh C, Fronius M, Chopovska Y, et al.: [“Fragebogen zum Kindlichen Sehvermögen (FKS)“. Assessment of quality of life with the German version of the Children‘s Visual Function Questionnaire]. Ophthalmologe 2009; 106: 420–6

4. Fronius M, Cirina L, Ackermann H, Kohnen T, Diehl CM: Efficiency of electronically monitored amblyopia treatment between 5 and 16 years of age: new insight into declining susceptibility of the visual system. Vision Res 2014; 103: 11–9

5. Levi DM, Li RW: Perceptual learning as a potential treatment for amblyopia: a mini-review. Vision Res 2009; 49: 2535–49

6. Vivid Vision. Virtual realitiy vision training for lazy eye. www.visus.de/software/vivid-vision/ (last accessed on 15 May 2023).

7. Bach M: [Dichoptic training for amblyopia]. Ophthalmologe 2016; 113: 304–7

8. Caterna Vision GmbH. Die neue digitale Amblyopietherapie. https://caterna.de/ (last accessed on 15 May 2023).

9. Lagrèze WA: App auf Rezept: Besorgt. Dtsch Arztebl 2014; 111: A-845 / B-727 / C-689

10. Schmucker C, Thoerel E, Flatscher-Thoeni M, et al.: Entwicklungsbedingte Sehstörungen: Profitieren Kinder und Jugendliche von aktivem Sehtraining? www.iqwig.de/download/ht21-03_sehtraining-bei-kindern-und-jugendlichen_hta-bericht_v1-0.pdf. HTA-Bericht 2022: HT21-03 (last accessed on 15 May 2023).

11. Moher D, Liberati A, Tetzlaff J, Altman DG: Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009; 6: e1000097

12. Holmes JM, Manh VM, Lazar EL, et al.: Effect of a binocular iPad game vs part-time patching in children aged 5 to 12 years with amblyopia: a randomized clinical trial. JAMA Ophthalmol 2016; 134: 1391–400

13. Manh VM, Holmes JM, Lazar EL, et al.: A randomized trial of a binocular iPad game versus part-time patching in children aged 13 to 16 years with amblyopia. Am J Ophthalmol 2018; 186: 104–15

14. Rajavi Z, Sabbaghi H, Amini Sharifi E, Behradfar N, Kheiri B: Comparison between patching and interactive binocular treatment in amblyopia: a randomized clinical trial. J Curr Ophthalmol 2019; 31: 426–31

15. Birch EE, Jost RM, Kelly KR, Leffler JN, Dao L, Beauchamp CL: Baseline and clinical factors associated with response to amblyopia treatment in a randomized clinical trial. Optom Vis Sci 2020; 97: 316–23

16. Yao J, Moon HW, Qu X: Binocular game versus part-time patching for treatment of anisometropic amblyopia in Chinese children: a randomised clinical trial. Br J Ophthalmol 2020; 104: 369–75

17. Rajavi Z, Soltani A, Vakili A, et al.: Virtual reality game playing in amblyopia therapy: a randomized clinical trial. J Pediatr Ophthalmol Strabismus 2021; 58: 154–60

18. Rajavi Z, Sabbaghi H, Amini Sharifi E, Behradfar N, Yaseri M: The role of interactive binocular treatment system in amblyopia therapy. J Curr Ophthalmol 2016; 28: 217–22

19. Xiao S, Angjeli E, Wu HC, et al.: Randomized controlled trial of a dichoptic digital therapeutic for amblyopia. Ophthalmology 2022; 129: 77–85

20. Holmes JM, Manny RE, Lazar EL, et al.: A randomized trial of binocular dig rush game treatment for amblyopia in children aged 7 to 12 years. Ophthalmology 2019; 126: 456–66

21. Gao TY, Guo CX, Babu RJ, et al.: Effectiveness of a binocular video game vs placebo video game for improving visual functions in older children, teenagers, and adults with amblyopia: a randomized clinical trial. JAMA Ophthalmol 2018; 136: 172–81

22. Herbison N, MacKeith D, Vivian A, et al.: Randomised controlled trial of video clips and interactive games to improve vision in children with amblyopia using the I-BiT system. Br J Ophthalmol 2016; 100: 1511–6

23. Iwata Y, Handa T, Ishikawa H, Goseki T, Shoji N: Comparison between amblyopia treatment with glasses only and combination of glasses and open-type binocular „occlu-Pad“ device. Biomed Res Int 2018; 2018

24. Dadeya S, Dangda S: Television video games in the treatment of amblyopia in children aged 4–7 years. Strabismus 2016; 24: 146–52

25. Bau V, Rose K, Pollack K, Spoerl E, Pillunat LE: Effectivity of an occlusion-supporting PC-based visual training programme by horizontal drifting sinus gratings in children with amblyopia. Klin Monatsbl Augenheilkd 2012; 229: 979–86

26. Kämpf U, Muchamedjarow F, Seiler T: Supportive amblyopia treatment by means of computer games with background stimulation; a placebo controlled pilot study of 10 days. Klin Monatsbl Augenheilkd 2001; 218: 243–50

27. Yeh WH, Lai LJ, Chang DW, Lin WS, Lin GM, Shaw FZ: Portable rotating grating stimulation for anisometropic amblyopia with 6 months training. Sci Rep 2021; 11: 11430

28. Jukes C, Bjerre A, Coupe J, Gibson J: Pilot study evaluating the feasibility of comparing computer game play with close work during occlusion in children aged 2–7 years with amblyopia. Br Ir Orthopt J 2019; 15: 115–24

29. Gao TY, Black JM, Babu RJ, et al.: Adherence to home-based videogame treatment for amblyopia in children and adults. Clin Exp Optom 2021; 104: 773–9

30. Kelly KR, Jost RM, Dao L, Beauchamp CL, Leffler JN, Birch EE: Binocular iPad game vs patching for treatment of amblyopia in children: a randomized clinical trial. JAMA Ophthalmol 2016; 134: 1402–8

31. Fronius M: Neue digitale Therapie- und Diagnostikmethoden für Amblyopie. Augenspiegel 2020: 42–4.

32. Tittes J, Baldwin AS, Hess RF, et al.: Assessment of stereovision with digital testing in adults and children with normal and impaired binocularity. Vision Res 2019; 164: 69–82

33. Pineles SL, Aakalu VK, Hutchinson AK, et al.: Binocular treatment of amblyopia: a report by the American Academy of Ophthalmology. Ophthalmology 2020; 127: 261–72

34. Brin TA, Chow A, Carter C, Oremus M, Bobier W, Thompson B: Efficacy of vision-based treatments for children and teens with amblyopia: a systematic review and meta-analysis of randomised controlled trials. BMJ Open Ophthalmol 2021; 6: e000657

35. Chen CW, Zhu Q, Duan YB, Yao JY: Comparison between binocular therapy and patching for treatment of amblyopia: a meta-analysis of randomised controlled trials. BMJ Open Ophthalmol 2021; 6: e000625

36. Kämpf U, Rychkova S, Lehnert R, Heim E, Muchamedjarow F: Visual acuity increase in meridional amblyopia by exercises with moving gratings as compared to stationary gratings. Strabismus 2022; 30: 99–110

37. Jost RM, Hudgins LA, Dao LM, et al.: Randomized clinical trial of streaming dichoptic movies versus patching for treatment of amblyopia in children aged 3 to 7 years. Sci Rep 2022; 12: 4157

38. Zheng CY, Xu W, Wu SQ, Han DX: A randomized study of network-based perception learning in the treatment of amblyopia children. Int J Ophthalmol 2022; 15: 800–6

e1. Berufsverband der Augenärzte (BVA): Leitlinie Nr. 26 a. http://augeninfo.de/leit/leit26a.pdf (last accessed on 13 April 2023).

e2. Arbeitsgemeinschaft der Wissenschaftlichen Medizinischen Fachgesellschaften (AWMF): S2k-Leitlinie 022–020: Visuelle Wahrnehmungsstörung. www.awmf.org/uploads/tx_szleitlinien/022-020l_S2k_Visuelle-Wahrnehmungsstoerungen_2017-12.pdf (last accessed on 13 April 2023).

e3. McGowan J, Sampson M, Salzwedel DM, Cogo E, Foerster V, Lefebvre C: PRESS peer review of electronic search strategies: 2015 guideline statement. J Clin Epidemiol 2016; 75: 40–6

e4. Sterne JAC, Savović J, Page MJ, et al.: RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ 2019; 366: l4898