Perhaps Meulengracht Icterus Index was the first attempt to use color index as an indirect method of assessing yellowish color intensity of serum in a jaundiced baby. Though the instrument was simple to operate, the results did not accurately reflect serum bilirubin levels. Subsequently, LaMotte-Pigford Icterus Index comparator, employing the method devised by Dr. Pigford (1928), consisted of comparing a sample of serum with a series of color standards representing different index values(1).

I read with interest the recent article on correlation of plasma color index with serum bilirubin(2). The technique described, though simple and not requiring sophisticated equip-ment, has both procedural and statistical lacunae.

The important procedural lacunae include (i) For weighing 2.0 g of potassium dichromate, an accurate chemical balance is required; (ii) Were the potassium dichromate diluted solutions prepared fresh each time (a cumbersome procedure) or used over pro-longed periods of time? In the latter case, how the color change over time period was monitored? (iii) Against what background the plasma and diluted potassium dichromate solution colors were compared? (iv) If the plasma color had yellowish intensity equally in between two color indices possessed by adjacent potassium dichromate solutions, how the index was calculated? (v) What was done to approximate the density of plasma to that of potassium dichromate solution; and (vi) What pathological conditions were considered capable of affecting color index of plasma?

The range of serum bilirubin level correlated with Plasma Color Index is not mentioned. From Fig. 1, it is apparent that extrapolation of graph showing linear cor-relation between serum bilirubin and plasma color index even for lower serum bilirubin values (<8.0 mg/dl) without actual values of serum bilirubin existing in that range may not be true. Moreover, upper as well as lower limits at 95% confidence interval of serum bilirubin level corresponding to plasma color index values should be depicted (in graphical form), in addition to correlation coefficient (r = 0.97; and p <0.001) between the two. It is because the correlation may not be the same over the entire range of serum bilirubin and plasma color index studied.

In my earlier study(3), I found a linear correlation between serum bilirubin level as determined by Diazo reaction and Icterus Index readings taken from LaMotte-Pigford Icterus Index comparator (coefficient of correlation - 0.8478, p <0.001), calculated on the basis of regression line drawn after computing 277 actual data obtained simul-taneously by diazo and Icterus Index method. However, actual serum bilirubin values were much lower than the Icterus Index values. Moreover, Icterus index values are affected by a number of pathological conditions.

Therefore, I feel that plasma color index is not a very useful method even for screening the neonates with jaundice.