Background
Schizophrenia is a mental disorder affecting approximately 1% of the world’s population and is a severe disorder that leads to functional deterioration [
1]. Despite cardinal features of schizophrenia, it remains the least understood psychiatric disorder owing to the lack of pathological hallmarks [
2,
3]. With the identification of schizophrenia susceptibility genes [
4], genetic traits have been considered to play important roles in schizophrenia occurrence [
5]. The relative contribution of genetic factors in schizophrenia is estimated to be up to 80% [
6].
Recently, the target genes of antihypertensive medications were reported to be associated with the risk of schizophrenia. Specifically, low angiotensin-converting enzyme (ACE) messenger RNA and protein levels, which are targets of ACE inhibitors, are associated with an increased risk of schizophrenia [
7]. In addition, according to Fan et al., genetically proxied ACE inhibitors were reported to be associated with an increased risk of SCZ in Europeans and East Asians [
8]. Contrary to ACE inhibitors, other antihypertensive medications such as BB and CCB were found to have no association. Animal experiments demonstrated that the brain RAS targeted by ACE inhibitors can regulate various brain functions such as sensory information processing, learning, memory, and emotional responses [
9]. However, it remains unclear whether this potential biological association translates into clinically significant difference of the schizophrenia occurrence in real-world scenarios. Given the widespread use of ACE inhibitors in hypertensive patients and their potential biological implications for schizophrenia risk, investigating this association using real-world data is essential.
Therefore, comparing the effects of antihypertensive drugs on schizophrenia may be a way to identify potential risk factors for schizophrenia occurrence. We aimed to conduct a head-to-head study comparing the occurrence of schizophrenia between antihypertensive drugs in patients with hypertension. Specifically, we investigated whether the use of ACE inhibitors increased the risk of schizophrenia compared with the use of angiotensin receptor blockers (ARBs) or thiazide diuretics in the US and Korea across the Observational Health Data Sciences and Informatics (OHDSI) network [
10].
Discussion
The potential association between ACE inhibitors and an increased risk of schizophrenia is highly relevant because of the number of affected patients and the burden of schizophrenia, warranting thorough investigation. In this study, we extensively estimated the comparative risks of ACE inhibitors and thiazide diuretics or ARBs on the occurrence of schizophrenia. No differences in risk were found between the use of ACE inhibitors versus ARB or between the use of ACE inhibitors versus thiazide diuretics. Although the use of ACE inhibitors was associated with an increased risk of schizophrenia compared with the use of ARB in the group aged > 45 years, the results were not consistent in the sensitivity analyses. Regarding the secondary outcome, no difference in risk was found among the antihypertensive drugs.
Schizophrenia imposes significant health, social, and economic burdens on individuals, families, caregivers, and society at large [
22]. Unfortunately, by the time schizophrenia becomes apparent behaviorally, neural damages may already be irreversible [
23]. Owing to the limited effectiveness of treatments, identifying psychosis risk factors for prevention and early detection has become crucial [
24]. Additionally, regarding antihypertensive medications, 31.1% of adults worldwide are affected by hypertension [
25]. ACE inhibitors are the most commonly used antihypertensive medications in the US. Previous studies on the relationship between ACE inhibitors and schizophrenia have limitations in terms of sample size or cross-sectional design [
26,
27]. Therefore, we conducted this well-designed longitudinal cohort study. First, we selected the fit-for-purpose databases for the two countries. The large claims database has less fragmentation than individual electronic medical records, allowing us to conduct longitudinal cohort studies for identifying genetic relationships in schizophrenia [
28]. Given the different prevalence of schizophrenia between countries [
29], we analyzed more than 2 million patients in the United States and South Korea. In particular, the HIRA database contains nationwide claims data for the entire Korean population; therefore, our results are sufficiently representative. Second, many robust designs and methods were applied to infer associations between study groups. Controlling biases is critical in observational studies using routinely collected observational databases [
30]. Using an active-comparator new-user design, large-scale PS methods can resolve biases arising from time-related design and comparability [
31,
32]. An assessment of systematic errors using falsification endpoints also provides a more reliable statistical interpretation and minimizes the effect of residual bias [
18].
Additionally, in the subgroup of individuals aged ≥ 45 years, a significant difference was observed between ACE inhibitors and ARBs. The possible biological pathway for the association between ACE inhibitors and schizophrenia is that ACE and the central RAS may play a role in inflammation and immunity [
33]. Immune dysfunction due to reduced ACE activity may contribute to the development of schizophrenia [
34]. Especially, according to studies on the pharmacokinetics of ACE inhibitors, the Area Under the Plasma Concentration-Time Curve has been reported to be greater in older individuals compared to younger ones, attributed to renal function decline and changes in body composition [
35]. These findings suggest that the impact of ACE inhibitors is more pronounced in older individuals, as in our results. As another possible explanation, a previous study suggested the therapeutic potential of ARBs in patients with schizophrenia through the anti-inflammatory properties of gamma-aminobutyric acid [
36], while thiazide diuretics had no effect on schizophrenia [
13]. This could explain why ARBs are associated with a lower risk than ACE inhibitors. Nonetheless, the results were not significant in the ITT setting and require further study, making it difficult to draw definitive conclusions.
Moreover, differences in the risk of schizophrenia based on antihypertensive medication existed in the US data at the ITT follow-up. Although the differences between ACE inhibitors and thiazide diuretics were inconsistent, ACE inhibitors were consistently associated with a higher risk of schizophrenia than ARBs in both subgroup analyses and secondary outcomes. This result appears to be consistent with the results of the subgroup analysis at the AT follow-up. However, ITT can overestimate the effects of treatment in the presence of differential adherence [
37].
Hypertensive patients are more likely to be diagnosed with mental disorders, and hypertension increases the severity of psychological distress. On the contrary, mental disorders are independent risk factors for hypertension. In other words, there is a clinically significant bidirectional relationship between hypertension and mental disorders [
38‐
40]. In these situations, it is important to clarify how and to what extent antihypertensives affect schizophrenia from a clinical perspective. Given our findings of no significant differences by antihypertensive medication, there is insufficient evidence to recommend clinically that antihypertensive medications be reduced or discontinued. From the patient’s perspective, information about hypertension medications associated with schizophrenia risk could impact treatment adherence for people with hypertension, given what has happened to them during COVID-19 [
41]. Considering our results, it does not appear that people with hypertension need to consider whether to use or change their antihypertensive medication because of the risk of schizophrenia.
This study had some limitations. First, there may be unmeasured risk factors for schizophrenia. For example, the balance for hypertension status (including blood pressure values) between the two groups could not be determined due to the nature of the claims data. A family history of schizophrenia and social history, such as immigration, are related to the development of schizophrenia [
42,
43]. In addition, economic variables (such as income status) may also be associated with the development of schizophrenia, but were not used in this study. However, we used large-scale PS methods that can help reduce the impact of measured confounders and balance the distribution of these variables between groups [
44]. Nevertheless, given the uncontrolled confounding by the propensity score method, further studies including social and family factors are needed. Second, the number of patients varies across the databases. Although data from 50 million people in Korea were used, only approximately 22,927 patients used ACE inhibitors, while data from the United States exceeded 5 million patients using ACE inhibitors. Such discrepancies in sample sizes could potentially impact the generalizability of our findings [
45]. However, it is important to note that this heterogeneity in clinical practice can also be seen as a strength of our study. By utilizing data with diverse prescribing patterns, we can generate more reliable and generalizable evidence that better reflects real-world clinical scenarios. Moreover, the use of multinational databases and studies with multiple databases of varying sizes have previously demonstrated feasible and consistent results [
46]. Third, the diagnostic system for schizophrenia has limitations. Prior reviews have shown variability in schizophrenia diagnosis [
47], which may be due to the complexity and heterogeneity of schizophrenia [
48]. These diagnostic problems appear not only in schizophrenia but also in other psychiatric diseases such as depression and bipolar disorder [
49]. For the strictness of diagnosis, we added prescriptions of antipsychotics and occurrences of psychiatry procedures. Lastly, more comprehensive analyses are still needed to generalize our findings. This study only included RAS inhibitors and thiazide diuretics among the main antihypertensive drugs, and additional analyses such as calcium channel blockers could be considered. It also excluded patients on two or more medications, which are prescribed to more than half of all patients with hypertension [
50], and further research is needed on patients on such combination therapies.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.