Long-Term Outcome and Predictor Variables in the Treatment of Acquired Esotropia with Botulinum Toxin

Abstract

purpose. To determine the long-term results of botulinum therapy in acquired esotropia and to identify predictors of a satisfactory outcome.

methods. Sixty-eight children (age range, 8–64 months) with acquired esotropia were enrolled in a prospective study. Botulinum toxin A was injected in the two medial recti. Motor and sensory statuses were evaluated at 1 and 2 weeks; 3, 6, and 12 months; and every year after the last injection. Univariate and multivariate logistic regression analyses were performed to relate motor and sensory outcome to variables recorded as potential predictors.

results. After an average follow-up of 4.8 years since the last injection, motor success was obtained in 36 children with one injection (52.9%), increasing to 48 (70.6%) and 60 (88.2%) children after two and three injections, respectively. Forty-eight (70.6%) patients had at least peripheral fusion (category 1 binocularity) and 32 (47.1%) had stereoacuity of at least 400 seconds of arc (category 2 binocularity). Higher hypermetropia, less severe amblyopia, and a smaller angle of esotropia were the best predictors of motor success. Minimal amblyopia and favorable motor alignment were associated with better binocularity outcome.

conclusions. Botulinum is an effective long-term treatment of acquired esotropia. It is especially useful in children with high hypermetropia, minimal amblyopia, and small esotropic deviation.

Esotropia that occurs in children after 6 months of age is usually considered “acquired esotropia.” This category of esotropia is undercorrected in 10% to 50% of patients after standard surgery—in part because the total amount of deviation is not always evident. ^{1 2 3} The preoperative wearing of prisms (prism adaptation) may disclose the full latent deviation in some cases. Surgery is then based on this deviation, with the inconvenience of cost and time. ^{1 3} Another suggested method to augment the amount of surgery is operating for an angle between the near deviation with correction and without correction. ² The alternative to conventional surgery is botulinum toxin injection. The frequent potential for fusion in children with acquired esotropia ⁴ theoretically makes them good candidates for this therapy. Botulinum toxin is considered effective in these patients, although there are different views. ^{5 6} The purpose of this report is to assess the long-term results in children with acquired esotropia who have been treated with botulinum injection and to elucidate which variables are useful predictors of outcome. 

Methods

Selection Strategy

Patients with newly diagnosed acquired esotropia arriving at our clinic were selected during a 3-year period of recruitment to participate in a prospective study. The study was approved by the institutional review committee and followed the tenets of the Declaration of Helsinki. Written informed consent was obtained from parents of participants. For inclusion in this study children had to be older than 6 months at the onset of nonparalytic, nonaccommodative, or partially accommodative esotropia; should not have had previous surgery; and had to be 8 years old or less at the time of corrective treatment. Age at onset was documented by historical information provided by the parents and confirmed by clinical records when possible. Criteria for exclusion were nystagmus, A-V pattern, dissociated vertical deviation, vertical deviation greater than 4 prism diopters, and a distance-to-near difference exceeding 10 prism diopters. Distance-near disparity was evaluated with patients wearing correction. 

Of the 163 esotropic patients examined, 68 were eligible (40 boys, 28 girls). Children were excluded by selection criteria (78 children) or parental rejection of therapy (17 children). Patients entered the study between the ages of 8 and 64 months. Characteristics of the study group are listed in Table 1 . 

Clinical Workup

Refraction was performed in patients under 1% cyclopentolate, 30 to 45 minutes after instillation of the drops. In children with darkly pigmented iris or when cyclopentolate retinoscopic findings were variable, we used atropine 1% twice daily for 3 days before refraction. Until children reached 4 years of age, we prescribed glasses with the full retinoscopic correction for children with hypermetropia of 2 diopters (D) or more, and myopia of 3 D or more. In children older than 4 years, both myopia and hyperopia of 1.5 D or greater were corrected. If the full hyperopic refraction was not tolerated by hypermetropes, it was reduced within 2 to 3 weeks, after subjective refining. In myopes, we used a lens of the minimum power that provided maximum visual acuity. ^{7 8} In children with residual esotropia after wearing their prescriptions for 6 weeks, we checked the effect of additional plus correction in reducing the deviation without visual blurring and/or repeated refraction with atropine. 

All described measures were mainly directed to avoid additional therapy in purely accommodative esotropia that may have been managed with spectacles alone. Whenever the residual deviation could not be reduced to less than 8 prism diopter at distance, we discussed with the children’s parents the possibility of botulinum injection and surgical correction and indicated botulinum treatment when the parents gave consent. We measured the preinjection deviation with correction by the alternate prism-and-cover test in 47 children, or by the Krimsky test when it could not be used (i.e., in 21 children). Visual acuity was determined by the Cardiff acuity test administered between 16 and 27 to 40 months of age and by the Allen cards or E test when children were able to cooperate (this occurred at ages between 27 and 40 months). The depth of amblyopia was measured as the difference in visual acuity between the two eyes (log minimum angle of resolution [logMAR] lines). In younger children, we also checked the severity of amblyopia by fixation preference at near, having the patient fixate on a small accommodative target, but data obtained with the Cardiff acuity test were incorporated into the statistical analysis. Amblyopia therapy was performed before botulinum injection, generally for 3 to 6 months (i.e., the period when it is maximally effective and after which further therapy probably will not be successful), ^{9 10} to avoid excessive delay of corrective treatment for esotropia. Thirty-six patients were amblyopes and received occlusive therapy (average duration, 130 days; range, 80–185 days). In 14 (38.8%) of them, the amblyopia was not completely responsive to treatment, and therefore they had some degree of amblyopia when injected with the toxin. 

Successful motor outcome was defined as a distance deviation of no more than 8 prism diopters by the simultaneous prism-and-cover test. The moment of alignment was considered when the patient reached a deviation within the successful range maintained for at least 6 months. 

The sensory fusion was evaluated by Bagolini lenses and the Worth 4-dot test and the stereo perception by the random-dot, with a random-dot ground (Randot Stereotests; Stereo Optical, Inc., Chicago, IL), Titmus, and TNO stereo tests. We distinguished two categories of binocularity. In category 1, children had peripheral fusion and gross or absent stereopsis. They were able to fuse at least on the Worth 4-dot test at near and did not show peripheral suppression on the Bagolini lenses, but stereopsis was not detected by the random-dot test (some of them had gross stereoacuity of >400 seconds of arc, as measured by the Titmus). In category 2, they had at least detectable stereopsis by the random-dot test (≤400 seconds of arc) and both peripheral and central fusion or a small central suppression scotoma on the Bagolini lenses. 

We injected botulinum toxin A (Botox; Allergan, Dublin, Ireland), with the children under brief nitrous oxide inhalation in the operating room, inserting the needle through the conjunctiva (without previous incision), and advanced with electromyographic control until the area of loudest sound, when we injected the fluid slowly. We used the maximal doses recommended by Scott et al., ⁶ depending on the amount of deviation (Table 2) . The dosage was always divided between the two medial recti, which were simultaneously injected. Children were observed for an average of 4.8 years after the last injection (range, 3.4–6.3 years), with evaluations at 1 week, 2 weeks, 3 months, 6 months, 1 year, and every year after the last injection. The decision to administer subsequent injections usually was made 6 months after the preceding injection (average time, 170 days; range, 121–192 days) or at the following visits when the desired alignment was not reached or maintained (i.e.,<8 prism diopters of esotropia at distance). It is important to note that refraction was performed after botulinum treatment during the 3-month follow-up visit. We dispensed the full hyperopic or minimal myopic correction (with subjective refinement at distance when possible). Residual esotropia was treated by repeating botulinum injection. 

Outcome Assessment and Statistical Analysis

The outcome assessment consisted of clinical examination of the motor and sensory status at 3 years and at last visit. Variables presumed to be associated with the motor and sensory outcomes (Table 1) were recorded before corrective treatment and during the follow-up sessions. 

The statistical analysis included investigation of association between variables of interest, mainly between the outcome variables at the end of follow-up and the potential predictors (univariate and multivariate stepwise logistic regression analysis using SPSS; SPSS, Inc, Chicago, IL). 

Results

Motor and Sensory Outcome

We recorded the motor and sensory results at 3 years after the last injection and at the last visit (i.e., at an average of 4.8 years after the last injection; range, 3.4–6.3 years) to evaluate the stability and long-term efficacy of this therapy in acquired esotropia (Fig. 1) . Sixty-one (89.7%) of 68 children were classified as having motor success at 3 years, and 60 (88.2%) maintained this success at last visit. One injection of the toxin was enough to produce a stable motor alignment through the entire follow-up in 36 (52.9%) children. With a second injection 12 (70.6%) more children obtained the desired stable alignment up to the end of the study, whereas 13 additional children reached desired alignment after a third injection, maintained it at 3 years, and 12 (88.2%) of them had maintained successful alignment at the last visit. Forty-six of the patients with successful motor alignment at the end of follow-up were orthotropic (67.6% of the total and 76.6% of the patients with motor success). Fifty-one (75%) children showed peripheral fusion (category 1 binocularity) at 3 years, and 48 (70.6%) children at the last visit. Of these, 32 (47.1% of the total sample) children showed stereopsis (category 2 binocularity), both at 3 years and at the last visit. Fifteen children did not have binocularity, and five were unable to cooperate. 

Transient ptosis occurred in 24 (35.2%) patients (average duration, 3.9 weeks; range, 2.1–7.3 weeks). There was only one (1.4%) child with overcorrection, with 20 prism diopters of exotropia as a consequence of botulinum injection and one (1.4%) with hypertropia greater than 4 prism diopters (7 prism diopters). New-onset amblyopia, resulting from the temporary ptosis or from overcorrection produced by the toxin (indicated by an increase in interocular logMAR line difference), and distance-near disparity were not found in any of the studied cases. Globe perforation did not occur. 

Analysis of Predictor Variables

We studied the effect of different variables (Table 1) with suspected clinical influence on the motor outcome. In the analysis, sex was treated as a categorical variable, whereas all other explanatory variables were treated as continuous variables. The motor outcome was significantly associated (P < 0.05) in univariate analysis with the depth of amblyopia and the angle of deviation, and was marginally associated (P < 0.25) with other variables (Table 1) . 

For multivariate stepwise analysis we included initially variables with a significance of P < 0.25 obtained in univariate analysis. In this model, depth of amblyopia and refraction remained, in combination, the two most valuable predictors of motor outcome (model 1 in Table 3 ) after elimination of the other variables. Patients with a lower level of amblyopia and greater hypermetropia were associated with a satisfactory stable motor alignment. If we included in the multivariate model only variables with strict statistical significance in the univariate analysis (P < 0.05 in Table 1 ), a lower level of amblyopia and smaller angle of esotropia were the best predictors of motor success (model 2, Table 3 ). The goodness of fit and predictability with the two models were good, but higher with the first model, and statistical power computation at a significance level of 0.05 yielded 96% and 67%, respectively. The calculated adjusted odds ratios of predictor variables were close to 1.00, although highly significant, because they provided a comparison of, for example, two children 1 logMAR line apart in depth of amblyopia (because these were the units used to measure the variable, such as 1 D or 1 prism diopter, for the other predictors). A more meaningful comparison would be obtained from model 1, comparing two children with 3 logMAR lines of difference in depth of amblyopia, which induces a 32% decrease in odds of motor success in the child with deeper amblyopia, after controlling for refraction. ¹¹  

The clinical characteristics of the eight patients who did not respond after multiple injections of the toxin are summarized in Table 4 . 

The effect of the same variables on binocularity outcome was also studied (in addition to the effect of the motor status 6 months after the last injection). The presence of category 1 binocularity was associated with lower level of amblyopia before treatment (P = 0.0006) and with a successful motor status at 6 months after the last injection (P = 0.002). The presence of category 2 binocularity was also significantly associated with a lower depth of amblyopia before treatment and with motor success at 6 months, but only the first of these variables remained significant in the multivariate analysis (P = 0.007). 

Discussion

Long-Term Efficacy of Botulinum Toxin in Acquired Esotropia

Botulinum toxin was effective in the long term in treating acquired esotropia. The overall motor success rate was 88.2%, with 70.6% of children having peripheral fusion. The main disadvantage was that only 52.9% of the treated children reached and maintained their alignment with a single injection, whereas 70.6% and 88.2% needed one or two and one to three injections, respectively. Children with hypermetropia of +2.50 logMAR or more, amblyopia of 3 logMAR lines or less, and angle of esotropia of 35 prism diopters or less (16 patients, 25%) have all gained adequate alignment after a single dose. 

Undesirable effects were transient and nondeleterious or infrequent: temporary ptosis in 24 (35.2%) children, which resolved spontaneously; only one (1.4%) case of overcorrected exotropia; and one (1.4%) case of significant postinjection hypertropia. Amblyopia did not develop as a consequence of these complications. 

The results reported herein are compatible with some preceding articles on botulinum toxin in childhood esotropia. ^{6 12 13} The long-term motor outcome after multiple injections is moderately improved compared with one of the studies, ⁶ in which different categories of esotropia are included, and is about the same as in the other two, of which one also included different esotropic categories ¹² and the other referred to infantile esotropia. ¹³ The motor success rates in these three studies are 66%, 85%, and 89%, respectively. 

Comparison between Botulinum and Surgical Treatment

These results give a hint of the possible results of a comparison of botulinum toxin with surgical treatment of acquired esotropia. Conclusions are not definitive because a comparison group treated with surgery is not included. Standard surgery is successful in 50% to 74% of cases, but these figures improve to 88% and 90%, respectively, after augmented surgery and prism adaptation. ^{1 2 3} Fusion results after one surgical procedure are approximately equivalent to those obtained after one to three botulinum injections: 60% to 75% of children have peripheral fusion with surgery ^{2 3} compared with 70% achieved after chemodenervation. Therefore, it seems that, from the motor and sensory point of view, a single operation may be as effective as one to three botulinum doses. Surgery is invasive, promotes scarring, and requires longer and deeper anesthesia. Botulinum is less invasive but more repetitive, does not produce scars, and demands brief superficial anesthesia, but it is not without risk for children. 

Prognostic Factors

Depth of amblyopia is a negative predictor of success with botulinum toxin, whereas this issue is controversial in recent conventional surgery literature. ^{3 14 15 16 17 18}  

Significant hypermetropia is a positive prognostic factor of satisfactory alignment with botulinum, and it has also been related to motor success after surgery for acquired esotropia. ^{3 16} In this study, changes in refractive prescription after botulinum injection were based on retinoscopic findings and subjective refinement, and we avoided overplussing with visual blurring to correct residual esotropia. Therefore, it is unlikely that such modifications are related to better results in high hypermetropes. It has been suggested that esotropes with an accommodative mechanism could have better binocularity. ³ Children with moderate to high myopia would not be as good candidates for this therapy as hypermetropes. 

A greater pretreatment angle is prognostic of failure, as is particularly evident in Table 3 . This applies also to conventional surgery. ^{3 16} The necessity of a higher total dose of the toxin is associated with worse motor results. Abbasoglu et al. ¹⁹ demonstrated that the percentage reduction of initial deviation was similar with each repeat botulinum application and that there was no cumulative effect. 

Binocularity in infantile esotropia after treatment with botulinum toxin has been suspected to be of poor quality, or at least of worse quality than in children who undergo surgery. ²⁰ Our data indicate, to a certain extent, that this effect, attributed to delay in correcting misalignment, would not occur in acquired esotropia. A stronger potential to regain bifoveal fixation and a more ready resolution of misalignment (due to smaller deviations) in acquired esotropia could account for this difference. 

In this investigation, depth of amblyopia before treatment was a negative predictor of binocularity in the long term. Previous studies have pointed out the independence between amblyopia and binocularity, ^{21 22 23 24} but it is also reported that amblyopia therapy may improve stereoacuity. ²⁵  

In conclusion, the present study indicates that botulinum treatment of acquired esotropia provides satisfactory long-term motor and sensory results. The most likely candidates for this therapy are children with minimal amblyopia, high hypermetropia, and a small angle of esotropia. 

 Presented in part at the 103rd Annual Meeting of the American Academy of Ophthalmology, Orlando, Florida, October 24–27, 1999.

 Submitted for publication April 23, 2001; accepted June 7, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Characteristics of the Study Group and Association with Motor Success

Variable	M	SD	Univariate Association with Motor Success (P)	Patients (n)
Age at onset (mo)	20.83	8.12	0.97
Age at start of botulinum treatment (mo)	36.29	11.42	0.54
Age at alignment (mo)	48.00	13.57	0.30
Sex			0.71
F				28
M				40
Duration untreated (mo)	15.46	11.42	0.14 (−)
Duration until alignment (mo)	27.16	13.57	0.10 (−)
Pretreatment angle (PD)*	35.42	17.57	0.02 (−)
Pretreatment refraction (D), †	1.38	2.68	0.21 (+)
Pretreatment amblyopia, ‡	0.56	1.29	0.004 (−)
Injections (n)	1.676	0.843	0.42
Total dose (U)	16.426	8.640	0.18 (−)
Maximal deviation (PD), §	−12.848	19.430	0.75

 Data are mean ± SD. D, diopters; PD, prism diopters; U, units; (−), negative relationship; (+), positive relationship.

*  Alternate prism and cover at distance (or Krimsky test).

^†  Mean spherical equivalent.

^‡  Interocular acuity difference in logMAR lines.

^§  Maximal deviation angle caused by the paralyzing effect of the toxin.

Table 2.

 

Botulinum Toxin Dosage

Esotropia (PD)
<20	1.25
>20	2.5–5
20–30	2.5
30–40	2.5, 5*
40–50	2.5, 5*
>50	5

 Data are units botulinum toxin per eye muscle. Bilateral medial rectus injection was always performed. PD, prism diopters.

*  2.5 U injected in one medial rectus and 5 U in the second.

Figure 1.

 

Percentage of children with successful motor result (distance manifest deviation ≤8 prism diopters) and who reached category 1 (peripheral fusion) or category 2 (stereoacuity ≤400 seconds arc) binocularity. Dashed lines: the 3-year outcome; solid lines: outcome at the end of follow-up (mean, 4.8 years) and at 3 years, when there is no difference between the two.

Table 3.

 

Effect of Predictor Variables^* on the Motor Outcome by Multivariate Logistic Regression

Variable	Coefficient	Standard Error	P	Adjusted Odds Ratio (95% CI)
Model 1
Constant	0.9639	0.0418
Amblyopia depth, †	−0.1254	0.0240	0.0012	0.882 (0.841–0.924)
Refraction (D)	0.0380	0.0118	0.0149	1.04 (1.014–1.063)
Model 2
Constant	1.1674	0.0979
Amblyopia depth, †	−0.0905	0.0346	0.0142	0.913 (0.853–0.977)
Deviation angle (PD), ‡	−0.0059	0.0025	0.0267	0.994 (0.989–0.999)

 Model 1 includes variables with significance of P < 0.25 in univariate analysis (Table 1) for stepwise building of the model. Model 2 includes variables with significance of P < 0.05 in univariate analysis (Table 1) for stepwise building of the model. D, diopters; PD, prism diopters; CI, confidence interval.

*  Variables measured before botulinum treatment.

^†  Interocular acuity difference in logMAR lines.

^‡  Alternate prism and cover test at distance (or Krimsky test).

Table 4.

 

Characteristics of Children with Unsatisfactory Motor Result

Patient	Age at Onset (mo)	Age at Start of Treatment (mo)	Sex	Pretreatment Angle* (PD)	Postreatment Angle, † (PD)	Pretreatment Refraction (D)	Pretreatment Amblyopia, ‡	Injections (n)	Total Dose (U)	Maximum Deviation, § (PD)
6	21	28	M	60	20	−7	5.2	3	25	0
10	13	23	F	35	12	+1	0	3	17.5	8
12	22	36	M	55	18	−1.75	2.2	3	22	10
17	25	42	F	70	20	+1.5	1.5	4	30	14
21	14	25	M	45	15	+2	0.97	3	19	10
32	19	39	M	50	−20	+3.75	0.7	2	20	−60
35	28	41	F	50	18	+1.5	0.45	3	20	12
65	16	30	F	65	25	−2.5	3.01	3	28	15

 PD, prism diopters; D, diopters.

*  Alternate prism and cover test at distance (or Krimsky test).

^†  Simultaneous prism and cover test at distance.

^‡  Interocular acuity difference in logMAR lines.

^§  Maximum deviation produced by the paralyzing effect of the toxin.

The authors thank Víctor Abraira, PhD, for his assistance with statistical analyses.