Go straight to some highlights on this page:

Overall light decrease nixed because it protects controls

Authors maximize irradiation with most dangerous light

Authors maximize difference between study groups

Early protection allegedly to difficult and costly

Babies remain unprotected for first 24 hours
> Page  3-4 >  06-07-95

3.3 LIGHT REDUCTION-METHODS
3.3.1 RATIONALE
FOR METHOD OF PARTIAL LIGHT OCCLUSION
Several alternatives were considered as a method of reducing the amount of light incident on the preterm eye. These included draping the isolettes similarly to Glass and co-workers; a head box; decreasing room light at its sources, i.e., ceiling, exam, and window lighting; patient goggles; and total occlusion, i.e., patching.

Draping the isolette yielded far too little light reduction and was not deemed reliable. Head boxes were impractical. Decreasing source light was highly variable and not easily controlled and very dependent on isolette stationing within a unit. It also couldn't separate controls on a practical basis. Both goggles used over the eyes or total occlusion were practical, efficient, reliable and easily separated control and treatment groups.

Partial light reduction via goggles was deemed preferable to total occlusion. Total occlusion presents the theoretic risk of impairing early visual development in the lateral geniculate bodies and occipital cortex, conceivably producing a bilateral amblyopia. Despite the absence of any evidence that total occlusion could produce amblyopia at such an early point in a premature infant's life, selecting goggles which allow form vision perception will further minimize the risk of amblyopia. This is covered more fully in section 3.5.2.

Finally, it was elected to apply goggles to the treatment group only. Using goggles with no filter on the controls was not effective in masking since the clear vs. dark goggles are easily seen. And even clear goggles reduce ambient light incident on the eyes, thus diminishing our control-treatment separation, especially in the crucial ultra violet range.

[Note: The authors are thus making every effort to achieve maximum irradiation of the babies in the most dangerous wavelength region, to better see any damage this exposure might cause to their eyes.]

> Page 3-5 > 06-07-95

3.3.2 RATIONALE FOR FILTER DENSITY

The magnitude of goggle light reduction was based on a compromise between maximizing light reduction and minimizing form vision deprivation. Neutral density (ND) filters are quantified in log units and percent transmission of light. These correspond as follows:

The visual impact of various filters was assessed using Snellen recognition acuity in 18 healthy adults. Acuity determination was done in a standard 20 foot lane with normal room illumination with best correction in place. All patients had 20/20 acuity without filters.

A neutral density filter of 1.0 screening 90% of ambient light had no impact on Snellen acuity. A filter of 1.3 screening 95% had little or no effect, with all subjects either 20/20 or 20/25. A filter of 1.5 log units screening 97% of ambient light reduced Snellen acuity to 20/25 or 20/30. Higher filters began to have more effect (Figure 1).

Although the applicability of adult standards to preterm infants is tenuous, there is no doubt that neutral density filters permit resolution of formed images. An infant's perception of these images is subject to considerable interpretation. Available evidence on evaluation of infant visual perception is discussed in section 3.5.2.

Once a quantitative sense of the degree of form deprivation produced by neutral density filter strength was achieved, a level of filtration based on lighting concerns was evaluated. Ambient NICU lighting levels vary a great deal. A typical nursery can vary from 500 lux to over 2000 lux in daytime conditions. Imposing neutral density filters reduces the magnitude of the entire range proportionately as follows:

> page 3-6 > 06 07 95

In order to achieve a balance between maximizing the degree of light reduction and minimizing the degree of form deprivation, a 1.5 log unit ND filter was designated as optimal based on the effects of neutral density filters on normal adult visual acuity.  Because of fragility and expense considerations, Kodak Wratten neutral density filters were not chosen for use in the study goggles. A more durable and less expensive polycarbonate material was chosen. The transmission characteristics of this "near neutral density" filter are shown in Appendix R.3. Snellen acuity through this chosen lens was 20/25 to 20/40 in normal adults.

3.3.3 RATIONALE OF TIMING OF GOGGLE WEAR

There have been no studies in human infants that have included an evaluation o

the postnatal age range over which preterm infants might be most susceptible to light-induced retinal injury. If prematurity is a major risk factor for this injury, it would seem reasonable that infants would be most susceptible in the first few days and weeks after birth. Ideally, we would like to maximize the likelihood of detecting a difference between the two study groups by beginning light reduction in the delivery room. This would require obtaining consent from the mother during labor and randomizing the infanl

prior to delivery. While not impossible, this approach would introduce considerable logistic and cost constraints into the study. In order to balance the goal of early entry with the practical constraints of a large multicenter randomized trial, we will require that

> page 3-7 > 06-07-95 randomization take place in the first 24 hours of life so that light reduction may begin by 24 hours of age.

The CRYO-ROP Study has produced a wealth of information on the natural history of retinopathy of prematurity in a study population of 4099 infants weighing less than 1251 grams. The utilization of this data allowed proper structuring of the proposed light reduction protocol. Table 3-1 is a summary of CRYO-ROP statistics regarding the incidence of any ROP, prethreshold, and threshold ROP by birth weight. ^5-9

Threshold ROP represents severe ROP requiring surgical cryotherapy. Prethreshold ROP is a degree of ROP which precedes threshold and represents the first significant danger point from which permanent structural changes in the retina may result leading to some degree of visual loss . For babies under 1000 grams, these risks are substantial with more than 26% of infants developing prethreshold ROP and more than 9% of infants developing threshold disease. More than 15% of those babies at greatest risk (birth weight less than 750 grams) reached threshold and hence would require cryosurgery. The benefit of a cost effective preventive therapy is obvious.

Continue reading