Article Text

Download PDFPDF

Original article
Impact of eyelid closure on the intraocular pressure lowering effect of prostaglandins: a randomised controlled trial
  1. Eugenio A Maul,
  2. David S Friedman,
  3. Harry A Quigley,
  4. Henry D Jampel
  1. The Wilmer Eye Institute at Johns Hopkins University, Baltimore, Maryland, USA
  1. Correspondence to Professor Henry D Jampel, Maumenee B-110, 600 N. Wolfe St, Baltimore, MD 21287, USA; hjampel{at}


Background/aims To determine if eyelid closure (ELC) after topical prostaglandin instillation provides greater intraocular pressure (IOP) reduction than prostaglandin instillation without ELC.

Methods Patients receiving chronic bilateral prostaglandin monotherapy were enrolled in this study. The study intervention, ELC, was randomly assigned to one eye, while the fellow eye served as control. ELC was performed for either 1 min or 3 min. After a 1-day washout, the IOP was measured in a masked fashion at baseline, 1 h and 24 h, and at a final visit that took place 7–14 days after enrolment. All visits were scheduled during the morning, and every individual patient's visits occurred at similar times during the day. The main outcome was difference between intervention eye and control eye in IOP-lowering from baseline.

Results 51 patients meeting eligibility criteria were enrolled: 25 were randomised to ELC for 1 min and 26 to ELC for 3 min in the intervention eye. The pooled IOP-lowering difference (95% CI, p value) in intervention versus control eyes was 0.24 mm Hg (−0.5 to 0.9, p=0.50), 0.24 mm Hg (−0.7 to 1.2, p=0.61) and 0.24 mm Hg (−0.7 to 1.2, p=0.61) in the overall group, 1 min ELC subgroup and 3 min ELC subgroup, respectively. The effect of ELC did not change significantly across visits.

Conclusions ELC did not provide significant additional IOP reduction compared with no ELC in patients using chronic prostaglandin monotherapy.

  • Randomised controlled trials
  • glaucoma
  • open-angle/drug therapy
  • intraocular Pressure/drug effects
  • Ophthalmic Solutions
  • Treatment Medical
  • Glaucoma
View Full Text

Statistics from

Systemic absorption of topical intraocular pressure (IOP)-lowering medications became an increased cause of concern after the recognition of systemic side effects that could be associated with the use of topical timolol.1–5 Most of the systemic side effects are caused by absorption into the bloodstream that takes place when timolol drains through the nasolacrimal duct onto the nasal mucosa.6 Nasolacrimal occlusion (NLO) and eyelid closure (ELC) have been shown not only to decrease clearance through the nasolacrimal system (therefore systemic absorption),7 8 but also to increase ocular contact time and enhance ocular penetration of topically applied agents.8–10 Although the literature evaluating the impact of increased ocular contact time on the effect of eye drops to lower IOP is small and inconsistent, studies suggesting that better IOP control can be achieved using these techniques are widely cited.11–13

While the American Academy of Ophthalmology's Preferred Practice Pattern on Primary Open Angle Glaucoma14 recommends education about ELC and NLO for all patients using eye drops to lower IOP, several gaps in our knowledge preclude us from making optimal evidence-based recommendations. First, it is not known whether either ELC or NLO should be recommended: (1) to prevent systemic side effects; or (2) to increase the IOP-lowering effect; or (3) both. Second, there are no data, to our knowledge, to recommend ELC alone, NLO alone, or both together as effective means to enhance IOP lowering. Third, only ELC and NLO durations of 3 or 5 min have been previously studied. Perhaps a shorter period of time would be as effective. Last, no research with ELC and NLO has been conducted with prostaglandin agents, which are the most widely prescribed agents for lowering IOP at present.

We therefore designed a study to measure the IOP-lowering effect of ELC in chronic users of prostaglandin eye drops.


Participants were recruited from the practices of the glaucoma faculty at the Glaucoma Center of Excellence Wilmer Eye Institute. Eligibility criteria included age >18 years, and use of one of the three topical prostaglandins (latanoprost, travoprost or bimatoprost) and no other IOP-lowering medications for at least 1 month. Exclusion criteria included previous incisional surgery for glaucoma, previous iridotomy, punctal occlusion, an abnormal slit lamp examination (except for cataract or intraocular lens implant) and trabeculoplasty or other incisional eye surgery during the past 6 months. All visits were scheduled during the morning before 11:00, and every individual patient's visits were scheduled within a maximum of 2 h of each other across days. The same prostaglandin agent that patients were using on a regular basis was maintained throughout the study.

Participants were requested not to use their drops the day before the baseline visit, and were given a reminder telephone call to ensure this. At the baseline visit, patients were asked about the use of their prostaglandin drops the day before and their appointment was rescheduled if they had mistakenly used their drops.

Prior to randomisation, IOP was measured using a calibrated Goldmann applanation tonometer. All IOP measurements in the study were obtained by a masked operator and a recorder who read the results independently. Three consecutive measurements were obtained for the right and then for the left eye at each visit. The median of the three was used to represent the IOP.

Participants were assigned by computerised random number to one of four groups: ELC for 1 or 3 min and intervention eye right or left. The allocation sequence was concealed in sequentially numbered, opaque and sealed envelopes from the researchers (DSF, EAM, HDJ, HAQ) enrolling and assessing participants. The envelopes were opened only after the enrolled participants completed all baseline assessments including IOP measurement. The fellow eye not randomised to intervention served as a control.

Drops were instilled in the intervention eye first, after which the subject was instructed to close both eyes gently for the assigned time. Immediately afterwards, drops were instilled in the control eye, but this time the subjects were instructed not to keep their eyes closed and to blink as they normally would. Any additional interventions such as nasolacrimal occlusion were discouraged. During the baseline visit a study member endured that the drops were used according to study protocol and subjects were asked to reproduce the same eye drop application technique in the intervention and control eye through the rest of the study. The IOP was measured again 1 h after eye drop instillation and the patient was sent home and instructed not to use their prostaglandin eye drop that evening. At the conclusion of the 24 h visit the following morning, patients were instructed to resume their drops that same day at the time they routinely used their drops before enrolling in the study and continue performing ELC through the rest of the study as assigned by the protocol. A handout with specific instructions was given to every subject. A final visit was scheduled 7–14 days after baseline during which the IOP was measured again. A summary of the study visits and procedures is displayed in figure 1.

Figure 1

Summary of study visits and interventions. ELC, eyelid cloasure; IOP, intraocular pressure.

The primary outcome measure was IOP change from baseline in the intervention eye compared with the control eye. Subgroup analysis for the effect of 1 min and 3 min ELC constituted a secondary outcome.

The projected sample size was 43 pairs of eyes to detect a difference of 1.5 mm Hg in the intervention compared with control eyes with a significance of 5% and a power of 90%, assuming a correlation between paired eyes of 0.5 and a SD of 3 mm Hg. We decided to enrol 20% more patients to compensate for potential losses to follow-up and a higher than expected SD.

The IOP change comparisons were performed using a mixed model with two levels of clustering at the patient and eye level. An exchangeable correlation was assumed and the coefficients were estimated using maximum likelihood estimation. All statistical analyses and sample size calculation were performed using Stata 11.0 (2009; Stata Corp., College Station, Texas, USA).


Patients were recruited from March 2009 to January 2010. Sixty-eight eligible patients agreed to consider participation before we achieved the projected sample size. After being contacted, 53 agreed to participate and scheduled their study visits. Two patients did not come to the baseline visit and when contacted again declined to participate in the study. Twenty-five patients were randomised to ELC for 1 min and 26 to ELC for 3 min in the intervention eye. Fifty-one patients completed all the study visits within protocol. The baseline demographic characteristics of enrolled patients are displayed in table 1.

Table 1

Demographic and clinical characteristics of study participants

The baseline IOP was 18.6±4.4 and 18.8±4.4 mm Hg (mean±SD) in intervention and control eyes, respectively. The IOP-lowering effect achieved in the intervention group was on average 0.24 mm Hg less than in the control group (95% CI −0.5 to 0.9 mm Hg, p=0.50) (table 2). Visit-specific subgroup results are depicted in figure 2. The IOP-lowering effect of the topical prostaglandins was similar at all visits. The model including subgroup analysis (shown in table 3) did not demonstrate differences in the effect of 1 min or 3 min ELC. The effect of ELC was not modified by visit in either model, suggesting that the effect of intervention did not appear to increase nor decrease through visits (likelihood-ratio test for nested models, p value of 0.98 and 0.62 for the overall model presented in table 2 and for the subgroup-model presented in table 3, respectively). The trends of the difference in effect at each of the visits are displayed in figure 3.

Table 2

Linear mixed effects regression measuring the effect of eyelid closure (ELC) on intraocular pressure (IOP) change*

Figure 2

Mean intraocular pressure (IOP) change from baseline in the 1 and 3 min eyelid closure (ELC) subgroups. IOP changes were markedly similar in intervention and control groups, and were not modified by ELC duration.

Table 3

Linear mixed effects regression measuring the effect of eyelid closure (ELC) subgroups on intraocular pressure (IOP) change*

Figure 3

Net effect of eyelid closure (mean intraocular pressure (IOP) change in the intervention—control eye) for the overall group, 1 min and 3 min subgroups. The 95% CI is expected to be fully below 0 (solid line) in case of statistically significant benefit from eyelid closure, and below −1.5 (dashed line) in case of clinically significant benefit from eyelid closure.


We found that ELC did not significantly potentiate the IOP-lowering effect of topically applied prostaglandins. There was no difference between ELC of 1 and 3 min duration. Ours is the first study to evaluate the effect of ELC on the IOP-lowering effect of prostaglandins and is the only published randomised controlled trial (RCT) powered to detect a 1.5 mm Hg difference, commonly accepted as a clinically valuable effect.15

Two previous RCTs using punctual plugs in an attempt to increase the IOP-lowering effect of topical glaucoma medications failed to detect a difference in IOP lowering,16 17 but they were underpowered. The first study by Simel and Simel17 attempted to determine if bilateral inferior punctal occlusion with dissolvable collagen inserts would benefit patients with IOP >21 mm Hg under different medical regimens that included beta blockers, adrenergic and cholinergic drugs. That study could only have detected a difference of 7 mm Hg or more due to NLO. The study by Bartlett et al16 evaluated the effect of inferior silicone punctal plugs on the IOP of patients treated with timolol 0.25%. Although no effect of punctal occlusion could be detected, only the pilot data of 17 patients were reported.

Other published studies have not provided optimal evidence that interventions involving the lacrimal drainage system can potentiate the IOP-lowering effect of topical agents. Ariturk et al11 reported an additional 1.8 mm Hg IOP lowering (p<0.05) by using an inferior silicone punctal plug in one eye of 20 patients. All patients had an IOP <21 mm Hg using one or more agents that included adrenergics, cholinergics and beta blockers. However, the intervention was intentionally assigned to the eye with the higher IOP. Huang and Lee12 reported on the effectiveness of inferior punctal plugs in 19 patients medically controlled under different combinations of topical therapy. An additional IOP reduction of 1.3 mm Hg (p<0.01) was observed in the intervention group. Although baseline IOP for both eyes was similar, the intervention was not randomly assigned.

Zimmerman et al13 published a case–control study suggesting that NLO increases the effectiveness of 0.25% timolol maleate eye drops. The NLO group had an IOP reduction that was 1.8 mm  Hg greater than control eyes at 24 h, but there appeared to have been no adjustment for significant baseline differences between groups. Another study by Zimmerman et al18 failed to demonstrate a benefit from NLO in subjects treated with epinephrine or dipivefrin. While the literature suggests that NLO or ELC prevent systemic side effects and increase the ocular penetration of topical agents,7–10 the evidence in favour of better IOP control is therefore inconclusive.

The limitations of our study include the fact that we studied only one class of ocular hypotensive agents, and our results might not be applicable to other glaucoma eye drops. Prostaglandins have a long-lasting ocular hypotensive effect well beyond 24 h; it is possible that other drugs with a shorter hypotensive effect may benefit from an intervention to increase ocular contact time such as ELC instituted in the context of chronic treatment. Second, we only studied ELC, and not NLO. It is possible that using NLO or a combination of NLO and ELC would yield different results. Third, the baseline IOP of 18 mm Hg, while it is typical for persons with open angle glaucoma, reduced the absolute level of change in IOP that would pertain to those with higher IOP. Fourth, the result might have been different with more than a 1-day washout. The rationale for not using a longer washout is that a minimal washout provides a better estimate of the expected effect of instituting ELC in a chronically treated patient. The acute effect of ELC in a fully washed out patient may become irrelevant in the longer term. Fifth, we did not directly supervise eye drop instillation through the study and relied on patient compliance with instructions for the effect observed at the final visit. A final limitation was our follow-up of 1–2 weeks: perhaps longer follow-up would have resulted in a different effect. However, the observed trend in ELC effect did not increase towards the final visit.

Since ELC carries no risk, one might ask what the harm is in instructing patients to perform it? However, every extra action that we ask of our patients runs the risk of reducing the patient's adherence to a prescribed medical regimen, and poor adherence is a common reason for reduced effectiveness of medical therapy. For a patient using two or three bottles of eye drops at a time, being told that 5 min of ELC for each drop is necessary could potentially reduce his/her adherence. In this context it may be desirable to focus the efforts on achieving ELC or NLO for drugs such as beta blockers, in whom systemic side effects can be reduced.5 8

In summary, this RCT in patients under chronic prostaglandin monotherapy failed to demonstrate that ELC results in additional IOP lowering. Further study is required to determine if ELC or NLO enhances the IOP-lowering effect of other agents.


We would like to acknowledge Jiangxia Wang from the Dana Center Statistical Consulting Team, supported by grant EY01765.


View Abstract


  • Funding Study partially supported by NIH grant EY01765 to conduct statistical analyses.

  • Competing interests None to declare.

  • Patient consent Obtained.

  • Ethics approval The study followed the tenets of the Declaration of Helsinki, and was approved by the Johns Hopkins Joint Committee for Clinical Investigation (protocol number NA_00021272). The study adhered to Health Insurance Portability and Accountability Act guidelines.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.