T a variety of ocular disorders with

ACTA OPHTHALMOLOGICA SCANDINAVICA 1997 - A comparative study of polyacrylic acid (Viscotears@) liquid gel versus polyvinylalcohol in the treatment of dry eyes J. Brodwall', G. Alme', S. Gedde-Dahl', J. Smith', N. P. Lilliedahl', P. A. Kunz and P. Sunderraj Oslo Centre of Ophthalmology', Oslo, Norway, Ciba Vision Ophthalmics*, Bulach, Switzerland ABSTRACT. We conducted a prospective, randomised, single-masked study comparing the safety and efficacy of polyacrylic acid 0.2% (PAA) and polyvinylalcoho1 1.4% (PVA) in 85 patients (PAA 43, PVA ) with dry eyes. The two groups were similar in patient demographics and study parameters at baseline. With treatment, the reduction in total symptoms (gritty or foreign body sensation, burning sensation, dry eye sensation, photophobia, others) and signs (conjunctival hyperaemia, ciliary injection, corneal and conjunctival epithelial staining) score on PAA was significantly greater than that on PVA at both two and four weeks. The daily frequency of instillation of PAA was significantly less than that of PVA on 1 of the 27 (59%) study days. For overall local tolerance there was a significant preference for PAA compared to PVA by both patients and doctors. Only one patient on each treatment had an adverse event and neither was serious. PAA (Viscotears") was as safe as, but better tolerated and more effective than PVA in the treatment of dry eye conditions. Key words: dry eyes - liquid gel - polyacrylic acid - polyvinylalcohol- tear substitutes - treatment - Viscotears". Acta Ophthalmol. Scand. 1997: 75: 457-41 tion, efficacy, safety and local tolerance of PAA and PVA in patients with dry eyes. However, in the meantime, Marner et al. (199) have published their multicentre, prospective, randomised, open, crossover comparison of carbomer gel with polyvinylalcohol in 1 patients with dry eyes. They found that the carbomer ocular gel was used significantly less frequently when compared with PVA eye drops and achieved significantly better reduction of total symptoms score and prolongation of tear break-up time, although significantly more patients complained of adverse events while using the gel. Unfortunately, the results of the Marner et al. (199) study cannot be compared with the present study, as the concentration of the carbomer used was not specified by Marner et al. (199). he term 'dry eye' is used to describe T a variety of ocular disorders with diverse aetiopathogenesis but similar symptoms: discomfort, a feeling of dryness, burning or stinging, grittiness, foreign body sensation and photophobia. The signs, which may vary according to the precise cause of dry eye, include conjunctival and lid margin hyperaemia, reduced marginal tear strip, tear film debris, reduced tear secretion as evidenced by Schirmer's test, tear film instability as evidenced by break-up time and corneal and conjunctival staining with Rose Bengal. Even after years of intensive research there is no cure for dry eyes. The supplementation of existing tears with artificial tear formulation (tear substitutes) continues to be the mainstay of treatment. Although most dry eye patients obtain marked improvement of their symptoms after treatment with artificial tears, this beneficial effect does not last long due to the short retention time of commonly used tear substitutes (Bach et al. 1972; Lemp 1973). However, Viscotears" (polyacrylic acid, also called carbomer 940, 0.2%/0), a clear liquid gel with high viscosity has been found to have a much longer ocular retention time of approximately 1 minutes compared to 2 minutes for the aqueous tear substitute polyvinylalcohol (PVA) 1.4% (Marquardt & Christ 198). This property of polyacrylic acid (PAA) may be expected to offer prolonged symptomatic relief and reduced frequency of application in patients with dry eyes. We therefore performed a prospective, randomised, investigator-masked, parallel-group study to compare the daily frequency of instilla- Material The test product evaluated was Viscotears" (polyacrylic acid 0.2%), a clear, colourless, highly viscous gel formed of high molecular weight, cross-linked polymers of acrylic acid. It is a slightly hypotonic (250 mosmol) formulation with a neutral ph of 7.3 and preserved with cetrimide 0.01 Yo. It is packaged in polyfoil tubes with application tips. The reference product studied was polyvinylalcohol 1.4% (PVA) preserved with benzalkonium chloride 0.01% and packaged in polyethylene bottles. Patients and Methods Eighty patients suffering from dry eyes were to be recruited for this study accord-

ACTA OPHTHALMOLOGICA SCANDINAVICA 1997 ing to the inclusion and exclusion criteria outlined below. The patient had at least two of the following symptoms in both eyes: gritty/ foreign body sensation, burning sensation, dry eye sensation and photophobia and at least one of the following signs in both eyes: Schirmer test that is < 8 mm/ 5 min on the day of recruitment or tear break-up time of < 10 seconds. Patients fulfilling one or more of the following criteria were excluded from the study: age < 18 and > 75 years, pregnancy or lactation, known hypersensitivity to polyacrylic acid, cetrimide, polyvinyl alcohol, benzalkonium chloride, systemic therapy which may induce corneal deposits (caradarone, chlorpromazine) or influence lacrimal secretion e.g. betablockers, naso-lacrimal obstruction, external eye disease inducing conjunctival inflammation and/or infection and corneal scars, dystrophies and infections, intraocular inflammation, contact lens wear, any local treatment with eye drops and/or ointment other than for dry eye, therapy resistant dry eye patient (non-responders), non-compliance with protocol (e.g. attendance and completion of patient diary) or participating in another trial. The study activities that were performed in connection with the recruited patients are outlined in Table 1. Recruited patients, who were already on treatment for dry eyes, underwent a wash-out period of 7 days during which they discontinued their previous treatment and instead used preservative-free single dose units of sodium chloride 0.9% before commencing on active study treatment. However, recruited patients who were newly diagnosed as having dry eyes started active study treatment without any wash-out period. One of the prime objectives of this study was to compare the frequency of instillation of the two study medications. The study medications were to be used daily on an as required basis - at the onset of any of the symptoms of dry eyes - and the frequency of ocular instillation was evaluated by using a diary card which was filled in daily by the study patients. The symptoms assessed at each visit were: foreign body or gritty sensation, burning sensation, dry eye sensation, photophobia and others. They were assessed on a four-point 0-3 score as follows: (0) = absent, (1) = mild (present but not distressing), (2) =moderate (distressing, but not interfering with daily life) and (3) =severe (very distressing and interfering with daily life). Table 1. Activities undertaken in connection with the study patients. Recruitment Day DayO Day 14 Day28 Check inclusion criteria x \ Check exclusion criteria x X X X Informed consent X Randomisation X Patient diary card X2 X X X Assess Symptoms X X X X Conjunctival injection x X X X Ciliary injection X X X X Epithelial staining X X X X Tear BUT X X X X Schirmer s test X X X Local tolerance X X Dispense medication x3 x4 X Collect bottleshbes X X I Start of treatment. Diary card filled in only by patients needing a wash-out period. Saline 0.9% eyedrops in single dose units. Active Medication. Both conjunctival and ciliary injections were evaluated at each visit. These signs were graded on a four-point 0-3 score as follows: (0) =vessels normal, (1) = some vessels definitely injected, (2) =diffusely injected eye but with individual vessels still discernible and (3)=intensely red eye with individual vessels not easily seen. Conjunctival and corneal epithelial staining were graded after instillation of a single drop of 1% Rose Bengal stain as follows depending on the number of punctate stains: 0=0 to 1, 1 = 2 to 10, 2 = 11 to 32,3 = 33 to 100,4= 101 to 315 and 5 =more than 31 or confluent staining. The cornea and the interpalpebral conjunctiva on either side of each cornea were graded separately and the grades from the three zones were summed up to give the definite score for that eye. Thus, the minimum score for each eye was zero and the maximum score 15. The tear film break-up-time (BUT) was assessed using fluorescein. Schirmer s test was performed without prior use of a local anaesthetic. Local tolerance was evaluated by grading the following elicited complaints arising immediately after instillation of the study medication: burning sensation, blurred vision and sticky eyelids. The grading was on a four-point 0-3 scale similar to that used for the symptoms of dry eyes. No study medication was used for at least two hours prior to the ophthalmic examination on days 14 and 28 of the study to enable a more accurate assessment of the improvement in the dry eye condition. Statistical methods and analysis Patients recruited into the study were considered to have completed the study if they had 28 days of treatment with the study medication and had attended both the 14 day and 28 day follow-up examinations. The daily frequency of instilling the study medication as recorded in the patient diary cards during the 28 days of the active treatment period was analysed by a repeated measures ANOVA (analysis of variance) using Bonferroni s criteria to avoid overstating significance levels in multiple comparisons. Changes in symptoms and signs from day 0 (start of treatment day) recorded after active treatment at day 14 and day 28 examinations were assessed by Chisquared tests for two-way and multi-way tables. However, those clinical tests that produced results which may be regarded as continuous, i.e. number of drops instilled per day, total signs and symptoms score, tear break-up time and Schirmer s, were analysed by ANOVA using the Bonferroni s criterion. The treatment unit was the eye so as to avoid any correlation between eyes in any one patient and this association was confirmed by analysis of co-variance. A probability of 5 0.05 was considered statistically signtficant. Results Eighty-five patients were randomised to treatment with one or the other study medication. The disposition and analyses of these patients are shown in Fig. 1. No

ACTA OPHTHALMOLOGICA SCANDINAVICA 1997 - Ended n=85 Randomlsed n = 85 PVA = 1 Analyaed for efficacy at 4 weeks n = 79 P M = 38 PVA = 41 I I Withdnw before week 4 visit n = 1 I PM=l PVA = 0 Fig. 1. The disposition and analyses of the study patients. Table 2. Baseline characteristics of the study patients. Parameters No. of patients Sex Age, years Race Diagnosis of dry eye Wash-out period Schirmer s test Tear break-up-time Total symptom & sign scores st. err = standard error. Males Females Caucasian Asian Keratoconjunctivitis sicca Primary Sjogren s Secondary Sjogren s Required Not required Polyacrylic acid Polyvinylalcohol (PA4 (PW 43 37 0.2 1 28 9 37 5.83 0.1 9.58 0.79 4.3 0.24 7.07 0.54 7.5 1 0.37 7.05 0.38 4 38 1.8 0 25 11 3 39.25 0.1 9.4 0.79 4.98 0.24.79 0.54 7.71 0.37 7.0 0.38 analyses was possible in four patients who did not attend any follow-up visit after commencement of the study medication and could not be traced. Table 2 shows the baseline characteristics of the 85 recruited patients. The two treatment groups were comparable as demonstrated by chi-squared tests for categorical variables and ANOVA for continuous variables. Of the 85 recruited patients, the number of patients eligible for study analysis varied from 79 (for efficacy at 4 weeks) to 81 (for safety and local tolerance). As the analysed patients differed from the recruited patients by at most six patients for any analysis it was considered redundant to repeat this analysis. At day 0 (start of treatment) and at 2 and 4 weeks after treatment, the symptoms and signs were recorded for both eyes. In the analysis of categorical variables the worse eye was treated separately from the other eye. Of the 85 recruited patients, five had only one eye affected. For of the 80 patients with both eyes affected, the worse eye was chosen on the basis of worse signs and symptoms, Schirmer s test and tear break-up time. The mean number of drops instilled each day from day 1 to day 27 by all the compliant patients was analysed. Day 0, the formal start of treatment was not included in the analysis as not all patients started treatment until the following day. Also, day 28, the formal end of treatment was not included in the analysis as patient diary cards for this day were incomplete. It should be noted that the diary cards recorded the frequency of usage of the drops on a per patient and not a per eye basis. It was observed that the only statistically sigdicant difference between the two treatment groups was on day 1 with the patients on PVA, on average, taking more drops than those on PAA. A single patient who used PAA up to 32 times per day was excluded as an outlier and the daily instillation of drops again plotted in Fig. 2. It was now found that there was a statistically significant difference (p < 0.05) between the two treatment groups on 1 of the 27 (59%) study days analysed, with the PAA patients instilling significantly less drops than those on PVA. For the individual symptoms and signs (symptoms: gritty or foreign body sensation, burning sensation, dry eye sensation, photophobia and others, signs: conjunctival hyperaemia, ciliary injection, epithelial staining) evaluated, the changes at 2 weeks were compared with day 0. The number of patients on PAA report-

ACTA OPHTHALMOLOGICA SCANDINAVICA 1997 rn B4.5 - '0 'ii 4-5 2 3.5-3 = 3-5 2 2.5'' ' ' ' ' ' ' ' ' ' ' ' ' I 0 2 4 8 10 12 14 1 18 20 22 24 2 28 Day of the study Fig. 2. The mean number of drops instilled daily by all compliant study patients excepting the outlier. ing that they were better was more than those on PVA for many of the individual symptoms and signs. This difference was statistically significant for dry eye sensation for both the 'worse' (p = 0.01) and 'other' (p = 0.04) eyes. By four weeks there was a marked improvement in individual symptoms and signs on PAA and this was significantly greater than on PVA for the following symptoms and signs: i) foreign body or gritty sensation for the worse (p = 0.02) and other (p=0.005) eyes; ii) burning sensation for the worse (p=o.o2) and other (p = 0.01) eyes; iii) dry eye sensation for the worse (p=o.ol) and other (p = 0.005) eyes, and iv) epithelial staining for the other (p = 0.04) eye. The total symptoms and signs scores, derived as a sum of the already mentioned individual symptoms and signs for day 0, week 2 and week 4 are shown in Table 3. Both treatments produced a statistically significant reduction in the total symptoms and signs score (p<o.ool at 2 weeks and p < 0.0001 at 4 weeks). Also, the reduction in total symptoms and signs score on PAA was significantly better than that obtained on PVA at both 2 weeks (25% reduction on PAA compared with 10% on PVA, p = 0.003) and 4 weeks (4% reduction on PAA compared with 1% on PVA, p < 0.0001). The tear film break-up-time was assessed at day 0, week 2 and week 4. For both study groups there was no statistically significant change in tear break-uptime after treatment (p = 0.8 at 2 weeks and p=o.1 at 4 weeks). Also, there was no statistically significant difference between the two treatment groups at either two (p = 0.) or four (p = 0.3) weeks. The Schirmer's test was performed on day 0 and week 4. For both treatment groups there was no statistically significant change in Schirmer's test values with time in the worse eye (p = 0.009), but not in the other eye (p = 0.9). Also, there was a statistically significant difference between the two study treatments favouring PAA in the 'other' eye (p = 0.02) but not in the 'worse' eye (p = 0.9). In terms of local tolerance, burning and stinging sensation was statistically significantly (p<o.ool) less on PAA than on PVA at both 2 weeks and 4 weeks. The reverse was true for blurred vision (p < 0.001 at 2 weeks and p=0.009 at 4 weeks). However, the majority of the symptoms of intolerance were usually mild and transient with a median duration of less than 0 seconds. An overall assessment of local tolerance was performed at two weeks and four weeks by the study patients and doctors as poor, fair, good or excellent. All these four assessments demonstrated a statistically significant preference for PAA over PVA (patients at 2 weeks p = 0.004 and 4 weeks p = 0.03; doctors at 2 weeks p < 0.001 and 4 weeks p= 0.02). For global assessment of treatment by the study patients, all compliant patients were asked the following question at 2 weeks and again at 4 weeks: compared to your last visit, did the treatment make your eye much better, no better, or worse? It was noted that at 2 weeks, in comparison to the last visit on day 0, the improvement on PAA was statistically significantly more than that on PVA (75% of PAA patients felt better compared with 39% on PVA; p = 0.004). At four weeks, in comparison to the last visit at week 2, the improvement on PAA was still greater than that on PVA (59% of PAA patients felt better compared with 37% of PVA patients) but this difference was not statistically significant (p = 0.12). Only two study patients had an adverse experience - one patient on PAA discontinued treatment due to complaints of poor tolerance and sticky eyelids, while 'one patient on PVA experienced a mild ocular hypersensitivity reaction. Discussion Marner et al. (199), in the only previously published prospective, randomised comparison of polyacrylic acid (carbomer) ocular gel, with polyvinylalcoho1 eye drops, studied 1 patients with dry eyes. However, we feel that a detailed comparison of the results of the Marner et al. study with that of ours is not appro- PAA PVA Study eye DayO Week2 Diff. DayO Week2 Diff. Worseeye 7.7 5. -2.1 7.7.8-0.9 0.3 0.3 (No.) (39) (41) 7.2 5.4-1.8 7.0.4-0. 0.3 0.3 (No.) (39) (41) Diff. = difference. = standard error. PAA PVA DayO Week4 Diff. DayO Week4 Diff. 7.8 4.3-3.5 7.7.3-1.4 0.4 0.4 (38) (41) 7.3 3.9-3.4 7.0.0-1.0 0.4 0.4 (38) (41) 40

~~~ ~ ~~ ACTA OPHTHALMOLOGICA SCANDINAVICA 1997 - priate as the concentration of the carbomer used by Marner et al. is not known. In our study of patients with dry eyes, the mean number of drops of PAA used per day was significantly less than that of PVA on 59% of the study days. This confirms the expectation raised by previous authors (Leibowitz et al. 1984; Marquardt 198) that the increased ocular retention time of PAA would result in a reduction in the number of drops required daily for the relief of dry eye symptoms. Although the exact mechanism of action of PAA in the treatment of dry eyes in still unclear, it has been proposed that this material persists in the conjunctival sac and slowly melts while absorbing fluid and retaining it locally (Leibowitz 1984). Following treatment, both PAA and PVA produced a significant improvement in the dry eye status of the study patients as indicated by the reduction in the total symptoms and signs score. The improvement obtained with PAA was at least twice as great as that obtained with PVA at both follow-up visits. Also, the improvement on PAA was significantly greater than that on PVA for several of the individual symptoms and signs like foreign body or gritty sensation, burning sensation, dry eye sensation and epithelial staining. The results are consistent with other reports that PAA was highly effective in relieving the symptoms and signs of patients with dry eyes of varying severity (Leibowitz 1984; Marquardt 198; Kessler 1991). PAA has been found to spread well across the corneal and conjunctival surfaces without forming streaks and to significantly increase the tear film break-uptime (Marquardt & Christ 198; van Bijsterveld 1988). In this study, however, although there was an increase in tear film BUT (break-up-time) in PAA patients in both the treated eyes, the increase was not statistically significant. Also, there was no statistically significant difference between the two study groups. These observations may be partly due to the fact that the study patients were instructed to apply their study medication at least two hours prior to their examination visit. We observed a statistically significant improvement in Schirmer s test values with time in both of the two treatment groups in the worse eye but not in the other eye. The clinical significance of this observation is not known. There was, however, a statistically significant difference between PAA and PVA in the non-worse eye in favour of PAA, but this difference was judged clinically insignificant because of the small magnitude of the difference, the lack of demonstrated effect on lacrimal secretion by the study treatments and the variability in Schirmer s test values. All the previous studies on PAA liquid gel (Leibowitz 1984; Marquardt & Christ 198; Marquardt 198; van Bijsterveld 1988; Kessler 1991) have commented on its excellent local tolerance even with extended use. In this study, the burning and stinging sensation on instillation was significantly less with PAA, while blurred vision and sticky eyelids occurred significantly more, probably due to the higher viscosity and increased ocular retention time of PAA. An overall assessment of local tolerance revealed a significant preference for PAA compared with PVA by both the patients and the doctors. This may be because the burning or staining sensation felt in instillation of PVA was perceived to be more troublesome than the transient blurred vision or sticky eyelids experienced by those on PAA. All the study patients were asked to make a global assessment of their dry eye condition following treatment at week 2 (in comparison with the baseline at day 0) and week 4 (in comparison with week 2). At both assessments, the improvement on PAA was more than that on PVA with the difference on the first assessment being statistically significant. This indicates that the PAA patients subjectively obtained a marked improvement in their dry eye condition in the first two weeks of treatment and continued to improve over the next two weeks, although not as dramatically owing to the marked improvement already obtained. In terms of safety, only two patients, one on each treatment, experienced an adverse event and neither was severe. In conclusion, both PAA and PVA were found to be effective in improving the symptoms and signs of dry eyes, with PAA being significantly more effective than PVA at a lesser daily frequency of instillation. Both PAA and PVA were safe and well-tolerated, with PAA being significantly better tolerated than PVA. Overall, the patients demonstrated a significant preference for PAA compared with PVA. We recommend PAA as the medication of first choice in patients with dry eyes requiring therapy with artificial tear substitutes. Acknowledgment We thank Mr Ian Clarke for the statistical analyses presented in this paper. References Bach FC, Adam JB, McWhirter HC & Johnson J E (1972): Ocular retention of artificial tear solptions. Ann Ophthalmol 4: 11-119. Kessler CH (1991): Therapeutic improvement of keratoconjunctivitis sicca by a gel containing polyacrylic acid. Spectrum Augenheilkd 5: 9-75. Leibowitz HM, Chang RK & Mandell AI (1984): Gel tears -A new medication for the treatment of dry eyes. Ophthalmology 91: 1199-1204. Lemp MA (1973): Artificial tear solutions. Int Ophthalmol Clin 13: 221-229. Marner K, Moller PM, Dillon M & Rask- Pederson E (199): Viscous carbomer eye drops in patients with dry eyes. Acta Ophthalmol Stand 74: 249-252. Marquardt R (198): Die Behandlung des trockenen auges mit einem neuer tropffahigen gel (The treatment of dry eye with new liquid gel). Klin Monatsbl Augenheilkd 189: 51-54. Marquardt R & Christ Th. (198): Untersuchungen zur verweildauer von tranenersatzmitteln (Studies on the retention time of tear substitutes). Klin Monatsbl Augenheilkd 189: 187-2. van Bijsterveld OP (1988): Action of carbopolymer gel therapy on lacrimal function parameters in keratoconjunctivitis sicca. Z Prakt Augenheilkd 9: 151-154. Received on August 14th, 1995. Corresponding author: Dr J Brodwall Oslo Center of Ophthalmology Sorkedalsvn. 10 A N-039 Oslo, Norway.