Editorials

Treatment of raised intraocular pressure and prevention of glaucoma

Evidence at last that treatment works

Two important randomised controlled trialsone from the United States, the other from Swedenwere published last year in the Archives of Ophthalmology, and their findings were a cause for celebration for ophthalmologists and subspecialists in glaucoma. ^{1 2} Intraocular pressure has traditionally been lowered pharmacologically or surgically in an attempt to prevent the disease destroying sight long before randomised controlled trials were conceived. The rationale was based on indirect evidence. However persuasive this might have been, it did not protect against lingering doubts caused by observing patients progress relentlessly towards blindness despite apparently successful control of intraocular pressure or the fact that a substantial proportion of people with glaucoma have pressure that is always within the normal range. Some even proposed that raised pressure was effect not causea failure of autoregulation because of interruption of biofeedback.

These doubts hindered advocates of population screening because evidence of effectiveness of treatment, a fundamental requirement, was not there. Eddy, in examining the economics of population screening in the United States, was one of the first to draw our attention to these deficiencies.³ For a while we were locked into an ethical dilemma. Evidence was lacking that lowering pressure was effective, yet it was considered unethical to withhold treatment from a control group. The ethical imperative to produce the evidence gained ground after Rossetti's systematic review of the effectiveness of the medical treatment of chronic open angle glaucoma.⁴ Out of 114 randomised controlled trials in the review, only eight attempted to assess effectiveness with a vision related outcome against placebo or control. Only three studies provided data on visual field and were included in a meta-analysis. This did not show a beneficial effect of lowering pressure on visual function.

Several comparative studies indicated an effect. The Moorfields laser medicine surgery trial indicated that best control was achieved by surgery, both for pressure and visual function, as did Jay's Glasgow trial. ^{5 6} Later, a more complex trial comparing different strategies for treatment of advanced disease, the advanced glaucoma intervention trial from the United States, could show that the better the control of pressure, the less likely it was that the visual field would deteriorate.⁷ But this was not from randomised comparisons, and similar evidence was available from observational studies.

One way round the ethical obstacle was to look at patients with normal pressure glaucoma because many of these were not treated anyway. The collaborative normal tension glaucoma trial did show that a percentage reduction in pressure achieved by medicine with or without surgery reduced the risk of progression in the loss of visual field, but this emerged only after controlling for the effects of cataract, which were more incident in the treatment group.⁸

The other ethical loophole was to tackle ocular hypertensionraised pressure but no evidence of optic nerve damage. An adequately powered multicentre trial was organised in the United States, the first results of which were reported in June 2002the ocular hypertension treatment study.¹

Patients were randomised to a percentage reduction in pressure with medicines or no intervention. This design prevented double masking (a term preferred to "blinding" in trials about ocular disorders) but outcome measures were assessed objectively by masked observers. A 20% reduction in pressure achieved a 5.1% absolute reduction in risk of progression to overt disease (from 9.5% to 4.4%) where this was defined as an incident loss of visual field and or demonstrable excavation of the optic nerve head (hazard ratio 0.4, 95% confidence interval 0.27 to 0.59). The effect was less but still significant if a stricter definition of onset of disease was used with demonstrable functional change onlyan absolute risk reduction of 2.4% and a number needed to treat of 42.

The question still remained whether treating pressure in overt disease prevents progression. The answer is provided by the second trial published in October 2002the early manifest glaucoma trial.² A population based survey in southern Sweden of more than 40 000 identified 255 people in the early stages of glaucoma who agreed to be randomised to treatment with laser and topical  blockers or nothing. Visual fields and pressure were monitored closely. Once progression occurred, appropriate treatment was promptly instituted. Progression was carefully specified and constituted subtle changes in the peripheral vision not normally subjectively noticeable. Safety limits for pressure were also set.

The trial report is exemplary and adheres closely to CONSORT requirements. Treatment reduced the risk of progression by 17% (7% to 23%, P=0.004, number needed to treat 6). Topical  blockers for glaucoma have been shown to be potentially harmful. ^{9 10} Two participants in the treatment arm developed asthma in the early manifest glaucoma trial, although respiratory symptoms occurred equally in either arm in the ocular hypertension treatment study. A new finding from the early manifest glaucoma trial is that the risk of nuclear cataract is clearly increased in the treatment arm. This could be the effect of topical  blockers, other additives such as preservatives, or laser trabeculoplasty. A similar finding from an observational study in Barbados has also been reported.¹¹ At least it is now possible to weigh up both the benefits and risks of treatment.

In both studies the participants were aware whether they have been treated or not. ^{1 2} This remains an issue despite the objectivity of the outcome measures. It is possible that a placebo effect is important in glaucoma. Being a control is potentially stressful. In the ocular hypertension treatment study, adverse effects were reported consistently more often in the control arm. Stress may have a (as yet undetermined) role in glaucoma. Circulating corticosteroids may influence pressure and circadian rhythms. A further trial, still under way, on pressure reduction in ocular hypertension in Europeansthe European glaucoma prevention trialis double masked and will be an important addition to the body of evidence.¹²

R Wormald, consultant ophthalmologist. 

Research and Development Department, Moorfields Eye Hospital, London EC1V 2PD (r.wormald.ac.uk)

Footnotes

Competing interests: RW has received honoraria for speaking at meetings and support for attending international meetings symposia from various pharmaceuticals companies, including Alcon, Allergan, Mercke, and Pharmacia.