You are here

Article Text

Article menu
Heart failure and cardiomyopathies
Original research
Outcomes in patients treated with loop diuretics without a diagnosis of heart failure: a retrospective cohort study
Free
  1. Joseph James Cuthbert1,2,
  2. Ireneous Soyiri3,
  3. Stephanie Jayne Lomax4,
  4. John Turgoose5,
  5. Ahmet Fuat6,
  6. Judith Cohen3,5,
  7. Andrew L Clark2
  1. 1 Clinical Sciences Centre, Hull York Medical School, University of Hull, Hull, UK
  2. 2 Department of Cardiology, Castle Hill Hospital, Cottingham, UK
  3. 3 Institute for Clinical and Applied Health Research, Hull York Medical School, University of Hull, Hull, UK
  4. 4 Cottingham Medical Centre, Cottingham, UK
  5. 5 Hull Health Trials Unit, Hull York Medical School, University of Hull, Hull, UK
  6. 6 GP, County Durham and Darlington NHS Foundation Trust, Darlington Memorial Hospital, Darlington, UK
  1. Correspondence to Dr Joseph James Cuthbert, Clinical Sciences Centre, Hull York Medical School, University of Hull, Hull HU6 7RU, UK; joe.cuthbert{at}hyms.ac.uk

Abstract

Background Loop diuretics are commonly prescribed in the community, not always to patients with a recorded diagnosis of heart failure (HF). The rate of HF events in patients prescribed loop diuretics without a diagnosis of HF is unknown.

Methods This was a propensity-matched cohort study using data from the Clinical Practice Research Datalink, Hospital Episode Statistics and Office of National Statistics in the UK. Patients prescribed a loop diuretic without a diagnosis of HF (loop diuretic group) between 1 January 2010 and 31 December 2015 were compared with patients with HF (HF group)—analysis A, and patients with risk factors for HF (either ischaemic heart disease, or diabetes and hypertension—at-risk group)—analysis B. The primary endpoint was an HF event (a composite of presentation with HF symptoms, HF hospitalisation, HF diagnosis (analysis B only) and all-cause mortality).

Results From a total population of 180 384 patients (78 968 in the loop diuretic group, 28 177 in the HF group and 73 239 in the at-risk group), there were 59 694 patients, 22 352 patients and 57 219 patients in the loop diuretic, HF and at-risk groups, respectively, after exclusion criteria were applied. After propensity matching for age, sex and comorbidities, patients in the loop diuretic group had a similar rate of HF events as those in the HF group (71.9% vs 72.1%; HR=0.92 (95% CI 0.90 to 0.94); p<0.001), and twice as those in the at-risk group (59.2% vs 35.7%; HR=2.04 (95% CI 2.00 to 2.08); p<0.001).

Conclusions Patients prescribed a loop diuretic without a recorded diagnosis of HF experience HF events at a rate comparable with that of patients with a recorded diagnosis of HF; many of these patients may have undiagnosed HF.

  • Heart failure
  • Epidemiology
  • Electronic Health Records
  • Pharmacology, Clinical

Data availability statement

Data may be obtained from a third party and are not publicly available. Data provided by CPRD.

https://doi.org/10.1136/heartjnl-2023-323577

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • The proportion of patients on primary care heart failure (HF) registers is much lower than expected from epidemiological data: there may be many patients missing from community HF registers. Furthermore, most patients only receive a diagnosis of HF after deteriorating to the point of requiring hospital admission; strategies to identify these missing patients may improve outcomes.

    WHAT THIS STUDY ADDS

    • Patients prescribed a loop diuretic without a recorded diagnosis of HF have an outcome profile similar to that of patients with a diagnosis of HF.

    • There were approximately three times as many patients prescribed a loop diuretic without a diagnosis of HF than were given a diagnosis of HF during the period January 2010–December 2015.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • There are many medications that cannot be prescribed without first performing a blood test; mandating natriuretic peptide testing before prescribing a loop diuretic would be an easy-to-implement change in practice. Downstream testing may identify HF at an early stage of the disease.

    • Establishing what proportion of patients taking a loop diuretic without a diagnosis of HF actually have underlying HF will allow for better planning of HF services and allow for adequate funding to support one of the biggest problems facing the modern National Health Service.

    Introduction

    Heart failure (HF) is a difficult diagnosis to make: symptoms are non-specific,1 clinical signs are difficult to elicit,2 natriuretic peptide testing is insensitive3 and echocardiography is prone to errors in measurement.4 5 Perhaps as a consequence, the true incidence and prevalence of HF are not clear.

    Epidemiological studies report the prevalence as being between 1% and 7% of the general population.6–10 The prevalence increases with age and affects more than 1 in 10 patients aged over 80 years.11 However, in the UK, the prevalence of HF recorded in primary care HF registers is 0.9%.12

    Loop diuretics are among the most commonly prescribed medications in primary care but they have few indications other than the treatment of venous congestion due to HF.13 An audit of clinical practice found that the prevalence HF according to the register was much the same as the prevalence of loop diuretic prescription among patients who did not have a diagnosis of HF. Patients prescribed a loop diuretic without a diagnosis of HF had a rate of HF hospitalisation or all-cause death of 25% after 2 years.14 It is possible a proportion of patients prescribed loop diuretics in the community have underlying HF.

    We used data from the Clinical Practice Research Datalink (CPRD), Hospital Episode Statistics (HES) admitted patient care (APC) records and Office of National Statistics (ONS) death records to assess the frequency of HF-related events in patients prescribed a loop diuretic without a diagnosis of HF compared with patients with a diagnosis of HF. Loop diuretics may also be prescribed for other conditions associated with an increased risk of HF such as hypertension or chronic kidney disease (CKD). Thus, in a second analysis, we also assessed the frequency of HF-related events in a control group of patients at risk of developing HF to compare outcome profiles.

    Methods

    Data sources

    The CPRD database contains anonymised longitudinal patient data on demographics, lifestyle, diagnoses, medications, investigations including blood tests and referrals collected from primary care across the UK.15

    Primary care electronic records were linked to the HES APC records and the ONS death records. HES APC and ONS death records are databases of all hospital admissions and deaths, respectively, in the UK. The cause of each is ascribed an International Classification of Diseases 10 (ICD-10) code.

    Study populations

    All patients aged over 18 years of age contributing data to the CPRD between 1 January 2010 and 31 December 2015, who had been registered with their practice for at least 1 year, with records deemed acceptable by CPRD quality control, and approved for linkage to the HES APC and ONS death record datasets were eligible for inclusion. The time frame was chosen to include the time when the National Institute for Health and Care Excellence introduced a guideline for the diagnostic process for HF (2010),16 and to allow the majority of patients to have at least 5 years of follow-up prior to the peaks of the coronavirus pandemic in 2020–2021.

    Patients were split into three groups: (1) patients prescribed a loop diuretic but who did not have a recorded diagnosis of HF (loop diuretic group); (2) patients with a new recorded diagnosis of HF (HF group); and (3) patients with a new diagnosis of HF risk factors—ischaemic heart disease (IHD) or a diagnosis of diabetes with a pre-existing diagnosis of hypertension, or vice versa (at-risk group).

    Case identification

    Read and ICD-10 code lists were generated from medical dictionary keyword searches, previously published literature6 and online clinical code repositories (online supplemental file appendix).17 The index dates were the date of the first medication code for a loop diuretic, the date of the first Read or ICD-10 code for HF for the HF group, and the date of the first Read code for IHD or the first Read code for diabetes in a patient with pre-existing hypertension (or vice versa) for the loop diuretic, HF and at-risk groups, respectively.18

    Supplemental material

    Various exclusion criteria were applied to the three groups (figure 1). Patients with an existing Read or ICD-10 code for HF predating 1 January 2010 (exclusion criteria 1–2), or a Read code specifically excluding HF before the index date (exclusion criterion 3), or whose date of death was prior to the index date (exclusion criterion 8) were excluded from all three groups.

    Figure 1
    Figure 1

    Cohort flow diagram and exclusion criteria. CPRD, Clinical Practice Research Datalink; DM, diabetes mellitus; HF, heart failure; HTN, hypertension; IHD, ischaemic heart disease.

    Patients with a Read code for HF in the 3 months after the index date (exclusion criterion 4), or hospitalisation for HF or death within 1 month of the index date (exclusion criterion 5) were presumed to have clinically evident HF at the time of loop diuretic prescription and were excluded from the loop diuretic group. To ensure a fair comparison between the groups, patients who met exclusion criterion 5 were also excluded from the HF and at-risk groups.

    Patients who underwent natriuretic peptide testing, echocardiography or referral to cardiology outpatient departments within 3 months of the index date were presumed to have followed an appropriate diagnostic pathway and were excluded from the loop diuretic group (exclusion criterion 6). Exclusion criterion 6 was also applied to the at-risk group on the assumption that some patients who underwent investigations may have had HF.

    Patients with Read codes for either aortic or mitral valve disease 3 months before or after the index date were presumed to have HF due to valve disease and were excluded from the loop diuretic group (exclusion criterion 7).

    Finally, patients who were prescribed a loop diuretic before or on the index date were excluded from the at-risk group (exclusion criterion 9).

    We extracted from the primary care electronic record the following: body mass index, smoking status, common comorbidities, HF medications, presentation with HF symptoms before the index date (peripheral oedema, breathlessness, orthopnoea, paroxysmal nocturnal dyspnoea or fatigue), natriuretic peptide testing, echocardiography referrals or results, and outpatient referrals. Cause-specific hospitalisation data were extracted using linkage with HES APC records. Cause-specific mortality was extracted from ONS death records.

    Statistical analysis

    We performed two separate comparisons: analysis A—loop diuretic group versus the HF group, and analysis B—loop diuretic group versus the at-risk group. In analysis A, patients in the loop diuretic group were matched with patients in the HF group using a propensity score using age as a continuous variable, sex, and the presence of IHD, diabetes, hypertension and atrial fibrillation (AF). The propensity score was calculated using a cumulative logit regression model. Matching was on a 1:1 nearest neighbour basis, without replacement, with a calliper width of 0.2 of the SD of the logit of the propensity score.

    In analysis B, patients in the study population were matched to patients in the at-risk group for age and sex. Matching was on a 1:1 nearest neighbour basis without replacement with exact matches only. Standardised mean difference and distribution plots were used to check the adequacy of the matching.

    Two sensitivity analyses were performed: one for analysis A in which patients were matched for loop diuretic prescription as well as age, sex, and the presence of IHD, diabetes, hypertension and AF; and one for analysis B in which patients were matched for the presence of IHD, diabetes, hypertension and AF, as well as age and sex.

    Continuous data are presented as medians (first–third quartiles), while categorical data are presented as numbers (%). Differences in baseline characteristics between unmatched groups and matched groups were tested using independent t-tests for continuous variables and Χ2 tests for categorical variables.

    Differences in outcome were assessed using univariable and multivariable Cox regression models and Kaplan-Meier curves. The two-tailed level of statistical significance was set at <0.05. All statistical analyses were performed using SPSS V.28.

    Outcome definitions

    The primary outcome was time to first HF-related event which comprised of presentation to primary care with symptoms of HF, or hospitalisation with HF or all-cause mortality in analysis A; and presentation to primary care with symptoms of HF, or incident HF (new diagnosis of HF made in either primary or secondary care), or hospitalisation with HF, or all-cause mortality in analysis B. Secondary endpoints of time to first HF hospitalisation or all-cause mortality, and time to first all-cause hospitalisation or all-cause mortality were also assessed in both analyses. Patients were followed up until the first HF event occurred or until 5 years.

    Patient and public involvement

    The Involve Hull patient and public involvement group provided written feedback on the study protocol during conception and prior to submission to the CPRD and guided plans for dissemination.

    Funding

    This study was funded by the Hull and East Riding Cardiac Trust Fund which had no input in the study design, data analysis or drafting of this manuscript.

    Results

    Of the 180 384 patients with either a first prescription of loop diuretics, or first diagnosis of either HF, IHD, hypertension or diabetes between 1 January 2010 and 31 December 2015, 78 968 had a new loop diuretic prescription, 28 177 had a new diagnosis of HF, 32 701 had a new diagnosis of IHD, and 40 538 had a new diagnosis of diabetes in the context of pre-existing hypertension or vice versa. After application of exclusion criteria, 139 265 were used in the analyses comprising of 59 694 in the loop diuretic group, 22 352 in the HF group and 57 219 in the at-risk group (figure 1).

    Patient characteristics

    Compared with patients with HF, patients taking a loop diuretic without a diagnosis of HF were younger, more likely to be female (38% male vs 52% female), and were less likely to have AF (11% vs 24%), CKD (20% vs 30%) and IHD (15% vs 30%) (p<0.001 for all). Furosemide was the most commonly prescribed loop diuretic in both groups (table 1).

    View this table:
    Table 1

    Patient characteristics: analysis A

    Compared with patients with HF risk factors, patients taking a loop diuretic without a diagnosis of HF were older (74 vs 64 years), more likely to be women (38% male vs 59% female), and were more likely to have either AF (11% vs 4%) or CKD (20% vs 12%) (p<0.001 for all) (table 2).

    View this table:
    Table 2

    Patient characteristics: analysis B

    Symptom burden

    Only one in five patients in the loop diuretic group and the HF group had a presentation of HF symptoms to primary care in the month before their index date. Of those who had a recorded presentation prior to the index date, patients in the loop diuretic group were more likely to present with oedema (80% vs 26%) and less likely to present with breathlessness (16% vs 65%) than those in the HF group (p<0.001 for both) (table 1).

    Outcome

    Analysis A

    In the propensity-matched cohorts, during a median follow-up of 65 (21–92) months, an HF-related event occurred in 71.9% of patients in the loop diuretic group and 72.1% of patients in the HF group. The proportion of patients presenting with an HF symptom was greater in the loop diuretic group (37.1% vs 27.8%; p<0.001), but both hospitalisation with HF (1.9% vs 4.0%; p<0.001) and all-cause mortality (55.6% vs 61.2%; p<0.001) were more frequent in the HF group (table 3, online supplemental figure 1 and online supplemental table 1).

    View this table:
    Table 3

    Outcomes in analyses A and B

    Patients prescribed a loop diuretic without a recorded diagnosis of HF were only 6% less likely to experience an HF event compared with those with HF after adjustment for baseline characteristics (HR=0.94 (95% CI 0.92 to 0.96; p<0.001) (figure 2 and table 4).

    Figure 2
    Figure 2

    Risk of HF event in patients taking a loop diuretic without a diagnosis of HF (loop diuretic group) compared with those with an HF diagnosis (HF group). HF, heart failure.

    View this table:
    Table 4

    Univariable and multivariable Cox regression analyses for heart failure (HF) events in analyses A and B

    The sensitivity analysis for analysis A found that HF events were significantly more likely in patients with a recorded diagnosis of HF compared with those taking a loop diuretic without a diagnosis of HF (online supplemental tables 2 and 3). However, HF events were still very common in patients prescribed a loop diuretic without a recorded diagnosis of HF (69.3% vs 73.7%; p<0.001).

    All-cause hospitalisation or death occurred in 67.7% of patients in the loop diuretic group and 71.5% of patients in the HF group (p<0.001), although all-cause hospitalisation was more common in the loop diuretic group (27.5% vs 23.0%; p<0.001) (online supplemental figure 2, table 3 and online supplemental table 4).

    Analysis B

    In the propensity-matched cohorts, during a median follow-up of 89 (66–109) months, an HF-related event occurred in 59% of patients in the loop diuretic group and 36% of patients in the at-risk group (HR 2.04 (95% CI 2.00 to 2.08); p<0.001) (table 3).

    Patients in the loop diuretic group were approximately twice as likely to experience an HF event compared with those in the at-risk group (unadjusted HR=2.04 (95% CI 2.00 to 2.08); p<0.001) (figure 3, table 4, online supplemental figure 3 and online supplemental table 1).

    Figure 3
    Figure 3

    Risk of HF event in patients taking a loop diuretic without a diagnosis of HF (loop diuretic group) compared with patients with HF risk factors (at-risk group). HF, heart failure.

    The sensitivity analysis for analysis B found a higher prevalence of HF events in both groups but similarly greater risk in patients in the loop diuretic group compared with the at-risk group (77.0% vs 52.6%; p<0.001) (online supplemental tables 2 and 3).

    All-cause hospitalisation or mortality occurred in 56% of patients in the loop diuretic group and 42% of patients in the at-risk group (p<0.001) (online supplemental table 4 and online supplemental figure 4).

    Discussion

    Using a large, representative sample of patients in primary care, we found that patients who are prescribed a loop diuretic but who do not have a recorded diagnosis of HF have a high rate of HF-related events—similar to that of those with a confirmed diagnosis of HF, and nearly twice that of age-matched and sex-matched patients with risk factors for developing HF. During the 5-year index period of our study (2010–2015), there were over twice as many patients prescribed a loop diuretic without a recorded diagnosis of HF than were given a formal diagnosis of HF. Contrasting the results of analysis A and analysis B, patients prescribed a loop diuretic are far more similar to patients with a recorded diagnosis of HF in terms of symptom burden and outcome than they are to patients with risk factors for HF; undiagnosed or ‘uncoded’ HF may account for many of the loop diuretic prescriptions.

    Patients prescribed a loop diuretic without a diagnosis of HF

    The demographics of patients prescribed a loop diuretic without a recorded diagnosis of HF and those with a recorded HF in this study are similar to those of patients with HF and a normal ejection fraction (HeFNEF): the majority were female and aged over 70 years with multiple comorbidities (AF, CKD and hypertension were the most common, while IHD was uncommon).19 20

    While we cannot infer what proportion of patients in the loop diuretic group had underlying HF as a cause of their symptoms, as a group, they were more likely than patients with HF to present to their general practitioner (GP) with symptoms, and only marginally less likely to be admitted to hospital or die.

    Defining HF

    The benefits of establishing a diagnosis of HF for an individual are numerous: in the case of HF with a reduced ejection fraction, there are multiple medical and device treatments which can enormously reduce the chance of serious morbidity and prolong life.21 In the case of HeFNEF, which may account for approximately half of all HF diagnoses,22 treatment with sodium glucose co-transporter 2 inhibitors can reduce morbidity related to HF.23 24 Regardless of HF phenotype, establishing a diagnosis provides clarity to the patient and removes clinical uncertainty for the non-specialist which may prevent delays to treatment.25

    The benefits of establishing a diagnosis of HF for the wider community are also numerous: having a proper understanding of the epidemiology of HF is essential for planning healthcare services. Establishing the diagnosis in the community is associated with lower healthcare costs and better clinical outcomes.26 However, the proportion of patients who receive their HF diagnosis in the community is decreasing26–28; the most recent data suggest that up to 80% of patients receive their diagnosis only after hospitalisation with HF.28 This may be due to the requirement for preliminary investigations prior to specialist referral in order to make a diagnosis, leading to uncertainty before a diagnosis is confirmed or refuted. Potentially compounding the problem are the diverse and complex diagnostic criteria, particularly for HeFNEF.1

    The symptoms of congestion are often, quite reasonably, treated before a definitive diagnosis is made. However, very few of the patients prescribed a loop diuretic had had appropriate investigations during 10 years’ follow-up. In contrast, the majority of patients with a diagnosis of HF have some form of investigation or referral before a diagnosis is made, regardless of the setting in which it is made.28

    Can we rely on coding?

    We found that only one in five patients who were prescribed a loop diuretic or were given a diagnosis of HF were coded as having an attendance with symptoms of HF in the month prior to the index date (including on the index date itself), which seems surprising and may represent absent coding. GPs in the UK are financially incentivised via Quality Outcomes Framework to keep and maintain a register of patients within their practice population with an HF diagnosis.12 While absent coding may account for some of the ‘missing’ patients in our analysis, this may only affect a minority of patients.14

    Estimates of the prevalence of HF in community settings vary greatly depending on the methods of diagnosis and the populations studied.6–11 Using clinical coding in the general population, the prevalence of HF in the UK is estimated to be 1.4%.6 If even a small proportion of patients prescribed a loop diuretic had underlying HF, regardless of whether it was clinically recognised by the clinician, using clinical coding to estimate prevalence would be an underestimation. Consequently, planning for and funding of HF services are unlikely to be adequate.

    Clinical implications

    The widespread use of loop diuretics without further investigations is an impediment to a timely diagnosis of HF. There are many cardiovascular medications which require blood tests (renal function or serum electrolyte concentrations, for example) to be checked prior to initiation. We believe that mandating measurement of natriuretic peptide concentrations prior to initiation of loop diuretics is necessary, clinically appropriate, straightforward to implement and may improve care.

    Patients with HF are being missed at present and, as a consequence, potentially not receiving life-prolonging and symptom-relieving medication. The wider healthcare economy is not receiving the benefits of early appropriate treatment that reduces the risk of hospitalisation and which would improve our understanding of HF epidemiology, allowing better planning and funding of HF services. A nationwide effort to review all patients currently taking loop diuretics without a diagnosis of HF may find many patients with a treatable condition who stand to gain much from their diagnosis.

    Limitations

    We had incomplete clinical information in the available electronic health records. Absent coding may account for the majority of missed diagnoses in patients prescribed a loop diuretic without a diagnosis of HF.14 If this finding is generalisable to the data from the CPRD, our findings are all the more important. If the absence of a clinical code does not mean the absence of the disease, then epidemiological reports using electronic data are destined to underestimate prevalence. However, the positive predictive value of a diagnosis recorded in CPRD being clinically present is approximately 89%.29

    It is likely that some patients in the loop diuretic group would have been prescribed a loop diuretic for the treatment of other causes of peripheral oedema such as hypoalbuminaemia, lymphoedema or venous stasis—the pattern of symptoms before the index date were notably different for those in the loop diuretic group compared with those in the HF group.

    We acknowledge a degree of immortal time bias affecting patients in the HF group. Although 50% of patients were recorded as being on a loop diuretic at the time of the HF diagnosis, approximately 25% were taking a loop diuretic before the diagnosis was made (25% started a loop diuretic at the same time as the HF diagnosis).

    Although we used propensity matching and multivariable Cox regression analyses, we cannot account for unmeasured clinical variables that may confound the results.

    Conclusion

    Patients who are prescribed a loop diuretic without a recorded diagnosis of HF are at high risk of HF-related events over long-term follow-up. Many patients in the community may have unrecognised HF. This has profound implications in our understanding of HF epidemiology.

    Supplemental material

    Data availability statement

    Data may be obtained from a third party and are not publicly available. Data provided by CPRD.

    Ethics statements

    Patient consent for publication

    Ethics approval

    Scientific approval for the present study was given by the CPRD Independent Scientific Advisory Committee. Ethics approval was given by the Hull York Medical School Ethics Committee (ref 21.28-RECORD-HF).

    References

    1. ↵
      2. McDonagh TA ,
      3. Metra M ,
      4. Adamo M , et al
      . 2021 ESC guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J 2021;42:3599–726. doi:10.1093/eurheartj/ehab368
      OpenUrlCrossRefPubMed
    2. ↵
      2. Breidthardt T ,
      3. Moreno-Weidmann Z ,
      4. Uthoff H , et al
      . How accurate is clinical assessment of neck veins in the estimation of central venous pressure in acute heart failure? Insights from a prospective study. Eur J Heart Fail 2018;20:1160–2. doi:10.1002/ejhf.1111
      OpenUrl
    3. ↵
      2. Shelton RJ ,
      3. Clark AL ,
      4. Goode K , et al
      . The diagnostic utility of N-terminal pro-B-type natriuretic peptide for the detection of major structural heart disease in patients with atrial fibrillation. Eur Heart J 2006;27:2353–61. doi:10.1093/eurheartj/ehl233
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      2. Letnes JM ,
      3. Eriksen-Volnes T ,
      4. Nes B , et al
      . Variability of echocardiographic measures of left ventricular diastolic function. The HUNT study. Echocardiography 2021;38:901–8. doi:10.1111/echo.15073
      OpenUrl
    5. ↵
      2. Dorosz JL ,
      3. Lezotte DC ,
      4. Weitzenkamp DA , et al
      . Performance of 3-dimensional echocardiography in measuring left ventricular volumes and ejection fraction: a systematic review and meta-analysis. J Am Coll Cardiol 2012;59:1799–808. doi:10.1016/j.jacc.2012.01.037
      OpenUrlFREE Full Text
    6. ↵
      2. Conrad N ,
      3. Judge A ,
      4. Tran J , et al
      . Temporal trends and patterns in heart failure incidence: a population-based study of 4 million individuals. Lancet 2018;391:572–80. doi:10.1016/S0140-6736(17)32520-5
      OpenUrlCrossRefPubMed
    7. ↵
      2. Norhammar A ,
      3. Bodegard J ,
      4. Vanderheyden M , et al
      . Prevalence, outcomes and costs of a contemporary, multinational population with heart failure. Heart 2023;109:548–56. doi:10.1136/heartjnl-2022-321702
      OpenUrlAbstract/FREE Full Text
    8. ↵
      2. Anguita Sánchez M ,
      3. Crespo Leiro MG ,
      4. de Teresa Galván E , et al
      . Prevalence of heart failure in the Spanish general population aged over 45 years. Rev Esp Cardiol 2008;61:1041–9. doi:10.1016/s1885-5857(09)60007-4
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      2. Ceia F ,
      3. Fonseca C ,
      4. Mota T , et al
      . Prevalence of chronic heart failure in southwestern Europe: the EPICA study. Eur J Heart Fail 2002;4:531–9. doi:10.1016/s1388-9842(02)00034-x
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      2. Seferović PM ,
      3. Vardas P ,
      4. Jankowska EA , et al
      . The heart failure association Atlas: heart failure epidemiology and management statistics. Eur J Heart Fail 2021;23:906–14. doi:10.1002/ejhf.2143
      OpenUrlPubMed
    11. ↵
      2. Savarese G ,
      3. Becher PM ,
      4. Lund LH , et al
      . Global burden of heart failure: a comprehensive and updated review of epidemiology. Cardiovasc Res 2023;118:3272–87. doi:10.1093/cvr/cvac013
      OpenUrl
    12. ↵
      1. National Health Service Digital
      . Quality and outcomes framework (QOF), 2019-20. Available: https://digital.nhs.uk/data-and-information/publications/statistical/quality-and-outcomes-framework-achievement-prevalence-and-exceptions-data/2019-20 [Accessed 13 Feb 2023].
    13. ↵
      1. National Institute for Health and Care Excellence
      . British national Formulary [Diuretics]. Treatment summaries. Available: https://bnf.nice.org.uk/treatment-summaries/diuretics [Accessed 28 Feb 2023].
    14. ↵
      2. Cuthbert JJ ,
      3. Gopal J ,
      4. Crundall-Goode A , et al
      . Are there patients missing from community heart failure registers? An audit of clinical practice. Eur J Prev Cardiol 2019;26:291–8. doi:10.1177/2047487318810839
      OpenUrlCrossRefPubMed
    15. ↵
      2. Herrett E ,
      3. Gallagher AM ,
      4. Bhaskaran K , et al
      . Data resource profile: clinical practice research Datalink (CPRD). Int J Epidemiol 2015;44:827–36. doi:10.1093/ije/dyv098
      OpenUrlCrossRefPubMed
    16. ↵
      1. National Institute for Health and Care Excellence (NICE)
      . Chronic heart failure in adults: diagnosis and management [NICE guideline NG106]. Overview | Chronic heart failure in adults: diagnosis and management | Guidance | NICE; 2018.
    17. ↵
      2. Springate DA ,
      3. Kontopantelis E ,
      4. Ashcroft DM , et al
      . Clinicalcodes: an online clinical codes repository to improve the validity and reproducibility of research using electronic medical records. PLoS One 2014;9:e99825. doi:10.1371/journal.pone.0099825
    18. ↵
      1. National Health Service Digital
      . Read codes, 2022. Available: https://digital.nhs.uk/services/terminology-and-classifications/read-codes [Accessed 3 Mar 2023].
    19. ↵
      2. Teramoto K ,
      3. Teng T-H ,
      4. Chandramouli C , et al
      . Epidemiology and clinical features of heart failure with preserved ejection fraction. Card Fail Rev 2022;8:e27. doi:10.15420/cfr.2022.06
    20. ↵
      2. Chioncel O ,
      3. Lainscak M ,
      4. Seferovic PM , et al
      . Epidemiology and one-year outcomes in patients with chronic heart failure and preserved, mid-range and reduced ejection fraction: an analysis of the ESC heart failure long-term Registry. Eur J Heart Fail 2017;19:1574–85. doi:10.1002/ejhf.813
      OpenUrlCrossRefPubMed
    21. ↵
      2. Vaduganathan M ,
      3. Claggett BL ,
      4. Jhund PS , et al
      . Estimating lifetime benefits of comprehensive disease-modifying pharmacological therapies in patients with heart failure with reduced ejection fraction: a comparative analysis of three randomised controlled trials. Lancet 2020;396:121–8. doi:10.1016/S0140-6736(20)30748-0
      OpenUrlCrossRefPubMed
    22. ↵
      2. Dunlay SM ,
      3. Roger VL ,
      4. Redfield MM
      . Epidemiology of heart failure with preserved ejection fraction. Nat Rev Cardiol 2017;14:591–602. doi:10.1038/nrcardio.2017.65
      OpenUrlCrossRefPubMed
    23. ↵
      2. Anker SD ,
      3. Butler J ,
      4. Filippatos G , et al
      . Empagliflozin in heart failure with a preserved ejection fraction. N Engl J Med 2021;385:1451–61. doi:10.1056/NEJMoa2107038
      OpenUrlCrossRefPubMed
    24. ↵
      2. Solomon SD ,
      3. McMurray JJV ,
      4. Claggett B , et al
      . Dapagliflozin in heart failure with mildly reduced or preserved ejection fraction. N Engl J Med 2022;387:1089–98. doi:10.1056/NEJMoa2206286
      OpenUrlCrossRefPubMed
    25. ↵
      2. Kwok CS ,
      3. Burke H ,
      4. McDermott S , et al
      . Missed opportunities in the diagnosis of heart failure: evaluation of pathways to determine sources of delay to specialist evaluation. Curr Heart Fail Rep 2022;19:247–53. doi:10.1007/s11897-022-00551-4
      OpenUrlCrossRef
    26. ↵
      2. Bachtiger P ,
      3. Kelshiker MA ,
      4. Petri CF , et al
      . Survival and health economic outcomes in heart failure diagnosed at hospital admission versus community settings: a propensity-matched analysis. BMJ Health Care Inform 2023;30:e100718. doi:10.1136/bmjhci-2022-100718
    27. ↵
      2. Koudstaal S ,
      3. Pujades-Rodriguez M ,
      4. Denaxas S , et al
      . Prognostic burden of heart failure recorded in primary care, acute hospital admissions, or both: a population-based linked electronic health record cohort study in 2.1 million people. Eur J Heart Fail 2017;19:1119–27. doi:10.1002/ejhf.709
      OpenUrlCrossRefPubMed
    28. ↵
      2. Bottle A ,
      3. Kim D ,
      4. Aylin P , et al
      . Routes to diagnosis of heart failure: observational study using linked data in England. Heart 2018;104:600–5. doi:10.1136/heartjnl-2017-312183
      OpenUrlAbstract/FREE Full Text
    29. ↵
      2. Quint JK ,
      3. Müllerova H ,
      4. DiSantostefano RL , et al
      . Validation of chronic obstructive pulmonary disease recording in the clinical practice research datalink (CPRD-GOLD). BMJ Open 2014;4:e005540. doi:10.1136/bmjopen-2014-005540

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • X @JJCuthbert

    • Presented at This work was presented as a poster at the British Cardiovascular Society Annual Conference 5–7 June 2023.

    • Contributors JJC, IS, SJL and ALC conceived the project. JJC and IS performed data cleaning and analysis. JC and JT were responsible for data handling and governance. JJC, SJL, AF, IS and ALC drafted the manuscript. JJC is responsible for the overall content as guarantor.

    • Funding This study was funded by the Hull and East Riding Cardiac Trust Fund.

    • Disclaimer The funder had no input into the study design, data analysis or drafting of this manuscript.

    • Competing interests None declared.

    • Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.