An easily missed cause for low GCS in a Scottish Intensive Care Unit

Abstract

A case report of a 36-year-old patient who presented with seizures, a large thyroid and decreased consciousness level requiring admission to the intensive care unit, where the final diagnosis had a prevalence of 2.1/100,000 and a female to male ratio of 4:1. The final diagnosis was suggested early by a focussed internet search. This is unusual because as clinicians we often tell the general public to ignore an internet search. The early suggestion demonstrated that in the right hands the use of internet searches can help us make more informed decisions and aid patient management when combined with traditional investigative medical practice. However, the lesson with this case is that there are many unusual and rare causes of presentations to critical care units and as clinicians we should keep an open mind and wide differential diagnosis so as not to miss the rarer reasons for these presentations.

Case report

A 36-year-old male presented to the emergency department (ED) following an unresponsive period at home where paramedics found him to have abnormal arm and leg movements in keeping with a seizure. Transfer time to ED was 34 min where he was found to have a left upward gaze, his right arm was in fixed flexion, he had continuous myoclonic type movements in his left arm and leg, and he had a Glasgow coma scale (GCS) of 9 (E4 V2 M3). He was managed with IV lorazepam, but quickly progressed to a rapid sequence induction. Intubation was difficult due to a large thyroid mass affecting his laryngeal position. He was empirically treated with IV tazocin, gentamicin, and acyclovir. He was subsequently changed to ceftriaxone upon intensive care unit (ICU) admission.

This gentleman had a medical history that was suspicious of autoimmune hypothyroid disease, but at the stage of presentation, this had not yet been diagnosed formally. Collateral history from the family revealed three months of increasing thyroid swelling which was awaiting investigation via fine needle aspiration. He was also undergoing thyroid replacement therapy. He was otherwise fit and well with no other medical problems.

Computed tomography (CT) head and neck showed normal intracranial appearances and a large bulky homogenous thyroid with small posterior mediastinal lymph nodes in keeping with viral thyroiditis. Collateral history from the family revealed three months of increasing thyroid swelling which was awaiting investigation via fine needle aspiration.

In ICU, continuing clinical and subclinical refractory seizures necessitated starting phenytoin, levetiracitam, sodium valproate, and a midazolam infusion. Blood investigation showed a white cell count of 14 × 10⁹/L, C-reactive protein of 6 mg/L, a thyroid-stimulating hormone (TSH) of 98 mU/L (0.4–4.5), free T4 of 16 pmol/L (10–24), and a free T3 of 2.6 pmol/L (4–8.3). Antithyroid peroxidase levels were high (>3000 U/ml). The high TSH prompted an increase in his prescribed levothyroxine. All sets of blood cultures were negative (four sets in total), and there were no significant electrolyte abnormalities. A lumbar puncture had a negative Gram stain, no organisms, a raised protein of 904 (0–700 mg/L), no cells, no blood, and a normal glucose. Viral polymerase chain reaction of this sample was also negative. Following a failed extubation, an electro-encephalogram showed no evidence of epileptiform activity, just slow wave activity. An extubation at 48 h was successful, but GCS did not improve beyond 11 (E3, M5, V3), necessitating re-intubation for magnetic resonance imaging (MRI) which showed a normal brain parenchyma with no evidence of intracranial bleed or mass. All images showed an absence of left transverse and sigmoid flow void, the radiologist concluding that venous sinus thrombosis was possible. Following MRI, the patient was successfully extubated with recovery to GCS 14 (E4, M6, V4) and discharged from ICU to the neurology ward. A total of four days was spent in ICU.

Subsequent investigations included a CT venogram demonstrating normal flow and no evidence of venous sinus thrombosis, a further lumbar puncture at seven days giving negative results and a repeat EEG emulating previous results. All antibiotics and antivirial treatments were stopped, and he was discharged from hospital at eight days with only levetiracetam, phenytoin, and levothyroxine prescribed regularly. He underwent an ultrasound scan of his thyroid, which confirmed a hetrogenous thyroid in keeping with viral thyroiditis. Clinical examination showed that his thyroid had markedly reduced in size by the time of discharge.

Unfortunately, over the next week he was re-admitted via his general practitioner with confusion and cognitive impairment. His symptoms deteriorated causing reduced co-ordination, reduced ability to follow simple commands, and an associated persistent tremor. Sodium valproate was again added into his antiepileptic regime. The neurology team extensively discussed his case and he was started on a regime of prednisolone resulting in significant clinical improvement. The initial dose of prednisolone was 1000 mg dropping to 60 mg on discharge. He was discharged a week after readmission and had no further seizure activity for a couple of months. A diagnosis of Hashimoto’s encephalopathy (HE) has tentatively been made.

Unfortunately, he then had a further two seizures requiring two intensive care admissions to different units. Following the second of these admissions, he was treated with intravenous immunoglobulin (IVIG) therapy and has since been seizure free but has had a long neurological recovery. He has suffered from significant fatigue, chronic pain, and anxiety but is now improving. Since IVIG he has successfully been weaned off of the antiepileptics and remains seizure free.

Discussion

HE is a rare and uncommon diagnosis often made on the basis of exclusion of other pathologies by means of extensive investigation. It was first mentioned in literature in 1966, where a case of encephalopathy was described in the context of concurrent Hashimoto’s thyroid disease.^1,2 A brief literature search shows that there are only five publications relating to intensive care admissions.^3–7 These publications describe a total of 13 patients between them. The age of these patients ranged between 13.5 and 83 years. The literature surrounding HE would suggest that the fifth decade of life is the most common time of onset and there is a prevalence towards females. ⁸ Although HE is the name coined, a more useful one is steroid responsive encephalopathy associated with autoimmune thyroiditis, a more descriptive name but one that maybe hinting at the lack of understanding surrounding pathophysiological processes that occur. ⁸ Naming aside the presence of antithyroid antibodies, encephalopathy and steroid response is a unique syndrome. ⁹

There have been multiple suggestions as to how and why these patients develop encephalopathy, but despite much debate, there is no firm conclusion. The most widely accepted theories are of autoimmune vasculitis, ¹⁰ global brain hypo-perfusion, ¹¹ or the presence of an antigen within both the brain and the thyroid. ¹² It may be more relevant to discuss what is not considered the cause, which is the presence of elevated thyroid stimulating hormone, or decreased T4 or T3 levels. In many respects, this is what makes this diagnosis unique as unlike many pathological processes linked to thyroid disease – euthyroid patients predominate in HE. ⁹

It is generally accepted that there are two common types of presentation in HE. The first is the more acute type, similar to the patient described here. It involves a presentation with seizures, stroke-like symptoms, transient focal neurology, or loss of consciousness. The second is much more subtle in its onset developing over weeks or months with symptoms not dissimilar to those that dementia patients suffer. Confusion, inappropriate aggression, and general cognitive decline have all been described. ¹⁰ It is not clear if age has a link to the type of symptoms that predominate; however, paediatric cases seem to present most commonly with seizures. ⁶

The diagnosis of HE presents a significant challenge to physicians as so many conditions have similarity in their symptom profile. It is probably fair to say that the diagnosis remains one of exclusion as well as correlation with clinical response to immunosuppressive treatment. As in this case, a lot of investigations are needed to exclude other causes of encephalopathy such as infection, metabolic, toxin or malignancy. The combination of lumbar puncture, blood tests and imaging are all required. Cerebro-spinal fluid (CSF) results can certainly lead clinicians towards HE as a diagnosis, with approximately 75% of the described cases having raised protein and 25% increased cell count, glucose often remains normal. ¹³ A confounding factor described in some case reports is the presence of brain 14-3-3 protein within CSF, which usually indicates Creutzfeldt-Jakob disease but has been seen in patients with HE. In these cases, it would be prudent to confirm with histology investigations so as to pursue the correct course of treatment. ¹⁴ However, in this case, the neurology team involved did not feel that it was necessary to exclude this, as the presentation had none of the hallmarks of Creutzfeldt-Jakod disease, his initial results led to a more likely diagnosis of HE.

Imaging studies are usually reported as normal, but there is occasional variation in findings. Bi-cortical T2-weighted enhancing lesions have been described on MRI but was not the case here. ¹³ In this case, sinus thrombosis was suggested by the MRI filling defect; however, it was ruled out by further imaging and it also did not fit the clinical picture. Key factors here are relative improvement to extubation within 24 h, a picture of improving cerebral function and lack of eye or facial swelling. ¹⁵

Electroencephalograms are usually unremarkable and certainly not diagnostic. Slow wave activity is often reported but can indicate many different clinical scenarios that are of limited help with diagnosis.

As in this case, the diagnosis is often made on response to steroid management. There are a small number of cases describing plasmapheresis or immunoglobulin treatment when steroid management fails. ⁴ The vast majority of cases do describe improvement or even resolution of symptoms within weeks of immunosuppression treatment with relapse relatively uncommon. ⁶

Aside from the clinical investigation to find the final diagnosis, it is important to mention that the correct diagnosis was made by doing an early internet search. This is not in any way a scientific approach to patient management, but it is important to discuss. The internet when used correctly by the right operator is a powerful aid to help the management of patients. It is of course a double-edged sword in as much that it can be just as harmful in the wrong hands. The suggestion in the literature is that simple or specific diagnosis is much more likely to be accurately found via internet search. Therefore, it is certainly a tool that can be invaluable to moving patient care forward. ¹⁶

Conclusion

On the night this gentleman was admitted, a brief internet search looking for a unifying diagnosis was undertaken by a junior registrar. A focused internet search suggested that HE was a possible cause for all the symptoms, but was almost unanimously dismissed by the clinical team as too unlikely to be possible. However, through a process of elimination, that diagnosis was proven to be correct. The gentleman eventually responded to steroid therapy and IVIG. He is slowly regaining neurological function and remains seizure free. The lesson to be learned here is that the unlikely diagnosis is possible, and although uncommon, it is important to cast a wide net in the search for diagnosis in our more complex patients. Certainly, we do not advocate the use of the internet for diagnosis but openness to unusual pathologies and causes for patients presenting to the ICU should always be considered, however, unlikely they at first seem.