Int Ophthalmol DOI 10.1007/s10792-013-9884-6
ORIGINAL PAPER
Reduced effect of anti-vascular endothelial growth factor agents on diabetics with vitreomacular interface abnormalities Dan Yoon • Irene Rusu • Irene Barbazetto
Received: 30 September 2013 / Accepted: 13 November 2013 Ó Springer Science+Business Media Dordrecht 2013
Abstract This study investigated whether the presence of vitreomacular interface abnormalities (VMIAs) affects the improvement in visual acuity and edema of patients with diabetic macular edema (DME) who received three anti-vascular endothelial growth factor (VEGF) injections. Fifteen eyes of 11 patients with clinically significant macular edema were retrospectively divided into either the control group (only DME) or the experimental group (DME and VMIA) based on optical coherence tomography images. We defined VMIA patterns as epiretinal membrane and/or anomalous vitreomacular adhesion. Changes in central macular thickness (CMT), total macular volume (TMV), and best-corrected visual acuity (BCVA) from the baseline to post third injection were compared between the two groups. After the third injection, the decreases in CMT and TMV were not statistically different between the two groups. The improvement in BCVA was larger in the control group (0.1742 ± 0.0508 logMAR) than in the experimental group (0.0766 ± 0.0562 logMAR; p \ 0.01). Our study showed that after the third anti-VEGF D. Yoon (&) I. Rusu Department of Ophthalmology, New York University School of Medicine, 462 First Ave. NBV 5N18, New York, NY 10016, USA e-mail: [email protected]; [email protected] I. Barbazetto Vitreous-Retina-Macula Consultants of New York, New York, NY, USA
injection, the BCVA of patients with both DME and VMIAs improved significantly less than that of patients with only DME. Our results suggest that VMIAs may play a crucial role in reducing the therapeutic effects of anti-VEGF agents. Keywords Diabetic macular edema Vitreomacular interface abnormality Anti-VEGF Epiretinal membrane
Introduction Diabetic macular edema (DME) is one of the most common causes of visual impairment in patients with diabetes [1]. Vitreomacular interface abnormalities (VMIAs) appear to occur in 7–16 % of eyes with DME, and the annual incidence may be as high as 4.5 % [2, 3]. VMIAs include epiretinal membrane and/or anamolous vitreomacular adhesion, which is sometimes referred to as vitreomacular traction or posterior vitreomacular separation and is characterized as having features of a taut, thick posterior hyaloid [4]. Both features are often better identified using optical coherence tomography (OCT) in addition to conventional fundus examination [4]. The presence or absence of these interface abnormalities potentially affect the response to treatment in DME patients. For many years, the gold standard treatment for DME was focal laser photocoagulation, which can
123
Int Ophthalmol
reduce the risk of moderate visual loss, defined as a loss of C15 letters, by 50 % compared to those with deferred treatment [5]. Nevertheless, after 3 years of treatment, approximately 12 % of treated patients still experience moderate visual loss. This unresponsiveness is thought to be due to persistent macular edema [5]. Recently, VMIAs were found to be associated with as high as 52–67 % of cases of persistent macular edema that was unresponsive to focal laser photocoagulation [3, 4, 6]. Whether VMIAs play a role in determining the efficacy of newer treatments, specifically anti-vascular endothelial growth factor (VEGF) treatments, has not been studied extensively. Ranibizumab (Lucentis, Genentech Inc, South San Francisco, CA, USA) was approved by the Food and Drug Administration (FDA) for the treatment of DME based on two phase III trials, RISE and RIDE [7]. Furthermore, recent results of the Bevacizumab or Laser Therapy (BOLT) study showed that bevacizumab (Avastin; Genentech Inc, South San Francisco, CA, USA) was also superior to laser treatment [8]. Only one study to date by Wu et al. [9] suggests that DME with features of VMIAs may be associated with poor results after anti-VEGF treatments. One hypothesis for this trend is that traction caused by VMIAs may reduce the potential benefits of treatments. Support for this theory comes from growing evidence that vitrectomy and the resulting separation of posterior hyaloid from traction may help in treating persistent DME with VMIAs [10, 11]. This retrospective study further investigated whether the presence of VMIAs affects the improvement in visual acuity and edema from anti-VEGF injections. In particular, while Wu et al. mostly followed patients who only received one injection for the duration of 3 months, we studied whether changes in best-corrected Snellen visual acuity (BCVA), central macular thickness (CMT), and total macular volume (TMV) were still reduced in patients with VMIAs after three injections. Although the frequency of anti-VEGF regimen for DME remains controversial, most clinical trials were based on monthly injections for the first three injections. Since many clinicians follow this practice, this study would give us additional information of whether VMIAs affect the efficacy of anti-VEGF in treating DME under the current treatment timeline.
123
Methods The study was approved by the Institutional Review Board of North Shore - LIJ Health System. Patient selection We reviewed the charts of patients from a private practice retinal referral center in an urban setting with the billing diagnosis of DME who received anti-VEGF intravitreal injections between 2010 and 2012. Each anti-VEGF injection was either 0.3 mg ranibizumab or 1.25 mg bevacizumab. The inclusion criteria were (1) clinically significant macular edema as defined by the early treatment of diabetic retinopathy study (ETDRS) [12], (2) CMT of [250 lm by OCT; (3) age C18 years, and (4) received three anti-VEGF injections consecutively with a maximum duration of 2 months between each injection. The exclusion criteria were (1) ocular disease apart from diabetic retinopathy such as glaucoma, cataract, or active vitreous hemorrhage, (2) previous intravitreal injection of anti-VEGF agents or steroids prior to the first documented anti-VEGF injection, (3) previous macular focal laser therapy (those with panretinal photocoagulation (PRP) were still included), (4) simultaneous injection of steroids with anti-VEGF, (5) interruption of antiVEGF treatments with focal laser therapy or PRP, and (6) BCVA at the time of first treatment that was better than 20/30 (due to a ceiling effect) or worse than 20/200 (due to the difficulty in accurately quantifying visual acuity). All patients underwent a complete ophthalmic examination that included BCVA recording using Snellen charts, slit-lamp examination, intraocular pressure measurement, and fundus examination with indirect ophthalmoscopy. Central retinal characteristics were evaluated using spectral domain OCT (SD-OCT) (Spectralis; Heidelberg Engineering, Heidelberg, Germany). All OCT images were obtained by volume scan option. Due to the retrospective nature of the study, the scanned area ranged from 5.9 9 4.4 to 9.1 9 7.5 mm. The number of B scans also ranged from 19–49 (spaced 120–240 lm apart) with automatic real time (ART) mean ranging from 13 to 23. OCT grading Based on the grading of the OCT images by two independent reviewers (DY and IR) patients were
Int Ophthalmol
divided into either the control group (only DME) or the experimental group (DME and VMIA). We defined VMIA patterns as epiretinal membrane (ERM) and/or anomalous vitreomacular adhesions. Both reviewers agreed on the categorization of each scan. In 24 of 30 OCT scans, each follow-up image was automatically registered at the time of image acquisition to the first visit image and, therefore, the ETDRS grids for CMT and TMV measurements were automatically placed in identical locations for both the first visit image and follow-up image. In the remaining six OCT scans, placement of the ETDRS grids was adjusted by reviewers based on landmarks seen on corresponding IR images so that the ETDRS grids would be placed in approximately identical locations. Both reviewers agreed on all manual placements of ETDRS grids. Epiretinal membrane (ERM) On OCT, ERM was defined as thin, highly reflective bands anterior to the neurosensory retina as defined by Wilkins et al. [13]. The OCT configurations of ERM were defined as either globally adherent or partially separated with focal areas of macular attachments [13]. There is no minimum size requirement for ERM in the literature. In order to make the categorization of images objective, we categorized a hyperreflectant band [1.5 mm in horizontal size in any B-scan to be an ERM. We used the threshold of 1.5 mm, because the dimension of the smallest ERM found with threedimensional SD-OCT among patients with DME was 1.5 9 3 mm2 [14].
Anomalous vitreomacular adhesion Anomalous vitreomacular adhesion was defined as an evident hyperreflective band that adhered to the surface of the retina at specific sites, was elevated elsewhere off the surface, and continuous with the posterior vitreous surface, in agreement with the report by Ghazi et al. [4]. Statistical tests BCVA, CMT of 1 mm, and TMV of 3 mm at the time of first injection (baseline) and post third injection were noted. BCVAs were converted to the logarithm of minimal angle of resolution (logMAR) format for statistical manipulation. The final BCVA, CMT, and TMV after the third injection were compared to the baseline BCVA, CMT, and TMV at the time of the first injection using paired t tests. Changes in the BCVA, CMT, and TMV from baseline to post third injection were compared between the control group (only DME) and the experimental group (DME with VMIA). Normality was checked, and unpaired t test was used (SPSS; IBM Corporation, Armonk, NY, USA) for comparing changes in BCVA, CMT, and TMV between the control group and the experimental group.
Results Baseline characteristics Patient characteristics are summarized in Table 1. A total of 15 eyes from 11 patients (8 male and 3
Table 1 Baseline characteristics of the 15 eyes from 11 patients with diabetic macular edema Control
Experimental
p value
No. of patients
4
7
Age (±SD)
56 (±8)
67 (±5)
0.078
Sex (male:female) Duration of DM (years ± SD)
3:1 14.50 (±7.85)
5:2 14.00 (±7.21)
0.898 0.920
HgA1C (± SD)
7.43 (±0.81)
6.75 (±0.58)
0.293
Follow-up (days ± SD)
106 (±13)
98 (±19)
0.400
No. of days between each injection
35.2 (±4.2)
32.8 (±6.3)
0.400
No. of eyes with three injections
5
10
Stage (NPDR:PDR)
4:1
6:4
0.439
No. of eyes with PRP
0
3
0.171
DM diabetes mellitus, NPDR nonproliferative diabetic retinopathy, PDR proliferative diabetic retinopathy
123
Int Ophthalmol Table 2 Change in best-corrected visual acuity (BCVA), central macular thickness (CMT) of 1 mm, and total macular volume (TMV) of 3 mm from baseline to post third injection Control (mean ± SD)
Experimental (mean ± SD)
Baseline BCVA
0.4165 ± 0.1386
0.4497 ± 0.1245
Post third injection BCVA
0.2423 ± 0.0924
0.3732 ± 0.1350
P value (Post third injection - baseline)
\0.01
\0.01
p value (control vs. experimental) 0.664
\0.01
Change in BCVA
0.1742 ± 0.0508
0.0766 ± 0.0562
Baseline CMT
509.8 ± 181.8
471.6 ± 107.8
Post third injection CMT
388.8 ± 151.2
389.1 ± 113.6
p value (post third injection-baseline)
0.098
\0.05
Change in CMT
121.0 ± 126.2
82.4 ± 106.8
0.581
Baseline TMV
3.34 ± 0.6
3.07 ± 0.45
0.396
Post third injection TMV
2.74 ± 0.50
2.71 ± 0.41
p value (post third injection-baseline) Change in TMV
\0.05 0.60 ± 0.32
\0.01 0.35 ± 0.34
0.683
0.229
female) with nonproliferative (n = 10) and proliferative diabetic retinopathy (n = 5) were enrolled in the study. The average duration (±SD) of diabetes was 14.18 ± 7.05 years, and the average HgA1c level (±SD) was 7.04 ± 0.72. The average age (±SD) of the patients was 63 ± 8 years. The mean follow-up was 100.8 ± 16.91 days. Three eyes that received prior PRP were included in this study. In the experimental group, four eyes had anomalous vitreomacular adhesion only, five eyes had ERM only, and one eye had both. Of those with ERM, one was characterized as partially separated with focal attachment and five were characterized as globally adherent. In the 11 patients who were included in the above analysis, no feature of the anomalous vitreomacular adhesion or epiretinal membrane changed during the three injections.
Fig. 1 Changes in best-corrected visual acuity (BCVA). The white bar shows the improvement of the control group (0.1742 ± 0.0508 (SD; logMAR) and the black bar shows the experimental group (0.0766 ± 0.0562 (SD; logMAR; p \ 0.01). The error bars denote the standard error of the mean (SEM)
Change in BCVA
Change in CMT
The baseline BCVA (±SD) was not statistically different between the control group (0.4165 ± 0.1386 logMAR) and the experimental group (0.4497 ± 0.1245 logMAR; p = 0.664; Table 2). After three injections, the final BCVA was statistically different from the baseline BCVA for both the control (0.2423 ± 0.0924 logMAR; p \ 0.01) and the experimental group (0.3732 ± 0.1350 logMAR; p \ 0.01). The improvement in BCVA was larger in the control group (0.1742 ± 0.0508 logMAR) than in the experimental group (0.0766 ± 0.0562 logMAR; p \ 0.01; Fig. 1).
The baseline CMT (±SD) was not statistically different between the control group (509.8 ± 181.8 lm) and the experimental group (471.6 ± 107.8 lm; p = 0.683; Table 2). After three injections, the final CMT was statistically different from the baseline CMT for the experimental group (389.1 ± 113.6 lm; p \ 0.05) but not for the control group (388.8 ± 151.2 lm; p = 0.098). The decrease in CMT was not statistically different between the control group (121.0 ± 126.2 lm) and the experimental group (82.4 ± 106.8 lm; p = 0.581).
123
Int Ophthalmol
Change in TMV The baseline TMV (±SD) was not statistically different between the control group (3.34 ± 0.6 mm3) and the experimental group (3.07 ± 0.45 mm3; p = 0.396; Table 2). After three injections, the final TMV was statistically different from the baseline TMV for both the control group (2.74 ± 0.50 mm3; p \ 0.05) and the experimental group (2.71 ± 0.41 mm3; p \ 0.01). The decrease in TMV was not statistically different between the control group (0.60 ± 0.32 mm3) and the experimental group (0.35 ± 0.34 mm3; p = 0.229).
Discussion Our study showed that after the third anti-VEGF injection, the BCVA of patients with both DME and VMIAs improved significantly less than that of patients with only DME. There was no change in the features of the anomalous vitreomacular adhesions or ERM during the treatment process. Our results add to the previous evidence that suggested that patients with VMIAs did not respond as well to anti-VEGF injections as those with only cystoid macular edema (CME) after one treatment. [9] We showed the reduced effect of antiVEGF after multiple injections in order to mimic the clinical setting used in many practices. Our results suggest that VMIAs play a crucial role in reducing the therapeutic effects of anti-VEGF agents. In contrast to the significant difference in BCVA improvements between the control group and the experimental group, the decreases in CMT and TMV after three injections were not significantly different between the two groups. This result was surprising since a similar trend was expected among all three measures based on previous results. For example, Wu et al. [9] previously showed that BCVA, CMT, and TMV did not improve for DME patients with VMIAs after receiving one injection while all three measures improved for those with only CME. In our study, all three measures of DME patients with VMIAs improved after three injections, but only BCVA was improved by a significantly less degree when compared to the control group. These results suggest that while three injections may reduce the edema thickness for DME patients with VMIAs, there are multiple factors in
addition to the reduction in edema thickness that determine the final improvement in vision. Among many possible contributions regarding VMIAs, there are two main hypotheses to explain why either ERM or vitreomacular adhesions interfere with the therapeutic effects of anti-VEGF agents in DME—mechanical and physiological. Mechanically, these abnormalities cause traction on the macula [15]. The tractional forces of a taut, thickened posterior hyaloid pull the retina forward and may even produce a deep subretinal space where fluid can pool [16, 17]. Based on the presence of multilayered cellular components in excised ERM from patients with DME, it is hypothesized that the accumulation of chemoattractant cytokines leads to the migration of cells to the posterior hyaloid and that these cells play a crucial role in creating tractional forces [18, 19]. The tractional force hypothesis is also supported by findings that the benefits of vitrectomy may be confined to patients with OCT signs of macular traction [11]. Physiologically, VMIA may support DME through growth factors such as VEGF, fibroblast growth factor, and interleukin-6 that can alter vascular permeability. These growth factors have been localized by immunohistochemistry to the cells of vascular and avascular ERM in patients with diabetic retinopathy [20–22]. In addition, it has been suggested that these growth factors may be produced by cells within the cortical vitreous that accumulate/concentrate in the macular region due to VMIA and induce or exacerbate macular edema. [18] It is likely that both mechanical and physiological factors of VMIA play a role in reducing the effect of anti-VEGF. In this study, we categorized all images based on the presence of VMIA and ignored other OCT patterns of DME such as diffuse retinal thickening, CME, and serous retinal detachment. VMIAs were the sole focus of this study, because VMIAs are the only subtype that was consistently associated with reduced benefits of either focal laser or anti-VEGF, regardless of which outcome (BCVA, CMT, TMV) was studied. [3, 9] We also did not limit enrollment to one eye per patient because there was evidence that injections of bevacizumab only minimally raise the concentration in serum in the fellow eye [23]. There are several limitations to our study. It was a retrospective, short-term, non-randomized, and
123
Int Ophthalmol
uncontrolled study. The number of patients in each group was small, mainly due to the exclusion criteria—no previous injection, no interruption of anti-VEGF treatments by focal laser therapy or PRP, and no previous focal laser. It is controversial whether focal laser causes premacular proliferation and therefore, our study excluded all patients with previous focal laser. [4, 24] In addition, there were more patients who received previous PRP in the experimental group. It has been suggested it the past that thermal laser photocoagulation such as PRP is a well-known risk factor associated with a high prevalence of VMIA formation [25] We acknowledge that this may lead to a potential imbalance of study groups. In summary, we showed that the BCVA of patients with both DME and VMIAs improved significantly less than that of patients with only DME after multiple injections. Since many clinicians give at least three monthly injections at the beginning of treatment, we believe our results reflect the course of treatment encountered in many clinics. Our results add to the growing body of literature that the OCT pattern of diabetics can predict the efficacy of a particular treatment. Further studies must be performed to determine the pathogenesis of VMIAs and also to determine the treatment guidelines for diabetics with VMIAs.
References 1. Zimmet P, Alberti KG, Shaw J (2001) Global and societal implications of the diabetes epidemic. Nature 414:782–787 2. Chang CK, Cheng CK, Bai CH, Peng CH, Hu CC (2012) Development of vitreomacular interface abnormality in patients with diabetic macular edema. Taiwan J Ophthalmol 2:93–98 3. Kim NR, Kim YJ, Chin HS, Moon YS (2009) Optical coherence tomographic patterns in diabetic macular oedema: prediction of visual outcome after focal laser photocoagulation. Br J Ophthalmol 93:901–905 4. Ghazi NG, Ciralsky JB, Shah SM, Campochiaro PA, Haller JA (2007) Optical coherence tomography findings in persistent diabetic macular edema: the vitreomacular interface. Am J Ophthalmol 144:747–754 5. Group ETDRS (1985) Photocoagulation for diabetic macular edema: early treatment diabetic retinopathy study report number 1. Arch Ophthalmol 103:1796–1806 6. Kaiser PK, Riemann CD, Sears JE, Lewis H (2001) Macular traction detachment and diabetic macular edema associated with posterior hyaloidal traction. Am J Ophthalmol 131:44–49
123
7. Nguyen QD, Brown DM, Marcus DM, et al RISE and RIDE Research Group (2012) Ranibizumab for diabetic macular edema: results from 2 phase III randomized trials: RISE and RIDE. Ophthalmology119:789–801 8. Rajendram R, Fraser-Bell S, Kaines A et al (2012) A 2-year prospective randomized controlled trial of intravitreal bevacizumab or laser therapy (BOLT) in the management of diabetic macular edema: 24-month data: report 3. Arch Ophthalmol 130:972–979 9. Wu PC, Lai CH, Chen CL, Kuo CN (2012) Optical coherence tomographic patterns in diabetic macula edema can predict the effects of intravitreal bevacizumab injection as primary treatment. J Ocul Pharmacol Th 28:59–64 10. Laidlaw DA (2008) Vitrectomy for diabetic macular oedema. Eye 22:1337–1341 11. Massin P, Duguid G, Erginay A, Haouchine B, Gaudric A (2003) Optical coherence tomography for evaluating diabetic macular edema before and after vitrectomy. Am J Ophthalmol 135:169–177 12. Group ETDRS (1987) Treatment techniques and clinical guidelines for photocoagulation of diabetic macular edema: early treatment diabetic retinopathy study report number 2. Ophthalmology 94:761–774 13. Wilkins JR, Puliafito CA, Hee MR, Duker JS, Reichel E, Coker JG et al (1996) Characterization of epiretinal membranes using optical coherence tomography. Ophthalmology 103:2142–2151 14. Ophir A, Martinez MR, Mosqueda P, Trevino A (2010) Vitreous traction and epiretinal membranes in diabetic macular oedema using spectral-domain optical coherence tomography. Eye 24:1545–1553 15. Lewis H, Abrams GW, Blamenkranz MS, Campo RV (1992) Vitrectomy for diabetic macular traction and edema associated with posterior hyaloidal traction. Ophthalmology 99:753–759 16. Kaiser PK, Riemann CD, Sears JE, Lewis H (2001) Macular traction detachment and diabetic macular edema associated with posterior hyaloidal traction. Am J Ophthalmol 131:44–49 17. Thomas D, Bunce C, Moorman C, Laidlaw AH (2005) Frequency and associations of a taut thickened posterior hyaloid, partial vitreomacular separation, and subretinal fluid in patients with diabetic macular edema. Retina 25:883–888 18. Gandorfer A, Messmer EM, Ulbig MW, Kampik A (2000) Resolution of diabetic macular edema after surgical removal of the posterior hyaloid and the inner limiting membrane. Retina 20:126–133 19. Jumper JM, Embabi SN, Toth CA et al (2000) Electron immunocytochemical analysis of posterior hyaloid associated with diabetic macular edema. Retina 20:63–68 20. Yamamoto T, Akabane N, Takeuchi S (2001) Vitrectomy for diabetic macular edema: the role of posterior vitreous detachment and epimacular membrane. Am J Ophthalmol 132:369–377 21. Chen Y, Hackett S, Schoenfeld C, Vinores M, Vinores S, Campochiaro P (1997) Localisation of vascular endothelial growth factor and its receptors to cells of vascular and avascular epiretinal membranes. Br J Ophthalmol 81: 919–926
Int Ophthalmol 22. Schneeberger SA, Hjelmeland LM, Tucker RP, Morse LS (1997) Vascular endothelial growth factor and fibroblast growth factor 5 are colocalized in vascular and avascular epiretinal membranes. Am J Ophthalmol 124:447–454 23. Bakri SJ, Snyder MR, Reid JM, Pulido JS, Singh RJ (2007) Pharmacokinetics of intravitreal bevacizumab (Avastin). Ophthalmology 114:855–859
24. Appiah AP, Hirose T (1989) Secondary causes of premacular fibrosis. Ophthalmology 96:389–392 25. Mester U, Volker B, Kroll P, Berg P (1988) Complications of prophylactic argon laser treatment of retinal breaks and degenerations in 2,000 eyes. Ophthalmic Surg 19:482–484
123
ORIGINAL PAPER
Reduced effect of anti-vascular endothelial growth factor agents on diabetics with vitreomacular interface abnormalities Dan Yoon • Irene Rusu • Irene Barbazetto
Received: 30 September 2013 / Accepted: 13 November 2013 Ó Springer Science+Business Media Dordrecht 2013
Abstract This study investigated whether the presence of vitreomacular interface abnormalities (VMIAs) affects the improvement in visual acuity and edema of patients with diabetic macular edema (DME) who received three anti-vascular endothelial growth factor (VEGF) injections. Fifteen eyes of 11 patients with clinically significant macular edema were retrospectively divided into either the control group (only DME) or the experimental group (DME and VMIA) based on optical coherence tomography images. We defined VMIA patterns as epiretinal membrane and/or anomalous vitreomacular adhesion. Changes in central macular thickness (CMT), total macular volume (TMV), and best-corrected visual acuity (BCVA) from the baseline to post third injection were compared between the two groups. After the third injection, the decreases in CMT and TMV were not statistically different between the two groups. The improvement in BCVA was larger in the control group (0.1742 ± 0.0508 logMAR) than in the experimental group (0.0766 ± 0.0562 logMAR; p \ 0.01). Our study showed that after the third anti-VEGF D. Yoon (&) I. Rusu Department of Ophthalmology, New York University School of Medicine, 462 First Ave. NBV 5N18, New York, NY 10016, USA e-mail: [email protected]; [email protected] I. Barbazetto Vitreous-Retina-Macula Consultants of New York, New York, NY, USA
injection, the BCVA of patients with both DME and VMIAs improved significantly less than that of patients with only DME. Our results suggest that VMIAs may play a crucial role in reducing the therapeutic effects of anti-VEGF agents. Keywords Diabetic macular edema Vitreomacular interface abnormality Anti-VEGF Epiretinal membrane
Introduction Diabetic macular edema (DME) is one of the most common causes of visual impairment in patients with diabetes [1]. Vitreomacular interface abnormalities (VMIAs) appear to occur in 7–16 % of eyes with DME, and the annual incidence may be as high as 4.5 % [2, 3]. VMIAs include epiretinal membrane and/or anamolous vitreomacular adhesion, which is sometimes referred to as vitreomacular traction or posterior vitreomacular separation and is characterized as having features of a taut, thick posterior hyaloid [4]. Both features are often better identified using optical coherence tomography (OCT) in addition to conventional fundus examination [4]. The presence or absence of these interface abnormalities potentially affect the response to treatment in DME patients. For many years, the gold standard treatment for DME was focal laser photocoagulation, which can
123
Int Ophthalmol
reduce the risk of moderate visual loss, defined as a loss of C15 letters, by 50 % compared to those with deferred treatment [5]. Nevertheless, after 3 years of treatment, approximately 12 % of treated patients still experience moderate visual loss. This unresponsiveness is thought to be due to persistent macular edema [5]. Recently, VMIAs were found to be associated with as high as 52–67 % of cases of persistent macular edema that was unresponsive to focal laser photocoagulation [3, 4, 6]. Whether VMIAs play a role in determining the efficacy of newer treatments, specifically anti-vascular endothelial growth factor (VEGF) treatments, has not been studied extensively. Ranibizumab (Lucentis, Genentech Inc, South San Francisco, CA, USA) was approved by the Food and Drug Administration (FDA) for the treatment of DME based on two phase III trials, RISE and RIDE [7]. Furthermore, recent results of the Bevacizumab or Laser Therapy (BOLT) study showed that bevacizumab (Avastin; Genentech Inc, South San Francisco, CA, USA) was also superior to laser treatment [8]. Only one study to date by Wu et al. [9] suggests that DME with features of VMIAs may be associated with poor results after anti-VEGF treatments. One hypothesis for this trend is that traction caused by VMIAs may reduce the potential benefits of treatments. Support for this theory comes from growing evidence that vitrectomy and the resulting separation of posterior hyaloid from traction may help in treating persistent DME with VMIAs [10, 11]. This retrospective study further investigated whether the presence of VMIAs affects the improvement in visual acuity and edema from anti-VEGF injections. In particular, while Wu et al. mostly followed patients who only received one injection for the duration of 3 months, we studied whether changes in best-corrected Snellen visual acuity (BCVA), central macular thickness (CMT), and total macular volume (TMV) were still reduced in patients with VMIAs after three injections. Although the frequency of anti-VEGF regimen for DME remains controversial, most clinical trials were based on monthly injections for the first three injections. Since many clinicians follow this practice, this study would give us additional information of whether VMIAs affect the efficacy of anti-VEGF in treating DME under the current treatment timeline.
123
Methods The study was approved by the Institutional Review Board of North Shore - LIJ Health System. Patient selection We reviewed the charts of patients from a private practice retinal referral center in an urban setting with the billing diagnosis of DME who received anti-VEGF intravitreal injections between 2010 and 2012. Each anti-VEGF injection was either 0.3 mg ranibizumab or 1.25 mg bevacizumab. The inclusion criteria were (1) clinically significant macular edema as defined by the early treatment of diabetic retinopathy study (ETDRS) [12], (2) CMT of [250 lm by OCT; (3) age C18 years, and (4) received three anti-VEGF injections consecutively with a maximum duration of 2 months between each injection. The exclusion criteria were (1) ocular disease apart from diabetic retinopathy such as glaucoma, cataract, or active vitreous hemorrhage, (2) previous intravitreal injection of anti-VEGF agents or steroids prior to the first documented anti-VEGF injection, (3) previous macular focal laser therapy (those with panretinal photocoagulation (PRP) were still included), (4) simultaneous injection of steroids with anti-VEGF, (5) interruption of antiVEGF treatments with focal laser therapy or PRP, and (6) BCVA at the time of first treatment that was better than 20/30 (due to a ceiling effect) or worse than 20/200 (due to the difficulty in accurately quantifying visual acuity). All patients underwent a complete ophthalmic examination that included BCVA recording using Snellen charts, slit-lamp examination, intraocular pressure measurement, and fundus examination with indirect ophthalmoscopy. Central retinal characteristics were evaluated using spectral domain OCT (SD-OCT) (Spectralis; Heidelberg Engineering, Heidelberg, Germany). All OCT images were obtained by volume scan option. Due to the retrospective nature of the study, the scanned area ranged from 5.9 9 4.4 to 9.1 9 7.5 mm. The number of B scans also ranged from 19–49 (spaced 120–240 lm apart) with automatic real time (ART) mean ranging from 13 to 23. OCT grading Based on the grading of the OCT images by two independent reviewers (DY and IR) patients were
Int Ophthalmol
divided into either the control group (only DME) or the experimental group (DME and VMIA). We defined VMIA patterns as epiretinal membrane (ERM) and/or anomalous vitreomacular adhesions. Both reviewers agreed on the categorization of each scan. In 24 of 30 OCT scans, each follow-up image was automatically registered at the time of image acquisition to the first visit image and, therefore, the ETDRS grids for CMT and TMV measurements were automatically placed in identical locations for both the first visit image and follow-up image. In the remaining six OCT scans, placement of the ETDRS grids was adjusted by reviewers based on landmarks seen on corresponding IR images so that the ETDRS grids would be placed in approximately identical locations. Both reviewers agreed on all manual placements of ETDRS grids. Epiretinal membrane (ERM) On OCT, ERM was defined as thin, highly reflective bands anterior to the neurosensory retina as defined by Wilkins et al. [13]. The OCT configurations of ERM were defined as either globally adherent or partially separated with focal areas of macular attachments [13]. There is no minimum size requirement for ERM in the literature. In order to make the categorization of images objective, we categorized a hyperreflectant band [1.5 mm in horizontal size in any B-scan to be an ERM. We used the threshold of 1.5 mm, because the dimension of the smallest ERM found with threedimensional SD-OCT among patients with DME was 1.5 9 3 mm2 [14].
Anomalous vitreomacular adhesion Anomalous vitreomacular adhesion was defined as an evident hyperreflective band that adhered to the surface of the retina at specific sites, was elevated elsewhere off the surface, and continuous with the posterior vitreous surface, in agreement with the report by Ghazi et al. [4]. Statistical tests BCVA, CMT of 1 mm, and TMV of 3 mm at the time of first injection (baseline) and post third injection were noted. BCVAs were converted to the logarithm of minimal angle of resolution (logMAR) format for statistical manipulation. The final BCVA, CMT, and TMV after the third injection were compared to the baseline BCVA, CMT, and TMV at the time of the first injection using paired t tests. Changes in the BCVA, CMT, and TMV from baseline to post third injection were compared between the control group (only DME) and the experimental group (DME with VMIA). Normality was checked, and unpaired t test was used (SPSS; IBM Corporation, Armonk, NY, USA) for comparing changes in BCVA, CMT, and TMV between the control group and the experimental group.
Results Baseline characteristics Patient characteristics are summarized in Table 1. A total of 15 eyes from 11 patients (8 male and 3
Table 1 Baseline characteristics of the 15 eyes from 11 patients with diabetic macular edema Control
Experimental
p value
No. of patients
4
7
Age (±SD)
56 (±8)
67 (±5)
0.078
Sex (male:female) Duration of DM (years ± SD)
3:1 14.50 (±7.85)
5:2 14.00 (±7.21)
0.898 0.920
HgA1C (± SD)
7.43 (±0.81)
6.75 (±0.58)
0.293
Follow-up (days ± SD)
106 (±13)
98 (±19)
0.400
No. of days between each injection
35.2 (±4.2)
32.8 (±6.3)
0.400
No. of eyes with three injections
5
10
Stage (NPDR:PDR)
4:1
6:4
0.439
No. of eyes with PRP
0
3
0.171
DM diabetes mellitus, NPDR nonproliferative diabetic retinopathy, PDR proliferative diabetic retinopathy
123
Int Ophthalmol Table 2 Change in best-corrected visual acuity (BCVA), central macular thickness (CMT) of 1 mm, and total macular volume (TMV) of 3 mm from baseline to post third injection Control (mean ± SD)
Experimental (mean ± SD)
Baseline BCVA
0.4165 ± 0.1386
0.4497 ± 0.1245
Post third injection BCVA
0.2423 ± 0.0924
0.3732 ± 0.1350
P value (Post third injection - baseline)
\0.01
\0.01
p value (control vs. experimental) 0.664
\0.01
Change in BCVA
0.1742 ± 0.0508
0.0766 ± 0.0562
Baseline CMT
509.8 ± 181.8
471.6 ± 107.8
Post third injection CMT
388.8 ± 151.2
389.1 ± 113.6
p value (post third injection-baseline)
0.098
\0.05
Change in CMT
121.0 ± 126.2
82.4 ± 106.8
0.581
Baseline TMV
3.34 ± 0.6
3.07 ± 0.45
0.396
Post third injection TMV
2.74 ± 0.50
2.71 ± 0.41
p value (post third injection-baseline) Change in TMV
\0.05 0.60 ± 0.32
\0.01 0.35 ± 0.34
0.683
0.229
female) with nonproliferative (n = 10) and proliferative diabetic retinopathy (n = 5) were enrolled in the study. The average duration (±SD) of diabetes was 14.18 ± 7.05 years, and the average HgA1c level (±SD) was 7.04 ± 0.72. The average age (±SD) of the patients was 63 ± 8 years. The mean follow-up was 100.8 ± 16.91 days. Three eyes that received prior PRP were included in this study. In the experimental group, four eyes had anomalous vitreomacular adhesion only, five eyes had ERM only, and one eye had both. Of those with ERM, one was characterized as partially separated with focal attachment and five were characterized as globally adherent. In the 11 patients who were included in the above analysis, no feature of the anomalous vitreomacular adhesion or epiretinal membrane changed during the three injections.
Fig. 1 Changes in best-corrected visual acuity (BCVA). The white bar shows the improvement of the control group (0.1742 ± 0.0508 (SD; logMAR) and the black bar shows the experimental group (0.0766 ± 0.0562 (SD; logMAR; p \ 0.01). The error bars denote the standard error of the mean (SEM)
Change in BCVA
Change in CMT
The baseline BCVA (±SD) was not statistically different between the control group (0.4165 ± 0.1386 logMAR) and the experimental group (0.4497 ± 0.1245 logMAR; p = 0.664; Table 2). After three injections, the final BCVA was statistically different from the baseline BCVA for both the control (0.2423 ± 0.0924 logMAR; p \ 0.01) and the experimental group (0.3732 ± 0.1350 logMAR; p \ 0.01). The improvement in BCVA was larger in the control group (0.1742 ± 0.0508 logMAR) than in the experimental group (0.0766 ± 0.0562 logMAR; p \ 0.01; Fig. 1).
The baseline CMT (±SD) was not statistically different between the control group (509.8 ± 181.8 lm) and the experimental group (471.6 ± 107.8 lm; p = 0.683; Table 2). After three injections, the final CMT was statistically different from the baseline CMT for the experimental group (389.1 ± 113.6 lm; p \ 0.05) but not for the control group (388.8 ± 151.2 lm; p = 0.098). The decrease in CMT was not statistically different between the control group (121.0 ± 126.2 lm) and the experimental group (82.4 ± 106.8 lm; p = 0.581).
123
Int Ophthalmol
Change in TMV The baseline TMV (±SD) was not statistically different between the control group (3.34 ± 0.6 mm3) and the experimental group (3.07 ± 0.45 mm3; p = 0.396; Table 2). After three injections, the final TMV was statistically different from the baseline TMV for both the control group (2.74 ± 0.50 mm3; p \ 0.05) and the experimental group (2.71 ± 0.41 mm3; p \ 0.01). The decrease in TMV was not statistically different between the control group (0.60 ± 0.32 mm3) and the experimental group (0.35 ± 0.34 mm3; p = 0.229).
Discussion Our study showed that after the third anti-VEGF injection, the BCVA of patients with both DME and VMIAs improved significantly less than that of patients with only DME. There was no change in the features of the anomalous vitreomacular adhesions or ERM during the treatment process. Our results add to the previous evidence that suggested that patients with VMIAs did not respond as well to anti-VEGF injections as those with only cystoid macular edema (CME) after one treatment. [9] We showed the reduced effect of antiVEGF after multiple injections in order to mimic the clinical setting used in many practices. Our results suggest that VMIAs play a crucial role in reducing the therapeutic effects of anti-VEGF agents. In contrast to the significant difference in BCVA improvements between the control group and the experimental group, the decreases in CMT and TMV after three injections were not significantly different between the two groups. This result was surprising since a similar trend was expected among all three measures based on previous results. For example, Wu et al. [9] previously showed that BCVA, CMT, and TMV did not improve for DME patients with VMIAs after receiving one injection while all three measures improved for those with only CME. In our study, all three measures of DME patients with VMIAs improved after three injections, but only BCVA was improved by a significantly less degree when compared to the control group. These results suggest that while three injections may reduce the edema thickness for DME patients with VMIAs, there are multiple factors in
addition to the reduction in edema thickness that determine the final improvement in vision. Among many possible contributions regarding VMIAs, there are two main hypotheses to explain why either ERM or vitreomacular adhesions interfere with the therapeutic effects of anti-VEGF agents in DME—mechanical and physiological. Mechanically, these abnormalities cause traction on the macula [15]. The tractional forces of a taut, thickened posterior hyaloid pull the retina forward and may even produce a deep subretinal space where fluid can pool [16, 17]. Based on the presence of multilayered cellular components in excised ERM from patients with DME, it is hypothesized that the accumulation of chemoattractant cytokines leads to the migration of cells to the posterior hyaloid and that these cells play a crucial role in creating tractional forces [18, 19]. The tractional force hypothesis is also supported by findings that the benefits of vitrectomy may be confined to patients with OCT signs of macular traction [11]. Physiologically, VMIA may support DME through growth factors such as VEGF, fibroblast growth factor, and interleukin-6 that can alter vascular permeability. These growth factors have been localized by immunohistochemistry to the cells of vascular and avascular ERM in patients with diabetic retinopathy [20–22]. In addition, it has been suggested that these growth factors may be produced by cells within the cortical vitreous that accumulate/concentrate in the macular region due to VMIA and induce or exacerbate macular edema. [18] It is likely that both mechanical and physiological factors of VMIA play a role in reducing the effect of anti-VEGF. In this study, we categorized all images based on the presence of VMIA and ignored other OCT patterns of DME such as diffuse retinal thickening, CME, and serous retinal detachment. VMIAs were the sole focus of this study, because VMIAs are the only subtype that was consistently associated with reduced benefits of either focal laser or anti-VEGF, regardless of which outcome (BCVA, CMT, TMV) was studied. [3, 9] We also did not limit enrollment to one eye per patient because there was evidence that injections of bevacizumab only minimally raise the concentration in serum in the fellow eye [23]. There are several limitations to our study. It was a retrospective, short-term, non-randomized, and
123
Int Ophthalmol
uncontrolled study. The number of patients in each group was small, mainly due to the exclusion criteria—no previous injection, no interruption of anti-VEGF treatments by focal laser therapy or PRP, and no previous focal laser. It is controversial whether focal laser causes premacular proliferation and therefore, our study excluded all patients with previous focal laser. [4, 24] In addition, there were more patients who received previous PRP in the experimental group. It has been suggested it the past that thermal laser photocoagulation such as PRP is a well-known risk factor associated with a high prevalence of VMIA formation [25] We acknowledge that this may lead to a potential imbalance of study groups. In summary, we showed that the BCVA of patients with both DME and VMIAs improved significantly less than that of patients with only DME after multiple injections. Since many clinicians give at least three monthly injections at the beginning of treatment, we believe our results reflect the course of treatment encountered in many clinics. Our results add to the growing body of literature that the OCT pattern of diabetics can predict the efficacy of a particular treatment. Further studies must be performed to determine the pathogenesis of VMIAs and also to determine the treatment guidelines for diabetics with VMIAs.
References 1. Zimmet P, Alberti KG, Shaw J (2001) Global and societal implications of the diabetes epidemic. Nature 414:782–787 2. Chang CK, Cheng CK, Bai CH, Peng CH, Hu CC (2012) Development of vitreomacular interface abnormality in patients with diabetic macular edema. Taiwan J Ophthalmol 2:93–98 3. Kim NR, Kim YJ, Chin HS, Moon YS (2009) Optical coherence tomographic patterns in diabetic macular oedema: prediction of visual outcome after focal laser photocoagulation. Br J Ophthalmol 93:901–905 4. Ghazi NG, Ciralsky JB, Shah SM, Campochiaro PA, Haller JA (2007) Optical coherence tomography findings in persistent diabetic macular edema: the vitreomacular interface. Am J Ophthalmol 144:747–754 5. Group ETDRS (1985) Photocoagulation for diabetic macular edema: early treatment diabetic retinopathy study report number 1. Arch Ophthalmol 103:1796–1806 6. Kaiser PK, Riemann CD, Sears JE, Lewis H (2001) Macular traction detachment and diabetic macular edema associated with posterior hyaloidal traction. Am J Ophthalmol 131:44–49
123
7. Nguyen QD, Brown DM, Marcus DM, et al RISE and RIDE Research Group (2012) Ranibizumab for diabetic macular edema: results from 2 phase III randomized trials: RISE and RIDE. Ophthalmology119:789–801 8. Rajendram R, Fraser-Bell S, Kaines A et al (2012) A 2-year prospective randomized controlled trial of intravitreal bevacizumab or laser therapy (BOLT) in the management of diabetic macular edema: 24-month data: report 3. Arch Ophthalmol 130:972–979 9. Wu PC, Lai CH, Chen CL, Kuo CN (2012) Optical coherence tomographic patterns in diabetic macula edema can predict the effects of intravitreal bevacizumab injection as primary treatment. J Ocul Pharmacol Th 28:59–64 10. Laidlaw DA (2008) Vitrectomy for diabetic macular oedema. Eye 22:1337–1341 11. Massin P, Duguid G, Erginay A, Haouchine B, Gaudric A (2003) Optical coherence tomography for evaluating diabetic macular edema before and after vitrectomy. Am J Ophthalmol 135:169–177 12. Group ETDRS (1987) Treatment techniques and clinical guidelines for photocoagulation of diabetic macular edema: early treatment diabetic retinopathy study report number 2. Ophthalmology 94:761–774 13. Wilkins JR, Puliafito CA, Hee MR, Duker JS, Reichel E, Coker JG et al (1996) Characterization of epiretinal membranes using optical coherence tomography. Ophthalmology 103:2142–2151 14. Ophir A, Martinez MR, Mosqueda P, Trevino A (2010) Vitreous traction and epiretinal membranes in diabetic macular oedema using spectral-domain optical coherence tomography. Eye 24:1545–1553 15. Lewis H, Abrams GW, Blamenkranz MS, Campo RV (1992) Vitrectomy for diabetic macular traction and edema associated with posterior hyaloidal traction. Ophthalmology 99:753–759 16. Kaiser PK, Riemann CD, Sears JE, Lewis H (2001) Macular traction detachment and diabetic macular edema associated with posterior hyaloidal traction. Am J Ophthalmol 131:44–49 17. Thomas D, Bunce C, Moorman C, Laidlaw AH (2005) Frequency and associations of a taut thickened posterior hyaloid, partial vitreomacular separation, and subretinal fluid in patients with diabetic macular edema. Retina 25:883–888 18. Gandorfer A, Messmer EM, Ulbig MW, Kampik A (2000) Resolution of diabetic macular edema after surgical removal of the posterior hyaloid and the inner limiting membrane. Retina 20:126–133 19. Jumper JM, Embabi SN, Toth CA et al (2000) Electron immunocytochemical analysis of posterior hyaloid associated with diabetic macular edema. Retina 20:63–68 20. Yamamoto T, Akabane N, Takeuchi S (2001) Vitrectomy for diabetic macular edema: the role of posterior vitreous detachment and epimacular membrane. Am J Ophthalmol 132:369–377 21. Chen Y, Hackett S, Schoenfeld C, Vinores M, Vinores S, Campochiaro P (1997) Localisation of vascular endothelial growth factor and its receptors to cells of vascular and avascular epiretinal membranes. Br J Ophthalmol 81: 919–926
Int Ophthalmol 22. Schneeberger SA, Hjelmeland LM, Tucker RP, Morse LS (1997) Vascular endothelial growth factor and fibroblast growth factor 5 are colocalized in vascular and avascular epiretinal membranes. Am J Ophthalmol 124:447–454 23. Bakri SJ, Snyder MR, Reid JM, Pulido JS, Singh RJ (2007) Pharmacokinetics of intravitreal bevacizumab (Avastin). Ophthalmology 114:855–859
24. Appiah AP, Hirose T (1989) Secondary causes of premacular fibrosis. Ophthalmology 96:389–392 25. Mester U, Volker B, Kroll P, Berg P (1988) Complications of prophylactic argon laser treatment of retinal breaks and degenerations in 2,000 eyes. Ophthalmic Surg 19:482–484
123
