The Ice Pack Test in the Differential Diagnosis of Myasthenic Diplopia

Results 

Eighty-nine patients, 56 males and 33 females aged 27 to 83 years (mean, 54 years), were included in the study and form the basis of this report. Thirteen patients were excluded because of loss to follow-up before establishment of diagnosis, history of posttraumatic restrictive strabismus, or inflammatory cause for blepharoptosis. Mean follow-up time was 9.3 months (range, 6–30 months). No adverse effects to the application of the ice pack, except for mild to moderate discomfort within the first 2 minutes of application, were noted in any of these patients.

Positive neostigmine test results were obtained in 5 of 28 patients tested. Among 62 patients submitted to serologic antibody testing, 9 had positive results for the acetylcholine receptor antibodies and 1 had positive results for MuSK.

Overall, 15 patients were diagnosed with myasthenia gravis, and 74 patients diagnosed with other entities served as controls. Seventy-three patients (13 myasthenic and 60 controls) were tested for ophthalmoparesis (diplopia) and 41 patients (13 myasthenic and 28 controls) were tested for blepharoptosis. Twenty-five of these patients (11 myasthenic and 14 controls) had both diplopia and blepharoptosis and were tested concurrently for these 2 conditions. Sensitivity and specificity of the ice pack test were calculated in the total population of patients tested for each condition, diplopia (n = 73) or blepharoptosis (n = 41).

Schematically, the patients can be divided into 3 groups according to the presenting symptomatology: group 1 comprises 48 patients with isolated diplopia, 2 of whom were diagnosed as myasthenic, 33 with paretic strabismus (6 with oculomotor, 16 with trochlear, and 11 with abducens nerve palsy), and 13 with diplopia of miscellaneous cause, including 4 patients with a recent onset of diplopia bearing the diagnosis of Graves' disease, 2 cases of decompensated intermittent exotropia (convergence insufficiency type), 2 patients with decompensated microtropia, and 1 case each of divergence insufficiency, late-onset comitant esotropia, skew deviation, fixation switch diplopia, and diplopia after cataract surgery.

Group 2 comprises 25 patients with both diplopia and blepharoptosis: 11 were diagnosed as myasthenic, 8 with oculomotor nerve palsy, 4 with chronic progressive external ophthalmoplegia, 1 patient with recurrent Tolosa-Hunt syndrome, and 1 patient with coexisting levator dehiscence and diplopia after cataract surgery.

Group 3 comprises 16 patients with isolated ptosis at presentation: 2 were diagnosed as myasthenic, 4 with Horner syndrome, and 10 with miscellaneous causes of ptosis, including levator dehiscence (n = 6) and congenital (n = 2) and contact lens-related (n = 2) blepharoptosis.

The results of the ice pack test in myasthenic versus nonmyasthenic patients is summarized in Table 1. The results of the ice pack test (5-minute application on initial presentation) were positive in 9 of 73 patients with diplopia, 8 of whom subsequently were identified as myasthenic and 1 of whom was diagnosed with fourth-nerve paresis (specificity, 98.3%; 95% confidence interval [CI], 90.3%–99.9%). The latter patient is a 71-year-old female with 6 prism diopters (PD) of right hypertropia compatible with right fourth-nerve paresis after herpes zoster ophthalmicus. The positive effect of the ice test was transient, lasting for less than 1 minute. A thorough work-up, including a brain computed tomographic scan, AchR and MuSK antibodies, antinuclear antibodies, and fluorescent treponemal antibody-absorption (FTA-ABS), were negative. C-reactive protein was elevated (30 mg/l; reference range, <5 mg/l). Diplopia resolved within 6 weeks and was attributed to inflammatory fourth-nerve paresis.

Table 1. Results of the Ice Pack Testing in Myasthenic versus Nonmyasthenic (Main Diagnostic Entities) Patients

		Positive⁎ Ice Test Results	Negative Ice Test Results	Total
	Myasthenia gravis (n = 15)
	Ophthalmoplegia	10	3	13
	Blepharoptosis	12	1	13
	Any of the above	14†	1‡	15
	Nonmyasthenic patients (n = 74)
	Ophthalmoplegia	1	59	60
	Paretic strabismus	1	40	41
	IIIrd nerve§	0	14	14
	IVth nerve§	1	15	16
	VIth nerve§	0	11	11
	CPEO	0∥	4	4
	Miscellaneous	0	15	15
	Blepharoptosis	2	26	28
	IIIrd nerve palsy	0	8	8
	Horner syndrome	0	4	4
	CPEO	0∥	4	4
	Miscellaneous	2	18	20
	Any of the above	3†	71‡	74

CPEO = chronic progressive external ophthalmoplegia.

⁎ Positivity by the lower threshold criterion of 50% reduction in deviation in primary position or at least 10 prism diopters for deviations or 25 prism diopters or more. † Positive for either diplopia or blepharoptosis (at least one), whichever was present in each patient. ‡ Negative ice test response for diplopia and blepharoptosis. § Paresis or paralysis. ∥ Partial response as evidenced by a decrease in deviation by 25% was noted in 3 of 4 patients, and an elevation of the upper eyelid by 1 to 1.50 mm was noted in all 4 patients.

The ice pack test results were positive in 15 of 41 patients with ptosis, 13 of whom were identified as myasthenic. In the sample of patients with nonmyasthenic ptosis, the test results proved to be false positives in 2 of 29 patients (specificity, 93.1%; 95% CI, 76.9%–99.1%). The first patient is a 39-year-old woman who also had diplopia and a history of recurrent attacks of painful ophthalmoplegia attributed to orbital myositis at presentation, and the second patient carried the diagnosis of common pemphigi. The neostigmine test and acetylcholine receptor antibodies analysis showed negative results in these patients. The ice pack test did not improve coexisting ocular misalignment in the former patient. A therapeutic trial of anticholinesterase medication (pyridostigmine) did not improve diplopia or ptosis in either patient.

Table 2 presents the clinical characteristics and results of diagnostic tests for 15 patients subsequently diagnosed as myasthenic. Thirteen (86.6%) of 15 myasthenic patients reported diplopia. Eight of these patients responded positively to the 5-minute test on initial evaluation by diminished primary position deviation of at least 50% (sensitivity, 61.5%; 95% CI, 35.4%–82.4%). One patient who had a partial response not fulfilling the 50% improvement criterion responded positively when the ice pack was applied for 10 minutes (sensitivity for the 10-minute application, 69.2%; 95% CI, 42.0%–87.6%).

Table 2. Results of the Ice Pack Test Application on Ptosis and Ophthalmoplegia in 15 Myasthenic Patients

	Patient No.	Diagnosis	Age (yrs)	Gender	Marginal Reflex Distance (mm)		Primary Deviation (MD)			Acetylcholine Receptor, Abs§, Muscle-Specific Kinase Results	Neostigmine Test Results
					Before Ice Test⁎	After Ice Test†	Before Ice Test⁎	5-Minute Ice Test‡	10-Minute Ice Test‡
	1	Generalized MG	70	F	0.50	+2.50	16 LHT, 6 LXT	6 LHT	6 LXT	+
	2	Generalized MG, SLE	55	F	—		16 LHT	7 LHT	5 LHT	+
	3	OMG	70	F	1.00	+3.00	10 RHpoT	5RHpoT	4 RHpT	+	+
	4	OMG	27	M	0.50	+3.50	4 HpoT	Ortho	Ortho		+
	5	Generalized MG	74	M	1.00 2.00	+4.00 +3.00	8 RET 4 RH	3 RET, 1 RH	2 RET	+++
	6	OMG	66	M	1.5	+2.50	Ortho	Ortho	Ortho	−	+
	7	Generalized MG	56	F	1.50	+3.00	11 RH	5 RH	5 RHT	+
	8	OMG	76	M	5.00, −0.50	+0.50 +4.00	14 RET, 5 RHT	14 RET, 5 RHT	12 RET, 2 RHT	+
	9	Generalized MG	70	M	1.5	+3.5	10 LHpT	4 LHypT	2 LHypt	+
	10	OMG	47	F	1.00	+2.50	—	—	—	−	+
	11	OMG	24	M	1.50	+1.50	35 LHT 10 XT	25 LHT, 8 XT	20 LHT	+	+
							16 LHT 8 XT	7 LHT, 5 XT
	12	Generalized MG	74	F	1.5	+3.50	25 LHT 10 LXT	14 LHT, 8 LXT	10 LHT, 8 LXT	+
	13	OMG (Graves + IVth nerve palsy), thymoma	33	M	1.00, 4.00	+4.50, +2.00	16 LHT 15 LXT	6 LHT, 10 LXT		+
					—	—	22 LHT 25 LXT	14 LHT, 14 LXT	12 LHT, 14 LXT
					—	—	20 LET 3 LHT	7 LET, 3 LHT
	14	Generalized MG, Graves, HIV+	52	M	Retraction	—	40 LX(T) 5 LHpT	35 LX(T), 5 LHpT	32 LX(T), 5 LHpT	MuSK+
							10 LX(T)	Ortho
	15	Generalized MG, Graves	79	M	6.00, −4.00	+1.50, +5.00	45 LXT, 10 LHpT	40 LXT, 6 LHpT	40 LXT	+

F = female; LHT = left hypertropia; LHpT = left hypotropia; LX(T) = intermittent left exotropia; M = male; MD = mean deviation; MG = myasthenia gravis; MRD = marginal reflex distance; OMG = ocular myasthenia gravis; RHpT = right hypotropia; RHT = right hypertropia; SLE = systemic lupus erythematosus.

⁎ MRD measurements and alternate prism and cover test measurements (in prism diopters) immediately before the application of the ice pack. † Difference in MRD measurements: MRD value immediately after removal of the ice pack minus MRD before the application of the ice test. ‡ Alternate prism and cover test measurements immediately after the application of the ice pack for 5 minutes and 10 minutes, respectively. § (+) or (—) indicates positivity for acetylcholine receptor antibodies. MUSK positivity is specifically indicated.

Three patients subsequently diagnosed as myasthenic had a recent diagnosis of thyroid eye disease at presentation (patients 13, 14, and 15, Table 2; Fig 1). Excluding these 3 patients from analysis, the sensitivity of the ice test for diplopia would be 70% (7 of 10 patients tested; 95% CI, 39.23%–89.67%).

Figure 1. Photographs of patient 15 (Table 2) showing coexisting myasthenia gravis and dysthyroid ophthalmopathy. Photographs showing (A) profound left blepharoptosis and ophthalmoplegia, which improved after (B) the application of the ice pack for 5 minutes. Photographs showing the same patient after the institution of antimyasthenic treatment (C) before and (D) after the application of the ice pack. Note bilateral eyelid retraction (D) resulting from concomitant thyroid eye disease, unraveled by the application of the ice pack.

Three patients (patients 11, 13, and 14, Table 2) were tested on several occasions with variable angles of deviation before treatment of myasthenia was initiated. One patient (patient 14) initially had a deviation of 40 PD extropia (XT) and a clinical picture of bilateral pseudointernuclear ophthalmoplegia that was not responsive to the application of the ice pack for either 5 or 10 minutes at presentation. Subsequent presentation a few days later with 10 PD of exotropia promptly reversed with application of the ice pack for 5 minutes.

Deviations larger than 20 PD did not respond by the criterion of 50% reduction of the presenting deviation. However, an improvement of 10 PD in at least one of the components of deviation occurred in 2 of 4 patients with a deviation of 25 PD or more at presentation (patients 11 and 12, Table 2).

Fourteen of 15 myasthenic patients showed positive results for at least one of the ocular manifestations of myasthenia, diplopia or blepharoptosis, bringing the overall sensitivity of the test for the detection of myasthenia gravis to 93.33% (95% CI, 68.05%–99.83%) if diplopia response is evaluated by the lower threshold criterion.

Ptosis in the context of oculomotor nerve palsy either remained unchanged or even worsened by 1 to 1.5 mm in 2 patients with third-nerve palsy of microvascular origin. None of these patients had an improvement in their diplopia, either.

The ice test invariably showed negative results in all 4 patients with Horner's syndrome. Three of 4 patients with chronic progressive external ophthalmoplegia reported a subjective improvement that was measurable in terms of 1 to 1.5 mm of elevation of the eyelids and a mild improvement in ocular deviation not exceeding 30% reduction of the amount of ocular deviation (none of 4 tested patients had deviations exceeding 20 PD) in primary position.

Increasing the duration of cooling to 10 minutes also was investigated in 14 patients with oculomotor nerve paresis and in 13 patients with various strabismic entities (3 patients with fourth-nerve paresis and 4 patients with sixth-nerve paresis, 1 patient with hypertropia after cataract surgery, 1 patient with a history of orbital myositis, and 4 patients with chronic progressive external ophthalmoplegia). None of these patients responded to the 10-minute test.

Discussion 

After physiologic studies demonstrating a temperature-sensitive effect on repetitive nerve stimulation from myasthenic patients with local cooling,11, 12 Saavedra et al6 suggested a cold test that favorably influences neuromuscular transmission for the diagnosis of ocular myasthenia.

Subsequent studies4, 8, 13, 14, 15 confirmed the diagnostic validity of the test in the evaluation of myasthenic ptosis and suggested 2 minutes as the optimum time of application of the ice pack, from the 5 to 10 minutes suggested initially.6, 16 The suggested threshold for a positive response in the evaluation of blepharoptosis is an elevation of the ptotic eyelid by 2.00 mm or more.4, 8, 14

The diagnostic value and the diagnostic criteria of the ice pack test for the assessment of myasthenic diplopia are not well defined, however, and despite its practicality, the test was reported by Thanvi and Lo10 as “rarely performed in the UK.” In the single reported series focusing on extraocular muscle responses, 15 of 15 myasthenic patients responded to orbital cooling “by dramatic improvement in alignment (no strabismus in primary position) or restoration of normal ductions.”5 No long-term follow-up was reported in the control group, and the authors acknowledge that some of these patients might have been myasthenic.

The authors' experience with the use of the ice pack has been quite different from the above. A partial rather than a complete resolution of diplopia was the rule for our patients. A decrease by half in at least one of the components of deviation in primary position was adopted as the threshold offering the optimum diagnostic value (sensitivity, 61.5%, and specificity, 98.2%, for the 5-minute application). Raising the threshold for a positive result to complete resolution of diplopia in primary position would have lowered sensitivity to 25% without increasing specificity. Had the threshold for improvement been lowered to 25%, sensitivity would have increased to 76.9% on initial application, but specificity would have decreased to 93.3% (95% CI, 83.61%–97.84%) because of 3 patients with chronic progressive external ophthalmoplegia responding to the test by this criterion.

The quantitatively small but consistent improvement in ptosis and extraocular muscle motility noted in the patients with chronic progressive external ophthalmoplegia has not been mentioned in the literature so far. It may denote a mild, nonspecific, facilitatory impact on muscle contractility in myopathic strabismus in general.

In patients with profound myasthenic impairment of ocular motility, the response to the ice pack tended to be less impressive. The authors suggest interpreting a decrease by 10 PD in the angle of strabismus after the ice pack test is applied for 5 minutes as diagnostic when the presenting deviation is larger than 20 PD. By incorporating this criterion, referred to as the lower threshold criterion, the sensitivity of the test reached 76.9% (95% CI, 49.06%–92.50%) without adversely affecting specificity.

Increasing the time of application to 10 minutes did not significantly improve sensitivity. The single patient who did respond to the 10-minute but not to the 5-minute application (patient 12, Table 2) would be considered a respondent anyway if the lower threshold for larger deviations were adopted.

The sensitivity for the assessment of myasthenic ptosis was superior to that for diplopia (92.3% vs. 76.9%, 5-minute application, lower threshold criterion), but not to a statistically significant degree (P = 0.593, Fisher exact test).

The blepharoptosis response was positive in all but 1 patient who tested positively for diplopia. This latter patient's ophthalmoplegia (patient 11, Table 2) responded partially to the ice test and was considered positive only by the lower threshold criterion. Two myasthenic patients with a positive response for blepharoptosis had a false-negative result for diplopia (patients 8 and 15; Fig 1).

The specific value of the test in addressing diplopia of myasthenic origin is emphasized further by testing 3 patients (patients 13, 14, and 15, Table 2) with signs of concurrent dysthyroid ophthalmopathy, one of whom (patient 13, Table 2) also had evidence of congenital mild fourth-nerve paresis. The test proved consistent in selectively addressing the part of deviation attributable to myasthenia during recurrences in the latter case.

The high prevalence of myasthenia gravis among patients with dysthyroid ophthalmopathy (reportedly 0.2% in the literature) presents an extra challenge to the clinician who needs a reliable tool to orient the differential diagnosis in this not uncommon scenario. This is the reason why those 3 patients with the apparently confounding variable of dysthyroid ophthalmopathy were included in the final sensitivity ratio. Antimyasthenic treatment (including steroid administration in 2 of them) resulted in prompt resolution of symptoms in all 3 patients, confirming the predominantly myasthenic origin of diplopia.

False-positive results were noted in 3 patients in this series, 1 diagnosed with inflammatory fourth-nerve paresis and 2 patients with blepharoptosis: one had a history of recurrent painful ophthalmoplegia and the other bearing the diagnosis of common pemphiga. There was an inflammatory-autoimmune common denominator in the etiopathogenesis of all 3 entities. No signs of external eye disease inflammation were documented in these patients. This association is noted, although a plausible explanation for a link between the peripheral site of action of the ice pack and the presumed inflammation-mediated etiopathogenesis of ptosis or diplopia in these 3 cases may not be offered. False-positive responses to the ice pack test have been reported very uncommonly in the literature.7

This study has to be viewed in light of its limitations. Variability in the presenting clinical picture is inherent to myasthenia and may pose methodologic problems in the interpretation of testing results and comparison of testing settings in the same patient. The authors thus believe that basing their conclusions on the initial evaluation is the best means to address the primary question in this study, that is, the value of this test as a preliminary test for myasthenia.

A major limiting factor in the interpretation of the results of this test is the standardization of the method of application of the ice pack. Bilateral simultaneous orbital cooling and the use of thin surgical gloves filled with ice cubes were adopted, as opposed to commercially available ice packs. The latter are associated with lower sensitivity, according to Ellis et al.5 Special care was taken each time a surgical glove was filled with ice cubes so that the configuration of the ice pack allowed for the largest possible surface of apposition with the convex contour of the eyeball, thus maximizing the cooling effect on the extraocular muscles. The need for customized care, however, may be considered suboptimal to standardization of testing conditions. Anatomic limitations in the exposure of extraocular muscles to the cooling effect of the ice pack may be responsible for lower sensitivity in the assessment of diplopia compared with blepharoptosis. With these anatomic restrictions in mind, the lower threshold criterion for larger angles of strabismus is suggested, although this is supported by evidence from a relatively small number of patients in this series.

Ensuring proper control of the head position is critical while cover test measurements are performed before and after the application of the ice pack test. Given the relatively small quantitative impact of the ice test in myasthenic deviation, one needs to eliminate gaze incomitance as a confounding factor in interpreting the results of the test. Immediate assessment of the ocular deviation after the ice pack is removed is critical because the effect of cooling is very short, in the order of 30 seconds to a few minutes. This is the reason the assessment was based on cover test measurements in primary position rather than on assessment of ocular motility.

Despite the availability of noninvasive tests such as lid fatigability in upgaze, the rest test, and the sleep test, establishing the diagnosis of myasthenia gravis necessitates the demonstration of positive AchR or MuSK antibodies, positive edrophonium or neostigmine test results, or a combination thereof.17 Although highly specific for myasthenia, a false-negative rate of 40% to 50% has been reported for acetylcholine receptor antibodies.18, 19 The edrophonium test long has been considered the gold standard for the diagnosis of myasthenia gravis with a sensitivity of 86% to 97% and a specificity of 83%.20, 21 However, it is time consuming with rare, but potentially serious, side effects, requiring careful monitoring of pulse and blood pressure. False-positive or paradoxical responses, like worsening of ocular alignment with edrophonium at a rate of 25%, have been reported.22, 23

Major advantages of the ice test include its accessibility as a quick, noninvasive, inexpensive, safe, simple24 preliminary test for orienting the diagnosis to myasthenia in the office. However, the physician should bear in mind that a partial rather than a complete response of ophthalmoparesis to the application of the ice pack may be expected. The threshold for a positive response offering the best diagnostic value in this series was a decrease by half in the angle of strabismus in primary position with the application of the test for 5 minutes. For angles of 25 PD or more, a decrease in the deviation by 10 PD could be considered diagnostic as well. By adopting this lower threshold criterion, the cumulative sensitivity of the test for assessing either manifestation of myasthenia gravis, ophthalmoparesis or blepharoptosis, was high (93.3%) in this series. However, the ice pack test proved of outstanding value in addressing diplopia of specifically myasthenic origin. A positive response when evaluating strabismus is highly suggestive of myasthenia even if multiple causes of diplopia coexist in the same patient. Potentially life-threatening causes of diplopia, ptosis, or both, such as oculomotor nerve palsy and Horner syndrome, invariably were nonresponsive or even aggravated by the ice pack.