of children operated once and 36% of children operated twice. In the studies by Costenbader [1], by Birch (1990) and in the ELISSS [57], children who had been scheduled for surgery but who had not been operated at final assess-

11 ment had better binocular vision than those who had been operated.

Spontaneous resolution of infantile strabismus has first been reported by Clarke & Noel [58]. In a study by the Pediatric Eye Disease Investigator Group [59], among 170 children with IE (age ±3 months at recruitment), of those who had had an angle of strabismus >21.8° during two examinations at least one week apart, 2.4% had an angle <4.6° at ±7 months. Among those children who had had an angle of strabismus >11.3° during two examinations at recruitment, 27% had an angle <4.6° at ±7 months.

Reduction of the angle within 5° frequently results in microstrabismus with peripheral fusion, central suppression, and a favorable appearance. Due to the peripheral fusion, the strabismus remains stable and rarely needs additional surgery, as has been found for small angles postoperatively in the study by Van de Vijver et al. [55].

11.3.6Spontaneous Reduction of the Angle

In the ELISSS, more than half of the children who were scheduled for surgery, but had not been operated at the age of 6 years, had a spontaneous reduction of the strabismus into a microstrabismus (Fig. 11.8).

There are few studies with similar longitudinal measurements of the angle of strabismus in a large group of children. In a recent study by Pediatric Eye Disease Investigator Group [60], the angle of strabismus was measured in 81 children with IE aged 6.0 ± 1.7 months (range 2.4–9.5) at baseline and at 6-week intervals for 18 weeks, using prism and alternate cover test at near (70% of the children) or a modified Krimsky at near (30%). In 20%, all four measurements were within 2.9° or less than one another. In 46%, any two of the four measurements differed by 8.5° or more.

Could we have distinguished the ELISSS children who were scheduled for surgery but, in the end, were never operated, at an early age? In other words, can the reduction of the angle be predicted and, hence, unnecessary operations be avoided in individual cases by waiting? This line of reasoning only pertains to the majority of cases where microstrabismus with peripheral fusion is the best possible result. One cannot exclude the rare possibility that an occasional child, with a pure motor cause of IE, would achieve full binocular vision with 60 arc seconds stereopsis by very early surgery.

11.3.7Predictors of Spontaneous Reduction into Microstrabismus

In the ELISSS, of all parameters assessed in the baseline examination at approximately 11 months, only the angle of strabismus at baseline predicted, to some extent, whether a child had been operated at the age of 6 years or not (Fig. 11.9). Among children with an angle equal or smaller than 13° at baseline at approximately 11 months, 34.9% had not been operated at the age of 6 years. Hypermetropia around spher. + 4 increased the likelihood of regression without surgery, emphasising the need for full refractive correction (there may have been some very early cases of accommodative esotropia). Age at recruitment, age that strabismus reportedly had started and degree of amblyopia at baseline examination seemed not predictive.

11.3.8Random-E ects Model Predicting the Angle and its Variation

In the 532 children of the ELISSS, the angle of strabismus, refraction, and visual acuity was assessed at baseline at approximately 11 months and every 6 months thereafter, until the final evaluation at the age of 6 years. The resulting, slightly more than 6,000, orthoptic exams were used to construct a random-e ects model [61] that forecasts the expected angle and its variation years ahead, on the basis of one or more measurements of the angle and refraction in infancy.

Angles of strabismus measured at di erent ages and the refraction of the patient can be entered in the model. On entering successive measurements of the angle of strabismus, the model adjusts the slope, i.e., yearly increase or decrease of the expected angle, according to the trend. The uncertainty about the slope decreases with additional measurements because the random e ect of the slope of the lines decreases. The uncertainty about the slope is compounded by additional variation of the angle around this slope for an individual child (Fig. 11.10).

In simulations with the random-e ects model, it was found that the chance of a spontaneous reduction of a strabismus into a microstrabismus is considerable when an angle of strabismus 14° or less is found repeatedly at the age of 1 or 2 years. In the ELISSS, esotropia 13° or less at baseline at approximately 11 months of age had not been operated at the age of 6 years in 35% of the cases (Fig. 11.7). If the angle is large on multiple measurements, the chance that the esotropia will decrease into a microstrabismus spontaneously is very small.

Horizontal angle of strabismus

Horizontal angle of strabismus

30

20

10

0

–10

0

30

20

10

0

–10

0

In the model, refractive error exerted its largest influence, i.e., causing the largest chance of spontaneous reduction into a microstrabismus, at a spher. + 4. Some children in the ELISSS study population may actually have been very early cases of accommodative esotropia. In case of hypermetropia, especially with convergence excess, a large reduction in the angle may occur after fitting full correcting glasses, thereby avoiding surgery.

Summary for the Clinician

■The chance of a spontaneous reduction of the esotropia into microstrabismus is considerable when an angle of strabismus of 13° or less is found repeatedly at the age of 1 year.

■Fit full-correcting glasses in case of hypermetropia accompanying esotropia at an early age because a large reduction of the angle of strabismus can be achieved without surgery and with better binocular vision.