VALID PARAMETERS FOR INVESTIGATION OF THE PUPILLARY LIGHT REFLEX IN NORMAL AND DIABETIC SUBJECTS SHOWN BY FACTOR-ANALYSIS AND PARTIAL CORRELATION

Pupillary test data of 103 normal and 119 diabetic subjects (47 IDDM, 72 NIDDM) were evaluated by factor analysis. From a total of nine pupi llary parameters three factors were extracted in the analysis. Factor 1 represents maximal pupillary area, contraction velocity at 1 s, dila tion velocity at 6 s and minimal pupillary area - static and simple dy namic parameters; factor 2 amplitude of pupillary unrest, area under t he detrended curve of pupillary unrest and period of pupillary unrest - parameters of pupillary unrest; factor 3 fusion frequency of pupilla ry response following flicker stimuli and latency time of pupillary li ght reflex - second order dynamic parameters. Factor analysis was then applied to investigate diabetic patients with a high percentage of au tonomic neuropathic participants (about 39 % had pupillary and about 3 5 % had cardiorespiratory function disorders), which revealed the same three factors as those identified in normal subjects. Furthermore, an age-related database of parameters of pupillary unrest is given. It d emonstrates that normal subjects and diabetic patients did not differ in the period of pupillary unrest (normal vs diabetic (mean +/- SEM): 1550 +/- 29 vs 1536 +/- 27 ms; 2p > 0.5). The difference in amplitude (47.8 +/- 2.8 vs 41.0 +/- 2.6 % percentile; 2p = 0.071) and area under the detrended curve of pupillary unrest (47.9 +/- 2.8 vs 40.8 +/- 2.6 % percentile, 2p = 0.062) seems to show a trend but was not significa nt. In conclusion, factor analysis revealed three different pupillary test factors. From the comparison of normal and diabetic subjects fact or 1 which accounts for the highest percentage of variance (congruent- to 43 %) and factor 3 (congruent-to 12 %) appear to be useful for inve stigating the pupillary light reflex. Factor 2 is not useful because o f the insignificant differences between the normal and diabetic group. From factor analysis and partial correlation we believe that pupillar y autonomic function in diabetic patients can be best assessed by usin g only two parameters, maximal pupillary area and latency time.