Efficacy of preoperative nonsteroidal anti-inflammatory drug and the re-dilation technique in minimizing miosis after femtosecond laser in cataract surgery

Submitted for publication: October 10, 2017 Accepted for publication: July 24, 2018 Funding: No specific financial support was available for this study. Disclosure of potential conflicts of interest: None of the authors have any potential conflicts of interest to disclose. Corresponding author: Bruna V. Ventura. Fundação Altino Ventura FAV Rua do Progresso, 71 Recife, PE 50070-020 Brazil E-mail: brunaventuramd@gmail.com Approved by the following research ethics committee: Fundação Altino Ventura (# 47333115.9.0000.5532). ABSTRACT | Purpose: To assess the efficacy of using a nonste­ roidal anti­inflammatory drug preoperatively and of applying the re­dilation technique when necessary to minimize pupil size variation when comparing the degree of mydriasis before femtosecond laser pretreatment with that at the beginning of phacoemulsification. Methods: This retrospective study included patients who underwent cataract surgery using the LenSx (Alcon Laboratories, Inc., Fort Worth, TX). Our routine dilating regimen with flurbiprofen, tropicamide, and phenylephrine was used. The re­dilation technique was applied on eyes that manifested with a pupillary diameter that was smaller than the programmed capsulotomy diameter after laser pretreatment. The technique consists of overcoming pupillary contraction by instilling tropi­ camide and phenylephrine before phacoemulsification. Pupil size was assessed before femtosecond laser application and at the beginning of phacoemulsification. Results: Seventy­five eyes (70 patients) were included. Nine (12%) eyes underwent the re­dilation technique. There was no significant difference in mean pupillary diameter and mean pupillary area between the two studied surgical time points (p=0.412 and 0.437, respectively). The overall pupillary area constriction was 2.4 mm2. Immediately before opening the wounds for phacoemulsification, none of the eyes presented with a pupillary diameter <5 mm, and 61 (85.3%) eyes had a pupillary diameter >6 mm. Conclusion: Preoperative administration of nonsteroidal anti­inflammatory drug and the re­dilation technique resulted in no significant pupil size variation in eyes that were pretreated with the femtosecond laser, when comparing the measurements made before the laser application and at the beginning of phacoemulsification. This approach can avoid the need to proceed with cataract extraction with a constricted pupil.

ABSTRACT | Purpose: To assess the efficacy of using a nonste roidal antiinflammatory drug preoperatively and of applying the redilation technique when necessary to minimize pupil size variation when comparing the degree of mydriasis before femtosecond laser pretreatment with that at the beginning of phacoemulsification.Methods: This retrospective study included patients who underwent cataract surgery using the LenSx (Alcon Laboratories, Inc., Fort Worth, TX).Our routine dilating regimen with flurbiprofen, tropicamide, and phenylephrine was used.The redilation technique was applied on eyes that manifested with a pupillary diameter that was smaller than the programmed capsulotomy diameter after laser pretreatment.The technique consists of overcoming pupillary contraction by instilling tropi camide and phenylephrine before phacoemulsification.Pupil size was assessed before femtosecond laser application and at the beginning of phacoemulsification.Results: Seventyfive eyes (70 patients) were included.Nine (12%) eyes underwent the redilation technique.There was no significant difference in mean pupillary diameter and mean pupillary area between the two studied surgical time points (p=0.412 and 0.437, respectively).The overall pupillary area constriction was 2.4 mm 2 .Immediately before opening the wounds for phacoemulsification, none of the eyes presented with a pupillary diameter <5 mm, and 61 (85.3%) eyes had a pupillary diameter >6 mm.Conclusion: Preoperative administration of nonsteroidal antiinflammatory drug and the redilation technique resulted in no significant pupil size variation in eyes that were pretreated with the femtosecond laser, when comparing the measurements made before the laser application and at the beginning of phacoemulsification.This approach can avoid the need to proceed with cataract extraction with a constricted pupil.

INTRODUCTION
Femtosecond laser in cataract surgery is useful for both routine and challenging cases (13) .However, its use is not exempt from adverse events, such as induction of significant miosis (48) .In fact, previous studies reported pu pil size reduction after laser pretreatment in up to 32% of cases (4) .Given the increased risk of surgical compli cations (9) , preventing miosis is an important issue among cataract surgeons.
Diakonis et al. (7) compared three laser platforms with regard to pupil size alteration in femtosecond laserassis ted cataract surgery (FLACS).They found that the LenSx (Alcon Laboratories, Inc., Fort Worth, TX); the Catalys Precision Laser System (OptiMedica, Abbott Me dical Optics, Santa Ana, CA); and the Victus (Bausch & Lomb, Inc., Rochester, NY) significantly decreased the pupillary diameter, with the LenSx inducing the highest degree of miosis, followed by the Catalys and, finally, the Victus.Another study reported a 29.7% decrease in pupillary area after laser pretreatment using the Catalys Preci sion Laser System (8) .Notably, the preoperative di lating regimen used in the previous studies did not include a nonsteroidal antiinflammatory drug (NSAID).Mo reover, the degree of miosis correlated with age, time for lens fragmentation, and time for the creation of the main incision.
NSAIDs inhibit the synthesis of prostaglandin, which is an intraocular mediator of inflammation (10) .Higher le vels of prostaglandin are seen during FLACS than during conventional surgery, and studies have suggested their important role in the increased risk of intraoperative mio sis after femtosecond laser use (1115) .NSAIDs are effective in lowering these prostaglandin levels (13,14) and have the potential benefit of decreasing intraoperative pupillary contraction during FLACS (15) .Aside from the use of NSAIDs, a greater surgeon experience and the improvement in laser software have decreased the occurrence of miosis after femtosecond laser application (5) .However, despite all efforts, some cases still develop a small pupil before cataract extraction (7) .We recently described redilation as a technique to manage significant miosis caused by the femtosecond laser in some eyes (16) .This technique consists of overcoming pupillary contraction with the use of more dilating drops before proceeding with pha coemulsification.
Since there is a lack of knowledge regarding the effec tiveness of using preoperative NSAID combined with the redilation technique, when necessary, the aim of the present study was to assess the efficacy of this approach to minimize pupil size variation by comparing the degree of mydriasis before femtosecond laser pretreatment with that at the beginning of phacoemulsification.

METHODS
This retrospective study was approved by the institu tional review board of the Altino Ventura Foundation, in Recife, Brazil and followed the tenets of the Declaration of Helsinki.All patients who underwent cataract surgery by two surgeons (MCV and BVV) who used the LenSx at the HOPE Eye Hospital, in Recife, Brazil between March 2015 and October 2015 were eligible for inclusion.All femtosecond laser pretreatments were done by a single doctor (BVV) who had extensive experience on surgeries using the LenSx, whereas the phacoemulsification was done by either one of the two surgeons (MCV and BVV).Patients who received topical treatment for glaucoma or any other disease and those with inflammatory eye di sease, previous ocular surgery or trauma, pseudoexfo liation syndrome, preoperative zonular weakness, history of treatment with an alphaadrenergic antagonist, history of poor pupillary dilation (<5 mm), or rheumatologic disease were excluded from the study.At our institution, patients for surgery are randomly assigned in one of two rooms, one of which has a video recording system.Therefore, we excluded cases that did not have surgical videos of both the femtosecond laser pretreatment and phacoemulsification.
The patient's charts were reviewed to collect informa tion on age, gender, hypertension status, diabetes mellitus history, preoperative intraocular pressure (IOP), endo thelial cell count, and central corneal thickness.In addi tion, we took note of the laser parameters, such as energy used to create the capsulotomy and lens frag mentation; crystalline lens thickness measured intrao pe ratively by the spectral domain optical coherence tomographer (OCT) of the LenSx; the distance measured by the OCT as the delta up and delta down, which corresponds to the cylindrical area where the laser is fired to perform the capsulotomy; whereas an arcuate incision was perfor med; and the total suctionon time.The total suctionon time was defined as the time period between turning on and off suction in the patient interface.Given that the LenSx did not automatically register this parameter, the videos from the LenSx treatment were reviewed and the suctionon time was measured manually and recorded to one decimal place of a second.

Surgical technique
Our routine pupildilating regimen was used in all cases and consisted of administration of 1% tropicamide (Mydriacyl, Alcon Lab. Inc., Fort Worth, TX) and 10% phe nylephrine (Fenilefrina, Allergan, Irvine, CA) four times one hour before the surgery and once between the laser application and the phacoemulsification.In addition, a drop of 0.03% flurbiprofen (Ocufen, Allergan) was ins tilled once an hour before the surgery.All proce dures were performed under topical anesthesia using 0.5% pro xymetacaine chloride (Anestalcon, Alcon Lab.).
The femtosecond laserassisted pretreatment inclu ded main corneal incision, side port corneal incision, capsulotomy, and lens fragmentation.Depending on the preoperative corneal astigmatism, an arcuate incision was also made.The energy used to perform the main incision and the side port incision was 5 μJ.The energy used to make the capsulotomy varied from 6.00 to 7.50 μJ, with 0.5μJ intervals, and was adjusted during surgery with the objective of avoiding incomplete capsulotomies.The delta up and delta down were preset to 275 μm and 350 μm, respectively, and were altered according to the discretion of the surgeon who made adjustments in the laser parameters intraoperatively.Table 1 shows the femtosecond laser parameters that were used.
The lens fragmentation pattern comprised a 2mm central cylinder combined with two 4.8mm chops.There were two energy protocols for lens fragmentation: 1) the standard protocol, in which 8 μJ was applied to the anterior 2 mm of the lens and 7 μJ was applied to the posterior 3 mm of the lens and 2) the dense cataract protocol, in which 11.50 μJ was applied to the anterior 2 mm of the lens and 9.50 μJ was applied to the posterior 3 mm of the lens.We routinely used the dense cataract protocol on eyes that had a nuclear opalescence ≥3 in the Lens Opacity Classification System III (17) .
After the laser pretreatment and before starting phacoemulsification, we routinely checked the patient's pupillary status.The redilation technique was applied on eyes that evolved with a pupillary diameter that was smaller than the programmed capsulotomy diameter, which hindered identification of the capsulotomy's bor ders (16) .This technique consisted of redilating the pupil using 1% tropicamide and 10% phenylephrine every 10 minutes for 30 minutes before proceeding with phacoe mulsification.
The surgical videos of both the femtosecond laser pretreatment and the phacoemulsification of all patients were reviewed.Two images were captured for each eye, one immediately after turning on the suction of the femtosecond laser at the start of the pretreatment and another under the surgical microscope, immediately be fore opening the incisions to start the phacoemulsifica tion.These images were used to calculate the pupillary and capsulotomy diameter and area using the ImageJ software (National Institute of Health) (Figure 1) (A) .Pu pillary diameter was assessed based on its maximum horizontal dimension.On the two video images that were obtained, we used our routine programmed cap sulotomy diameter of 4.9 mm to determine the pupillary diameter and area, based on the following formulae: Pupillary area (mm 2 )=(pupillary area on video/capsulotomy area on video) × π × (programmed capsulotomy diameter/2) 2 Pupillary diameter (mm)=(pupillary diameter on video/ capsulotomy diameter on video) × π × (programmed capsulotomy diameter) A pupillary diameter <6 mm was considered small for cataract extraction, whereas <5 mm was considered clinically significant for cataract extraction (7) .
In addition, the distance (mm) between the capsu lotomy and pupillary borders was calculated as the pu pillary diameter minus the capsulotomy diameter divided by 2.

Statistical analysis
Statistical analyses were performed using SPSS for Windows (version 18.0, SPSS Inc., Chicago, Illinois, USA) and Microsoft Office Excel (Microsoft, Redmond, Washington, USA).The results of the qualitative varia bles were expressed by their absolute and relative fre quencies; whereas the results of the quantitative variables were expressed by their minimum and maximum values and mean and standard deviation (SD).The paired Student's t-test was used to compare the pupillary area and dia meter before the laser pretreatment with those at the beginning of phacoemulsification.The Pearson corre lation coefficient and the independent sample ttest were used to identify statistically significant correlations between the reduction in the pupillary diameter with the quantitative and qualitative variables, respectively.Normality was checked by the Kolmogorov-Smirnov test.A p<0.05 was used throughout this study to reject the null hypothesis.

RESULTS
The total number of FLACS done during the study pe riod was 262, of which 133 were done in the room with a video recording system.Of these, 58 were excluded following the exclusion criteria.If the second operating room had a video recording system, 61 of the 129 eyes operated would have been excluded based on our ex clusion criteria.

A B D C
Finally, 75 eyes of 70 patients were analyzed in this study.The participants' mean age was 69.3 ± 7.9 years (range, 5090 years); 49 (70.0%) were women; 38 (54.3%) had hypertension; and 17 (24.3%)had diabetes mellitus.The patients' preoperative and femtosecond laser data are shown in table 2. Nine (12%) eyes evolved with a pupillary diameter that was smaller than the programmed capsu lotomy diameter after the laser pretreatment and before moving the patient into the operating room for phacoe mulsification.In these eyes, the redilation technique was applied before proceeding with cataract extraction.
Table 3 shows the pupillary diameter and area of all the eyes before the laser pretreatment and immediately before opening the wounds for phacoemulsification.The mean pupillary diameter and area did not significantly differ between these two surgical time points (p=0.412 and 0.437, respectively), even in the subgroup compari son between eyes that did not need the redilation and those in which the technique was used.The mean pu pillary diameter before laser application and before pha coemulsification in the first subgroup was 6.9 ± 0.6 mm and 6.9 ± 0.7 mm, respectively, (p=0.4) and the mean pupillary area was 37.2 ± 6.0 mm 2 and 37.1 ± 7.2 mm 2 , respectively (p=0.5); a similar comparison in the second subgroup showed a mean pupillary diameter of 6.8 ± 0.5 mm and 6.4 ± 0.8 mm, respectively (p=0.1) and a mean pupillary area of 35.8 ± 5.3 mm 2 and 31.9 ± 7.9 mm 2 , respectively (p=0.1).Both subgroups had statistically similar mean pupillary diameter (p=0.2) and area (p=0.3)before femtosecond laser application, but they differed at the beginning of phacoemulsifi cation with regard to both parameters (p=0.024 and 0.025, respectively).
When analyzing all eyes, the overall pupillary area constriction was 2.4 mm 2 .The mean distance between the capsulotomy and the pupillary border did not signi ficantly differ before femtosecond laser application and at the beginning of phacoemulsification [1.0 ± 0.3 mm (range, 0.41.7 mm) vs. 1.0 ± 0.4 mm (range, 0.12.0 mm), respectively; p=0.219].Immediately before opening the wounds for phacoemulsification, none of the eyes presented with a pupillary diameter <5 mm, and 64 (85.3%) eyes had a pupillary diameter >6 mm.Table 4 shows the distribution of eyes with regard to pupillary diameter before the laser pretreatment and immediate ly before opening the wounds for phacoemulsification.None of the analyzed variables correlated with reduc tion in pupillary diameter when comparing the measu rements made before the laser application with those at the beginning of phacoemulsification (Table 5).

DISCUSSION
Miosis secondary to femtosecond laser pretreatment in cataract surgery imposes surgical challenges.Mecha   nical dilation devices, such as iris retractors and Malyugin ring (Microsurgical Technology, Redmond, WA), can be used to manage the constricted pupil (18) .However, since the capsulotomy has already been done by the laser and the small pupil hinders the visualization of the capsulotomy's edge, the insertion and positioning of these devices can cause capsular damage and further complications.We have recently published the redila tion technique as an alternate solution for these cases (16) , with the goal of obtaining a second dilation that is bigger than the anterior capsulotomy's edge, allowing sufficient visualization as if the miosis had not occurred.In the present study, we assessed the pupillary size before laser pretreatment and at the beginning of phacoemul sification in eyes that received NSAID preoperatively and underwent the redi lation technique if the pupillary diameter was <4.9 mm after laser application.There were no significant differences in the mean pu pillary diameter, pupillary area, and distance between the capsulotomy and the pupillary border before the laser pretreatment and immediately before opening the wounds for phacoemulsification.These results are in contrast with those of previous studies (7,8) .Diakonis et al. (7) reported a statistically significant decrease of 1.42 mm in mean pupillary diameter when using the LenSx.Although that study and ours included eyes that underwent FLACS with the LenSx and were pretreated with NSAIDs, our different results were probably from the variation in the pupildilating regimen and the appli cation of the redilation technique in 12% of our eyes before proceeding with phacoemulsification.
The 12% incidence of clinically significant miosis before cataract extraction after femtosecond laser pretreatment was similar to that described by some au thors (19) , but other studies reported incidences as high as 32% (4) and as low as 1.23% (5) .The fact that none of the eyes in this study started phacoemulsification with a pupillary diameter <5 mm, despite the 12% incidence of secondary miosis, implies the effectiveness of the redilation technique in reversing laserinduced pu pillary constriction in eyes that were pretreated with NSAIDs.This approach allows a satisfactory pupillary diameter for cataract extraction and discards the need to use iris retractors or iris expansion rings.
The effectiveness of our approach is also seen when comparing the percentage of eyes with a pupillary dia meter <5 mm, between 5 and 6 mm, and >6 mm imme diately before starting phacoemulsification in FLACS.In a previous paper, the pupillary diameter was <5 mm in 7.6% of the eyes after laser pretreatment and before starting cataract extraction and >6 mm in only 58.2% of the eyes (7) .On the other hand, in the present study, none of the eyes presented with a pupillary diameter <5 mm immediately before phacoemulsification and 85.3% of the eyes had a pupillary diameter >6 mm.The refore, while the surgeon proceeded with surgery with a clinically significant small pupil for cataract extraction in 7.6% of the eyes in the previous study (7) , increasing the risk of having further complications, this did not happen in any of the eyes included in our paper.When comparing the measurement made before the laser application and immediately before phacoe mulsification, the overall pupillary area constriction in our study was 2.4 mm 2 .This was in contrast with the results of Jun et al. (8) , who reported a pupillary constric tion of 29.7% from the initiation of femtosecond laser pretreatment with the Catalys Precision Laser System to the initiation of phacoemulsification.They did not use tropicamide and phenylephrine between the laser application and the phacoemulsification, preoperative NSAID, or the redilation technique.Furthermore, the different peculiarities among the femtosecond laser platforms in terms of the docking system, total time of laser application, and time and energy to perform each of the pretreatment steps probably had a varying impact on the pupil size during surgery (7) .However, we postu lated the importance of preoperative NSAID and the redilation technique to achieve our results with the LenSx, since another paper has shown that this femto second laser platform induced a higher degree of mio sis compared with that of the Catalys Precision Laser System (7) .Therefore, our data suggested that both the preoperative use of NSAID and the asneeded redilation technique were effective in minimizing pupil size variation, based on the comparison of mydriasis before laser pre treatment and at the beginning of phacoemulsification.
When comparing the two surgical time points in this study, none of the analyzed variables correlated with reduction in pupillary diameter in the eyes that received preoperative NSAID combined with the redilation technique, as necessary.This result differed from the findings of a previous study that reported a correlation of the degree of miosis with patient's age, time for lens fragmentation, and time for main incision creation, when assessing the Catalys Precision Laser System (8) .These dif fe rent results were probably due to the varia tions in the dilating regimen, the combined approach of NSAID and the redilation technique, and the different femtosecond laser platforms used in each study.
The main limitations of our study include its retros pective nature and all the drawbacks associated with this design.We did not have a way to retrospectively measure the pupil size after femtosecond laser application and before moving the patient into the operating room for phacoemulsification.Therefore, we could not report the exact pupil size at this time point in the nine eyes that received the redilation technique.In addition, we did not have a record of the time lapse between the ter mination of the laser treatment and the assessment of the pupil size before proceeding with cataract extraction.Although a previous study did not find a higher degree of miosis with a longer time lapse between laser applica tion and phacoemulsification, this would have been an interesting variable to evaluate (8) .Furthermore, some of the eyes that were operated on during the study period were not included in the analysis, because they were operated in a surgical room without a video recording system.Although inclusion of all eyes that were opera ted on would have been ideal, the cases were randomly assigned to each room, which reduces the potential for sampling bias.
In conclusion, the approach of preoperative NSAID use and the redilation technique, when necessary, was effective in minimizing pupil size variation before the laser pretreatment and at the beginning of phacoemulsi fication.No significant difference in pupil size was seen when comparing these two surgical time points.This approach avoided the need to proceed with phacoemul sification with a constricted pupil and the need to use iris retractors or expansion rings.Furthermore, it allowed a satisfactory pupillary diameter for the surgeons, in order to decrease the risk of complications during FLACS.

Figure 1 .
Figure 1.ImageJ software calculations.Screenshots of the femtosecond laser display immediately after turning the suction on show the capsulotomy border (A) and the pupillary border (B) demarcated by the yellow circle (yellow arrow) created by the ImageJ software.Screenshots of the surgical video immediately before opening the incisions to the start phacoemulsification show the capsulotomy border (C) and the pupillary border (D) demarcated by the yellow circle (yellow arrow) created by the ImageJ software.

Table 3 .
Pupillary diameter and area during femtosecond laser-assisted cataract surgery (n=75 eyes) SD= standard deviation.

Table 4 .
Distribution of eyes according to the pupillary diameter during femtosecond laser-assisted cataract surgery (n=75 eyes)

Table 5 .
Correlation between the patients' demographics and femtosecond laser data with decrease in pupillary diameter during femtosecond laser-assisted cataract surgery Pearson correlation analysis was used in all evaluations, except for hypertension and diabetes mellitus history, arcuate incision confection, energy for primary incision confection, and lens fragmentation protocol, for which Spearman correlation analysis was used. *