Do geomagnetic disturbances of solar origin affect arterial ... - Nature

Journal of Human Hypertension (1998) 12, 749–754  1998 Stockton Press. All rights reserved 0950-9240/98 $12.00 http://www.stockton-press.co.uk/jhh

ORIGINAL ARTICLE

Do geomagnetic disturbances of solar origin affect arterial blood pressure? S Ghione1, L Mezzasalma1, C Del Seppia1 and F Papi2 1

CNR Institute of Clinical Physiology, Pisa; 2Department of Ethology, Ecology and Evolution, University of Pisa, Pisa, Italy

Objective: Episodic reports suggest that geomagnetic disturbances of solar origin are associated with biological and clinical events, including increased arterial blood pressure (BP). We reassessed this aspect by relating solar activity levels to ambulatory BP measured in our out-patient population. Patients and methods: The ambulatory BP measurements of 447 consecutive untreated patients attending a hypertension out-patient clinic who did a monitoring for diagnostic purposes over 5 years were retrieved. The mean daytime, night-time and 24-h BP and heart rate values were related to the temporally corresponding geomagnetic index k-sum obtained by the nearest observatory. K-sum is a local measurement of the irregular disturbances of the geomagnetic field caused by solar particle radiation. Results: Significant to highly significant positive corre-

lations were observed for k-sum with systolic (daytime and 24 h) and diastolic BP (daytime, night-time and 24 h), but not with heart rate. No correlations were found with the k-sum of 1 or 2 days before the monitorings. Multiple correlations which also included other potential confounding factors (date, age) confirmed a significant effect of k-sum on BP. Comparison made in seasonmatched subgroups of quiet and disturbed days (using three different criteria of definition), always showed significantly higher values in the disturbed days for all BP parameters except systolic night-time pressure. The difference between the quietest and the most disturbed days was of about 6 to 8 mm Hg for 24-h systolic and diastolic BP. Conclusion: These results are unlikely to be due to unrelated secular trends, but seem to reflect a real relation between magnetic field disturbances and BP.

Keywords: arterial blood pressure; geomagnetic activity; arterial blood pressure monitoring

Introduction The possibility that geomagnetic activity influences biological and clinical phenomena has been the object of curiosity on the part of the general public and, although with some scepticism, also on the part of scientists.1 Solar flares and other solar activity can generate so-called magnetic storms, which, through modulation of the solar wind, can cause magnetic anomalies in the Earth’s environment.2 Information on these geomagnetic disturbances can be obtained from various geomagnetic observatories around the world. Several biological and clinical phenomena have been reported to be associated with geomagnetic disturbances3–7 but the association between clinical events and geomagnetic disturbances is far from clear, as may be shown for the case of heart attacks. Strongly significant positive correlations between the daily levels of geomagnetic activity indices and daily admissions of cardiac emergency cases were observed in a retrospective study conducted in India8 but a similar enquiry carried out in England was unable to show any such correlation.9 In reviewing the literature we were struck by the brief report of Stoupel et al10 that blood pressure Correspondence: Dr Sergio Ghione, Institute of Clinical Physiology, Via Savi 8, 56126 Pisa, Italy Received 25 June 1998; revised and accepted 6 August 1998

(BP) values obtained in 39 patients with hypertension by means of ambulatory monitoring carried out on days with high geomagnetic activity were significantly higher than values obtained in 35 other patients on days with low geomagnetic activity. The aim of our paper is to reassess this problem in a larger untreated out-patient population.

Subjects and methods Subjects The data of all 1310 ambulatory BP monitorings carried out over 5 years (1 January 1992 to 31 December 1996) in our out-patient clinic for diagnostic purposes were electronically retrieved. Of these 220 were excluded because of insufficient duration (less than 12 hours) or insufficient number of recordings (less than 30), or age less than 20 years. Thirty-nine monitorings, done in July 1996 were also excluded because geomagnetic data were not available for this month. Of the remaining 1051 monitorings, 468 were done in 447 subjects (181 females and 266 males) without any hypotensive treatment and are the object of the present report. The average age of the subjects was 47.9 ± 13.7 years (mean ± s.d.; range 21–85 years).

Solar geomagnetic disturbances and BP S Ghione et al

750

Geomagnetic data Geomagnetic data over the same 5 years were kindly provided by the closest geomagnetic observatory in L’Aquila, Central Italy (287 km from Pisa). The geomagnetic parameter used was the k-index which is commonly employed for estimating the solar magnetic activity and is based on the local measurement of the irregular disturbances of the geomagnetic field caused by solar particle radiation.11 It is determined by measuring the range (difference between the highest and lowest value) for the most disturbed horizontal magnetic field component during 3hourly time intervals. We used three indexes obtained from the k-index tables: (1) the sum of the 3-hourly k-indexes over 24 h (24-h k-sum), starting from 06.00 to 06.00 Greenwich time (GT) of the following day; (2) the sum of the k-indexes from 06.00 to 21.00 GT (defined as daytime k-sum); and (3) the sum of the k-indexes from 21.00 to 06.00 GT (defined as night-time k-sum). Monitoring data Two Spacelab recorders were used. Each recorder was checked every 3 months by comparison with manual readings. The mean duration of monitoring and the mean number of recordings were 24.2 ± 2.0 hours and 68.2 ± 10.3 measurements. The recorders were applied in the morning of working days, between 07.30 and 09.00 and programmed to perform a BP measurement every 15 min from 07.00 to 22.00 and every 30 min from 22.00 to 07.00. All subjects returned to their usual activities. The data analysed were: mean systolic and diastolic BP and heart rate over 24 h, during daytime (from 07.00 to 22.00) and during night-time (from 22.00 to 07.00). Heart rate values were not available for the first year. Statistical analysis The data obtained by ambulatory monitoring were correlated with k-sum values for the same time interval (24 h, daytime and night-time) as outlined above. Correlations were also determined with the 24-h k-sum of 1 and 2 days before the individual monitorings. In order to perform the multiple regressions, simple regressions were first analysed for the ambulatory data against time (the date of monitoring), sex, age, body weight and environmental temperature. Maximal and minimal temperature were obtained from the metereological station of the airport at Pisa, and the mean value was estimated as the average of the two. Only those variables for which a statistical significant correlation was found, were used as components of multiple regression. Data were also analysed in subgroups of monitorings done in days of quiet and disturbed geomagnetic activity and matched for season with a procedure which requires a brief description. For the days during which monitorings were available, the frequency distribution of daily k-sum values was constructed and its percentile limits determined. Comparisons were made between monitoring data obtained during quiet and disturbed days defined by

three different criteria: (1) days with 24-h k-sum value below the median against those above the median; (2) days with 24-h k-sum value below the 25th percentile against those above the 75th percentile; and (3) days with 24-h k-sum value below the 10th percentile against those above the 90th percentile. In order to compare groups of similar days, for each comparison the data were automatically matched with a computer program. In order to maximise the comparison, the data were sorted by the kvalue and matched starting from the highest available k-sum value (in the higher percentile group) and finding (in the lower percentile group) the lowest available k-sum value for which the date of recordings was done in the same day (±5 days), but not necessarily in the same year. The groups were then compared by an unpaired t-test. Statistical analysis was done on an Apple computer with a statistical package (Statview, Abacus Concepts; Berkeley, CA, USA).

Results The annual mean values of 24-h ambulatory BP and heart rate and of geomagnetic activity are reported in Table 1. K-sum correlated highly significantly with 24-h, daytime and night-time diastolic BP and with 24-h and daytime systolic BP (Table 2). For heart rate no significant correlation was observed. No significant correlation was observed when the monitoring values were correlated with the k-sum values of 1 and 2 days before. A graphical representation of the relationship between k-sum and ambulatory values is shown in Figure 1 which depicts as line charts the average values of 24-h systolic and diastolic BP and heart rate for classes of 24-h k-sum of increasing value. The corresponding numerical data are reported in Table 3. When ambulatory BP data were correlated with other potential confounding factors, significant inverse correlations were observed for all BP parameters with time. In addition, all heart rate parameters correlated inversely with age. Finally, a weak inverse correlation was found for age with 24 h and daytime diastolic BP. No significant correlations were observed with body weight and environmental temperature. When monitoring data were related in a multiple regression model with those variables for which a statistical significance was found and with the corresponding k-sum, statistically significant partial regression coefficients with k-sum were found for diastolic BP (24-h: t = 2.664, P = 0.008; daytime: t = 2.515, P = 0.012; night-time: t = 2.353, P = 0.019). For systolic BP, the partial regression coefficient with k-sum was significant for daytime values (t = 2.057, P = 0.040), fell short of statistical significance for 24-h values (t = 1.880, P = 0.059) and was not significant for night-time values. For heart rate, the partial regression coefficients with k-sum were never significant. In order to exclude a potential effect of the season, comparisons were made, as outlined in the Methods section, between ambulatory BP data measured in quiet and disturbed days. As reported in Table 4, in all comparisons, significantly higher values were


Table 1 Mean values of 24 h k-sum index in the years 1992–1996 and of ambulatory monitoring data available for the same period Year

K-sum

K-sum*

No. recordings

24 h SBP

24 h DBP

24 h HR

1992 1993 1994 1995 1996

17.9 ± 7.4 16.9 ± 7.3 18.3 ± 7.9 16.4 ± 7.2 14.8 ± 6.4

16.6 ± 6.9 17.3 ± 7.4 19.3 ± 7.4 18.6 ± 7.5 14.8 ± 6.4

76 97 81 84 130

134.5 ± 12.8 133.2 ± 13.0 130.4 ± 12.1 129.9 ± 11.9 126.2 ± 12.6

85.7 ± 8.9 85.9 ± 9.9 83.7 ± 8.4 83.4 ± 10.0 80.3 ± 10.5

n.a. 75.9 ± 10.6 77.4 ± 8.2 75.3 ± 9.5 75.6 ± 10.3

K-sum refers to the mean value ± s.d. of all the days of the year and k-sum* to the mean value ± s.d. of the days when a monitoring was done. SBP = systolic blood pressure (mm Hg). DBP = diastolic blood pressure (mm Hg); HR = heart rate (bpm); n.a. = not available. Table 2 Simple regressions of ambulatory BP and heart rate data with k-sum both averaged over the same time-interval Correlations

Time-interval

Intercept

Slope

R

P-value

Systolic BP

24 hours daytime night-time

127.1 130.7 119.0

0.193 0.284 0.310

0.109 0.106 0.067

0.019 0.022 NS

Diastolic BP


80.0 83.7 71.2

0.205 0.300 0.428

0.150 0.144 0.120

0.001 0.002 0.001

Heart rate


74.8 78.3 65.6

0.070 0.085 0.250

0.053 0.041 0.080

NS NS NS

BP = blood pressure NS = not statistically significant.

observed in the disturbed days for all BP values, except for night-time systolic pressure. For heart rate, the values during disturbed days were always higher than during quiet days, but this difference attained statistical significance only in three comparisons. The difference of the ambulatory values between the groups was increasingly greater the more extreme was the comparison.

Discussion This study provides evidence that arterial BP tends to be higher on days with more disturbed geomagnetic activity compared with quieter days, confirming the results of a similar study done in a smaller group of subjects by Stoupel et al.10 Several aspects of our results sustain, in our opinion, this conclusion. Since measurements on days with different geomagnetic activity were done in different subjects, a substantial part of the variability is expected to be due to inter-individual differences. These would have made conditions worse for the detection of an environmental effect on BP, such as that of disturbed geomagnetic activity. Despite this, the results pointing to a relation between k-sum as a measure of geomagnetic disturbance and BP were consistent when analysed in different ways. Almost all relations were still significant when other potential confounding factors were taken into account, including season, which is related to geomagnetic activity (because of the varying distance of the earth from the sun in the different seasons)8 and has been reported to influence BP.12 Inspection of Figure 1 and calculations from the estimated regression slopes are consistent in indicating that, for extreme values of geomagnetic disturbances a difference of about 6 to 8 mm Hg in systolic

and diastolic BPs may be expected. It may be interesting to note that these effects are of the same order of magnitude as those reported for the influence of other individual, lifestyle and environmental factors on BP such as sodium consumption,13 body weight,14 alcohol consumption15 and external temperature.16 Limitations of the study We are aware that, when relations are analysed in a retrospective fashion between data collected over a prolonged period of time, the findings obtained are exposed to potential flaws of various origin. A first problem is the risk that the results are due to spurious associations, because of unrelated secular trends of the variables under examination. Although independent studies are needed to exclude such a possibility, the consistency of the results when analysed in different ways lead us to believe that the association found between BP and k-sum may reflect a real effect of geomagnetic disturbances on BP. In fact firstly the correlations of k-sum with BP were still significant when data were analysed in a multiple regression model which took also the time (date) into account. Secondly, if the relation found were due to the effect of a secular trend of BP, one would expect similar relations between BPs and the k-sum of 1 or 2 days before the day of measurements. The fact that a relation was found with k-sum for all ambulatory BPs (except for night-time systolic pressure) when data were analysed for the corresponding time interval and no relation whatsoever with the k-sum of 1 or 2 days before, strongly suggests that the relations found were not due to a secular trend. Thirdly, in the 5 year period examined ksum tended to increase over the first 3 years and

751


752

not been raised by previous retrospective studies such as the present one and, probably, if it were not confirmed by other independent retrospective studies.

Possible interpretations What could be the reasons for abnormalities of the geomagnetic field affecting arterial BP? A possible key of interpretation may be found from studies which suggest that naturally occurring geomagnetic disturbances and/or artificially altered magnetic fields may affect the response to environmental stress in animal and man. Under this context, an aspect particularly studied was pain perception. A significant increase of pain threshold was in fact reported in mice during a geomagnetic storm4 and exposure to artificial magnetic fields has been shown by others and ourselves in several species, including man, to increase nociception probably by suppressing stress-induced hypoalgesia.17–21 On the other hand, exposure to abnormal magnetic fields in the mouse has been reported to increase exploratory mobility, which is thought to be an expression of anxiety.17 In pigeons, a similar exposure disturbs the initial orientation of the animal, probably as a consequence of an abnormal emotional state (see references in Luschi et al).22 In fact, similar behaviour of disturbed initial orientation is observed after stressful exposure, such as light deprivation or immobilisation,23 and pretreatment with the tranquilliser promazine abolishes the effects of both magnetic and stressful treatment.22 In addition, evidence has emerged that magnetic treatment may affect endogenous opioid function.24 Taken together these observations are consistent with the idea that abnormalities of the magnetic field may in some way, possibly through endorphinergic pathways, interfere with normal mechanisms activated in the organism in situations of stress, increasing perceived stress. The observations that not only BP but in some analyses also heart rate is significantly increased in disturbed compared to quiet days may be in line with this idea. We cautiously propose that increased average BP in days of high geomagnetic disturbance may reflect a slight increase in the ‘vulnerability’ of subjects to everyday life stress due to an impaired coping capacity provoked by abnormal magnetic fields. Figure 1 Average 24-h systolic (SBP) and diastolic BP (DBP) and heart rate (HR) plotted in classes of increasing k-sum value (for further details see Table 3).

decrease in the following 2 (Table 1), whereas BP tended to decrease over the whole period. Therefore in the first 3 years k-sum and BP had a divergent trend; however, when analysis was limited to these years, positive significant relations were still found between k-sum and BP (data not shown). Another limitation of this study is its retrospective nature. A prospective study is certainly needed to provide an established validation. However it is difficult to imagine that a prospective study in a field like this would be planned if an interest had

Acknowledgements We would like to thank Prof. P Palangio (National Geophysics Institute of L’Aquila) for supplying us the geomagnetic data, Dr S Degl’Innocenti (Department of Physics, University of Pisa), Dr M Velli (Astrophysics Observatory of Arcetri of Florence) and Dr M Messerotti (Astronomical Observatory of Trieste) for helpful discussions and suggestions, Mr G Ceccanti (Institute of Clinical Physiology, CNR) for his excellent technical assistance and Ms A Ciurli (Institute of Clinical Physiology, CNR) for secretarial support.


753

Table 3 Mean and s.e.m. values of 24 h ambulatory BP and heart rate in grouping intervals with increasing 24 h k-sum value K-sum

24-h SBP

24-h DBP

24-h HR

Range

Counts

Mean

s.e.m.

Mean

s.e.m.

Mean

s.e.m.

Counts

Mean

s.e.m.

1–8 9–12 13–16 17–20 21–24 25–28 29–32 33– 40

41 110 100 69 64 44 30 10

6.63 10.43 14.52 18.65 22.33 26.00 30.33 34.80

0.26 0.10 0.11 0.13 0.13 0.18 0.20 0.70

129.83 128.16 130.30 129.93 131.63 130.91 135.40 136.70

2.19 1.19 1.29 1.38 1.67 1.84 2.63 3.37

82.10 81.23 83.99 83.74 83.63 84.66 87.13 90.00

1.51 0.91 0.98 1.02 1.41 1.46 1.92 3.01

33 90 79 49 47 40 28 9

74.52 74.27 77.86 77.71 74.28 76.05 78.36 75.67

1.36 0.93 1.15 1.64 1.38 1.53 2.17 1.98

SBP = systolic blood pressure (mm Hg); DBP = diastolic blood pressure (mm Hg); HR = heart rate (bpm). The table neports in a more detailed way the data shown in Figure 1. The column ‘Range’ represents the limits of the classes of the 24 h k-sum distribution, ‘Counts’ the frequency in each class. ‘Mean’ and ‘s.e.m.’ are the mean values and standard error of the means found in the classes. Each class has a grouping interval of 4 Units of 24 h k-sum except for the first two and the last two which had to be grouped together because of too small numbers of data (class 0– 4 and class 36– 40 had only two counts each). The counts for heart rate were smaller because heart rate data were not available for 1992. Table 4 Comparison between ambulatory BP and heart rate data obtained in days (matched for the day ±5, but not necessarily for the year) above and below and certain percentile (see text for further details) Lower vs upper half percentiles Quiet

Disturbed

139 154 10.2 ± 2.8

139 158 24.6 ± 5.0

24-h SBP Daytime SBP Night-time SBP

129.0 ± 13.5 132.2 ± 13.7 119.9 ± 14.6

132.1 ± 12.8 135.5 ± 13.1 122.6 ± 14.5

24-h DBP Daytime DBP Night-time DBP

82.0 ± 9.4 85.3 ± 9.6 72.6 ± 10.9

24-h HR Daytime HR Night-time HR

75.3 ± 9.2 78.2 ± 9.7 66.6 ± 9.5

No. of days No. of monitorings k-sum

25th vs 75th percentiles

Pvalue

Quiet

Disturbed

64 71 8.3 ± 2.0

64 76 27.9 ± 4.2

0.034 0.029 NS

128.7 ± 12.7 131.8 ± 12.8 120.0 ± 14.3

134.3 ± 12.9 138.0 ± 13.1 123.8 ± 14.4

85.1 ± 10.0 88.6 ± 10.4 75.4 ± 11.0

0.006 0.005 0.025

81.3 ± 9.7 84.5 ± 9.8 72.0 ± 11.8

76.6 ± 10.0 79.8 ± 10.7 67.9 ± 9.6

NS NS NS

74.6 ± 8.5 77.4 ± 8.8 66.3 ± 9.5

10th vs 90th percentiles Pvalue

Quiet

Disturbed

Pvalue

22 25 6.8 ± 1.2

22 25 31.3 ± 3.3

0.008 0.004 NS

128.4 ± 13.3 131.6 ± 13.5 119.7 ± 14.9

135.3 ± 11.0 139.2 ± 11.6 123.9 ± 12.2

0.050 0.038 NS

87.1 ± 10.2 90.7 ± 10.8 77.1 ± 10.9

⬍0.001 ⬍0.001 0.008

81.3 ± 10.3 84.8 ± 10.7 71.9 ± 10.8

88.8 ± 9.6 92.1 ± 10.4 79.1 ± 10.5

0.011 0.018 0.021

78.2 ± 10.5 81.3 ± 11.0 69.5 ± 9.9

0.043 0.030 NS

75.1 ± 8.0 78.7 ± 7.9 65.5 ± 8.5

80.3 ± 10.5 83.2 ± 10.4 71.5 ± 11.2

NS NS 0.048

Values are mean ± s.d. SBP = systolic blood pressure (mm Hg). DBP = diastolic blood pressure (mm Hg), HR = heart rate (bpm).

References 1 Macklis R. Magnetic healing, quackery and the debate about health effects of electromagnetic fields. Ann Int Med 1993; 118: 376–383. 2 Bruzek, Durrant. Illustrated glossary for solar and solar-terrestrial physics. D. Reidel Publishing Co, Amsterdam, 1977. 3 Persinger MA. Day time wheel running activity in laboratory rats following the geomagnetic event of 5–6 July, 1974, Int J Biometeorol 1974; 20: 19–22. 4 Ossenkopp KP, Kavaliers M, Hirst M. Reduced nocturnal morphine analgesia in mice following a geomagnetic disturbance. Neurosci Lett 1983; 40: 321–325. 5 Keshavan MS et al. Convulsive threshold in humans and rats and magnetic field changes: observation during total solar eclipse. Neurosci Lett 1981; 22: 205–208. 6 Rajaram M, Mitra S. Correlation between convulsive seizure and geomagnetic activity. Neurosci Lett 1981; 24: 187–191. 7 Ronald W. Geomagnetic storms: association with incidence of depression as measured by hospital admission. Br J Psychiat 1994; 164: 403– 409. 8 Malin SRC, Srivastava BJ. Correlation between heart attacks and magnetic activity. Nature 1979; 277: 646–648.

9 Knox EG, Armstrong E, Lancashire R. Heart attacks and geomagnetic activity. Nature 1979; 281: 564 –565. 10 Stoupel E, Wittenberg C, Zabludowski J, Boner G. Ambulatory blood pressure monitoring in patients with hypertension on days of high and low geomagnetic activity. J Hum Hypertens 1995; 9: 293–294. 11 Menvielle M, Berthelier A. The k-derived planetary indeces: description and availability. Rev Geophys 1991; 29: 415– 432. 12 Giaconi S et al. Seasonal influences on blood pressure in high normal to mild hypertensive range. Hypertension 1989; 14: 22–27. 13 Intersalt Cooperative Research Group. Intersalt: an international study of electrolyte excretion and blood pressure. Results for 24 hour urinary sodium and potassium excretion. Br Med J 1988; 297: 319–328. 14 Kaplan N. Clinical hypertension. Williams & Wilkins: Dallas, Texas, 1990. 15 Klatsky AL, Friedman GD, Siegelaub AB, Ge´rard MJ. Alcohol consumption and blood pressure Kaiser-Permanente Multiphasic Health Examination data. N Engl J Med 1977; 296: 194 –200. 16 Brennan P, Greenberg G, Miall W, Thompson S. Seasonal variation in arterial blood pressure. Br Med J 1982; 285: 919–923. 17 Kavaliers M, Ossenkopp KP. Stress-induced opioid


754

analgesia and activity in mice: inhibitory influences of exposure to magnetic fields. Psychopharmacology 1986; 89: 440– 443. 18 Kavaliers M, Ossenkopp KP. Opioid systems and magnetic field effects in the land snail, Cepaea nemoralis. Biol Bull 1991; 180: 301–309. 19 Del Seppia C, Ghione S, Luschi P, Papi F. Exposure to oscillating magnetic fields influences sensitivity to electrical stimuli. I. Experiments on pigeons. Bioelectromagnetics 1995; 16: 290–294. 20 Papi F et al. Exposure to oscillating magnetic fields influences sensitivity to electrical stimuli. 2. Experiments on humans. Bioelectromagnetics 1995; 16: 295–300.

21 Sartucci F et al. Changes in pain perception and painrelated somatosensory evoked potentials in humans produced by exposure to oscillating magnetic fields. Brain Res 1997; 769: 362–366. 22 Luschi P, Del Seppia C, Crosio E, Papi F. Pigeon homing: evidence against reliance on magnetic information picked up en route to release sites. Proc R Soc Lond B 1996; 263: 1219–1224. 23 Del Seppia C, Luschi P, Papi F. Influence of emotional factors on the initial orientation of pigeons. Anim Behav 1996; 52: 33– 47. 24 Papi F, Luschi P, Limonta P. Orientation-disturbing magnetic treatment affects the pigeon opioid system. J Exp Biol 1992; 166: 169–179.