Survival of the Heart

---

Hospital data and screening studies however fail to explore the survival of patients who are diagnosed in a community setting and not necessarily admitted to hospital nor actively involved with screening for research purposes. Electronic primary care records provide a valuable source of data directly relevant to community populations The computerisation of primary care in the UK and increasingly robust coding of medical information has led to the creation of large datasets which, following anonymization, can be used to explore epidemiological trends, including survival rates.

The aim of this study was to determine the 1, 5 and year survival rates of patients with heart failure in a primary care population and examine whether prognosis has improved over time. Survival analysis was carried out using an open matched retrospective cohort from THIN database for the period from 1 January to 31 December At each consultation, the GP records details of the medical encounter, including diagnosis.

Demographic details such as age, sex and linked deprivation scores also form part of the electronic record. Practices that contributed at least 1 year of clinical data were included in the study.

This is the date after which recorded mortality in the practice is consistent with predicted mortality. This was to ensure that the main outcome of interest was being accurately recorded by the participating practices. The dataset extracted from THIN database included all persons aged 45 years and over, registered at a practice for at least 12 months during the study period.

This age cut-off was chosen because the types of heart failure affecting children and younger people are pathologically distinct from those found in older adults. Eligible cases had a clinical code of heart failure. The index date was the first date of a recorded heart failure code. Patients with a previous diagnosis of heart failure, either prior to age 45 or the study start date, were excluded. Cases were matched with up to five comparator patients who were registered in the same practice on the index date and did not have a diagnosis of heart failure on that date but could become a case later.

Patients with a diagnosis of heart failure were identified using clinical codes input by GPs to record new diagnoses in the medical record. Heart failure is a clinical syndrome and the diagnosis requires the presence of symptoms and objective evidence of a structural or functional abnormality of the heart.

Patients with a clinical code of heart failure and either an echocardiograph report or a hospital letter were classified as being a confirmed case of heart failure and those with just a clinical code alone as an unconfirmed case. Demographic variables including age, sex, ethnicity, area deprivation quintile, cardiovascular risk factors and co-morbidities were extracted. The latest deprivation quintile prior to the index date was used or, if unavailable, the most recently recorded after the index date. Cardiovascular risk factors smoking, blood pressure, cholesterol, body mass index BMI were the most recently recorded prior to the index date.

Cardiovascular co-morbidities angina, myocardial infarction MI , ischaemic heart disease, diabetes, hypertension, stroke, atrial fibrillation AF , valve disease , were defined by the presence of a clinical code at any time prior to the index date. The outcome measure was overall survival rate for the cohort which was determined using the date of death for patients who had died from any cause all-cause mortality.

Data were extracted directly from THIN database using clinical codes. Analysis was carried out using Stata versions 10 and The absolute number of heart failure cases, and overall incidence rate by age and sex, was calculated. Heart failure patients and matched patients without heart failure, included in the survival analyses were followed until the earliest of the following dates: Kaplan-Meier curves were used to compare survival in the heart failure and no heart failure cohorts.

Log rank tests were used to compare survival between groups. One-, five- and ten-year survival rates for the heart failure cohort were calculated by year age band from the age of For comparison, survival rates of the no heart failure cohort were also determined. To examine trends over time, age-sex specific and overall survival rates were calculated by year of diagnosis. Case definition was also explored by calculating overall survival rates by year of diagnosis for the confirmed and unconfirmed heart failure groups separately.

Five hundred and sixty four practices contributed at least 1 year of data between 1 January and 31 December A total of patient records were included in the dataset; patients had a new clinical code of heart failure during the study period. Overall heart failure incidence was 3.

Introduction

Incidence of heart failure in age group 65 and older was 6. The characteristics of patients with and without heart failure are shown in Table 1. All five Townsend deprivation quintiles were similarly represented except for the most deprived group which had around one-third less cases than the other four groups. BMI was similar but there were 1. Cardiovascular co-morbidities, particularly ischaemic heart disease, AF and valvular disease were more common in patients with heart failure.

Characteristics of participants with heart failure and matched participants without heart failure from The Health Improvement Network THIN database between 1 January and 31 December Overall, heart failure cases had a significantly worse prognosis than comparators log rank test, chi-square 1 The survival rates for the heart failure cohort, and age, sex and practice matched cohort without heart failure, are shown in Table 2. Overall survival rates in the heart failure group were One-, five- and ten-year survival rates in participants with heart failure and matched participants without heart failure from The Health Improvement Network THIN database between 1 January and 31 December The survival rates of all heart failure cases by year of diagnosis are shown in Figure 1.

One-, five- and ten-year survival rates for men and women by age group are shown in Figure 2a—c , respectively. Survival rates for all groups did not change significantly over time. The number of confirmed cases made up just over a quarter The proportion of patients in each Townsend quintile was similar for confirmed and unconfirmed cases. Ischaemic heart disease, angina and MI were all more common in the confirmed group. The overall survival rates were lower in the unconfirmed heart failure group than in the confirmed heart failure group log rank test, chi-square 1 Figure 3 shows 1-, 5- and year survival rates in the confirmed and unconfirmed heart failure groups over time.

This study provides contemporary survival rate estimates for patients with a first diagnostic code of heart failure in their GP record. The outlook for patients with heart failure did not improve over time. Patients with a confirmed diagnosis had a slightly better outcome than those that were unconfirmed. The cohort in this analysis included patients with heart failure, much larger than most prospective studies, and the follow-up period was over a year period.

The NHS provides healthcare to the entire population of the UK, free at the point of access and most patients are registered with a primary care provider. Unlike screening cohort studies, primary care databases do not rely on participants volunteering to take part in the study rather they represent a cross-section of the entire population. These results are therefore likely to be generalizable to the community population as a whole The large number of patients in each age and sex category also improves the accuracy of the survival rate estimates.

The main limitation of the study is the reliability of GP coding. Research using general practice databases is reliant on the accuracy of clinical coding input by GPs during the consultation. Heart failure is a chronic condition which is often insidious in onset and can masquerade as other conditions making early and accurate diagnosis difficult. Over time the use and accuracy of clinical coding has improved, particularly since the introduction of the QOF in The benefit of THIN and similar primary care databases are that they provide an insight into real-life general practice and the survival rates of patients with a clinical code of heart failure in their record is likely to be an important statistic for practising GPs.

There is a paucity of contemporary epidemiological information on heart failure from a primary care setting. Annual QOF data shows a prevalence of 0.

Survival rate estimates have therefore come from screened cohorts and hospital databases. All deaths in the cohort were collated from routinely collected mortality data. The overall year survival rate of In patients with HF of both reduced and preserved EF, the influences of readily available predictors of mortality can be quantified in an integer score accessible by an easy-to-use website www. The score has the potential for widespread implementation in a clinical setting. See page for the editorial comment on this article doi: Heart failure HF is a major cause of death, but prognosis in individual patients is highly variable.

Quantifying a patient's survival prospects based on their overall risk profile will help identify those patients in need of more intensive monitoring and therapy, and also help target appropriate populations for trials of new therapies. There exist previous risk models for patients with HF. Each model's development is from a limited cohort size, compromising the ability to truly quantify the best risk prediction model. Also most models are restricted to patients with reduced left-ventricular ejection fraction EF , thus excluding many HF patients with preserved EF.

We use readily available risk factors based on 39 patients from 30 studies to provide a user-friendly score that readily quantifies individual patient mortality risk. Here one registry is excluded since it had only median 3-month follow-up. The Coordinating Centre at the University of Auckland assembled the database for 29 studies. In these studies, such rounded values were re-allocated within 2.

Poisson regression models were used to simultaneously relate baseline variables to the time to death from any cause, with study fitted as a random effect. Since mortality risk is higher early on, the underlying Poisson rate was set in three time bands: For binary and categorical variables, dummy variables were used.

Quantitative variables were fitted as continuous measurements, unless there was a clear evidence of non-linearity, e. Also two highly significant statistical interactions were included in the main model: Each variable's strength of contribution to predicting mortality was expressed as the z statistic. The larger the z the smaller the P -value, e.

Missing values are handled by multiple imputations using chained equations. First, for each variable with missing values, a regression equation is created. This model includes the outcome and follow-up time, in this case the Nelson—Aalen estimator as recommended by White and Royston 10 , an indicator variable for each study and other model covariates. For continuous variables, this is a multivariable linear regression, for binary variables, a logistic regression, and for ordered categorical variables, an ordinal logistic regression.

Once all such regression equations are defined, missing values are replaced by randomly chosen observed values of each variable in the first iteration. For subsequent iterations, missing values are replaced by a random draw from the distribution defined by the regression equations.

Survival analysis of heart failure patients: A case study

This was repeated for 10 iterations, the final value being the chosen imputed value. This is similar to Gibbs sampling. This entire process was repeated 25 times, thus creating 25 imputed data sets. The next step was to estimate the model for each of these data sets. Finally, the model coefficients are averaged according to Rubin's rule. We converted the Poisson model predictor to an integer score, which is then directly related to an individual's probability of dying within 3 years.

A zero score represents a patient at lowest possible risk. Having grouped each variable into convenient intervals, the score increases by an integer amount for each risk factor level above the lowest risk. Each integer is a rounding of the exact coefficient in the Poisson model, making log rate ratio 0. This report is based on 39 patients from 30 studies: Supplementary material online Table S1 describes each of the 30 studies.

Overall, 15 There were 31 baseline variables available for inclusion in prognostic models. A multiple imputation algorithm see Methods was used to overcome this problem. Consequently, all results are based on average estimates across 25 imputed data sets. For continuous variables, potential non-linearity in the prediction of survival was explored, as were potential statistical interactions between predictors.

The mortality association of increased age was more marked with higher EF, whereas the inverse association of systolic blood pressure with mortality became more marked with lower EF. The impact of age which varies with EF is particularly strong, and hence is shown on a different scale to the other plots. All charts are on the same scale except that for the interaction between ejection fraction and age, where the impact on mortality is more marked.

For each patient, the integer amounts contributed by the risk factor's values are added up to obtain a total integer score for that patient. For instance, scores of 10, 20, 30, and 40 have 3-year probabilities 0. Groups 1—4 comprise patients with scores 0—16, 17—20, 21—24, and 25—28, respectively, approximately the first four quintiles of risk.

To give more detail at higher risk, groups 5 and 6 comprise patients with scores 29—32 and 33 or more, approximately the top two deciles of risk. The marked continuous separation of the six Kaplan—Meier curves is striking: Cumulative mortality risk over 3 years for patients classified into six risk groups. Risk groups 1—4 represent the first four quintiles of risk integer scores 0—16, 17—20, 21—24, and 25—28, respectively. Risk groups 5 and 6 represent the top two deciles of risk integer scores 29—32 and 33 or more, respectively.

Associated Data

In the bottom two groups, the observed mortality is slightly lower than that predicted by the model, but overall the marked gradient in risk is well captured by the integer score. For most predictors, the strength of mortality association is similar in both subgroups. However, the impact of age is more marked and the impact of lower SBP is less marked in patients with preserved left-ventricular function, consistent with the interactions in the overall model. In this meta-analysis of 30 cohort studies, we explored between-study heterogeneity in mortality prediction.

From fitting separate models for each study, we observe a good consistency across studies re the relative importance of the predictors data not shown.

This reveals substantial between-study differences in mortality risk not explained by predictors in our model. However, a comparison of the seven randomized trials with the 23 patient registries reveals no significant difference in their mortality rates.

Characteristics of Patients With Survival Longer Than 20 Years Following Heart Transplantation

This study identifies 13 independent predictors of mortality in HF. Although all have been previously identified, the model and risk score reported here are the most comprehensive and generalizable available in the literature. They are based on 39 patients from 30 studies with a median follow-up of 2.

Also, we include patients with both reduced and preserved EF, the latter being absent from most previous models of HF prognosis. Given the wide variety of different studies included, with a global representation, the findings are inherently generalizable to a broad spectrum of current and future patients.

Conversion of the risk model into a user-friendly integer score accessible by the website www. All 13 predictors in the risk score should be routinely available, though provision will be made in the website for one or two variables to be unknown for an individual. We included serum creatinine rather than creatinine clearance or eGFR.

The latter involve formulae that include age, which would artificially diminish the huge influence of age on prognosis. While others report heart rate as a significant predictor of mortality, 21 we find that once the strong influence of beta blocker use is included, heart rate was not a strong independent predictor. Cardiovascular disease history e. What mattered most was the time since first diagnosis of HF, best captured by whether this exceeds 18 months.

Besides the powerful influence of diabetes, the other disease indicator of a poorer prognosis was prevalence of COPD. Previous myocardial infarction, atrial fibrillation, and LBBB were not sufficiently strong independent predictors of risk to be included in our model. Nearly all predictors display a similar influence on mortality in both subgroups. Two exceptions are age better prognosis of preserved EF compared with reduced EF HF is more pronounced at younger ages and systolic blood pressure, which have a stronger inverse association with mortality in patients with reduced EF.

Our meta-analysis of 30 cohort studies enables exploration of between-study differences in mortality risk. Separately, for each of the 10 largest studies, we calculated Poisson regression models for the same 13 predictors. Informal inspection of models across studies shows a consistent pattern to be expected, given there are no surprises among the selected predictors. An additional model, with study included as a fixed effect rather than a random effect , reveals some between-study variation in mortality risk not captured by the predictor variables.

This may be due to geographic variations or unidentified patient-selection criteria varying across registries and clinical trials, though overall patients in registries and trials appear at similar risk. Also, calendar year may be relevant since improved treatment of HF may enhance prognosis in more recent times. We will explore these issues in a subsequent publication. Specifically, the score facilitates the identification of low-risk patients, e. We recognize some limitations.

To overcome this problem, we used sophisticated computer-intensive multiple imputation methods. In addition, we have checked the robustness of our overall findings for each predictor by separate analyses within each cohort where full data for that predictor were available. Conventional good practice seeks to validate a new risk score on external data. That is important when a risk score arises from a single cohort in one particular setting, especially when that cohort has limited size.

Here, the circumstances are different. We have a global meta-analysis of 30 cohorts with the largest numbers of patients and deaths ever investigated in HF.

• Allure Book 2 State Of Romance (Contemporary Romance)
• 3D and 4D Ultrasound: A Text & Atlas
• Dark Awakening (Crimson City)
• False Flag Murdering Neocon Crazies (Evil Empire Series Book 1)
• FAITH: The Imperative Ingredient
• On the Beach at Ala Moana
• 35 morti (Italian Edition)