If you have diabetes, you have almost certainly heard about the annual dilated eye exam. It is one of the few pieces of preventive-care advice that turns out, on inspection, to be both well-evidenced and worth following exactly as written. The reason is that diabetic retinopathy — the most common cause of working-age blindness in many countries — is largely silent in its early stages. People rarely notice it themselves. The dilated exam, sometimes alongside optical coherence tomography (OCT) and fundus photography, is what catches it.
There is a second, smaller observation worth knowing. Visual function can change in diabetes before any visible retinopathy appears on a dilated exam. Contrast sensitivity is one of the functional measurements that picks this up. It is not a substitute for the dilated exam — those exams look for structural change (microaneurysms, dot-and-blot hemorrhages, exudates, edema, new vessels) that contrast sensitivity cannot see. But contrast sensitivity is a real, repeatable, on-your-own-screen measurement that can sometimes register early functional change. It is reasonable to use it as a between-exams self-tracking adjunct.
The rest of this post is about why that signal exists, what the literature shows, and the specific limits of what an at-home contrast sensitivity test can and cannot tell you if you are diabetic. The advice at the end is short and unambiguous: keep doing the dilated exams.
What diabetes does to the retina, briefly
Diabetes is a systemic disease — its consequences ripple far beyond blood sugar. Chronic hyperglycemia damages the small blood vessels throughout the body, and the retinal capillaries are particularly vulnerable. Over years to decades, that damage produces the classic vascular picture of diabetic retinopathy: capillary closure and leakage, microaneurysms, hemorrhages, ischemia, and — in advanced stages — neovascularization (new, fragile vessels that can bleed) and tractional retinal detachment.
Two other distinctions matter for what follows. Diabetic macular edema (DME) is fluid leakage into the macula, the central retina responsible for fine vision. DME can occur at any stage of retinopathy and is one of the most common reasons diabetic patients lose central vision. It is also one of the things a dilated exam combined with OCT detects most reliably. Proliferative diabetic retinopathy (PDR) is the late stage, defined by new vessels on the disc or elsewhere; it is what most aggressively threatens sight. Most diabetic vision loss is preventable when DME and PDR are caught early — which is why the annual dilated exam exists.
A piece that is easier to miss: the neural retina starts to suffer before the vessels look bad on an exam. Electroretinography, microperimetry, and contrast sensitivity studies have all picked up functional change in diabetics whose dilated exams looked normal or showed only minimal retinopathy. Reviewers in the early-detection literature frame this as a "neurovascular" picture rather than a purely vascular one: neural-retinal dysfunction can precede ophthalmoscopically visible vascular damage (Safi, Safi, Hafezi-Moghadam & Ahmadieh, 2018). Contrast sensitivity sits in that "functional change you can measure before the structural change is visible" category.
This is not a reason to skip the dilated exam. It is the opposite — it is a reason the exam catches things sooner when paired with functional information, and a reason contrast sensitivity is interesting at all in diabetes.
What the literature shows on CSF in diabetes
Contrast sensitivity in diabetes has been studied for forty years. A few honest summaries of what that work supports.
Contrast sensitivity is reduced in diabetes, on average, before retinopathy is visible. One of the earliest careful studies — Sokol, Moskowitz, Skarf, Evans, Molitch and Senior (1985), in the Archives of Ophthalmology — measured contrast thresholds across six spatial frequencies (0.5 to ~22.8 cycles per degree) in 64 patients with insulin-dependent and non-insulin-dependent diabetes. Patients with no visible retinopathy in the non-insulin-dependent group already showed abnormal contrast sensitivity at the highest spatial frequency tested. Patients with even minimal background retinopathy showed abnormalities across all frequencies. Visual acuity was, in the same patients, normal. The finding has been replicated and extended many times since.
The effect is detectable on letter-based contrast tests, not only on grating tests. Stavrou and Wood (2003), in Clinical & Experimental Optometry, compared letter contrast sensitivity and visual acuity in type 2 diabetic patients (with and without retinopathy or macular edema) against age-matched controls. Letter contrast sensitivity discriminated diabetics with minimal or no retinopathy from controls. Visual acuity did not. Their conclusion mirrors Sokol et al.: contrast sensitivity catches functional change earlier than the acuity chart.
Larger cohort work confirms the pattern and links it to systemic measures. The SN-DREAMS-II cohort (Gella, Raman, Pal et al., 2017, Eye) measured contrast sensitivity in 653 diabetic patients from the Sankara Nethralaya Diabetic Retinopathy Epidemiology and Molecular Genetics Study. CS was correlated with the duration of diabetes, severity of any retinopathy, severity of any cataract, refractive error, hemoglobin level, vibration perception threshold (a peripheral-neuropathy measure), and best-corrected visual acuity. CS in diabetes is shaped by the systemic disease — not only by what is visible on the retina.
Two honest caveats. First, the pattern across spatial frequencies in diabetes is more broad reduction than the textbook glaucoma "mid-frequency loss" picture. Some studies report mid-frequency emphasis; some report low- and high-frequency emphasis. The mechanism is multi-layered (neural-retinal dysfunction plus emerging microvascular change plus, for many older diabetics, some lens involvement), and the pattern reflects that. Second, no contrast sensitivity test — clinical or online — has been validated as a stand-alone screening test for diabetic retinopathy against a clinical-diagnosis reference standard. The literature supports CS as a complement, picking up functional change that the structural exam may not yet show. The annual dilated fundus exam remains the standard for staging and for catching the structural findings that prompt treatment.
This is also where contrast sensitivity in diabetes differs structurally from contrast sensitivity in glaucoma. Glaucoma is, in a sense, a single-pathway story: retinal ganglion cells and the magnocellular projection. Diabetes is a systemic disease whose retinal effects play out over many years and many tissues. The CSF reduction in diabetes is one output of a system; HbA1c, blood pressure, lipid panel, duration of disease, and the dilated exam findings are some of its inputs. Tracking one without the others doesn't tell you very much.
What an online CSF test can usefully do for someone with diabetes
The honest framing: take a baseline; retake at intervals; bring meaningful changes into the conversation with your eye doctor and your endocrinologist. That is the use case.
A few practical specifics.
Take a baseline early. If you have just been diagnosed or have not measured before, doing the test now gives you a number that future tests can be compared against. Type 1 and type 2 diabetes have somewhat different age-of-onset and duration-to-retinopathy profiles, but the baseline argument applies the same way to both: a number from when retinopathy was not yet a feature is a more useful comparison point than the population average.
Retake on a regular cadence — at minimum yearly, alongside your dilated exam. A reasonable rhythm for many diabetic patients is annually under stable conditions, more frequently when HbA1c control has changed substantially (in either direction), or when you have new vision concerns. Use the same device, similar lighting, and similar viewing distance each time, so the change you see reflects vision and not setup.
If your CS is dropping over multiple sessions, bring the result to your dilated exam. Two or three below-baseline sessions on the same device, in similar conditions, is more meaningful than any single number. The natural conversation to have with the eye doctor is whether your current OCT and fundus findings still match the pattern they would expect — not whether contrast sensitivity is "the" diagnostic. The contrast result is one signal; the structural exam is the answer.
HbA1c is the systemic input the team optimises. CSF is one output among several. Tighter control reduces but does not eliminate the risk of progression to retinopathy. Many patients with well-controlled diabetes still develop some retinopathy over time, particularly with longer disease duration. So a stable contrast sensitivity result alongside a well-controlled HbA1c is reassuring, but neither is sufficient on its own to skip the dilated exam.
Cross-link: the underlying primer on what contrast sensitivity is and why it matters lives in What contrast sensitivity actually measures; the calibration and adaptive-procedure details for the test we built live on the methodology page.
What it cannot tell you
Note. A contrast sensitivity test is a screening signal of overall visual function. It does not diagnose diabetic retinopathy — or stage it, or rule it out — and it cannot replace the annual dilated fundus exam.
A normal contrast sensitivity result does not mean your retina is free of diabetic change. Early retinopathy is often asymptomatic and can be present despite a normal contrast result. Diabetic macular edema in particular requires OCT or careful funduscopy to detect; a screening test will not show it.
A reduced contrast sensitivity result does not mean you have diabetic retinopathy. The list of other things that lower CS is long: refractive error (the most common cause by a wide margin — wear your current prescription when testing), cataract (very common in long-standing diabetes), dry eye, glaucoma, post-concussion vision change, normal aging, fatigue, medications, low room lighting, and test-setup issues. Several of these are themselves more common in diabetic populations, which is part of why CS reductions are non-specific in diabetes.
A contrast sensitivity test cannot stage retinopathy. The clinical staging framework — mild, moderate, severe non-proliferative; proliferative; with or without diabetic macular edema — comes from the structural exam, OCT, and (where indicated) fluorescein angiography or OCT angiography. CSF gives a single functional axis; staging needs the picture.
A single result is a snapshot, not a trend. Test-retest variability is real even with well-validated clinical instruments — the Pelli-Robson chart, for instance, has a published test-retest repeatability of about ±0.15 log units and a smallest-clinically-meaningful change of around ±0.30 log units (Mäntyjärvi & Laitinen, 2001). One below-typical result on one afternoon is much weaker evidence than a sustained shift across multiple sessions on the same setup.
The shape this leaves you with: useful tracking signal, real biological basis, modest specificity. That is what makes it an honest screening adjunct rather than a stand-alone test.
Practical next steps
What to do, depending on where you land:
If you have diabetes (type 1 or type 2): keep your annual dilated eye exam. That advice is older than this website and will outlast it. The recommended interval may be more frequent than annual depending on disease duration, retinopathy stage, pregnancy, or recent rapid changes in HbA1c — your ophthalmologist or optometrist sets that. The exam catches microaneurysms, hemorrhages, edema, and neovascularization that no functional test can see. Functional testing complements it; it does not replace it.
Take a contrast sensitivity baseline and retake yearly. Save the result. The point is the trend on your own setup, not any single absolute number. If your HbA1c control changes substantially, or if you notice new vision symptoms, an extra retest is sensible.
If your CS starts dropping across multiple sessions: first, check the easy things. Is your refraction current? Have you been more fatigued? Is the room lighting different? Are you testing on a different screen? Have you developed dry eye? If a sustained drop persists across two or three sessions on the same device in similar conditions, mention it at your next eye-care visit. A more frequent dilated exam may be warranted; that conversation belongs in the clinic, not in a self-test app.
Keep an eye on the systemic numbers. HbA1c, blood pressure, lipid panel, kidney function — these are the inputs that determine whether the retina sees more damage over time. Good systemic control reduces but does not eliminate retinopathy risk. The exam is still annual.
A briefer way to say all of that: the dilated exam is the gold standard. Contrast self-tracking is a useful adjunct between exams. Both, not one.
Take the test
Take the test now. Save the result. Retake annually — or after any sustained change in HbA1c — and bring the result to your next dilated exam. If your contrast is dropping but the eye exam is clean, the answer is not "stop worrying" and not "panic"; it is "discuss the trend with your eye doctor, and keep watching."
Diabetic retinopathy is a slow disease that responds well to early intervention. The single most useful thing you can do for your eyes if you are diabetic is to keep showing up for the dilated exam. Self-tracked contrast sensitivity is, at best, one extra thread the eye doctor's appointment can be built around. The thread is not the rope.
References
- Sokol, S., Moskowitz, A., Skarf, B., Evans, R., Molitch, M., & Senior, B. (1985). Contrast sensitivity in diabetics with and without background retinopathy. Archives of Ophthalmology, 103(1), 51–54. Early careful study showing contrast sensitivity reductions across multiple spatial frequencies in diabetic patients with minimal or no visible retinopathy, despite normal visual acuity.
- Stavrou, E. P., & Wood, J. M. (2003). Letter contrast sensitivity changes in early diabetic retinopathy. Clinical & Experimental Optometry, 86(3), 152–156. Letter-based contrast sensitivity discriminated type 2 diabetic patients with minimal or no retinopathy from age-matched controls, while visual acuity did not.
- Gella, L., Raman, R., Pal, S. S., Ganesan, S., & Sharma, T. (2017). Contrast sensitivity and its determinants in people with diabetes: SN-DREAMS-II, Report No 6. Eye, 31(5), 743–750. Large-cohort analysis (n = 653) showing contrast sensitivity in diabetes is associated with disease duration, retinopathy severity, cataract severity, refractive error, and systemic measures including peripheral neuropathy.
- Safi, H., Safi, S., Hafezi-Moghadam, A., & Ahmadieh, H. (2018). Early detection of diabetic retinopathy. Survey of Ophthalmology, 63(5), 601–608. Review of functional and structural early-detection methods, including contrast sensitivity, color vision, microperimetry, multifocal ERG, and OCT-based measures; supports the neurovascular framing of preclinical diabetic retinopathy.
- Mäntyjärvi, M., & Laitinen, T. (2001). Normal values for the Pelli-Robson contrast sensitivity test. Journal of Cataract and Refractive Surgery, 27(2), 261–266. Source of the age-stratified normative Pelli-Robson values and the published test-retest repeatability figures referenced above.