Descriptive analysis of mortality due to Parkinson’s disease and Alzheimer disease in Argentina

Abstract

Introduction

Parkinson’s disease (PD) and Alzheimer disease (AD), along with other nervous system disorders, are the leading causes of global disease burden. Although global trends have remained stable in recent decades, an increase in years of life lost due to these diseases has been recorded. In Argentina, though no prevalence records are available for PD and AD, reliable mortality data do exist. Similarities and differences in mortality caused by PD and AD are analysed in the present study.

Methods

The study was performed in the total population between 2000 and 2019. Age-standardised mortality rates (ASMR) were calculated. At the national level, annual trends were analysed and comparisons between provinces were made. Subsequently, spatial autocorrelation in the Centre Region (CR) was examined.

Results

Between 2000 and 2019, there were no statistically significant changes in ASMR for PD or AD, in either sex. In most cases, ASMR in the CR and its provinces were higher than the national average. At the department level, a trend towards the concentration of similar values was observed, with the exception of mortality due to PD among women.

Conclusions

The CR exhibits a higher ASMR than the national average, with rates of mortality due to AD in men in Santa Fe province being particularly noteworthy. It also shows a distinct spatial concentration pattern, with 2 distinct clusters. Further research is required to investigate the underlying sociodemographic and environmental factors contributing to this pattern.

Resumen

Introducción

A nivel mundial, las enfermedades de Parkinson (EP) y Alzheimer (EA), junto con otras enfermedades del sistema nervioso, son la principal causa de carga global de enfermedad. Si bien las tendencias mundiales se han mantenido estables en las últimas décadas, se han registrado aumentos en los años de vida perdidos a causa de dichas enfermedades. En Argentina, aunque los registros de prevalencia son inexistentes, se cuenta con datos confiables de mortalidad. Este trabajo analiza las similitudes y diferencias de la mortalidad por EA y EP en Argentina.

Métodos

El estudio se realizó en la población total entre 2000 y 2019. Se calcularon tasas ajustadas de mortalidad por edad (TMAE). A nivel país, se analizaron las tendencias anuales y se establecieron comparaciones entre las provincias. Posteriormente, se exploró la autocorrelación espacial de la mortalidad al interior de la Región Centro (RC).

Resultados

Entre 2000 y 2019, no se registraron cambios estadísticamente significativos en las TMAE por EP y EA en ambos sexos. En la RC y sus provincias, las TMAE son, en general, superiores al promedio nacional. A nivel departamental se observa una tendencia hacia la concentración en el espacio de valores similares de las TMAE, excepto en el caso de EP para las mujeres.

Conclusiones

La RC evidencia TMAE por encima del promedio nacional, especialmente la provincia de Santa Fe en EA en hombres, y un patrón espacialmente concentrado, con la presencia de dos conglomerados disimiles sobre cuyas condiciones sociodemográficas y ambientales se requiere profundizar.

Introduction

Diseases of the central nervous system, including neurodegenerative diseases, are among the leading causes of morbidity and mortality worldwide, and constitute a top public healthcare priority.¹ Neurodegenerative diseases affect the nervous system in a range of ways, and have physical and mental repercussions. The best developed and most explored hypothesis is that different situations can cause accumulation of different types of aberrant proteins (according to the resulting pathological picture) in the nervous system from very early ages, continuing throughout an individual’s life.²

The 2 most frequent neurodegenerative diseases are Alzheimer disease (AD) and Parkinson’s disease (PD). AD, first described in 1906,³ is the result of neuronal atrophy and death due to amyloid deposition; women are affected approximately twice as frequently as men. The age-standardised mortality rate (ASMR) for AD was estimated at 25.2 (95% confidence interval, 6.7-64.4) deaths per 100 000 population in 2021.¹ PD, first described in 1861,⁴ results from the loss of dopaminergic neurons, possibly due to protein modifications; PD mainly affects men and is less prevalent than AD, with an ASMR of 4.8 (4.3-5.2) deaths per 100 000 population in 2021.¹

According to recent estimates, the worldwide ASMRs of both diseases remained stable between 1990 and 2021.¹ However, other studies into mortality due to AD and PD in the United States, and due to AD in the European Union, show a significant increase in ASMR in recent decades.^5–7 Globally, a publication on PD based on World Health Organization (WHO) data reported a significant increase in ASMR between 1994 and 2019.⁸ In the Region of the Americas, estimated ASMR for these diseases remained practically unchanged between 2000 and 2019.⁹ In Argentina, a recent publication reported ASMRs of 1.4 and 4.9 deaths per 100 000 population for PD and AD, respectively, in 2019.¹⁰

The purpose of this study is to analyse the similarities and differences between mortality due to AD and PD in the different provinces of Argentina, with a particular focus on the Centre Region (CR), in the light of its specific sociodemographic and environmental characteristics.

Material and methods

We conducted a descriptive study of mortality due to PD and AD in the total population of Argentina between 2000 and 2019. Annual data on deaths were provided by the Argentinian Ministry of Health’s Department of Health Information and Statistics,¹¹ and population estimates were obtained from the provincial population projections developed by the National Institute of Statistics and Censuses.¹² Due to the lack of official estimates based on sex and 5-year age groups, data at the department level (sub-divisions of provinces) were estimated through the application of interpolation formulae to calculate the intercensus population. These calculations were performed using the AGEINT template on the Population Analysis System Software.¹³ These interpolation formulae were applied to the populations reported in the national population censuses from 1991, 2001, and 2010.

Mortality data included the following variables: basic cause of death, place of residence of the deceased (province and department), age, and sex. Basic causes of death were coded according to the 10th revision of the International Classification of Diseases (ICD-10).¹⁴ The study analyses 2 specific diseases, PD (code G20) and AD (code G30), and excludes other disorders with different aetiology, diagnosis, and progression, such as secondary parkinsonism (G21), vascular dementia (F01), unspecified dementia (F03), and other degenerative diseases of the nervous system (G31).

ASMRs for the total population were calculated with the direct method, with comparisons against the WHO’s proposed standard population.¹⁵ Standardisation was performed using the EPIDAT software,¹⁶ based on the number of deaths observed and the total estimated population for each region analysed. Rates are expressed per 100 000 population, and presented with the corresponding 95% confidence intervals.

ASMRs are estimated for each year and for the entirety of the period 2000-2019. While annual rates were used for the analysis of trends over time for the country as a whole, province-/department-level comparisons were based on the ASMRs calculated for the 20-year study period, given the low incidence of the basic causes of death analysis in some provinces and nearly all departments in the CR, which includes the populations of the provinces of Córdoba, Entre Ríos, and Santa Fe.

To identify specific periods in which statistically significant differences in ASMRs occurred, trends were modelled using the Joinpoint Regression program.¹⁷ This method enables the identification of changes in trend, and provides the average annual percentage change (AAPC) for the total period, and the statistical significance of these values. Regression models were built using the weighted Bayesian information criterion (BICw); confidence intervals for AAPC were estimated using the empirical quantiles method.

To analyse the spatial distribution of mortality due to AD and PD in the CR, and to identify clusters with higher or lower values, we conducted spatial autocorrelation with the Moran index (Moran’s I), using the GeoDa software¹⁸. This index evaluates the existence of clusters in the spatial distribution of a variable, and may be calculated for the total space (global) or for individual units (local). The global Moran’s I, whose values range from –1 (perfect distribution) to 1 (perfect concentration), summarises the degree of spatial autocorrelation and results in a coefficient (the slope of the regression line), representing the degree of association between a given level in the variable of interest in a given geographic region with respect to the weighted average value of the same variable in the contiguous or neighbouring areas.¹⁹

The local Moran’s I, in turn, enables visual exploration of clustering patterns formed as a function of the values of a given variable in the observation units analysed and in the neighbouring units, identifying the magnitude of the contribution of each spatial unit to the overall value. This calculation yields a map in which each spatial unit is differentiated according to its spatial autocorrelation with respect to neighbouring spatial units.¹⁹

Considering the high variability in population size between the departments of the CR, Bayesian statistical standardisation was applied to adjust values in order to control for the effect of these differences.²⁰ We also used the expected number of deaths obtained from direct standardisation to control for the effect of age on mortality rates.

Results

A total of 39 105 people in Argentina died due to PD and AD in the period 2000-2019, with a mean of 1955 deaths per year; women accounted for approximately 6 out of 10 deaths. The analysis of trends over time identified no statistically significant differences (P > .05) in ASMR trend in this time period for either disease, in either sex (Fig. 1; annual mortality statistics are presented in Supplementary Tables 1 and 2).

Figure 1.

Argentina (2000-2019). Trends over time in age-standardised mortality rates (per 100 000 population) for Alzheimer disease and Parkinson’s disease, by sex.

AD: Alzheimer disease; PD: Parkinson’s disease.

Descriptive analysis of province-level data

Fig. 2 shows the ASMRs (and 95% confidence intervals) for PD at the province level, for both sexes, compared against the mean for the country as a whole. In all provinces, these adjusted values are higher in men than in women, with values ranging from 1.1 times (La Rioja) to 2.2 times greater (Santa Fe and Jujuy). In both sexes, the CR and the provinces of Chaco, Mendoza, Neuquén, Río Negro, and Santa Fe present significantly higher ASMRs than the national average. The provinces of Córdoba and Formosa also presented higher rates among women.

Figure 2.

Argentina (2000-2019). Age-standardised mortality rate (per 100 000 population) for Parkinson’s disease in each province, compared with the national average, for each sex.

ASMR: age-standardised mortality rate.

Fig. 3 presents the same comparisons, but for AD. The difference between sexes in ASMR is the opposite of that observed for PD, with women presenting ASMRs ranging from 1.1 times (Chaco, Entre Ríos, Misiones, and Santa Fe) to 5.6 times greater (Tierra del Fuego) than men. ASMRs from the provinces of Córdoba, Entre Ríos, Formosa, La Pampa, Mendoza, Neuquén, Río Negro, and Santa Fe are significantly higher than the national average in both sexes. Among men, the ASMR for Misiones was also significantly higher than the national average, whereas among women, rates were significantly higher in Chubut, San Luis, and Tierra del Fuego.

Figure 3.

Argentina (2000-2019). Age-standardised mortality rate (per 100 000 population) for Alzheimer disease in each province, compared with the national average, for each sex. ASMR: age-standardised mortality rate.

The CR, and the provinces that make up the region, present significantly higher ASMRs for PD and AD than those observed at the national level, in both sexes, with the exception of ASMR for PD in Entre Ríos (both sexes) and Córdoba (men only). Particularly relevant data were those observed in Santa Fe, where rates were not only higher than the national average, but also higher than those observed in Córdoba and Entre Ríos, in the case of deaths due to PD in men and due to AD in women. For instance, the ASMR for PD in men in Santa Fe is 1.4-1.6 times higher than that of Entre Ríos and 1.3-1.4 times higher than that of Córdoba (Table 1).

Table 1.

Centre Region (2000-2019). Age-standardised mortality rates (per 100 000 population) for Parkinson’s disease and Alzheimer disease, by province, for the Centre Region, and Argentina as a whole, by sex.

Sex	Region	PD (ICD-10 code G20)			AD (ICD-10 code G30)
		ASMR	95% CI		ASMR	95% CI
			LL	UL		LL	UL
Men	Córdoba	1.97	1.80	2.11	2.60	2.44	2.78
	Entre Ríos	1.74	1.53	2.00	2.91	2.61	3.22
	Santa Fe	2.58	2.43	2.78	3.37	3.17	3.56
	Centre Region	2.20	2.07	2.32	2.98	2.85	3.10
	Argentina	1.77	1.71	1.85	1.86	1.80	1.90
Women	Córdoba	1.10	1.02	1.19	3.28	3.14	3.45
	Entre Ríos	0.86	0.71	0.99	3.16	2.89	3.41
	Santa Fe	1.16	1.07	1.24	3.66	3.49	3.83
	Centre Region	1.09	1.01	1.19	3.43	3.33	3.56
	Argentina	0.93	0.87	0.99	2.28	2.19	2.32

AD: Alzheimer disease; ASMR: age-standardised mortality rate; CI: confidence interval; ICD-10: 10th revision of the International Classification of Diseases; LL: lower limit; PD: Parkinson’s disease; UL: upper limit.

Values in italics are significantly higher than the national mean.

Spatial analysis of departments of the Centre Region

Several spatial differences at the department level were observed in mortality due to PD and AD in the CR, with a trend towards a concentration in space of similar ASMRs, with a very low likelihood that the distribution of these rates was random. Application of Moran’s I values to the expected number of deaths and the estimated population of each department, with a view to controlling the variance instability, yielded positive values (indicating presence of spatial concentration), showing statistical significance at the P < .01 level for AD, in both sexes, and at the P < .1 level in PD, in men only. In women, deaths due to PD presented a positive Moran’s I value, without reaching statistical significance at the P < .1 level (I = 0.044; pseudo P = .211).

With a view to detecting clusters with concentrations of higher or lower values than would be expected in a homogeneous distribution, we also calculated the local Moran’s I. Figs. 4 and 5 present the results as spatial autocorrelation maps, identifying clusters with high and low ASMRs (Supplementary Tables 3 and 4 present the mortality statistics for each department in the CR).

Figure 4.

Centre Region (2000-2019). Local analysis of spatial autocorrelation of age-standardised mortality rates for Parkinson’s disease in men, at the department level.

Figure 5.

Centre Region (2000-2019). Local analysis of spatial autocorrelation of age-standardised mortality rates for Alzheimer disease in men and in women, at the department level.

As shown in Fig. 4, mortality due to PD in men presents one cluster of high values and 2 clusters of low values. The cluster of high mortality rates includes 4 departments at the border of Córdoba (Marcos Juárez) and Santa Fe (San Martín, Belgrano, and Caseros), whereas the clusters of low mortality are located in the capital department of the province of Córdoba and 2 departments to the north (Colón and Totoral), on the one hand, and in the north-east of Entre Ríos (Feliciano, Federación, Federal y Concordia), on the other. No statistically significant clusters in PD mortality rates were observed in women.

Fig. 5 shows data on mortality due to AD in both sexes. In both men and women, a cluster of high values was detected in departments in the south-east of Córdoba and the south-west of Santa Fe. In men, the cluster extended across the majority of departments in southern Santa Fe (San Martín, San Jerónimo, Belgrano, Iriondo, Caseros, San Lorenzo, and General López) and 2 departments in the south-east of Córdoba (Marcos Juárez and Presidente Roque Sáenz Peña). In women, on the other hand, the cluster included the departments of General Roca, Río Cuarto, and R. Sáenz Peña, in southern Córdoba, and General López, in southern Santa Fe.

For both sexes, a cluster of low values was observed in the north-west of Córdoba; this cluster was larger for men, including (in addition to the departments of Ischilín, Totoral, and Punilla) the departments of Colón, Cruz del Eje, Capital, Río Primero, Santa María, San Javier, and Tulumba.

Discussion

This article summarises the findings of a descriptive study of rates of mortality due to AD and PD in Argentina between 2000 and 2019, with an analysis of data at the provincial and (in the CR) department levels.

When considering variations over time in mortality due to these diseases, our results coincide with those reported in a recent analysis from the 2021 Global Burden of Disease study,¹ which reports that global ASMRs for these diseases remained stable in the period 1990-2021. However, differences have been reported in the trends observed in other countries and regions, with statistically significant increases in recent decades.^5–8 Pan American Health Organization data for the Region of the Americas suggests that ASMRs did not significantly vary between 2000 and 2019. A similar trend is reported in Argentina,⁹ and is also supported by our own findings.

Despite the high coverage of mortality records in Argentina (> 95%), almost one-third of deaths recorded between 2000 and 2019 corresponded to “garbage codes,”²¹ including intermediate, terminal, and unspecified complications and poorly defined causes, which are listed as the basic cause of death. This high rate of garbage codes raises the suspicion that deaths due to neurodegenerative diseases may be under-reported, particularly due to the fact that the direct cause of death is often cardiorespiratory failure (or other causes), with the underlying cause not being reported.

No study to date has analysed overall rates of mortality due to these diseases in Argentina; although the WHO has published estimates, they differ greatly from the results of the present study.¹⁰ There are 2 possible reasons for these differences. Firstly, the basic causes of death included under the denominations AD and PD: unlike the WHO studies,²² our study only included the specific diseases AD (G30) and PD (G20) as basic causes of death, and excluded other entities that frequently present different pathophysiology, diagnosis, and progression, such as those listed in the material and methods section. Secondly, we used a different study period: the WHO estimates are based on data from the last year available (2019), whereas we calculated an average for the entirety of the period 2000-2019.

Turning our attention to the relationship between province-level and national ASMRs, nearly one-third of provinces presented PD mortality rates above the national mean; for AD, this was the case for nearly half of provinces. In the CR as a whole, both sexes showed mortality rates significantly higher than the national mean. In the analysis of individual provinces in this region, the rates observed in Santa Fe were significantly higher than the national mean for both sexes and diseases, and higher than other provinces in the CR for PD in men and AD in women.

In the spatial analysis of clusters of higher or lower values than would be expected in a random distribution, the CR presented 2 equidistant geographical areas. Firstly, we observed an area with departments presenting very high rates, in south-eastern Córdoba and south-western Santa Fe, for AD in both sexes and for PD in men. Secondly, a group of departments was detected in the centre-north of Córdoba province, for both diseases, and in north-eastern Entre Ríos for PD. A partial overlap was observed in the regional distribution of clusters of high risk, for both diseases and in men (Figs. 4 and 5).

Populations in the Centre Region of Argentina are similar in terms of demographic growth and population density, which differentiate them from the other neighbouring provinces; these populations present some of the highest percentages of people aged 65 years or older.²³ Simultaneously, the region is the country’s most important agro-industrial centre; due to its significant environmental exposure and natural conditions, its population presents one of the highest rates of exposure to drinking water with high levels of arsenic in the country. This is reflected in cancer mortality rates, as reported in previous studies.²⁴ This series of adverse conditions, as well as growing population ageing, justified the decision to focus our study on neurodegenerative diseases, with a particular focus on mortality in the CR rather than other Argentinian regions.

Analysis of the sociodemographic and environmental conditions in the departments making up each of the clusters identified would shed light on the characteristics associated with mortality due to these diseases. In this regard, exploring possible associations between these diseases, in terms of environmental exposure and sociodemographic circumstances, might yield information on the possible causes in terms of the pathogenesis and development of these diseases in Argentina.

In general terms, we must underscore a fundamental limitation in our study: the use of mortality data, rather than incidence data. However, given the lack of representative prevalence registries for these diseases in Argentina,²⁵ mortality records constitute an important source of information for the analysis of healthcare, particularly in a global context in which neurological diseases, including neurodegenerative diseases, have become the leading cause of disease burden, given the increase in recent decades in the number of years of life lost due to these diseases, making them a top public health priority.¹

Funding

This study has received no specific funding from any public, commercial, or non-profit organisation.

Acknowledgements

The authors thank Agustín Ibáñez and colleagues for their valuable comments and suggestions on the preliminary version of this article.