2.1 EudraVigilance

The ADR database EudraVigilance of the European Medicines Agency (EMA)²⁴ includes all spontaneously reported ADRs by health care professionals (HCP) such as physicians and pharmacists or non-health care professionals (non-HCP) such as patients and their relatives from the member states of the European Economic Area. In EudraVigilance, ADRs are coded in accordance with MedDRA terminology²⁵ and drugs in accordance with the EudraVigilance medicinal product dictionary.²⁶ MedDRA terminology is a standardised medical terminology with five different hierarchical levels to code, among others, diagnoses, laboratory results and symptoms. The structure of the five different hierarchical levels enables specific and aggregated analyses of the reported ADRs. Among others, the preferred term (PT) level describes the symptoms, laboratory results or diagnoses and the system organ class (SOC) level the groupings of these by aetiology, manifestation site or purpose. Furthermore, standardised MedDRA queries (SMQ)²⁷ are pre-determined sets of MedDRA terms (e.g. symptoms) grouped together to identify ADR reports describing specific diagnoses.

2.2 vARIANCE study

The vARIANCE Study (Angioedema Risk Under Angiotensin-Converting Enzyme Inhibitors) aims to identify factors associated with ACEi-/ARB-induced angioedema.³¹ The patients were recruited in Germany and Austria in several hospitals and primary care centres.³² The data were collected using a specific questionnaire. A more detailed description of the study can be found on the study website (www.variance-studie.info). All participants provided written informed consent prior to inclusion, and the study was approved by the respective ethics committees.

2.2.1 Identification of angioedema cases in the vARIANCE study

All cases of adults aged 18–90 years received between 01/01/2018 and 31/01/2023 were identified in the database of the vARIANCE study (n = 142). Among these, cases with a complete questionnaire and an at least possible causal relationship between the intake of an ACEi or ARB, and the occurrence of angioedema according to WHO criteria,²⁸ assessed by the reporting physician and the study committee, were included. Cases with self-reported urticaria or other causes more likely to be associated with the angioedema were excluded. The final dataset consisted of n = 114 cases (Figure 1).

In 80 of these 114 cases (70.2%), the presence of hereditary angioedema (HAE) could be ruled out with high probability by screening either available sequencing data on five HAE-associated genes (SERPING1, F12, PLG, ANGPT1, KNG1; n = 64, 80.0%)³² or relevant laboratory parameters (n = 16, 20.0%). However, cases of HAE with normal C1 esterase inhibitor concentration and activity may have been missed due to their normal C1 esterase inhibitor values and since not all to date known genes (e.g. MYOF, HS3ST6, CPN1) associated with this rare condition were investigated.

2.3 Drug prescription data

Prescription data according to 300 SGB V were provided by the Central Research Institute for Ambulatory Health Care in Germany³³ for patients ≥16 years with at least one prescription of an ACEi, ARB or their combination products between 01/01/2018 and 30/06/2022. The number of drug prescriptions represents the total number of out-patient prescriptions for patients with statutory health insurance (almost 90% of the German population³⁴) dispensed in German pharmacies. Notably, in-patient prescriptions are not covered.

2.4 Analysis criteria

2.5 Statistical analyses

3.1 Descriptive analyses of the patient populations

Patients who developed angioedema were on average 67 years old in EudraVigilance reports and 64 years old in the vARIANCE study and a slightly higher proportion were males (EudraVigilance: 52.0%; vARIANCE: 55.3%) (Table 1). Likewise, the BMI of the patients was on average higher than 25 (‘overweight’) in both datasets (EudraVigilance: 70.8%, vARIANCE: 77.5%). Most of the patients in the vARIANCE study with information about their self-reported ancestry were Europeans (93.6%).

In the vARIANCE study, about half of the patients reported consumption of alcohol (52.6%) and that they previously smoked (48.2%). Only 16.7% stated that they were current smokers. In contrast, only 4.1% of the patients in the EudraVigilance reports reported consumption of alcohol and current smoking, respectively, and one person was known to have smoked in the past (0.6%). In both datasets, former and current smoking as well as consumption of alcohol were more often reported for males than for females (Supporting Information S3: Appendix 3).

About one third of the patients from the vARIANCE study reported an allergy (34.2%) (Table 1). Most of them had allergies to pollen and dust mites (48.7%). Allergies against drugs were present in one third of these patients (33.3%). Allergies and intolerances were reported in 11.7% of the patients in EudraVigilance reports. Among these, allergies against drugs (50.0%) and pollen and dust mites (35.0%) were also reported most frequently.

A previous angioedema was known for 57.9% and 8.8% of the patients from the vARIANCE and the EudraVigilance datasets, respectively. Most of them were attributed to ACEi/ARBs (vARIANCE: 63.6%; EudraVigilance: 73.3%).

Chronic diseases were reported by 80.7% of the patients from the vARIANCE study, and by 64.9% of the patients from EudraVigilance. In both datasets, hypertension and type 2 diabetes mellitus ranked first and second among the chronic diseases. In EudraVigilance reports, coronary artery diseases ranked second among the chronic diseases most frequently reported in males (Supporting Information S3: Appendix 3).

About two-thirds of the EudraVigilance reports were classified as serious (67.3%), in 31.0% a hospitalization or a prolongation thereof was mentioned, and in 5 reports (2.9%) a fatal outcome was reported. Slightly more than half of the patients from the vARIANCE study were hospitalized due to the angioedema (51.8%). In both datasets, angioedema seemed to be more serious in males than in females (Supporting Information S3: Appendix 3). In EudraVigilance reports, 55.9% of ARB- and 84.1% of ACEi-induced angioedema were classified as serious (Supporting Information S4: Appendix 4).

3.2 Descriptive analyses of drug therapy

ACEi were reported as suspected in 84.2% of cases of the vARIANCE study and in 40.4% of the EudraVigilance reports (Table 2). In contrast, ARBs were mentioned as suspected in 59.6% of EudraVigilance reports compared to 11.4% of cases from the vARIANCE study. Ramipril (61.9%) was clearly most often stated as suspected ACEi/ARB in the vARIANCE study, followed by lisinopril (13.3%) and candesartan (9.7%). In the EudraVigilance reports, sacubitril/valsartan (29.8%) was the ACEi/ARB most frequently reported as suspected followed by ramipril (27.5%) and candesartan (17.5%). Sacubitril/valsartan was also the suspected drug most frequently reported for males in EudraVigilance reports, but ranked only third in reports for females (Supporting Information S2: Appendix 2).

In both datasets, hypertension was the indication, and the standard dose was the dose most frequently reported (Table 2). Heart failure was reported as an indication in roughly one-third of EudraVigilance reports compared to none of the cases from the vARIANCE study. Additionally, heart failure was reported as an indication in about half of the EudraVigilance reports with respective information referring to males and to ARBs (Supporting Informations S3 and S4: Appendices 3 and 4).

In the vARIANCE study, the duration of ACEi/ARB therapy until the occurrence of angioedema was longer than 1 year in 73.4% of the patients with respective information and only 12.0% of these patients developed angioedema in the first 2 months of drug therapy (Table 2). In contrast, in the EudraVigilance reports with respective information, angioedema occurred most frequently in the first 3 days of drug therapy (27.7%) and in nearly two-thirds of all cases within the first 2 months. In only 20.8%, it occurred after one year (mean: 478.5 (+/−1086) days; median: 14 [3–300] days). The respective ACEi/ARB therapy was discontinued after the occurrence of angioedema in more than 70% of the patients with respective information in both datasets. However, in 29.2% of these patients in the vARIANCE study, the respective ACEi/ARB therapy was continued. In 31.7% of the patients from the vARIANCE study with respective information, re-challenge was carried out and 65.6% of them developed angioedema again.

3.3 Descriptive analyses of angioedema

Even if the proportions differed clearly, lips, face and tongue were the locations most frequently affected by the angioedema in both datasets (Table 3). The angioedema lasted for 0–3 days in 83.2% of the patients from the vARIANCE study and 59.5% of the EudraVigilance reports (median: 3 [1–5] days) with respective information. Infections and substitution by a different brand with the same active ingredient were the most frequently reported associated factors in the vARIANCE study and EudraVigilance reports. Corrective treatment of angioedema was initiated in 83.3% of the patients in the vARIANCE study and in 20.8% of the patients in EudraVigilance reports. In both datasets, antihistamines and corticosteroids were most frequently used for the treatment of angioedema. In 55.8% and 25.5% of these patients from the vARIANCE study and from EudraVigilance reports, the treatment was considered effective.

3.4 Comparative analyses of angioedema and reference reports from EudraVigilance

3.5 Number of angioedema reports from HCPs in relation to their number of drug prescriptions (=reporting rates)

The annual number of angioedema reports associated with ACEi/ARBs reported by HCPs decreased between 2018 and 2021 (from 46 to 19 reports, −41.3%), whereas the total number of ACEi/ARB prescriptions increased slightly (from ∼58 to ∼61 million, +4.9%). Summarized for 2018–2021, 4 angioedema reports per 10 million ACEi prescriptions and 7 angioedema reports per 10 million ARB prescriptions were reported by HCPs (Table 5). Considering the ACEi/ARBs with more than 3 reports from HCPs, the highest reporting rates per 10 million prescriptions were observed for sacubitril/valsartan (542 reports), followed by lisinopril (21 reports) and valsartan (18 reports).

4.1 Association with demographic parameters

The average age of the patients was 64.2 (+/−11.3) (vARIANCE) and 67.5 (+/− 14.2) (EudraVigilance) years which are consistent with other studies reporting that age >65 is a risk factor (OR 1,36 (KI 1.07–1.73)), but may also reflect the average age of the patients typically taking ACEi and ARB.^{14, 16, 38}

In previous studies, female sex was found to be positively associated with ACEi-induced angioedema probably due to increased levels of oestrogen.^{13, 39} In contrast, our study findings across both datasets consisted of slightly more cases reported in males. Likewise, 51% of the patients were males in a study by Mahmoudpour et al. (2016).¹⁴ Considering the number of reports per 10 million drug prescriptions from HCP from EudraVigilance, sex differences could only be observed for some individual ACEi/ARBs but not for the classes of ACEi and ARB themselves. Angioedema in males were more often classified as serious than in females in our analysis and thus may be reported more often.⁴⁰ Further studies are needed to clarify whether males more often develop serious angioedema than females.

4.2 Association with allergy

The proportion of patients with allergies in the vARIANCE study (34.2%) was roughly similar to the proportion expected in the German population (females: 34.7%; males 27.0%).⁴¹ In the conditional logistic regression analysis, allergy was clearly more often reported in angioedema compared with reference reports (adj. OR 3.3 [1.6–7.1]). Likewise, a history of drug rash and seasonal allergies was reported to be positively associated with ACEi-induced angioedema by others.¹¹

4.3 Association with underlying diseases

In both datasets, hypertension and type 2 diabetes mellitus were reported most frequently among chronic diseases. Hypertension is often treated with ACEi or ARBs, thus explaining the high number of reports on this comorbidity.^{38, 42} In Germany, a slightly higher prevalence of coronary artery diseases was observed for males than for females.⁴¹ Heart failure and coronary artery diseases may favour the occurrence of angioedema.¹ Likewise, coronary artery diseases were reported most often for males in EudraVigilance reports suggesting a possibly higher risk to develop angioedema and probably explaining the slightly higher number of angioedema reports for males in general and of those classified as serious in our analysis.

Diabetes was reported to be negatively associated with ACEi- and ARB-induced angioedema.^{1, 11} The tendency of a lower risk for patients with type 2 diabetes mellitus when comparing angioedema and reference reports may be assumed based on our results of logistic regression analysis (adj. OR 0.6 [0.3–1.0]).

More than two-thirds of patients in both datasets were overweight. Overweight was associated with the re-occurrence of angioedema and was expected to be a risk factor for more severe angioedema by others.¹⁶ However, differences in BMI could not be confirmed between angioedema and reference reports in our analysis.

4.4 Association with co-medication

In a retrospective study, angioedema because of a combination of mTORi and ACEi was diagnosed in 6.6% (n = 9) versus 1.9% (n = 1) of patients with kidney transplants exposed to mTORi but not exposed to ACEi.¹⁸ Likewise, angioedema occurs 1.3%–5.9% (n = 1–15)^43-47 more frequently when ACEi are concomitantly used with fibrinolytics compared to controls⁴⁸ probably due to increased production of bradykinin.⁴³ Adding to these observations, we found that everolimus and fibrinolytics were more frequently reported in angioedema compared to reference reports, albeit with wide confidence intervals due to small sample sizes (mTORi, n = 2 and fibrinolytics, n = 1, Table 4). The combination of ACEi with everolimus or alteplase is already listed as contraindication in the respective product information.^{49, 50}

4.5 Association with life-style factors

Smokers have an increased risk of developing angioedema due to significantly lower serum DPP-IV activity.²⁰ Likewise, alcohol-related disorders may play a role in ACEi-induced angioedema.⁵¹ In summary, however, we did not find a clear association of angioedema with current smoking status and alcohol consumption in our analysis. Such an association, however, could also be related to co-morbidities such as coronary artery disease, heart failure, and hypertension. The number of cases with respective information in both datasets and the degree of information available did not allow investigation of possible interactions.

4.6 Previous angioedema

57.9% (vARIANCE) and 8.8% (EudraVigilance) of the patients had a known history of previous angioedema and more than two-thirds of these were attributed to ACEi/ARBs. A 10-fold increased risk of further angioedema has been reported with continuous ACEi treatment despite the occurrence of previous angioedema.² Previous ACEi- or ARB-induced angioedema are already listed as contraindications in the product information of ACEi and ARBs. The re-occurrence of angioedema was also seen in our data, namely in two of three patients with angioedema after re-exposure. Likewise, almost 90% of the patients in the vARIANCE study with a known history of angioedema developed more than one angioedema episode in the past (unknown if associated with ACEi/ARB therapy). However, it may be possible that the previous angioedema was not attributed to the ACEi/ARB intake since about three quarters of all angioedema in the vARIANCE study occurred after 1 year of drug therapy. Nevertheless, in almost one third of all patients (29.2%) from the vARIANCE study, the respective ACEi/ARB therapy was continued after angioedema occurrence, although the responses might partly result from a misunderstanding of the respective question in the questionnaire. As a translation into clinical practice, physicians should question the patients with regard to the occurrence of angioedema when filling a repeat-prescription for ACEi or ARB.

4.7 Clinical phenotype and time to onset of angioedema

In both datasets, the face, lips and tongue were most frequently affected by angioedema, thereby confirming respective findings in the literature.⁵ In accordance with literature,⁵ we assume that there is no relevant difference between the phenotypes of ACEi-versus ARB-induced angioedema, since the clinical phenotypes did not differ between the vARIANCE study and the EudraVigilance reports.

Divergent results were observed concerning the time to onset of the angioedema. While in 70% of patients in the vARIANCE study, the angioedema occurred after 1 year of treatment, the angioedema occurred within the first year in about 80% of the EudraVigilance reports. According to the literature, ACEi-induced angioedema most frequently develops within the first weeks of treatment but may occur at any time, even years after initiating ACEi therapy.² Our findings in the vARIANCE study support the occurrence of angioedema later than a year from initiation, and this, as a clinical translation, is an aspect physicians should be aware of.

4.8 Severity and treatment of the angioedema cases

Overall, ACEi- and ARB-induced angioedema were quiet serious, necessitating hospital admission in every second patient in the vARIANCE study and every third patient from the EudraVigilance reports. ACEi-compared to ARB-induced angioedema were reported as more serious and resulted in hospitalization or prolongation thereof more often.^{23, 52} This was also reflected in the EudraVigilance reports in our analysis.

In patients with respective information, antihistamines and corticosteroids were most frequently administered as a therapy, as also reported in the literature.⁵³ Treatment with antihistamines and corticosteroids was considered to be effective in some cases, which may indicate possible effectiveness although stated otherwise.^{2, 53, 54} Another explanation could be the discontinuation of the ACEi/ARBs (especially for those lasting longer than 3 days) as well as the spontaneous remission, which could lead to an improvement of symptoms.

4.9 Proportion of ACEi-versus ARB-induced angioedema in both datasets

In Germany, ACEi is prescribed more frequently than ARB (Supporting Information S7: Appendix 7). With an incidence of 1.79 for ACEi and 0.62 for ARBs in 1000 people, ACEi-induced angioedema is almost three times more common than ARB-induced angioedema.⁵² A clearly higher proportion of ACEi- than ARB-induced angioedema was also observed in the vARIANCE study but not in the EudraVigilance reports. The combination sacubitril/valsartan was the ARB most frequently reported in EudraVigilance reports. Other studies did not show a higher risk of angioedema associated with sacubtiril/valsartan compared with ACEi or ARBs.^55-58 Thus, reporting bias has to be considered.⁴⁰