Figure 2 shows the time course of prescribed DDD of the various NKA-inhibitors in Germany from 1990 to 2023. Figure 3 shows that all clinical trials registered on ClinicalTrials focused on digoxin. We identified eight NKA-inhibitor-trials within the investigated timespan. The study analysis found one major clinical study investigating the efficacy and prognostic treatment effect (The Digitalis Investigation Group – DIG 1997). Three post-hoc analyses of the DIG study were performed. The DIG (1997) represents the only randomized controlled trial (RCT) that was conducted with a masked design and a placebo group (Table 1). Although the DIG-trial showed a reduction in hospitalizations, it failed to prove a significant mortality reduction for digoxin [3]. Figure 2A and B, indicate that this finding only had minimal impact on the clinical use of this substance. Subsequent post-hoc analyses showed that digoxin may be able to reduce mortality in patients with heart failure and reduced ejection fraction (HFrEF) when administered in the correct dosage [6, 7]. Nevertheless, this effect has to be verified in a clinical RCT. In addition to the DIG, there were two other clinical RCTs that investigated digoxin: AFFIRM (2002) and RATE-AF (2020).

A Key events for NKA-inhibitors within the investigated timespan. According key-event legend on the right. (Data source: WidO). B Key events for NKA-inhibitors within the investigated timespan. According key-event legend on the right. (Data source: WidO)

However, these two studies showed deficiencies in placebo- and masking aspects [8, 9]. Although, the RATE-AF study had a relatively small population size, with only 160 participants [9]. Looking ahead, the multicenter, randomized, double-blind, placebo-controlled DIGIT-HF trial had the objective to evaluate the clinical evidence of digitoxin in patients with chronic HFrEF in addition to guideline-based therapy [10]. The primary composite endpoint was all-cause mortality or hospitalization for heart failure [10]. The DIGIT-HF trial identified a significant reduction in the composite endpoint for certain digitoxin plasma concentrations for patients with chronic HFrEF [5]. The DIGIT-HF trial currently is the only clinical trial for the investigation of digitoxin as the primary substance and the only trial providing scientific evidence for the use of digitoxin in clinical practice. In general, there seems to be a clear predominance of studies investigating digoxin, compared to digitoxin (Fig. 3). Until publication of the DIGIT-HF trial, most of the scientific evidence relied on the results of the DIG trial in 1997, which investigated digoxin as primary substance. Since the DIGIT-HF trial was registered on EudraCT, it was therefore not included in our study count on ClinicalTrials.gov in Fig. 3. In 2023 Merck, the main producer of digitoxin drugs permanently discontinued the production of Digimerck® due to economic reasons and a production material shortage (Bauersachs et al., 2023; Gelbe Liste, 2022). This led to significant DDD-reduction in all substances after 2021 (Figs. 2B).

Search results of total study count of NKA-inhibitors on ClinicalTrials displayed in a column chart. blue: total studies with indications relevant to this analysis found for the individual NKA-inhibitor, orange: total studies with results available. DIGIT-HF was registered on EudraCT and is therefore not represented in our ClinicalTrials.gov study count

For the investigated timeframe (1990–2023), the DDD of all four NKA-inhibitor substances are continuously decreased. The pre-/post-study analyses of the five major studies DIG (1997), AFFIRM (2002), Rathore et al. (2003), Ahmed et al. (2006) and Castagno et al. (2012) found negative APC values after each study, indicating a decline in DDD after the releasing date of the individual study. Although the pre/post analyses suggest a temporal association, they did not formally test the differences between the periods. The actual changes in DDD-trends were identified by our whole-Joinpoint trend analyses over the complete investigation: For digitoxin the analysis identified significant DDD-trend-changes after 1994 and 2012 (Fig. 4A). Digoxin showed significant DDD-trend-changes after 2006, 2015 and 2021 (Fig. 4B). For acetyldigoxin the analysis identified significant DDD-trend-changes after 2003 and 2021 (Fig. 5A). For Metildigoxin the significant DDD-trend-changes were identified after 1996, 2015 and 2021 (Fig. 5B). The detailed data-results of the whole-joinpoint analyses are displayed in Tables S6-9 in the supplement. Our analysis found no direct association between the study release dates and a subsequent change in DDD-trends. All DDD numbers showed a statistically significant DDD-trend change after 2021: increased DDD-volumes for digoxin, acetyldigoxin and metildigoxin, digitoxin-DDD are decreased.

Multi-panel display of joinpoint analyses for digitoxin and digoxin use (DDD/1000-inhabitants/day) from 1990–2023. (A) Digitoxin; (B) Digoxin. Brown square points display the DDD/1000-inhabitants/day. The coloured lines represent the automatically recognized trend segments. A different colour represents a different trend segment. Detailed data tables according the DDD/1000-inhabitants/day for the four NKA-inhibitors (APC, Upper- and lower CI, P-values) are displayed in Tables 2–5 in the supplementary material. Y-axis was adjusted individually for each substance to improve visualization of substance specific trends. Comparison should focus on relative trends

Multi-panel display of joinpoint analyses for digitoxin and digoxin use (DDD/1000-inhabitants/day) from 1990–2023. (A) Acetyldigoxin; (B) Metildigoxin. Brown square points display the DDD/1000-inhabitants/day. The coloured lines represent the automatically recognized trend segments. A different colour represents a different trend segment. Detailed data tables according the DDD/1000-inhabitants/day for the four NKA-inhibitors (APC, Upper- and lower CI, P-values) are displayed in Tables 2–5 in the supplementary material. Y-axis was adjusted individually for each substance to improve visualization of substance specific trends. Comparison should focus on relative trends

Total annual number of reported exposure cases with NKA-Inhibitors of the GIZ-Nord catchment area presented in a combined column chart: digitoxin (blue), digoxin (orange) and unspecified digitalis glycosides (green). (Data source: GIZ-Nord Poison Centre)

Although the absolute number of exposure cases remains low, reported exposure cases for digitoxin increased from 8 to 10 cases per year to 30–35 cases per year from 1996 to 2022 (Fig. 6), therefore tripling over the investigated timeframe. The reported exposure cases by the GIZ-Nord poisons center show a markedly different age distribution for NKA-inhibitors compared to other reported substances (Fig. 7). Reported cases with NKA-inhibitors are disproportionally more represented in older age groups, especially in patients > 70 years. Reported exposure cases with other substances are more uniformly distributed across age groups.

Percentage distribution of reported exposure cases with NKA-inhibitors across the individual age groups presented in a column chart based on data received by the GIZ-Nord: Registered exposure cases with NKA-inhibitors (orange), registered general drug exposure cases (blue) (Data source: GIZ-Nord)

More than 50% of the reported exposure cases with NKA-inhibitors are found within the age group >70 years, compared to less than 10% in the exposures with other drugs. This highlights the overrepresentation of elderly patients in the reported exposure cases with NKA-inhibitors.

Joinpoint analyses for digitoxin and digoxin related exposure cases from 1996–2022 based on GIZ-Nord data. (A) Digitoxin; (B) Digoxin. Brown square points display the reported exposure cases per 100.000-inhabitants. The coloured lines represent the automatically recognized trend segments. Detailed results (APC, Upper- and lower CI, P-values) are displayed in Tables 10–13 in the supplementary material. Absolute case numbers shown in Fig. 6. Y-axis was adjusted individually for each substance to improve visualization of substance specific trends.Comparison should focus on relative trends

Joinpoint analyses for digitoxin and digoxin related exposure cases per 1 mio. DDD from 1996–2022 based on a model calculation of GIZ-Nord data and exposure-adjusted DDD values. (A) Digitoxin; (B) Digoxin. Brown square points display the reported exposure cases per 1 mio. DDD. The coloured lines represent the automatically recognized trend segments. Detailed data tables (APC, Upper- and lower CI, P-values) are displayed in Tables 10–13 in the sup–plementary material. Absolute case numbers presented in Fig. 6.Y-axis was adjusted individually for each substance to improve visualization of substance specific trends. Comparison should focus on relative trends

The population adapted joinpoint calculation for reported digitoxin-related exposure cases/100.000 inhabitants shows a significant and continuous increase from 1995 to 2022 (APC = + 4.65%/year, p < 0.00001) (Fig. 8A). For digoxin our population adapted analysis showed a strong increase in reported exposure-cases/100.000 inhabitants from 1996 to 2005 (APC = 14.97%, p < 0.004). From 2005 to 2022 the reported exposure cases/100.000 inhabitants are decreasing (APC = -6.01%, p < 0.000001) (Fig. 8B). The exposure adapted joinpoint analysis per 1mio. digitoxin DDD showed an even more drastic increase (APC = + 11.30%/year, p < 0.000001), indicating a substantial increase in reported exposure cases relative to the estimated prescription volume (Fig. 9A). The exposure adapted joinpoint analysis per 1mio. digoxin DDD found a strong and significant increase from 1996 to 2006 (APC = 38.32%, p < 0.000001) followed by an insignificant stable trend from 2006 to 2022 and large variabitlity of individual values (APC = 1.18%, p = 0.79) (Fig. 9B). The reported exposure cases show noticeable year-to-year variability especially for digoxin, which should be considered for the interpretation of the magnitude of the overall trend. Detailed results are shown in the supplementary material (Tables 10–13).

In the DDD-half-time calculations for NKA-inhibitors, Digitoxin exhibited the highest value whereas Digoxin demonstrated the lowest calculated half-time:

Calculation of the DDD-half-time for the individual NKA-inhibitor substances

With the longest DDD-half-time digitoxin showed the flattest and most steady decline curve, whereas digoxin showed the steepest decline and therefore the shortest DDD-half-time (Fig. 10). Acetyldigoxin and Metildigoxin lie in-between with declines intermediate to digoxin and digitoxin (Fig. 10).