Recurrent Painful Ophthalmoplegic Neuropathy: Diagnostic Controversies. Single Case Report with a Systematic Review

Background: Ophthalmoplegic migraine, renamed “Recurrent Painful Ophthalmoplegic Neuropathy (RPON) in 2013 by the International Headache Society is a rare neurologic disorder characterized by recurrent attacks of ophthalmoplegia associated to ipsilateral headache. The etiology is still unknown. Typical magnetic resonance imaging ndings show a focal nerve thickening and contrast enhancement. In the majority of cases, there is a full recovery within days or weeks. There is no evidence supporting a specic treatment. The review denes the characteristics of the recurrent painful ophthalmoplegic neuropathy in patients within 2 years of age underlying the importance of the role of magnetic resonance imaging even in presence of the rst attack. Thus, an emblematic case report is presented. Case presentation: The authors present a case of third cranial nerve paresis in a 17-month-old male child, presenting a neuroradiological pattern highly suggestive of schwannoma, aneurism or recurrent painful ophthalmoplegic neuropathy. Thus, a review of the literature with the pediatric casuistry of recurrent painful ophthalmoplegic neuropathy occurred within 2 years of age focusing on diagnostic considerations is presented. The authors highlight the importance to consider recurrent painful ophthalmoplegic neuropathy in presence of magnetic resonance imaging ndings and clinical symptoms referable to aneurysm or schwannoma. Thus, the review denes the characteristics and the neuroradiological ndings at the rst RPON attack occurred under 2 years of age. Conclusion: Although two attacks are necessary, the review strongly suggests to consider recurrent painful ophthalmoplegic neuropathy even at the rst attack, in presence of described characteristics and the aforementioned magnetic resonance imaging ndings. Two eligible studies consensus on which studies to include. The methodological quality of this systematic review has been assessed using the AMSTAR 2(12) tool as a quality review”, no randomized controlled trials (RCTs) abstracted the following information: age at onset and current age; sex; cranial nerve (CN) involved (side); headache (side); associated symptoms: photophobia, phonophobia, nausea, vomiting, irritability, other ndings; ocular ndings: ocular symptoms/signs, diplopia, ophthalmoplegia, palpebral ptosis, pupillary dilation; MRI ndings in the acute phase: nerve thickening, post-contrast enhancement; altered cerebrospinal uid (CSF) if lumbar puncture performed; headache duration; ophthalmoplegia duration; interval between headache onset and ophthalmoplegia; time to resolution of symptoms/signs; therapy in the acute phase; follow-up; prophylactic therapy; control MRI, number of acute episodes; interval between episodes; comorbidity; family history of migraine. a patient affected by RPON who developed, after years, a schwannoma of the third cranial nerve. They proposed two explanations of the aforementioned correlation. According to the rst one, tumor could intermittently release chemical substances which stimulate trigeminal nerve receptors, leading to the headache. In this case, schwannoma would mimics RPON and it would be an initial manifestation of the tumor, which would be too small to be found in MRI. The second hypothesis, instead, considers RPON as an inammatory cranial neuralgia and not a migraine. In this case, episodes of inammation lead to demyelination and remyelination. Schwann cells proliferation could lead to the transformation into schwannoma. As a result, isolated oculomotor schwannoma could be considered as a long-term complication of RPON. Both hypothesis suggest the importance of serial brain MRIs in the long-term follow-up of RPON. In conclusion, our case, compared to the reviewed literature, a diagnosis of RPON was highly suggestive even at the rst attack. Our work highlights the importance to consider RPON in presence of MRI ndings and clinical symptoms referable to aneurysm or schwannoma. This review denes the characteristics of MRI ndings at the rst RPON attack occurred under 2 years of age. Although two attacks are necessary, it strongly suggests to consider RPON even at the rst attack, in presence of described characteristics. Thus, as mentioned by Wong(11) and Yinglu(14), we suggest to add into the diagnostic criteria the MRI ndings, including enhancement and thickening of the nerve involved. We analyzed the relationship between RPON and schwannoma. As proposed by Petruzzelli et al.(15), we are performing a long-term follow-up at our institute in order to prevent any complications.


Background
Ophthalmoplegic migraine (OM), renamed "Recurrent Painful Ophthalmoplegic Neuropathy (RPON) in 2013 by the International Headache Society(1) is a rare neurologic disorder characterized by recurrent attacks of ophthalmoplegia associated to ipsilateral headache, that can be migranous or not. In population, incidence is estimated as 0.7/million every year. Laboratory tests such as erythrocythe sedimentation rate, rheumathoid factor, antinuclear antibody, Venereal Disease Research Laboratory (VDRL) and cerebrospinal uid are normally negative (2). In 80% of cases it involves the third cranial nerve (3). Typical Magnetic Resonance Imaging (MRI) ndings show a focal nerve thickening and contrast enhancement. The etiology of this rare syndrome is still unknown (4). In the majority of cases, there is a full recovery within days or weeks, but less frequently, patients have persistent neurologic de cits (5). According to the International Classi cation of Headache Disorders (ICHD), it is possible to diagnose RPON/OM with at least 2 attacks of a migraine-like headache associated with paresis of the ocular cranial nerves that occurs within 4 days from the beginning of symptoms. Possible ocular ndings are ophthalmoparesis, ptosis, or mydriasis. In order to perform the diagnosis, causes as tumors, infections, and thrombosis must have been excluded (4). According to the aforementioned classi cation, diagnosing this disorder during the rst attack is not possible, even if the MRI ndings are highly suggestive of RPON. Speci cally, thickening of the interested cranial nerve, with a reduced post-contrast enhancement isa typical pattern. The role of the imaging is challenging because frequently MRI is negative even in case of con rmed RPON.
It is worthy to mention that ocular nerve palsies are rare in childhood(6). Speci cally, the third cranial nerve is the less affected in children (7).
RPON is one of the rarest causes of third cranial nerve palsy(8). Herein, we report a case of third cranial nerve paresis in a 17-month-old male child, presenting a neuroradiological pattern highly suggestive of schwannoma, aneurism or RPON. Even if the MRI at the rst attack was highly suggestive of RPON, the diagnosis according to the ICHD was not possible. Thus, we systematically reviewed the pediatric cases of RPON/OM occurred within 2 years of age comparing them with our case.

Case Presentation
Herein, we report the case of a 17-month-old male child referred to our Institute presenting gradual onset of mild eyelid ptosis and divergent strabismus of the left eye, preceded two days before by an episode of vomiting. A week prior to the hospitalization, an episode of inconsolable crying, lasting about two hours, occurred with loss of appetite during the following days. Neither fever nor other clinical ndings were evident.
The patient, third son, was born at term from Cesarean section after pregnancy complicated by placenta previa. Neonatal period was regular.
Spherocytosis was diagnosed during the rst months of life. His family history revealed spherocytosis (mother and sister) and Hashimoto's thyroiditis (mother). At admission, physical examination was normal, except for eyelid ptosis and lateral deviation of the left eye due to mild medial rectus muscle de ciency and without pupillary dilation, suggesting the involvement of the third cranial nerve. Fundus examination was normal. C-reacting protein (CRP) was negative. Moreover, serological tests and autoimmune panel were negative. Brain magnetic resonance imaging (MRI), enhanced after contrast administration, and magnetic resonance angiography (MRA) were performed. They suggested a vascular anomaly, along the medial side of the left cerebral peduncle, referable to an arterial aneurysm nearby the ipsilateral third cranial nerve ( gure 1).
However, the angio-CT examination did not con rm the vascular anomaly, highlighting a minimal size irregularity of the P1 tract of the left posterior cerebral artery ( gure 2). On the basis of MRI ndings, a third cranial nerve neuropathy was suspected. About three weeks after hospital admission, left third oculomotor nerve ophthalmoplegia was no longer appreciable. One month later, a brain MRI was repeated and con rmed a sectorial slight thickening of the emergence of the left third cranial nerve, with a reduced post-contrast enhancement compared with the previous exam ( gure 3).
One year later, a brain MRI was repeated, showing a complete resolution of the previous neuroradiological lesions ( gure 4). In relation to MRI ndings and clinical situation, the reported case was highly suggestive of an episode of recurrent painful ophthalmoplegic neuropathy.
Nonetheless, according to the diagnostic criteria proposed by the International Classi cation of Headche Disorders (ICHD) (2018) at least two attacks are necessary to con rm the diagnosis (9). Other considerable, even highly improbable, hypothesis was a schwannoma of the third nerve.
For a correct evaluation of the case, we planned a strict follow-up: after 18 months from the diagnosis, the patient had an episode of headache with inconsolable crying treated with paracetamol. During this episode, neurological examination was negative. No other similar episodes with ophtalmoplegia occurred and the neurological examination was negative. After 30 months, the child was conducted at our emergency department presenting again eyelid ptosis and divergent strabismus of the left eye, associated with vomiting and headache. During the hospitalization symptoms gradually resolved spontaneously with a total resolution. This second acute attack con rmed our already strongly suspected diagnosis of RPON.

Discussion And Conclusions
A diagnosis of RPON is always challeging, especially under 24 months of age. In fact, children are not able to describe headache and the rst symptom is often irritability. Our case is highly suggestive of RPON but a de nite diagnosis was not possible at the rst attack(9).
Third nerve thickening and post-contrast enhancement are suggestive of oculomotor nerve schwannoma, carcinomatosis, granulomatosis, in ammatory or infectious neuritis (10). The case presented was from the beginning highly suggestive for RPON but MRI ndings still have no relevance in the diagnostic criteria. Our case reported represents a diagnostic controversial: all symptoms and neuroradiological lesions were highly suggestive but no diagnostic possibilities were admitted during the rst attack. In 1997 Wong and Wong (11) suggested to include these MRI ndings associated to a single reversible episode of ophthalmoplegia as a supportive diagnostic criteria of RPON. Notwithstanding, neuroradiological imaging during the rst episode have only suggested a probable and not a de nitive diagnosis so far. It is important to highlight that in our case CT was positive, unlikely Ambrosetto et al. (2) reviewed, showing that CT is normally negative. We systemically reviewed in Pubmed all cases of RPON occurred within 24 months of age and we compared it with our patient. PubMed was searched for all cases of RPON using the search terms "ophthalmoplegic migraine OR recurrent painful ophthalmoplegic neuropathy". Only articles in English or Spanish have been ltered. We performed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement ( gure 5). Inclusion criteria were age (up to 24 months), presence of MRI ndings and diagnosis of RPON, for these reasons, we excluded 208 records from database searching. Two reviewers independently agreed on selection of eligible studies and achieved consensus on which studies to include. The methodological quality of this systematic review has been assessed using the AMSTAR 2(12) tool as a "low quality review", since no randomized controlled trials (RCTs) are available to date on the scienti c literature.
We abstracted the following information: age at onset and current age; sex; cranial nerve (CN) involved (side); headache (side); associated symptoms: photophobia, phonophobia, nausea, vomiting, irritability, other ndings; ocular ndings: ocular symptoms/signs, diplopia, ophthalmoplegia, palpebral ptosis, pupillary dilation; MRI ndings in the acute phase: nerve thickening, post-contrast enhancement; altered cerebrospinal uid (CSF) if lumbar puncture performed; headache duration; ophthalmoplegia duration; interval between headache onset and ophthalmoplegia; time to resolution of symptoms/signs; therapy in the acute phase; follow-up; prophylactic therapy; control MRI, number of acute episodes; interval between episodes; comorbidity; family history of migraine. The median age at the rst attack was 14,3 months. 69,2% of patients were females and 30,8% males. The cranial nerve involved was always the third one, except for two patients where it was not mentioned (13). headache was the most frequent symptom, followed by nausea and vomiting. One third of patients presented associated symptoms such as photophobia, phonophobia and irritability. Ocular symptoms/signs were always present: ophthalmoplegia and palpebral ptosis were the most frequent ones, followed by diplopia. Unlikely reported by Huang (3), we found some patient who presented pupillary dilation.
It is interesting to highlight the MRI ndings: it had been showed a nerve thickening in 61,5% of cases and a post-contrast enhancement in 53,8% of patients. In cases where MRI was negative, it is important to understand whether the imaging was really negative or the timing was wrong. In fact, in some patients there is no evidence of MRI abnormalities neither during the interictal phase nor during the rst attack and it could only be found after attacks (14). Therapy in acute phase had been administered in 70% of patients using corticosteroids. In 50% of cases, at follow-up examination was noted a periodic recurrence of migraine with or without ophthalmoplegia. A limited number of patients (20%) had permanent neurological damage. Control MRI had been performed in 50% of cases. It showed in a limited number of patients (20%) normal ndings and in the majority of them (80%) a persistent enlargement but to a lesser degree. Notably that family history of migraine was positive in 46,1% of patients and in most cases was on the mother side. We compared the characteristics of our patient with the ones of the review (MRI ndings in the acute phase, symptoms and its duration and response to therapy). For the aforementioned reasons, we strongly supported from the rst attack that this case was highly suggestive of RPON. We highlight that the rst diagnostic hypothesis were aneurysm and schwannoma. Our work let us to extend the knowledge about RPON, suggesting to think at this diagnosis at its very rst attack, even in presence of initial MRI ndings referable to vascular anomaly or tumors as schwannoma. A relationship between RPON and schwannoma has been often discussed. In 2019 Petruzzelli et al. (15) reported a patient affected by RPON who developed, after years, a schwannoma of the third cranial nerve. They proposed two explanations of the aforementioned correlation. According to the rst one, tumor could intermittently release chemical substances which stimulate trigeminal nerve receptors, leading to the headache. In this case, schwannoma would mimics RPON and it would be an initial manifestation of the tumor, which would be too small to be found in MRI. The second hypothesis, instead, considers RPON as an in ammatory cranial neuralgia and not a migraine. In this case, episodes of in ammation lead to demyelination and remyelination. Schwann cells proliferation could lead to the transformation into schwannoma. As a result, isolated oculomotor schwannoma could be considered as a long-term complication of RPON. Both hypothesis suggest the importance of serial brain MRIs in the long-term follow-up of RPON. In conclusion, our case, compared to the reviewed literature, a diagnosis of RPON was highly suggestive even at the rst attack. Our work highlights the importance to consider RPON in presence of MRI ndings and clinical symptoms referable to aneurysm or schwannoma. This review de nes the characteristics of MRI ndings at the rst RPON attack occurred under 2 years of age. Although two attacks are necessary, it strongly suggests to consider RPON even at the rst attack, in presence of described characteristics. Thus, as mentioned by Wong (11) and Yinglu (14), we suggest to add into the diagnostic criteria the MRI ndings, including enhancement and thickening of the nerve involved. We analyzed the relationship between    Complete resolution of the previous neuroradiological lesions PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement