  |
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Journal of the New Zealand Medical Association,
22-August-2003, Vol 116 No 1180
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Screening for diabetic retinopathy using the mobile retinal
camera: the Waikato experience
Elham Reda, Peter Dunn, Clive Straker, David Worsley, Keith
Gross, Isla Trapski and Susan Whitcombe
In 1993 we reported the introduction of a mobile retinal
photoscreening programme for the detection of diabetic
retinopathy.1 In this report we document our
experience with the programme from August 1993 to December 2001.
Diabetic retinopathy is the leading cause of blindness and
visual impairment in adults in developed
societies.2,3 Laser photocoagulation therapy
has been demonstrated to be effective in the treatment of proliferative
retinopathy and diabetic maculopathy,4,5,6 and
if implemented in time is able to reduce visual impairment and
blindness.7 However, significant retinopathy is
commonly asymptomatic in its early stages,4,8
and therefore regular, scheduled eye screening is
recommended.6 Retinal photoscreening has been
accepted as a cost-effective and adequate methodology for this
purpose.9,10
The population of the Waikato District Health Board (DHB)
(312 918 in the 1996 Census, at the approximate midpoint of the programme
period) constitutes nearly 9% of the total New Zealand population. By 2010, the
total population is projected to increase to 338 150. One fifth of the Waikato
DHB population (62 584 people) identify themselves as Maori. Forty four per cent
reside in rural areas, compared with the New Zealand average of 23%. The
absolute number of subjects with diabetes in the Waikato DHB region is not
known, but applying age- and ethnicity-specific diabetes prevalence
figures11 gives an estimate of 10 326 for the
1996 Census. Prevalence is highest among the Maori, estimated to be 12% of the
total Maori population, compared with 10% of the Pacific Islands population and
only 6% of all other ethnic groups. Access to ophthalmologists is difficult due
to distance and the large number of diabetic subjects in comparison with the
restricted number of ophthalmologists. These factors dictated the need for a
mobile retinal photoscreening programme for the area.
The prevalence of diabetic retinopathy has been reported in
a number of populations,9,12 but published data
on the prevalence of diabetic eye disease in New Zealand is
sparse.1,13,14 All diabetic individuals are
prone to develop eye complications, with particular subgroups at higher risk.
The prevalence of diabetic retinopathy is strongly related to the duration of
diabetes. After 20 years, nearly all patients with Type 1 diabetes and over 60 %
of patients with Type 2 diabetes have some degree of
retinopathy.15 Up to 21% of Type 2 diabetics
are reported to have a degree of retinopathy at the time of diagnosis. Regular
screening for the detection of retinopathy is recommended in all diabetic
subjects. Patients over 10 years of age with Type 1 diabetes should have an
initial eye examination by an ophthalmologist or optometrist within five years
of the onset of diabetes. Patients with Type 2 diabetes should have an initial
dilated eye examination by an ophthalmologist shortly after the diagnosis of the
diabetes is made. Subsequent examinations for both Type 1 and Type 2 diabetic
patients should be repeated annually. Examinations will be required more
frequently if the retinopathy is
progressing.16
MethodsThe methodology for this study
has been reported previously.1 All patients
with diabetes mellitus, excluding those with significant eye disease and under
the continuing care of an ophthalmologist, were invited to join the programme.
Referrals were mostly obtained from general practitioners, members of the
Waikato Regional Diabetes Team, the Waikato Hospital ophthalmology clinic, and
ophthalmologists. A standard retinal camera was used in conjunction with
mydriasis throughout the programme, initially a Kowa FX 50 R, currently a Topcon
TRC 50X. All photographs were taken by a qualified specialist medical
photographer. The camera was transported to the screening location in the back
of a station wagon by the diabetes service’s specialist medical
photographer. The cameras used were not designed for transport and a
pneumatically cushioned transport frame was specially designed by local
engineers to enable the cameras to be transported safely in a light motor
vehicle. Screening took place at 28 venues throughout the Waikato, including
community centres, marae, medical centres, hospital outpatient clinics and a
local prison. Specialist diabetes educators within the service arranged
screening days for their local area in collaboration with the local general
practices. Twenty to sixty patients were booked per screening day.
All patients were registered on a database maintained within the Waikato Regional Diabetes Service centre. The database was used to schedule appointments, record the results of the photoscreens, and report the results and actions recommended by the reviewing ophthalmologist to the patients and their general practitioners. At photoscreening, the patient’s pupils were dilated with 1% tropicamide and 2.5% phenylephrine eye drops administered by the specialist diabetes educator. Two 50° arc colour retinal photographs were taken of each eye, one centred on the macula with the optic nerve head at the periphery of the view, and one centred nasally to the optic nerve head again with the optic nerve head at the periphery of the view. A third photograph, focused at the lens plane, was taken if there was evidence of cataract. Patients were provided with written instructions at the time of making their appointment to bring a driver to the appointment, not to drive on the day of pupillary dilatation, and to contact their general practitioner in the event of any concerns following the procedure. All slides were identified with the patient’s name and the date of photoscreening. Developed films with work sheets were presented to the ophthalmologist for grading. Each image was assigned one of five categories: NDR = no significant diabetic retinopathy; NVR = non-vision-threatening diabetic retinopathy; VTR = vision-threatening retinopathy; OP = other pathology; FNA = films not assessable. In addition, the ophthalmologist could add a free-text comment. The ophthalmologist selected one of four action categories: re-screen in one year; re-screen in two years; re-screen in three years; or refer to the Waikato Eye Clinic. Patients who did not attend were recorded in the database from September 1999 onwards. These patients were then re-booked or the general practitioner requested to re-refer. Prior to September 1999 patients were commonly rescheduled by the local diabetes nurse specialist but there was no formal process for doing so. ResultsFrom 5 August 1993 to 7 December
2001, 8172 patients were screened for retinopathy one or more times. A total of
15 555 screening episodes were recorded, and 15 334 films were assessable. Of
these, 9.8% of patients were rescheduled for screening at one year, 74.9% at two
years, and 0.4% at three years. A further 14.7% were referred on to the eye
clinic for specialist assessment. Failure-to-attend rates were high throughout
the study: European 13.1%, Maori 32.2%, Asian 23.9%, Indian 26.9%, Pacific
Islands 34%, with an overall rate of 18.7% of 7188 scheduled
appointments.
Numbers of first and subsequent photoscreens per calendar
year are shown in Table 1. Currently, approximately 1000 new patients are being
enrolled every year.
Table 1. Photoscreens per calender year of
programme
Table 2. Numbers of subjects, gender and mean age per
ethnic group
The demographic characteristics and ethnicity of the group
are listed in Table 2. Males and females are equally represented. Europeans and
Maori are the largest subgroups numerically, and the Europeans are older. The
group is further categorised according to ethnicity and number of photoscreens
per patient in Table 3. Patients were screened up to seven times in the course
of the programme.
Table 3. Distribution of number of screens per patient
in each ethnic group
As seen in Table 4 the overall prevalence of NVR and VTR are
9.3% (n = 1448) and 3.1% (n = 474) respectively. Maori patients have a higher
prevalence of VTR (4.3%) than Europeans (2.5%). The prevalence figures for each
retinopathy category per calendar year of the programme are shown in Table 5.
The highest prevalence of VTR was recorded in the first two years, with a marked
fall over the course of the programme.
Table 4. Distribution of retinopathy findings based on
ethnicity
FNA = films not assessable; NDR = no significant
diabetic retinopathy; NVR = non-vision-threatening diabetic retinopathy; OP =
other pathology; VTR = vision-threatening retinopathy
Table 5. Percentage distribution of findings per
calender year of programme
FNA = films not assessable; NDR = no significant
diabetic retinopathy; NVR = non-vision-threatening diabetic retinopathy; OP =
other pathology; VTR = vision-threatening retinopathy
To provide an estimate of the incidence of VTR during the
period of the study, patients were split into those with NVR on first
photoscreening (643) (Table 6), and those with NDR on first photoscreening
(6376) (Table 7). There were 13 618 patient years of follow up in the initial
NVR group, and 888 patient years of follow up in the initial NDR group. The
incidence of VTR in the initial NDR group equalled 76 unique patient events in
13618 patient years (5.58 events per 1000 patient yrs). The incidence of VTR in
initial NVR group equalled 49 unique patient events in 888 patient years (55
events per 1000 patient yrs).
Table 6. Photoscreen results on patients with
non-vision-threatening retinopathy (NVR) at initial review
FNA = films not assessable; NDR = no significant
diabetic retinopathy; NVR = non-vision-threatening diabetic retinopathy; OP =
other pathology; VTR = vision-threatening retinopathy
Table 7. Photoscreen results on patients with no
significant diabetic retinopathy (NDR) at initial review
FNA = films not assessable; NDR = no significant
diabetic retinopathy; NVR = non-vision-threatening diabetic retinopathy; OP =
other pathology; VTR = vision-threatening retinopathy
DiscussionIt has been estimated that the
total number of people with diabetes in New Zealand will increase from 147 300
in the year 2000 to 180 400 by the year 2010.17
The explosive growth of the diabetic population demands greater efficiencies in
the management of our patients with potential or actual vision-threatening
conditions. A screening programme should aim to detect patients at risk when
they can still be effectively treated, and this can be achieved by regularly
checking the patients’ eyes. The 1988 service planning guidelines for
diabetes recommended regular eye review for all patients with diabetes. This was
not occurring due to the fact that New Zealand doesn’t have enough
ophthalmologists to meet the extra demand that comprehensive systematic review
would put on the service. Our mobile retinal photoscreening programme reduced
the burden of normal eye screening on our ophthalmology service and allowed
community-based capture of retinal images, ensuring coverage of those subjects
who do not regularly attend hospital clinics. Images were then assessed
centrally by specialist ophthalmologists.
Various methods of screening have been shown to be
sufficiently sensitive and specific for the detection of sight-threatening eye
disease (STED)12,18 at justifiable
costs.10,19 However, the sensitivity and
specificity of a screening programme are not the only important considerations.
The extent of the population coverage and the screening intervals are vital to
the success of a screening programme. The mobile service makes the screening
more accessible to a larger number of patients. Our mobile retinal
photoscreening programme has achieved good ascertainment of the estimated
diabetic population for at least one photo screen. Approximately 79% of the
estimated 1996 diabetic population in the Waikato region were screened one or
more times.
The percentage of the non-assessable images was 6.2%. The
British Diabetic Association recommends a maximum failure rate of 5% for any
screening programme to be acceptable.20 It is
important to differentiate between technical reasons for poor photography and
physical causes such as cataract. The photograph should be repeated if possible
if the reason is technical. If there is any other eye disease the patient should
be referred to an ophthalmologist. Overall, the number of non-assessable images
has been acceptable.
We found that 12 132 eyes (78.0%) didn’t show any
retinal lesion and, therefore, these patients were rescheduled for repeat
photoscreening in two years, thus reducing the waiting list for a specialist
ophthalmologist examination. The percentage referred for specialist review was
14.7%.
The true prevalence of sight-threatening diabetic
retinopathy (STDR) is unknown, but current evidence suggests that 40–45%
of the diabetic population have some diabetic retinopathy (DR) with 10–14%
having STDR.12 Other investigators recently
found a baseline prevalence of any retinopathy, proliferative diabetic
retinopathy (PDR), and sight-threatening eye disease (STED) of 45.7%, 3.7% and
16.4% respectively, in Type 1 diabetes.21 In
Type 2 diabetes, the baseline prevalence of any retinopathy, PDR and STED was
25.3%, 0.5% and 6.0% respectively. A large study of DR has been conducted from
1977 in Newcastle. Over 11 years, 5519 diabetic subjects were assessed. Thirty
five per cent had some evidence of retinopathy and in 11.4% VTR was
found.22
Our data provide an estimate of the baseline prevalence of
grades of diabetic retinopathy and VTR in the Waikato diabetic population, at
the point of acquisition into a community-based retinopathy screening programme.
The prevalence of any DR was 9.3% (1448 of 15 555) and of VTR was 3.1% (474 of
15 555). Maori patients showed a higher prevalence of VTR (4.3%). The prevalence
of VTR has fallen since inception of the programme, as incipient problems were
detected in the initial cycles of screening. The reduction in prevalence of
retinopathy in our programme indicates that the prevalence of undetected and
untreated significant retinopathy in the community is falling.
Our results also suggest that the risk of subsequent VTR is
tenfold higher in the NVR than the NDR group. This is in agreement with the
previous epidemiological studies describing the onset and progression of
diabetic retinopathy. The WESDR (Wisconsin Epidemiologic Study of Diabetic
Retinopathy) established that progression of retinopathy was a function of
baseline retinopathy.23,24 The more severe the
baseline retinopathy, the greater the frequency of progression to
vision-threatening retinopathy. Conversely, among Type 2 diabetic patients whose
baseline photographs showed no retinopathy, there was less PDR or progression to
severe macular oedema over four years. Furthermore, although some guidelines
suggest annual screening is required,16,23,24
screening every two years has proved clinically satisfactory for monitoring NDR
patients in our series.
It should be noted that, in the Waikato management
programme, patients with significant retinopathy are followed in specialist
ophthalmology clinics and do not appear on the photoscreening register once they
have been referred. We do not know precisely how many patients are in the
specialist clinic programme at the moment, but expect this to be a significant
number. We intend to identify these patients to enable us to assess the
percentage of expected diabetic patients in the region receiving either
specialist review or photoscreening review.
Several cross-sectional studies show a higher prevalence of
microalbuminuria, proteinuria and end-stage renal failure in Maori and Pacific
Island patients with Type 2 diabetes compared with
Europeans.25 The reasons for these ethnic
differences in diabetic nephropathy have not been defined. Similar issues can be
raised in relation to diabetic retinopathy. Our results demonstrated a higher
prevalence of VTR in Maori (4.3%) than Europeans (2.5%). This is in agreement
with a household survey of known diabetics in South Auckland, which indicated
that diabetic eye disease had resulted in blindness in at least one eye in 7% of
Maori, 8% of Pacific Island patients, but only 2% of
Europeans.14 Among patients with Type 2
diabetes attending the Wellington Diabetes Clinic, Pacific Islands people were
more likely to have any retinopathy by direct ophthalmoscopy (40%) than
Europeans (23%) or Maori (16%).13 It remains to
be determined how much of this apparent increase in susceptibility is due to
modifiable metabolic factors, which may in turn be related to socioeconomic
status and lack of access to culturally appropriate medical care, and how much
is due to genetic factors.
In the future, the use of digital imaging systems and
development of automated assessment systems should improve the efficiency and
effectiveness of the management of patients with diabetic
retinopathy26 and reduce the frequency of
results that cannot be assessed.
In conclusion, the Waikato Regional Mobile Diabetic
Retinopathy Photoscreening Programme has shown that mobile retinal photography
is a practical and effective method of screening for diabetic eye disease in
rural areas.
Author information:
Elham Reda, Endocrinology Registrar; Peter Dunn, Endocrinologist and Director of
Diabetes and Endocrinology, Department of Diabetes and Endocrinology; Clive
Straker, Ophthalmologist; David Worsley, Ophthalmologist, Department of
Ophthalmology; Isla Trapski, Medical Photographer; Susan Whitecombe, Diabetes
Team Leader, Waikato Hospital, Hamilton; Keith Gross, Ophthalmologist,
Department of Ophthalmology, Rotorua Hospital, Rotorua
Acknowledgements: We
thank the Waikato Regional Diabetes Service staff for their enthusiastic
contribution to the programme.
Correspondence: Dr
Peter Dunn, Endocrinologist, Waikato Diabetes Service, Waikato Hospital, Private
Bag 3200, Hamilton. Fax: (07) 839 8811; email: dunnp@waikatodhb.govt.nz
References:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| | Current
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads) Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals |
|