In this retrospective study, a large real-world cohort of more than 90,000 adult patients with NAFLD in Germany was compared with a 1:1 cohort of patients without NAFLD for incidence of CKD over a 10-year period using the Disease Analyzer database (IQVIA). Patients were matched for sex, age, index year, annual visit frequency, hypertension, and diabetes. Our results show that patients with NAFLD have a significantly increased risk of developing CKD. Within 10 years of the index date, 19.1% of patients with NAFLD were newly diagnosed with CKD, compared with only 11.1% of patients without NAFLD. This association was most pronounced in the 18- to 50-year age group.

Similar to our findings, a meta-analysis by Musso et al. with a total of 33 studies and over 63,902 participants showed that NAFLD is associated with an increased incidence and prevalence of CKD [22]. In addition, a retrospective study of 8,329 non-diabetic, non-hypertensive Korean men with normal renal function at baseline found that NAFLD was associated with an increased incidence of CKD after adjustment for age, cholesterol, and other factors over a 3-year period [23]. Another meta-analysis examined the magnitude of the association between NAFLD and the risk of developing CKD and comprised a total of 9 observational studies with a total of 96,595 adults of predominantly Asian descent and 4653 cases of stage ≥ 3 CKD over a median period of 5.2 years. The authors concluded that patients with NAFLD have a significantly higher risk of developing CKD than patients without NAFLD and that NAFLD is associated with a nearly 40 percent increased long-term risk of CKD [22]. A recently published updated large meta-analysis by Mantovani et al. of observational studies involving a total of 1,222,032 individuals (28.1% with NAFLD) and 33,840 cases of stage ≥ 3 CKD with a median follow-up of 9.7 years again confirmed that NAFLD is associated with a significantly increased risk of CKD. The authors demonstrated an approximately 1.45-fold increased long-term risk of developing stage ≥ 3 CKD in patients with NAFLD. All risks were independent of age, sex, obesity, hypertension, diabetes, and other common CKD risk factors [19].

Of note, in contrast to the study by Mantovani et al., which focused only on CKD ≥ 3 stages, our analysis included all 5 CKD stages as the primary end point because coding for each CKD subclass was not available. This fact may explain the higher risk of developing CKD in our study compared with the work of these authors (HR 1.8 vs 1.43). Furthermore, although the risk of CKD was numerically slightly increased in female NAFLD patients in our study, this observation could not be interpreted as a significant sex difference. While Mantovani et al. hypothesized that the observation of a sex-independent association was due to the fact that some of the underlying studies did not adequately adjust for sex differences, our data rather suggest that sex and most likely menopausal status are not effect modifiers associated with NAFLD and CKD [19].

The apparent decline in CKD incidence with increasing age demonstrated in our study (from 2.13 OR for 18–50 year olds to 1.66 (1.56–1.95) for > 70 year olds) also suggests that NAFLD adds to the risk of CKD at a time when traditional risk factors are less prominent.

Several pathophysiological mechanisms for the association between NAFLD and CKD have been discussed in the literature. Although our analysis cannot elucidate the mechanisms by which NAFLD contributes to the development of CKD, our study and others, including the recent meta-analysis by Mantovani et al., suggest that, far beyond its phenotype, NAFLD is a systemic disease whose pathophysiology and prognosis are determined by the involvement of multiple organ systems. The heterogeneity of effect sizes and incomplete penetrance imply that the pathogenetic pathways linking NAFLD to CKD are complex and determined by a variety of metabolic, genetic, epigenetic and dietary factors that are currently not fully understood and require further investigation [24, 25]. Recent evidence suggests that various metabolic processes in NAFLD may promote atherogenic dyslipidemia, induce hypertension, and trigger a chronic systemic inflammatory response leading to the development and progression of CKD [26].

There is increasing evidence of liver–kidney interactions in patients with NAFLD, including altered renin–angiotensin system (RAS) activation, impaired antioxidant defenses, and dysfunctional lipogenesis [27]. In this context, RAS may represent a possible link between NAFLD and CKD. It has been reported that RAS activation in the liver promotes insulin resistance, lipogenesis, and the production of proinflammatory cytokines such as interleukin-6 (IL-6) and tumor growth factor-β (TGF-β) [27, 28], which induces fibrogenesis and causes histological changes typical of NASH [27]. In the kidney, RAS activation plays a key role in the development of ectopic lipid deposition, which in turn leads to glomerulosclerosis through oxidative stress and inflammatory processes [29]. Understanding these mechanisms could help identifying therapeutic targets for the prevention and treatment of NAFLD and CKD. Furthermore, the role of the energy sensor 5′-AMP-activated protein kinase (AMPK) and its regulation of fetuin-A and adiponectin in liver and kidney cells was recently investigated [28, 30]. Fetuin-A is a serum protein mediated by AMPK as an important promoter of insulin resistance in both podocytes and hepatocytes [30] and has been found at elevated serum levels in patients with NAFLD and CKD [31]. Increased caloric intake and obesity are thought to trigger an inflammatory cascade between adipocytes in the liver and kidney via AMPK, fetuin-A and adiponectin, leading to end-organ damage [30].

The rapidly increasing prevalence of adipositas, type 2 diabetes and metabolic syndrome is a major challenge for the healthcare system. Currently, there is no approved drug therapy for NAFLD. The cornerstones of NAFLD treatment are lifestyle interventions (e.g., physical activity, weight reduction, diet modification) and control of metabolic syndrome and cardiovascular risk factors [32]. Based on the presumed pathophysiological mechanisms, several pharmacotherapeutic interventions are under investigation for the treatment of NAFLD. Although very few studies have examined the use of medications and behavioral modifications in both NAFLD and CKD, the common cardiometabolic risk factors and underlying pathophysiology may suggest that these therapies are applicable to both conditions.

Limited data suggest that RAS blockade with angiotensin receptor blockers (ARBs) reduces insulin resistance and inflammatory markers in patients with NAFLD steatosis independent of blood pressure lowering [27]. In addition, reductions in necroinflammation, NAFLD activity score, NASH fibrosis stage, and microalbuminuria were observed [33]. A recent cross-sectional study also showed that CKD-NAFLD patients taking angiotensin-converting enzyme inhibitors (ACE-I) or ARBs had less liver stiffness than patients not taking any medication [34]. Other research suggests that insulin-sensitizing agents, including thiazolidinediones (TZDs) such as pioglitazone, may be beneficial in the treatment of NAFLD [30].

We acknowledge that our study has some limitations that are mainly due to the study design and are therefore unavoidable. Because diagnoses were documented using ICD codes, we cannot exclude the possibility that incorrect or inadequate coding could lead to potential bias. Another limitation is that the analyses were not based on laboratory, imaging, or histologic findings, so the validity of the NAFLD codes in the Disease Analyzer database could not be confirmed. As a further limitation, the Disease Analyzer database does not provide more detailed information, such as laboratory data, liver or kidney histology, or clinical course, in addition to ICD-10 codes, which would have allowed more accurate stratification of NAFLD or more precise definition of CKD stage. The database also does not capture the mortality data, making it impossible to calculate survival of study patients. In addition, other risk factors for CKD, such as family history, smoking, or medication use, were not available for further analysis. The database contains electronic medical records from office-based physicians and no data from hospitals or dialysis centers. It should also be noted that the analyses are purely descriptive in nature, with only exploratory observations showing an independent association between CKD and NAFLD, but not proving a causal relationship.

However, the strengths of this work include the large number of patients included, the long study period of 10 years, and the use of a database whose representativeness and validity have already been proven [21].