Treating socket discharge and discomfort: a crossover randomised trial of vitamin A ointment, low-dose steroid eyedrops and artificial tears

Abstract

Purpose

This study evaluated the effects of artificial tears, vitamin A ointment and low-dose steroid drops on chronic irritation and discharge in anophthalmic patients wearing ocular prostheses. It also aimed to identify patient subgroups that may benefit from specific regimens.

Methods

This prospective, crossover randomised trial used a Williams design with 40 ocular prostheses wearing participants aged 16 years or older who experienced chronic discharge or discomfort.

Each participant underwent four 2-week intervention periods: artificial tears, vitamin A ointment, low-dose steroids (fluorometholone) and no medication as a control, separated by 2-week washout periods. Visual analogue scale questionnaires assessed discharge and irritation frequency and severity. Data were analysed using linear mixed models with Bonferroni correction for multiple comparisons and subgroup analyses were descriptive.

Results

In the total group, vitamin A ointment most consistently improved all outcome parameters: irritation frequency (−1.766, 95% CI (−2.857 to –0.675), p<0.001), irritation severity (−1.448, 95% CI (−2.564 to –0.332), p=0.004), discharge frequency (−1.186, 95% CI (−2.295 to –0.077), p=0.029) and discharge severity (−1.146, 95% CI (−2.177 to –0.114), p=0.021). Artificial tears significantly reduced irritation frequency (−1.279, 95% CI (−2.370 to –0.188), p=0.013). Fluorometholone showed no significant reduction in the total group. Individually, there was considerable variation among the outcomes, indicating that the optimal regimen differs between individuals.

Conclusions

Vitamin A ointment demonstrated the most consistent benefit in anophthalmic patients wearing ocular prostheses, whereas artificial tears and low-dose steroid drops were associated with clinically relevant improvement in a subset of individual patients with chronic irritation and discharge.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Chronic mucopurulent discharge and discomfort frequently affect anophthalmic patients with ocular prostheses due to tear film instability, epithelial damage, inflammation and mechanical irritation. Existing treatments—artificial tears, vitamin A ointment and low-dose steroids—have been used empirically without strong comparative evidence or standardised protocols.

WHAT THIS STUDY ADDS

• This is the first randomised controlled trial to systematically compare these therapies, finding vitamin A ointment significantly more effective in reducing irritation and discharge than other treatments.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• The findings support vitamin A ointment as first-line treatment and emphasise the importance of personalised management. This evidence fills a critical gap and can guide clinical decision-making and future research towards optimised, patient-tailored therapies for ocular prosthesis complications.

Introduction

The anophthalmic socket, resulting from the surgical removal of an eye due to trauma, infection, malignancy or congenital conditions, presents significant challenges in both functional and aesthetic rehabilitation. Ocular prostheses play a crucial role in restoring the cosmetic appearance and psychological well-being of patients who have undergone enucleation or evisceration. However, a substantial proportion of prosthetic eye wearers experience chronic complications such as socket discomfort and excessive mucopurulent discharge—among the most frequently reported issues postsurgery.¹ These complications can lead to reduced quality of life and compromised prosthesis tolerance.² Despite the high prevalence of these issues, there remains a lack of standardised, evidence-based treatment protocols for managing socket irritation and discharge.³

The pathophysiology of the anophthalmic socket involves substantial anatomical and physiological adaptations following eye removal. Alterations in tear film composition occur and spaces behind the prosthesis can allow for the pooling of socket fluids, functioning as growth medium for bacteria, which increases mucous secretion and contributes to excessive recurrent discharge.^{3 4} Wearing an ocular prosthesis can also induce mechanical irritation and a reduction in the meibomian gland surface area, ultimately contributing to dry anophthalmic socket syndrome (DASS).^{4 5} This condition is characterised by tear film instability, increased conjunctival keratinisation and persistent irritation, which collectively result in discomfort and challenges in prosthetic wear.⁵ Additionally, factors such as a history of radiation therapy or chemical burns may cause conjunctival inflammation or scarring further exacerbating DASS.^{6 7}

In severe cases, progressive orbital fat atrophy, inflammation and extraocular muscle dysfunction can contribute to fibrosis, socket contracture, inadequate prosthesis motility and lagophthalmos impeding the natural wiping properties of the eyelids. An ill-fitting prosthesis can exacerbate tissue changes, further disrupting the socket microenvironment and leading to secondary complications.⁷ While common causes of discomfort or discharge include old, unpolished or improperly fitted prostheses as well as medical conditions such as implant exposure or nasolacrimal duct obstruction, some patients continue to experience persistent symptoms despite corrective interventions.⁸ In such cases, long-term pharmacological management may be required to alleviate discomfort and maintain ocular surface health.

Current strategies for managing chronic socket discomfort and discharge often involve the use of artificial tears, vitamin A ointment and low-dose topical corticosteroids, along with minimising the handling and cleaning of the prosthesis.⁹ Artificial tears are widely used for their lubricating properties, which help maintain ocular surface hydration and mechanically flush out irritants.¹⁰ Vitamin A ointment has been recognised for its role in epithelial repair and ocular surface stabilisation, potentially offering therapeutic benefits for the conjunctival surface of the anophthalmic socket.¹¹ Conversely, low-dose topical corticosteroids, such as fluorometholone (FML), are prescribed to mitigate inflammatory responses and irritation; however, their efficacy in reducing chronic discharge in prosthetic eye wearers has not been rigorously evaluated.¹²

Given the variability in treatment approaches and the absence of robust evidence guiding clinical practice, there is a pressing need to evaluate and compare the efficacy of these therapeutic options. This study addresses this gap by systematically comparing the effectiveness of artificial tears, vitamin A ointment and low-dose topical corticosteroid drops (FML) in managing chronic mucopurulent discharge and socket discomfort in prosthetic eye wearers. Furthermore, the study aims to identify patient subgroups that may derive greater benefit from specific treatments, thereby informing personalised clinical management strategies.

Methods

Study design

This study was a randomised crossover trial utilising a 4×4 Williams design to assess the effects of topical artificial tears (Hylan, Tramedico, the Netherlands), vitamin A eye-ointment (Hylo-Night, Ursapharm, Germany) and FML eyedrops (FML Liquifilm, Allergan) on chronic mucopurulent discharge and discomfort in anophthalmic patients wearing eye prostheses. A crossover design was chosen over a parallel-group design to allow each participant to serve as their own control, thereby minimising interparticipant variability and increasing the reliability of treatment comparisons. The Williams design further ensured a balanced allocation of treatment sequences across participants, minimising potential biases from treatment order effects. Each participant underwent a 16-week study period, beginning with a 2-week run-in period during which any ongoing medication was discontinued to establish a consistent baseline across all participants.

The intervention phase consisted of four 2-week periods: three involving one of the active treatments and one period serving as a control (no medication). Each intervention period was concluded with questionnaires and separated by a 2-week washout period to mitigate carryover effects. No important changes to the methods were made after trial commencement.

Patients

Patients were recruited from February 2024 till May 2024 according to the International Conference on Harmonization Good Clinical Practice Guidelines found in the Declaration of Helsinki, including approval by the Amsterdam University Medical Centre Ethics Committee. Possible participants were identified either through their treating ocularist or ophthalmologist or through the ‘OOG-in-OOG’ association; a Dutch patient organisation dedicated to supporting individuals with eye loss and those wearing an ocular prosthesis. All participants provided written informed consent before any study procedures were conducted.

Eligibility criteria

Patients aged 16 years or older with ocular prostheses who suffered from chronic excessive mucopurulent discharge and/or discomfort were included in the study. We excluded patients unable to understand the study protocol or complete questionnaires, and patients with treatable medical conditions of the ocular socket, such as conjunctival cysts, exposed implants or (sub)acute infections requiring antibiotics. Patients with issues related to scratched or irregular prostheses, overdue polishing or poor prosthetic fit (eg, large empty spaces or pressure points) were also excluded. However, after adequately addressing these underlying conditions, some of these patients were included if symptoms of mucopurulent discharge and discomfort persisted. Finally, participants experiencing a common cold during the study were asked to pause participation until symptoms had resolved for at least 2 weeks to minimise potential confounding effects on study outcomes.

Procedures and main outcome measures

Demographic and clinical data were collected, including type of eye removal surgery, number of eye and/or socket surgeries, prosthesis and implant type, prosthesis cleaning frequency, presence of lagophthalmos and conjunctival damage, and characteristics of mucopurulent discharge (presence, frequency, colour and volume) as well as DASS. All participants underwent a comprehensive ophthalmic examination, and data were stored securely in accordance with local data protection regulations.

Primary outcomes—irritation severity and frequency, and discharge severity and frequency—were assessed using a digital continuous visual analogue scale (VAS) administered via an online Castor EDC survey. Frequency was scored from 1 (‘barely’) to 10 (‘constant’), and severity from 1 (‘mild’) to 10 (‘severe’). Participants rated discharge-related symptoms (tearing, crusting or discharge) and irritation-related symptoms (dryness or irritation). A reduction of >2 points was predefined as clinically relevant. The questionnaire was adapted from the Symptom Assessment in Dry Eye (SANDE) questionnaire to include mucopurulent discharge-related outcomes.¹³

Secondary outcomes examined associations between patient characteristics and treatment response. The ophthalmic examination assessed the presence of lagophthalmos, conjunctival damage (fluorescein staining), predominant mucopurulent discharge complaints and DASS, defined as a DEQ-5 score ≥6 plus at least one objective sign (conjunctival staining, reduced tear meniscus height and/or blepharitis).¹⁴

Participants self-administered artificial tears four times per day, fluorometholone eyedrops two times per day and vitamin A ointment one time per day before sleep; all medications were prescribed through the participants’ own pharmacy.

Sample size

The sample size for this study was calculated using Power Analysis and Sample Size (PASS) software. For a 4×4 Williams crossover design, a sample size of eight participants per sequence (32 total participants) provided 81.5% power to detect a 2.0-point difference on the VAS scale (prespecified as the minimum clinically important difference by expert consensus), using a two-sided t-test with an overall significance level of 0.05 and a Bonferroni-adjusted individual test significance level of 0.00833, assuming an SD of paired differences of 3.0. As there are no prior pilot studies or similar investigations evaluating these regimens with the same outcomes, these parameters were determined based on ophthalmologist consensus. To account for an estimated dropout rate of 20%, an additional eight participants were added, resulting in a total target enrolment of 40 participants (10 per group).

Randomisation

Participants were randomised by the coordinating investigator, using an electronic randomisation tool (Castor EDC). Block randomisation was performed with block sizes of 4, creating a total of 10 blocks. Each participant within a block was assigned a letter (A through D) that determined in which order the treatment regimens were followed.

Statistical methods

The effectiveness of each treatment regimen compared with no treatment was evaluated using a linear mixed model with treatment as a fixed effect (no-medication as reference), random intercepts for participants and sequences, and autoregressive covariance structure to account for repeated measures and crossover design correlations. The 4×4 Williams design with block randomisation (block size 4) balanced sequences to minimise order/period effects; carryover was ruled out via residual diagnostics and model fit statistics. Two-week washout periods matching treatment duration were confirmed adequate by consistent baseline VAS scores across periods. Pairwise comparisons used Estimated Marginal Means with Bonferroni correction. The study adhered to Consolidated Standards of Reporting Trials 2010 crossover guidelines.¹⁵ Secondary endpoints were analysed descriptively.

Results

Study population

The demographic and clinical characteristics of the study participants, both per sequence and overall, are summarised in table 1. During the recruitment period (February 2024–May 2024), a total of 61 individuals were assessed for eligibility. Of these, 21 participants were excluded prior to trial initiation: 14 subjects did not respond to the invitation for the physical intake appointment, 2 did not meet the inclusion criteria and 5 declined participation following a discussion of the trial procedures. The patient disposition is illustrated in figure 1. Each of the four groups, comprising 10 subjects, successfully completed the trial, with no losses to follow-up and no (severe) adverse events. Among the 40 enrolled participants, only the presence of lagophthalmos (20 with vs 20 without) and the presence of DASS (22 with vs 18 without) were evenly distributed, ensuring balance for descriptive secondary outcome assessments.

Table 1. Baseline demographic and clinical characteristics by sequence and total.

Characteristics	Group A (N=10)	Group B (N=10)	Group C (N=10)	Group D (N=10)	Total (N=40)
Mean (SD) age (years)	60 (18.4)	48 (16.7)	62 (6.4)	63 (11.3)	58 (15.4)
Gender
Male	3 (30)	3 (30)	3 (30)	3 (30)	12 (30)
Female	7 (70)	7 (70)	7 (70)	7 (70)	28 (70)
Eye removal surgery type
Evisceration	3 (30)	5 (50)	5 (50)	6 (60)	19 (47.5)
Enucleation	4 (40)	2 (20)	3 (30)	1 (10)	10 (25)
Other*	0 (0)	1 (10)	1 (10)	1 (10)	3 (7.5)
Unknown	3 (30)	2 (20)	1 (10)	2 (20)	8 (20)
Prosthesis type
Synthetic	9 (90)	9 (90)	10 (100)	10 (100)	38 (95)
Glass	1 (10)	1 (10)	0 (0)	1 (10)	2 (5)
Implant type
Non-integrated	7 (70)	6 (60)	6 (60)	8 (80)	27 (67.5)
Dermis fat graft	2 (20)	1 (10)	0 (0)	0 (0)	3 (7.5)
Other†	0 (0)	1 (10)	1 (10)	1 (10)	3 (7.5)
Unknown	1 (10)	2 (20)	3 (30)	1 (10)	7 (17.5)
Cleaning frequency
Daily	3 (30)	4 (40)	4 (40)	3 (10)	14 (35)
Weekly	1 (10)	1 (10)	0 (0)	1 (10)	3 (7.5)
Monthly	3 (30)	1 (10)	1 (10)	2 (20)	7 (17.5)
Yearly	3 (30)	4 (40)	5 (50)	3 (30)	15 (37.5)
No cleaning	0 (0)	0 (0)	0 (0)	1 (10)	1 (2.5)
Lagophthalmos
Yes	6 (60)	7 (70)	6 (60)	1 (10)	20 (50)
No	4 (40)	3 (30)	4 (40)	9 (90)	20 (50)
Conjunctival damage
Yes	0 (0)	1 (100)	0 (0)	0 (0)	1‡ (2.5)
No	0 (0)	0 (0)	0 (0)	0 (0)	39 (97.5)
Mucopurulent discharge
Yes	10 (25)	10 (25)	10 (25)	10 (25)	100 (100)
No	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
DASS
Yes	8 (80)	6 (60)	5 (50)	3 (30)	22 (55)
No	2 (20)	4 (40)	5 (50)	7 (70)	18 (45)

Yes if, Dry Eye Questionnaire (DEQ-5) > 6 with conjunctival staining or reduced tear meniscus height or blepharitis.

Data are number (%) unless stated otherwise.

^*Group B and C: one patient has microphthalmia. Group D: one patient had a phthisis bulbi.

^†Group B: no implant.

^‡Caused by radiation therapy.

DASS, dry anophthalmic socket syndrome.

Figure 1. Study design overview of the randomised 4×4 Williams crossover trial.

Primary outcomes

Of the 40 participants, 29 (72.5%) showed a clinically relevant response (≥2-point reduction) to ≥1 treatment in ≥1 outcome, while 33 participants (82.5%) had >1-point improvement in ≥1 outcome. All three interventions reduced symptoms compared with no medication (figure 2), with vitamin A ointment achieving statistically significant effects across all four outcomes: irritation frequency (−1.77 points, 95% CI −2.86 to −0.68, p=0.001), irritation severity (−1.45 points, 95% CI −2.56 to −0.33, p=0.004), discharge frequency (−1.19 points, 95% CI −2.30 to −0.08, p=0.029), discharge severity (−1.15 points, 95% CI −2.18 to −0.11, p=0.021). Artificial tears significantly reduced irritation frequency (−1.28 points, 95% CI −2.37 to −0.19, p=0.013). FML showed no statistically significant group-level effects. Individual response variation is shown in figure 3.

Figure 2. Intervention effects on outcomes compared with no medication control group are shown as mean reduction differences (with 95% CI). FML, fluorometholone.

Figure 3. Distribution of individual-level differences of discharge and irritation scores for each treatment compared to no medication. FML, fluorometholone.

Secondary outcomes

No specific treatment appeared to provide a greater benefit for participants with lagophthalmos. However, participants without DASS showed a greater response to FML drops, particularly in reducing irritation severity and frequency. Additionally, vitamin A ointment appeared to be beneficial in reducing irritation frequency, while artificial tears demonstrated efficacy in alleviating discharge severity.

Discussion

Chronic mucopurulent discharge and discomfort in anophthalmic patients with ocular prostheses is a clinical challenge, often requiring tailored therapeutic approaches. This randomised crossover trial evaluated the efficacy of three commonly used interventions in managing these complications. Our findings provide a very much needed insight into the effectiveness of these treatments and emphasise the importance of individualised management strategies.

Interpretation of primary outcomes

Vitamin A ointment emerged as the most effective treatment in this study, with significant improvements across all four primary outcome measures: irritation frequency, irritation severity, discharge frequency and discharge severity. Its performance stands out compared with both artificial tears and low-dose steroid drops, which showed more limited or inconsistent effects.

A key difference between the interventions lies in their formulation. The ointment has a substantially higher viscosity than the water-based drops, allowing it to remain in the socket longer and provide more sustained lubrication at the interface between the prosthesis and the conjunctiva. This prolonged contact can reduce mechanical shear forces and may create a more stable socket environment in patients whose symptoms are largely driven by friction.

However, formulation characteristics alone may be unlikely to account fully for the observed benefits. Vitamin A is known to support epithelial health and repair, offering a pharmacological effect that complements its lubricating properties.¹⁶ As the trial did not include a bland ointment without vitamin A, it is not possible to distinguish precisely how much of the treatment effect is attributable to viscosity and how much to the vitamin A component. A non-vitamin A ointment with similar rheological properties could theoretically improve symptoms through mechanical protection, but whether such a preparation would match the overall effect of vitamin A ointment remains uncertain and would need to be addressed in a dedicated comparative study.

In contrast, artificial tears showed limited overall efficacy, despite reducing irritation frequency in some patients. Their primary function as lubricants addresses surface dryness but fails to target the underlying inflammatory or epithelial repair mechanisms necessary for managing chronic discharge.¹⁷ Additionally, the need for frequent application (four times daily) may hinder patient compliance, further limiting their effectiveness.

Surprisingly, analysis of the cohort as a whole showed no statistically significant improvement with FML drops, despite their known anti-inflammatory properties. This finding may indicate that inflammatory processes within the anophthalmic socket are relatively unresponsive to low-dose corticosteroid therapy, that the two times per day dosing regimen was inadequate to achieve sustained symptom control, or that mechanical factors such as friction play a more prominent role than inflammation in symptom generation for a substantial proportion of patients.

Notably, FML Liquifilm was the only intervention containing a preservative. In contrast, both artificial tears (Hylan) and vitamin A ointment (Hylo-Night) were preservative-free, whereas FML contains benzalkonium chloride (BAK) as its standard preservative. BAK is known to exert cytotoxic effects on the ocular surface and to promote the expression of inflammatory mediators, including matrix metalloproteinase-9 (MMP-9). Elevated MMP-9 levels have been demonstrated in anophthalmic sockets and are associated with increased conjunctival inflammation.¹⁸

In this setting, repeated exposure to BAK may have attenuated the potential anti-inflammatory effects of low-dose FML by sustaining epithelial stress and MMP-9 activity, thereby contributing to its limited efficacy compared with preservative-free formulations. Although this represents a plausible explanation, the present study was not designed to isolate the impact of preservatives, and evaluation of a preservative-free corticosteroid formulation would be of interest in future research.

Further studies are also warranted to determine whether higher corticosteroid doses or alternative anti-inflammatory therapies may yield improved clinical outcomes.

While vitamin A ointment was the most effective overall statistically, a small subset experienced worsening of symptoms with its use. Conversely, some patients who did not benefit from vitamin A ointment showed improvement with FML drops or artificial tears. There was no uniform response within each medication group, as seen in figure 3. This variability highlights the need for a personalised approach, where treatment choices are guided by individual patient characteristics and responses, further emphasising the complexity of managing anophthalmic socket symptoms and the importance of tailoring interventions to individual needs. Nevertheless, a notable finding was that 72.5% of participants exhibited a clinically relevant response of >2 point reduction to at least one treatment in at least one outcome parameter. 82.5% showed improvement when measured with a threshold of >1 point reduction. Although mean reductions were below the prespecified clinically relevant threshold, the majority of participants experienced meaningful improvements with at least one treatment, highlighting that benefit is concentrated in individual responders rather than across the entire cohort.

Secondary outcomes and subgroup analysis

Descriptive analysis of secondary outcomes revealed no significant differences in treatment response between participants with and without lagophthalmos or DASS. For example, vitamin A ointment was equally effective in both lagophthalmos and non-lagophthalmos groups, suggesting that eyelid closure abnormalities do not significantly influence its therapeutic impact, which is in line with earlier findings of lagophthalmos not impacting the frequency of irritation or discharge complaints.¹⁹ Similarly, while participants without DASS appeared to derive slightly greater benefit from FML drops, these observations were not statistically tested and should be interpreted cautiously. Overall, the findings suggest that predetermining treatment based on lagophthalmos or DASS is not advantageous.

Regular follow-up is essential to monitor treatment efficacy and adjust regimens as needed. Clinicians should also prioritise patient education on proper prosthesis care and the importance of adherence to prescribed therapies, as these factors are critical for optimising outcomes and improving quality of life.²⁰

Limitations

Several limitations should be acknowledged. First, this study focused solely on three medications, excluding other long-term therapeutic options such as gel-based formulations or ocular sprays. These alternatives may offer benefits in terms of user convenience and adherence; for example, gels provide prolonged lubrication compared with artificial tears, while sprays enable easier application with less dosing precision required.

Additionally, the relatively small sample size (n=40) may limit generalisability. The unmasked no-medication control and self-reported outcomes represent limitations. Expectation or behavioural changes during the identifiable control period may have inflated active treatment effects, despite the crossover design mitigating interindividual bias. Although the crossover design enhances efficiency, potential carryover effects remain a concern despite washout periods. Future studies could incorporate objective measures to strengthen validity.

Conclusion

In conclusion, this study is the first of its kind to investigate and compare the available treatments for managing chronic mucopurulent discharge and discomfort in anophthalmic patients with ocular prostheses. It demonstrates that vitamin A ointment is the most effective intervention and should be considered as the first-line treatment due to its broad efficacy and dual mechanism of action. Yet, some individuals may experience better effects from either artificial tears or low-dose steroids. Artificial tears may be suitable for patients with mild irritation but limited discharge, while FML drops could be reserved for cases where inflammation is the primary concern, despite their limited overall efficacy in this study. The findings highlight the importance of personalised treatment strategies and the limitations of current therapeutic options. By addressing these challenges, clinicians can improve symptom management, prosthesis tolerance and overall quality of life for this vulnerable patient population. Future research should focus on refining treatment protocols, exploring novel therapies and developing evidence-based guidelines to standardise care.

Data availability statement

Data are available upon reasonable request.