OCT proves that vitreomacular adhesion is significantly more likely to develop vision-threatening retinal complications than vitreomacular separation

The Institutional Review Board at Cheng Hsin General Hospital, Taipei, Taiwan [approval number: (607)106–15; Supplementary file 1] gave permission for this chart review study. During the baseline examination by SD-OCT (Spectralis, Heidelberg Engineering, Heidelberg, Germany) between January 1st 2016 and July 31st 2017, one-hundred and twenty-four eyes (one or two eyes per individual) with either vitreomacular separation (n = 8) or VMA with perifoveal vitreous detachment (n = 116) were selected and followed up by SD-OCT monthly. The scan type was a volume (cube) scan. Additionally, scan length [6 × 6 mm (20°× 20°)], scan density (25 line scan/20°× 20°) and resolution [3.9 μm (digital)] were set as indicated. The incidence of floaters was presently assessed. Patients were excluded if their medical records indicated that they were not asked “are you suffering from floaters in your vision?”. The incidence rates for various complications related to either of the defined conditions were also evaluated. To evaluate the complication incidence, the cases with defined complications were also pinpointed after an initial SD-OCT was able to prove diagnosis. Observation of the disease course was carried out in order to establish how the two defined entities are related to the various defined complications. Furthermore, patients were excluded if they had any of the following previous ocular diseases, there were glaucoma, proliferative diabetic retinopathy, uveitis, ocular trauma, vitreous hemorrhage as well as aforementioned complications. In addition, patients that did not undergo regular follow-up were excluded.

The medical records included whether or not patients had an onset of VMA or vitreomacular separation related symptoms, such as floaters, a distorted image and blurred vision. In addition to SD-OCT, every enrolled eye received conventional ocular/vitreoretinal examinations by indirect ophthalmoscope [6], and slit-lamp biomicroscopy with the help of a contact lens to detect a range of signs such as the presence of a retinal break, a retinal atrophic hole and lattice degeneration.

When VMA was present, in contrast to unilateral posterior hyaloid detachment (J-shaped pattern), the partial posterior vitreous might be detached both nasally and temporally, but persistent attachment could be observed to be maintained at the fovea, which resulted in a V-shaped pattern [11].

For clinical practicability, the vitreomacular angle of the present cases were measured as follows. Two lines were drawn from the foveal center: one horizontal line along the inner retina and the other oblique line along the separated posterior hyaloid. The angle between the inner retinal surface and posterior hyaloid was then directly measured using the On-Screen Protractor Program (Minimum Java Runtime Environment version 1.7, Gnu’s Not Unix General Public License version 3); the largest angle for VMA, with either V-shaped or J-shaped posterior hyaloids, of an eye that received both vertical and horizontal OCT scans was defined as the vitreomacular angle of the examined eye. The directly measured critical vitreomacular angle was 13.3°, which is considerably lower than that 45.0° of the study of Tsai et al. [12]. The above vitreomacular angle was retrieved by utilizing the point on the receiver operating characteristics curve with a minimum value of [(1 − sensitivity)² + (1 − specificity)²] for macular hole in all of the selected eyes.

Under the present hypothesis that there are more complications with VMA [8], the one-tailed Fisher’s exact test was used to analyze the enrolled eye results obtained during the present study. This allowed an analysis of whether VMA shows a higher incidence of defined vision-threatening retinal complications than vitreomacular separation. All data collected during this study were analyzed using Statistical Package for the Social Sciences 20.0 (Statistical Package for the Social Sciences, Incorporation, Chicago, Illinois, United States of America). The data are expressed as mean ± standard deviation. The unpaired Student’s t-test was used when comparing two independent groups. A probability value of < 0.05 was considered significant.

Compared to the results of Abdolrahimzadeh et al. (a frequency of 85% for incomplete posterior vitreous detachment) [10], 93.5% of the eyes in the present study presented with VMA (116 out of 124). Ito et al. [13] stated that posterior vitreous detachment starts from the perifoveal region, next it continues to the fovea and finally, it develops into complete posterior vitreous detachment. Nevertheless, how the exact dynamic evolution of VMA takes place needs to be further investigated.

As implied by previous reports [3, 12, 14‐21], VMA can be complicated by a variety of defined visual dysfunction retinal diseases. The present results support the hypothesis that the incidence of defined retinal complications among VMA eyes seems to be significantly greater than that among vitreomacular separation eyes (Tables 1, 2 and 3; Figs. 1, 2, 3 and 4; Supplementary file 2). This agrees with the widely accepted concept that the former have persistently stronger adhesions that pull at the fovea [3, 12, 14‐21], and induce detrimental complications. By way of contrast, the latter group seems to have lost this strong traction [3, 12, 17, 22]. Thus, accurate diagnosis, long term observation and appropriate management needs to be emphasized when treating patients with VMA associated adverse sequelae. As mentioned earlier, SD-OCT seems to be a clinically useful and an easily accessible approach that is able to detect various specific vitreomacular interface abnormalities that might not be observed by ultrasonography and indirect ophthalmoscopy [3, 12, 17, 22].

As mentioned at the final paragraph of the Results, in contrast to the defined low hyperopia (1.0 ± 2.5) of vitreomacular separation, low myopia (− 1.5 ± 4.1) or high myopia (− 5.4 ± 9.7/− 7.0 ± 11.7) seems to be related, respectively, to VMA without complications and VMA with complications/macular holes. A further larger scale study is needed in order to investigate the role of the high myopia in the development of VMA-induced complications/macular holes.

Vitreomacular traction seems to be able to produce tangential (horizontal) [12, 22] and anteroposterior (vertical) traction [12, 22] that causes foveal conformational changes and subsequent visual dysfunctional symptoms such as blurred vision, floaters/flashes and/or a distorted image. In a twenty-year cross-sectional study, namely the Beaver Dam Eye Study [21], it was found that the incidence of vitreomacular traction was 1.6% among all eyes with maculae scanned by SD-OCT. In contrast, the incidence of vitreomacular traction was presently estimated to be 0.8% (1 out of 124 eyes; Table 1). Classic vitreomacular traction could be divided into two types: V-shape and J-shape [11]; while, another method subclassified vitreomacular traction into focal (vitreomacular attachment diameter ≤ 1500 μm) and broad subtypes (> 1500 μm) [23]. Kishi et al. demonstrated that the margin of the 1500 μm foveal region had a firm vitreoretinal adhesion [24]. As recommended by Bottos et al. [25], the classification of vitreomacular traction, based on the diameter of the adhesion, appears to be more precisely than that based on its morphology.

Over the past two decades, the literatures [3, 12, 22, 26‐32] available on the incidence of adverse retinal disorders such as macular hole that are associated with VMA has been limited. Furthermore, the findings are quite variable due to differences in a range of factors; these include the study design (cross sectional or longitudinal; prospective or retrospective), the diagnostic tools used (ultrasound, OCT, etc) and the different criteria utilized to define the classification and diagnosis of VMA or vitreomacular separation (with or without acute symptoms) [2‐4, 8, 12, 22, 27]. Additionally, a recent retrospective multicenter study [27] has reported that full thickness macular holes occurred in 7 out of 168 eyes (4.2%) with vitreomacular traction. This differs to some extent to the present retrospective study in which the incidence of full thickness macular holes was 6.5% (8 out of 124 eyes). Compared to the present results, it has been suggested that VMA and vitreomacular traction appear to be two different entities in view of the difference in the incidence (6.5% vs. 4.2%) of full thickness macular hole [27]. There are also differences in SD-OCT-based findings, namely without or with interruption of all foveal retinal layers and distortion of the foveal contour (normal depression vs. dome-shaped elevation; see also Supplementary file 2) [2, 3, 11, 12, 27]. In contrast to the vitreomacular traction associated idiopathic full thickness macular hole that occurs sometimes in the presence of foveal detachment due to vertical traction [12], none of the present eight VMA eyes with subsequent full thickness macular hole complications showed foveal detachment. Gass [28] has postulated that premacular vitreous tangential contraction might lead to a macular hole. This hypothesis regarding tangential vitreoretinal traction has been replaced by a newer theory, which suggests that macular holes are the result of anteroposterior vitreofoveal traction in the perifoveal area [12, 22, 29, 30]. However, the postulation of Gass seems to explain better the present finding, which is that six out of eight eyes (Tables 1 and 2) displayed V-shaped VMAs (Figs. 2a, 3a and 4a) and that this state might have induced tangential traction resulting in the complication, specifically, full thickness macular hole. Compared to VMA with unilateral hyaloid membrane separation, V-shaped VMAs perhaps indicate a bilateral pull with greater traction force. This might then be more likely to induce macular hole formation. However, this hypothesis needs further investigation. In a prospective report of Johnson, Van Mewkirk and Meyer [32], it was shown by the SD-OCT that perifoveal vitreous detachment was the initial pathogenesis associated with grade 1–2 macular hole (n = 26). Moreover, persistent vitreofoveolar adherence, namely VMA [3], was evident in 18 eyes of those eyes with defined macular hole. This evidence [32] also strongly supports why eyes with VMA in the present study have a significant greater possibility of ocular complications, such as macular hole (n = 8), than eyes with vitreomacular separation. In contrast to the above, the study of Carrero [8] was prospective and the recruited eyes had acute symptoms; this might have contributed to his finding that macular holes were not found.

Tsai et al. [12] and Theodossiadis et al. [22] have reported that the vitreomacular angles between the posterior vitreous membrane and the horizontal lines of the retina (inner retinal surface [12] or retinal pigment epithelium [22]) are proportionally associated with the severity of vitreomacular traction [22, 31, 33]. Their findings [12. 22] seem to support the present results whereby the mean vitreomacular angle (Fig. 2; Table 2) of the eight eyes with VMA and consequent sequelae, i.e. macular holes, was significantly (p = 0.04) higher at 24.2°, with possible stronger traction force, compared to those of patients without adverse complications (n = 104; at 13.6°). This is not inconsistent with a recent prospective study [31] where out of fifty one cases of VMA (2%), one case (Case 2) with an initial vitreomacular angle 27° nasally or 20° temporally [31] [vs. 31.2° nasally or 24.6° temporally (measured by the present measurement method)] showed an evolved vitreofoveal separation associated with a highly suspected complicated macular hole. This was in spite of an intact external limiting membrane (presumably due to healing process), 10 days after the last examination in the VMA stage. To determine whether VMA eyes with a vitreomacular angle more than the present critical vitreomacular angle of 13.3° (mean) have a greater possibility of associated macular hole complication requires a larger scale investigation. On the other hand, Tsai et al. [12] and Spaide et al. [26] have suggested that the wider the diameter of the VMA, the stronger the traction exerted on the fovea. However, the adhesion diameters (mean: 239.0 ± 52.5 μm) in the present study are somewhat difficult to be definitively measured, for example the broad typed VMA [3] nasal to the fovea presented in Case 2 (Fig. 1a). V-shaped vitreomacular traction has been reported to lead to tractional macular hole [25]. This seems also to be the case in our study and the above study when there are initial V-shaped VMAs, specifically for six out of 8 eyes with complicated macular holes in the present study (Tables 1 and 2) and for the above mentioned Case 2 with a suspected complicated macular hole in Figure 6 (middle) of a recent publication [31]. Classification of VMA induced macular hole, based on the morphology, seems to be at present the most practical approach, rather than one based on the diameter of the adhesion.

Most incomplete posterior vitreous detachments with retinal detachment are likely to be diagnosed as being the rhegmatogenous type [14]. Abdolrahimzadeh et al. [10] suggested that some peripheral vitreoretinal adhesions were not easily visible clinically and that retinal breaks could occur by sudden traction. It is reasonable to postulate that the posterior vitreous membrane might be pulled from its residual peripheral attachment sites, which will then cause retinal tears and this could then extend into the rhegmatogenous retinal detachment. As demonstrated by the present study, because the incidence of retinal detachment is 1.6% (2 out of 124 eyes), retinal detachments are likely to be due to a new onset of peripheral retinal lattice degeneration, which is then followed by a vitreous traction-induced retinal break. This seems to be the situation with Case 3 and 9. However, the ability to detect the peripheral lesions, such as a break, is beyond the scan length limitation of SD-OCT.

Apart from those limitations, the present study does indicate that there is a greater proportion (93.5%) of VMA than vitreomacular separation among these enrolled eyes. Moreover, as compared to vitreomacular separation, VMA does have a significantly greater possibility of progression into various defined vision-threatening retinal complications. Whether eyes with VMA might undergo horizontal (tangential) traction, with this playing a role in the pathogenesis of a consequent macular hole, needs further evaluation. What is more, whenever general ophthalmologists identify a patient with VMA, it might be useful that regular follow-ups are carried out, especially by means of tests such as monthly SD-OCT, if this is available, that targets the vitreomacular interface. Moreover, other conventional fundus examinations that can detect peripheral retinal lesions, including breaks or lattice degeneration, need to be carried out, if possible with the pupils being fully dilated. Prompt referral to a vitreoretinal specialist is warranted when any defined visual dysfunction retinal complication occurs.