The evaluation of patient demographics, etiologies and apraclonidine test results in adult Horner's syndrome

We aimed to demonstrate the patient demographics, etiologies and apraclonidine test results in adult Horner's syndrome. This retrospective study was performed by the analysis of medical data of patients who were given 0.5% apraclonidine test. Patients' past medical history, demographic data, etiologies, accompanying neurological findings and pharmacological test results were assessed. Forty patients (21 females and 19 males) with a mean age of 50.3 ± 11.6 years were evaluated. Apraclonidine 0.5% test was positive in 37 patients (92.5%). An etiology could be identified in 20 patients (central [9 patients, 45%], preganglionic [9 patients, 45%] and postganglionic [2 patients, 10%]). Neurological findings accompanying Horner’s syndrome were present in 8 patients. Despite detailed investigations, in a significant number of patients with Horner's syndrome an underlying cause may not be detected. Among the identifiable lesions, central and preganglionic involvements are still the first leading causes of Horner's syndrome. In addition, apraclonidine test may not be positive in all patients and a negative response does not exclude Horner's syndrome.


Introduction
Horner's syndrome (HS) also termed as ''oculosympathetic paresis'' was first reported by Claude Bernard in 1854. However, Johann Friedrich Horner described this syndrome in detail for the first time in 1869. The syndrome can occur at any age or ethnic group without significant gender predisposition. The estimated cumulative incidence is 2.4/100.000 person and the incidence peaks at age range of 50-54 years. HS results from interruption of the sympathetic pathway anywhere from the hypothalamus to the orbit [1][2][3][4][5][6]. This pathway is made up of three-neuron arc and the conditions causing the syndrome are clinically classified as central (first-order neuron), preganglionic (second-order neuron), and postganglionic (thirdorder neuron) according to the lesion localization.
Cerebral vascular accidents, Chiari malformation, syringomyelia, lateral medullary syndrome, multiple sclerosis, meningitis, encephalitis, intracranial tumors and spinal cord tumors may lead to central HS. Preganglionic lesions are associated with various causes including the mediastinal lymphadenopathy, apical lung cancer, subclavian artery lesions, neuroblastoma of the paravertebral sympathetic chain, dental abscess involving the mandibular region, brachial plexus traumas, cervical traumas and iatrogenic traumas. On the other hand, postganglionic lesions include internal carotid artery dissections or aneurysms, carotid-cavernous fistulas, cluster headaches or migraines, herpes zoster infection, Raeder's paratrigeminal syndrome and temporal arteritis [5]. Classical triad of HS consists of ipsilateral ptosis, miosis and anhidrosis though anhidrosis is unlikely to be present in postganglionic lesions [1]. It is easy to diagnose HS. However, determining the lesion localization is often challenging sometimes requiring great effort. Besides clinical findings, pharmacological testing is helpful for both the diagnosis and lesion localization [1][2][3][4]. Cocaine and apraclonidine are the most commonly used agents for the diagnosis [3]. However, drug choice for the diagnosis is still controversial. Furthermore, many researchers have reported that apraclonidine or cocaine tests could be negative in some patients with HS.
This clinical study aimed to demonstrate the baseline demographics, underlying etiologies and concomitant neurological findings in patients with adult HS. In addition, we reported the rate of apraclonidine positivity in adult HS.

Study design and patient's selection
This retrospective study was adhered to the Declaration of Helsinki. The data have been carefully collected by examining the medical records of the patients with adult HS who applied to the Neuroophthalmology clinic at Ege University and Bozyaka Training and Research Hospital between 2014 and 2019. We examined 45 patients' files. After the initial assessment of the records, five patients were excluded due to incomplete data. We identified 40 patients who had undergone apraclonidine 0.5% test. Past medical history, demographic data, underlying etiology, accompanying neurological findings and pharmacological tests used for the diagnosis were recorded. The results of comprehensive radiological imaging (chest X-ray, thorax computed tomography (CT), head CT, magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA) of the head and neck ultrasound) and laboratory tests were noted. Detailed neuro-ophthalmologic examination including eye movements, pupil size and its light and near responses were also examined.

The diagnosis of HS
In this study, patients who had anisocoria with normal pupillary constriction to the light and near in both pupils were pre-diagnosed with HS, and then, the diagnosis was confirmed by clinically and pharmacologically. Clinically, it was defined as miosis with or without ptosis with one or more of the following: (a) the presence of pupillary dilation lag of the smaller pupil, (b) anhidrosis, (c) hypochromic heterochromia iridis. On the other hand, apraclonidine 0.5% eye drop (Iopidine, Alcon, Fort Worth, TX, USA) was used for pharmacological test. Baseline photographs in dim light in a darkened room was obtained at first, and then, apraclonidine 0.5% drop was placed in each eye. Approximately 30-45 min after the administration, the repeated photographs were taken under the same light conditions. When the reversal anisocoria were determined, the diagnosis of HS was supported.
All statistical analysis was performed using SPSS (Statistical Package for the Social Sciences) for Windows (version 21.0; Chicago, SPSS, Inc). All data are expressed as the mean ± standard deviation.

Results
Forty patients' medical data were assessed in this study (21 females and 19 males). The mean age was 50.3 ± 11.6 years. Patients' demographic, underlying etiologies, accompanying neurologic findings and the results of 0.5% apraclonidine test are summarized in Table 1. When the patients were assessed according to the underlying etiologies, 16 patients (40%) were idiopathic and 4 patients (10%) were congenital. An underlying etiology was present in 20 patients (50%); central lesions in 9 patients (45%), preganglionic  [5]. MRI/MRA is usually recommended to rule out vascular events such as carotid artery dissection or non-vascular masses of the neck or chest. However, some authors suggest that these lesions can also be assessed by CT/CT angiography (CTA). While CT can provide an adequate image for soft tissue masses, carotid artery lumen can be visualized by CTA. On the other hand, an advantage of MRI/MRA is that shows not only vascular lumen but also its wall [7]. Furthermore, advanced magnetic resonance sequences (dynamic contrast enhancement, diffusion-weighted imaging, etc.) in the era of modern noninvasive brain imaging may aid to demonstrate neurovascular compression and to differentiate malignant lesions from the benign lesions of the head and neck [8][9][10]. Despite advances in neuroimaging and other diagnostic tests, in some patients with HS an underlying etiology cannot be identified. In this study, we detected an underlying etiology in 60% (24 cases) of the patients with HS, which was consistent with the previous studies. This rate was reported as 58.9% by Wilhelm et al. [11], 61% by Sabbagh et al. [12] and 65% by Maloney et al. [13]. In our study, most of the cases had central (45%) or preganglionic (45%) lesions, but postganglionic causes were found only in 10% of the cases. Contrary to our study, Maloney et al. [13] reported that preganglionic (44%) and postganglionic (43%) causes were found in similar rates, whereas central lesions (13%) were very infrequent. Common causes of HS differ depending on the population and/or ethnic groups in which the study was conducted. We demonstrated that the most common cause of HS was stroke (17.5%, 7 patients), followed by malignancies (12.5%, 5 patients) and previous surgery (7.5%, 3 patients). On the other hand, Sabbagh et al. [12] reported that the most common cause of pharmacologically confirmed HS was postprocedure (surgical, central venous line) (21%, 34 patients), followed by carotid dissection (9%, 14 patients), trauma (8%, 13 patients), cluster headache (8%, 12 patients), tumors (7%, 11 patients) and stroke (4%, 7 patients), respectively. Maloney et al. [13] demonstrated that tumors (13%), cluster headache (12%) and iatrogenic (10%) (neck surgery and carotid angiography) lesions are among the most frequent causes associated with HS. In a very recent study conducted by Han et al. [6], nationwide populationbased incidence and possible etiologies of Horner syndrome were reported. They included 1331 adult patients with HS. The peak incidence age of this cohort was 50-54 years. The authors demonstrated that the most common underlying etiology was iatrogenic causes secondary to surgical procedures. Similarly, most of the patients in our study were over 50 years, but HS secondary to surgical trauma was infrequent and observed in only 3 patients (7.5%).
Malignancy is one of the most important lifethreatening causes of HS. This syndrome may rarely be the first presentation of the malignancy, but it usually occurs long after the diagnosis of malignancy [2]. In a large series reported by Thompson et al. [14], malignancy was responsible from approximately 25% of the preganglionic HS cases. The most common malignancies associated with preganglionic lesions have been reported as breast and lung cancers. However, Han et al. [6] stated that the most common malignancy associated with this syndrome was thyroid tumor. Wilhelm et al. [11] reported malignant tumor in only one of 33 preganglionic lesions and 6 of 20 postganglionic lesions. Sabbagh et al. [12] found tumors in 11 of the patients (11%) as the underlying etiology of pharmacologically confirmed HS. Unlike the previous studies, we reported higher malignancy rates (4 of 9 patients, 44%) in patients with preganglionic HS. These malignancies consisted of lung cancer, laryngeal cancer, orbital invasion by neuroblastoma and nasopharyngeal cancer. In posterior cavernous sinus or brainstem involvement, HS may be accompanied by abducens paralysis. This condition, also called Parkinson's syndrome [15], was observed in two cases. One was cavernous sinus invasion due to nasopharyngeal tumor and the other one was brain stem metastasis.
Apraclonidine, an alpha-2 adrenergic agonist with weak alpha-1 adrenergic activity, is used for decreasing intraocular pressure in glaucoma patients. Although it has no obvious effect on pupil's size, iris dilation is observed due to supersensitivity of the postsynaptic alpha-1 adrenergic receptors located on the dilator muscle affected by oculosympathetic paresis [1,3,[16][17][18]. It is still so controversial which test used in the diagnosis of HS is more reliable. Early studies stated that apraclonidine and cocaine had similar sensitivity and specificity for detecting HS [17]. In a research conducted by Bremner [3], apraclonidine was found to have higher sensitivity (93%) when compared with cocaine (40%). The other disadvantages of cocaine are being more expensive and less available. Therefore, it is recommended that apraclonidine should be the ''gold standard'' pharmacological test for diagnosing HS [3,19]. However, the test must be performed cautiously within hours after symptom onset and infants under 1 year of age [3]. There are two main issues restricting the clinical use of apraclonidine. The first issue is the potential side effects, especially in children. In a study conducted by Watts et al. [19], drowsiness and unresponsiveness have been reported in infants under the age of 6 months after topical administration of 1% apraclonidine. Furthermore, hypotension, bradycardia, somnolence and lethargy have been reported as other side effects related to alpha-adrenergic receptor agonism [2,[19][20][21]. Although there are limited reports on the safety of apraclonidine in children younger than 10 years old, it is still the preferred drug diagnosing HS in infants and young children [1,22]. Another concern associated with topical apraclonidine usage is that the test results are negative in the early period after interruption of the sympathetic innervation. Because upregulation of the postsynaptic adrenergic receptors may require a certain period after sympathetic denervation, pupil dilation may not be achieved in an acute case [23]. The apraclonidine test can be positive between 1 month and 10 years after sympathetic interruption [12,16,[24][25][26][27][28][29][30][31]. Therefore, it may require several weeks or months to confirm a negative test result detected during the acute stage of HS. Several studies have reported that aproclonidine positivity varies between 88 and 100%. Brown et al. [27] suggested that apraclonidine positivity was 88%. On the other hand, this positivity was found to be 100% by Morales et al. [16], Koc et al. [18] and Bacal et al. [26]. However, the most important limitation of these studies mentioned above was the relatively small number of patients. In our study, we found a high rate of apraclonidine positivity (92.5%) similar to Bremner's study (93%) [3]. When we evaluated apraclonidine negative cases, we noticed that 2 patients were idiopathic and 1 patient was related to a central metastasis. However, the test was positive in all patients with second-and third-order neuron involvement. As the clinical picture was very typical, we did not perform further evaluation such as cocaine test in 3 patients who had negative response to 0.5% apraclonidine.
This study has some limitations. First, it has limited statistical power due to relatively small sample size. Second, standard apraclonidine test was used for the diagnosis of Horner's syndrome. Dilute phenylephrine, cocaine and hydroxyl-amphetamine could not be used in this study since either it is not generally available as a proprietary formulation or it is difficult to access. We therefore could not compare apraclonidine with cocaine. Third, we could not detect whether the apraclonidine test became positive after the acute period in patients who had a negative response initially. Fourth, our findings do not reflect the entire population as we only evaluated adult HS patients.

Conclusion
HS, an important clinical entity, has a large number of causes. Though often idiopathic or benign, the causes can sometimes be complex and life-threatening. Central and preganglionic disorders are still the first leading causes in the identifiable group. Although apraclonidine test is positive with high sensitivity in the majority of patients, the clinicians should keep in mind that negative apraclonidine test does not exclude the syndrome.