Medicina (Kaunas) 2013;49(8):379-85 379 Associations Between Air Temperature and Daily Suicide Counts in Astana, Kazakhstan

Andrej M. Grjibovski1, 2, Gulmira Kozhakhmetova3, Aliya Kosbayeva4, Bettina Menne5 1Department of International Public Health, Norwegian Institute of Public Health, Norway, 2International School of Public Health, Northern State Medical University, Russia, 3Medical University of Astana, Kazakhstan, 4WHO Office in Kazakhstan, Kazakhstan, 5WHO European Centre for Environment and Health, Bonn Office, Germany

Key Words: suicides; temperature; apparent temperature; Central Asia; Kazakhstan.

Summary. Background and Objective. Seasonal variations in suicide mortality and its associa- tion with ambient air temperature have been observed in many countries. However, the evidence from Central Asia is scarce. The aim of the study was to assess the relationship between 4 indicators of air temperature and daily suicide counts in Astana, Kazakhstan. Material and Methods. The daily counts of suicides (ICD-10 codes, X60–X84) for the popu- lation of Astana in 2005–2010 were collected using death certificates and medical records at the Municipal Bureau of Forensic Medicine. Associations between the number of cases and , maxi- mum, mean apparent, and maximum apparent temperatures were studied using negative models controlling for the effects of month, year, weekends, holidays, wind velocity, baro- metric pressure, and relative humidity. Results. Altogether, there were 685 suicides in Astana in 2005–2010. A clear seasonal pattern with the peak in summer was observed. In crude analyses, significant associations between suicide counts and all 4 temperatures were found. After adjustment for other variables, only apparent tem- peratures remained significantly associated with the outcome. An increase in the mean apparent temperature by 1°C was associated with an increase in suicide counts by 2.1% (95% CI, 0.4–3.8). Similar results were obtained for the maximum apparent temperature (1.2%, 95% CI, 0.1–2.3). Conclusions. The results suggest a linear relationship between apparent temperatures and daily suicide counts across the whole spectrum of temperatures. Factors behind this association need further research with a further going aim to develop mitigation strategies in the period of climate change.

Introduction in spring with social interactions but not weather Suicide is among the top 10 causes of death ac- factors being the most important factors behind this counting for more than 1 million lives lost each year pattern (4). Although most of the frequently cited worldwide (1). Moreover, the incidence of suicide studies have reported seasonal variations in suicide is increasing making it a considerable public health rates with peaks during spring or summer months problem. According to the World Health Organiza- (5–7), several other studies failed to reveal a season- tion (WHO) estimates, 1.8% of all the global deaths al pattern (8–10). This heterogeneity in the results were attributed to suicide in 1998, and this propor- could be partly explained by the latitude of study tion is expected to increase to 2.4% by 2020 (2). sites. While seasonal variations were found in tem- Various socioeconomic, demographic, and psy- perate climates, no associations were reported from chological factors have been studied in relation to the equatorial regions (11). Moreover, studies from suicide during the last decades, although seasonal England and Wales (12) and Switzerland (13) us- variations in suicide rates have been documented ing the data for the periods of more than 100 years already in the 19th century. Morselli collected data have shown a decrease in the effect of seasons on on suicide from 28 European countries and found the incidence of suicide. Thus, the results of the as- that suicides were more common during summer sociations between suicide and seasons are still con- and less frequent during winter months suggesting tradictory warranting further research (14). that air temperature could increase the risk of sui- Air temperature has been the most studied cli- cide through the alterations of brain functions (3). matic variable that could explain seasonal variations Durkheim, however, in his famous book on sui- in suicide rates. Several studies from the 1980s– cides reported that suicides occur more frequently 1990s (15–17) reported associations between tem- perature and suicide, but these analyses were based Correspondence to A. M. Grjibovski, Department of Inter- on monthly and weekly data, threatening the valid- national Public Health, Norwegian Institute of Public Health, ity of the results since the effect of temperature on Postbox 4404, Nydalen, 0403 Oslo, Norway E-mail: [email protected] suicide-related behavior has been reported to be im-

Medicina (Kaunas) 2013;49(8) 380 Andrej M. Grjibovski, Gulmira Kozhakhmetova, Aliya Kosbayeva, Bettina Menne mediate (18). During the 2000s, a number of studies Material and Methods using daily data were published. Positive associations This is an ecological study using daily data on cli- between suicide and temperature on the same day matic factors and suicide counts in Astana. Data on and a day before were found in Germany (19), while all suicides (ICD-10 codes X60–X84) from Febru- temperatures on the same day only were associated ary 11, 2005, to December 31, 2010, were obtained with suicides in England and Wales (8), Taiwan (20), from the Municipal Bureau of Forensic Medicine. and Japan (21). Moreover, Lin et al. studied tem- Data on mean and maximum daily temperatures, perature, humidity, atmospheric pressure, rainfall humidity, barometric pressure, and wind speed were and sunshine, but only temperature was significantly obtained from the Kazakhstani Hydrometerological associated with suicide rates (20). However, no as- Service (Kazhydromet). In addition to traditionally sociation between meteorological variables and sui- used mean and maximum daily temperatures, we ap- cides was found in other studies (22). plied mean and maximum apparent temperatures as Given heterogeneity of the findings on the asso- the indices of discomfort related to the individual’s ciations between temperature and suicides, studies perceived air temperature since it combines the ef- in other regions are warranted (19). Moreover, most fects of temperature and humidity and is associated of the studies were performed in economically stable with body’s ability to cool by perspiration and evap- countries with temperate, subtropical, and tropical oration during hot days (28). Mean and maximum climates, while the evidence from transitional coun- apparent temperatures were calculated using the fol- tries, particularly with arid- or semiarid climates is lowing formula: nonexistent to the best of our knowledge. As in many other former Soviet Republics, sui- AT=–2.653+0.994*T+0.0153*DT2, cide rates in Kazakhstan (25.6 per 100 000 in 2008) where T is the mean or the maximum daily tem- are much greater than in other countries with simi- perature and DT is the dew point temperature (29). lar economic indicators. Male and female suicide Daily suicide counts were used as a discrete de- rates in 2008 were 43.0 per 100 000 and 9.3 per pendent variable. The mean daily temperature was 100 000, respectively, placing Kazakhstan among the primary independent variable. An exploratory the top 5 countries with the highest suicide rates in analysis using the countfit function in the Stata soft- the world (23). However, no studies on suicide and ware (30) was used to select the best fitting model its association with climatic factors originating from for a dependent variable. and par- Kazakhstan or other Central Asian republics of the tial autocorrelations of the dependent variable were former USSR have been published internationally. used to find the appropriate autoregressive term for The capital of Kazakhstan – Astana – is one of the regression models. A negative binomial regres- the fastest growing capitals in the world. Its popula- sion model with no autoregressive terms provided tion increased from 281 thousand in 1999 to 709 the best fit. Moreover, it accounts for the overdisper- thousand in 2010. According to the Köppen-Geiger sion of the data. Robust standard errors were used to classification, Astana is located in the steppe region additionally account for heterogeneity as described on the border between a humid continental and a elsewhere (30). Months, years, and holidays were semiarid climate and has cold winters and warm included in the models as dichotomous variables. summers (24). Mean temperatures for January and July are −17.3°C and 20.2°C, respectively. However, The analyses were repeated using a maximum daily winter temperatures below –30°C and summer tem- temperature as well as mean and maximum apparent peratures above 30°C are not uncommon providing temperatures. Additional adjustment was performed vast opportunities for studying the effects of tem- for barometric pressure, wind speed, and humidity. perature on various health outcomes (25–27). First, a curvilinear relationship between tem- A unique geographic location of this relatively perature and each of the outcomes was modeled by large city with one of the greatest of tem- fitting cubic splines with knots spaced every 5°C peratures across the world combined with rapid eco- across the whole temperature spectrum using the nomic changes and an initially high level of suicides uvrs function in the Stata software with a linear ad- make the population of Astana particularly vulner- justment for covariates (31). Then the most parsi- able to the effects of climatic factors and provides monious model was selected by the stepwise reduc- a rare opportunity to study associations between tion of the number of knots to the minimum using a temperature and suicide in a semiarid climate in 5% level of alpha error. The best fitting models were a country undergoing a rapid social and economic linear models (no knots) across the whole tempera- transition. ture spectrum. SPSS v. 17 was used for basic data The aim of this study was to assess the rela- processing (SPSS Inc., Chicago, IL, USA). Mod- tionship between air temperature and daily suicide eling was performed using the Stata 10.0 software counts in Astana, Kazakhstan. (Stata Corp, TX, USA).

Medicina (Kaunas) 2013;49(8) Relationship Between Temperature and Suicide Mortality in Kazakhstan 381

Results Table 1. Basic for Daily Suicide Counts and Climatic Altogether, 685 cases of suicide occurred in Variables in Astana, Kazakhstan, in 2005–2010 Astana between February 11, 2005, and Decem- ber 31, 2010. Daily suicide counts varied between Variable Mean SD Min Max 0 and 4 with a mean of 0.32 (SD, 0.58). A clear Suicide counts 0.3 0.6 0 4 seasonal pattern of suicides in Astana was observed: Mean temperature, °C 4.6 14.2 –36.2 31.7 0.40 suicides on the average occurred during sum- Maximum temperature, °C 10.1 15.1 –32.7 40.1 Mean apparent mer months while the corresponding numbers for temperature, °C 4.1 12.4 –16.2 30.7 spring, autumn, and winter were 0.37, 0.26, and Maximum apparent 0.23, respectively. Basic data on daily suicides and temperature, °C 11.2 17.0 –15.7 52.6 climatic variables are summarized in Table 1. Al- Relative humidity, % 64.9 16.4 14 98 together, 32% and 30% of all the suicides occurred Wind velocity, m/s 3.0 1.6 0 13.4 during summers and springs, respectively. The mean Barometric pressure, hPa 977.8 9.3 951.2 1010.7 numbers of suicides per day are presented in Fig. 1. No clear pattern between the 4 temperatures and daily suicides counts were revealed (Fig. 2), and 0.5 statistically, the linear model provided the best fit to the data. In crude analyses, all 4 temperatures 0.4 were positively and significantly associated with the outcome. After adjustment for the effects of years, months, and holidays, the coefficients remained 0.3 significant for all temperatures but daily maximum temperature. Further adjustment for wind velocity, 0.2 humidity, and barometric pressure resulted in the fact that only mean apparent and maximum appar- ent temperatures remained significantly associated 0.1 with daily suicide counts (Table 2). An increase in Mean Number of Suicides per Day 1°C in the mean apparent temperature was associ- 0.0 ated with an increase in the number of suicides by Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 2.1% (95% CI, 0.4–3.8). A corresponding increase Month in the maximum apparent temperature was 1.2% (95% CI, 0.1–2.3). Moreover, when seasons were Fig. 1. Mean daily suicide counts in Astana, Kazakhstan, included in the model with any of the studied tem- in 2005–2010 by month of the year peratures, the coefficients for seasons were reduced to nonsignificant values (data not shown) suggest- maximum, mean apparent and maximum apparent ing that the temperature is likely to explain seasonal temperatures were associated with the number of variations in suicide counts in the study setting. suicides in crude analyses, but after controlling for seasonal, long-term effects and other climatic vari- Discussion ables, only mean apparent and maximum apparent This is to our knowledge the first study on the temperatures remained significantly related to the associations between air temperature and suicides in outcome. Moreover, we did not find any threshold Central Asia. Moreover, this is one of the first stud- for the effect of temperature that suggests a linear ies that assessed the effect of 4 different temperature relationship between all the temperature and daily estimates on the daily suicide counts. Both mean, suicide counts in Astana.

Table 2. A Percentage Increase in the Number of Suicide Cases Associated With an Increase in Mean, Maximum, Mean Apparent, and Maximum Apparent Temperatures by 1°C in Astana, Kazakhstan, in 2005–2010

Variable1 Crude 95% CI Adjusted2 95% CI Adjusted3 95% CI Mean temperature 1.42 0.84; 2.00 1.53 0.15; 2.92 1.37 –0.21; 2.94 Maximum temperature 1.25 0.71; 1.78 1.11 –0.02; 2.38 0.91 –0.58; 2.39 Mean apparent temperature 1.62 0.98; 2.25 2.01 0.39; 3.62 2.114 0.42; 3.79 Maximum apparent temperature 1.12 0.66; 1.58 1.11 0.03; 2.19 1.194 0.07; 2.31 1Each temperature was studied using separate models. 2Adjusted for years, months, and holidays. 3Adjusted for years, months, holidays, wind velocity, humidity, and barometric pressure. 4Apparent temperatures reflect the effects of both temperature and humidity; therefore, humidity was excluded from the model.

Medicina (Kaunas) 2013;49(8) 382 Andrej M. Grjibovski, Gulmira Kozhakhmetova, Aliya Kosbayeva, Bettina Menne

Regression Spline (–25 –20 –15 –5 0 5 10 15 20 25) Regression Spline (–15 –5 0 5 10 15 20 25 30 35) Adjusted for Covariates Adjusted for Covariates 2 2

0 0

–2 –2

Log Daily Counts of Suicides Log –4 Daily Counts of Suicides Log –4 –40 –20 0 20 40 –40 –20 0 20 40

Mean Temperature Maximum Temperature

Regression Spline (–15 –5 0 5 10 15 20 25) Regression Spline (–15 –5 0 5 10 15 20 25 30 35) Adjusted for Covariates Adjusted for Covariates 2 2

1 1

0 0

–1 –1

–2

Log Daily Counts of Suicides Log Daily Counts of Suicides Log –2 –20 –10 0 10 20 30 –20 –10 0 10 20 30 40 50 60

Mean Apparent Temperature Maximum Apparent Temperature

Fig. 2. Relationship between the daily suicide counts and mean (top left), maximum (top right), mean apparent (bottom left), and maximum apparent temperatures (bottom right) in Astana, Kazakhstan, in 2005–2010

The main strength of this study is the use of the to confound associations between the temperature data on daily counts and meteorological variables and suicide. from one setting that reduces the likelihood of geo- Another limitation of the study is a relatively graphical bias, which might threaten the validity of small number of suicides that occurred during the the findings is several previous studies where aver- study period resulting in limited statistical power for age temperatures for the whole country were used. stratified analyses. Given that the male-to-female Moreover, all cases of suicide as registered in the ratio of suicides in Kazakhstan is more than 4 to Municipal Bureau of Forensic Medicine during the 1, our results mostly reflect the pattern of suicide study period that minimized selection bias. How- among men, while the models for women did not ever, the data from January 1, 2005, to February 10, converge due to a low number of cases. 2005, were unavailable. This limitation results in a One more limitation, which is typical of most slight underestimation of the proportion of suicides similar studies on climatic factors and health out- in winter. However, the degree of underestimation comes, is its ecological design, i.e., the use of ag- is small. Moreover, our finding on a greater mean gregated data on both exposure and outcome with number of suicides per day by month (Fig. 1) re- no possibility to assess an association between the mains unaffected as well as the associations between temperature and suicide at the individual level. the temperature and the outcome. Another strength However, it is reasonable to assume that those who is that we adjusted for other meteorological vari- died were at least to the same degree exposed to the ables, such as wind velocity, barometric pressure, effects of temperature as the rest of the population. and humidity (the latter was not used for the analy- Misclassification of the cause of death has been sis of apparent temperatures), which have a potential mentioned as a potential limitation of several simi-

Medicina (Kaunas) 2013;49(8) Relationship Between Temperature and Suicide Mortality in Kazakhstan 383 lar studies (19–22). In our study, we used the data ences in the methods, statistical from the Municipal Bureau of forensic medicine, modeling techniques, and environmental variables, which generally provide more valid data on the ex- which were used for adjustment. Most studies (8, ternal causes of death than the in 10, 12, 13), although not all (19–22), used the data the post-Soviet states (32), although no evaluation of at the national level ensuring adequate statistical the validity of the registration of suicides in Kazakh- power, although at the expense of geographical bias, stan has been performed. However, even there is the i.e., the geographical differences between the place potential for the underestimation of the number of where the temperature was measured and where sui- suicides in this study, it is unlikely that the quality cide took place. In our study, the distance between of registration is related to the temperature. Thus, if the meteorological station and the place of suicide present, this bias is nondifferential and would influ- is less than 10 km, but at the expense of statisti- ence the width of the confidence intervals, but not cal power. Modeling techniques vary considerably the point estimates. among studies. Our approach is similar to the ones Our main results are in line with most previous widely used in studies on the associations between studies showing a clear seasonal pattern of suicides cardiovascular, cerebrovascular, and respiratory dis- (3, 5–7, 12, 13, 21). However, while the peak of sui- eases with first modeling curvilinear relationships cide mortality occurs in spring in several countries and comparing them with linear models (28). The (4, 21), our results clearly show that although there main difference is that we used dichotomous “dum- are more suicides in spring than autumn and winter, my” variables for months and years, while other re- summer months, particularly June and July, are the searchers used various smoothing functions to con- months with the highest numbers of suicides per trol for secular and seasonal trends (8, 21). day (Fig. 1). Interestingly, in multivariable analyses, Nearly all previous studies used a mean tempera- the introduction of the temperature in the model ture as the main exposure variable (15–22). Our attenuates the effect of the season, which becomes study seems to be among the first to evaluate the ef- close to zero and nonsignificant. Thus, we one can fect of not only the mean temperature, but also mean speculate that air temperature or other factors close- apparent, maximum, and maximum apparent tem- ly associated with air temperature may at least partly peratures. Interestingly, only apparent temperatures explain the observed seasonal variation in Astana. remained significantly associated with the outcome However, the effects of temperature were found in after adjustment for seasonal effects and other mete- other settings even in the absence of the seasonal orological variables. Moreover, after adjustment for patterns of suicide (8). In the latter study, contrary other factors, the effect of mean and maximum tem- to our findings, a nonlinear relationship between the peratures was slightly reduced, but the association temperature and suicides was found. An increase in with apparent temperatures strengthened. However, the temperature above 18°C was associated with a the coefficient for the mean temperature decreased 3.8% increase, on the average, in the daily suicide only marginally, and the loss of significance can be counts. In our study, a linear model provided the attributed to limited power. Apparent and maximum best fit, although given the low number of cases, a apparent temperatures were used in studies on car- nonlinear relationship could not be detected due to diovascular or respiratory diseases, but not suicides limited power. However, looking at Fig. 2, one may (28). The apparent temperature is an index of dis- suspect thresholds at 25°C for mean and mean ap- comfort related to the individual’s perceived air tem- parent temperatures and at 35°C for maximum and perature since it combines the effects of temperature maximum apparent temperatures. We attempted to and humidity and is associated with the body’s ability repeat our analyses for the parts of the temperature to cool by perspiration and evaporation during hot range above the thresholds, but the number of cases days. Biological mechanisms behind the observed was too small for the models to converge. associations between apparent temperatures, particu- Most studies reported associations between the larly the mean apparent temperature and suicide, in temperature and the number of suicides with no de- Astana require further research. layed effects. Muller et al. found an association be- Mechanisms through which suicide counts are tween suicide counts and the temperature not only greater during the days with a higher temperature at the day of the event, but also with a lag of one remain unknown. Several mechanisms were dis- day. We repeated our analyses using temperatures cussed and summarized in 3 groups: sociological, with lags 0–3, but did not find a significant asso- biological, and psychological (8). Kazakhstan as ciation between suicide and lagged variables sup- many other former Soviet Republics has a high per porting the hypothesis that the effect of temperature capita consumption of alcohol particularly among seems to be immediate (18). men, and one may speculate that more alcohol Several other studies reported positive associa- could be consumed in warmer days (8), although tions between the temperature and suicide counts, there is no evidence that this might be the case in but direct comparisons are difficult due to differ- Medicina (Kaunas) 2013;49(8) 384 Andrej M. Grjibovski, Gulmira Kozhakhmetova, Aliya Kosbayeva, Bettina Menne

Kazakhstan. Anderson et al. showed that high tem- with findings from other settings and suggest a lin- peratures may influence the level of aggressiveness ear relationship between daily suicide counts and and violence (33), and Page et al. suggested that this mean apparent and maximum apparent tempera- may increase the likelihood for committing suicide tures with no thresholds. Factors and mechanisms in these days, because in their study, the strongest behind this association warrant further research with association with temperature was observed for vio- a further going aim to develop mitigation strategies lent suicides (8). Several studies have shown altera- in the period of climate change. tions in the noradrenergic, dopaminergic systems and particularly serotonin systems in the pathogen- Acknowledgments esis of suicide (34). Lower serotonergic activity in- This publication has been developed within the creases aggressive behavior and impulsiveness and WHO/BMU project on protecting health from cli- thus may induce suicidal behavior (35). However, it mate change in Europe, coordinated by Dr. Bettina remains unknown whether an increase in tempera- Menne and Dr. Joanna Nurse, WHO Regional Of- ture is associated with a decrease in serotonin levels fice for Europe. We are grateful for the financial and requires further research. support received from Germany.

Conclusions Statement of Conflict of Interest The results from Astana, Kazakhstan, are in line The authors state no conflict of interest.

References 1. Claassen CA, Yip PS, Corcoran P, Bossarte RM, Lawrence 16. Linkowski P, Martin F, De Maertelaer V. Effect of some BA, Currier GW. National suicide rates a century after Dur- climatic factors on violent and non-violent suicides in Bel- kheim: do we know enough to estimate error? Suicide Life gium. J Affect Disord 1992;25:161-6. Threat Behav 2010;40:193-223. 17. Preti A, Miotto P. Seasonality in suicides: the influence of 2. World Health Organization. Suicide prevention, 2005 [Cit- suicide method, gender and age on suicide distribution in ed 22 October 2012]. Available from: URL: http://www.who. Italy. Psychiatry Res 1998;81:219-31. int/mental_health/prevention/suicide/suicideprevent/en/. 18. Basu R, Samet JM. Relation between elevated ambient tem- 3. Morselli J. Suicide: an essay on comparative moral statistics. perature and mortality: a review of the epidemiologic evi- London: Kegan Paul; 1881. dence. Epidemiol Rev 2002;24:190-202. 4. Durkheim E. Suicide: a study in sociology. New York: The 19. Müller H, Biermann T, Renk S, Reulbach U, Ströbel A, Free Press; 1966. Kornhuber J, et al. Higher environmental temperature and 5. Petridou E, Papadopoulos FC, Frangakis CE, Skalkidou A, global radiation are correlated with increasing suicidality – Trichopoulos D. A role of sunshine in the triggering of sui- a localized data analysis. Chronobiol Int 2011; 28:949-57. cide. 2002;13:106-9. 20. Lin HC, Chen CS, Xirasagar S, Lee HC. Seasonality and 6. Lambert G, Reid C, Kaye D, Jennings G, Esler M. In- climatic associations with violent and nonviolent suicide: creased suicide rate in the middle aged and its association a population-based study. Neuropsychobiology 2008;57: with hours of sunlight. Am J Psychiatry 2003;160:793-5. 32‑7. 7. Hakko H, Rasanen P, Tiihonen J. Seasonal variation in sui- 21. Likhvar V, Honda Y, Ono M. Relation between temperature cide occurrence in Finland. Acta Psychiatr Scand 1998;8: and suicide mortality in Japan in the presence of other con- 92-7. founding factors using time-series analysis with a semipara- 8. Page L, Hajat S, Kovats RS. Relationships between daily metric approach. Environ Health Prev Med 2011;16:36-43. suicide counts and temperature in England and Wales. Br J 22. Chiu LP. Do weather, day of the week, and address affect Psychiatry 2007;191:106-12. the rate of attempted suicide in Hong Kong? Soc Psychiatry 9. Kevan SM. Perspectives on season of suicide: a review. Soc Psychiatr Epidemiol 1988;23:229-35. Sci Med 1980;14:369-78. 23. WHO: suicide rates per 100,000 by country, year and sex 10. Ajdacic-Gross V, Lauber C, Sansossio R, Bopp M, Eich D, [Cited 22 October 2012]. Available from: URL: http://www. Gostynski M, et al. Seasonal associations between weather who.int/mental_health/prevention/suicide_rates/en/. conditions and suicide – evidence against a classic hypoth- 24. Grjibovski AM, Kosbayeva A, Menne B. The effect of am- esis. Am J Epidemiol 2007;165:561-9. bient air temperature and precipitation on monthly counts 11. Parker G, Gao F, Machin D. Seasonality of suicide in Singa- of salmonellosis in four regions of Kazakhstan, Central pore: data from the equator. Psychol Med 2001;31:549-53. Asia, in 2000-2010. Epidemiol Infect 2013;2:1-8. 12. Yip PS, Chao A, Chiu CW. Seasonal variation in suicides: 25. Peel MC, Finlayson BL, McMahon TA. Updated world map diminished or vanished. Experience from England and of the Köppen-Geiger climate classification. Hydrol Earth Wales, 1982–1996. Br J Psychiatry 2000;177:366-9. Syst Sc 2007;11:1633-44. 13. Ajdacic-Gross V, Bopp M, Sansossio R, Lauber C, Gos- 26. Grjibovski AM, Nurgaliyeva N, Kosbayeva A, Menne B. No tynski M, Eich D, et al. Diversity and change in suicide association between temperature and deaths from cardiovas- seasonality over 125 years. J Epidemiol Community Health cular and cerebrovascular diseases during the cold season 2005;59:967-72. in Astana, Kazakhstan – the second coldest capital in the 14. Deisenhammer EA. Weather and suicide: the present state world. Int J Circumpolar Health 2012;71. of knowledge on the association of meteorological factors 27. Grjibovski AM, Nurgaliyeva N, Kosbayeva A, Sharbakov A, with suicidal behavior. Acta Psychiatr Scand 2003;108: Seysembekov T, Menne B. Effect of high temperatures on 402‑9. daily counts of mortality from diseases of circulatory system 15. Souetre E, Wehr TA, Douillet P, Darcourt G. Influence of in Astana, Kazakhstan. Medicina (Kaunas) 2012;48:640-6. environmental factors on suicidal behavior. Psychiatry Res 28. Baccini M, Biggeri A, Accetta G, Kosatsky T, Katsouyanni 1990;32:253-63. K, Analitis A, et al. Heat effects on mortality in 15 Euro-

Medicina (Kaunas) 2013;49(8) Relationship Between Temperature and Suicide Mortality in Kazakhstan 385

pean cities. Epidemiology 2008;19:711-9. 32. Väli M, Lang K, Soonets R, Talumäe M, Grjibovski AM. 29. Kalkstein LS, Valimont KM. An evaluation of summer dis- Childhood deaths from external causes in Estonia, 2001- comfort in the United States using a relative climatological 2005. BMC Public Health 2007;7:158. index. Bull Am Meteorol Soc 1986;67:842-8. 33. Anderson C. Temperature and aggression: ubiquitous ef- 30. Long S, Freese J, editors. Regression models for categorical fects of heat on occurrence of human violence. Psychol Bull dependent variables using stata. 2nd ed. Tex: Stata Corp; 1989;106:74-96. 2006. p. 527. 34. Mann JJ. Neurobiology of suicidal behavior. Nat Rev Neu- 31. Royston P, Sauerbrei W. Multivariable modelling with cu- rosci 2003;4:819-28. bic regression splines: a principled approach. Stata Journal 35. Bloom FF, Kupfer DJ, editors. Psychopharmacology. New 2007;7:45-70. York: Raven Press; 1995.

Received 2 June 2013, accepted 30 August 2013

Medicina (Kaunas) 2013;49(8)