Article Text

Download PDFPDF
Long-term reconstruction of foveal microstructure and visual acuity after idiopathic macular hole repair: three-year follow-up study
  1. Masahiro Kitao,
  2. Taku Wakabayashi,
  3. Kentaro Nishida,
  4. Hirokazu Sakaguchi,
  5. Kohji Nishida
  1. Department of Ophthalmology, Osaka University Graduate School of Medicine, Suita, Japan
  1. Correspondence to Dr Taku Wakabayashi, Department of Ophthalmology, Osaka University Graduate School of Medicine, Suita, Osaka 565-0871, Japan; taku.wakabayashi{at}ophthal.med.osaka-u.ac.jp

Abstract

Aims To evaluate long-term reconstructive changes in foveal microstructures and their associations with visual improvement in eyes with surgically closed macular holes (MHs).

Methods Twenty-eight eyes of 28 patients who underwent successful idiopathic MH repair were retrospectively studied. Best-corrected visual acuity (BCVA) and spectral-domain optical coherence tomography images were examined preoperatively and 1, 3, 6, 12, 24 and 36 months postoperatively. Correlations between postoperative BCVA and parameters relating to the reconstruction of the foveal photoreceptor layer including the external limiting membrane (ELM), ellipsoid zone (EZ) and cone interdigitation zone (CIZ) as well as changes in glial cells were evaluated.

Results Logarithm of the minimum angle of resolution BCVA improved continuously during 3-year follow-up (baseline 0.70±0.27, 1 month 0.36±0.34, 3 months 0.29±0.30, 6 months 0.22±0.24, 12 months 0.18±0.25, 24 months 0.14±0.22, 36 months 0.10±0.19) (p=0.015). Continuous reconstruction of the foveal microstructure was apparent throughout the 3-year follow-up. The reconstruction process was initiated by glial proliferation, followed by ELM bridging, glial elimination with EZ reconstruction and CIZ reconstruction. Better BCVA at the 3-year time-point was significantly associated with early ELM bridging, early glial disappearance and photoreceptor integrity defined as complete reconstruction of the ELM, EZ and CIZ.

Conclusions Integrity of the photoreceptor layer was correlated with better long-term visual outcomes after MH repair. Reconstruction of the foveal ELM and disappearance of glial proliferation in the early postoperative period predicted better visual recovery.

  • macular hole
  • external limiting membrane (ELM)
  • ellipsoid zone (EZ)
  • cone interdigitation zone (CIZ)
  • glial cells
  • visual recovery

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Macular holes (MHs) are full-thickness retinal defects in the fovea resulting in visual impairment.1 Advances in surgical techniques and instrumentation have significantly improved the anatomic and visual success rates of MH treatments. In recent decades, anatomic success rates of 85% to 100% and visual improvement rates of 85% to 95% have been reported.2–6 However, even after successful MH closure, some patients may occasionally have limited visual improvement. In such cases, spectral-domain optical coherence tomography (SD-OCT) can be used to investigate the presence of photoreceptor disruption.7 Several studies have reported that the integrity of the external limiting membrane (ELM), ellipsoid zone (EZ) (previously termed the inner segment and outer segment) and the cone interdigitation zone (CIZ) (previously termed the cone outer segment tips (COST) line) correlated with postoperative visual acuity, indicating that reconstruction of the photoreceptor layer may be critical for visual recovery in cases of surgically closed MHs.8–13 However, most relevant studies have only reported data derived from follow-up periods of 1 year, and associations between foveal reconstructive changes and postoperative visual acuity extending for up to 3 years have not been extensively reported.

Previous SD-OCT studies have investigated the healing process in surgically closed MHs.7 8 14–16 A MH is closed by proliferating glial cells filling the foveal defect, and centripetal bridging of the ELM with subsequent reconstruction of the EZ. Excessive glial proliferation prevents reconstruction of the photoreceptor layer and has been shown to be associated with worse visual acuity. However, the long-term relationships between this glial cell proliferation and restoration of the foveal photoreceptor layer have not been reported. In addition, it is unclear how long the glial cells are present for and how long photoreceptor reconstruction lasts in the fovea in the long term.

The purpose of this study was to investigate the long-term restoration of foveal microstructures and its association with visual recovery in patients with successful MH closure. The predictive factors for visual acuity 3 years postoperatively were also evaluated.

Methods

One hundred and fifty-five patients who had undergone anatomically successful MH surgery at Osaka University Hospital in Japan between June 2009 and September 2012 were initially enrolled in this observational study. All patients included had been diagnosed with a stage 2, 3 or 4 idiopathic MH according to the Gass classification.17 We excluded eyes with a traumatic MH, a retinal detachment-associated MH, a high myopia-associated MH with a refractive error of more than −8 D or a follow-up length less than 36 months postoperatively. All patients had a standard ophthalmological examination before and 1, 3, 6, 12, 24 and 36 months after surgery. Examinations included best-corrected visual acuity (BCVA), binocular indirect ophthalmoscopy, slit-lamp biomicroscopy, fundus photography and SD-OCT (Cirrus HD-OCT, Carl Zeiss Meditec, Dublin, California, USA).

The surgery was performed via standard 25-gauge pars plana vitrectomy consisting of a core vitrectomy with intravitreal injection of triamcinolone acetonide (TA) to clarify the vitreous gel, surgical creation of a posterior vitreous detachment if it was not present, internal limiting membrane (ILM) peeling using TA or ‘Brilliant Blue G’ (0.25 mg/mL) and gas tamponade with 20% sulfur hexafluoride. Cataract surgery and intraocular lens implantation were performed simultaneously in patients with a cataract. Patients were instructed to remain face down for 2 to 4 days postoperatively. MH closure was confirmed via postoperative SD-OCT images.

Preoperative data included a complete medical and ophthalmic history, preoperative BCVA, lens status, and clinical staging and minimum diameter of the MH on SD-OCT. In all patients, the SD-OCT was performed on the same day as the clinical examination. Foveal microstructures were evaluated via a horizontal 6 mm scan image using five-line raster mode and serial B-scan cross-sectional images using macular cube 512×128 scanning. Restoration of the photoreceptor layer was defined as the continuous back-reflection line corresponding to the ELM, EZ and CIZ. The glial proliferation was defined as the presence of a foveal hyper-reflective lesion on SD-OCT based on the previous reports of histopathologic studies of autopsy eyes and OCT images of surgically closed MHs.18–21 Two masked investigators (MK, TW) interpreted the SD-OCT images. In the event of disagreement, a third investigator (HS) was consulted to derive a final determination. Central retinal thickness was measured automatically as the average retinal thickness in the central area within a diameter of 1 mm. We also evaluated the inner retinal defect known as the dissociated optic nerve fibre layer (DONFL) appearance using en face SD-OCT images.22 En face SD-OCT images at the level of the ILM were created via advanced visualisation of the macular cube 512×128 scan to visualise the inner retinal layers.23 The area of the DONFL within a 6×6 mm area was measured manually using ImageJ software (National Institutes of Health, Bethesda, Maryland, USA).

For statistical analysis, BCVA was measured using the Landolt C acuity chart and analysed on a logarithm of the minimal angle of resolution (logMAR) scale. The results were analysed using a one-way analysis of variance (ANOVA) when quantitative parameters were compared among baseline, 1, 3, 6, 12, 24 and 36 months. If the parameter was not normally distributed, the non-parametric Kruskal-Wallis one-way ANOVA on ranks was applied. χ2 tests or the Fisher’s exact test was performed as appropriate to compare the proportions of baseline and postoperative data at 1, 3, 6, 12, 24 and 36 months. Univariate regression was used to investigate associations between logMAR BCVA and age, gender, symptom duration, preoperative BCVA, stage and diameter of the MH, and postoperative foveal microstructures on SD-OCT at the follow-up visit to determine whether they were predictive of BCVA at 3 years. All analyses were conducted using SigmaStat software V.3.1 (SPSS) and JMP Pro software (SAS), and p<0.05 was deemed to indicate statistical significance.

Results

Twenty-eight eyes of 28 patients with a surgically closed MH met the study criteria for data analysis. The baseline characteristics of the 28 eyes are summarised in table 1. The mean patient age was 67.0±5.5 years (range 53–79 years). Preoperative mean logMAR BCVA was 0.70±0.27. Preoperatively, the MHs were classified as stage 2 in 9 eyes (32%), stage 3 in 13 eyes (46%) and stage 4 in 6 eyes (21%). The mean preoperative minimum diameter of the hole was 308±140 µm (range 89–609 µm).

Table 1

Baseline patient characteristics

Visual outcomes

Changes in logMAR BCVA are shown in table 2. At 1, 3, 6, 12, 24 and 36 months, the respective mean postoperative BCVAs were 0.36±0.34, 0.29±0.30, 0.22±0.24, 0.18±0.25, 0.14±0.22 and 0.10±0.19 indicating that BCVA continuously improved over 3 years postoperatively (p=0.015).

Table 2

Postoperative time-dependent changes in logMAR BCVA and foveal microstructure

Reconstruction of foveal microstructure

Changes in the status of the ELM, EZ, CIZ and glial proliferation are shown in table 2. Subretinal fluid (SRF) at the fovea was observed in seven eyes (25%) at 1 month and two eyes (7%) at 3 months. However, SRF was completely absent in all eyes at 6, 12, 24 and 36 months postoperatively.

Reconstruction of the ELM by centripetal bridging was first observed before reconstruction of the EZ and CIZ (figures 1 and 2). At 1, 3, 6 and 12 months, reconstruction of the ELM was seen in 14 eyes (50%), 21 eyes (75%), 24 eyes (86%) and 28 eyes (100%), respectively. The ELM line was completely restored continuously in all eyes at 1 year postoperatively, and was continuously reconstructed at 24 and 36 months postoperatively. The EZ was reconstructed centripetally after ELM bridging. At 1, 3, 6, 12, 24 and 36 months postoperatively, reconstruction of the EZ was observed in no eyes (0%), 2 eyes (7%), 3 eyes (11%), 5 eyes (18%), 15 eyes (54%) and 22 eyes (79%), respectively. Reconstruction of the CIZ was observed after EZ had been reconstructed. Reconstruction of the CIZ was not observed at 1 month or 3 months, but at 6, 12, 24 and 36 months it was observed in 1 eye (4%), 1 eye (4%), 4 eyes (14%) and 11 eyes (39%), respectively. The complete CIZ line was only observed in eyes with an intact ELM and EZ.

Figure 1

Long-term reconstruction of the photoreceptor layer at the fovea after MH repair in a 63-year-old woman. (A) A fundus photograph before surgery. Decimal VA was 0.1 in the left eye. (B) A SD-OCT image obtained before surgery showed a full-thickness MH. The ELM was visualised along the elevated outer retina around the hole. The minimum diameter of the hole was 296 µm. (C) A SD-OCT image obtained 1 month after surgery showed a restored ELM. A foveal hyper-reflective lesion indicating glial tissue (yellow asterisks) was observed at the inner retina above the ELM. The photoreceptor EZ and CIZ were disrupted. VA was 0.6. (D) A SD-OCT image obtained 3 months after surgery. The ELM was restored, although EZ and CIZ photoreceptors were disrupted. The glial tissue had disappeared. VA was 0.7. (E) A SD-OCT image obtained 6 months after surgery. The disrupted EZ remained, but its area had decreased. VA was 0.9. (F) A SD-OCT image obtained 12 months after surgery. The ELM and EZ were restored, but the CIZ was not completely reconstructed. VA was 1.0. (G) A SD-OCT image obtained 24 months after surgery. The ELM, EZ, and CIZ were reconstructed. VA was 1.2. (H) A SD-OCT image obtained 36 months after surgery. The ELM, EZ and CIZ remained reconstructed. VA was 1.2. 1M, 1 month; 3M, 3 months; 6M, 6 months; 12M, 12 months; 24M, 24 months; 36M, 36 months; CIZ, cone interdigitation zone; ELM, external limiting membrane; EZ, ellipsoid zone; MH, macular hole; SD-OCT, spectral-domain optical coherence tomography; VA, visual acuity.

Figure 2

Long-term reconstruction of the photoreceptor layer at the fovea after MH repair in a 62-year-old woman. (A) A fundus photograph before surgery. Decimal VA was 0.08 in the left eye. (B) A SD-OCT image obtained before surgery showed a full-thickness MH. The ELM was visualised along the elevated outer retina around the hole. The minimum diameter of the hole was 609 µm. (C) A SD-OCT image obtained 1 month after surgery showed disruption of the ELM, EZ and CIZ. A foveal hyper-reflective lesion indicating glial tissue (yellow asterisks) replaced all intraretinal layers. VA was 0.2. (D) A SD-OCT image obtained 3 months after surgery. The ELM was restored with bridging formation. Photoreceptors were disrupted in the EZ and CIZ. The glial tissue (yellow asterisks) was observed at the inner retina above the ELM. VA was 0.2. (E) A SD-OCT image obtained 6 months after surgery. The EZ and CIZ were still disrupted, but the relevant areas had decreased. The glial tissue (yellow asterisks) was still observed at the inner retina above the ELM but the area decreased. VA was 0.3. (F) A SD-OCT image obtained 12 months after surgery. The area of disrupted EZ and CIZ had decreased further. VA was 0.3. (G) A SD-OCT image obtained 24 months after surgery. The EZ and CIZ were slightly disrupted. The glial tissue (yellow asterisk) shifted upward to the inner retinal layer. VA was 0.3. (H) A SD-OCT image obtained 36 months after surgery. The ELM and EZ were reconstructed. The CIZ remained disrupted. The glial tissue had disappeared. VA was 0.4. 1M, 1 month; 3M, 3 months; 6M, 6 months; 12M, 12 months; 24M, 24 months; 36M, 36 months; CIZ, cone interdigitation zone; ELM, external limiting membrane; EZ, ellipsoid zone; MH, macular hole; SD-OCT, spectral-domain optical coherence tomography; VA, visual acuity.

At 1, 3, 6, 12, 24 and 36 months postoperatively, the presence of glial proliferation defined as a hyper-reflective lesion on SD-OCT was observed in 25 eyes (89%), 20 eyes (71%), 12 eyes (43%), 10 eyes (36%), 7 eyes (25%) and 3 eyes (11%), respectively (table 2). Thus, most eyes had glial proliferation at the fovea in the early postoperative period, but thereafter glial cells gradually decreased over time. Disappearance of the glial cells was first observed in the outer nuclear layer just above the ELM, and the remaining glial cells migrated upward to the inner retinal layer and eventually disappeared (figure 2). During this process, the EZ was concomitantly reconstructed with subsequent restoration of the CIZ.

Inner retinal defect

The inner retinal defect known as the DONFL was not detected before surgery. Variable area of the DONFL became evident within the area of ILM peeling 3 months after surgery (table 2 and figure 3). The area of the DONFL further increased by more than 10% in 7 (25%) eyes, was unchanged (within 10%) in 21 (75%) eyes and decreased by more than 10% in 0 (0%) eyes at 3 years compared with 6 months.

Figure 3

Changes in the inner retinal layers after macular hole repair in a 73-year-old woman. (A) En face SD-OCT and horizontal B-scanning SD-OCT before surgery revealed a full-thickness macular hole. Decimal VA was 0.1. (B) One month after surgery, multiple dark dots indicating DONFL had appeared within the area of internal limiting membrane peeling. The % area of the DONFL within a 6×6 mm area was 5.25%. With regard to the ELM, EZ and CIZ, photoreceptor status was ELM(+)/EZ(−)/CIZ(−). VA was 0.1. (C) Three months after surgery, the appearance of DONFL was clearly evident. The % area was 8.14%. Photoreceptor status was ELM(+)/EZ(−)/CIZ(−). VA was 0.2. (D) Six months after surgery, the % area of DONFL appearance was 8.99%. Photoreceptor status was ELM(+)/EZ(−)/CIZ(−). VA was 0.2. (E) Twelve months after surgery, the % area of DONFL appearance was 9.29%. Photoreceptor status was ELM(+)/EZ(−)/CIZ(−).VA was 0.2. (F) Twenty-four months after surgery, the % area of DONFL appearance was 8.60%. Photoreceptor status was ELM(+)/EZ(−)/CIZ(−). VA was 0.2. (G) Thirty-six months after surgery, the % area of DONFL appearance was 8.61%. Photoreceptor status was ELM(+)/EZ(−)/CIZ(−). VA was 0.2. 1M, 1 month; 3M, 3 months; 6M, 6 months; 12M, 12 months; 24M, 24 months; 36M, 36 months; CIZ, cone interdigitation zone; DONFL, dissociated optic nerve fibre layer; ELM, external limiting membrane; EZ, ellipsoid zone; SD-OCT, spectral-domain optical coherence tomography; VA, visual acuity.

Predictive factors for visual acuity at 3 years

Preoperative factors significantly associated with BCVA at 3 years included preoperative BCVA (p=0.019) and preoperative MH size (p=0.014) (table 3). Specifically, better preoperative BCVA and smaller MH size were associated with better BCVA at 3 years. The area of the DONFL was not associated with visual acuity at 3 years.

Table 3

Univariate regression analyses of associations between postoperative BCVA at 3 years and variables

The median time at which ELM bridging first appeared was 2 months (range 1–12 months). The median time at which glial cells disappeared was 6 months (range 1–36 months). Both the time until ELM bridging and glial cell disappearance were significantly associated with better BCVA (p<0.001 and p=0.033, respectively) and photoreceptor reconstruction (p=0.008 and p<0.001, respectively) at 3 years, indicating that eyes that achieved ELM reconstruction and glial disappearance faster had a higher chance of achieving better BCVA and photoreceptor restoration at 3 years (figure 4). Postoperative BCVA was significantly correlated with photoreceptor integrity at 3 years.

Figure 4

Schematic representation of the hypothesis of the reconstructive process at the fovea after macular hole repair. 1. Macular hole before surgery. 2. In the early postoperative period, the foveal defect was filled by proliferating glial cells that replaced the entire intraretinal layer. 3. Centripetal bridging of the ELM. Glial proliferation was observed at the inner retina above the ELM. 4. The numbers of glial cells reduced with proximity to the inner retinal layer, and they ultimately disappeared. 5. Following reconstruction of the ELM and glial elimination, the EZ was completely reconstructed. 6. Following the reconstruction of the EZ, the CIZ was subsequently reconstructed. Eyes exhibiting more rapid ELM reconstruction were more likely to exhibit complete reconstruction of the photoreceptors and better best-corrected visual acuity. CIZ, cone interdigitation zone; ELM, external limiting membrane; EZ, ellipsoid zone ; RPE, retinal pigment epithelium.

We categorised the eyes in the study into three groups based on integrity of the foveal photoreceptor layer on SD-OCT at 3 years; ELM(+)/EZ(+)/CIZ(+), 11 eyes with restoration of ELM, EZ and CIZ; ELM(+)/EZ(+)/CIZ(−), 11 eyes with restoration of ELM and EZ but with disrupted CIZ; and ELM(+)/EZ(−)/CIZ(−), 6 eyes with restoration of ELM, and disrupted EZ and CIZ (table 4). At 3 years, postoperative BCVA was best in ELM(+)/EZ(+)/CIZ(+) eyes. In addition, ELM(+)/EZ(+)/CIZ(+) eyes exhibited significantly earlier reconstruction of the ELM and disappearance of glial cells postoperatively.

Table 4

Visual acuity and foveal microstructure 36 months after surgery

Discussion

In the current study, we evaluated long-term microstructural and visual changes over time after MH repair. The integrity of the photoreceptor layer determined by reconstruction of the ELM, EZ and CIZ progressively improved during the 3-year follow-up period. Concomitantly, visual acuity improved during the 3 years, indicating that both microstructure and visual acuity improve continuously for up to 3 years after surgery.

The current study elucidated how the foveal microstructure is reconstructed and return to normal retinal architecture in the long term in surgically closed MHs (figure 4). In the early postoperative period at 1 month, most eyes (25/28, 89%) showed proliferating glial cells defined as a hyper-reflective lesion filling the foveal defect as previously described.8 14–16 In 14 (56%) of 25 eyes, the foveal defect was filled by the proliferating glial cells that replace the entire intraretinal layer without bridging of the ELM as shown in figure 2. In contrast, 3 eyes without glial proliferation and 11 (44%) of 25 eyes with glial proliferation exhibited reconstruction of the ELM at the fovea, indicating that centripetal bridging of the ELM proceeded faster than glial cell proliferation into the foveal defect. In such cases, glial proliferation was absent or was only evident at the inner retina above the ELM, as shown in figure 1. Eyes with faster ELM reconstruction had a higher chance of achieving complete reconstruction of the photoreceptors and better BCVA at 3 years. Although the importance of ELM reconstruction for better visual acuity at 12 months has been reported,8 current study indicated that the status of the ELM in the early postoperative period was predictive of 3-year postoperative BCVA in surgically closed MHs.

Reconstruction of the ELM was achieved within 12 months postoperatively in all eyes in the current study, although the time required for reconstruction varied. Following the reconstruction of the ELM, EZ reconstruction occurred with subsequent restoration of the CIZ. No eyes with a disrupted ELM had an intact EZ or CIZ, and no eyes exhibited a restored CIZ without reconstruction of the ELM and EZ. Those results indicate that the foveal microstructure is reconstructed in the order ELM, EZ and then CIZ, similar with the reconstructive process reported in epiretinal membrane (ERM), retinal detachment and uveitis.24–26 During this process, glial proliferation in the outer nuclear layer (ONL) decreased to just above that of the ELM, and the remaining glial cells shifted toward the inner retinal layer and eventually disappeared in most eyes (89%) at the 3-year time-point. The time until glial cells disappeared was significantly associated with complete photoreceptor reconstruction and visual acuity at 3 years. Therefore, the disappearance of glial tissue may facilitate structural and functional recovery of the photoreceptor cell bodies at the ONL, resulting in recovery of cone inner and outer segments as indicated by restored EZ and CIZ.

At 1, 2 and 3 years postoperatively, complete reconstruction of photoreceptors defined as reconstruction of the ELM, EZ and CIZ in conjunction with the disappearance of glial proliferation at the fovea was achieved in 1 eye (4%), 4 (14%) eyes and 11 (39%) eyes, respectively. Therefore, even though successful MH closure was achieved in the early postoperative period, it took several years for the photoreceptor layer to undergo complete subsequent restoration. The eyes with completely reconstructed photoreceptors at 3 years had better visual acuity than those without. This result is consistent with a previous report suggesting that reconstruction of the CIZ, previously described as the COST line, is important for visual acuity 12 months after surgery.13 Therefore, the speed of complete photoreceptor reconstruction may account for faster visual recovery, and delayed reconstructive changes may result in limited visual recovery in surgically closed MHs (figure 4). However, because postoperative visual acuity has been shown to improve continuously for up to 5 years in this context,27 further photoreceptor restoration and concomitant visual improvement may subsequently occur in some eyes in the present study that did not exhibit complete restoration at 3 years.

The current study suggests that promoting early ELM bridging may promote more rapid visual recovery after MH repair. In cases involving central nervous system trauma such as spinal cord injury, the repair is initiated by bridging of glial cells followed by further axon regeneration.28 In that study, local delivery of connective tissue growth factor (CTGF) protein increased the bridging activity of glial cells whereas decreased CTGF resulted in deficits in glial bridging and inhibited spinal cord regeneration. Although it is not clear whether CTGF is involved in the healing process of MH, investigation of the mechanism of the initial bridging events and elucidating the way to promote them may yield further insights into the treatment of MH.

The limitations of the present study include the relatively small sample size and the fact that it was retrospective. In addition, the precise mean time of ELM reconstruction and glial elimination may be difficult to determine since we did not examine patients every months. Therefore, further studies are recommended to confirm our results. Nevertheless, the study provided valuable information pertaining to the foveal microstructure reconstruction process, and suggested that faster ELM reconstruction and glial elimination in the early postoperative period may predict complete photoreceptor restoration and visual recovery in patients with surgically closed MHs. Without definitive strategy for facilitating photoreceptor reconstruction especially in eyes with large MHs resulting in delayed reconstruction (table 3), our findings may only provide a better understanding of the reconstructive process and a possible reason for the differences in visual recovery after MH surgery. However, this information may be valuable for developing novel therapeutic strategies to promote foveal reconstruction and improve visual recovery after MH surgery in the future.

In conclusion, early postoperative restoration of the ELM and disappearance of glial cells seems important for morphologic and functional recovery of the foveal photoreceptor layer in closed MHs. Long-term visual acuity at 3 years was significantly correlated with integrity of the photoreceptor layer, defined as restoration of the ELM, EZ and CIZ.

References

Footnotes

  • Contributors Study design, data acquisition and statistical analysis: MK and TW; interpretation of data: MK, TW, KeN, HS; writing (original draft): MK; writing (review and editing): TW, KeN, HS, KoN; final approval of manuscript: all authors; supervision: HS and KoN.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval The study adhered to the tenets of the Declaration of Helsinki and was approved by the institutional review board committee of Osaka University Medical School.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • At a glance
    Keith Barton James Chodosh Jost B Jonas