Introduction
Cardiovascular diseases (CVDs) are the leading cause of death in Europe. (1) Peripheral artery disease (PAD) is, along with ischemic heart disease (IHD) and stroke, one of the three most significant diagnostic subgroups of CVDs. Today, the term PAD is used in clinical practice to denote partial or complete obstruction of one or more peripheral arteries of the lower extremities. (2) Morbidity and mortality from this disease is primarily influenced by traditional risk factors for CVDs such as for example smoking, diabetes, and hyperlipidemia. (3) Globally, 230 million people suffer from PAD, with both sexes represented approximately equality, although the clinically asymptomatic form of the disease is more common in women. (4)
Peripheral artery disease is characterized by a continuum of clinical presentations that includes the initial asymptomatic forms of the disease and ultimately life-threatening critical ischemia. (3) In clinical practice, intermittent claudication is the most common form of the disease, and its prevalence increases with age. It is estimated that the prevalence is <1% in the population aged <50, whereas it is 6% in those aged >65. (5) The true prevalence, estimated based on the ankle brachial index (ABI), indicates that the asymptomatic clinical form of the disease in the general population is several times more common than intermittent claudication. (6)
As early as 2015, the Croatian Academy of Sciences and Arts and the Croatian Society for Atherosclerosis recognized the clinical significance of asymptomatic atherosclerosis at a scientific congress on the prevention of atherosclerosis. They emphasized the need for personalized PAD prevention and more frequent application of diagnostic methods such as the ABI, pulse wave velocity, and flow-mediated dilatation, which have still not entered routine clinical application for asymptomatic disease in the Republic of Croatia. (7, 8)
Peripheral artery disease is one of the more significant causes of amputation worldwide, but it also severely impacts quality of life in patients and increases risk of severe cardiovascular (CV) events and death. It should be emphasized that, starting with the 1980s, a reduction in mortality from leading CVDs – IHD and stroke – was achieved in Europe. However, despite the analysis of population dynamics for PAD indicating reduced incidence in EU15+ countries in the last 27 years, mortality has nevertheless increased. (9)
The CV burden in Croatia stands out in comparison with other Mediterranean EU member countries, such as for example Italy, Spain, and Portugal, due to high morbidity and mortality found in Croatia. Croatian CVD mortality patterns are more similar to transitional countries in central and eastern Europe. (10) However, despite the leading role held by CVDs in total morbidity and mortality, no systematic epidemiological studies have been performed on the dynamics of the main diagnostic subgroups in Croatia, which includes PAD. (11) It is to be assumed that in Croatia, as in other countries, PAD represents a suboptimally diagnosed and treated disease in comparison with IHD and stroke. Other than the existence of individual hospital registries, medical centers do not systematically track treatment outcomes in patients with advanced stages of the disease treated with endovascular and surgical procedures. (12, 13) Given that the only data available are aggregate vital statistics on mortality, and considering all the limitations of such data, it is only possible to describe their trends in order to identify regions in which mortality from this disease is increasing.
This study was based on a regression analysis of age-standardized mortality rates due to PAD according to NUTS-2 regions in Croatia from 2011 to 2020 in order to determine existing trends and timepoints in which trend changes could be noted.
Discussion
In this study, we used a selected set of diagnoses to calculate age-standardized mortality rates from PAD from 2011 to 2020 by NUTS-2 regions based on mortality statistics. Mortality increased by 3.09% in men and 2.94% in women, and there were large regional differences in trend, which were statistically significant in men in Adriatic Croatia and in both sexes in Northern Croatia.
When discussing CVDs, it should be emphasized that the Republic of Croatia is among the EU member states that are considered transitional countries with a high burden of CVD mortality and morbidity. (29) Differences by sex indicate higher mortality in men in all three CVD diagnostic subgroups. Our results also confirm regional differences, with, for example, the highest mortality from CVD and IHD for both sexes being found in Pannonian Croatia, whereas the highest mortality from PAD was found in the City of Zagreb. This is in agreement with previous studies, according to which the continental part of Croatia has a greater CVD burden in comparison with the Mediterranean region. (30) Since the City of Zagreb, along with Adriatic Croatia, has lower mortality rates for IHD and CVD, we would expect PAD mortality to be along the same gradient. (30, 31) However, the City of Zagreb stands out with the highest mortality from PAD for both sexes (Table 2).
Regional differences in PAD mortality trends, although pronounced, were not statistically significant based on the average annual percentage change of age-standardized rates. In men, the greatest increase was in the City of Zagreb, 3.71%, while the lowest and also the only negative value was found in Northern Croatia, namely -0.24%. In women, the increase was highest in the City of Zagreb and lowest in Northern Croatia: 3.57% and 1.67%, respectively. These differences in mortality, in line with previous studies, are largely explainable by the distribution of risk factors and healthcare. (32-34) However, we should not ignore the role played by correct identification of this disease both by medical professionals and the patients themselves, as well as the significant role of socioeconomic determinants in the distribution of health and disease. The results obtained in the present study are difficult to explain precisely, due to the lack of systematic and methodologically more challenging epidemiological studies that provide causal explanations of interregional differences in CVD morbidity and mortality. Namely, since the data available on incidence and prevalence are based on assessments, it is ultimately statistics on mortality, the provision of which is obliged by law, that represents the foundation of such public health research, despite its limitations. (35) Despite the lack of national studies, studies conducted globally are largely generalizable to our population as well. According to these studies, CVD mortality is primarily influenced by risk factors, with the rest of the mortality rate, less than 50%, being determined by the availability and quality of medical care. Availably/equality and quality of healthcare are, along with the funds allocated to healthcare, key domains for assessing the successfulness of a healthcare systems. It should be emphasized that the funds allocated to healthcare in the Republic of Croatia are the third lowest per capita in the EU, reaching 1 272 USD (adjusted for purchasing power parity) in 2016, which was 52% of the EU average and 85% of the EU13 average. (10, 36) In addition to low funding, data on healthcare quality and the effectiveness of the applied medical technologies are not available at the level of medical institutions, and thus their contribution to the measurement of treatment outcomes cannot be evaluated. (37) However, the Eurostat database provides methodologically standardized and aggregated indicators for clinical outcomes in myocardial infarction and stroke, which allows comparisons between countries. For example, standardized 30-day hospital mortality from myocardial infarction is several times higher than this mortality statistic in the neighboring country of Slovenia, with a tenfold difference in mortality ratio between medical institutions. (36)
Given the above, mortality due to PAD in the present study can be explained only by simultaneously considering the inadequacy and limitations of epidemiological and clinical data as well as indicators on treatment quality. Analysis of medical care should certainly differentiate between the healthcare structure, determined by expenses and the availability of specialist physicians, and the process of providing vascular care, where the annual volume of interventions should be emphasized. Despite the lack of data, the differences in treatment outcomes in clinical practice are clear, and it can be assumed that larger medical and academic centers provide higher quality treatment. (38) Additionally, when analyzing the availability of endovascular and surgical interventions, the presence of several high-volume centers in the City of Zagreb should certainly be emphasized, as they may have influenced the study results, primarily in the sense of better registration of PAD as the cause of death. Such centralization of human and technological potential is not a characteristic exclusive to the treatment of PAD, and such trends have been observed for other specialties as well. (38)
The Republic of Croatia is severely burdened by lifestyle-related risk factors, which are the cause of more than half of deaths. (10) Among risk factors that have the highest influence on mortality from PAD, elevated fasting blood glucose values, arterial hypertension (AH), smoking, and a diet rich in salt stand out as those with the greatest effect. (33) Their significance for mortality is indicated by the high Croatian mortality rates from causes that can be prevented and treated, which are above the EU average. (10, 39) Furthermore, smoking represents a large public health issue in Croatia. Despite a prevalence of 20% in adults, which is among the highest in Europe, no effective measures for reducing this unwanted trend have been implemented. (10) The leading position of Croatia among EU countries with an overall prevalence of AH of 37% should also certainly be emphasized, and the Croatian average systolic pressure of 137.5 mmHg (95% CI 131.2.143.8) in 2015 was the highest in the world. (40) However, despite its great significance for public health, no systematic screening programs for AH have been implemented, and the recent rate of undiagnosed AH is 7.1%. (41, 42) Given that AH, both diagnosed and undiagnosed, is one of the major risk factors for the development of PAD, it is to be assumed that its synergistic effect together other traditional risk factors, primarily smoking and diabetes, explains the large mortality ratio. (3, 43) Studies conducted in the 2000s showed a significantly higher prevalence of AH in continental Croatia in comparison with Adriatic Croatia, whereas the prevalence of smoking was higher in Adriatic Croatia. (44) There was a notable correspondence with lower mortality rates due to PAD in the present study with the lower prevalence of AH in Adriatic Croatia in earlier studies, but, on the other hand, the rate of undiagnosed AH is significantly associated with living in Adriatic Croatia. The loss of the Mediterranean diet pattern in Adriatic Croatia, which is sometimes also associated with higher salt intake, has been previously described in some studies. (45) Furthermore, persons with undiagnosed AH are more often smokers and consume alcohol more often in comparison with diagnosed hypertonic patients. (42) Although population attributable fractions for individual traditional risk factors and their individual influence on CV mortality have not been calculated for the Republic of Croatia, there was a notable correspondence between AH and higher mortality due to PAD in continental Croatia. Despite that, and regardless of age and sex, patients with diagnosed PAD are significantly more frequently long-term smokers and hypertonic in comparison with coronary patients without PAD. (46)
The higher mortality rates in the City of Zagreb can also be partially explained by better identification of PAD on part of healthcare professionals (Table 2). Correctly recognizing this disease affects morbidity and mortality indicators, and it is likely that our family physicians, like physicians from other countries, are probably poorly acquainted with this chronic disease. (47) Diabetes, which often causes polyvascular involvement in patients with PAD, had a significant effect on all-cause mortality in previous studies, and a higher rate of unwanted cardiac events was also observed in symptomatic hospital patients during long-term follow-up. (12, 48) However, diagnosed coronary patients had almost the same ratio of PAD prevalence in comparison with those in whom there was suspicion of PAD based on clinical symptoms, but no currently established PAD diagnosis (6.4% vs. 6.3%). (46) Data showing how familiar family medicine physicians are with coordinating the clinical process and the diagnostic and treatment procedures for patients suffering from PAD would be extremely useful, but are not available. Such data would contribute to improved screening for PAD in coronary patients. Furthermore, there are no data based on which a clear determination of the quality of primary healthcare could be made, but its fragmentation is clear in practice, as is the inadequate identification and treatment of PAD. (10)
Economic development and its inequality contribute to the distribution of CV mortality for all diagnostic subgroups (Table 2). Psychosocial risk factors are considered to be of approximately equal significance since their risk is familiar to traditional CV risk factors. (49, 50) The Republic of Croatia is a country with pronounced regional differences, and the 2.5 times higher income in the City of Zagreb in comparison with Pannonian Croatia must be emphasized. (51, 52) Social stratification can be clearly seen in the high and increasing economical inequalities of among inhabitants, as well as in the distribution of health, health-related behavior, and disease. (10) However, despite pronounced social stratification, inequalities regarding the availability of healthcare and differences in clinical outcomes are given little attention, and it is thus difficult to assess the causal relationship between the two due to the availability of mostly aggregated data. (53, 54) Furthermore, the level of medical care in different institutions also cannot be compared since there is no systematic data gathering on clinical outcomes based on indicators specific to the disease. (55)
Our results indicate an increase in mortality trends that is more pronounced in men, which is in agreement with globally reported trends in the last 20 years. (56) By time period, a drop in mortality was present is both sexes from 2011 to 2018, after which the period from 2018 to 2020 had a pronounced, but non-significant increase of 25.36% and 29.94%.
Trends by time period and various joinpoints differed by NUTS-2 region and by sex. Significant trend changes were observed in Adriatic Croatia in men (Figure 2, Table 4) and in Northern Croatia in both sexes (Figure 2, Table 3).
The mortality trend in women in Adriatic Croatia was stable at 2.43%. In contrast, a negative trend was observed in men from 2011 to 2014, with an average annual percentage change of -9,14%, but in the following period, from 2014 to 2020, there was a significant joinpoint and a positive trend of 7.67% (Figure 4, Table 4). This Mediterranean region traditionally has lower CVD mortality rates, which is ascribable to risk factors, but also to the protective effect of the Mediterranean lifestyle and diet. (31, 57) It is certainly important to note the superior economic status of counties in this region, which are among the most highly developed, for example Istria County, Zadar County, and Dubrovnik-Neretva County. (30) Income, together with individual education levels, are significant determinants of population health, and a positive correlation has been reported between self-assessed health and income. (36) However, this region has a high risk of undiagnosed AH, which is explained in the literature as due to reduced blood pressure measurements on part of family physicians. (34)
The trend for men in the City of Zagreb was stable at 3.71% during the whole study period, as opposed to women, in whom a joinpoint was observed between two periods, namely from 2011 to 2018 and from 2018 to 2020. In the period from 2011 to 2018, a negative trend of -3.12% was found in women, whereas the trend from 2018 to 2020 was extremely positive at 30.85% (Figure 1).
There were no significant trend changes in Pannonian Croatia. The trend in men was constant at 1.14%, while a joinpoint was found in women between different periods. The trend was negative (-3.78%) from 2011 to 2018, followed by a positive trend of 28.71% from 2018 to 2020 (Figure 3).
Among all the regions, the largest changes in mortality trends were found in Northern Croatia. In men, a significant negative trend of -7.58% was found for the period between 2011 to 2018, which was followed by a significant but positive trend of 30.38%. Similar dynamics were observed in women – after a negative trend of -6.91% (2011-2018) there was a significant positive trend of 38.45% from 2018 to 2020 (Figure 2, Table 3).
When interpreting the epidemiological indicators of PAD, primarily prevalence and mortality, it should be emphasized that such estimates, both in Croatia and globally, are substantially undervalued. According to experience from clinical practice, the application of ABI in the general population is insufficient and the ratio of asymptomatic cases with clinically significant ABI values is large. (58) Clinically significant ABI of ≤0.9 affects mortality and causes a 3- to 6-fold increased risk of death in comparison with healthy persons. (4) Furthermore, morbidity and mortality after PAD diagnosis have been shown to be higher than after myocardial infarction. (42) ABI screening for PAD is rarely conducted in the general population, and the role of PAD in increasing the risk of death and complications in patients with other CVDs, such as IHD and CVI, is often neglected. Despite the efforts of the Working Group on Angiology and Peripheral Vascular Diseases of the Croatian Cardiac Society (CCS), which are based on the initiative of the Working Group on Peripheral Circulation of the European Society of Cardiology (ESC), measuring ABI in the assessment of CV risk and the functional assessment of the peripheral arteries of the lower limbs is still a neglected test in the Republic of Croatia. (7) Among physicians actively participating in the diagnostic process and assessment of PAD, only 7% apply ABI as a screening method and the initial method of assessing disease severity. Furthermore, as many as 78% of cardiologists believe additional education in angiology is needed for the treatment and diagnosis of PAD. All the above indicates the necessity of adhering to ESC guidelines for the diagnosis and treatment of PAD that have been presented by the CCS. (59) The need to increase patient awareness regarding prevention and early diagnosis has also been recognized by the Section for Interventional Radiology of the Croatian Society of Radiology, which published a Handbook for patients suffering from peripheral artery disease that is primarily aimed at the general population. (60)
Finally, it is equally important to emphasized the clinical significance of early PAD diagnosis and the association between thrombosis of the arteries in the lower extremities and increased risk of cancer. The risk for the development of cancer from any site is 2.5% within 6 months of follow-up after the establishment of the diagnosis and remains elevated by up to 17.9% after 20 years, with an especially pronounced association with lung cancer. (61, 62) The latter fact is not negligible in countries with a high burden of oncological diseases.
The present study had both advantages and limitations. An advantage that should be emphasized is the application of the new National Classification of Statistical Regions from January 2019, which divides the Republic of Croatia into the four second-level statistical regions (NUTS-2) we have described above, whereas the country was previously divided into only two such regions – Pannonian and Adriatic Croatia. In the earlier classification, less-developed counties in continental Croatia were in the same category as the City of Zagreb, which has significantly higher GPD per capita. The new classifications places all NUTS-2 regions except the City of Zagreb in the category of less-developed regions, since their GDP is below 75% of the EU average. Funds allocated to healthcare per capita are the third lowest in the EU in absolute terms, reaching 1 272 USD (adjusted for purchasing power parity) in 2016, which was 52% of the EU average and 85% of the EU13 average. (43, 47) The City of Zagreb must be analyzed separately in epidemiological studies due the higher average income and availability and quality of healthcare, which affects morbidity and mortality indicators. In addition to the use of the current NUTS-2 classification, one advantage of the present study that should certainly be emphasized is the use of the publicly available joinpoint program, which excels at effective performance of linear and joinpoint trend analysis. (63) The limited capacity to discover smaller trend changes in the period between joinpoints is the weakness of this program. However, joinpoint analysis is a statistical method used for the identification and quantification of significant changes in data trends, which is extremely useful in epidemiological research and making decisions on public health. The ecological design of the study also represents a limitation, since it makes it very difficult to reach conclusions on causal relationships. However, there is no unified national registry or coordinated data gathering at the hospital level that could be used to monitor treatment outcomes in patients using the chosen indicators. Additionally, it was not possible to ascertain the accuracy of the coding of causes of death between the NUTS-2 regions, and it should also be emphasized that patients with PAD have numerous comorbidities that certainly make it more difficult to correctly code the cause of death. The comparability of the study was also limited by the large differences between different countries in the practice of the coding for patients with PAD in whom revascularization is indicated. (64) ICD-10 is not appropriate for differentiating the individual clinical stages of PAD, leading to the coding of this disease, as opposed to IHD and stroke, being very prone to misclassification bias. Some countries, such as for example Germany or the United States, use the expanded ICD-10 classification in clinical practice, which is much better suited to the clinical presentation of PAD. However, the diagnoses selected for this study are applied in two large clinical hospital centers that perform endovascular and surgical procedures for the treatment of patients with PAD and can be considered representative of clinical practice. The literature emphasizes the shortcoming of ICD-10 and the advantages of ICD-11, which provides more opportunities for registering the individual clinical stages of PAD since it includes diagnoses that use the Rutherford and Fontaine classification for PAD. (64) However, it should be emphasized that the Republic of Croatia does not have a consistent and unified system for gathering data at the level of secondary healthcare. Every medical center in Croatia has its own hospital information system, and the existence of different diagnoses and different coding practices cannot be excluded, nor can possible changes in computer systems (and the available ICD codes) in healthcare institutions during the study period be ruled out. However, vital statistics on mortality in Europe are among the best in the world, and such routinely gathered data often form the basis of epidemiological studies. The Republic of Croatia has a high completeness in mortality data, estimated to be above 65%. (65, 66) Previous scientific studies have reported good practices for coding causes of death for IHD and CVD, and we may therefore assume the same level of quality for PAD, despite the lack of specific studies. (67) In the Croatian healthcare system, there are no administrative limitations for the use of healthcare outside the place of residence, and many patients are treated in the larger medical and academic centers, which increases waiting lists. (68) This practice causes a burden on the healthcare system, but does not effect mortality statistics since the causes of death are registered by place of residence. The difference in mortality trends between men and women indicate that there was no change in the way causes of death related of PAD were coded, as that would have led to similarities in the trends for both sexes. Although previous studies indicate an advantage for men for certain treatment procedures, the design of this study and the obtained data limit reaching such a conclusion. (69) Despite the limitations listed above, this study quantified the regional differences in mortality and emphasized the need for further clinical and epidemiological research on PAD.
Peripheral artery disease has a high prevalence, but it also represents a complex chronic disease that is challenging to diagnose and treat, and difficult to code. It is characterized by a large number of asymptomatic cases that have a highly increased risk of unwanted CV events. Its mortality, although undoubtably underestimated, has been increasing in the Republic of Croatia from 2011 to 2020. The City of Zagreb stands out among the NUTS-2 regions as having increased mortality in both sexes, and Pannonian and Northern Croatia had significant trend changes. Epidemiological and clinical studies on PAD in Croatia are severely lacking, despite the great clinical and public health significance it has among CVDs. Although larger medical centers use technologically advanced endovascular and surgical methods for treating patients with PAD, systematic studies on the application of diagnostic and treatment guidelines and the measurement of clinical treatment outcomes for these patients have not been conducted. Regardless of the lack of individual patient data, this regional study on mortality trends, in addition to raising awareness, facilitates a better understanding of the epidemiological dynamics of this insufficiently diagnosed and treated, complex, and chronic disease. Systematic screening and gathering individual and clinically significant data in medical institutions that treat patients with PAD and their unification in regional and national registries are the only way to improve the clinical care for patients with PAD in the future. From an epidemiological perspective, early diagnosis and reducing risk factors as well as secondary prevention are of the highest importance in reducing mortality.