Hospitalisation for acute heart failure and in-hospital mortality before, during and after the COVID-19 pandemic in France: a nationwide cohort study from 2013 to 2024

Abstract

Introduction

Healthcare systems were reorganised in 2020 to manage the COVID-19 pandemic. Despite their urgent status, hospital admissions for acute heart failure (AHF) were reported to decline from 9% to 66% worldwide between 2020 and 2021, with divergent findings regarding in-hospital mortality. This study aimed to investigate in detail the evolution of AHF hospitalisations and in-hospital mortality in France from 2013 to 2024.

Methods

Based on the 2.9 million AHF hospitalisations recorded in France from 2013 to 2024, yearly numbers of hospitalisations and deaths expected in years 2020–2024 were estimated using a Poisson regression model, with 2013–2019 as the reference period. The differences between observed and expected event counts in the years 2020–2024 were used to quantify the disruptions that occurred since the emergence of the pandemic.

Results

A total deficit of −222 913 (−223 908 to −221 926) (mean (95% CI)) AHF hospitalisations was estimated for the years 2020–2024, corresponding to a 16.1% decrease compared with pre-pandemic trends. The yearly reduction in AHF hospitalisations worsened over time, from −39 268 (–39 685 to –38 847) fewer cases in 2020 to –55 521 (–55 984 to −55 051) in 2024. Between 2020 and 2024, 7794 (7557 to 8028) excess in-hospital deaths were estimated, corresponding to an 8.4% excess compared with pre-pandemic trends. From 2021 to 2024, this excess ranged from 9.6% to 16% for females compared with 7.1% to 11.1% for males.

Conclusions

The apparent long-lasting changes in the management of patients with AHF in France observed since the COVID-19 pandemic emergence, particularly among females, suggest further research for better understanding the sustained observed disruptions.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• In 2020 and 2021, hospitalisations for acute heart failure were reported to decline worldwide following the onset of the COVID-19 pandemic. However, findings on concomitant in-hospital mortality have remained unclear, and little is known about whether these disruptions persisted through 2022–2024.

WHAT THIS STUDY ADDS

• Analyses of exhaustive French national data indicate that the decline in admissions observed in 2020 persisted and even worsened through 2024, with an overall decrease of 16.1%. In parallel, in-hospital mortality was estimated in each year from 2020 to 2024, and the resulting excess corresponded to a cumulative increase of 8.4%. Females were more impacted than males by both disruptions.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• This study highlights critical warnings on ongoing disruptions affecting patients hospitalised for acute heart failure in France and identifies the subpopulations most impacted. These findings might contribute to guide targeted mitigation strategies and to enhance the preparedness of national health systems for future health crises.

Introduction

At the end of 2019, COVID-19 emerged in Wuhan, China,¹ and by 30 January 2020, the WHO declared it a Public Health Emergency of International Concern (PHEIC) due to its rapid global spread.² In response to the crisis, many countries implemented unprecedented measures, including lockdowns and significant reorganisation of healthcare systems.³ In France, non-urgent hospitalisations were deferred to mitigate the virus spread and prioritise the care of COVID-19 patients.⁴ However, despite the urgent nature of acute heart failure (AHF), a decline in hospitalisations for this condition was observed early in the pandemic period, in March 2020.⁵ A decrease in the incidence of AHF hospitalisations, not only in 2020, but also in 2021 and 2022, was highlighted in a more recent study published in 2024, which focused on the epidemiology of hospitalised heart failure in France.⁶ However, this latter study considered only the first hospitalisation of the year without any prior hospitalisations in the preceding 5 years and did not provide precise quantification of either the decrease in hospitalisations or in-hospital mortality in the pandemic period. Several studies have reported similar reductions in AHF hospital admissions across Europe^{7 8} and worldwide,9,11 with decreases ranging from 9% to 66% between 2020 and 2021. However, the reported in-hospital mortality of AHF admissions during the pandemic period varied widely, with some studies reporting an increased mortality,12,14 while others reported no significant change.15,17 Additionally, little is known about the persistence of the disruptions, including trends in hospitalisation rates and mortality in 2023 and 2024, while the WHO declared the end of the phase of PHEIC on 5 May 2023.¹⁸

Given the uncertainties regarding the extent of the decline in AHF hospitalisations, the corresponding in-hospital mortality rates, and the evolution of these patterns in 2023 and 2024, this study aimed to provide detailed investigations of the disruptions during and after the pandemic period in France. Hospitalisation data from 2013 to 2019 were analysed and used as a reference to model the historical evolution in the numbers of AHF hospitalisations and corresponding in-hospital deaths, allowing for estimates of expected values from 2020 to 2024 in the absence of the pandemic. The primary metrics to quantify the disruptions occurring during the 2020–2024 period were the differences between observed and expected hospitalisations and deaths. One may hypothesise that the decline in AHF admissions observed during the early phase of the pandemic has likely tapered as the pandemic’s intensity decreased, potentially with a catch-up phenomenon. By 2023 and into 2024, a return to pre-pandemic trends might have been expected. Hypothesising what should be the evolving trends in corresponding in-hospital mortality is more challenging. On the one hand, hospital admissions may have been restricted to the most severe emergencies, which could have potentially increased in-hospital mortality. On the other hand, it is possible that more critical patients succumbed to COVID-19, leaving fewer severe AHF cases, which could lower in-hospital mortality. In any scenario, we expect mortality rates to return, at most, to pre-pandemic levels after 2023.

In summary, the overall objective of this study was to provide a comprehensive overview of the evolution in the management of patients with AHF over the past 12 years and to evaluate the full impact of the COVID-19 pandemic period on hospitalisations and associated mortality in France, from 2020 to 2024.

Materials and methods

Data sources

Data on hospitalisations for AHF come from the French Hospital Discharge Database (PMSI), which is part of the National Health Data System (SNDS). The SNDS is a national claims database designed to reimburse care via the French Health Insurance System, covering nearly 100% of the population receiving healthcare in France.¹⁹ With the growing interest in the scientific medical community for real-world data derived from large administrative healthcare databases, the SNDS has, for instance, previously been used to study the impact of the pandemic period on acute cardiovascular diseases20,22 and other diseases.^{23 24} Data on the French population structure were also used in this study and are available in open access from the French National Institute for Statistics and Economic Studies (Insee).²⁵ There was no missing variable value in our database queries, and therefore, analyses assumed the absence of missing data. Nevertheless, any lack in the recording process of data in the SNDS cannot be verified.

Population under study

This open cohort study, covering the entire French population from 1 January 2013 to 31 December 2024, is reported following the Strengthening the Reporting of Observational Studies in Epidemiology guidelines.²⁶ As the age structure of the French population is reported on 1 January of each year, the age considered for participants was that on 1 January and data were aggregated by year. Patients aged 99 years or more were handled in a single category. Following the Sex and Gender Equity in Research guidelines,²⁷ sex was always taken into account in the study design and report.

Outcomes of interest

Two outcomes were identified and analysed during the study period: the annual number of AHF hospitalisations and the number of associated in-hospital deaths. AHF hospitalisation events were identified in the SNDS database based on the International Classification of Diseases, 10th revision (ICD-10), using an algorithm developed by the healthcare expenditures and conditions mapping²⁸ (see online supplemental text 1). In-hospital death was defined as an AHF hospital stay that ended with the death of the patient. Deaths attributed to COVID-19 among patients who died during hospitalisation for AHF were assessed in a supplemental analysis (see online supplemental analysis 1).

Analysing the evolution of the numbers of AHF hospitalisations and associated in-hospital deaths.

To quantify the impact of the pandemic period on the outcomes of interest, a two-step approach was used, comparing the expected and observed numbers of AHF hospitalisations and associated in-hospital deaths. In the first step, the pre-pandemic period (2013–2019) was used as a reference to model the natural evolution of both outcomes just before the pandemic onset. The modelling was based on a Poisson regression, detailed in the next subsection. In the second step, for each outcome of interest, the corresponding regression model was used to predict the expected values for years 2020–2024, assuming the natural evolution observed during the pre-pandemic period had not been disrupted by the pandemic. The expected values were then compared with the observed occurrences, enabling us to characterise how the natural evolution of AHF hospitalisations and corresponding in-hospital deaths was disrupted following the emergence of the COVID-19 pandemic (see online supplemental figure 1). To summarise, the models were fitted using data from 2013 to 2019 and then applied to generate expected numbers of hospitalisations and deaths for 2020–2024, with the annual differences between observed and expected values quantifying the changes that occurred following the onset of the pandemic. A similar approach has been previously applied to estimate expected mortality levels for the calculation of cause-specific excess mortality²⁹ or all-cause excess mortality.^{30 31}

Statistical analysis

A Poisson regression model was used to estimate the numbers of expected AHF hospitalisations and in-hospital deaths after the arrival of the pandemic, considering pre-pandemic years 2013–2019 as the reference period.

The model included age as a year-specific categorical variable to accurately capture potential non-linear effects of age on the numbers of hospitalisations and in-hospital deaths, an interaction between age and sex to precisely estimate sex differences across age groups and calendar year as a numerical variable to assess yearly temporal trends in the numbers of hospitalisations and deaths. To estimate expected AHF hospitalisations while accounting for changes in population size and age structure over the study period, the natural logarithm of the population was used as an offset term. For in-hospital mortality among AHF admissions, the number of AHF admissions was used as an offset term to account for changes in hospital load over time. Consequently, the mortality estimates were interpreted taking into account the changes in admission volumes during the pandemic. The Akaike information criterion was employed as a selection criterion to optimise goodness-of-fit and avoid overfitting (online supplemental table 2). Calibration of the model was visually assessed by comparing observed and predicted data between 2013 and 2019 (online supplemental figure 1). To investigate seasonal periodicity, additional analyses were performed at both monthly and weekly temporal resolutions, employing sinusoidal and spline-based models. The details of these models, along with graphical visualisations of the predicted trends and their impact on the final estimates, are available in the (online supplemental material, analysis 2 and figures 2-4). The Poisson regression used for modelling the number of hospitalisations was:

$${\displaystyle {\displaystyle \begin{array}{l}\log \left({\widehat{eH}}_{i,j,k}\right)=\log \left({\text{Population}}_{i,j,k}\right)+{\beta }_{h_0}+{\beta }_{h_1}{\text{Sex}}_i+\\ {\beta }_{h_2}{\text{Age}}_j+{\beta }_{h_3}{\text{Year}}_k+{\beta }_{h_4}{\text{Sex}}_i{\text{Age}}_j\end{array}}}$$

With$H_{i,j,k}$ the number of observed AHF hospitalisations (in the SNDS database) during year k ($kϵ\left\{2020,2021,2022,\mathrm{2023,2024}\right\}$) in the stratum of persons with sex i and age j and ${\widehat{eH}}_{i,j,k}$ the corresponding number of AHF hospitalisations expected to occur with the model during year k in the same population stratum.

The resulting difference between observed and expected values is ${\widehat{dH}}_{i,j,k}=H_{i,j,k}-{\widehat{eH}}_{i,j,k}$, and the disruption observed year k compared with the expected natural evolution of the number of AHF hospitalisations was calculated as follows:

$${\displaystyle {\displaystyle {\displaystyle {\widehat{dH}}_k=\sum _{i\in \{\text{males},\text{females}\}}\sum _{j=0}^{99}{\widehat{dH}}_{i,j,k}}}}$$

Similarly, the Poisson regression used for modelling the number of in-hospital deaths among AHF hospitalisations was:

$${\displaystyle {\displaystyle \begin{array}{l}\log \left({\widehat{eM}}_{i,j,k}\right)=\log \left(H_{i,j,k}\right)+{\beta }_{m_0}+{\beta }_{m_1}{\text{Sex}}_i+\\ {\beta }_{m_2}{\text{Age}}_j+{\beta }_{m_3}{\text{Year}}_k+{\beta }_{m_4}{\text{Sex}}_i{\text{Age}}_j\end{array}}}$$

and the disruption observed year k compared with the expected natural evolution of the number of in-hospital deaths was calculated as follows:

$${\displaystyle {\displaystyle {\displaystyle {\widehat{dM}}_k=\sum _{i\in \{males,females\}}\sum _{j=0}^{99}{\widehat{dM}}_{i,j,k},with{\widehat{dM}}_{i,j,k}=M_{i,j,k}-{\widehat{eM}}_{i,j,k}}}}$$

All analyses were performed with statistical software R V.4.0.2 (R Foundation for Statistical Computing, Vienna, Austria), package ggplot2 was used to generate figures, package stats was used for the generalised linear models and 95% CIs were estimated over 10 000 bootstrap replications.

Results

Population under study

Between 2013 and 2024, the French population increased from 65.6 million to 68.4 million, and the median age rose from 40 to 42 years (see study flow diagram in figure 1). During the study period, which included a total of 805 627 212 person-years, 2 922 319 AHF hospitalisations occurred at 83 (74–88) years old (median age (IQR)), including 246 080 corresponding in-hospital deaths at 86 (80–91) years old.

Figure 1. Study flow diagram. Abbreviations used: P corresponds to the population size on 1 January of this given year, H corresponds to the total number of acute heart failure hospitalisations and D to the total number of death events that occurred in France each year. Age is reported by its median value (IQR), and the percentage of data concerning females is also reported.

Figure 2 presents the number of AHF hospitalisations that occurred each week over the study period, with the corresponding distribution by age and sex. This synthesis picture of exhaustive national data in France over time provides key insights into AHF hospitalisation trends. First, hospitalisations predominantly involved patients aged between 70 and 95 years. Second, data present seasonal patterns, with a peak at the beginning of each year and a decrease during summer months. Third, the number of AHF hospitalisations gradually increased from 2013 to 2019. Finally, a sharp decline in hospital admissions was notable during the first wave of the pandemic in April 2020, followed by a series of fluctuations through to 2024, without ever returning to pre-pandemic levels. These visual trends are corroborated by the numerical data in the flow diagram (see figure 1).

Figure 2. Occurrences of AHF hospitalisations observed each week in males and females according to age in France, years 2013–2024 (n=2 922 319). Main panel: weekly number of AHF hospitalisations according to age. For example, in January 2019, the colour gradient is red for individuals aged between 85 and 90, corresponding to a weekly number of around 300 hospitalisations. Top panel: cumulative numbers (in males and females) of AHF hospitalisations each week in France from 2013 to 2024, broken down by sex. Right panel: cumulative number (over time) of AHF hospitalisations according to age in males and females. AHF, acute heart failure.

Evolution in the number of AHF hospitalisations

As illustrated in figure 3, the numbers of AHF hospitalisations observed in years 2020–2024, that is, after the emergence of the pandemic, consistently remained below the expected numbers based on the trend from the previous seven pre-pandemic years (2013–2019), for both males and females. The decline pattern intensified over time, starting with −39 268 (95% CI −39 685 to −38 847) fewer hospitalisations than expected in 2020 and peaking at −55 521 (95% CI −55 984 to −55 051) fewer in 2024. These estimates correspond to annual reductions ranging from −11.3% to −19.5%. Considering the overall pandemic period, there were −222 913 (95% CI −223 908 to −221 926) fewer AHF hospital admissions, representing a −16.1% reduction from the expected figures. Females were more affected than males each year, with a total of −117 459 (95% CI −118 142 to −116 774) fewer hospitalisations than expected (−17.2%) compared with −105 454 (95% CI −106 169 to −104 755) fewer hospitalisations (−15.0%) for males between 2020 and 2024. Detailed annual estimates of the decline in AHF hospitalisations, stratified by sex, are provided in table 1. As shown in online supplemental figure 5, the drop in hospitalisations involved patients aged 83 (77–88) years between 2020 and 2024, while the median age of hospitalised patients during the same period was 82 (74–88) years.

Figure 3. Differences between observed and expected numbers of acute heart failure hospitalisations, France, 2013–2024. *Observed minus expected. Main panel, annual difference between observed and expected hospitalisations according to age. For example, in 2024, the colour gradient is dark blue for individuals aged 85, corresponding to 1500 fewer admissions than expected. Top panel, annual cumulative (in males and females) difference between observed and expected hospitalisations, years 2013–2019 was used as the reference period. Right panel, cumulative (over time, years 2020–2024) difference between observed and expected hospitalisations according to age and sex.

Table 1. Differences (observed − expected) in the numbers of hospitalisations for acute heart failure, and in the numbers of corresponding in-hospital deaths, France, 2020−2024.

	Year
Difference (observed – expected)*	2020	2021	2022	2023	2024
Number of hospitalisations	−39 268 (−39 685 to −38 847) (−14.5%)	−31 041 (−31 467 to −30 621) (−11.3%)	−46 349 (−46 788 to −45 901) (−16.7%)	−50 734 (−51 193 to −50 275) (−18.1%)	−55 521 (−55 984 to −55 051) (−19.5%)
Females	−22 750 (−23 044 to −22 451) (−16.9%)	−17 976 (−18 275 to −17 676) (−13.3%)	−24 805 (−25 117 to −24 490) (−18.2%)	−24 781 (−25 098 to −24 460) (−18.1%)	−27 147 (−27 474 to −26 819) (−19.6%)
Males	−16 518 (−16 817 to −16 226) (−12.1%)	−13 065 (−13 373 to −12 765) (−9.4%)	−21 544 (−21 857 to −21 231) (−15.3%)	−25 953 (−26 277 to −25 627) (−18.1%)	−28 374 (−28 719 to −28 044) (−19.4%)
Number of in-hospital deaths	828 (729 to 928) (4.4%)	1625 (1517 to 1731) (8.3%)	2427 (2323 to 2531) (13.2%)	1739 (1634 to 1844) (9.5%)	1175 (1068 to 1281) (6.5%)
Females	402 (331 to 471) (4.3%)	932 (857 to 1007) (9.6%)	1473 (1400 to 1546) (16.0%)	1025 (950 to 1101) (11.1%)	729 (651 to 805) (8.0%)
Males	426 (354 to 499) (4.5%)	693 (615 to 768) (7.1%)	954 (881 to 1026) (10.4%)	714 (641 to 788) (7.9%)	446 (371 to 522) (4.9%)

In the first part of the table, the negative difference between observed and expected AHF hospitalisations corresponds to a drop in hospitalisations.

In the second part of the table, the positive difference between observed and expected number of deaths corresponds to excess mortality during AHF hospitalisation.

^*Differences expressed as: mean difference (95% CI) (disruption of the evolution relative to the expected trend, ie, difference/expected number×100); reference period for estimating expected numbers=years 2013–2019.

AHF, acute heart failure.

Evolution in the number of in-hospital deaths among AHF hospitalisations

Figure 4 illustrates the evolution of in-hospital mortality among patients hospitalised for AHF in the years 2020–2024. Throughout the overall pandemic period, in-hospital mortality exceeded expectations for both males and females. This excess mortality steadily increased from 828 (95% CI 729 to 928) excess deaths in 2020 to 2427 (95% CI 2323 to 2531) in 2022, before decreasing to 1175 (95% CI 1068 to 1281) in 2024. These excess deaths represented relative increases ranging from 4.4% to 13.2% compared with the pre-pandemic trend. The cumulative toll over the years 2020–2024 reached 7794 (95% CI 7557 to 8028) excess deaths (8.4%). Since 2021, the excess in-hospital mortality has affected more females than males, with corresponding relative increases ranging from 9.6% to 16.0% for females, compared with 7.1% to 10.4% for males. Over the 2021–2024 period, a total of 4159 (95% CI 4009 to 4308) in-hospital deaths in excess among females was estimated (relative excess of 11.2%), compared with 2807 (95% CI 2660 to 2955) corresponding excess deaths among males (relative excess of 7.6%), yielding a female-to-male ratio of 1.47 (95% CI 1.38 to 1.57). Detailed annual estimates by sex are provided in table 1. As shown in online supplemental Figure 6, patients contributing to the excess in-hospital mortality had a median age of 86 years (78–91), compared with 87 years (80–91) for those observed.

Figure 4. Differences between observed and expected numbers of in-hospital deaths, France, 2013–2024. *Observed minus expected. Main panel, annual difference between observed and expected in-hospital deaths according to age. For example, in 2022, the colour gradient is red for individuals aged 85, corresponding to 100 excess in-hospital deaths. Top panel, annual cumulative (in males and females) difference between observed and expected in-hospital deaths, years 2013–2019 was used as the reference period. Right panel, cumulative (over time, years 2020–2024) difference between observed and expected in-hospital deaths according to age and sex.

Discussion

Key results

In this open cohort study of the French population from 2013 to 2024 (805 627 212 person-years), trends in AHF hospitalisations and corresponding in-hospital mortality were analysed to accurately quantify the extent of disruption during and after the COVID-19 pandemic period (2020–2024). The study provides evidence of a major and prolonged disruption since the onset of COVID-19, which remained substantial in 2024, even after the end of the pandemic. Specifically, a 14.5% decline in hospital admissions occurred in 2020, and, surprisingly, this decline worsened to 19.5% in 2024. The decrease, observed throughout the pandemic period and persisting even in 2024, affected both sexes but was more pronounced among females. As shown in online supplemental figure 5, the decline in hospitalisations was more pronounced among older patients. This trend is further supported by the flow diagram, which illustrates that the median age of hospitalised patients increased from 2013 to 2019, reflecting the ageing population in France. However, during the pandemic period, a decrease in the median age of patients hospitalised for AHF was observed, suggesting that the decline in hospitalisations may have been more pronounced among the oldest patients. As shown in online supplemental Analysis 3, the decline in hospitalisations appears to be specific to AHF. Indeed, while the healthcare system experienced global disruptions since 2020, the reduction in hospitalisations was even more pronounced in patients with AHF compared with the general population and this gap widened over time (online supplemental figure 7 and table 3).

In contrast to the decline in AHF hospitalisations, in-hospital mortality exceeded expectations, with excess mortality ranging from 4.4% to 13.2% between 2020 and 2024. This corresponded to an estimated total of 7794 (95% CI 7557 to 8028) excess in-hospital deaths, representing an 8.4% increase compared with the expected mortality based on the pre-pandemic trend. Similar to the decline in hospital admissions, the increase in in-hospital mortality affected both sexes throughout the pandemic and after and was more pronounced in females. As illustrated in online supplemental figure 6, the rise in in-hospital mortality extended beyond the oldest patient groups.

Interpretation

The study confirms previous national observations regarding the decline in AHF hospitalisations between 2020 and 2022, while providing more detailed insights into the magnitude of the disruption. Additionally, it extends the analysis to include 2023 and 2024, offering a comprehensive assessment of the pandemic period’s long-term impact thanks to nationwide data spanning 12 consecutive years. The drop in hospitalisations, ranging from 11.3% to 14.5% in 2020 and 2021, aligns with figures reported in Europe^{7 8} and worldwide.9,11 Importantly, while it was anticipated that after the widespread vaccination against COVID-19, and the reduction in pandemic severity from 2022 to 2024, the situation would return to the pre-pandemic state, the study evidenced that the disruption worsened in the following years, peaking at 19.5% fewer hospitalisations in 2024.

Two contrasting effects of the pandemic may have contributed to the substantial decline in AHF hospitalisations. First, a series of protective effects related to the pandemic period could have led to a reduction in AHF incidence and, consequently, fewer hospitalisations. Key changes in population behaviours since 2020, such as widespread mask usage, have reduced the transmission of respiratory viruses, which are well-known to increase the risk of acute decompensation in patients with AHF.^{32 33} Shifts in healthcare practices, including the rapid expansion of telemedicine,^{34 35} the introduction of new medications³⁶ and major initiatives by the National Health Insurance since 2022,³⁷ may have improved patient care and further contributed to the reduction in hospitalisations for AHF. While these factors likely explain part of the decrease in hospitalisations, they do not fully account for the abrupt and persistent nature of this change, nor do they clarify why older individuals and females were particularly affected.

Second, a series of detrimental effects from the pandemic may have led to premature deaths before hospitalisation, thus reducing the overall number of hospitalisations. The widespread impact of the pandemic period resulted in a sudden and prolonged period of excess mortality in France,^{38 39} which is consistent with the observed decline in hospital admissions for the years 2020–2022. The major risk factors for COVID-19 mortality, such as older age and comorbidities,⁴⁰ closely match the profile of patients typically hospitalised for AHF. However, this direct effect of the pandemic cannot fully explain the decrease in AHF hospitalisations. COVID-19-related deaths officially decreased from 69 000 in 2020⁴¹ to 5572 in 2023,⁴² while the present study results indicate that the decline in the number of AHF hospitalisations was 39 268 fewer admissions in 2020 and increased to 50 734 in 2023. More importantly, in the post-pandemic period, in 2024, although the number of officially reported COVID-19-related deaths dropped to just 96,⁴³ the decline in AHF hospitalisations reached its peak, with 55 521 fewer admissions. Beyond direct competitive effects, some negative indirect consequences of the pandemic may also have contributed to premature death.⁴⁴ One cannot exclude that some factors, such as fear of seeking hospital care⁴⁵ or the documented reduction in the initiation of cardiovascular risk-related medications in France in 2020,⁴⁶ might have contributed to an increase in out-of-hospital AHF-related deaths before hospitalisation. Indeed, studies have reported a significant increase in out-of-hospital AHF-related deaths occurring at home during 2020, with rises ranging from 25% to 35% compared with the pre-pandemic period.^{47 48} Numerous studies have documented gender inequality in the management of cardiovascular care,49,51 making it likely that, during a crisis, females were among the first to suffer the negative consequences. Notably, a rise in cardiovascular mortality was observed in France from 2020 to 2022, particularly among females.⁵²

Study results indicate that in-hospital deaths among AHF hospitalisations exceeded expected levels throughout the pandemic. While some previous studies found no significant change in in-hospital mortality,12,14 others reported an increase.15,17 These discrepancies may be due to differences between countries or to the greater statistical power provided by the exhaustive data used in the present study. The increase in in-hospital mortality during the early phase of the pandemic could be attributed to COVID-19-related deaths. In 2020, for instance, 929 AHF-hospitalised patients died of COVID-19, potentially explaining the 828 excess deaths estimated for that year (online supplemental table 1). However, COVID-19 mortality among patients with AHF predominantly affected males and has declined over time. In contrast, in-hospital excess mortality has risen, particularly among females. Notably, in 2022, COVID-19 mortality could account for only 27% of the excess in in-hospital mortality observed in females. While COVID-19-related mortality, as a new cause of death emerging in 2020, could have explained the increase in in-hospital mortality among patients with AHF, this possibility does not fully align with the observed worsening mortality over time in this study, especially in females. A reduction in hospital capacity in France,⁵³ which has worsened since the start of the pandemic, may have restricted hospitalisations to the most severe cases, thereby increasing in-hospital mortality rates. The persistent excess mortality observed in this study is concerning and suggests a profound and long-lasting disruption in the healthcare system’s ability to manage patients with AHF, particularly females.

Strengths and limitations

This study is based on exhaustive data of AHF hospitalisations in France, and it covers not only the entire pandemic period but also one full post-pandemic year. The evolving patterns reported are reliably representative of the whole country. Moreover, the analyses include 7 years of pre-pandemic data, allowing historical trends before the onset of the pandemic to be taken into account and enabling reliable estimates of the subsequent perturbations. Most studies published to date have shorter reference periods and/or only estimate disruptions for parts of the pandemic. This study not only offers detailed results on the decline in hospital admissions and the increase in in-hospital mortality by year, sex and age, but also documents the evolution of these indicators over 12 consecutive years at the national level.

This study has several limitations. The main limitation lies in the lack of access to key clinical and behavioural variables. Important confounding factors such as weight, tobacco use and blood pressure, as well as clinical information such as biological and radiological results or the New York Heart Association function, were not available for analysis. This is due to the nature of the database, which is primarily administrative and does not yet incorporate clinical data. In addition, patient identification relied on ICD-10 codes, which are susceptible to coding errors; however, this method has been previously assessed and validated for the identification of acute heart failure cases.^{54 55} ICD-10 codes referring to left ventricular ejection fraction, including heart failure with reduced ejection fraction, heart failure with mildly reduced ejection fraction and heart failure with preserved ejection fraction, have only been reported in the national database since 2019.⁶ Consequently, these complementary clinical characteristics could not be taken into account in our longitudinal analyses covering the period 2013–2024. Additionally, as healthcare systems vary between countries, the findings of this national study cannot be directly generalised to other nations. However, as previously mentioned, similar patterns have been reported in studies conducted elsewhere. Several hypotheses were proposed regarding the impact of the COVID-19 pandemic on AHF hospitalisations and mortality. These include both protective and negative effects, direct impacts on mortality and indirect disruptions to the healthcare system. However, due to the complexity and multifactorial nature of these causes, it is not possible to draw causal conclusions. Nevertheless, despite the potential limitations of the study, the large sample size and the magnitude of the estimated disruptions strongly support the conclusion that a major disturbance in the management of patients with AHF occurred at the onset of the pandemic and has continued through to 2024.

Conclusion

This study reveals significant disruptions in the healthcare pathway for AHF, marked by a decline in hospitalisations and an increase in in-hospital mortality from the onset of the COVID-19 pandemic in 2020 through to 2024, with a particular impact on females. Further research devoted to exploring the underlying causes of these persistent disruptions is of primary importance for developing strategies to mitigate their impact.

Data availability statement

Data are available upon reasonable request.