This study was performed as a randomized, double-blinded, placebo-controlled, intraindividual, two-period, and crossover clinical trial. After fully explaining the study protocol and its probable safety and efficacy, written informed consent was obtained from all patients. This clinical trial was approved by the review board/ethics committee of the Ophthalmic Research Center of Shahid Beheshti University of Medical Sciences.

All patients with severe non-PDR and early or high-risk characteristic PDR who were considered as candidates for bilateral PRP by a retina specialist were included in this trial. Patients with a history of intraocular surgery, peripheral retinal laser therapy, or any eye disease that may interfere with the study protocol in either eye were excluded. Vitreous hemorrhage or other media opacity precluding proper laser therapy in either eye was also used as exclusion criteria. A data sheet was completed for all included patients.

Enrolled patients were scheduled to receive three or four sessions of PRP. However, the study protocol was performed only for the first two sessions of treatment. For each person, randomization was performed before the first session. The right eye of each person was randomly assigned to one group based on a randomized permuted block with randomly selected block lengths of 4 and 6. The left eye was assigned to the other group. At the first session and based on randomization, one eye received topical sodium diclofenac ophthalmic solution 0.1% and the other eye topical artificial tear drop (as placebo) three times, 5 minutes apart, at least 20 minutes before laser treatment. At the second session, the eyes were given the alternate drug with the same schedule. An argon laser was used with a wavelength of 532 nm, spot size of 200 µm, exposure time of 0.2 ms, and energy level of 100 to 500 mW.

In this study, pain level was scored using a visual analogue scale (VAS). It consisted of a 10 cm scale arranged from no pain to the worst pain imaginable. Markings of the subjects on the scale are translated to a number from 0 to 10. Before applying the laser, patients were instructed on how VAS works. Patients unable to cooperate with the VAS test were excluded.

Laser therapy was performed based on the Early Treatment Diabetic Retinopathy Study protocol, and started with the right eyes. Surgeons were encouraged to use nearly the same power and amount of laser for both eyes at each stage and to apply the laser on the same quadrants of both eyes, avoiding the area of long ciliary nerves. At the end of laser treatment in each session, patients scored the pain in each eye.

Patients, the nurse who instilled the drops, and the surgeons who performed the laser operation and asked about the pain were all blinded to the drops given. The difference of pain level was compared between the two drugs as the main outcome measure.

A total of 200 eyes of 100 diabetic patients enrolled in and completed the study. The mean age was 54 ± 10 years (range, 26 to 75) and 63% were female. The laser procedure data for each group are presented in Table 1. Groups were comparable regarding the mean number of laser applications, mean energy level of laser, and frequency of the treated quadrants. The mean time interval between instillation of drops and laser procedure was 119 ± 55 minutes, with 114 ± 55 minutes for the first session and 124 ± 55 minutes for the second session.

Pain sensation according to VAS at the first session was 5.9 ± 3.1 in the diclofenac group and 5.5 ± 3.0 in the placebo group (Table 1). The corresponding values for the second session were 5.4 ± 2.8 and 5.5 ± 2.9. Fig. 1 demonstrates change in pain scores in each subject in every session. Collecting both sessions together, pain sensation was 5.6 ± 3.0 and 5.5 ± 3.0 based on VAS in the treated and control groups, respectively. In total, the calculated treatment effect was 0.15 (95% confidence interval [CI], −0.27 to 0.58; p = 0.486) and therefore, there was no significant difference between the treatments. Carryover effect was statistically insignificant (p = 0.283) and the estimated period effect was 0.24 (95% CI, −0.50 to 0.98; p = 0.530).

This trial demonstrated that pretreatment with topical sodium diclofenac ophthalmic solution 0.1% did not diminish pain during PRP in diabetic patients. Pain experienced during PRP is very variable but appears to be more common with retreatment and with treatment applied anterior to the equator, especially in the 3, 6, 9, and 12 o'clock positions (corresponding to the location of the long ciliary nerves) [3]. As suggested by various authors, there are several ways of mitigating this pain. Some patients respond to pretreatment oral analgesia. Others may even require local anesthetic blockade or general anesthesia in order to tolerate treatment. However, these interventions carry the risk of side effects and require additional monitoring of the patient [3]. Inhaled Entonox was also evaluated as an analgesic during PRP in a randomized, crossover, and double-blinded pilot study. Mean pain scores, based on VAS, from Entonox treatment compared to air inhalation (as a placebo) were 2.94 ± 2.73 vs. 3.73 ± 3.20, respectively (p < 0.03). The investigators concluded that Entonox can be used as a safe and effective analgesic agent during PRP treatment [3]. The feasibility of manual acupuncture for reducing pain during PRP treatment was also investigated in a prospective, comparative nonrandomized study and was found to be helpful [12]. In a recent published randomized controlled trial, single spot short duration time (20 ms) was compared with conventional laser therapy as another option for diabetic retinopathy treatment. The investigators found that short pulse laser was significantly less painful but just as effective as conventional laser during 6 months of follow-up [13].

The analgesic effect of NSAIDs is believed to apply through inhibition of the arachidonic acid cascade. This cascade divides into the cyclo-oxygenase and the lipo-oxygenase pathways. The main products of the cyclo-oxygenase pathway are prostaglandins that take part in maintaining and amplifying the cellular and humoral phases of the inflammatory response. During an inflammatory response, several mediators are released that stimulate pain producing nerve fibers [11].

Sodium diclofenac 0.1% belongs to the phenylacetic acid chemical class and appears to have a dual effect on both cyclo-oxygenase and lipo-oxygenase pathways [14]. It is believed that it acts on the posterior segment of the eye either by diffusion into the vitreous from the aqueous or by “desensitization” of the entire distribution of the trigeminal nerve fibers around the globe [11].

The analgesic effect of topical sodium diclofenac 0.1% during PRP and macular laser photocoagulation was evaluated in a prospective, double-blinded, crossover, randomized, and clinical study [11]. In this report, the study population included 87 patients, 45 with PDR treated with PRP (group A) and 42 with non-PDR and clinically significant macular edema (group B) who received grid treatment of the posterior pole. Treatment of the posterior pole was associated with no, mild, or negligible pain and this was attributed to the lower power levels used for grid photocoagulation. However, they found a statistically significant effect (p = 0.01 by paired t-test) in group A by using topical diclofenac. It should be noted that even in this group, nine patients out of 45 had more pain when using diclofenac than using sodium chloride and seven patients had similar levels of pain using both treatments. Nevertheless, the mean reported level of pain was 44.2% when sodium diclofenac 0.1% drops were used and 53.1% when sodium chloride 0.9% drops were used. Although the difference between the groups was statistically significant, the reported level of pain sensation did not differ significantly from the clinical point of view. The mean pain score in our study was similar to the sodium chloride group of this study. It was 5.4 in the treated group and 5.3 in the control group, and we found no significant difference between the groups. The difference between two studies might be due to the different times of drop installations before PRP and the various characteristics of laser application.

In another randomized, double-masked, and placebo-controlled clinical trial, the authors compared oral diclofenac, topical diclofenac, and placebo in pain reduction during PRP and concluded that a single dose of oral diclofenac was an effective pretreatment analgesic agent for reducing pain experienced during PRP for PDR. They found a significantly lower pain level in patients receiving pretreatment of oral diclofenac compared to the controls. However, the difference between topical diclofenac and placebo was not significant in univariate analysis [15]. Although this latter result was comparable to our result, multivariate regression analysis for age, gender, and total laser energy in their study demonstrated a significantly lower pain level for topical diclofenac versus placebo [15].

Despite the above reported beneficial effect of the topical sodium diclofenac 0.1% during PRP, there is another study with a conclusion similar to our results. In this trial, the investigators evaluated the effect of topical ketorolac 0.5%, another topical NASAID, for ocular pain relief during PRP [16]. It was a prospective, randomized, double-blinded, and controlled study with 60 eyes of 30 consecutive PDR patients. One hour before laser treatment, ketorolac tromethamine 0.5% was instilled to one eye and artificial tear drop to the fellow eye. Directly after treatment, patients were asked for the severity of pain in both eyes using VAS. Mean pain level for placebo-instilled eyes was 4.8 and 4.4 for ketorolac-instilled eyes. There was no significant difference for pain levels between the groups (p = 0.29). The conclusion of this study was parallel to ours, but instead using a different type of topical NSAID.

In the present trial, the mean time interval between instillation of the drops and laser treatment was about 115 minutes. Although this waiting time would not be convenient for daily practice, it was an appropriate interval in order to achieve near maximum analgesic effect of topical sodium diclofenac 0.1%. The highest average concentration of this drug to be found in the aqueous humor was 82 ng/mL at 2.4 hours after instillation [17]. This time interval was 68.3 minutes in the study using diclofenac [11] and 1 hour in the trial using ketorolac 0.5% [16].

Our study was sufficiently powered with 100 cases, 200 eyes, and 400 interventions. Performing a randomized double-masked placebo-controlled intraindividual crossover clinical trial enabled us to overcome many biases and limitations as the subjective-dependent main outcome of this study, which was pain sensation. Nonetheless, we used the VAS test, which has been found to be correlative and reproducible [18]. It is an acceptable and widely used method for measuring pain sensation. Additionally, surgeons were encouraged to try to apply the laser on the same quadrants of both eyes in each session since the level of pain sensation may vary based on different quadrants receiving the laser.

In summary, we report that using topical diclofenac sodium as an analgesic before PRP in patients with diabetic retinopathy was not helpful in pain reduction during the procedure. As previously reported, the drug is supposed to act on the posterior segment of the eye either by diffusion into the vitreous from the aqueous or by “desensitization” of the entire trigeminal nerve fibers around the globe [11]. However, we did not observe any effect using this type of intervention in our study. We believe that the concentration of the drug around the corresponding nerves with the aforementioned method of administration would not be enough to diminish pain level during a relatively painful procedure like PRP. Further studies evaluating other modalities and routes of administration are warranted.