We reviewed the article by [1] which presented a retrospective study of 220 patients with sudden unexpected death in epilepsy (SUDEP) from the North American SUDEP Registry (NASR), concluding that closer collaboration between death investigators and epilepsy communities is necessary to enhance the understanding of SUDEP. While the study is valuable, we raise concerns about the lack of brain autopsy data, which may miss alternative causes of death like ischemic stroke or intracerebral bleeding. Additionally, the absence of family history collection limits the identification of hereditary disorders. Another shortcoming is the lack of consideration of brain-heart connections, such as Takotsubo syndrome (TTS) and neurogenic pulmonary edema (NPE), which may contribute to SUDEP. We suggest comprehensive brain autopsy, detailed family history, and inclusion of TTS and NPE in future studies to improve the understanding of SUDEP mechanisms.

We read with interest the article by Visuttijai et al. (2020) on two unrelated patients with progressive external ophthalmoplegia (PEO) due to novel variants in the MT-TN gene. The authors concluded that these variants are pathogenic and that MT-TN is a mutation hot spot for sporadic PEO. However, we raise several concerns. First, the pathogenicity of the variants was assessed without applying the Yarham score, a standardized method for evaluating mitochondrial tRNA mutations. Our application of the modified Yarham score classifies the variants as only "possibly pathogenic." Additionally, the study focuses solely on extraocular symptoms of PEO, yet both patients exhibited manifestations beyond the eye muscles, suggesting they had PEO plus. We believe the study would benefit from a more thorough investigation of multisystem involvement and further evidence supporting the pathogenicity of the MT-TN variants. Such investigations should include brain imaging, neurologic, and organ system assessments to identify subclinical manifestations typical of PEO plus.

Starting from the time series expressed in the image space, the medical images in the 3D space are uniformly registered to the standard space through the registration conversion algorithm, and then the data in the standard space is fed back to the source image data to achieve image registration. Any images obtained on the same acquisition equipment and the same parameter settings have the same spatial sequence, and the spatial sequence is mapped to a unified spherical standard space to form an image in the standard space to achieve the image registration function. The calculation and realization of the project are based on the Matlab platform as the experimental environment.

With interest we read the article by Edwards et al. about a 19 years old female with Leigh syndrome due to the variant m.9176T>C in MT-ATP6 [1]. The patient experienced sudden onset cerebellar edema leading to posterior fossa compression, tonsillar descent, and death despite immediate surgical decompression [1]. We have the following comments and concerns.

The report has a number of shortcomings. First, it is not mentioned if cerebellar edema occurred bilaterally or unilaterally. Knowing this is crucial for identifying the cause of the edema.

Second, we do not agree with the classification of the disease as “adult” Leigh syndrome [1]. The patient was 19 years old at presentation but is described with long-standing bilateral hypoacusis [1]. Since hearing impairment is a frequent manifestation of a mitochondrial disorder (MID), the initial manifestation of the disease was the hearing problem before age 19 years. Furthermore, the patient was described with bilateral ptosis, which presumably was present also already before age 19 years. Thus, the patient cannot be classified with adult Leigh syndrome, at most as paediatric Leigh syndrome with survival into adulthood.

Third, we do not agree with the classification of the cerebellar edema as vasogenic. The authors only present a CT-scan which shows edema, tonsillar descent, and the emerging hydrocephalus but does not allow classification of the edema. Of interest would be an MRI, to see if there was hyperintensity on diffusion weighted imaging (DWI) and hyperintensity on apparent diffusion coefficient (ADC), which would be compatible with a vasogenic edema. The vasogenic edema is characteristic for stroke-like lesions (SLLs). Additionally, SLLs show up as hyperintensity on perfusion weighted imaging (PWI) and hypointensity on oxygen-extraction fraction MRI (OEF-MRI) [2].

Fourth, the cause of the edema remains unclear. Though we agree that a SLL, as has been previously reported [3], cannot be excluded, various differential diagnoses of a cerebellar edema have not been considered. These include ischemic stroke, cerebellitis, posterior reversible or encephalopathy syndrome (PRES), bleeding, lymphoma, or metastasis. A strong argument against a SLL is that SLLs usually do not go along with edematous expansion of the lesion and do not exhibit a mass effect. 

Missing is a prospective investigation for multisystem involvement [4]. Though the patient was young, affection of organs other than the brain and the muscle is conceivable. Of particular interest is if there was any type of cardiac involvement, such as hypertrophic respectively dilated cardiomyopathy, or noncompaction. Assuming that cerebellar edema was due to an ischemic stroke, it is conceivable that there was spontaneous dissection of vertebral artery or thatthere was embolic occlusion of a vertebral artery secondarily leading to ischemic stroke. Missing in this respect is a description of the intra- and extra-cerebral arteries. An acute ischemic stroke would show up on multimodal MRI as DWI hyperintensity and as ADC hypointensity.

It is unusual that blood gas analysis revealed alkalosis despite lactic acidosis. How do the authors explain this discrepancy? Was alkalosis due to compensatory hyperventilation?

Missing is the lactate level in the cerebrospinal fluid. If increased, it would be an argument in favour of a SLL. Was cerebral lactate elevation documented by MR-spectroscopy?

Overall, this interesting case report would profit from addressing the shortcomings above. More likely than a SLL was an ischemic lesion responsible for the cerebellar edema.

With interest we read the review article by Ellul et al. about neurological abnormalities in SARS-CoV-infected patients [1]. We have the following comments and concerns. 

A neurological disease not addressed in the review is myasthenia. There are several reports indicating that SARS-CoV-2 can deteriorate myasthenia [2] or even trigger myasthenic crises [3]. Since patients with myasthenia are at increased risk due to their mostly immunological disease and immune-modulatory treatment, actions need to be taken to avoid a SARS-CoV-2 infection in these patients.              

Other neurological manifestations not addressed in the review are muscle cramping, tics, and tremor reported in <1% of 917 Chinese SARS-CoV-2-infected [4]. In some patients the virus may also trigger rhabdomyolysis. 

We do not agree that Guillain-Barre syndrome has been reported in only in 19 patients. In a recent meta-analysis, 24 patients with polyradiculitis were presented [5]. Fourteen patients were diagnosed with AIDP, four with AMAN, three with MFS, and two with AMSAN [5]. In none of them was the virus detected in the CSF, why a cross-reaction of the immune-system against the virus and components of the radices or the extensive immunological response to the virus were made responsible [5]. Since this analysis (end of May) at least three more patients have been reported.   

Encephalopathy is a hazy term indicating cerebral disease in general, that should be avoided. In a recent review of SARS-CoV-2-associated meningitis/encephalitis (submitted end of June) it turned out that the majority of the patients with SARS-CoV-2-associated encephalopathy has para-infectious, non-vascular, non-hypoxic, presumably immune-mediated CNS-disease, classified as auto-immune encephalitis (n=11), acute, hemorrhaghic, necrotising encephalopathy (n=2), or acute disseminated encephalomyelitis (n=3). No virus was found in the CSF in any of them. Only in 14/48 patients CNS-disease was classified as infectious (meningitis (n=1), encephalitis (n=5), meningo-encephalitis (n=5), myelitis (n=3)). The virus was present in the CSF in only 4/14 patients. Fifteen/48 patients received virostatics, 21 antibiotics, 11 steroids, 5 immunoglobulins, 7 plasma exchange, 8 antiepileptics, 8 chloroquine, and 11 required mechanical ventilation. Twenty-two patients recovered, 2 did not, and 6 died.

In conclusion, infection with SARS-CoV-2 definitively concerns the neurologist and has therapeutic implications. Peripheral neuronal networks (pulmonary, enteric, pharyngeal, nasal) are affected before the CNS. There are indications that CNS/PNS are more widely and frequently directly or indirectly involved than anticipated. In most cases, SARS-CoV-2-associated CNS/PNS disease is immune-mediated. Thus, in most of these patients immuno-suppressive / immune - modulatory treatment, immunoglobulins, or plasma exchange should be considered.

With interest we read the article by Shand, J. A. et al., [1] about renal involvement in monozygotic twins carrying the mtDNA variant m.3243A>G at heteroplasmy rates 20%/40% (twin-1) respectively 10%/40% (twin-2) in blood lymphocytes respectively buccal mucosa [1]. Variable onset of end-stage kidney disease (ESKD) in both patients was attributed to their different heteroplasmy rates [1]. We have the following comments and concerns.

Heteroplasmy rates from blood lymphocytes/buccal mucosa (not affected) are not representative for disease severity in clinically affected organs (kidneys) [2]. More appropriate than heteroplasmy rates from a tissue not affected, such as lymphocytes, would be urinary epithelial cells or renal tissue. We should know heteroplasmy rates in both twins from urinary epithelial cells and in the explanted kidneys. In patients with renal involvement in a mitochondrial disorder (MID) due to the variant m.3243A>G heteroplasmy rates in the kidneys may as high as 89% [3]. 

Phenotypic expression, and thus disease severity, progression, and outcome, may not only depend on the heteroplasmy rate but also on mtDNA copy number [4]. We should know if mtDNA copy number was reduced (mtDNA depletion) or not in index patients.

Since the grandfather of the index patients had deafness but not the grandmother and since two uncles had deafness as well, it is more likely that the phenotype of the index patients is not only attributable to the mtDNA variant alone but to a second mutation in a yet unidentified nuclear gene with autosomal dominant inheritance. Thus, genetic work-up should be expanded to whole exome sequencing (WES). 

We do not agree with the statement that “The level of organ impairment relates to its metabolic demand, rate of cell turnover, and individual tissue-level mutation load, or heteroplasmy” [1]. Other factors that determine the phenotype include mtDNA copy number, haplotype, polymorphisms, drugs, oxidative stress, diet, and environmental toxins. Renal function may also depend on drinking habits. Were they the same in both twins?

The m.3243A>G variant may manifest in the kidneys not only as focal segmental glomerulosclerosis (FSGS) or as ESKD [1]. but also with nephrolithiasis [5], granular swollen epithelial cells [6], macroalbuminuria ([7], interstitial fibrosis [3], renal insufficiency [8], non-specific glomerular changes [9], abnormal mitochondria in podocytes and tubular epithelial cells [10], or renal cysts [11].

The variable phenotypic expression even in monocygotic twins can be explained by the bottleneck effect and tissue segregation. The copy number of mutated mtDNA may be different in both halfs of the fertilised oocyte after the first division and mutated mtDNA may be distributed to different progenitor cells in both twins.

We do not agree with the statement in the abstract that “The mtDNA mutation mt.3243A>G is most commonly associated with maternally inherited diabetes and deafness” [1]. The syndrome most frequently associated with the m.3243A>G variant is mitochondrial encephalopathy, lactic acidosis, and stroke-like episodes (MELAS) [12].

Overall, this interesting study has a number of shortcomings, which need to be addressed before conclusions as those presented, can be drawn. 

Heteroplasmy rates in urinary epithelial or renal cells need to be determined, mtDNA copy numbers need to be presented, all factors influencing disease severity should be considered, and heteroplasmy rates and mtDNA copy numbers should be determined in other first degree relatives.

With interest we read the article by Lopes et al. about a 3 years old female infant with Leigh syndrome due to the variant m.8993T>G (heteroplasmy >90%), with disease onset at age 8 months with generalised hypotonia and psychomotor retardation [1]. The condition deteriorated suddenly one day after performing a cerebral MRI under anesthesia when the patient developed hyporeactivity, aggravated hypotonia, exaggerated tendon reflexes, and cutaneous pallor [1]. Subsequently a transient, self-limiting episode of apnea and cyanosis with somnolence and increased prostration occurred [1]. Phenytoin was begun. Serum lactate and citrulline were elevated and the MRI showed T2-hyperintesities of the putamen bilaterally [1]. She later developed ataxia, dystonia, and epilepsy and received phenobarbital (PB), levetiracetam (LEV), and a vitamin cocktail [1]. We have the following comments and concerns.

Patients with a mitochondrial disorder (MID) may develop adverse reactions to drugs for general or local anesthesia or for muscle relaxation [2,3]. Thus, we should be informed which drugs were applied for anesthesia prior to the MRI and if deterioration could be attributed to side effects of these agents. Deterioration of the phenotype after anesthesia was more likely due to side effects to the drugs used for anesthesia than to metabolic stress induced by the anesthetic procedure, an infection, vaccination, or fasting, as suspected by the authors. Generalised anesthesia is usually well tolerated by MID patients if mitochondrion-toxic agents are avoided. 

The subsequent episode of apnea was obviously interpreted as epileptic seizure in the absence of epileptiform discharges on electroencephalography (EEG) as the patient received phenytoin (PHT) [1]. From PHT it is well known that it can be mitochondrion toxic [4], why it should not be applied as first line medication for mitochondrial epilepsy. We should know if PHT was replaced by PB and LEV because of side effects or because of ineffectivity. Furthermore, the therapeutic range of PHT is narrow and the majority of patients has serum levels below or above the therapeutic range [5]. There are also concerns regarding the application of PB to patients with a MID [4]. A further option for the anti-seizure management of mitochondrial epilepsy is the ketogenic diet (KD) [6].

Leigh syndrome may not only be associated with elevated lactate in the serum but also in the cerebrospinal fluid (CSF) [7]. Thus, we should know if the patient was investigated for concentrations of CSF lactate or if MR-spectroscopy (MRS) was carried out to document an increased lactate peak. Deterioration of the clinical condition may be attributed to lactic acidosis, why we should know if serum/CSF lactate was determined prior to the deterioration and if serum/CSF lactate was normal before. 

Since patients with Leigh syndrome may not only present with involvement of the basal ganglia but also with involvement of the cerebellum [8], we should know if ataxia was rather attributed to a cerebellar than to a basal ganglia lesion. In this respect it would be interesting to know if the patient had dysarthria, tremor, dysmetria, or nystagmus.

Since the variant m.8993T>G was obviously inherited from the mother (heteroplasmy rate: 75%) we should know if the mother manifested clinically or subclinically, or not. 

Overall, this interesting report could be more meaningful by providing supplementary information and discussion of unsolved issues, such as the trigger of the initial deterioration, toxic effect of PHT and PB, cause of ataxia, and presence of elevated lactate in the CSF. The mother should undergo a detailed neurologic exam.

With interest we read the article by Sanderson, K. G. et al., [1] about a 14yo Indian male with sensory ataxic neuropathy, dysarthria, and ophthalmoplegia (SANDO) plus syndrome due to the previously reported variant c.911T>G in POLG1, who was described as the first documented case with generalised rod ON-bipolar dysfunction but electronegative electroretinography (ERG) [1]. We have the following comments and concerns. 

We do not agree with the diagnosis of SANDO [1]. SANDO is characterised by sensory neuropathy, dysarthria, and ophthalmoplegia. However, the index patient had cyclic vomiting syndrome, anorexia, developmental delay (speech at age 2y), epilepsy (treated with oxcarbazepine (OXC)), myopathy affecting the extra-ocular, limb and respiratory muscles (ophthalmoparesis, ptosis, quadruparesis with contractures, respiratory insufficiency), sleep apnoea syndrome, and dysphagia requiring nutrition via a G-tube [1]. The patient did not fulfil the diagnostic criteria for SANDO, since he was not described as having ever developed the cardinal phenotypic features sensory neuropathy and dysarthria [1]. We should know the results of the clinical neurological exam at the last follow-up and the results of nerve conduction studies. 

POLG1 mutations may secondarily cause reduction of the mtDNA copy number (mtDNA depletion) [2]. Thus, we should know if there was mtDNA depletion and to which degree the mtDNA copy number was reduced.   

There is a discrepancy between the description of the phenotype in the case report section and the discussion. In the case report section nothing is mentioned about headache or hypotonia. Surprisingly, the discussion indicates that the patient additionally had sporadic headache since age 4y, and hypotonia from the age of 4y.   

The patient was diagnosed with sleep apnoea syndrome (SAS). However, the patient was reported to have experienced weight loss and affection of the respiratory muscles. Is it conceivable that SAS was rather chronic respiratory failure due to affection of respiratory muscles? We should know the results of sleep laboratory findings and of cerebral MRI to see if there was brainstem involvement in the disease.

The patient presented with chronic coughing. We should know the cause of chronic coughing. Is micro-aspiration conceivable or was it due to heart failure? Since POLG1-related disorders may go along with cardiomyopathy [3], it is crucial that the results of echocardiography and long-term ECG recordings are presented. Additionally, we should know the serum levels of creatine-kinase, lactate, and proBNP. 

Since POLG1 variants may manifest with optic atrophy [4], we should know the results of visually-evoked potentials (VEPs) and the magnetic resonance imaging (MRI) of the optic nerve. Though only rarely reported, POLG1 variants may also go along with retinopathy, particularly pigmented retinopathy [5].

OXC is potentially mitochondrion-toxic and may have an adverse effect on the mitochondrial metabolism [6]. We should know the daily dosage and serum levels, and if any side effect developed. The patient additionally received coenzyme-Q (CoQ), lipoic acid, and L-carnitine since the age of 9y [1] but the mitochondrial disorder (MID) was diagnosed not earlier than at age 10y. Which was the rationale for prescribing antioxidants and cofactors in the absence of a documented MID at that time?             

Overall, this interesting case of a POLG1-related mitochondrial disorder (MID) has a number of shortcomings, which need to be addressed before drawing final conclusions. The patient had no SANDO but rather an early-onset, non-syndromic, multisystem, POLG1-related MID, manifesting in the brain, skeletal muscle, intestines, and the eyes.

With interest we read the article by Rein, N. et al., [1] about the management of three patients with seropositive myasthenia gravis, who got infected with SARS.CoV-2 [1]. Only one of these patients developed symptomatic SARS-CoV-2 infection (COVID-19) and received appropriate treatment [1]. This patient also experienced deterioration of myasthenic manifestations during the viral infection [1]. Interestingly, the relapse in this patient started already prior to the infection with SARS-CoV-2, suggesting that COVID-19 had no adverse effect on myasthenia [1]. In two of the myasthenia patients the SARS-CoV-2 infection remained asymptomatic and both did not experience a relapse of myasthenia, why they required neither a treatment for COVID-19 nor additional treatment for myasthenia [1]. We have the following comments and concerns.

Patient-1 was treated with an increase in prednisone and additionally received lopinavir, ritonavir, and chloroquine. Lopinavir is an anti-retroviral agent applied in the treatment of HIV-infected patients. Anti-retroviral agents generally inhibit the uptake of L-carnitine [2]. The myotoxic effect of anti-retroviral agents has been also attributed to suppression of the fatty acid oxidation and to impaired fatty acid handling and partioning in myocytes [3]. Particularly in combination with other myotoxic drugs, lopinavir may cause myopathy and rhabdomyolysis [4]. Another side effect of the anti-retroviral treatment can be hyperthyroidism [5]. Hyperthyroidism on the other hand can cause hypokalemic thyreotoxic periodic paralysis, which may mimic deteriorations of myasthenia [5]. Thus, we should know if thyroidea-stimulating hormone levels were reduced or not in patient-1. Also, ritonavir may cause severe myopathy [6].

Chloroquine commonly prescribed for inflammatory arthritis, can be myotoxic as well. Chloroquine myopathy manifests clinically as proximal muscle weakness and can be associated with neuropathy and cardiomyopathy [7]. Muscle biopsy consistently reveals curvilinear bodies and muscle fiber atrophy with vacuolar changes [7]. In rare cases, chloroquine may cause vacuolar myopathy [8]. Potential predisposing factors include Caucasian race and concomitant renal failure [7]. Resolution of chloroquine myopathy is slow after discontinuation of therapy and may be incomplete [7]. 

It is unusual that patient with a long-term course of myasthenia receive steroids as a long-lasting therapy. From prednisone it is well-known that it triggers the development of a mitochondrial myopathy, and thus should be given for years but only as a bridging therapy until immunsuppressants exhibit their anti-myasthenic effect. Steroids should be given in long-term myasthenia patients only during acute exacerbations or for myasthenic crises.

The authors regard ptosis, respiratory insufficiency, and proximal limb muscle weakness in patient-1 as clinical manifestations of myasthenia triggered by the SARS-CoV-2 infection. We should know if the decremental response, jitter, and titers of acetyl-choline-receptor antibodies truly increased during the SARS-CoV-2 infection. If these parameters did not worsen, worsening of muscle manifestations can be rather attributed to myotoxic side effects of the anti-SARS-CoV-2 treatment than to myasthenia.

Particularly the combination of various myotoxic drugs (anti-retro-viral agents, chloroquine, steroids) could lead to clinical manifestations of a myopathy. Renal or liver insufficiency may predispose for the occurrence of these myotoxic effects. In this respect we should know which other drugs the three included patients were regularly taking and if patient-1 had any indications for renal or liver failure. We should know the entire medication to assess if the clinical manifestations in patient-1 were due to myasthenia or due to myotoxic side effects of any drugs.   

Concerning the deterioration of muscle manifestations in patient-1 it should be considered that SARS-CoV-2 itself my cause myopathy [9], manifesting as myopathy, creatine-kinase elevation, rhabdomyolysis, or muscle weakness [10]. Deterioration of muscle manifestations in patient-1 prior to SARS-CoV-2 may, nonetheless, being attributed to the infection, which may have gone subclinical in its early stages. 

Overall, the study about three cases with myasthenia and SARS-CoV-2 infection reported by Rein et al. has some shortcomings, which should be addressed before drawing final conclusions. All drugs the three patients were taking before and during the infection need to be mentioned, an explanation for the chronic use of prednisone should be provided, and myopathy due to the myotoxic effect of anti-SARS-CoV-2 medication should be considered as cause of the deterioration of muscle manifestations in patient-1.

With interest we read the article by Reed, E. M. et al., [1] about a 25yo male who was diagnosed with ischemic stroke upon the clinical presentation (acute aphasia) and a cerebral computed tomography (left temporo-parietal subacute lesion) [1]. Additionally, his history was positive for diabetes, attention-deficit/hyperactivity disorder, and pervasive development disorder [1]. The patient was later diagnosed with mitochondrial encephalopathy, lactic acidosis, and stroke-like episode (MELAS) syndrome upon lactic acidosis, recurrent seizures, laminar cortical necrosis, and documentation of the m.3243A>G variant in MT-TL1 [1]. Nine months after diagnosing MELAS the patient was found dead on the floor and the cause of death was attributed to ketoacidosis after forensic autopsy [1]. We have the following comments and concerns.

We do not agree with ketoacidosis as the cause of death in this patient. MELAS is a multisystem disease, affecting not only the brain and the spinal cord, but also the eyes, ears, vestibulum, endocrine organs, the gastrointestinal tract, and the heart. More rarely, the lungs, kidneys, bone marrow, skeleton, and the skin may be affected. Differential diagnoses that should be considered as a cause of death in addition to ketoacidosis include sudden unexplained death in epilepsy (SUDEP), Takotsubo syndrome (TTS), cardiac arrest, ventricular arrhythmias, drug side effects or drug interactions, and pulmonary embolism. High vitreous glucose and β-hydroxy-butyrate not necessarily imply derailed diabetes. These elevated parameters could also indicate extreme stress. We should know the last HbA1c value and the last antidiabetic medication prior to decease.

Since the patient had epilepsy, we should know the types of seizures, the seizure frequency, the EEG results, the anti-seizure drug (ASD) regimen, the adherence to ASD treatment, and the ASD effect. Since stroke-like episodes (SLEs) frequently go along with either seizures or epileptiform discharges on EEG [2], we should know the results of electroencephalography (EEG) during hospitalisation for the “ischemic” stroke. With regard to TTS we should know if post-mortem catecholamine levels were determined. TTS is characterised by a massive elevation of serum catecholamines (catecholamine storm) [3]. With regard to cardiac arrest and ventricular arrhythmias we should know if ECG, long-term ECG, and echocardiography prior to decease were normal and if the history was positive for palpitations, syncopes, or heart failure. Was pulmonary embolism excluded on autopsy?

A shortcoming of the study is that the “stroke” was diagnosed only upon CCT but not upon multimodal cerebral MRI. Most likely the initial stroke was not ischemic in nature but a so called stroke-like episode (SLE), which is the hallmark of MELAS but may occur also in other MIDs [4]. The morphological equivalent of a SLE on cerebral MRI is the so called stroke-like lesion (SLL). A SLL is characterised by T2-, DWI-, PWI- hyperintensity, and hypointensity on oxygen-extraction fraction MRI, which is not confined to a vascular territory [4]. SLLs show a progressive course with regard to intensity and extension and end up as normal brain tissue, cysts, white matter lesion, toenail sign, or laminar cortical necrosis [5]. SLLs may recur in the same or other locations, supra- or infra-tentorially. A strong argument in favour of a SLL in the index patient is that the MRI two months after the event showed laminar cortical necrosis. We should know if the onset of aphasia was truly acute or subacute. 

Missing is the medication the patient was regularly taking prior to decease. Missing are the post-mortem creatine-kinase and lactate values.

Overall, this interesting case could profit from a discussion about alternative causes of sudden death including SUDEP, TTS, ventricular arrhythmias, cardiac arrest, and pulmonary embolism. High vitreous glucose and β-hydroxy-butyrate not necessarily imply ketoacidosis from diabetes.