Bacterial metabolites originating in the gut traverse to the liver via the portal vein before entering systemic circulation. The crosstalk between the gut and liver ultimately leaded to insulin resistance and even diabetes (29). Koh et al. investigated amino acid-derived microbial metabolites potentially linked to insulin resistance and T2D (12). In their initial study involving 15 obese subjects (body mass index [BMI] > 40), higher concentrations of ImP were observed in both portal and peripheral blood of 5 T2D subjects compared to 10 BMI-matched controls. This finding was corroborated in a larger cohort of 649 middle-aged individuals from the Swedish community, where ImP levels remained significantly elevated in treatment-naive T2D subjects after adjusting for BMI, sex, and age (12). Subsequently, numerous studies successively reported the association between ImP levels and T2D. In a large European multicentric cohort (MetaCardis) comprising 1,958 subjects from France, Germany, and Denmark, progressively elevated ImP levels were observed across patients with normal glucose tolerance, prediabetes, and overt T2D (13). Additionally, circulating ImP levels exhibited positive correlations with HbA1C (13, 30), HOMA-IR (13), insulinemia (13), fasting glucose (13, 15) and postprandial glucose (13, 31). These findings indicate that ImP is not only associated with impaired glucose metabolism but also with diabetic status.

The analysis of gut microbiota in T2D patients effectively addressed the reasons for the changes in circulating ImP levels. In an in vitro gut simulator experiment monitoring ImP production kinetics, the gut microbiota of T2D patients demonstrated ImP production capability, unlike non-T2D patients (12). Further investigation revealed the enrichment of ImP-producing bacteria, characterized by “Y” or “M” UrdA homologs, in the intestines of T2D patients (12), as well as in the intestine and skin of T2D mice (22). These included strains previously associated with an elevated risk of T2D in large population cohorts, such as Streptococcus mutans (32), Eggerthella lenta (33), and Lactobacillus gasseri (33).

However, ImP serves not only as a disease marker but also correlates with an increased risk of prediabetes and T2D, as demonstrated in cohort studies (13, 15), suggesting its biological impact on T2D progression. In animal experiments, ImP injection induced glucose intolerance and decreased hepatic insulin signaling (12). Mechanistically, ImP disrupts insulin signaling by activating p38γ mitogen-activated protein kinase (MAPK), leading to p62 phosphorylation and subsequent activation of mechanistic target of rapamycin complex 1 (mTORC1). This results in the phosphorylation and degradation of insulin receptor substrates 1 and 2 (IRS1 and IRS2). Consistently, phosphorylation of p62 and S6K1 were elevated in the human liver compared to healthy controls (12), highlighting the role of ImP in impairing insulin signaling through the p62/mTORC1 pathway.

In addition to its impact on T2D itself, ImP has also been found to influence the hypoglycemic effects of metformin, the first-line therapy for T2D. Metformin exhibits substantial variability in efficacy among individuals, and genetic variations, particularly in genes encoding transporters such as organic cation transporter 1 (OCT1) (34) and glucose transporter 2 (GLUT2) (35), have been identified as influencing metformin response. In addition to gene polymorphisms, Koh et al. discovered that intestinal ImP levels contribute to this variability (36). T2D patients on metformin with persistently high blood glucose levels showed elevated ImP concentrations.

Further experimental studies demonstrated that ImP diminishes the glucose-lowering effect of metformin and inhibits metformin-induced activation of adenosine 5′-monophosphate-activated protein kinase (AMPK) by impeding AMPK serine phosphorylation through the p38γ/Akt pathway. However, as the authors point out, the study has certain limitations worth noting. Given its cross-sectional design, it remains unclear whether individuals with higher blood glucose values (and consequently higher plasma ImP levels) actually responded poorly to metformin or had more severe diabetes prior to treatment initiation. Hence, a longitudinal cohort study is warranted to ascertain whether ImP directly undermines metformin efficacy and whether metformin contributes to the proliferation of ImP-producing bacteria.

The sub-analysis of the MetaCardis study, including 20% of participants with CVD, revealed a significant increase in circulating ImP concentrations among CVD patients after adjusting for traditional risk factors (age, gender, BMI, ethnicity), kidney function, and presence of T2D (13). Correlation analysis demonstrated a strong positive correlation between ImP levels and serum inflammatory markers, including total leukocyte count, high-sensitivity C-reactive protein (hs-CRP), and interferon gamma-induced protein 10 (IP-10) (13), suggesting a potential association between ImP, inflammation, and CVD progression, warranting further investigation.

Recent large-scale clinical studies have linked ImP to atherosclerosis (18–20), a chronic inflammatory vascular disease and the major cause of CVD (37). The role of gut microbiota in atherosclerosis has been supported by increasing mechanistic evidence (38). In fact, significant changes in gut microbiota have been observed in patients with subclinical coronary atherosclerosis before plaque formation. A study involving 8, 973 participants without overt atherosclerotic disease revealed significant alterations in gut microbiota, including oral microbial species like Streptococcus spp, which correlated significantly with ImP levels and systemic inflammation markers (hs-CRP levels and neutrophil counts) (20). However, due to its cross-sectional design and lack of experimental evidence, the study failed to extrapolate the predictive value of ImP on plaque formation and its potential implications in atherogenesis.

Nevertheless, data from HIV patients have provided indications of a possible association between ImP and the presence of plaques in carotid and coronary arteries (18, 19). HIV infection is linked to chronic inflammation and immune activation, critical factors in atherosclerosis and thrombosis development (39). HIV-induced disruptions in gut microbiota exacerbate chronic inflammation and metabolic irregularities, increasing atherosclerosis risk (40, 41).

In a study of 320 females living with or at risk of HIV infection, with 26% having carotid artery plaque, a distinct shift in gut microbiota composition and increased ImP plasma levels were observed (18). Circulating ImP levels were inversely correlated with potentially beneficial microbial species linked to reduced carotid artery plaque. Additionally, ImP levels positively correlated with serum inflammatory markers, including CX3CL1, TNFSRF9, and LIF-R, associated with immune activation and inflammation pathways related to atherosclerotic plaques (42–44). Notably, after adjusting for plasma ImP levels or inflammatory markers, the association between gut bacterial species and plaque weakened, indicating that circulating ImP levels and related inflammatory markers may partly explain these associations. Further analysis identified 17 ImP-associated species (as illustrated in Table 1 ), with 8 correlating with the functional enzyme hutH. A gut microbiota score derived from these ImP-associated species showed a positive correlation with plaque formation and several pro-inflammatory markers, even after adjusting for multiple factors. Collectively, gut bacteria may contribute to plaque formation by modulating host immune activation and inflammation through elevated IMP levels.

Recent research suggests that ImP is not only associated with plaque formation but also with plaque obstruction. HIV-infected patients with obstructive coronary artery disease (CAD) exhibited lower gut microbiota diversity and significant compositional changes compared to HIV-infected individuals without CAD or non-obstructive CAD, with increased abundance of known ImP producers such as Rumiococcus gnavus and Veillonella (19). ImP plasma levels were associated with this dysbiosis, significantly elevated in participants with obstructive CAD (19). However, after adjustment for traditional and HIV-related risk factors, gut dysbiosis but not plasma ImP was independently associated with obstructive CAD, indicating that the effects of gut microbiota may extend beyond ImP in plaque obstruction. Longitudinal studies are needed to establish a causal relationship between ImP levels and plaque formation, shedding light on its predictive value for atherosclerosis. Additionally, exploring ImP in the context of HIV-related CVD is crucial, highlighting the importance of understanding the role of ImP and gut dysbiosis in driving CVD risk among HIV patients.

The recent findings offer a strong foundation for future investigations, yet additional experiments are necessary to move beyond associations and establish causal evidence linking ImP to atherosclerosis. Firstly, further research is needed to elucidate the mechanistic connection between ImP and inflammation, addressing the significant correlations observed in multiple studies. Secondly, Atherosclerosis initiates with endothelial injury, leading to the accumulation of macrophage foam cells and infiltration of smooth muscle cells, resulting in fatty streak formation (45). Inflammatory processes play a pivotal role in the development of vulnerable plaques (46). Rupture of the plaque’s cap triggers platelet aggregation, precipitating thrombosis and vascular obstruction (47). Therefore, understanding whether ImP is involved in endothelial injury, foam cell formation, and platelet aggregation should be the central focus of future research into the progression of atherosclerosis.

Individuals with T2D face more than a twofold increased risk of developing HF compared to non-T2D patients, with higher risks of incident cases and mortality among diagnosed patients (48). Profiling the metabolic connections and shared components of T2D and HF could unveil new disease pathways, enhance risk prediction, and enable tailored prevention and management strategies (49). The elevation of intestinal metabolite ImP in the circulation of T2D patients, along with its induction of insulin resistance in animal models (12), offers a novel perspective on unraveling the molecular signatures and metabolic remodeling of HF and its associated factors.

Targeted metabolomic analysis of 260 individuals with diverse glucose metabolism from the Risk Evaluation and Management of Heart Failure (REM-HF) cohort in China identified ImP as a microbial signature contributing to the shared etiologies of T2D, HF, and CKD (21). Data from the Boston Puerto Rican Health Study (BPRHS) cohort (50) and European Prospective Investigation into Cancer (EPIC)-Norfolk study (49) further corroborated this finding. Impressively, serum ImP levels increased by 1.1–1.6 fold with each additional chronic HF comorbidity (21), supporting ImP as a component of the metabolic connections among T2D, HF, and CKD.

Recently, Molinaro et al. from Sweden (51) investigated the association between circulating ImP levels, HF, and incident mortality risk. In the population-based MetaCardis cohort, significantly higher ImP levels were observed in individuals with established CVD or HF compared to those without, with the highest levels detected in HF patients. Individuals in the highest quartile of ImP levels had a threefold increased risk of HF compared to those in the lowest quartile, even after adjusting for multiple traditional cardiovascular risk factors. Moreover, ImP levels were inversely associated with left ventricular ejection fraction (LVEF) and positively correlated with pro-atrial natriuretic peptide (proANP) and N-terminal pro–B-type natriuretic peptide (NT-proBNP) levels. These findings were consistent across the GeneBank cohort from North America, predominantly comprising patients with HF (n = 407), CVD (n = 1,331), and without CVD or HF (n = 417). Additionally, longitudinal follow-up data from the North American cohort revealed that the highest quartile of ImP was independently associated with an increased risk of overall mortality, even after adjusting for traditional risk factors and baseline covariates (adjusted HR = 1.85, 95% CI [1.20, 2.88], P < 0.01). Overall, this study, drawing from two large independent cohorts, offers compelling evidence supporting a substantial correlation between ImP levels and CVD, HF, and HF-associated phenotypes including reduced left ventricular ejection fraction and heightened natriuretic peptide levels. Crucially, this correlation persists regardless of obesity and T2D, known contributors to disease progression. In vitro experiments conducted on H9c2 cardiomyoblast cells pretreated with hypoxia/reoxygenation further supported a causal link between ImP and distinct HF-relevant phenotypes. In this study, the intervention of 0.1 μM ImP for 24 hours significantly elevated the expression of the Natriuretic Peptide B gene (NPPB), which encodes the B-type natriuretic peptide (BNP), and disrupted cardiomyoblast functions, as indicated by significantly reduced mitochondrial membrane potential (21).

The current research on the association between elevated circulating ImP levels and HF presents valuable insights, yet it also reveals several limitations and areas for future investigation. One major concern is the lack of clarity regarding the underlying reasons for the elevation of ImP in HF patients, posing a significant gap in our understanding. Moreover, the reliance on cross-sectional data underscores the need for longitudinal research to establish definitive causal relationships. Furthermore, the predominantly focused research on specific populations calls for more diverse cohorts to ensure the generalizability of findings. Future studies should prioritize exploring the mechanistic understanding of ImP’s role in HF development and conducting interventional trials to assess the therapeutic potential of targeting ImP levels. Consideration of confounding factors such as medication use and lifestyle variables is essential, and efforts to validate ImP as a diagnostic and prognostic biomarker are warranted. Addressing these gaps through comprehensive research endeavors will enhance our understanding of the involvement of ImP in HF and facilitate the development of effective therapeutic strategies.

NAFLD, affecting approximately one quarter of the global population, encompasses a spectrum of conditions ranging from simple hepatic steatosis, often linked to obesity, to NASH, which can progress to fibrosis, cirrhosis, and hepatocellular carcinoma (52). The gut and liver are interconnected through the portal vein, forming the gut-liver axis, which serves as a direct pathway for gut microbiota and their metabolic by-products to reach the liver (53). Bidirectional communication along the gut-liver axis plays a pivotal role in NAFLD pathogenesis (54). In NAFLD, microbial dysbiosis in the gut, particularly a decrease in SCFAs-producing microbiota, has been documented (55). This reduction in SCFAs production leads to an elevation in intestinal pH (56), influencing bacterial metabolite production and subsequent absorption into the host circulation (57).

In Göttingen minipigs fed a choline-deficient amino acid-defined high-fat diet (CDAHFD), serving as a NASH animal model, notable increases in serum ImP concentration and pancreatic glucagon levels were observed. These changes were accompanied by liver activation of mTORC1, as indicated by increased expression of liver RHEB and MTOR genes, along with impaired hepatic insulin signaling, demonstrated by decreased expression of IRS1 and IRS2 (16). Moreover, multiple linear regression analysis identified ImP as a statistically significant predictor for glucagon levels (P = 0.0068). Additionally, 16S rRNA analysis showed significant downregulation of intestinal SCFAs-producing bacteria in CDAHFD-fed minipigs, including butyrate-producing members of Lachnospiraceae and propionate producers from the Muribaculaceae family (16). This resulted in an elevation of colon luminal pH, creating an environment conducive to enzymatic activity of bacterial urocanate reductase, thereby facilitating ImP production from histidine metabolism (12, 16, 26, 28). Following production, ImP is likely transported from the intestines to the liver via the portal vein (58), where it subsequently contributes to impaired hepatic insulin signaling, hyperglucagonemia, decreased expression of the glucagon receptor, and disruption of the liver-α-cell axis (16). This aligns with emerging evidence showing a negative correlation between ImP levels and fibroblast growth factor 21 (FGF-21), an endogenous regulator of lipid and glucose metabolism (30). Recently, ImP were enriched in cirrhotic patients with chronic hepatitis B compared to healthy subjects, signifying its potential role in chronic hepatitis B progression, but its role remains to be further clarified.

Future research should delve into elucidating the exact role of ImP in the pathogenesis of NAFLD and other liver disease. Specifically, efforts should focus on determining the association of ImP with the progression of liver diseases and its potential diagnostic and therapeutic implications. Moreover, researchers could explore strategies to modulate ImP levels by manipulating gut microbiota composition or intervening in gut SCFAs production, thereby developing novel treatment approaches.

Globally, over 10% of the population suffers from CKD, characterized by kidney damage, typically indicated by urinary albumin, or decreased kidney function, measured by glomerular filtration rate (59). Individuals with CKD face a significantly elevated risk of CVD and cardiovascular-related mortality (60), underscoring the importance of early detection for effective management. The gut microbiota and their associated metabolites play a crucial role in the microbiota-gut-kidney axis, offering a promising avenue for early diagnosis and personalized treatment to slow renal progression (61).

Recent studies have identified ImP as another metabolite linked to kidney traits, prospectively associated with CKD incidence over time (17). In a large study involving 2,438 Hispanic/Latino adults (12% with CKD), elevated ImP levels were correlated with worsening kidney traits, including reduced eGFR, increased urinary albumin-to-creatinine (UAC) ratio, and CKD incidence over approximately 6 years (17). Similarly, in another cohort from China with varying glucose tolerances, ImP levels exhibited a strong association with creatinine, Cystatin C, and estimated glomerular filtration rate (eGFR). Furthermore, serum ImP levels increased by 1.5 times with the occurrence of CKD in patients with T2D and chronic HF (21). Notably, ImP was more strongly associated with biomarkers of CKD than with those of HF and T2D, indicating its potential pathogenic role in all three conditions and suggesting shared etiologies mediated by ImP among these diseases.

The understanding of the mechanisms underlying increased ImP levels in CKD is still evolving. Alterations in the gut microbiota may be a contributing factor, as CKD progression leads to factors such as sodium and water retention, increased circulatory system pressure, visceral congestion, intestinal wall edema, and impaired intestinal barrier function, resulting in bacterial translocation and gut dysbiosis (62). Studies like the Hispanic Community Health Study/Study of Latinos (HCHS/SOL) study (17) have shown that higher CKD incidence rates and UAC ratios, along with lower eGFR, are associated with reduced gut microbiota diversity and alterations in overall microbial composition, which may directly contribute to increased ImP production. Future research should further explore the relationship between CKD progression, gut dysbiosis, and ImP production to uncover potential therapeutic strategies for managing CKD-related complications and reducing the risk of ImP-associated health issues.

Another interesting discovery is the observed prospective association of ImP with changes in renal function and impairment in CKD, particularly prominent in patients with diabetes (17). This is consistent with animal studies where serum ImP levels were significantly elevated in diabetic mice models (63). ImP levels positively correlated with renal functional parameters like the UAC ratio. In cellular studies, ImP was found to stimulate inflammation and fibrosis by promoting toll-like receptor 4 (TLR4)-mediated phosphorylation of NF-κB and Stat3 and expression of IL-6, TGF-β1, and MyD88 (63). Moreover, downregulating ImP-producing bacteria, such as certain genera of Bacteroides and Unidentified Ruminococcaceae, was shown to ameliorate diabetic kidney disease by suppressing ImP-induced protein expression of the TLR4 signaling pathway in vivo and in vitro (63). Understanding the mechanistic role of ImP in CKD progression and its associations with other renal conditions, along with validating these mechanisms through extensive cell and animal model studies, could lead to improved clinical applications in the future.

The interconnection among obesity, hypertension, diabetes, and CVD underscores the importance of investigating shared metabolic pathways to improve risk assessment and develop personalized prevention and management strategies. Studies exploring the association between ImP, T2D, and CVD, such as those by Koh and Molinaro, et al. (12, 13) have included patients with overt metabolic diseases or CVD, complicating the differentiation of ImP’s correlation with known risk factors like weight gain, hypertension, and cholesterol. Insights from microbiome community typing analyses within the MetaCardis cohort provided some clues, revealing elevated ImP levels in individuals with the Bacteroides2 (Bact2) enterotype (13), a gut microbiota profile associated with systemic inflammation and obesity (64). Research involving 1,018 females from the UK Adult Twin Registry (TwinsUK) cohort further supports the correlation between ImP and obesity (15). This study found that serum ImP levels were positively correlated with BMI (Pearson correlation coefficient [r] = 0.18, P = 5E-9), visceral fat mass (r = 0.067, P = 0.06), and an increased risk of obesity (r = 0.2, P = 8E-9), indicating circulating ImP levels as a potential marker for obesity.

In another cohort of overweight/obese subjects without T2D and not on any CVD medication, the association between circulating plasma ImP concentrations and CVD risk factors, including blood pressure, HDL cholesterol, and LDL cholesterol, was investigated (14). This study revealed a positive correlation between plasma ImP concentrations and diastolic blood pressure (Spearman rank correlation coefficient [rs] = 0.285, P = 0.004), with borderline significance for systolic blood pressure (rs = 0.187, P = 0.060). Interestingly, no significant association was found between plasma ImP concentrations and peripheral or hepatic insulin resistance, contrary to the findings of Koh et al. (12). This discrepancy could be attributed to differences in the study populations, as Koh et al.’s study included patients with varying BMI and metabolic disease severity. However, the data revealing the association between ImP and blood pressure comes from cohort comprising subjects without T2D, with a very homogeneous range of BMI, and without any medication or overt chronic diseases except for metabolic syndrome (14), suggests that if ImP influences blood pressure, it may do so through mechanisms other than insulin resistance.

While the association between ImP and obesity, as well as blood pressure has been observed, the underlying mechanisms remain unclear. Future research should delve deeper into the molecular pathways linking ImP to obesity and hypertension and explore potential therapeutic targets to mitigate its effects on cardiovascular health. Additionally, the discrepancy in findings regarding the association between ImP and insulin resistance highlights the need for further clarification through well-controlled studies involving diverse populations.

The association between elevated ImP levels and cardiovascular metabolic impairment underscores the importance of exploring the mechanisms underlying ImP’s effects, which may elucidate the potential for targeting ImP to improve cardiovascular metabolism and identify novel drug targets. Insights from the work of Koh and Molinaro et al. (12, 13, 36) have shed light on this field. Their research team elucidated the molecular mechanism by which ImP impairs glucose tolerance and insulin signaling through the activation of the p38γ/p62/mTORC1 signaling pathway. Further studies have revealed that ImP inhibits the hypoglycemic activity of metformin via the p38γ/Akt/AMPK pathway. Structural analysis has unveiled the interaction between ImP and the adenosine triphosphate (ATP) binding pocket of p38γ. In silico analysis suggests that pirfenidone, used to treat idiopathic pulmonary fibrosis, may compete with ImP for binding to this site of p38γ, implying its potential as a candidate for combination therapy in individuals with T2D who are unresponsive to metformin. Future research should focus on elucidating the detailed mechanism of ImP binding to p38γ for structure-based drug design. While pirfenidone shows promise, clinical trials are needed to confirm its efficacy and safety in T2D patients, along with investigations into potential drug interactions.

Besides opposing the interaction involving ImP, strategies aimed at reducing ImP production, like inhibiting enzymes such as urocanate reductase, present promising new therapeutic avenues for T2D. In this regard, Venskutonytė and Koh et al. elucidated the X-ray structures of its ligand-binding domains (26), providing valuable insights for structure-based drug design and aiding in the development of inhibitors for potential treatment of metabolic disorders.

While the existing research provides valuable insights into the role of ImP in cardiometabolic diseases, several limitations and areas for future investigation should be acknowledged. Firstly, much of the current research focuses on observational and cross-sectional studies, limiting our ability to establish causal relationships between ImP and cardiometabolic diseases. Longitudinal follow-up studies are needed to elucidate the temporal relationship between ImP levels and disease onset and progression.

Secondly, there is a noticeable lack of diversity in study populations, with a predominant inclusion of individuals of European descent in many studies. Given that the gut microbiome exhibits significant variation across geographical regions (84, 85), and considering that the production of ImP is highly dependent on bacterial activity, it becomes imperative for future studies to encompass more diverse populations. This approach will not only enhance the generalizability of findings but also allow for a comprehensive understanding of how ethnic and geographic factors may influence ImP production and its implications for cardiometabolic diseases.

Thirdly, while circulating ImP levels have been correlated with systemic inflammation in metabolic disorder (13, 18, 20), the underlying mechanisms remain unclear, highlighting the necessity for further research into ImP’s regulation of inflammatory pathways. Notably, rectal administration of ImP in mice resulted in a significant increase in NF-κB, iNOS, and IL-6 expression, accompanied by a reduction in goblet cell count (23). These findings suggest ImP’s potential to induce intestinal inflammation, disrupt the intestinal barrier, and alter goblet cell proliferation. Given the crucial role of intestinal barrier integrity in cardiometabolic diseases, targeting ImP to modulate intestinal barrier function holds promise for cardiometabolic disease treatment. However, additional evidence is required to substantiate ImP’s regulatory role and elucidate its molecular mechanisms in this context.

Fourthly, optimal enzymatic activity of UrdA occurs under neutral pH conditions (28), while the proximal colon tends to maintain an acidic environment. Reduced production of SCFAs and increased protein fermentation in the gut contribute to the elevation of colonic pH (86), potentially facilitating ImP production. This inference partially elucidates the observed association between decreased fiber intake and increased ImP levels in certain populations, along with the correlation between a Western diet and elevated ImP content (13, 70, 71). Therefore, future research should focus on elucidating the contributions and mechanisms of factors influencing intestinal pH on ImP levels, thereby providing theoretical insights into targeting ImP-mediated pathways of cardiovascular metabolic damage. Additionally, the inhibitory effect of ImP on the hypoglycemic efficacy of metformin (36) underscores the profound impact of gut microbiota and their metabolites on drug therapy. Exploring the effects of drugs related to the treatment of cardiovascular metabolic diseases on ImP-producing bacteria and ImP content presents a novel and intriguing topic, offering potential new insights into the therapeutic mechanisms of drugs.

Furthermore, while diet has been shown to significantly influence ImP production via its effects on gut microbiota composition, it should be noted that other potential factors influencing ImP levels, such as genetic predisposition or medication use, have not yet been thoroughly investigated. Future research will be needed to explore these dimensions, as well as to further elucidate the mechanistic links between ImP and cardiometabolic diseases.

Lastly, as an increasing number of studies shed light on the correlation between gut bacteria and ImP production, as demonstrated in Table 1 , it becomes evident that, like many other investigations focused on the gut microbiome, these studies raise more questions than they answer. Unveiling the precise contributions and mechanisms of these bacteria to ImP production is a vast and intricate endeavor, yet it constitutes a crucial aspect of research on ImP and warrants significant attention in future studies.