Article Text

Download PDFPDF

Associations of selected medications and visual function: the Beaver Dam Eye Study
  1. B E K Klein,
  2. R Klein,
  3. M D Knudtson,
  4. K E Lee,
  5. L G Danforth,
  6. J O Reinke,
  7. A M Adler
  1. Department of Ophthalmology and Visual Sciences, University of Wisconsin Medical School, Madison, WI 53726–2397, USA
  1. Correspondence to: Barbara E K Klein, MD, MPH, Department of Ophthalmology and Visual Sciences, University of Wisconsin-Madison, 610 North Walnut Street, 460 WARF, Madison, WI 53726–2397, USA; kleinb{at}


Aim: To investigate association of drug use and visual function.

Methods: A cross sectional population based study was carried out on participants in the 1993–5 examination phase of the Beaver Dam Eye Study. All drugs in current use by study participants were recorded. Performance based and self assessed visual functions were obtained at the time of the study evaluation. The main outcome measure was the relation of levels of visual functions by use of specific drugs.

Results: Many classes of drugs were associated with decreases in at least two performance based visual functions. For example, high blood pressure drugs were significantly associated with poorer best corrected visual acuity, poorer near vision, and poorer contrast sensitivity (p<0.001 for all). Patterns of association for self assessed visual functions were not as strong. However, use of glaucoma drops and benzodiazepines were associated with poorer self assessed visual functions in most circumstances cited.

Conclusions: Many commonly used medications are inversely associated with visual functions in a middle and older aged population. This may influence the ability to perform complex tasks and quality of life.

Statistics from

Normal visual function results from a complicated series of neural pathways resulting in detection, tracking, and recognition of stimuli. It can be disturbed in many ways, including factors that interfere with perceiving visual stimuli, integrating them, and interpreting them. Most commonly used measures of visual function are psychophysical so that even in the presence of intact anatomy and physiology, measured visual function may not be optimal. Drugs have been evaluated as potential risk factors associated with specific age related eye diseases.1–4 However, it is possible that drugs may alter results of performance based visual function tests and self assessed visual function without obvious evidence of ocular disease. We explored this possibility in data from the Beaver Dam Eye Study.


The Beaver Dam Eye Study is a population based study of age related ocular disorders. A private census was conducted to identify all people between the ages of 43–84 years who were residents of the city or township of Beaver Dam, WI, USA, in 1987–8.5 Of the 5925 eligible people identified, 4926 (83.1%) were examined during the baseline examination (1988–90), 225 (3.8%) died before examination, 91 (1.5%) had moved out of the area, 23 (0.4%) could not be located, 269 (4.5%) completed a questionnaire only, and 391 (6.6%) refused to participate. Non-participants were older than participants.6–9

During 1993–5, a 5 year follow up examination of the cohort was conducted using the same methodology as the first examination.10 Of the 4541 people surviving to the beginning of the 1993–5 period, 3684 (81.1%) actually participated in the follow up examination, in addition to 38 who failed to participate in the baseline examination. Of the 857 not participating, 423 (9.3%) refused to participate, 259 (5.7%) completed a questionnaire only, 171 (3.8%) died before examination, and four people could not be located. Surviving participants who were not re-examined in the follow up eye study (n=686) were older, had less education, lower income, poorer visual acuity, were more likely to have a history of cardiovascular disease, and smoked more at baseline than participants in the 5 year follow up examination.10 Procedures at the follow up were the same as at the baseline examination. Details are summarised below.

A refraction was performed for each eye and the best corrected visual acuity was measured according to an adaptation of the Early Treatment of Diabetic Retinopathy Study Protocol.11 Near vision was measured using a specified protocol with appropriate correction in place.11 Pelli-Robson letter charts were used to measure contrast sensitivity.12

During the interview, we asked a series of questions related to visual function. First, we asked about vision in general with the following possible responses: excellent, very good, good, fair, poor, or don't know.11 Thereafter, the following specific questions were asked: “When wearing glasses (if you need them), how much of the time does your vision limit you in the following areas: (1) reading small print like the telephone book or the classified ads?, (2) reading regular print like in a newspaper, magazine, recipe, menu, or numbers on the telephone?, (3) reading road signs (such as speed limit signs) or counting pins at the end of a bowling alley?, (4) recognising people or objects across the street (can you tell the features of their face)?” The participant selected a response of “none,” “little,” “some,” “most,” “all,” or “don't know” for each of the four questions. The participant was then asked “Do you drive at night?” If the response was “no,” the participant was asked “Is this because of your vision?” If the response was “yes” or “don't know,” the participant was asked “When you drive at night, about how much of the time are you limited by your vision?” The participant selected a response for this question of “none,” “little,” “some,” “most,” “all,” or “don't know.”

The questionnaire was administered at each examination. Participants were asked to bring to the examination all medications (prescription and over the counter) that they were regularly taking. The study examiner listed all medications that were brought to the examination. When participants failed to bring all their medications, a follow up phone call was made to obtain the missing information. Only data on current use at the second examination are reported because this was the first examination at which some of the visual function data were available. In addition, during the medical interview subjects were asked whether a doctor had ever told them that they had diabetes or high blood pressure, as well as some other conditions.

Ocular evaluation included measurement of pupil size with a hand held template in ambient illumination, pupillary dilatation if it was judged to be safe to do so, and standardised photographs of the lenses and retina according to protocols developed for this study. The photography and grading protocols for lens and retina for age related lesions13,14 have been previously reported. For these analyses, central cataract refers to nuclear cataract or posterior subcapsular or cortical cataract involving 25% or more of the central circle of the grading grid.

During the course of grading there was a continuous quality control procedure to monitor intragrader and intergrader variability. There was no evidence of marked or consistent change in variability during the course of either the prevalence or follow up studies or between the two study periods. Reproducibility of gradings using these systems has been published and was similar for intragrader and intergrader comparisons.15


Hypertension was defined as a systolic blood pressure of 160 mm Hg or more or a diastolic blood pressure of 95 mm Hg or more at the time of the examination, or a history of hypertension and current use of medication for hypertension.

Diabetes was defined as a previous history of diabetes treated with insulin, oral hypoglycaemic agents, or diet. Newly diagnosed diabetes was defined as no previous medical history of diabetes or glycosylated haemoglobin at the baseline examination with elevated glycosylated haemoglobin16 at this, the second visit.

In analyses of performance based visual functions, we used the following categories: visual acuity/near vision (using Snellen equivalents), good refers to vision of 20/20 or better, fair refers to vision of 20/25–20/32, and poor refers to 20/40 or worse, all in the better eye; contrast sensitivity, good refers to logarithm greater than 1.55 dB, poor refers to 1.45–1.55 dB, and bad refers to less than 1.45 dB. In analyses of the self assessed functions (as described in detail previously), participants were asked how much of the time their vision limited their abilities in specific circumstances, the responses were categorised in three groups as none, little or some, and most or all of the time. For the general self reported visual function question, responses were grouped as excellent or very good, good or fair, and poor.

Statistical analyses were performed using sas.17 Trends in proportions were tested for significance using the Mantel-Haenszel procedure.


There were approximately 671 different drugs preparations used in the population. We present data for only selected preparations based on number of users (n>50) and on the potential for relations to visual function.

The associations of the performance based and self assessed visual functions with the drugs of interest are given in Tables 1 and 2, respectively. Use of antihypertensives, antidepressants, antipsychotics, antianxiety drugs (including benzodiazepines), benzodiazepines, glaucoma drops (excluding β blockers, which were included in analyses of all β blockers which had no effects), narcotic analgesics, cardiac glycosides, oral hypoglycaemics, and insulin were associated with decrease in at least two of the performance based measures. Use of aspirin was associated with better visual acuity. Use of sedatives (barbiturate and non-barbiturate), antidepressants, antipsychotic drugs, antianxiety drugs, benzodiazepines, glaucoma drops, narcotic analgesics, cardiac glycosides, oral hypoglycaemics, and insulin were associated with poorer self assessed visual functions in at least two categories. We did not have information about the diagnoses that prompted use of most of these drugs except for antihypertensives and hypoglycaemic agents.

Table 1

Performance based visual functions in the better eye by drug use. The Beaver Dam Eye Study, 1993–5

Table 2

Self assessed visual function by drug use. The Beaver Dam Eye Study, 1993–5

Many people taking glaucoma drops were taking pilocarpine or a derivative of it (n=30). People taking glaucoma drops had smaller pupils than those not taking such medications (p<0.001 crude analysis, p=0.051 after age adjustment). There were 89 (2.6%) people not taking glaucoma drops who had pupil diameters of 2 mm, while 16 (14.7%) of those taking glaucoma drops had pupils of this diameter. Use of narcotic analgesics was not associated with pupil size.

No classes of drugs were associated with better self assessed visual functions in at least two categories. Confining these analyses to people who were free of central cataract or late stages of age related maculopathy did not alter the patterns of association of either performance based or self assessed visual functions with these drugs. Further subcategorisation of those taking antihypertensives by specific type of agent (for example, β blockers including those used topically as eye drops, thiazide or loop diuretics, etc) suggests that associations with visual functions were more consistent for loop diuretics (data not shown). We were not able to determine whether presence of diabetes or its treatment explained the associations of hypoglycaemic agents to visual functions.


We found significant associations of visual functions with several drugs or classes of drugs. Most of these are used for psychotropic disorders. One might anticipate that drugs result in poor performance on psychophysical functions. An effect of lorazepam, a benzodiazepine, on visual perception has been reported.18 McCarty et al have assessed the role of various psychotropic drugs on visual function and cataract in the Melbourne Visual Impairment Project.1 They found that phenothiazines were associated with cataract, but reported no specific association of psychotropic medications with vision. Nearly all neuroleptic (antipsychotic) drugs decrease motor activity and decrease interest in the environment.19 Thus, whether or not such drugs actually alter the physiology of vision, they may decrease visual perception or attention, resulting in poorer measured (and self reported) visual functions. Similar results may occur from antianxiety drugs.19,20

Some antidepressants have been noted to cause sedation, blurry vision, and difficulty in concentrating. These known side effects could have an effect on visual function.19,20 Opioid or narcotic analgesics may produce drowsiness and “mental clouding”,21 as may some non-narcotic analgesics. In addition, morphine may cause constriction of the pupil which may in turn decrease vision especially in conditions of decreased illumination.21 However, we did not find pupil size to be smaller in those taking narcotics in our population. Pupils were on average smaller in those taking drops (excluding β blockers) for glaucoma. It is plausible that this accounts for the association of glaucoma drops with the visual function.

Aside from these drugs, which have intended effects on the nervous systems, there are other widely used drugs prescribed for other reasons whose desired effects are mediated by the nervous system. β Blockers are widely used in our population, often for blood pressure control,2 although some participants use drops (primarily timolol) to lower the intraocular pressure. Use of β blockers is associated with fatigue, sleep disturbances, and depression.22 These may influence performance in vision testing and this may be reported as diminished self assessed function as well.

That poorer visual functions were more common in those with glaucoma is not surprising. Average pupil size in those taking drops, aside from β blockers, for glaucoma was smaller than those not taking such medications as stated above. It is not possible in our data to sort out the relative importance of the disease from its treatment in these analyses. A limitation to our analyses is that we do not have information on duration of use. Our implicit assumption was that the effects of most of the preparations were not duration related but only reflective of visual functions at the time the medications were being used. That is why our analyses were cross sectional. We analysed the relation of these medications to change in visual function 5 years later (data not shown). We found no evidence of an association.

It is possible that dose could affect the associations we examined. We have no information to evaluate this. It is possible that some medications may interact with others to cause diminished visual functions or to counteract such effects. Because of the plethora of drugs that our participants took, the number of comparisons would be excessively large and the possibility of many chance associations high. Similarly, drugs may interact synergistically or antagonistically with systemic disorders to effect visual function. Again, because of the number of comparisons, chance may yield spurious findings. Instead, we have chosen a relatively simple approach, limiting the number of drugs to those which we hypothesised would influence visual functions and evaluating each individually. In addition, we have no measures of cognitive ability or depression, nor knowledge of the diagnoses associated with the use of many of these drugs, so that we cannot evaluate the possibility of confounding as a result of these conditions.

In conclusion, we have found significant association between the use of several drugs and visual functions in a population of older adults. These associations may be important when best possible vision is needed to perform complex or possibly dangerous tasks such as operating machinery. They may also influence vision related quality of life.


This research is supported by National Institutes of Health grant EY06594 (RK, BEKK).


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.