Updating Cardiovascular Risk Functions for the Italian Population: The Cuore Project

The project represents a crucial milestone in the updating of the celebrated Progetto Cuore — a initiative that for decades has provided the Italian medical community with fundamental tools for assessing individual cardiovascular risk and personalizing care and prevention interventions for each patient.

The Epidemiological Context: Positive Trends but Persistent Challenges

Dr. Palmieri opened by illustrating the epidemiological picture of cardiovascular diseases in Italy through macro-level indicators showing encouraging trends. Mortality data for coronary and cerebrovascular diseases from 1980 to 2022 — the most recent year for which official mortality figures are available — reveal a steadily declining trend. Fortunately, thanks to the continuous efforts of researchers, physicians, and the national health system, this downward trajectory has persisted over time.

A particularly noteworthy finding concerns the effect of the Covid-19 pandemic: in 2020, a small rebound in cardiovascular mortality was recorded, likely as a secondary effect of the pandemic, followed by a return to the declining trend for both conditions in subsequent years. Overall, since the 1980s, mortality rates have fallen by approximately 70% for both coronary and cerebrovascular diseases.

Despite this significant progress, circulatory system diseases remain the leading cause of death in Italy, accounting for approximately 30–31% of all deaths — exceeding the 24% attributable to cancer. A relevant aspect from a gender perspective is that men exhibit higher mortality rates than women for both ischemic heart disease and cerebrovascular disease. For ischemic heart disease in particular, the male mortality rate is double that of women, while for cerebrovascular diseases the excess mortality in men stands at around 20%.

Trends in Hospital Admissions

Analysis of hospital discharge records (SDO) over the past six years also confirms a downward trend in hospitalizations. Particularly striking is the dramatic drop in 2020 due to Covid-19, when the healthcare system was essentially brought to a standstill and the population was effectively “frozen” with respect to these conditions, as hospital resources were redirected to Covid admissions. This situation unfortunately also had negative consequences in terms of mortality, as reflected in the data presented earlier.

Following the collapse of 2020, hospitalization rates are recovering slowly, but by 2023 figures were already below the last pre-Covid data point, confirming the general declining trend. As expected, admission rates are higher in men than in women for both cardiovascular conditions.

Absolute Cardiovascular Risk: Methodological Foundations

Dr. Palmieri then introduced the central concept of his presentation: absolute cardiovascular risk — a system for predicting cardiovascular events over the following 10 years that proves extremely useful in terms of prevention. Accurately estimating this risk requires following a precise methodological pathway.

The process first requires enrolling population cohorts at baseline, measuring a range of risk factors, risk conditions, and lifestyle variables, and — crucially — following these populations over time. The core of these risk assessment tools lies in risk functions: mathematical models that relate baseline conditions to the occurrence of cardiovascular events during follow-up.

The Three Fundamental Elements of Risk Functions

Risk functions depend on three main elements. The first is the average distribution of risk factors within the population, which varies from one population to another. The second is represented by risk coefficients — essentially the etiological weight of each individual risk factor — which are in fact similar across different populations. The third is the probability of surviving a cardiovascular event in the population, which also depends on the specific population under consideration.

This theoretical framework leads to a fundamental methodological conclusion: it is crucial to have risk functions estimated from data drawn from the very population to which the function will subsequently be applied. The more closely this general principle is respected, the more reliable and consistent the results will be. It is precisely this principle that guided the development of the Italian Progetto Cuore, avoiding the simple adoption of risk functions derived from foreign populations.

The Progetto Cuore: Italy’s First Risk Function

The Progetto Cuore marked a milestone in Italian preventive medicine. Through collaboration with several research groups across four areas of the country — north, center, and south — working in a standardized manner using the same methodologies derived from an international project (the MONICA project) and validating events in the same way, it became possible to produce the first risk function based on Italian data.

The original database drew on approximately 20,000 men and women aged 35 to 69, who over a median follow-up period of more than 10 years generated a total of approximately 920 cardiovascular events. The final model produced at the time revealed interesting differences between men and women. By using the hazard ratio relative to one standard deviation for each risk factor, it is possible to compare the contribution of factors with different units of measurement and assess the weight of each in risk prediction.

In men, the predominant factor with the greatest weight was age, while in women the factor with the highest hazard ratio was cigarette smoking. This gender difference is particularly significant and underscores the importance of personalized approaches in cardiovascular risk assessment.

Risk Charts and the Individual Score

From these analyses, tools applicable in primary prevention were developed — that is, for populations “healthy” with respect to cardiovascular disease, meaning individuals who had not experienced a major cardiovascular event such as myocardial infarction, stroke, or revascularization. The so-called Risk Charts were developed, based on six risk factors, along with the Individual Score, based on eight risk factors: sex, age, blood pressure, antihypertensive treatment, smoking habit, diabetes, total cholesterol, and HDL cholesterol.

On the basis of these factors, it is possible to individually assess the probability of experiencing a fatal or non-fatal coronary or cerebrovascular event over the following 10 years. The advantages of these tools are clear: they are straightforward to use, respect the multifactorial etiology of cardiovascular diseases, offer multiple and therefore customizable options for both preventive and therapeutic interventions. They also allow this assessment to be compared over time and against other individuals or situations, and enable a form of cost-benefit evaluation of each intervention.

The Limitations of Current Tools and the Need for an Update

Despite their undoubted utility, these tools present certain limitations that make updating necessary. As Dr. Palmieri emphasized, risk functions must be periodically revised. As previously noted, the closer the data used to develop the functions are to the target population, the greater the reliability and consistency of the risk assessment when applied to that same population.

Furthermore, the original Progetto Cuore functions were developed some time ago and do not account for newer risk factors, or do not incorporate certain risk factors in ways that could now be refined and improved. For example, diabetes was treated as a dichotomous variable (presence or absence), whereas it would be more informative to evaluate the continuous risk factor associated with blood glucose levels. To overcome these limitations, it is necessary to enroll more recent and larger cohorts.

The New Cohorts: The Health Examination Survey

Through participation in the PNRR HEAL ITALIA program, the team at the Istituto Superiore di Sanità was able to evaluate more recent cohorts and collect both risk factors and events in greater quantity and with more extended follow-up. This made it possible to assess the feasibility of broadening, for example, the age range to which the risk function could be applied, and to include additional risk factors using more sophisticated approaches.

The Progetto Cuore periodically implements the Health Examination Survey, which collects information not only on cardiovascular diseases but on the general health of the population. Three editions have been conducted: in 1998–2002, 2008–2012, and 2018–2019, while a new wave is currently underway, launched in 2023. These surveys collect population samples from all Italian regions, with approximately 10,000 participants per wave in the first three editions, while the current edition — following WHO directives — plans to enroll approximately 4,000 individuals across the national territory.

A fundamental aspect emphasized by Dr. Palmieri concerns the fact that these factors are “measured” rather than simply self-reported. While certain risk factors, risk conditions, and lifestyle variables must necessarily be collected through questionnaire, for the majority of parameters where it is feasible, direct measurements are performed using standardized instruments and methodologies. Both cardiovascular-specific risk factors and those associated with non-communicable diseases are assessed.

The Population Biological Bank

In addition to standard measurements, the project also collects 24-hour urine samples and maintains a biological bank that has enabled the collection and preservation of serum, plasma, buffy coat, and packed red blood cell samples — along with 24-hour urine specimens — from over 25,000 subjects. These samples are stored in the Biological Bank of the Istituto Superiore di Sanità, one of the few population-based biological banks in existence in Italy. This resource represents an invaluable scientific asset for future research.

The PNRR Work: Updating the Follow-up

The work conducted under the PNRR focused on updating the follow-up of previously enrolled populations. This was made possible by the fact that the Progetto Cuore is included in the National Statistics Plan, and therefore has access — albeit with considerable difficulty — to nominal (non-anonymous) mortality data. As Dr. Palmieri underscored, this represents the “conditio sine qua non” for implementing the follow-up of a cohort enrolled in scientific research: it is indispensable to know exactly what happened to each specific individual over time.

The undertaking was monumental: given that the cohorts were enrolled from the mid-1980s onwards, and that an average of approximately 600 deaths are recorded per year, this work had to be conducted year by year for approximately 32,000 individuals enrolled from the mid-1980s through the 2018–19 cohort — all forming part of the Progetto Cuore cohorts.

The Objectives and Challenges of Mortality Follow-up

The objective is to update the follow-up in order to determine what happened in terms of all-cause mortality, while also identifying the specific and primary cause of death. Achieving this objective requires not only having the data, but also undertaking an enormous effort of checking, cleaning, and verifying the consistency and completeness of data originating primarily from ISTAT.

As Dr. Palmieri highlighted, databases of this scale and complexity inevitably conceal countless pitfalls and present multiple problems. Extensive work is required in order to obtain data that are as reliable as possible. Thanks to this painstaking effort, the team succeeded in updating the follow-up for the period between 1999 and 2021 — with deaths prior to 1999 still to be recovered.

Results of the Mortality Follow-up Update

The data emerging from this initial analysis are significant. With over 32,000 enrolled subjects, all-cause mortality was first assessed. Between 1999 and 2021, approximately 7,000 deaths were recorded in total — 4,000 among men and 3,000 among women. Assessments were then conducted at different time points: at 5 years, 7.5 years, and 10 years, with incidence rates calculated on the basis of person-years.

The same approach was followed to specifically identify deaths from coronary disease and cerebrovascular disease — the outcomes of greatest relevance to the project. However, as Dr. Palmieri clarified, it is nonetheless necessary to have data on all deaths regardless of cause, for the purpose of a rigorous epidemiological analysis.

An Even Greater Challenge: Follow-up of Non-Fatal Events

The other major undertaking — still in progress but well advanced — concerns the updating of non-fatal events. This activity, based on hospital discharge records (SDO), has proven even more complex than the mortality follow-up. Dr. Palmieri recounted how, for the previous version of the Progetto Cuore risk charts, the team had managed, through painstaking work, to directly contact hospitals and municipal offices to collect the necessary data.

This enormous effort has grown progressively more complicated due to well-known issues surrounding privacy and data protection. Hospitals have become increasingly “closed” — justifiably so, in Dr. Palmieri’s view — out of concern over regulatory and liability implications, even when the requesting body was the Istituto Superiore di Sanità, the technical-scientific arm of the Ministry of Health. This meticulous approach was no longer viable, and the project was left essentially in limbo for several years.

The Miraculous Breakthrough: Record Linkage with the Ministry

As Dr. Palmieri himself described it — “miraculously” — and he stressed that this occurred only within the context of this project, thanks to a “stellar confluence of circumstances” — the Ministry agreed to provide an updated linkage between the Progetto Cuore cohort registry and hospital discharge records spanning several years, from 2001 to 2022.

The technical process is complex: the team provided the identifiers of enrolled individuals, and the Ministry returned a person identifier associated with SDO record identification codes. This identifier must then be matched to the anonymized SDO file, to which the Progetto Cuore has access by virtue of its inclusion in the National Statistics Plan.

The problem, as Dr. Palmieri frankly acknowledged, is that this linkage performed by the Ministry was likely outsourced to an external service provider rather than carried out internally — and the result must be taken “as is.” As he put it succinctly: “you’ve been granted a miracle, that’s what it is, and that’s what you’ve got.” Inevitably, as soon as the team began analyzing the data, a series of issues emerged. The team is working “like Sherlock Holmes” on these data, identifying systematic anomalies and attempting to trace them back to correct values — with no further support beyond the starting dataset.

Uncertainties Regarding Data Completeness

A critical issue concerns the impossibility of assessing how complete or incomplete the linkage may be, since no reference benchmark exists. The team must necessarily “trust” the work carried out by the Ministry. It is also important to consider the complexity of the underlying dataset: the starting point is hospital discharge files amounting to approximately 10 million records per year. Given that each individual may have multiple admissions for any cause and any diagnosis, the complexity of the linkage work is enormous. While for mortality the event is unique — a person can only die once, even if duplicates were found even in that dataset — for hospitalizations the situation is far more intricate.

Database Preparation Work

Dr. Palmieri briefly outlined the database preparation work, which involved addressing a series of technical and methodological challenges — including the year-by-year reconciliation of records that overlap across time. With the exception of 2006, which presented a particular situation, the team succeeded in linking approximately 120,000 hospital admissions from the base of 30,000 subjects enrolled across the various cohorts.

Future Perspectives

Looking ahead, as Dr. Palmieri summarized in his conclusions, several crucial activities remain to be completed. The obtained database must be carefully cleaned and systematically verified, and — most importantly — linked to the database of risk factors and risk conditions measured at baseline.

Subsequently, it will be necessary to correctly identify the events of interest, selecting only those of cardiovascular relevance to the study. Finally, all the statistical analyses will be performed to associate baseline conditions with the 10-year cardiovascular event risk — thereby producing the new, updated risk functions for the Italian population.

Conclusions

The work presented by Dr. Palmieri represents a scientific and organizational undertaking of remarkable proportions, testifying to the commitment of the Istituto Superiore di Sanità to providing the Italian medical community with precision tools for cardiovascular risk assessment. The updating of the Progetto Cuore — made possible through PNRR HEAL ITALIA funding — will yield more current risk functions, based on more recent and larger cohorts, and potentially applicable to broader age ranges.

The methodological challenge faced by the team — particularly with regard to record linkage with hospital discharge records — highlights the concrete difficulties researchers encounter in the era of privacy and data protection. At the same time, the success in gaining access, despite the described limitations, to these data represents a near-unique opportunity for Italian epidemiological research.

The approach described by Dr. Palmieri perfectly embodies the precision medicine principles promoted by the HEAL ITALIA project: developing tools based on Italian population data to personalize preventive and therapeutic interventions for the individual patient. Updated Risk Charts and Individual Scores will enable physicians to assess their patients’ risk profiles with greater accuracy and to tailor therapeutic and preventive strategies more effectively.

As Prof. Moroncini emphasized in his concluding remarks, this type of research is fundamental — especially for clinicians — offering practical, validated tools for everyday practice. The availability of risk functions specific to the Italian population, accounting for the epidemiological characteristics, risk factors, and survival probabilities particular to our country, represents an invaluable added value compared to the use of algorithms developed on foreign populations.

The work of the Progetto Cuore — begun decades ago and continuously updated through initiatives such as PNRR HEAL ITALIA — represents a virtuous example of longitudinal epidemiological research that, despite multiple operational and regulatory challenges, continues to make fundamental contributions to cardiovascular prevention in Italy. Integration with the new Health Examination Survey cohorts and the availability of the Population Biological Bank also open up prospects for future research capable of incorporating biomarkers and innovative risk factors into risk assessment functions — bringing us ever closer to the ideal of a truly personalized and precision-oriented cardiovascular medicine.