Gamma-glutamyl transferase levels are associated with the occurrence of post-stroke cognitive impairment: a multicenter cohort study

Globally, stroke is a leading cause of disabilities and mortalities, affecting one in every four people [1‐3]. Cognitive impairment, which is a common stroke complication, has attracted numerous research attention. According to the Vascular Impairment of Cognition Classification Consensus Study (VICCCS), vascular cognitive impairment (VCI) refers to cognitive disorders caused by underlying vascular factors and can be associated with obvious cerebrovascular diseases [4]. Vascular cognitive impairment no-dementia (VCI-ND) is based on the proposed criteria of small vessel ischemic disease and cognitive deficits without dementia [5]. Furthermore, vascular depression can lead to decreased intracortical facilitation and disruption of glutamate neurotransmission, which plays a major role in synaptic plasticity, and might contribute to the cognitive deterioration [6‐8]. These cognitive and mood symptoms are associated with vascular damage in the white matter connecting the prefrontal cortex and basal ganglia as well as those connecting the prefrontal cortex and cerebellum. Transcranial doppler ultrasound revealed a hemodynamic pattern of cerebral hypoperfusion and increased vascular resistance [9]. Cognitive impairment manifests as memory decline, abstract thinking, and judgment impairment, but, the ability for daily life is normal. However, it presents a higher risk for more severe cognitive impairments, especially after recurrent strokes, which can seriously affect a patient’s quality of life. As a subtype of VCI, post-stroke cognitive impairment (PSCI) emphasizes that stroke events trigger cognitive dysfunction. Approximately 50% of stroke survivors manifest cognitive dysfunctions, 6 months after stroke, and are more likely to develop dementia within the following 3 years, which significantly affects their quality of life [10, 11]. Moreover, a community-based epidemiological survey in China reported that incidences of PSCI and dementia were 56.6 and 23.2%, respectively, 3 months after stroke [12].

Currently, the diagnosis of PSCI is mainly based on clinical manifestations and on structural changes in brains of patients. This diagnostic criteria formed the basis for construction of SIGNAL₂ and CHANGE risk models [13, 14]. The Leukoaraiosis and Disability Study (LADIS) revealed that the severity of changes in white matter is associated with worse performances on overall cognitive tests [15]. Alterrations in mean diffusivity of normal-appearing white matter, corpus callosum atrophy, the presence of lacunes in the thalamus, gray matter, and hippocampal volumes are significantly associated with speed, memory performance, and executive functions. Combined measurement of these imaging metrics can be used as a comprehensive neuroimaging marker for predicting vascular cognitive impairment [16‐19]. However, the use of biomarkers for the diagnosis and prognosis of PSCI remains a challenge [20, 21].

Gamma-glutamyl transferase (GGT) is a serum metabolic biomarker that is mainly used to assess liver function [22, 23]. GGT is involved in maintenance of physiological concentrations of glutathione in cells and reflects the oxidation-antioxidant balance in the body [24, 25]. It has been reported that GGT levels are correlated with decreased cognitive function in diabetics [26, 27]. In addition, a Korean retrospective study found that GGT variability is associated with Alzheimer’s disease, implying that serum GGT levels are potential predictors of cognitive decline [28]. Moreover, serum metabolites, including GGT, have been shown to be differentially expressed in patients with PSCI and post-stroke non-cognitive impairment [29, 30], suggesting that GGT may affect PSCI occurrence.

However, the role of GGT in PSCI has not been conclusively determined, and to date, only a handful of models for predicting PSCI have been constructed. Notably, these models are mainly constructed based on cerebrovascular risk factors, with the effects of non-cerebrovascular risk factors on PSCI remaining unclear. Therefore, the relationship between GGT and PSCI should be evaluated further. In addition, expert consensus states that the diagnosis of PSCI refers to cognitive dysfunction after a stroke event in 6 months, and most patients suffer cognitive impairment within 3 months after stroke [4‐6]. Therefore, we aimed to investigate the association of serum GGT with PSCI during 3 months of follow-up. This study is presented in accordance with the STROBE reporting checklist.

This large prospective cohort study demonstrated that baseline GGT levels were inversely associated with PSCI occurrence. Specifically, extremely low GGT levels were established to be risk factors for PSCI, even after adjusting for confounding factors including age, sex, educational level, smoking, drinking, BMI, some laboratory indicators, and medical history. Interestingly, the incorporation of GGT into the conventional model resulted in an 11.87% increase in predicting PSCI. Furthermore, PSCI showed a stronger inverse association with GGT especially in individuals with a lower BMI. Correlations between GGT and PSCI in other subgroups revealed no significant change after testing for interactions. This finding indicates that GGT exerted a consistent effect on PSCI, regardless of patients’ age, sex, alcohol drinking habits, and stroke type [48].

As a common complication after stroke, PSCI is associated with serious disabilities. Studies have found that a variety of serum biomarkers are associated with PSCI. Moreover, previous studies reported contrasting findings in terms of the relationship between GGT and cognitive impairment. However, studies have not evaluated the role of GGT in PSCI. According to previous studies, oxidative stress is one of the pathogenic mechanisms of PSCI [30, 49]. After cerebral ischemia and hypoxia, endogenous antioxidants are decreased and oxygen free radicals are overproduced during perfusion of low cerebral blood flow. The body’s oxidative and antioxidant systems are out of balance. Free radicals lead to cell death by damaging proteins, fats, and DNA, which in turn leads to systemic vascular endothelial dysfunction, increases the permeability of the blood-brain barrier and leads to extravasation of blood substances and leakage of serum proteins. These abnormalities are thought to lead to subsequent neuronal damage, such as grey matter atrophy and cortical thinning, leading to cognitive dysfunction [50]. GGT as a biomarker reflecting the oxidation-antioxidant balance in the body should be paid attention to in PSCI related studies. In this multicenter cohort study, we established an association between GGT and PSCI.

This result may be explained by the following mechanism. Biologically, GGT is critical for antioxidant defenses [18]. It is involved in the maintenance of physiological concentrations of glutathione and plays a vital role in protecting cells from oxidative stress damage. GGT induction can be used as a protective adaptation mechanism in physiological and pathological processes [51]. In the initial development of stroke, inflammatory cytokines levels in the body increase, and GGT levels can compensatory increase when catabolism of inflammatory cytokines containing glutathione. In this process, the glutamic acid and the strong reducing agent (dipeptide cysteinyl glycine) are produced. The latter is hydrolyzed by dipeptidase to cysteine and glycine, which are then taken up by cells for intracellular glutathione resynthesis [22, 25]. Among the catabolic products mentioned above, glutamate can be used as the energy material of brain tissue to improve and maintain neurological function. Both glycine and cysteine are constituent amino acids of endogenous antioxidant reduced glutathione, which can protect nerve cells from oxidative stress and reduce the oxidative damage caused by Aβ deposition. The amino acid neurotransmitter is an important transmitter system in the brain. GGT is thought to contribute to the transport process of amino acids across the blood-brain barrier due to the tight junctions of the endothelium cells that prevent the free diffusion of substances [52, 53]. In the brain, GGT mainly exists in the microvascular endothelial cells and in the choroid plexus where the blood-cerebrospinal fluid barrier exists. This enzyme plays a role in regulating the uptake and transport of amino acids, facilitating amino acid transport across the blood-brain barrier and intracellular glutathione regeneration. GGT has been shown to have a protective effect on brain cells. The main deficiency of the vascular endothelial barrier is its inability to prevent the entrance of lipophilic xenobiotic substances, while the intracellular glutathione synthesis catalyzed by GGT can detoxify such substances. Therefore, GGT plays an important role in defending cells from oxidative-induced damage. Moreover, increased GGT levels in the normal range can also indicate that the liver is better at dealing with oxidative stress. When the degree of oxidative stress in the body decreases, PSCI incidence decrease accordingly. Notably, very low GGT levels are indicators of poor liver functions. In patients with more severe chronic liver disease, especially in advanced cirrhosis, the GGT level is continuously at a low value, possibly due to the loss of glutamyl transpeptidase synthesis in hepatocytes. Deficiency or absence of GGT leads to impaired glutamate cycle, which affects the absorption, transport, and utilization of amino acids causes glutathione resynthesis disorder, and progressive neurological symptoms [54]. Therefore, a better GGT reserve is essential for generating enough glutathione to maintain the redox balance in the body.

However, it still had some limitations. First, we adopted a relatively short follow-up period which may have influenced the observed outcomes. Previous studies mainly focused on the relationship between GGT with AD and cognitive decline in later stages of life over 10 year follow-up periods. However, the relationship between GGT and PSCI has not been conclusively determined. Since PSCI is closely associated with stroke and is characterized by its fluctuations, the outcome of the association between GGT and PSCI depends on the length of follow-up. Second, we found that in the population with high GGT levels, levels of biochemical indicators for liver functions, such as ALT and AST were also high. It is possible that this population paid more attention to their health conditions, and on their own, they could be adopting certain measures to protect their liver functions during follow-up. This may have weakened the potentially dangerous relationship between GGT and PSCI. Third, due to missing data for some variables of interest, the potential impact of residual confounders such as a more accurate assessment of the emotional state, partial drug use and dosage, and location of stroke lesions on the results were not excluded completely. Moreover, this study only included patients with minor stroke (low NIHSS score) and TIA, which could not represent all stroke cohorts. To give greater relevance and breadth to the interesting results of this research, further research should also take into account patients with stroke of greater severity and longer follow-up. Furthermore, as an index of biological metabolism, GGT is affected by many factors and presents a dynamic change characteristic. Nevertheless, we only assessed the relationship between GGT at baseline and PSCI. Since baseline GGT levels may be temporarily altered by stroke events, we cannot rule out the possibility of changes in physical health state. The effect of GGT dynamic change on PSCI should be further studied in the future.

In summary, our results revealed that baseline GGT levels are inversely associated with PSCI, with extremely low GGT levels considered to be a risk factor for PSCI. However, GGT levels dynamically change and it plays a two-sided role in vivo. Therefore, relying solely on GGT to predict PSCI should be carefully considered and further longitudinal studies are needed to clarify the mechanism of GGT affecting neuroplasticity.