Acute stroke-like deficits associated with nonketotic hyperglycemic hyperosmolar state: an illustrative case and systematic review of literature

Nonketotic hyperglycemic hyperosmolar state (NKHHS) is a life-threatening condition characterized by severe hyperglycemia, hyperosmolarity, and dehydration, without significant ketoacidosis. Despite a low estimated incidence, NKHHS accounts for 1% of all hospital admissions in patients with diabetes, and carries an overall mortality rate of up to 20% [1]. NKHHS results from an imbalance between insulin demand and availability, with enough insulin to inhibit free fatty acid mobilization and ketogenesis, but not enough to enhance glucose transport into cells. Severe hyperglycemia with glycosuria in turn provokes osmotic diuresis and progressive glial dehydration [2]. Neurological manifestations associated with NKHHS encompass a wide range of syndromes, spanning from seizures (15–20% of cases, usually focal motor) to hyperglycemic nonketotic hemichorea (“diabetic striatopathy”) and alterated levels of consciousness [2, 3]. Rarely, NKHHS can present with acute neurological deficits mimicking stroke [4]. The pathophysiologic mechanism of negative symptoms during NKHHS is still unclear. An ictal or post-ictal phenomenon (Todd’s palsy-like) has been speculated [5]. It has been suggested that a metabolic disruption of neuronal function could lead to the occurrence of neurological deficits in patients with NKHHS. This hypothesis is supported by many pieces of evidence, including the occurrence of neurological deficits even in patients with normal electroencephalogram (EEG), rarely persisting after NKHHS resolution [6‐8], and the identification of peculiar magnetic resonance imaging (MRI) alterations [9].

Here we report an illustrative case of NKHHS presenting with reversible total anterior circulation infarct syndrome (OCSP-TACI, according to Bamford classification [10]), without evidence of underlying stroke or epilepsy. We also provide a systematic review of the current available literature on acute stroke-like deficits in NKHHS, in order to inform clinicians on peculiar features and work-up findings.

A 77-year-old woman with grade III obesity (body mass index 41 kg/m²) and a past medical history of poorly controlled adult-onset diabetes mellitus (AODM), hypertension, and hyperlipidemia was referred to our emergency department (ED) for the acute onset of left limb weakness shortly after the awakening (at about 6:00 AM). The neurological examination (9:00 AM) illustrated mild drowsiness, left side severe hemiparesis and sensory loss, left homonymous hemianopia, right gaze deviation, and dysarthria (National Institutes of Health Stroke Scale, NIHSS 27). These findings were consistent with a total anterior circulation infarct syndrome (TACI). Her blood pressure was 220/110 mmHg and did not decrease despite repeated administration of IV alpha blockers. A brain computed tomography (CT) was unremarkable for ischemic/hemorrhagic lesions and a CT angiogram excluded large vessel occlusions. Severe hyperglycemia (524 mg/dL) was found. Considering her pre-morbid functional status (modified Rankin Scale—mRS 3 for severe obesity with orthopedic comorbidity) and the elevated blood pressure and glucose levels, the patient was excluded from urgent revascularization with intravenous thrombolysis. Medical treatment was immediately started with intravenously hydration and 24-h infusion of urapidil and insulin. On day 1, the patient became less somnolent and complained of seeing flickering lights and criticized well-structured images (e.g., her daughter) appearing on her left visual field, which was normal when tested. A second brain CT was still unremarkable, while an EEG demonstrated right hemispheric slow wave activity, without epileptiform discharges (Fig. 1). A transcranial Doppler (TCD) was unremarkable. Over days 2 and 3, paralleling normalization of blood glucose levels, the patient gradually started to move her left limbs. Right gaze paralysis slowly resolved. Brain MRI performed at day 3, when neurological examination was normal, revealed the presence of disseminated right frontal cortical lesions with increased signal on diffusion-weighted imaging (DWI) sequences and reduced apparent diffusion coefficient (ADC). T2-weighted imaging and fluid-attenuated inversion recovery (FLAIR) showed ipsilateral subcortical hypointensities (Fig. 2). These alterations suggested an injury of the anterior cerebral regions induced by severe hyperglycemia [6]. Three days after symptom onset, her neurological status completely normalized (NIHSS 0). She was discharged a week after being admitted, without any neurological deficit. The timeline of events and significant investigations is shown in Fig. 3.

We identified a total of 37 articles published from 1968 to 2021 (Fig. 4), including 49 patients with NKHHS and ≥ 1 acute neurological defect. Thirty-one patients from 22 articles were excluded for concomitant epileptic activity at the EEG (n = 16) or for its unavailability (n = 15). Overall, 19 patients (18 from 15 articles [3, 6‐20] and our case) were included in the analysis (Table 1).

We reported the case of a patient presenting with NKHHS and clinical symptoms mimicking acute TACI syndrome, which resolved completely after the correction of hyperglicemia. Besides the very high levels of blood glucose and osmolarity, key findings of our case included the observation of MRI alterations suggestive for NKHHS and the absence of epileptic abnormalities at the EEG. Given the clinical presentation resembling acute ischemic stroke with negative brain CT, the urgent diagnosis of NKHHS is often challenging, but it should always be considered in patients with either unknown or poorly controlled AODM.

The etiology of negative neurological symptoms in patients with NKHHS is still debated. In particular, both ischemic and metabolic pathogeneses have been hypothesized (Fig. 5). Pivotal studies on animal models showed that both chronic and acute hyperglycemia may provoke a global decrease in cerebral blood flow (CBF), by enhancing focal or diffuse vasoconstriction [21]. This hypothesis is supported by the observation of increased peak systolic and mean flow velocities [22‐24], along with impaired cerebral autoregulation [25], in patients with diabetes or acute glucose exposure. However, perfusion or TCD studies have not been systematically performed in patients with NKHHS-related neurological manifestations. In our analysis, brain perfusion CT/MRI showed findings consistent with ischemic penumbra in congruous brain areas of 3 patients, while in a single case TCD disclosed a mild increase of the MCA mean flow velocity [6]. Anyway, in our case TCD was unremarkable, suggesting that neuronal impairment could not be explained by a pure hemodynamic basis.

Alternatively, a metabolic etiology is suggested by the frequent finding of peculiar MRI abnormalities [9]. These include the presence of subcortical T2-FLAIR hypointensities, which represent the most frequent alteration (100% of patients), and of disseminated cortical lesions with increased DWI signal (64%) and restricted ADC. T2-FLAIR hypointensities have been related to free radical damage, iron accumulation, and early cortical ischemia [26‐28], which can be promoted by either hyperosmolarity or hyperglicemia. In particular, altered mechanics of intracellular enzymes, prompted by osmotic dehydration, may generate reactive oxygen species [28]. Concomitant acidosis may reduce transferrin affinity, thus increasing unbound iron deposition [28]. Increased signal on DWI sequences indicates cytotoxic edema, potentially related to ischemia, but it may also reflect osmotic damage or hyperviscosity [6]. At a metabolic level, in NKHHS, gamma aminobutyric acid (GABA) is metabolized to succinic acid via the succinic acid semialdehyde pathway (GABA shunt), in order to provide intermediates for the Kreb’s cycle and provide energy [29]. GABA is the main inhibitory transmitter in the central nervous system and its depletion has been related to “positive” symptoms in NKHHS (e.g., focal seizures [29]) (Fig. 5). The observation of dramatically increased peaks in cerebral metabolites at MRS, without changes in lactate levels (typical of ischemic stroke), suggests that diabetes-related alterations may play a major role in inducing neuronal dysfunction. However, MRS was performed during symptoms only in 2 patients with NKHHS-related focal deficits, and larger studies are needed to confirm these findings.

In our analysis, the most frequent neurological deficit was homonymous hemianopia, which was reported (alone or in association with other symptoms) in 73% of patients. This result, along with the more frequent observation of MRI abnormalities in patients with hemianopia (91%) vs. anterior involvement (20%), may reflect an intrinsic vulnerability to hyperosmolarity of posterior brain areas. Two mechanisms could account for this phenomenon: (a) a lower density of sympathetic fibers regulating vasomotricity [30] and/or (b) a higher permeability of the blood–brain barrier (BBB) [31]. Indeed, a less pronounced capacity of blood vessels to adapt their supply in response to metabolic stress may foster neuronal dysfunction by provoking transient hypo or hyperperfusion, as occurs in posterior reversible encephalopathy syndrome (PRES) [32]. Inefficient sympathetic response might be enhanced by concomitant dysautonomia, which frequently complicates diabetes mellitus [33]. Additionally, the expression of tight junctions has been demonstrated to vary across different brain areas, potentially influencing BBB endurance to hyperosmolarity [34, 35]. The observation of Gd enhancement (an MRI surrogate of BBB disruption) only in NKHHS patients with posterior involvement might support this latter mechanism.

Anterior involvement in NKHHS seems less frequent and results more often in symptoms consistent with partial anterior circulation infarct syndrome (26% of all patients with negative deficits). Speech disturbances, and especially aphasia, were the most frequent symptoms in this subgroup (100%), probably reflecting a higher metabolic demand of related brain areas [36]. We found slightly higher blood glucose levels in patients with anterior circulation involvement, whereas statistical significance was not reached. Whether this observation will be confirmed by larger cohorts, it might imply that a higher degree of metabolic disruption is necessary to determine an impairment of anterior brain areas, perhaps for more abundant collaterals or for vasomotor differences, as discussed above. Furthermore, we found that patients with symptoms suggestive of anterior circulation infarct syndrome were significantly older compared to those with hemianopia, possibly reflecting age-related differences in brain microvasculature and cerebral autoregulation [37]. Visual hallucinations (VHs) were almost constant in patients with hemianopia (100% of cases from literature). Of interest, our patient was the sole without MRI or clinical signs of posterior involvement, thus suggesting a subtle dysfunction of occipital areas.

EEG abnormalities were common (68%), especially in patients with anterior involvement, and consisted mainly in focal or diffuse slow wave activity. Reported patterns were unspecific, and tended to be more diffuse in patients with higher blood glucose, suggesting a more severe metabolic impairment of the neuronal electrogenesis. We found a high prevalence (85%) of VHs in patients with EEG evidence of slow wave activity. VHs have been associated with the presence of occipital epileptiform discharges in some cases of NKHHS [38]. However, their frequent occurrence in the hemianopic visual field of patients without concomitant epileptic abnormalities, along with their well-criticized nature, may suggest a metabolic deficit of the cortical modulatory control (functional differentiation) with consequent abnormal activation of connected occipital areas, similarly to Charles Bonnet syndrome [39].

Prognosis of NKHHS is generally favorable, with complete recovery in the majority of cases (78%), following aggressive treatment leading to glucose and hosmolarity normalization. The persistence of visual symptoms has been reported in only 4 cases, all presenting with hemianopia, abnormal EEG, and MRI findings. A patient with persisting (although improved) hemianopia developed temporal-occipital atrophy at a brain MRI performed after 1.5 year [14]. Other two patients without persisting symptoms showed focal cortical gliosis and mild white matter loss at follow-up MRI [9]. Taken together, these findings may suggest that neuronal dysfunction could revert upon correction of the metabolic trigger, or result in irreversible damage once a critical threshold is reached. However, a follow-up MRI was performed only in 35% of patients and the timespan from symptom onset was highly variable (3 weeks–3 years). Therefore, the exact evolution of cerebral injury in NKHHS remains largely unknown. Additionally, the small number of patients with symptoms persistence and altered follow-up MRI does not enable to identify any reliable predictor of irreversible damage.

Our study has some relevant limitations. First, the small review sample, reflecting the rarity of NKHHS-related acute neurological deficits, does not allow us to draw firm conclusions on the etiology and clinical picture. Second is the inclusion of data accruing only from retrospective case reports or series, with limited impact deriving from publication bias.

Neurological deficits represent a rare but clinically relevant consequence of NKHHS which can raise important diagnostic and therapeutic challenges. The clinical spectrum of NKHHS encompasses symptoms mimicking more frequently posterior circulation stroke, although the involvement of anterior brain areas is not uncommon. The exact etiology of acute focal deficits in NKHHS is still to be completely characterized, but it may rely on metabolic alterations that are driven by hyperosmolarity. Brain MRI may provide suggestive findings, while EEG abnormalities are unspecific. Prognosis is generally favorable, whereas irreversible damage may develop in a minority of cases, especially with late diagnosis. Larger cohorts are needed, possibly extracted from existing AODM registries, to verify our findings. Future studies should also investigate diagnostic and prognostic predictors in NKHHS, potentially enhancing the early initiation of proper treatment and an accurate risk stratification. From this perspective, the longitudinal measurement of blood biomarkers related to neuroaxonal and astroglial damage (e.g., neurofilament light chain protein, glial fibrillary acidic protein) could help to clarify some pathogenic and prognostic aspects. To date, the achievement of an optimal glycemic control in patients with AODM remains the best strategy to prevent severe complications, including neurological symptoms associated with NKHHS.