To determine the relationship between plasma Lp(a) concentration and the ri
sk of developing diabetic retinopathy, 341 Type 1 diabetic patients underwe
nt an annual retinal fluorescein angiography and were assigned to one of 3
groups according to the stage of their diabetic retinopathy: no retinopathy
(NR), non-proliferative diabetic retinopathy (N-PDR), or proliferative dia
betic retinopathy (POR). One hundred and twenty-three Type 1 diabetic patie
nts had no retinopathy, 188 had N-PDR and 30 had PDR. The ages of the three
groups and the duration of diabetes were significantly different. Hyperten
sion, microalbuminuria and diabetic nephropathy were more frequent in FOR t
han in NR or N-PDR (p < 0.0001). Plasma HbA1c was higher in PDR than in NR
or N-PDR (p < 0.01). Type 1 patients who had been diabetic far at least 20
years included 30 NR, 108 N-PDR and 24 PDR. Type 1 diabetic patients with P
DR had microalbuminuria and macroproteinuria more frequently than other pat
ients (p < 0.0001 and 0.01, respectively). Type I diabetic patients with PD
R had the highest median plasma Lp(a) and the highest frequency of Lp(a) ab
ove SO mg/dl(p < 0.05). Multivariate analysis carried out in Type 1 diabeti
c patients with a duration of diabetes of at least 20 years showed that mic
roproteinuria, HbA1c and Lp(a) accounted significantly for 21% of variance
in retinal status. Lp(a) above 30 mg/dl was related to the risk of developi
ng FOR (OR = 8.40, p < 0.05). Lipoprotein(a) appears to be associated with
the severity of diabetic retinopathy in Type 1 diabetic patients, and parti
cular attention should be paid to those with Lp(a) above 30 mg/dl and pre-p
roliferative retinopathy.