Graefe's Archive for Clinical and Experimental Ophthalmology

, Volume 246, Issue 5, pp 653–659

Time-dependent effects on contrast sensitivity, near and distance acuity: difference in functional parameters? (Prospective, randomized pilot trial of photodynamic therapy versus full macular translocation)


    • Department for Ophthalmology, University Eye HospitalEberhard-Karls University
  • Matthias Lüke
    • Department for Ophthalmology, University Eye HospitalEberhard-Karls University
  • Karl U. Bartz-Schmidt
    • Department for Ophthalmology, University Eye HospitalEberhard-Karls University
  • Faik Gelisken
    • Department for Ophthalmology, University Eye HospitalEberhard-Karls University
Retinal Disorders

DOI: 10.1007/s00417-007-0726-y

Cite this article as:
Ziemssen, F., Lüke, M., Bartz-Schmidt, K.U. et al. Graefes Arch Clin Exp Ophthalmol (2008) 246: 653. doi:10.1007/s00417-007-0726-y



To report the change of contrast sensitivity (CS) after photodynamic therapy (PDT) vs full macular translocation (FMT) for neovascular age-related macular degeneration (AMD), and to relate this to other measures of visual function (distance and near acuity).


Fifty patients (50 eyes) with predominantly classic subfoveal choroidal neovascularisation (CNV) secondary to AMD were randomized to PDT or FMT. CS was measured with Pelli-Robson charts. Acuity scores of near visual function (NVS) were calculated after testing with visual acuity cards of the Swiss National Association of and for the Blind (SNAB). Best corrected distance visual acuity (DVA) was determined according to a standardized protocol with EDTRS charts. Primary end point was the change of CS at 12-month examination from baseline. The interaction of the CS with NVS and DVA was analysed.


Mean CS showed a decrease in both treatment groups (FMT: −2 letters, PDT: −3 letters, p = 0.969) at 12-month examination from baseline. While mean NVS improved by seven letters in the FMT group, a decrease of more than ten letters was seen in the PDT group (p < 0.05). We found no agreement between CS and high-contrast acuity (NVS, DVA). In FMT patients, the parameters at baseline (CS, NVS, DVA) correlated poorly with the corresponding 12-month results, therefore providing no informative basis to predict the later functional development. In contrast, PDT patients showed strong baseline-to-outcome coherence with baseline measures also associated with better final values.


Although FMT can initiate recovery of near and distance acuity over the period of 1 year in selected patients with classic CNV, CS did not differ between FMT and PDT. We found no close connection of CS with DVA or NVS, especially after FMT. Knowledge about the unequal variation of visual parameters can provide more comprehensive information when advising patients on different therapeutic options. That also applies in particular to vascular endothelial growth factor inhibitors, which seem to promise an even higher extent of gain in CS and to reach the peak of recovery at an earlier time.


Age-related macular degenerationContrast sensitivityFull macular translocationNear visionPhotodynamic therapyVisual acuitySurgery


Rehabilitation of visual function, or at least limiting the degree of vision loss, is the primary target in the treatment of neovascular age-related macular degeneration (AMD), a frequent cause of blindness in the elderly part of the Western population [1]. The growth of subfoveal choroidal neovascularisation (CNV) frequently leads to a dramatic loss in central vision. Distance visual acuity based on the assessment of high-contrast optotypes on a retro-illuminated background is most often used in clinical prospective studies as a main measurement outcome [2]. However, contrast sensitivity (CS) and near visual acuity are important independent functional parameters to determine the visual disability in patients suffering from CNV.

CS has been shown to be of great significance for predicting reading speed, mobility performance and recognition of targets such as faces and traffic signs [3]. A two-fold reduction in CS is associated with a three- to five-fold higher likelihood of reporting difficulties with daily tasks [4]. Near function is of utmost importance for reading [5]. Usage of near visual acuity cards does not only help to assess the exact magnification demand, but provides a near visual acuity score highly correlated to vision quality.

Distance visual acuity measured with a high-contrast letter chart may overestimate the visual capability of some patients, e.g. those with media opacities or various neural problems. Although there are some hints in the literature that the parameters of visual function can substantially differ—especially in eyes with neovascular AMD [6]—the development of individual values following two complementary treatment modalities has not been analyzed in longitudinal detail yet. Therefore, we examined the change of the CS in a group of patients with neovascular AMD treated in a prospective and randomized fashion either by photodynamic therapy (PDT) with verteporfin or by full macular rotation (FMT). The analysis of functional parameters at different time points aimed to clarify time-dependent effects in the course of both treatments.

Materials and methods

The FMT-PDT study was a prospective, randomized, controlled non-masked, mono-centre pilot trial for comparison of FMT and PDT in patients with classic subfoveal CNV. The study was completed in compliance with the Declaration of Helsinki and the International Committee on Harmonization guidelines. Institutional review board approval and written informed consent were obtained from patients who consented to participate in the study from 2001 through 2005. One-year results and related information such as demographic characteristics and inclusion/exclusion criteria of the FMT-PDT pilot trial have been published [7].

Best corrected distance visual acuity (DVA) was measured using a standard refraction protocol, standard retro-illuminated Early Treatment Diabetic Retinopathy Study (ETDRS) charts (one line = five letters) by an unmasked examiner. CS was measured using the Pelli-Robson chart (Clement Clarke International Ltd., Harlow, UK) which consists of 16 letter triplets. Each of the triplets is equivalent to one logMar unit. For determining CS, the lens power of the eye being tested was increased by 0.5 dioptre. The patient viewed the chart at 1-m distance with the fellow-eye occluded. The SNAB (Swiss National Association of and for the Blind) test cards were originally developed to determine the magnification requirements. Under optimal illumination of cards, the test was conducted in a distance of 25 cm with the best, predetermined near correction. Fluent reading of text passages was necessary to meet the logarithmic steps of magnification need. The near visual acuity scores (NVS) were then calculated using the calibrated nomogram in dependence on the modified Snellen formula and critical print size [8]. The modified Snellen formula is based on the reciprocal viewing distance, indicating the amount of magnification needed to bring the subject to standard performance.

Statistical analysis

Overall pre-treatment and post-treatment letter scores were calculated. For patients who discontinued the study, the last clinic visit was carried forward in the analysis (intention to treat) to avoid selection bias [9]. Completing the analysis of the collected data, the results of the letter scores before treatment (baseline) were compared with the scores 1 year after undergoing PDT or FMT. Data analysis was done with the statistical software package JMP5.0.1.2 (SAS Institute, Cary, NC, USA). For analysis of score changes, Wilcoxon signed rank test was performed. The significance level was set to α = 0.05. Pearson product-moment correlation included correlation coefficients and significant probability (obtaining, by chance alone, a correlation with greater absolute value than the computed value if no linear relationship exists).


At baseline, the two treatment groups (FMT vs PDT) did not differ in any of the functional parameters (Table 1, DVA: p = 0.490, CS: p = 0.626, NVS: p = 0.447, Wilcoxon signed rank test, Table 1).
Table 1

Baseline parameter and change within 12 months after treatment



3-month FMT

6-month FMT

9-month FMT

12-month FMT

1-year FMT


3-month PDT

6-month PDT

9-month PDT

12-month PDT

1-year PDT



95% CI

Mean change

95% CI

Mean change

95% CI

Mean change

95% CI

Mean change

95% CI


95% CI


95% CI

Mean change

95% CI

Mean change

95% CI

Mean change

95% CI

Mean change

95% CI


95% CI



























CS Pelli-Robson




















































Letter scores at baseline and after 12 months, including the mean change in score and the corresponding confidence interval (95% CI), are given. The p-value represents the matched pair analysis (FMT - PDT, Wilcoxon signed rank test).

While mean DVA gained 2.7 ETDRS letters (p = 0.43) in the 1st year after FMT, PDT patients showed a significant decrease in DVA (−11.9 letters, p < 0.05) at the 12-month examination. CS exhibited only a very slight diminishment over the postoperative year in both treatment groups (FMT: p = 0.59, PDT: p = 0.34, Fig. 1). The score of near function dropped in the PDT group in dependence on time (−10.4, p = 0.18), whereas a mean improvement of 7.4 was seen in the FMT group (p = 0.13) at 12-month examination from baseline. The outcome of DVA and NVS following the FMT was significantly better in comparison to PDT. The two treatment modalities did not vary in the recognition of low-contrast patterns (p = 0.969).
Fig. 1

Time course of functional parameters following FMT (black line) and PDT (grey, dotted line) with lines connecting means (diamonds). For DVA (a), CS (b) and NVS (c) the box plots indicate median and the 25th and 75th percentiles respectively. The whiskers extend to the outermost data point that falls within the upper/lower quartile +/− the 1.5-fold of the interquartile range

At baseline, the correlation between DVA and CS (Pearson coeff.: 0.459, p = 0.008), between DVA and NVS (Pearson coeff.: 0.577, p = 0.0001) and between NVS and CS (Pearson coeff.: 0.579, p = 0.0001) was very similar in both treatment groups. When comparing the correlation between the functional parameters at 12-month examination, correlation coefficients were even higher (DVA-CS: 0.697, DVA-NVS: 0.856, NVS-CS: 0.711) than the preoperative values, also without a significant difference between the two groups. Regarding the change in scores, the PDT group nearly showed the same agreement (DVA-CS Pearson coeff.: 0.347, p = 0.0893, DVA-NVS Pearson coeff.: 0.466, p = 0.0190, NVS-CS Pearson coeff.: 0.530, p = 0.0064) as seen in the MT group (DVA-CS Pearson coeff.: 0.593, p = 0.0018, DVA-NVS Pearson coeff.: 0.783, p = 0.0001, NVS-CS Pearson coeff.: 0.676, p = 0.0002).

Predictors of functional development

When evaluating which baseline parameter was most predictive for the 12-month examination, we found a clear difference between the FMT and PDT group. Whereas better baseline correlated better with final DVA in PDT patients (DVA Pearson coeff.: 0.546, p = 0.005, CS Pearson coeff.: 0.610, p = 0.0012, NVS Pearson coeff: 0.485, p = 0.0141), a negative correlation was seen between initial NVS and later outcome of PDT (Fig. 2). Lower probability values indicated an overall more inconsistent relationship between baseline and the follow-up parameter in the FMT group. For example, there was only weak correlation between DVA at 12-month examination and the baseline parameters (DVA Pearson coeff.: 0.079, p = 0.7099, CS Pearson coeff.: 0.134, p = 0.5237, NVS Pearson coeff.: −0.207, p = 0.322). Among the FMT patients the highest correspondence was seen between baseline DVA and NVS after 12 months (Pearson coeff.: 0.229 probability: 0.2700). The best agreement in the PDT group was assessed between baseline DVA and 12-month NVS (Pearson coeff.: 0.654, p = 0.0004).
Fig. 2

Comparison of baseline values and 12-month outcome specifies the potentially predictive values of DVA (a), CS (b), NVS (c). Linear regression models (FMT: black, PDT: grey) are indicated by the lines superposing the scatter plots

We found no correlation of age and visual function tests (CS, NVS, DVA), neither in terms of the absolute scores nor the postoperative change, with no difference between the treatment groups.


Our study delineated clear differences of visual function parameters between the two treatment modalities. The FMT achieved an increase in DVA and NVS, whereas only stabilization of these parameters was seen following PDT. Despite the higher number of gainers in visual acuity, which has to be carefully interpreted in the light of operative burden and potential complications, CS did not recover in both groups, showing very similar development in time.

The results of the visual outcome of FMT should be interpreted carefully. We were surprised at the less pronounced perturbation of CS by the extensive surgical procedure. All patients received temporary retinal detachment and even postoperative silicon oil filling for the maximum period of 3 months [10]. Nonetheless, the small change for the worse shows a similar extent at the 3-month follow-up as the course of non-invasive PDT. Although known postoperative factors contribute to photoreceptor damage, media opacities and disturbance of ocular surface, early CS did not differ on average between the treatment modalities, just as change in DVA is the same for FMT and PDT. The final difference between CS and DVA only originates from the later improvement of DVA in FMT patients. NVS even gets better at the first control examination in comparison to the baseline assessment. The discrepancy between the results of low- and high-contrast tests is hard to explain. In the literature, a six-letter loss of CS is translated into a doubling of the amount of contrast required, and is comparable to a 15-letter loss on the ETDRS chart [4, 11]. Similar considerations and calculations are obsolete, when a linear relationship is missing between CS and DVA. Despite the significant group correlation between different spatial vision measures, it is impossible to predict the outcome on an individual basis from baseline parameters. The time course of CS makes immediate and selective damage to the low-contrast resolution by FMT very unlikely. The data supported the hypothesis that the discrepancy might be caused by a differing incidence of recovery. Stopping further leakage and growth of the CNV is the purpose of both treatments. However, the FMT approach mostly extirpates the whole membrane and adjacent scar tissue, and thus is more successful in remodelling an anatomic structure which promises greater similarity to the healthy macula. While high-luminance acuity might be improved primarily by recovered photoreceptors, one can speculate that the pre-existing impairment might not allow an improvement in low-contrast discrimination. But the recent data disproved this hypothesis. In the ANCHOR study, the average CS gained 4.1 letters at 12-month examination. Over the PDT, a mean improvement of 7.1 letters in CS was achieved by 24 ranibizumab injections [12]. Although the exact numbers (percentage of losers vs gainers) have not been published yet, there is strong evidence that medicinal approaches can enhance the aligning discrimination independent of the contrast, while the damage to the retina, caused by the inherent factors of the surgical treatments, does not allow similar rehabilitation.

In population-based assessments, visual function measures usually decrease with age [13]. During the short follow-up of 12 months, we did not find a similar association of age and baseline function or therapy response when evaluating the selected group with active classic CNV in this study. However, we can not exclude the possibility that the psychophysical parameters assessed in this study are not influenced by age-dependent factors.

Independent of the CS, near and distance function at high contrast are known to be relevant factors of visual quality of life [14, 15]. Reading speed is directly dependent on NVS. The decrease in reading ability has been reported to most exactly represent the extent of relevant impairment [16, 17]. NVS is also known to be closely related to face recognition [18].

However, there are also other important vision measures (visual fields, dynamic visual acuity, image displacement threshold) making an impact on daily living tasks. For instance, mobility performance measured with preferred walking speed seems to be quite independent of DVA or CS [19].

Impaired CS is known to be associated with other difficulties experienced by AMD patients, particularly with regard to tasks requiring judgment of space and distance. As CS reflects the ability to detect the presence of objects in the travel path as a function of object size and its contrast with the surroundings, a deficiency in CS can primarily manifest as an increasing tendency to bump into obstacles. Poor CS is associated with reduced body sway and increased risk of falls in elderly patients [20, 21]. Thus, it is not astonishing that worse visual acuity is linked to a higher mortality rate [22]. Although even the visual loss seen in the PDT group in this pilot trial was not on a scale for which previous studies have described a direct short-term interaction with dying, influence on numerous intervening variables and long-term relevance have to be assumed. We found no baseline parameter to be a reliable predictor of response to the treatment modalities. In the FMT group, even patients with lower NVS before surgery showed a greater benefit than others. In consequence, making the decision of FMT still must follow other factors, such as duration of the symptoms, the progression of lesion size and the anatomic situation.

For FMT, previous studies already reported that the increase in NVS may exceed the effect on DVA [2325], but we could not reproduce a gain in CS described earlier [25]. The absence of randomized control groups (PDT) and the less restrictive inclusion criteria (not only classic CNV, no limit to the amount of subretinal haemorrhage or fibrosis) could have favoured a superior outcome in the past.

If putting this study into perspective of the TAP study [26], more patients of our PDT group lost ≥6 letters in CS (32.0 % vs 21.4%) than reported in the TAP trial. A loss ≥15 letters has been described in 4.5% of the TAP study population, in this mono-centre trial even 8% showed a similar decrease in CS. When just including the predominantly classic CNVs of TAP, the mean change in CS is −0.3 at 12-month examination vs −3.0 in our assessment. There was a similar ratio of CS losers in the FMT group (≥6 letters: 24.0%, ≥15 letters: 8%).

Nowadays, access to another dimension of visual rehabilitation has been achieved by intravitreal VEGF-inhibitors. Although it makes no sense to compare the visual gain of different studies and populations, the results of the ANCHOR study—with an increase in DVA of ten ETDRS letters already after 3 months [27]—might indicate a superior outcome. Even the improvement of CS, seen in patients with predominantly classic CNV after receiving monthly ranibizumab injections, was maintained over 24 months (+3.7 letters) [12].

In summary, the benefit of FMT is reflected in a significant increase in high contrast acuity (near and distance), but no improvement of CS could be achieved by FMT or PDT. Patients have to be informed about the unequal changes and variation of functional parameters in time.

Copyright information

© Springer-Verlag 2007