| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
From the Glaxo Department of Ophthalmic Epidemiology, Moorfields Eye Hospital, London, United Kingdom.
Abstract
PURPOSE. To identify the risk factors for having advanced glaucomatous visual field loss on the first visit at three hospital eye services.
METHODS. This was a hospital-based, case–control study involving patients newly diagnosed with glaucoma at first visit to one of three ophthalmic departments in the United Kingdom. Patients with a previous history of ocular hypertension or any documented suspicion of glaucoma (within the hospital eye service) were excluded.
RESULTS. Occupational group, initial intraocular pressure (IOP), family history of glaucoma, method of referral to hospital, and the number of years since the last visit to an optometrist were found to be independently associated with late presentation. A linear trend of increasing odds of late attendance was associated with increasing Standard Occupational Classification. Those in managerial (category II) and skilled (category III) groups estimated (95% confidence intervals) to be, respectively, 0.2 (0.00, 0.16) and 0.27 (0.1, 0.8) as likely to attend with advanced glaucomatous field loss as unskilled (category V) people with similar initial IOP, family history, referral route, and time since last optometrist visit. The data strongly suggest an association between IOP and advanced field loss at initial hospital examination. There was a 1.2 (1.12, 1.28) increase in the OR of late presentation per unit increase in millimeters of mercury after adjustment for the other mentioned factors. People with a family history of glaucoma were estimated to be almost one third (adjusted OR, 0.29 [0.12, 0.74]) as likely to have advanced field loss as those with no family history. People referred by any source other than an optometrist who has made the correct diagnosis of glaucoma were 4.5 times (adjusted OR, 4.53 [1.52, 13.48]) more likely to be late attenders than patients so referred but similar in other mentioned factors. These data also provide strong evidence that the more years since the last visit to an optometrist, the greater the likelihood of having advanced glaucomatous visual field loss on the first visit to the eye service (adjusted OR per year, 1.25 [1.10, 1.42]).
CONCLUSIONS. These data strongly suggest that certain subgroups of people with glaucoma were at greater risk of having advanced and irremediable field loss on first visiting the eye services studied.
Glaucoma is a disease with major public health implications. It is known to afflict approximately 2% of white people aged 40 or more and at occurs in at least four times that many African Americans and African Caribbeans.1 Populations are aging, and the increase in the prevalence of glaucoma with increasing age has been well documented.1 2
Despite new medical and surgical strategies to control intraocular pressure (IOP), blindness registration from glaucoma continues to increase, and glaucoma remains the second or third most common cause of blindness in the world.3 4
Why do people become blind from glaucoma? Grant and Burke5 have suggested that there are two reasons: 1) People seek medical attention late in the course of the disease, when the rate of visual field loss is increased, and there is less field left to lose. 2) People seek help early but receive suboptimal treatment.
Published work on the first of these reasons—late presentation—is scarce. This study was conducted to examine this and to attempt to identify the factors that put a person at risk of having advanced glaucomatous visual field loss on the first visit to the hospital eye service (HES).
Methods
This was a hospital-based, case–control study with recruitment at three independent eye departments in the United Kingdom (see Appendix). Patients were eligible for study if they were diagnosed with glaucoma according to the criteria listed below when first examined by an ophthalmologist at one of the three sites. Any patients with a previous history of glaucoma, ocular hypertension, or suspected glaucoma (diagnosed by an ophthalmologist or optometrist) were excluded from the study. All fields were examined by one author (SF) to ensure consistency of case definition.
Eligability Criteria for Recruitment
To be considered for the study, all patients had 1) visual field
loss that was consistent with a known pattern of glaucomatous cause
(e.g., arcuate scotomas), was compatible with the patient’s disc
changes and in which there was no suggestion of other optic nerve
disease (e.g., defects that crossed the horizontal midline); 2)
Glaucoma of any chronic type: primary open-angle, pseudoexfoliative,
normal tension, chronic angle closure, aphakic, or pigment dispersion;
and 3) two consecutive fields (threshold or suprathreshold) confirming
the loss, except when field loss was so advanced that perimetry was not
possible. 4) Patients who had more than one-third loss of fixation or
false-positive or false-negative results on the initial visual field
test were not enrolled. Fields were examined almost exclusively with
the Humphrey (San Leandro, CA) 24-2 or 10-2 threshold strategies.
If the patients fit these criteria then a decision was made about whether they could be cases or controls, based on visual field and optic disc anomalies. Cases were defined as fulfilling the above criteria but having field loss within 5° of fixation and beyond 30° in one or both eyes and with a cup-to-disc ratio greater than 0.8 in the same eye. Controls were defined as fulfilling the above criteria but with the visual field having no absolute scotomas within 20° of fixation in either eye and a cup-to-disc ratio assessed as higher than 0.5 or a difference of more than 0.2 between the discs.
In an attempt to minimize the problems of recall bias, patients were recruited after the first field examination. If, during follow-up, visual fields improved to the extent that the patient no longer met the criteria, they were removed from the study. Six patients were excluded on this basis, each after their second visual field trial. The disc parameter criteria were shown, by pilot study, to be sensitive indicators of cases and controls.
Basic demographic, referral, and ophthalmic information was collected at the initial consultation, and suitable patients consented to further evaluation. These patients were then telephoned by an interviewer masked to their case–control status, who validated demographic data and asked a series of standard questions regarding socioeconomic status, general medical health, and attendance and use of sight testing (i.e., optometrist) services.
For each patient identified as eligible for the study, the following information was collected: Age when first assessed by the ophthalmologist, sex, ethnic origin (white, all those from the United Kingdom and Ireland; black, African and African Caribbean; Asian, from the Indian subcontinent); coexisting ocular or medical disease; IOP according to the standard Goldmann tonometer reading at initial examination; frequency of sight testing; and Standard Occupational Classification.
Frequency of Sight Testing
In the United Kingdom most patients with glaucoma have their
referrals to the HES initiated by optometrists in commercial practice.
Optometrists do not usually refer directly to an ophthalmologist but
through the patient’s general medical practitioner (GP) who then
formally refers the patient to the HES. It is rare for a GP to have the
expertise or equipment necessary to test for glaucoma. By review of
referral letters and telephone confirmation, two principal
referral sources were identified: 1) an optometrist who made a
diagnosis of suspected glaucoma, and 2) an optometrist who did not
mention glaucoma in the referral letter, or the GP directly with no
prompting by an optometrist’s findings (and therefore without a
diagnosis of glaucoma).
Occupational Group
The patient’s occupational group was determined using the
Standard Occupational Classification outlined in Table 1
. If the patient was retired, unemployed, or had a long-term
illness, the occupation performed for the major part of his or her
working life was entered. For married women who had never worked, their
husbands’ occupation was entered per Standard Occupational
Classification guidelines.6
In practice, those belonging
in IIIN (nonmanual skilled occupations) and IIIM (manual skilled
occupations) were difficult to distinguish using the information
collected, and a pooled category III was used.
|
Statistical Analyses
Power calculations were used to determine the sample size, and 110
cases and 110 controls were recruited to enable an 80% chance of
detecting as statistically significant at the 5% level a doubling in
the odds of late attendance in a factor present among 10% of the
controls.
The data were analyzed using a software program7 to investigate the effects of each study factor on the odds of late attendance. Estimates of the odds ratios (ORs) of late attendance with approximate 95% confidence intervals, by study factor, were computed by logistic regression. In each case, the first category (unless otherwise stated) of each categorical study factor was used as a baseline, either because its selection appeared to provide the most meaningful results or because it contained the greater number of observations, and its choice therefore favored precision. Factors significant at univariate level were entered into a multivariate model but were subsequently omitted if not statistically significant in this model. This was referred to as model I. Unadjusted and adjusted ORs are presented with adjustment made for all factors found to be statistically significant in model I.
Results
The study was undertaken over a 9-month period from September 1996 through May 1997 and was discontinued when 110 cases and 110 controls had been reviewed. None of the patients who fit the study criteria declined to take part in the study, and there was no loss of patients during the study.
Table 2 shows the characteristics of the study population. The median (interquartile range; IQR) age of controls was 63 years (IQR, 52–71 years) and of cases was 72 years (IQR, 63–78 years). Similar numbers of men and women were studied. Most (n = 171) of the patients recruited were white, 35 were of African or African Caribbean origin, and 14 were of Asian origin. The most common diagnosis was primary open angle glaucoma (n = 155) followed by normal tension glaucoma (n = 27). Fifteen patients had diagnoses of pseudoexfoliative glaucoma, and 13 of chronic angle-closure glaucoma. The median IOP in right eyes was 24 mm Hg (IQR, 22–30 mm Hg) and in left eyes was 25 mm Hg (IQR, 22–30 mm Hg). The median maximum IOP (i.e., the higher of the right and left IOPs at initial examination) was 26 mm Hg (IQR, 23–31.5 mm Hg). Most subjects had no significant ongoing medical problems. Of those who reported problems, systemic hypertension was the most common disorder. Most subjects (n = 155) had no family history of glaucoma, 31 had a single first-degree relative with glaucoma, 14 had a second-degree relative with glaucoma, and 20 had two or more first-degree relatives with glaucoma. The majority (n = 163) of patients were referred from optometrists who had made a presumptive diagnosis of glaucoma. Table 2 also shows that most subjects had visited an optometrist no more than 5 years before the referral visit. Thirty-one subjects had either never seen an optometrist or had not seen one for more than 10 years.
|
|
Maximum IOP (i.e., the greater of the right and left IOPs) at initial examination was shown by these data to be strongly associated with the odds of having advanced field loss. We estimate a 1.2 (1.11, 1.28) increase in the odds of late attendance per increased millimeter of mercury after adjustment for model I factors.
The stronger the patient’s family history, the lower the odds of late attendance. Overall, a patient with a family history of glaucoma was estimated to be almost one third (0.29 [0.12, 0.74]) as likely to attend with advanced field loss as a patient with no family history but of the same occupational group and referral source and similar IOP and time since last visit to an optometrist.
Referral source is shown by these data to be strongly associated with late attendance. A patient referred through any source other than an optometrist who had made the correct diagnosis was estimated to be 4.5 times (4.53 [1.52, 13.48]) more likely to be a late attender than a patient so referred but similar in other model I variables. These data provide strong evidence that the greater the number of years since last visiting an optometrist, the greater the likelihood of delaying attendance at the eye service. (adjusted OR per year, 1.25 [1.10, 1.42]).
Univariate analysis suggested a strong association between ethnic origin and late attendance, estimating that people of African Caribbean origin were more than six times likely to attend with advanced field loss than were white patients. Ethnic origin was not, however, statistically significant in the multivariate model. Inspection of a partial correlation matrix of all study factors revealed moderate correlation between ethnic origin and referral source and number of years since last visit to an optometrist. Much of the apparent effect of ethnicity on late attendance appeared to be caused by less frequent sight testing. However the statistical nonsignificance may well have been a consequence of relatively low power. These data provided little evidence of association between late attendance and any of the other factors studied.
Discussion
Several investigators have shown that patients who have advanced glaucoma are at a substantial risk of blindness.5 8 9 10 11 12 Grant and Burke5 found that eyes with a visual field defect at the beginning of treatment are more likely to progress to blindness than eyes in which treatment is started at the stage when there is no field loss (although whether all the patients in the second group in their study had glaucoma is difficult to ascertain). Wilson et al.8 looked at risk factors for rate of progression of glaucomatous visual field loss in 57 patients and found that visual field loss at initial examination was the strongest determinant of rate of further visual field loss. Patients in their study deteriorated 11.7 times faster in eyes with more advanced field loss. Mikelburg et al.12 measured the scotoma mass of fields and compared them with the rate of subsequent decline. They found that when scotoma mass is small (i.e., early disease), rate of visual field loss is slow, but when the scotoma mass is large, rapid linear progression of visual fields loss occurs.
Thus, there is good (if inferential) evidence that late attendance (i.e., after substantial field loss) is a significant risk factor for subsequent blindness caused by glaucoma. There is also a biologic plausibility underpinning this, in that the fewer axons the optic nerve has, the fewer it can afford to lose without significant field changes.
Another important aspect of late attendance that should be examined is the extent of the problem. Late attendance may be important to the patient, but does it occur in significant numbers?
There have been a number of studies that have estimated the proportion of patients with glaucoma who have substantial visual field loss. Grant and Burke5 calculated that one third of the patients who had become blind from glaucoma had done so before they had sought medical attention for their eyes. Elkington et al.13 reported a figure of 33% delaying medical attention, a study in the west of Ireland14 found 10% of those with glaucoma were severely visually impaired at first examination, and Sheldrick et al.15 showed 20% with severe impairment. The published evidence seems to indicate that late presentation of glaucoma is an important and not uncommon risk factor for blindness. This study was an attempt to elucidate the risk factors for late presentation.
It was suggested in our pilot study that whether a person seeks attention late in the course of glaucoma is likely to be a function of the rapidity of visual field deterioration and the frequency of sight tests.16 A person with a rapid decline could lose significant field even with two yearly sight tests, unless tested during an early but detectable phase of the disease. Conversely, a person with a slowly declining field but who does not have sight testing for some years (or does not have a glaucoma examination during the sight tests) is at risk for a different reason. The pilot study found evidence to support this, but the findings were limited by the retrospective nature of the data collection. Do the results of this present study support the results of the pilot study?
Consistent with our earlier pilot study, these data provide strong evidence (P < 0.05) of an association between maximum IOP at initial examination and the odds of late attendance. There is evidence in the literature that higher IOPs lead to more rapid visual field loss (and thus increased likelihood of delayed attendance). David et al.,17 Hart et al.,18 and Armaly et al.19 all found more rapid visual field loss at higher levels of IOP. Jay and Murdoch20 calculated that for pressures of 21 to 25 mm Hg, untreated disease was likely to progress from early field changes to end stage in approximately 14 years, whereas for pressures higher than 30 mm Hg the interval to end-stage disease was as short as 3 years.
Related to this are the results from the category of glaucoma diagnosis. Tables 2 and 3 show that patients with pseudoexfoliative and chronic angle-closure glaucoma were in fact all categorized as cases (producing infinite odds ratios). Because these conditions are usually associated with higher IOPs than primary open-angle glaucoma, this is a further indication that those with higher IOPs are more likely to attend late.
Our results support the concept that those expected to have a rapid visual field decline (i.e., those with higher IOPs at the first visit) are at risk of having advanced visual field loss at initial examination in the eye service. Are the results also consistent with the alternative reason for patients’ attending late—that of infrequent sight testing?
In the United Kingdom, the optometrist plays a pivotal role in glaucoma detection. One study showed that 90% of patients with glaucoma are referred to the hospital on the basis of abnormal findings by an optometrist.15 Our results estimate that a patient who had not been correctly referred to the hospital by an optometrist was 4.5 times more likely to attend late than a comparable patient who had been properly referred. This result is very similar to that found in the pilot study.16
Referral source appears to be an important factor in early attendance, with patients referred from optometrists with a diagnosis of glaucoma more likely to be in the earlier stages of the disease. This suggests one of two things: that those who attend late are treated by optometrists who do not test for glaucoma, or more probably, they are people who tend not to have regular sight tests.
The evidence for the latter interpretation is the finding that the risk of late attendance is proportional to the number of years since the patient last visited an optometrist (adjusted OR per year since last visit, 1.25 [1.10,1.42]. Thus, those who attend for regular sight tests seem more likely to seek medical attention in the early stages of glaucoma. Although this may seem self-evident, there is little published evidence to support the protective effect of regular sight testing. These results lend weight to the concept that those who do not have regular sight tests are at greater risk of late attendance. However, the results contain a number of potential confounding factors.
Socioeconomic status (as measured by occupational categories) was shown to be strongly associated with the risk of late attendance, with those of highest socioeconomic status estimated to be at lowest risk of late attendance. Corroborating evidence for this comes from The General Household Survey: Analysis of Ophthalmic Data 1990–91 to 1993–94, which was published in July 1995.21 This was a nationwide survey of private households in the United Kingdom, with a sample size of between 23,000 and 24,000 people per year. In the years studied, questions were included on use of spectacles, contact lenses, and attendance for sight testing. The survey data indicate that the highest percentage of those who attended for regular sight testing were in occupational categories I and II with the lowest in categories IV and V, consistent with our figures, which indicate a lower risk of late attendance in these groups.
It has long been known that socioeconomic factors influence the access to adequate medical care as well as patient compliance with treatment,22 and this extends to screening services. Loehrer et al.23 found that late attendance in various cancers was directly related to lower socioeconomic status. They thought that this was caused by both risk-promoting lifestyles and beliefs about cancer based on incomplete or erroneous information. Other studies have confirmed the link between later presentation and social deprivation in those with breast,24 colorectal,25 and skin26 cancers. It has also been shown that lower socioeconomic class is associated with poor use of screening services in mammography and cervical smears.27
As far as eye disease is concerned, Smith et al.,28 examined the relationship between social deprivation and age of attendance for amblyopia. They found that the more deprived children are much more likely to attend late with anisometropic amblyopia. The Baltimore Eye Study examined socioeconomic status and visual impairment among urban Americans29 and found that lower status is associated with higher rates of visual impairment.
The finding in the present study that higher socioeconomic status was associated with a reduced risk of late attendance is largely, but not completely, explained by differences in sight testing rates. Even when sight testing is adjusted for, as in Table 3 , the difference in socioeconomic status remains, indicating there may well be other differences between the groups.
Family history of glaucoma was also found to be an important determinant of disease stage at which medical attention is sought. Those who reported a family history of glaucoma were one third as likely to attend late as those who did not, and the more members of the family with glaucoma, the more likely the patient was to attend early. The reason for this result may be twofold: First, those with a positive family history are more likely to appreciate the need for regular sight testing because of information passed on from other members of the family. Second, in the United Kingdom, all first-degree relatives of patients with glaucoma are encouraged to have regular sight tests by provision of free sight testing by the National Health Service.
People of African Caribbean origin were found, in the univariate model, to have a much greater risk of late attendance than whites, consistent with results of the pilot study. However, this risk was not statistically significant after adjustment. As described in the results section, this appears to be related to a lower attendance for sight testing in the African Caribbean subjects.
Although there is no evidence that there was any statistical difference between the IOPs of African Caribbeans and whites, this does not rule out a faster rate of visual field loss. Lower socioeconomic status, which is an important confounding factor, shows some correlation with ethnic origin and is probably a further reason for a lower rate of sight testing (and therefore late attendance).
Wilensky et al.30 suggest that African Americans, for socioeconomic reasons, do not seek medical care until glaucoma is more advanced, although their report presents no evidence to support this. There is, however, evidence that, compared with age-matched white patients, black patients attend later in the course of certain diseases, including prostate cancer,31 32 breast cancer,33 and colorectal cancer.34 Similarly, there is evidence that African American populations use health-screening services less than white patients.35 Wells and Horm33 showed that the racial differences of disease stage at initial examination disappeared if socioeconomic class was considered. Similarly, the Baltimore Eye Study identified an association between race and blindness that was reduced, but not eliminated, after adjustment for socioeconomic factors.29 The results from the present study seem to indicate that poor use of sight testing is an important risk factor for African Caribbean patients attending the HES with advanced glaucomatous field loss.
Our pilot study16 found a strong, linear relationship of increasing age and increasing risk of late attendance. This is unsurprising, because prevalence and incidence of glaucoma increase with age, and the General Household Survey indicated a reduction in the likelihood of sight testing in those aged more than 65 years in the United Kingdom. Although not statistically significant, the results of the present study show a general trend of increasing risk of late attendance with age, with or without adjustment.
The pilot study also indicated that men are more likely to attend late than women. Although there was a small protective effect for women in the present study, it was lost after adjustment and, again, is not statistically significant.
Patients who had a history of concurrent medical problems (the assumption being that they had an increased involvement with medical services and thus were more likely to be tested for glaucoma) were not found to seek medical attention during earlier stages of glaucoma than those who did not report any problems.
There are a number of potential biases in this study. Recall bias could be a problem in collecting data on the previous sight test. This information is related to the optometrist visit before the visit that resulted in referral (if the patient was referred by an optometrist), which may have been some time beforehand. This bias was reduced, however, because the subjects recruited were unaware of their own case–control status.
It is important to note that the data regarding IOP are one-off measurements on first examination in the HES and, with later regression to the mean, the individual IOPs could be different in subsequent examinations, although this is less likely to affect the overall median IOP. The nature of this study required that patients be recruited at initial examination, which means that prolonged follow-up of the patients has not yet occurred. Therefore, the findings regarding the normal-tension glaucoma group (Table 3) should be interpreted with caution, because some of the patients may have had conversion to primary open-angle glaucoma after longer follow-up or phasing.
Using three different sites in the study to recruit patients is unlikely to have introduced any bias into the results, because the patients were chosen with standard criteria listed earlier. Using three geographically distinct centers (and using equal numbers of cases and controls from each center) reduced the probability that the results were from one unrepresentative center.
In conclusion, published evidence indicates that late attendance for glaucoma appears to be associated with risk of subsequent blindness. In this study the risk factors for late attendance itself were higher IOP, infrequent sight testing, and lower socioeconomic status. Conversely, a positive family history was likely to be associated with earlier attendance.
Appendix 1
The current system for glaucoma detection in the United Kingdom relies almost exclusively on optometrists in private practice. A patient requiring a sight test attends an optometrist who, if glaucoma is suspected, refers the patient to a GP with the intention of onward referral to an ophthalmologist in the HES. Although an optometrist cannot refer a patient directly to the HES, the GP may. It is very rare for the GP to have the expertise or equipment to examine for glaucoma, which is why most referrals originate with the optometrist.
Participating centers: Moorfields Eye Hospital, London; Sunderland Eye Hospital; Harold Wood Hospital, Essex, United Kingdom.
Acknowledgements
The authors thank Rod Daniel, Peter Phelan, and Charles Claoué for their help in recruiting the patients.
Footnotes
Supported by grants from the International Glaucoma Association and Moorfields Locally Organised Research Scheme.
Submitted for publication September 29, 1998; revised March 18, 1999; accepted March 29, 1999.
Proprietary interest category: N.
Corresponding author: Scott Fraser, Glaxo Department of Ophthalmic Epidemiology, Moorfields Eye Hospital, City Road, London EC1V 2PD, UK.
References
This article has been cited by other articles:
|
|
 |
|
  R Shah, D F Edgar, P G Spry, R A Harper, A Kotecha, S Rughani, and B J W Evans Glaucoma detection: the content of optometric eye examinations for a presbyopic patient of African racial descent Br. J. Ophthalmol., April 1, 2009; 93(4): 492 - 496. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A Azuara-Blanco, J Burr, R Thomas, G Maclennan, and S McPherson The accuracy of accredited glaucoma optometrists in the diagnosis and treatment recommendation for glaucoma Br. J. Ophthalmol., December 1, 2007; 91(12): 1639 - 1643. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S-M Saw, G Gazzard, D Friedman, P J Foster, J G Devereux, M L Wong, and S Seah Awareness of glaucoma, and health beliefs of patients suffering primary acute angle closure Br. J. Ophthalmol., April 1, 2003; 87(4): 446 - 449. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Fraser, C. Bunce, R. Wormald, and E. Brunner Deprivation and late presentation of glaucoma: case-control study BMJ, March 17, 2001; 322(7287): 639 - 643. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
