Birthplace predicts risk for prehospital sudden cardiac death in middle-aged men who migrated to metropolitan area: The Helsinki Sudden Death Study

Abstract

Background. Eastern-born male Finns, irrespective of their place of residence, have high mortality from coronary heart disease (CHD), and half of such deaths are sudden.

Aim. To study whether eastern birthplace alone or combined with life-style factors predicts risk for prehospital sudden cardiac death (SCD) in the new (west) low-mortality area of residence.

Method. Prospective case-control autopsy study of all (700) out-of-hospital deaths of men aged 35–69 years in metropolitan Helsinki during 1981–82 and 1991–92. Data on CHD risk factors were obtained for 405, of whom 149 died of SCD (cases) and 256 of other causes (controls).

Results. A birthplace-by-age interaction with SCD (P=0.024) and with myocardial infarction (P=0.005) appeared. Men ≤54 years born in the east were more often victims of SCD (odds ratio 2.99, 95% confidence interval 1.38–6.49, P=0.006) than were men born in the west, independently of CHD risk factors. SCD was predicted also by alcohol consumption, age, smoking, and hypertension. Amongst older (>54 years) men no association with birthplace was any longer evident, but alcohol and socio-economic status predicted SCD.

Conclusions. Birthplace-based risk for SCD suggests the contribution of early life environment or genetic east-west differences, reflecting Finns’ two-phase settlement history.

• Birthplace
• Finland
• risk factors
• sudden cardiac death

Introduction

Sudden cardiac death (SCD) accounts for half of all deaths due to cardiovascular causes 1 and is resistant to the trend toward improvement in cardiovascular mortality 2, 3. Risk for SCD can be only partially predicted or prevented, because specific determinants of SCD are poorly known, in contrast to the well known risk factors of the more slowly developing symptomatic coronary heart disease (CHD) 4. Of SCDs, three-quarters take place out-of-hospital, and half the time the SCD is the first and only manifestation of underlying disease 5. Recently, unfavourable atherogenic and prothrombotic variants of cardiovascular genes have been found to cluster amongst SCD victims 6, 7, and some of these genes show an east-west difference among late and early settlements in Finland 8–13.

The world's highest CHD mortality in the economically less developed rural east Finland led to the North Karelia project 14, aimed at reducing dietary fats, smoking, and other cardiovascular risk factors at community level and later nationwide. From 1972 to the mid-1980s, major improvements occurred in population risk factor profiles, associated with simultaneous changes in CHD mortality, not only in North Karelia but also in other areas of Finland, although the east-west difference remained 15. However, in men with no prior history of myocardial infarction (MI), the annual average decline in CHD deaths was only half the CHD decline in men with previous MI 5, with no significant decrease in prehospital mortality 16.

Since the 1950s, Finland has undergone rapid economic and structural changes, with the younger population migrating from the rural east to the urban south and south-west. This resulted in a 26% increase in the Helsinki population between 1951 and 1970 17; for instance, 31.7% of the Helsinki residents in late 1981 were born in the eastern provinces 18. Besides changes in risk profiles, mortality trends may also have shown effects of this migration. If the excess (eastern) SCD mortality is mainly associated with cardiovascular risk factors, migrants would benefit from a move to an urban low-mortality area—assuming that the new environment will gradually lead individuals to adopt healthier life-styles. Conversely, migrants may carry their early susceptibility or less favourable genetic heritage 19 and contribute negatively to mortality rates in the environment of their new residence, as well. In migration studies in Finland and in the USA, of greater importance in determining risk for CHD death was birthplace 20–22. No data were, however, available on any involvement of life-style. Birthplace-based risk for CHD has also appeared in Finnish migrants to Sweden 23, 24. On the contrary, in the British Regional Heart Study, area of adult residence and factors in adulthood were more important determinants for CHD event and cardiovascular mortality 25, 26, whereas a national study from England and Wales showed the significant influence of birthplace on mortality from CHD 27. In a Hawaiian-Japanese study, risk for SCD of migrants converged over time with that of the local population 28. Data on association of birthplace and adult life-style factors on prehospital coronary SCD among Caucasians has, in fact, to our knowledge, never appeared.

We investigated whether birthplace alone or combined with CHD risk factors predicts prehospital coronary SCD in men born in a region either of high or of low CHD mortality and dying in the low-mortality metropolitan region.

Key messages

• Men aged ≤54 years born in a high-CHD (coronary heart disease) mortality east Finland were more often victims of sudden cardiac death (SCD) than were men born in a low-mortality west, independently of CHD risk factors, whereas amongst older (>54 years) men no association with birthplace was any longer evident.

• Birthplace-based risk suggests the contribution of early life environment or genetic east-west differences, reflecting the country's two-phase settlement history.

Abbreviations

Table

BMI	body mass index
CHD	Coronary heart disease
HSDS	Helsinki Sudden Death Study
MI	Myocardial infarction
OR	odds ratio
SCD	sudden cardiac death

Methods

The study was approved by the Ethics Committee of the Department of Forensic Medicine, University of Helsinki, Finland. The Helsinki Sudden Death Study (HSDS) was launched in 1981 as an independent study parallel to the national WHO MONICA project 3 to focus on risk factors for SCD and to follow changes in its epidemiology. The series was collected with a 10-year interval in 1981 and 1982 (n = 400) and in 1991 and 1992 (n = 300), and comprised all men aged 35–69 years (median 54 years) who died out of hospital and were subjected to forensic autopsy in the city of Helsinki. The Helsinki metropolitan area comprises several municipalities with a total population of about 1.5 million people. Indications for forensic autopsy were: out-of-hospital death of a previously healthy man or a deceased man who had not seen a doctor for more than one year; violent out-of-hospital death caused by accident, suicide, or homicide. Individuals treated in hospital immediately before death, as well as decomposed or mutilated bodies were excluded. As a result, the HSDS autopsy series covered 60.6% of all forensic autopsies on men within the chosen age-group. Furthermore, the series comprised 24.6% of all deaths of men aged 35–69 years in Helsinki during the study period. Of these 700 men, 288 (41.1%) died of cardiovascular diseases, 140 (20.0%) of other diseases, and 272 (38.9%) of violent causes. According to the Statistical Yearbook of the city of Helsinki 18 at the end of 1981, 88,911 men aged 35–69 were living in the metropolitan area; at the end of 1992, this figure rose to 100,725. Of these men, in 1981, 28,185 (31.7%) were eastern-born, and in 1992, 32,131 (31.9%).

Diagnosis of myocardial infarction at autopsy

The autopsies were performed by experienced forensic pathologists using a detailed protocol. Any myocardial infarction (MI) was confirmed by macroscopic and histologic examination and nitro-blue-tetrazolium (NBT) staining of the myocardium. The presence/absence of neutrophil granulocytes was considered as diagnostic of an acute MI and presence/absence of fibrous scar tissue as diagnostic of an old MI. Any recent or organizing macroscopic coronary (red) thrombosis was recorded during opening of the coronary arteries 7. Prehospital sudden cardiac death (SCD) was classified into coronary (codes I20–I25) and non-coronary by the International Classification of Diseases (ICD) 10. Of the 700 men, 228 (32.6%) suffered prehospital SCD due to CHD, comprising 79.2% of all deaths due to cardiac causes. Of coronary SCDs, 194 (85.1%) died of acute or old MI. The 60 (8.6%) non-coronary SCDs were due to cardiomyopathy, myocarditis, valvular disease, non-classifiable hypertrophy, or dilatation of the heart in the absence of significant coronary disease.

Risk factors underlying prehospital cardiac death

Within 2 weeks after death, a spouse, a relative, or a close friend of the deceased was interviewed by pathologists using a structured questionnaire comprising 50 questions. Data on hypertension, diabetes, smoking, alcohol consumption, and socio-economic status were collected and to recorded as binary variables (absence or presence). Daily dose of alcohol was estimated by the Alcohol Use Disorders Identification Test and by Michigan Alcoholism Screening Test 29, 30 and transformed into a daily average (absolute ethanol g/day, one drink = 12 g) 31. The men were then classified as non-drinkers to moderate drinkers (0–5 drinks/day) or as heavy drinkers (>5 drinks/day). Their occupations were categorized as non-manual or manual. Body mass index (BMI) was calculated (kg/m²), and birthplace of the deceased as well that of his parents was classified in Finland as west (early settlement) with low CHD mortality, and east (late settlement) 13 with high CHD mortality (Figure 1).

Figure 1.  Age-standardized mortality of middle-aged men from coronary heart disease by province in Finland, 1971–1975 20 (upper figures) and 1997–2003 (lower figures) 32. The mean value for the entire male population of Finland is 100. Diagram shows average population by western and eastern birthplace among male residents of Helsinki 18 and out-of-hospital deaths in the Helsinki Sudden Death Study series during 1981–1982 and 1991–1992.

Statistical analysis

Statistical analyses were performed by taking coronary SCDs and the subcategories as the cases, and deaths due to other causes as the controls. Differences between means were tested by analysis of variance (ANOVA) and significance between percentages by Pearson chi-square test. Logistic regression analysis served to investigate odds ratios (OR) and 95% confidence intervals (CI) between birthplace and SCD with age, BMI, alcohol consumption, smoking, hypertension, diabetes, and socio-economic status as confounders. All analyses were performed with SPSS statistical software (v. 12.01, SPSS Inc., Chicago, IL, USA). Statistical significance was set at P < 0.05.

Results

Of the 700 men, data on birthplace was available for 692 (98.9%), of whom 340 (49.1%) were born in the east and 352 (50.9%) in the west. Between 1981 and 1992 the average male population in this age-group was 94,818, of whom on average 30,158 (30.3%) were born in the east and 64,660 (69.7%) in the west 18. From these figures, calculation shows that the ratio of eastern-born men among out-of-hospital deaths greatly (chi-square = 94.89, P < 0.0001) exceeded their percentage of the metropolitan population (Figure 1). CHD risk factors were available for 405 men (57.9%), and of these more than half (n = 259, 64.0%) died of diseases, of which 149 (57.5%) were SCDs due to CHD with or without MI. The second most common cause was non-natural death (n = 146, 36.0%). Of the 405 men with background data, 215 (53.1%) were born in the west (low CHD) and 190 (46.9%) in the east (high CHD) (Table I). Of the 287 men with no background data 137 (47.7%) were born in the west and 150 (52.3%) were born in the east. There were no differences in geographic origin between those from whom interview data were obtained compared to those of whom no relatives could be contacted (P = 0.165). There were no significant differences in age, BMI, or measured CHD risk factors between those with or without interview data, with the exception of a slightly higher percentage of men with a history of alcoholism (P = 0.074) and manual workers (P = 0.067) among those from which interview data could not be obtained.

Table I.  Characteristics of the study population. Values are means±SD for normally distributed continuous variables, and numbers of cases (percentages) for categorical variables.

	All			≤54 Years			>54 Years
	West (n = 215)	East (n=190)	P-value	West (n = 108)	East (n = 102)	P-value	West (n = 107)	East (n=88)	P-value
Age (years)	53.36±10.07	53.24±9.58	0.902	44.62±5.65	45.78±5.98	0.149	62.18±4.02	61.89±4.15	0.609
Height (cm)	175.56±6.83	175.08±7.03	0.492	176.56±6.64	176.58±6.76	0.988	174.54±6.90	173.35±6.98	0.235
Body mass index (kg/m^2)	25.05±4.88	24.82±4.45	0.630	24.59±5.02	24.75±4.06	0.806	25.51±4.71	24.91±4.89	0.388
Alcohol use
< 1 drink/day	64 (29.8)	41 (21.6)	0.061	22 (20.4)	12 (11.8)	0.071	42 (39.3)	29 (33.0)	0.364
1–5 drinks/day	50 (23.3)	48 (25.3)	0.223	25 (23.1)	24 (23.5)	0.570	25 (23.4)	24 (27.3)	0.208
> 5 drinks/day	101 (47.0)	101 (53.2)	0.240	61 (56.5)	66 (64.7)	0.270	40 (37.4)	35 (39.8)	0.775
Smokers	173 (80.5)	162 (85.5)	0.184	85 (78.7)	89 (87.3)	0.152	88 (82.2)	73 (83.0)	0.896
Hypertension	54 (25.1)	36 (18.9)	0.136	19 (17.6)	11 (10.8)	0.159	35 (32.7)	25 (28.4)	0.517
Diabetes	39 (18.1)	49 (25.8)	0.062	18 (16.7)	20 (19.6)	0.580	21 (19.6)	29 (33.0)	0.034
Manual workers	129 (60.0)	127 (66.8)	0.154	65 (60.2)	65 (63.7)	0.597	64 (59.8)	62 (70.5)	0.122
Cause of death
Cardiac disease	96 (44.7)	86 (44.3)	0.811	30 (27.8)	36 (35.3)	0.249	66 (61.7)	50 (56.8)	0.489
Coronary heart disease	73 (34.0)	76 (40.0)	0.208	16 (14.8)	30 (29.4)	0.011	57 (53.3)	46 (52.3)	0.889
Other disease	41 (19.0)	36 (18.9)	1.000	21 (19.4)	23 (22.5)	0.581	20 (18.7)	13 (14.8)	0.471
Intoxication and violent	78 (36.3)	68 (35.8)	0.811	57 (52.8)	43 (42.2)	0.126	21 (19.6)	25 (28.4)	0.151

No differences existed in causes of death or CHD risk factors between men born in the east or the west, with the exception of a slightly higher (25.8% versus 18.1%, P = 0.062) occurrence of diabetes and slightly lower percentages of non-drinkers (<1 drink/day) (21.6% versus 29.8%, P = 0.061) among men born in the east. Of the series, 76 (40.0%) men from the east and 73 (34.0%) from the west died of SCD due to CHD (P = 0.208 for east-west difference). Significant age-by-birthplace interactions occurred with risk for prehospital coronary SCDs (P = 0.024), especially regarding all (acute or old) MIs (P = 0.005) and acute MIs (P = 0.004).

In logistic regression analysis (Table II) among men ≤54, with height, age, BMI, smoking, alcohol, diabetes, hypertension, and socio-economic status as confounders, eastern birthplace remained in the model as a strong predictor of SCD (OR 2.99, 95% CI 1.38–6.49, P = 0.006). Other significant variables were drinking (OR 0.23, 95% CI 0.11–0.49), age (OR 1.08, 95% CI 1.01–1.16), smoking (OR 3.94, 95% CI 1.10–14.09), and hypertension (OR 2.62, 95% CI 1.01–6.78). Eastern birthplace also showed similar associations with all (acute or old) MIs (OR 3.02, 95% CI 1.27–7.15), with acute MI with or without thrombosis (OR 2.86, 95% CI 1.09–7.51), and with the subset of MI due to coronary thrombosis (OR 4.24, 95% CI 1.15–15.47) (Figure 2). To exclude the possibility of an uncontrolled interaction between CHD risk factors in logistic regression analysis, we did the same analysis without any confounding factors among men ≤54 years. The association with eastern birthplace also emerged in this model, since eastern-born men were twice as often victims of SCD (P = 0.011). Furthermore, the association with birthplace was also evident if we split the series using only men of working age (35–64 years): 55 men out of 153 (35.9%) from the east and 41 out of 161 (25.5%) from the west died from SCD (P = 0.046).

Figure 2.  Frequency distribution and multivariate birthplace-based risk (odds ratio (OR) and 95% CI; above bars) for prehospital sudden cardiac death due to coronary heart disease including information on all myocardial infarction (acute or old) and coronary thrombosis among ≤54-year-old (upper figure) and >54-year-old men (lower figure) born in east and west Finland in the Helsinki Sudden Death Study series during 1981–1982 and 1991–1992.

Table II.  Risk for prehospital sudden cardiac death (SCD) due to coronary heart disease (CHD) associated with birthplace with and without CHD risk factors in men aged 35–69 years in metropolitan Helsinki. Values are means for normal distributed continuous variables, percentages for categorical variables and odds ratio (95% CI) for risk for SCD due to CHD (codes I20–I25).

	Men ≤54 years (n = 210)				Men >54 years (n = 195)
	Controls (n = 164)	SCD (n = 46)	OR (95% CI) (multivariate)	P-value	Controls (n = 92)	SCD (n = 103)	OR (95% CI) (multivariate)	P-value
	East/West (n = 72/92)	East/West (n = 30/16)^a			East/West (n = 42/50)	East/West (n = 46/57)
Birthplace (%)	43.9/56.1	65.2/34.9^a	2.99 (1.38–6.49)	0.006	45.6/54.3	44.7/55.3	0.87 (0.47–1.60)	0.652
Age (years)	45.13/44.17	47.37/47.18	1.08 (1.01–1.16)	0.017	61.69/61.70	62.07/62.61	1.03 (0.95–1.12)	0.424
Height (cm)	176.61/176.35	176.50/177.81	1.02 (0.96–1.08)	0.447	173.74/174.10	173.00/174.92	0.99 (0.95–1.04)	0.730
BMI (kg/m^2)	24.39/24.70	25.61/23.95	1.01 (0.94–1.10)	0.733	24.52/25.02	25.28/25.94	1.05 (0.98–1.12)	0.198
Smoking	87.5/75.0 ^a	86.7/100	3.94 (1.10–14.09)	0.035	76.2/86.0	89.1/78.9	0.95 (0.42–2.16)	0.905
>5 drinks/day	73.6/62.0	43.3/25.0	0.23 (0.11–0.49)	0.000	40.5/54.0	39.1/22.8	0.43 (0.22–0.84)	0.013
Hypertension	6.9/15.2	20.0/31.3	2.62 (1.01–6.78)	0.047	21.4/30.0	34.8/35.1	1.43 (0.73–2.82)	0.302
Diabetes	20.8/15.2	16.7/25.0	1.07 (0.41–2.77)	0.889	21.4/18.0	43.5/21.1^a	1.70 (0.82–3.52)	0.152
Manual workers	66.7/63.0	56.7/43.8	0.63 (0.29–1.36)	0.238	66.7/50.0	73.9/68.4	2.07 (1.06–4.07)	0.034

^aP < 0.05.

CI = confidence interval; BMI = body mass index; OR = odds ratio.

Among men aged >54 years, birthplace was no longer associated with prehospital SCD or its subcategories (univariate P = 0.889, multivariate P = 0.652) (Figure 2). In this age-group, drinking reduced risk for SCD (OR 0.43, 95% CI 0.22–0.84), whereas here low socio-economic status (OR 2.07, 95% CI 1.06–4.07) seemed to predict SCD.

Discussion

In this population-based autopsy study of middle-aged metropolitan men, birthplace in a high-CHD-mortality region was associated with prehospital coronary SCD independently of CHD risk factors and socio-economic status. We found no significant difference in CHD risk factors in favour of men born in the west, with the exception of a slightly lower frequency of reported diabetes at an older age. Our results may be explained by other unknown risk factors not possible to control for, such as early life environment or by genetic differences reflecting the two-phase settlement history of Finland.

In a Hawaiian-Japanese study 28, the pattern of SCD of migrants converged over time with that of the local population, suggesting the influence of adult life-style factors. Similar findings have been reported in Great Britain 25, 26 relative to CHD event and cardiovascular mortality, but conflicting results also exist 27. In Finland, middle-aged men born in the east who moved to the west maintained their higher CHD mortality independently of socio-economic status 20, 21. Birthplace-based risk for CHD has also appeared in Finnish migrants to Sweden 23, 24 and in southern-born African-Americans in New York 22. In this study, we could for the first time assess the contribution of birthplace to prehospital SCD and note whether this was associated with adult CHD risk factors and socio-economic status, using data obtained by interviews with the next of kin.

Birthplace was associated with SCD in men ≤54 years, but was thereafter no longer significant. This is in accordance with previous findings regarding CHD 20. Genetic influence on SCD 33 and atherosclerosis 6 may exert its strongest effects at a younger age, and other factors may play a more significant role in later life. Moreover, the mechanism and occurrence of SCD seem to differ at different ages. Recent reports suggest that younger SCD victims suffered more often from transmural (Q-wave) infarction caused by occluding fibrin-rich (red) coronary thrombosis, whereas in older victims with unstable angina pectoris who develop non-transmural (non-Q-wave) infarction, the main reason is severe coronary disease 7, 34. Thus, risk factors for SCD at younger ages may differ from those at older ages.

Of CHD risk factors, smoking and hypertension predicted SCD for men ≤54 years, whereas drinking in excess of 5 drinks/day seemed to play a protective role against coronary SCD, possibly by inducing reverse cholesterol transport 35 affecting thrombogenesis 36. Many other publications have shown that excessive alcohol intake may increase the risk for SCD, thus further studies are warranted 37–39.

After adjusting for these CHD risk factors, birthplace-based risk strengthened for men ≤54 years, to weaken to non-significant among men >54 years. This suggests that birthplace-based risk may interact with known CHD risk factors and with an unknown third determinant contributing to out-selection of men at risk. We found no significant birthplace-by-conventional-risk-factor interaction with SCD.

Interestingly, the decline in CHD mortality in North Karelia province between 1969 and 1979 was 24% but was 12% in the country as a whole 40, and this has been related to the North Karelia project. During the period from 1950 to 1970, however, this province lost 29% of its population, mostly to the capital region studied here 17. No differences in CHD mortality have been shown between eastern-born men who migrated to the capital area or who stayed in the east in the 1971–1985 period 20. We thus suggest that at least part of the less rapid decline in CHD mortality in the capital region compared to North Karelia might be due to the effect of population migration on CHD mortality.

As deficiencies in maternal nutrition and poor infant growth have been associated with CHD death in later life 41, 42, lower socio-economic status in the east may be connected with increased CHD risk in adulthood. We could not confirm that regional differences in socio-economic status, in BMI, or in height as markers of early life experience contributed to the risk for SCD. Our findings are supported by studies showing that the east-west contrast between socio-economic groups has varied little despite continuous economic growth throughout the 20th century. In our study, lower socio-economic status was associated with SCD only among men over 54 years old and had no effect on the association between birthplace and SCD. All this suggests that lower socio-economic status in childhood may not play a major role in the geography of SCD in Finland.

Atherogenic or prothrombotic gene variants, such as the S2 allele of apolipoprotein A-I/C-III gene complex 8, the D allele of angiotensin-converting enzyme (ACE) 10, and the familial hypercholesterolaemia mutation (FH-North Karelian) 11, are more common and factor-VII coagulant activity higher 9 in eastern Finland. A dual theory of Finns exists, supported by Y-chromosome markers of two distinct (east-west) male lineages between the early and late settlement areas 12, 13. Thus, the mechanism by which CHD risk factors affect the development of SCD may in part be genetically regulated.

Limitations inherent in forensic autopsies cluster in heavy alcohol consumers, although alcohol indeed is one of the risk factors for out-of-hospital death 38. Our series is therefore not a representative sample of the general population, but refers to a population at a high risk of sudden death outside hospital. Our results might be explained by assuming a greater risk of western men for violent deaths, masking a similar SCD risk among men from east and west Finland. However, there are no studies suggesting that western men are more prone to die violently. On the contrary, in a previous study eastern-born men who migrated to the capital region had slightly higher risk for suicide and other type of violent death compared to western-born men 20. We thus suggest that our finding is a real one and reflects the more rapid progression of coronary disease among men originating from the east. The present series is biased in that way that subjects of whom relatives could be contacted were in a slightly more favourable socio-economic situation, and they were less often known to have a history of alcoholism. However, there were no east-west differences regarding the lack of information on risk factors.

Data on CHD risk factors collected from a questionnaire may underestimate the true magnitude of risk profiles. Nor do we have cholesterol values measured from our subjects. However, previous reports have shown no east-west differences in levels of total cholesterol, in systolic blood pressure, or in smoking prevalence during the study period 3. Data on duration of stay in the area of residence was unavailable for any of the deceased. Mean age of interprovincial and international immigrants was approximately 25 years 17, and in Swedish migration studies, higher mortality from CHD in Finnish men persists in second-generation immigrants 24. Thus, duration of stay may not be considered a major confounder in the present study.

Our results suggest that factors associated with birthplace offer a significant contribution to risk for SCD in first-generation migrants within Finland, reflecting its two-phase settlement history. These factors most probably associate with population genetic differences stemming from this history. Tracking of birthplace in the high mortality CHD area seems to play an important role in identifying individuals at high SCD risk among a migrant population in its new place of residence.

Acknowledgements

The authors would like to thank Paula Maasilta, for her valuable comments on the manuscript.

This work was supported by the Yrjö Jahnsson Foundation, the Finnish Foundation of Alcohol Research, the Elli and Elvi Oksanen Fund of the Pirkanmaa Fund under the auspices of the Finnish Cultural Foundation, the Medical Research Fund of Tampere University Hospital, the Finnish Medical Foundation and the Finnish Foundation for Cardiovascular Research, the Maili Autio Fund, Kaisa Ahonen Fund, and Ester and Ilmari Pohto Fund under the auspices of the Finnish Cultural Foundation.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.