Article Text

Download PDFPDF

The North Jutland County Diabetic Retinopathy Study (NCDRS) 2. Non-ophthalmic parameters and clinically significant macular oedema
  1. L L Knudsen1,
  2. H-H Lervang2,
  3. S Lundbye-Christensen3,
  4. A Gorst-Rasmussen3
  1. 1
    Department of Ophthalmology, Aarhus University Hospital, Aalborg Sygehus Syd, Aalborg, Denmark
  2. 2
    Department of Medical Endocrinology, Aarhus University Hospital, Aalborg Sygehus Syd, Aalborg, Denmark
  3. 3
    Department of Mathematical Sciences, Aalborg University, Aalborg Øst, Denmark
  1. L L Knudsen, Department of Ophthalmology, Aarhus University Hospital, Aalborg Sygehus Syd, Hobrovej 18–22, DK-9100 Aalborg, Denmark; u19204{at}aas.nja.dk

Abstract

Background: The influence of non-ophthalmic parameters on the prevalence of clinically significant macular oedema has not been unambiguously established. The present study was initiated with the aim of clarification.

Methods: This cross-sectional study comprised 656 type 1 and 328 type 2 diabetic subjects undergoing retinopathy screening in the county of North Jutland. The association between the presence of clinically significant macular oedema and blood pressure, HbA1c, BMI, age, onset of diabetes, duration of diabetes, blood-pressure-reducing medication, lipid-lowering medication, neuropathy and urinary albumin excretion was explored using multiple logistic regression analysis.

Results: We found no significant association between the presence of clinically significant macular oedema and any of the examined parameters in type 1 diabetic subjects. In type 2 diabetic subjects, the duration of diabetes, HbA1c, neuropathy and increased urinary albumin excretion was significantly associated with the presence of clinically significant macular oedema.

Conclusions: The risk factors for clinically significant macular oedema differ in type 1 and type 2 diabetic subjects and can account only in part for this manifestation.

View Full Text

Statistics from Altmetric.com

Diabetic maculopathy is a leading cause of visual acuity reduction and blindness.13 The understanding of possible risk factors is therefore of interest, and a number of large-scale studies have explored the subject. Some studies suggest an association with metabolic regulation and other non-ophthalmic parameters4 5 while others disprove these results.6 7 The risk factors and causes of the diabetic maculopathy therefore still remain unclear.8

Recently, a large-scale cross-sectional study from the county of North Jutland explored the prevalence of proliferative retinopathy and clinically significant macular oedema (CSMO).9 The prevalence of proliferative retinopathy was reported to be reduced to less than a tenth of that of previous studies, possibly as a result of improved blood glucose regulation.9 Still, the prevalence of CSMO was reported to be relatively high and possibly increased, although the regulation of blood glucose had improved.9 The present study was initiated to explore the influence from non-ophthalmic risk factors on the prevalence of CSMO in the present population.

MATERIAL AND METHODS

In the period 1 April 2000 to 30 April 2004, 656 subjects with type 1 diabetes and 328 subjects with type 2 diabetes underwent diabetic retinopathy screening in the county of North Jutland. The type 1 diabetic subjects were almost exclusively from larger Aalborg (an urban area in the County of North Jutland), representing 70–75% of all adult type 1 diabetic subjects in this region. The type 2 diabetic subjects were enrolled from the entire County of North Jutland mainly due to poor regulation of diabetes. These individuals accounted for less than 5% of registered type 2 diabetic subjects in the County. The over-riding participants were Caucasians.

Diabetic retinopathy screening

The method has previously been described9 and is briefly summarised as follows. First, a standardised visual acuity was measured using a decimal progression scale. Second, the retina was photographically recorded using a digital camera (Zeiss DSC 420 resolution 1524×1012). One photo was centred at the macular region, and the other included the optic disc and the nasal part of the retina. Simultaneously, a number of selected non-ophthalmic parameters, as indicated below, were registered. Third, the digitised retinal recordings and registered non-ophthalmic parameters were electronically transferred to the Department of Ophthalmology for additional evaluation. Fourth, the retinal recordings were examined on a high-resolution screen (Nokia 446 PRO) for lesions in the macular region. In cases of any detectable pathology in the macular region, subjects were called for a clinical examination (22% of all subjects) to determine the presence of CSMO. If a subject failed to appear, they were summoned again, resulting in a 100% participation.

Definition of diabetes type

In the study, we defined type 1 and type 2 diabetes as follows:

  • type 1 diabetes: diabetic subjects less than 30 years of age at diagnosis, usually normal or underweighted at diagnosis or with a history of keto-acidosis

  • type 2 diabetes: diabetic subjects aged above 30 years at diagnosis, normally overweight at diagnosis and without a history of keto-acidosis

Non-ophthalmic parameters

The recorded non-ophthalmic parameters and their methods of measurement are described below:

  • regulation, blood pressure and DM status: HbA1c (%), systolic blood pressure (mm Hg), diastolic blood pressure (mm Hg), mean arterial blood pressure (mm Hg), neuropathy (±) and u-alb excretion (normal/micro-alb/proteinuri)

  • person characteristics: age (years), height (m), weight (kg), body mass index (BMI) (kg/m2)

  • characteristics of diabetes: duration of diabetes (years), age of DM-start (years)

  • medication: insulin (±), oral antidiabetics (±), diet only (±), blood-pressure-reducing medication (±), lipid-lowering medication (±); median and inter-quartile range for continous parameters are displayed in table 1.

Table 1 Crude and corrected odds ratios (OR) among type 1 and type 2 diabetic subjects for age, duration of diabetes, blood pressure, BMI and HbA1c

Definition of CSMO

The presence of CSMO was established from a clinical three-dimensional evaluation of the macular region using the ETDRS criteria.10

Data analysis

Data were composed from several sources. Typing errors and mismatch of cases were controlled for by visual inspection of scatter plots and and by validation of selected cases. Multiple logistic regressions were used for the calculation of prevalence, odds ratios and confidence intervals adjusted for various parameters. To avoid possible bias, the statistical analysis only included clinically significant macular oedema on the right eye. The statistical analysis was carried out using SPSS 12.0.2 for Windows and R R.2.1.1.11

RESULTS

The crude prevalence of CSMO was 9.6 (7.9 to 11.6)% for all diabetic subjects, 7.9 (6.1 to 10.3)% for type 1 and 12.8 (9.6 to 16.9)% type 2 diabetic subjects.

Among type 1 diabetic subjects, the presence of CSMO was not significantly associated with any of the examined parameters. An additional subdivision into prepuberty (<15 years) and postpuberty (⩾15 years) onset revealed no differences between these two groups.

Among type 2 diabetic subjects, the presence of CSMO was significantly associated with the duration of diabetes (p = 0.035; adjusted (for age, duration of diabetes, blood pressure, BMI and HbA1c) OR = 1.05, HbA1c (p = 0.036; adjusted OR = 1.26) and neuropathy (p = 0.047; adjusted OR = 2.60). Additional sub-analyses among patients with registered nephropathy revealed that microalbuminuria influenced the prevalence of CSMO insignificantly (p = 0.92; adjusted OR = 1.06), while proteinuria influenced it significantly (p = 0.004; adjusted OR = 5.18). Systolic blood pressure (p = 0.063; adjusted OR = 1.02) and blood-pressure-reducing medication (p = 0.068; adjusted OR = 2.59) were found to be close to the 5% confidence limit.

DISCUSSION

The presence of CSMO was not found to be associated with any of the examined parameters among type 1 diabetic subjects. In type 2 diabetic subjects, it was associated with the duration of diabetes, HbA1c, neuropathy and proteinuria. The study thus suggests differences between type 1 and type 2 diabetic subjects with respect to the presence and risk factors for CSMO. However, a selection bias among the type 2 diabetic subjects might be present. The study also suggests that the present risk factors account for only a minor fraction of subjects with CSMO.

The metabolic control of diabetic subjects has generally been improved to date, and these subjects also comprise the present population.9 Still, the prevalence of CSMO in the present study was found to be relatively high and possibly increased compared with previous studies. A multiple regression analysis of the present data revealed no significant influence of HbA1c in type 1 diabetic subjects and a significant but modest increased risk (26%) among type 2 diabetic subjects. The blood pressure did not influence the presence of CSMO significantly in any group of diabetic subjects. Consequently, the influence of regulatory parameters on the presence of CSMO in the present population was limited and could only account for a small number of these cases. However, it should be noted that the present population was relatively well regulated. Less well-regulated diabetic subjects could still have a significant increased risk for CSMO.

The potential advantages of a very tight regulation have previously been discussed. The present study results suggest that such additional improved regulation will have little additional benefit on the presence of CSMO. The identification of additional risk factors should have a high priority in future studies.

In the light of the present study, the causes of the development of CSMO therefore still remain unclear, but at least two possibilities seem plausible: previous glycaemic malregulation or genetic factors.

It is well documented that the function of various organic systems is influenced by previous malregulation, which makes it obvious to suspect such an association. It is also well known that the thickness and the composition of basement membranes in several organic systems differ in diabetic subjects. These changes in the basement membranes are, in general, the result of the surrounding cells and their function, thus suggesting some genetic influence. In the future, we will focus on previous regulation and possible genetic risk factors.

Table 2 Median, interquartile range and max/min values among type 1 and type 2 diabetic subjects for age, duration of diabetes, blood pressure, HbA1c and BMI
Table 3 Crude and corrected odds ratios (OR) among type 1 and type 2 diabetic subjects for blood-pressure-reducing medication, lipid-reducing medication neuropathy and nephropathy

REFERENCES

View Abstract

Footnotes

  • Competing interests: None declared.

  • Abbreviations:
    CSMO

    clinically significant macular oedema

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.