Decreased heart rate variability (HRV) has been associated with increased mortality after acute myocardial infarction [52] and after coronary artery bypass grafting surgery [53]. A one-standard deviation (SD) drop in the SD of total normal RR intervals was associated with a hazard ratio of 1.47 (95 % CI: 1.16, 1.86) in the Framingham Heart Study [54]. Together with increased sympathetic tone, it has been related to poor prognosis in several patient populations [55] and an increased risk of sudden cardiac death in patients with myocardial infarction, congestive heart failure, or hypertension [56]. A decreased nighttime HRV was reportedly a strong marker for the development of stroke in apparently healthy subjects [57]. In healthy young adults, reduced HRV indices were independently associated with increased CRP concentrations [58]. Understanding how the natural environment influences HRV is thus important, both on earth and in extraterrestrial space, where crews are spending ever extending journeys on the International Space Station.

In Stockholm, Sweden, the 24-hour SDs of hourly estimates of heart rate were obtained by Holter monitoring from 50 men who had had an acute myocardial infarction or had angina pectoris. Compared with the 24-hour SD of heart rate from 50 clinically healthy men of similar age, it was 20.5 % lower in the patients (P = 0.002) [59]. Not only is HRV decreased in the presence of ischemic heart disease, a low HRV is also a predictor of cardiovascular disease risk. Indeed, in Tokyo, Japan, the heart rate of 121 normotensive subjects and 176 treated hypertensive patients was measured around the clock by ambulatory monitoring for 2 days [60]. At the time of monitoring, all 297 participants were free of heart disease. The incidence of ischemic heart disease was recorded prospectively for 6 years. As compared to participants free of adverse outcome, patients who had developed coronary artery disease within the 6 years of follow-up have a 20.0 % lower HRV (P = 0.04). Results from these two studies were aligned with those of a retrospective analysis of archived data on all crews of the Soyuz spacecraft for 1990–1994. Based on the ECG records from 47 male and 2 female cosmonauts, the SD of heart rate was 30 % lower in cosmonauts examined during a magnetic storm, as compared to cosmonauts monitored on quiet days (P = 0.041) [61], suggesting that exposure to geomagnetic disturbances may increase cardiovascular disease risk (Fig. 13.4).

A reduced HRV on days of high magnetic activity vs. quiet days has also been documented in the longitudinal electrocardiographic (ECG) record of a clinically healthy man, both in terms of the coefficients of variation of R-R intervals and of the total spectral power [62]. In the frequency domain, the decrease in HRV was statistically significant in the very-low-frequency region (periods in the range of 6.5–333 s) but not in the low-frequency (periods of about 10.5 s) or high-frequency (periods of about 3.6 s) region, suggesting an underlying physiological mechanism other than the parasympathetic system as responsible for changes in HRV in response to magnetic activity [62, 63].

To follow up on these results, effects of geomagnetic disturbance on HRV were assessed in adults living at high latitude, in the auroral oval, where magnetic storms are more frequent and more intense than at lower latitudes. Nearly continuous, 7-day beat-to-beat ECG records were obtained between December 1998 and November 2000, on clinically healthy men (N = 15) and women (N = 4) from Finnmark College in Alta, Norway (69°56′ N, 23°22′ E). Their age ranged from 21 to 59 years. A geomagnetic record was obtained from the Auroral Observatory of the University of Tromsø, in Tromsø, Norway (69°39′ N, 18°56′ E), including 1-min data of total intensity (F, in nT), declination (D, angle between geographic and magnetic north, in degrees), inclination (I, angle between horizontal plane and magnetic direction, in degrees), horizontal intensity (H in nT), and vertical intensity (Z in nT). A decrease in HRV (P = 0.002) and an increase in the 24-hour average of heart rate (P = 0.020) were documented on days of high geomagnetic disturbance [64] (Fig. 13.5).

As expected, the decrease in spectral power was found primarily at frequencies lower than 0.04 Hz and was not statistically significant around one cycle in 3.6 s. The decrease in HRV was more pronounced in the VLF (21.9 % decrease, P < 0.001) than in the ULF (15.5 % decrease, P = 0.009) region of the spectrum [64–67]. This effect was more pronounced in spring and autumn, in the presence of a daily light-dark cycle, than in summer (continuous light) or in winter (continuous darkness) [64, 68]. This finding is clinically important because a reduction in the VLF range is associated with an increased risk of cardiac and dysrhythmic death [64]. Several mechanisms have been proposed for the VLF component, namely, thermoregulation [69], the renin-angiotensin-aldosterone system, and other humoral factors [70].

Investigation of 7-day ECG records from five clinically healthy young men living above the arctic circle compared HRV measures assessed over separate 24-hour spans among days of low, middle, and high geomagnetic activity, defined as days with values of the geomagnetic index ap < 7, 7–20, and 20–45, respectively. A graded response was observed, the extent of HRV decrease depending on the degree of geomagnetic activity. Specifically, as compared to days when ap < 7, TF was decreased by 18.1 % and 31.6 %, ULF by 18.1 % and 27.5 %, and VLF by 12.9 % and 28.6 % on days when 7 < ap < 20 and 20 < ap < 45, respectively [64, 71–73]. A graded decrease in HRV to geomagnetic activity suggests the existence of human magnetoreceptors. Phillips et al. [74] proposed a light-dependent magnetoreception mechanism and established a link between magnetic field sensitivity and the visual system in eastern red-spotted newts. The results suggest the operation of a light-darkness-alternation-influenced, if not light-dependent, magnetoreception mechanism in humans.

The existence of such a mechanism is supported by results from cross-spectral analyses. Group-averaged coherences at frequencies of 0.001–0.0001 Hz (0.25–5 hour) differed with statistical significance (P < 0.001) among the three lighting conditions, being on the average smaller during long days or long nights than during spring or autumn. Intermittent fields and components in the ultralow-frequency band of the electromagnetic field have been considered to be more bioeffective than the predominantly sinusoidal power line fields in the extremely low-frequency band. The ultralow-frequencies of the electromagnetic field are composed of six kinds of continuous pulsations, including Pc 1 (period range: 0.2–5 s), Pc 2 (5–10 s), Pc 3 (10–45 s), Pc 4 (45–150 s), Pc 5 (150–600 s), and Pc 6 (10 min–5 hours). The cross-spectral analysis results thus suggest that the magnetoreception mechanism may involve the Pc 6 geomagnetic pulsations [64].

Associations between geomagnetic storms and HRV were observed not only on the day of a storm but also for 2 days immediately following a geomagnetically disturbed day. Whereas changes in HR, NN interval, r–MSSD, and pNN50 were found to be statistically significant only during the day of the storm, changes in SDNNIDX (30 min) and in the power-law scaling of the power spectra remained statistically significant for 1 or 2 days following the day with a geomagnetic storm [64]. The 1/f slope was statistically significantly steeper not only on the day of the storm but also on the 2 days following a geomagnetic storm (P < 0.001), suggesting that geomagnetic disturbances have a lasting influence on the fractal scaling of HRV (Fig. 13.6).

A decrease in HRV and an increase in heart rate in association with solar activity, as gauged by Wolf numbers, are also observed in a 26-year longitudinal record of around-the-clock data collected mostly at 30-min intervals by a clinically healthy man (Figs. 13.7 and 13.8). Magnetic storms are usually more frequent at times of high solar activity. The data are modeled by the least-squares fit of cosine curves with periods of about 1.0 and 10.5 years. Nonlinearly, the periods and their 95 % confidence limits (CI) are estimated as 1.02 [95 %CI: 1.00, 1.04] and 10.60 [95 %CI: 9.21, 11.98] years for the monthly means and as 0.98 [95 %CI: 0.96, 1.00] and 10.44 [95 %CI: 8.85, 12.03] years for the monthly SDs of heart rate. In both cases, the about-11-year solar cycle length is included in the 95 % CI. An about-10.5-year component in the SD of heart rate had already been detected in the first 11 years of this record [63].

The mapping of chronomes, and particularly of cycles shared between physiology and the environment, may thus prove to be a valuable approach to address the goals specified at the beginning of this chapter. For this approach to be successful, reference values from womb to tomb are essential, a project started by the International Chronome Ecology of HRV (ICEHRV) Working Group [64]. For this purpose, around-the-clock ambulatory ECG records have already been obtained from infants (25 days–3months of age), children (3–9 years of age), boys and girls (10–14 years of age), pubertal boys and girls (15–20 years of age), adults of both genders (2–92 years of age), and pregnant women in Japan. In addition, 7-day or longer ECG records were obtained from clinically healthy adults in different geographic locations. Similar records from patients with coronary artery disease and with dilated cardiomyopathy have already found marked differences in their time structure [65, 75, 76].

The influence of space weather on blood pressure (BP) was noticed in a record of weekly measurements taken mostly on Sunday mornings by a 71-year-old psychiatrist [77]. His data were characterized by a large-amplitude circannual variation peaking in the summer. Although some studies reported higher BP values in summer than in winter in elderly populations [78], BP is usually higher in winter than in summer [79–81]. It seems likely, however, that circannual changes in BP are partly endogenous, even if they are statistically significantly correlated with weather and temperature [82, 83] or wind [79].

Circannual changes characterize not only BP but also a variety of hormones and their receptors [84–87], electrophysiological and other cardiovascular variables, as well as morbidity and mortality statistics [88–93]. For instance, changes in ventricular cardiac function, both during systole and diastole, exhibit a circannual rhythm in clinically healthy normotensive subjects [84]. The renin-angiotensin system may underlie circannual changes in BP in view of the fact that a very prominent circannual variation characterizes the response of renin release to calcium deprivation in isolated rat glomeruli [85, 86]. The surge in late spring coincides with a time of changes in the human circannual rhythm of circulating aldosterone [85]. A circannual rhythm in circulating aldosterone is a major classifier of risk of developing high BP and/or related cardiovascular disease [85]. Noteworthy is the observation that the circannual MESOR of circulating aldosterone is higher in women at low familial and personal cardiovascular disease risk, as compared to women at high risk [85]. This result is contrary to the view that as an aldosterone-secreting tumor of the adrenal cortex is associated with a high BP, a high cardiovascular disease risk should also be associated with elevated circulating aldosterone [77, 94]. The chronobiologic result in health is in keeping with the report of higher plasma aldosterone concentrations in children of parents with vs. without hypertension [95].

Melatonin may also be involved in the coordination of BP as an endogenous central hypotensive factor [96]. Melatonin administration in the evening has been associated with a decrease in BP [97]. Urinary melatonin excretion was also positively correlated with cardiovascular disease risk assessed by questionnaire [98]. Melatonin has been implicated in a variety of cardiovascular pathophysiological processes, including anti-inflammatory, antioxidant, antihypertensive, and antilipidemic functions [99]. In the circulation, the concentration of melatonin follows a circadian rhythm, with high values at night, thus providing time clues to target tissues endowed with melatonin receptors [99]. It has indeed been suggested that cardiovascular effects of melatonin are at least partly mediated via MT1 and MT2 receptors [100]. It was found that MT1 was expressed in healthy and diseased human coronary arteries and that MT2 was present in human arteries and left ventricles, its expression being altered in coronary heart disease [101]. Reportedly, melatonin is also able to synchronize neuronal firing in the suprachiasmatic nuclei (SCN) via membrane receptors, and clock gene expression is apparently sensitive to melatonin in the pars tuberalis [102].

The pineal being a sensitive receptor that can respond to variations of the order of a few nanoteslas (nT) induced by space weather [103–109] may also be involved indirectly in relation to BP changes and their related cardiovascular complications. In Alta, Norway (70° N), changes in geomagnetic activity above 80 nT/3 hours were associated with a reduced circadian amplitude of salivary melatonin [109]. This result is in keeping with a reduction in a melatonin metabolite excretion in the presence of increased geomagnetic activity, which was greater when the ambient light exposure was also reduced [108]. In this study as well, decreased melatonin secretion was observed when geomagnetic activity exceeded a threshold value of 30 nT. Further evidence in the experimental laboratory supports an effect of magnetic storms on heart function. Indeed, the ultrastructure of cardiomyocytes of rabbits was found to be drastically altered at a time coinciding with strong magnetic storms by comparison with that usually observed during quiet geomagnetic conditions [110].

In addition to the circannual rhythm in BP of the 71-year-old psychiatrist, an about-11-year variation similar to the solar activity cycle was resolved nonlinearly in his 26-year record spanning from 1969 to 1994 [111]. The fit of a model containing both a solar cycle and a 1-year cosine curve was statistically significant for solar cycles 20 (11.6 years), 21 (10.3 years), and 22 (7 years) covered by his record. The solar cycle component was statistically significant during each solar cycle for diastolic BP and during the last two solar cycles for systolic BP, acrophases occurring consistently approximately 2 years after maximal solar activity [111]. These results were compared to about-11-year and about-1-year changes in BP already reported in two other cases by the mid-1990s, one a clinically healthy man, the other an elderly man treated for MESOR-hypertension [111].

Since then, the record of self-measurements from the clinically healthy man was extended to span over four decades, and longitudinal time series from six other subjects covering 10 years or longer became available for analysis. These seven biomedical scientists included five physicians, four of them treated for high BP. While two of the subjects provided self-measurements, around-the-clock measurements were obtained by ambulatory monitoring from the other five researchers. One more subject contributed data on urinary excretion of 17-ketosteroids for 15 years [112, 113]. Nonlinear analyses of the original data and of the circadian (or in one case, circaseptan) MESORs and amplitudes revealed the presence of cycles with periods clustering around 5 and 10 years [114] (Fig. 13.9). These results strongly suggest that human physiology may be influenced by nonphotic environmental cycles related to solar activity.

Additional evidence stems from the detection in longitudinal BP and HR records of other nonphotic cycles featured in terrestrial-space weather. One such cycle is the transyear detected in solar wind speed [115], which has a period of about 1.3 years. As nonphotic cycles are wobbly, their periods are not stable [116]. While transyears can have periods in a relatively wise range, they are statistically significantly longer than one year. An about-1.3-year transyear was invariably detected in all 43 time series of BP and HR from 13 subjects [117, 118] (Fig. 13.10). When a gliding spectrum was obtained for the SBP record of an elderly man and compared to that of solar wind speed, similarities in the time course of the transyear could be followed between these two variables [119]. Results showed that the transyear in solar wind speed was expressed prominently between 1989 and 1996, after which it splits into two separate spectral bands until 1997 when the signal fades away and is no longer detectable. In SBP, the transyear remains prominently expressed until 1998, after which its amplitude is decreased but remains detectable until the end of the record in 2005, suggesting that this component, while still resonating with its environment, may also be partly genetically anchored [119].

Transyears have also been documented in transverse data series from neonates, where they are more prominent than the circannual component [120]. Whereas the about-yearly rhythm predominates over transyears in adulthood on the average, transyears regain prominence in the elderly, as gauged by amplitude ratios of these two cycles [120]. A similar time course as a function of age was also found for the about-weekly component in BP and HR [121–123]. The circaseptan period of neonatal BP and HR measured around the clock for the first few weeks after birth, estimated by nonlinear least squares, correlated positively with the corresponding period of the local index of geomagnetic disturbance (K) analyzed during spans matching each neonatal record [124]. These hybrid neonatal data, obtained around the clock for several days from 32 babies monitored during the first month of life, also showed cross-spectral coherence (>0.70) with K at periods of about 3.3 and 8.8 days [24].

Cross-spectral coherence was also documented in a baby boy born prematurely and monitored longitudinally between Jan 26, 1989 and Jan 22, 1990. His systolic BP was coherent with Bz, the vertical component of the interplanetary magnetic field, at a period of about 5 days, and his diastolic BP was coherent with Bz at a period of about 3.5 days [125] (Fig. 13.11). In adulthood as well, associations between BP and space weather were observed. In a 3-year record (1988–1990) of around-the-clock BP measurements obtained by ambulatory monitoring from a clinically healthy man, cross-spectral coherence with the planetary index of geomagnetic disturbance (Kp) was statistically significant for systolic and diastolic BP at periods of about 27.7 days (SBP: 0.567, P = 0.010; DBP: 0.516, P = 0.025) and 4.2 days [126] (Fig. 13.12). Associations of the BP and HR of another clinically healthy man with Kp and of HR of a 28-year-old woman who spent 267 days in social isolation with Kp and with cosmic rays have also been found [63].

Effects of space weather were also observed in the experimental laboratory. The circadian rhythm of systolic and mean intracardiac pressure measured around the clock in the left and right ventricles of male Chinchilla rabbits was found to be markedly altered in September 1984 when the study coincided with the occurrence of two strong geomagnetic storms, as compared to three other studies carried out between June 1984 and March 1985, when quiet geomagnetic conditions prevailed [110]. Considering the involvement of melatonin in relation to blood pressure, it may also be pertinent that in rats sampled around the clock for 7 days, alterations in the circadian rhythm of pineal and hypothalamic melatonin were found following the second extremum of a moderate double magnetic storm [127]. As compared to quiet conditions, during the storm, the MESOR and circadian amplitude were reduced in the case of pineal melatonin and increased in the case of hypothalamic melatonin [127].

Already by 1922, there was evidence for an adverse effect of space weather on human health. Two physicians, Sardou and Faure, and Vallot, an astronomer, meteorologist, and generous supporter of research, then published their finding that severe and fatal illnesses as well as common head cold and other aches were more frequent when there were many sunspots [128]. Whether all or only severe symptoms were considered, a higher incidence when sunspots were present than absent can be validated statistically. Clinical symptoms considered diseases of the heart, blood vessels, liver, kidney, and nervous system, ranging from mild to severe, including excitability, insomnia, tiredness, aches, muscle twitches, polyuria, digestive troubles, jitteriness, shivering, spasms, neuralgia, neural crises, asthma, dyspnea, fever, pain, vertigo, syncope, high blood pressure, tachycardia, arrhythmia, and true angina pectoris [128].

Much credit is also due to Bernhard and Gertraud (“Traute”) Düll for their pioneering work in documenting about-27-day cycles in human mortality and relating these patterns to space-terrestrial weather [129–135]. Using superposed epoch analysis, they showed how space/terrestrial weather may differentially influence human mortality from various causes, the 27-day mortality pattern being different whether death was from cardiac or respiratory disease or from suicide. Specifically, differences in the phase of a common 27-day cycle characterized these mortality patterns, determined on the basis of thoroughly assembled death certificates covering 68 stacked consecutive 27-day solar rotations from January 1928 to December 1932 (5 years) [136].

An association between the incidence of various diseases, myocardial infarctions and strokes in particular, and geomagnetic storms and other electromagnetic variations has long been claimed by scientists in the former Soviet Union [137–139]. As reviewed by Halberg et al. [140], results of direct experiments in which biological specimens are exposed to artificially induced electromagnetic fields support studies of correlation between geomagnetic activity and human morbidity and mortality [141]. Electromagnetic fields have been reported to affect blood pressure [142], baroreflex function [143], and HRV [144]. Effects of geomagnetic disturbances have also been reported on the capillary blood flow of patients with ischemic heart disease [145].

Such effects upon the cardiovascular system may contribute to observed associations between space/terrestrial weather and the incidence of adverse events, such as myocardial infarction. By superposed epochs, a consistent association of excessive geomagnetic activity with cardiovascular disease and myocardial infarctions particularly was documented in data from the 1960s to 1970s in two geographic locations [146]. Despite other reports of increased incidence of heart attacks in the presence of increased geomagnetic disturbances [147–155], not all studies found such associations [156, 157]. In particular, two large studies in the USA, carried out with impeccable inferential statistics, reported no solar effect in their investigation of associations of solar activity and magnetic disturbances with mortality [158, 159]. While there is no good explanation thus far for the controversy, mostly because underlying mechanisms are not well understood, several possible contributing factors have been proposed. They include the usually higher level of artificial magnetic disturbances from industry and domestic appliances in the USA and the fact that mortality from myocardial infarction does not necessarily coincide with the onset of symptoms, should they have been triggered by the magnetic storm [147]. Not following a sequence of several Horrebow-Schwabe (about 11-year) cycles and studying their variability may be another reason [1]. Lipa et al. [159] cautiously considered the possibility that any causal relationship between geomagnetic disturbances and coronary heart disease may be sensitive to geographic location, to the phase of the solar cycle, or to some other parameter which might distinguish their data from those of Gnevyshev and Novikova [141]. Distinguishing between mortality and morbidity data has also been recommended, as an about 10.5 % increase in mortality from myocardial infarction in association with big geomagnetic storms contrasted with a lack of effect on the incidence of morbid events in one study [152].

In a database of over 6.3 million cases requiring the call for an ambulance during the span from Jan 1, 1979 to Dec 31, 1981 in Moscow, Russia, including 85,819 diagnoses of myocardial infarction, cross-spectral coherence was found in association with the local index K of geomagnetic disturbance and with Bz, the north–south component of the interplanetary magnetic field at the same frequency of one cycle in about 3.15 days [125, 160] (Fig. 13.13). A north-to-south Bz turn triggers aurorae in the upper atmosphere and storms in the magnetosphere, the total current in the magnetosphere increasing up to ten times within a few minutes [161]. Defining a storm as a north-to-south daily mean changing from ≥1nT to < −1.5nT, there were 32 events during the 3-year span. Superposed epochs analysis indicated that the daily incidence of myocardial infarctions increased by 7.6 % after a Bz turn (P = 0.027), followed, however, by a 5.9 % decrease on the following day (P = 0.029) (Fig. 13.14). This result was consistently observed during each year considered separately [162]. Using a different method, the result was independently corroborated on the same data set [150], an even larger effect being documented in relation to two other criteria (antipodal index aa > 60 and Forbush decrease in cosmic ray intensity >1.5 %), with additional evidence stemming from another set of data [151].

In another database of 129,205 deaths from myocardial infarction in Minnesota during the span from 1968 to 1996 (29 years), an about-10.5-year component was detected with statistical significance after removal of a linearly decreasing trend (Fig. 13.15). During years of maximal solar activity, there was an excess of 220 deaths per year (about 5 %) as compared to years of minimal solar activity (P = 0.023) [163]. When the data are analyzed separately for each of the three solar cycles, statistically significant differences are found, however, perhaps providing some answer for the different results reported above [140]. Fitting a 10.5-year cosine curve to the detrended data, the model is statistically significant for the spans from 1968 to 1977 (P = 0.006) and from 1978 to 1986 (P < 0.001), but not for the span from 1987 to 1996 (P > 0.50). Whereas acrophases are very similar, differing by only 10° (2.8 % or 3.5 months), amplitudes vary from 131 and 278 deaths/year during the first two cycles to 25 deaths/year during the third cycle. The about-10.5-year cycle can be resolved nonlinearly overall and for the first two cycles, but not for the third, the period and its 95 % confidence interval being estimated as 10.83 [10.00, 11.66] years overall and as 11.48 [3.23, 19.72] and 10.64 [8.60, 13.28] years during the first two cycles, respectively [163]. As seen from Fig. 13.15, mortality from myocardial infarction in Minnesota is also characterized not only by a prominent annual variation, but also by a transyear and a cis-half year with periods of about 1.3 years and 0.42 year. These nonphotic cycles are resolved nonlinearly. They reflect signatures found in solar wind speed [115] and in the solar flare index [164], respectively.

Other diagnoses in the 1979–1981 Moscow database included strokes, hypertensive crises, arrhythmia, and sudden cardiac death, among others [125]. In contrast to the other cardiovascular morbidity statistics, all showing a very prominent weekly pattern, the circaseptan component characterizing sudden cardiac death, although statistically significant, was much less pronounced. Instead, an about-half-monthly component was detected with a period of about 2 weeks in least squares and population-mean cosinor spectra (Fig. 13.16). Nonlinearly, its period is estimated as 15.2 [95 % CI: 15.15–15.30] days. According to the criterion of a nonoverlap of 95 % confidence intervals, this period is statistically significantly different from the half-monthly period characterizing local K analyzed during the same 3-year span. For K, the corresponding period is estimated to be 14.0 [95 % CI: 13.94–14.17] days. Whereas the about-half-monthly period of both sudden cardiac death and local K is estimated with a relatively tight 95 % confidence interval when analyzed over the entire 3-year span, this component’s period shows great wobbliness when the data are analyzed by moving spectra over a shorter interval of 3 months, progressively displaced by 1.5 weeks throughout the time series. This component is not consistently detected either for sudden cardiac death or for local K (Fig. 13.16). A possible resonance with occasional frequency trapping between the two variables at this frequency is suggested by the fact that the half-monthly component is seemingly more prominently expressed in sudden cardiac death when it is detected in the spectrum of K [37, 165, 166].

Since the tenth revision of the International Classification of Diseases (ICD10), sudden cardiac death (code I46.1) has been separated from cardiac arrest associated with a myocardial infarction and is defined as unexpected death from cardiac causes within 1–24 hours of the onset of acute symptoms. Daily incidences of sudden cardiac death (ICD10, code I46.1, excluding death from myocardial infarction) in Minnesota during 1999–2003 (5 years) were compared to those of daily incidences of mortality from myocardial infarction during the same span in Minnesota and with those of daily (or monthly) incidences of sudden cardiac death in some other geographic locations.