BMJ Open Is Committed to Open Peer Review. As Part of This Commitment We Make the Peer Review History of Every Article We Publish Publicly Available

BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from

BMJ Open is committed to open peer review. As part of this commitment we make the peer review history of every article we publish publicly available.

When an article is published we post the peer reviewers’ comments and the authors’ responses online. We also post the versions of the paper that were used during peer review. These are the versions that the peer review comments apply to.

The versions of the paper that follow are the versions that were submitted during the peer review process. They are not the versions of record or the final published versions. They should not be cited or distributed as the published version of this manuscript.

BMJ Open is an open access journal and the full, final, typeset and author-corrected version of record of the manuscript is available on our site with no access controls, subscription charges or pay-per-view fees (http://bmjopen.bmj.com).

If you have any questions on BMJ Open’s open peer review process please email [email protected] http://bmjopen.bmj.com/ on September 25, 2021 by guest. Protected copyright. BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from

Epidemiology of Accidental Hypothermia in Denmark – a Nationwide Cohort Study

Journal: BMJ Open ManuscriptFor ID peerbmjopen-2020-046806 review only Article Type: Original research

Date Submitted by the 09-Nov-2020 Author:

Complete List of Authors: Wiberg, Sebastian; Rigshospitalet, Department of Cardiology Mortensen, Asmus; Bispebjerg and Frederiksberg Hospital, Department of Anaesthesia and Intensive Care Kjærgaarda, Jesper; Rigshospitalet, Department of Cardiology Hassager, Christian; Rigshospitalet, Cardiology Wanscher, Michael; Rigshospitalet, Department of CardioThoracic Anaesthesiology

ACCIDENT & EMERGENCY MEDICINE, EPIDEMIOLOGY, GENERAL Keywords: MEDICINE (see Internal Medicine)

http://bmjopen.bmj.com/ on September 25, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 22 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 6 7 8 9 I, the Submitting Author has the right to grant and does grant on behalf of all authors of the Work (as defined 10 in the below author licence), an exclusive licence and/or a non-exclusive licence for contributions from authors 11 who are: i) UK Crown employees; ii) where BMJ has agreed a CC-BY licence shall apply, and/or iii) in accordance 12 with the terms applicable for US Federal Government officers or employees acting as part of their official 13 duties; on a worldwide, perpetual, irrevocable, royalty-free basis to BMJ Publishing Group Ltd (“BMJ”) its 14 licensees and where the relevant Journal is co-owned by BMJ to the co-owners of the Journal, to publish the 15 Work in this journal and any other BMJ products and to exploit all rights, as set out in our licence. 16 17 The Submitting Author accepts and understands that any supply made under these terms is made by BMJ to 18 the Submitting Author Forunless you peer are acting as review an employee on behalf only of your employer or a postgraduate 19 student of an affiliated institution which is paying any applicable article publishing charge (“APC”) for Open 20 Access articles. Where the Submitting Author wishes to make the Work available on an Open Access basis (and 21 intends to pay the relevant APC), the terms of reuse of such Open Access shall be governed by a Creative 22 Commons licence – details of these licences and which Creative Commons licence will apply to this Work are set 23 out in our licence referred to above. 24 25 Other than as permitted in any relevant BMJ Author’s Self Archiving Policies, I confirm this Work has not been 26 accepted for publication elsewhere, is not being considered for publication elsewhere and does not duplicate 27 material already published. I confirm all authors consent to publication of this Work and authorise the granting 28 of this licence. 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 25, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 22

1 2 3 Epidemiology of Accidental Hypothermia in Denmark – a Nationwide 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 Cohort Study 6 7 8 9 Authors 10 Sebastian Wiberg* MD PhD1, Asmus Friborg Mortensen* MD2, Jesper Kjaergaard, 11 12 MD PhD DMSc1, Christian Hassager MD DMSc1, Michael Wanscher MD PhD3. 13 14 15 16 Affiliations 17 1) Department of Cardiology, The Heart Centre, Copenhagen University 18 For peer review only 19 Hospital Rigshospitalet 20 21 2) Department of Anaesthesia and Intensive Care, Bispebjerg and 22 23 Frederiksberg Hospital, University of Copenhagen, Denmark 24 3) Department of Cardiothoracic Anaesthesiology, The Heart Centre, 25 26 Copenhagen University Hospital Rigshospitalet, Denmark 27 28 29 30 Corresponding Author 31 Sebastian Wiberg, MD, PhD 32 33 Department of Cardiology 34 35 The Heart Centre

36 http://bmjopen.bmj.com/ 37 Copenhagen University Hospital Rigshospitalet 38 [email protected] 39 40 41 42 *) Equal author contribution 43

44 on September 25, 2021 by guest. Protected copyright. 45 46 Research Question 47 To investigate the reported incidence of accidental hypothermia in a nationwide 48 49 registry, and to further investigate the associations between accidental 50 51 hypothermia and the associated outcomes. 52 53 54 Study Design 55 56 Retrospective cohort study 57 58 59 60

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 22 BMJ Open

1 2 3 Abstract 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 Objectives 6 7 To investigate the reported incidence of accidental hypothermia (AH) in a 8 9 nationwide registry, and to further investigate the associations between 10 accidental hypothermia and the associated outcomes. 11 12 13 14 Design 15 16 Nationwide retrospective cohort study 17 18 For peer review only 19 Participants 20 21 All adult patients admitted to hospitals in Denmark with a diagnosis of AH from 22 23 January 1996 to November 2016. Patients were identified as all patients 24 admitted with an ICD-10 code of T689. Other recorded diagnoses were included 25 26 in the analyses. 27 28 29 30 Primary and secondary outcome measures 31 The primary outcome was one-year fatality. 32 33 34 35 Results

36 http://bmjopen.bmj.com/ 37 Throughout the inclusion period, 5242 patients were admitted with a diagnosis 38 of AH, corresponding to an average annual incidence of 4.4 (range 2.9 – 6.4) per 39 40 100,000 inhabitants. A total of 2230 (43%) had AH recorded as the primary 41 42 diagnosis without any recorded secondary diagnoses (Primary AH), 1336 (25%) 43 had AH recorded as the primary diagnosis with other recorded secondary

44 on September 25, 2021 by guest. Protected copyright. 45 46 diagnoses (Secondary AH, type I), and 1676 (32%) had AH recorded as a 47 secondary diagnosis with another recorded primary diagnosis (Secondary AH, 48 49 type II). Alcohol intoxication was the most prevalent diagnosis associated to AH. 50 51 Overall one-year fatality was 27%. In patients with primary AH, one-year fatality 52 53 was 22%, compared to 26% in patients with secondary AH type I, and 35% in 54 patients with secondary AH type II (p < 0.001). 55 56 57 58 Conclusions 59 60

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 22

1 2 3 The incidence of AH was 4.4 per 100.000 per year in the period from 1995 to 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2016, and the diagnosis was associated with a high comorbidity burden. AH 6 7 patients had a one-year fatality of 27%. In otherwise healthy patients with 8 9 primary AH, the one-year fatality was 22%. 10 11 12 Article Summary 13 14 Strengths and Limitations of this study 15 16 - Largest contemporary study to investigate the incidence, comorbidity 17 burden and outcome in patients diagnosed with AH. 18 For peer review only 19 - Data originated from national registers, which limited the data 20 21 granularity to what has been presented in the present paper. 22 23 24 Funding 25 26 The study was funded by Trygfonden - Accidental Hypothermia (ID: 117579). 27 28 The sponsor had no involvement in study design, collection, analysis and 29 30 interpretation of data, writing of the manuscript or decision to submit the 31 manuscript for publication. 32 33 34 35 Competing interests

36 http://bmjopen.bmj.com/ 37 None 38 39 40 41 42 43

44 on September 25, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 22 BMJ Open

1 2 3 Introduction 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 Accidental hypothermia (AH), defined as an involuntary drop in core body 6 7 temperature below 35C, is a serious condition with considerable fatality1,2. AH 8 9 can result from isolated cold exposure (primary hypothermia) or in relation to 10 acute or chronic illness (secondary hypothermia)2. The severity of AH is graded 11 12 by the Swiss Staging System into four stages based on core temperature: 32°C - 13 14 35°C (Stage I, mild AH), 28°C - < 32°C (Stage II, moderate AH), and < 28°C (Stage 15 16 III, severe AH). Stage IV is characterized by an accentuated drop of core 17 temperature and absence of vital signs3-5. In-hospital fatality after Stage III 18 For peer review only 19 hypothermia has been reported to be as high as 30%, and it is estimated that 20 21 approximately 1500 people die from hypothermia every year in the United 22 1,4 23 States . In secondary hypothermia, the risk of death is considered to be caused 24 25 by the underlying condition rather than the hypothermia itself. However, our 26 knowledge of the incidence, demography and outcomes after AH remains sparse. 27 28 The aim of this study was to investigate the reported incidence of AH in a 29 30 nationwide registry, and to further investigate the associations between AH and 31 32 comorbidities, and the associated outcomes in adult patients admitted with AH. 33 34 35 Methods

36 http://bmjopen.bmj.com/ 37 The present study is a nationwide retrospective cohort study including all adult 38 39 patients admitted to hospitals in Denmark with a diagnosis of AH from January 40 1996 to November 2016. The primary outcome was one-year fatality after 41 42 diagnosis. We chose a priori to include only those cases with a first-time 43 diagnosis of AH. The rationale behind this decision was to ensure that a patient

44 on September 25, 2021 by guest. Protected copyright. 45 46 with multiple hospital admissions for AH only counted once in the analysis. 47 All Danish residents are given a personal identification number (CPR, central 48 49 person registry number) at the time of birth or immigration. This number is used 50 51 in all contacts with the health care system, and it is used as identifier in national 52 53 registries. The Danish health care system is federally funded, and citizens are 54 eligible without payment. The Danish government maintains several nationwide 55 56 registries to keep track of expenses. We leveraged data from the Danish National 57 58 Patient Register (NPR) to identify all patients admitted with a final recorded 59 60 diagnosis of AH (i.e. recorded upon discharge from hospital), defined by the

4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 22

1 2 3 International Classification of Diseases, 10th edition (ICD-10, which has been used 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 in Denmark since 1993) code T689. 6 7 We stratified patients into three groups based on whether AH was recorded 8 9 as the only diagnosis (Primary AH), whether AH was recorded as the primary 10 diagnosis with other recorded secondary diagnoses (Secondary AH, type I), or 11 12 whether AH was recorded as a secondary diagnosis (Secondary AH, type II) for a 13 14 given admission. The rationale behind this stratification was to distinguish 15 16 patients with isolated AH and no comorbidity versus patients with comorbidity, 17 where the admitting physician believed that the AH diagnosis was the primary 18 For peer review only 19 cause of the admission versus patients with comorbidity, where the admitting 20 21 physician believed that the AH diagnosis was a secondary cause of admission. 22 23 The primary diagnosis is defined as the diagnosis primarily causing the need 24 for the admission of the patient to the hospital, while the secondary diagnoses 25 26 are defined as other conditions that required attention during a given admission. 27 28 Hospitals are reimbursed based on correct reporting of diagnoses, and 29 6 30 accordingly, the registry has a high accuracy for most diagnoses . Main treatment 31 categories, such as ‘intensive care admission’, and ‘respirator treatment’ were 32 33 also leveraged from NPR, and data were reported. Furthermore, the AH 34 35 diagnosis was linked to survival (i.e. the primary outcome) using the nationwide

36 http://bmjopen.bmj.com/ 37 Danish Register of Causes of Death (RCD). The study was approved by the Danish 38 Patient Safety Authority Authority (Institutional review board, ref. no 3-3013- 39 40 1906/1/) and was conducted in accordance with the Helsinki Declaration. 41 42 43 Demography

44 on September 25, 2021 by guest. Protected copyright. 45 2 46 Denmark covers a relatively small area of 43.000m but has a comparatively long 47 coastline of approximately 7300km. The highest point is 173 meters above sea 48 49 level. It is located on the 55th to 57th parallel north in the temperate climate zone. 50 51 Mean temperatures ranges from 0C in January and February to 15.7C in 52 st 53 August. The population consisted of 5,797,251 inhabitants per January 1 , 2016. 54 A total of 1,789,174 (31%) of the population lived in the capital region of 55 56 Denmark (Copenhagen) as per January 1st, 2016. 57 58 59 60 Statistical analyses

5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 22 BMJ Open

1 2 3 Throughout, categorical variables are presented as counts (%), normally 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 distributed continuous variables are presented as mean ± SD and skewed 6 7 continuous variables are presented as median (25th percentile - 75th percentile). 8 9 Normality was assessed visually by QQ plots. Associations between categorical 10 variables and AH types were tested by the Chi-Square test. Associations between 11 12 age and AH types were tested by the analyses of variance (ANOVA). 13 14 To assess any change in the incidence of AH over time, we applied linear 15 16 models with calendar year being included as a continuous covariate assuming a 17 linear trend, and we presented the absolute change in AH incidence per calendar 18 For peer review only 19 year  standard deviation. We repeated this analysis after stratification for AH 20 21 type. 22 23 The Kaplan-Meier estimator was applied to visualize fatality over time with 24 25 the Log-Rank test being applied to assess possible differences between strata. 26 We applied Cox proportional hazard models adjusting for isolated AH versus AH 27 28 with other secondary diagnoses, sex, age at admission, calendar year, and 29 30 geographical region, and presented hazard ratios with 95% confidence intervals 31 32 (95%CI). To assess possible collinearity, we included the following interaction 33 terms to the analyses: ‘sex*age group’, ‘sex*AH type’, ‘sex*calendar year’. In case 34 35 of a significant interaction term, we repeated the analyses after stratification into

36 http://bmjopen.bmj.com/ 37 groups. A significance level of 0.05 was applied throughout, and SAS software, 38 39 version 9.4 (SAS Institute, Cary, North Carolina, USA) was used for all statistical 40 analysis. 41 42 43 Results

44 on September 25, 2021 by guest. Protected copyright. 45 46 Over the inclusion period of almost 22 years, the Danish Health Care system 47 received a total of 5242 adult patients admitted with a diagnosis of AH. A total of 48 49 2230 (43%) had AH recorded as the final primary diagnosis without any 50 51 recorded secondary diagnoses (Primary AH), 1336 (25%) had AH recorded as 52 53 the final primary diagnosis with other recorded secondary diagnoses (Secondary 54 AH, type I), and 1676 (32%) had AH recorded as a final secondary diagnosis with 55 56 another recorded primary diagnosis (Secondary AH, type II). 57 58 A total of 1981 (38%) of the population were female, and mean age was 59 60 6121 years. Female sex was less prevalent among patients with primary AH

6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 22

1 2 3 compared to patients with secondary AH, type I and II (35% vs. 39% vs. 40%, p< 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 0.001). Further, patients with primary AH were younger compared to patients 6 7 with secondary AH, type I and II (5922 years vs. 6220 years vs. 6220 years, p 8 9 < 0.001). 10 11 12 Comorbidity 13 14 Both secondary AH type I and II were associated with a significant comorbidity 15 16 burden. A list of associated diagnoses with a proportion higher than 1% is 17 provided in Table 1. The most frequent diagnosis associated with AH was alcohol 18 For peer review only 19 intoxication (Table 1). 20 21 A total of 816 (16%) patients with AH received care in an intensive care unit. 22 23 A total of 305 (5.8%) were treated with vasopressors and/or inotropic agents, 24 25 397 (7.6%) received respirator treatment, and 479 (9.1) received dialysis. A total 26 of 21 (0.4%) patients were treated with extracorporeal circulation. 27 28 29 30 Incidence 31 32 The average annual per capita incidence of AH was 4.4 per 100,000 (95%CI 3.9 – 33 4.9 per 100,000) inhabitants with a range from 2.9 to 6.4 per 100,000 34 35 inhabitants (Figure 1). The overall incidence of AH was increasing from 2000 to

36 http://bmjopen.bmj.com/ 37 2016 (absolute increase 0.160.02 per 100,000 inhabitants per calendar year, p 38 39 < 0.001), which was driven by an increasing incidence of secondary AH type I 40 (absolute increase 0.040.006 per 100,000 inhabitants per calendar year, p < 41 42 0.001) and an increasing incidence of secondary AH type II (absolute increase 43

44 0.090.007 per 100,000 inhabitants per calendar year, p < 0.001). We found no on September 25, 2021 by guest. Protected copyright. 45 46 significant increase in the incidence of primary AH (absolute increase 0.020.01 47 48 per 100,000 per calendar year, p = 0.06). 49 The incidence of AH followed a seasonal trend, with the highest percent of AH 50 51 occurring in January (16%) and the lowest percent of AH occurring in August 52 53 (3.0%, Figure 2). Of all AH diagnoses, 29% were recorded in the Danish capital 54 55 region inhabited by 31% of the total Danish population. 56 57 58 Outcome 59 60

7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 22 BMJ Open

1 2 3 Overall fatality within the first 7 days of admission was 11%, increasing to 16% 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 at 30 days, and 27% at one year. Having secondary AH, type II, was associated 6 7 with a significantly higher fatality compared to having secondary AH, type I, 8 9 which was associated with a significantly higher fatality compared to having 10 primary AH (p < 0.001, Figure 3a), which remained unchanged after 11 Log-Rank 12 adjustment for confounding factors (Table 2). One-year fatality was 22% in 13 14 patients with primary AH, 26% in patients with secondary AH type I, and 35% in 15 16 patients with secondary AH type II (p < 0.001). 17 Female sex was associated with a higher fatality compared to male sex in 18 For peer review only 19 univariable analysis (HR 1.3 (1.2 – 1.4), p < 0.0001, Figure 3b), however, after 20 21 adjustment, female sex was associated with a lower fatality compared to male 22 23 sex with a HR of 0.89 (0.80 - 0.99, p = 0.04, Table 2). The interaction between sex 24 and age group was non-significant in the multivariable analysis (p=0.06). The 25 26 interactions between sex and AH type, and between sex and calendar year were 27 28 non-significant in multivariable analyses (p = 0.30, p = 0.45, respectively). 29 30 Higher age at admission was associated with higher fatality in both 31 univariable (Figure 3c) and multivariable analyses (Table 2). Calendar year was 32 33 not associated with outcome in univariable analysis, however, after adjustment, 34 35 calendar year was associated with decreased fatality with a HR of 0.98 (0.98 –

36 http://bmjopen.bmj.com/ 37 0.99, p < 0.001) per year (Table 2). We found no significant associations between 38 geographical region and outcome in multivariable analyses. 39 40 41 42 Discussion 43 In this retrospective cohort study, we found an annual incidence of AH of 4.4 per

44 on September 25, 2021 by guest. Protected copyright. 45 46 100,000. Overall, the AH diagnosis is associated with a high comorbidity burden, 47 and with an overall one-year fatality of 27%. 48 49 To the best of our knowledge, this is the largest contemporary study to 50 51 investigate the incidence, comorbidity burden and outcome in patients 52 53 diagnosed with AH. Other studies have previously reported AH incidences of 1.1 54 cases per 100,000 per year in the Netherlands from 1987-19907, 3.4 cases per 55 56 100,000 per year in Sweden from 2000-20078, and 6.9 per 100,000 per year in 57 58 New Zealand from 1977-19869. The presented incidences are in accordance with 59 60 our results. While the overall incidence of AH was increasing throughout our

8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 22

1 2 3 study period from 1994 to 2016, we did not observe a significant increase in the 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 incidence of primary AH. The increase in AH incidence was mainly driven by a 6 7 higher incidence of patients with another recorded primary diagnosis, and AH 8 9 being recorded as a secondary diagnosis. This could either be a result of an 10 actual increase in incidence of AH secondary to chronic disease, or it could be a 11 12 result of increased clinician awareness towards diagnosing AH correctly as a 13 14 secondary diagnosis. 15 16 In our cohort, the incidence of AH is low during summer and increases during 17 the winter months, which likely is explained by the country’s temperate climate 18 For peer review only 19 with a cold fall and winter. This is in accordance with findings from other studies 20 21 of AH in cold climate8,10-13. Due the limited elevation and temperate climate in 22 23 Denmark, avalanches do not occur, which is in contrast to AH cohorts presented 24 from countries with mountainous terrain. 25 26 In accordance with previous findings, more males were admitted with AH10,14- 27 28 17, but females had a higher fatality in univariable analyses. However, females 29 8 30 with AH were significantly older than men , and had a significantly higher 31 frequency of secondary AH compared to primary AH, i.e. a higher comorbidity 32 33 burden. After adjustment for known confounding/mediating factors, including 34 35 age, AH type, and calendar year, females had a significantly lower fatality

36 http://bmjopen.bmj.com/ 37 compared to men. Interaction analyses and following stratification into age 38 groups revealed, however, that this difference in fatality between female and 39 40 male sex was driven by a lower female fatality, only in patients older than 80 41 42 years of age. We found no association between female sex and fatality in the 43 other age groups. Whether the lower female fatality in patients older than 80

44 on September 25, 2021 by guest. Protected copyright. 45 46 years is driven by unmeasured confounding/mediation, or whether it reflects the 47 longer life expectancy in females is unknown. 48 49 The AH diagnosis was associated with a high comorbidity burden and having 50 51 secondary AH (i.e. other recorded diagnoses) was associated with a significantly 52 53 higher fatality compared to having primary AH. This corresponds well with 54 previous data, suggesting a high comorbidity burden in patients admitted with 55 56 cardiac arrest and AH18. However, even in patients with primary AH, the one- 57 58 year fatality was 22%. Kaplan-Meier curves showed that the rate of death was 59 60

9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 22 BMJ Open

1 2 3 highest the first 30 days after admission, where after the rate of death was 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 relatively constant out to one year. 6 7 The following limitations should be considered. Primarily, data originated 8 9 from national registers, which limited the data granularity to what has been 10 presented in the present paper, and we were not able to establish for example 11 12 the specific core temperatures at hospital admission, specific cause of AH, etc. 13 14 While the AH diagnosis likely has a high specificity, a lower sensitivity may have 15 16 resulted in underreporting of patients with hypothermia, especially in presence 17 of competing diagnoses. Accordingly, the incidence of secondary AH may be 18 For peer review only 19 higher than here reported. While diagnoses were classified by treating 20 21 physicians as primary or secondary, it is unknown to which extent this 22 23 classification corresponds to the definition of primary AH (AH from isolated cold 24 exposure) and secondary AH (AH in relation to acute or chronic illness)2. 25 26 However, the group of patients with AH recorded as the only diagnosis should 27 28 correspond to having primary hypothermia. The exclusions of patients with 29 30 more than one admission with AH recorded as the diagnosis could result in 31 underestimation of the incidence during the study period. 32 33 34 35 Conclusion

36 http://bmjopen.bmj.com/ 37 The incidence of AH was 4.4 per 100.000 per year in the period from 1995 to 38 2016. The AH diagnosis was associated with a high comorbidity burden, which 39 40 was more pronounced in women. Alcohol intoxication was the most prevalent 41 42 associated diagnosis. Overall, AH patients had a one-year fatality of 27%. In 43 otherwise healthy patients with primary AH, the one-year fatality was 22%.

44 on September 25, 2021 by guest. Protected copyright. 45 46 47 Ethics approval and consent to participate 48 49 The study was approved by the Danish Patient Safety Authority and was 50 51 conducted in accordance with the Helsinki Declaration. As we conducted a 52 53 national retrospective study over a time period of 21 years, the Danish Patient 54 Safety Authority approved the study without need for obtaining consent from 55 56 each individual. 57 58 59 60 Availability of data and materials

10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 22

1 2 3 Data was leveraged from national Danish registries. Accordingly, assess to data 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 can be obtained by application to the relevant Danish authorities. 6 7 8 9 Authors contributions 10 Author MW initiated the project. All authors were involved in planning the study 11 12 including the statistical analyses. Authors SW and AFM wrote the first draft of 13 14 the manuscript. All authors were involved in the production of the final 15 16 manuscript. 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 25, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 22 BMJ Open

1 2 3 References 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 1. van der Ploeg G-J, Goslings JC, Walpoth BH, Bierens JJLM. Accidental 6 hypothermia: Rewarming treatments, complications and outcomes from 7 one university medical centre. Resuscitation. 81, 1550-1555 (2010). 8 9 10 2. Brown DJA, Brugger H, Boyd J, Paal P. Accidental Hypothermia. N Engl J 11 Med. 376, 1930-1938 (2012). 12 13 3. Durrer B, Brugger H, Syme D. The Medical On-site Treatment of 14 Hypothermia* ICAR-MEDCOM Recommendation. High Altitude Medicine 15 Biology. 4, 99-103 (2003). 16 17 4. Truhlář A, Deakin CD, Soar J, et al. European Resuscitation Council 18 For peer review only 19 Guidelines for Resuscitation 2015: Section 4. Cardiac arrest in special 20 circumstances. Resuscitation. 95, 148-201 (2015). 21 22 5. Pasquier M, Zurron N, Weith B, et al. Deep Accidental Hypothermia with 23 Core Temperature Below 24°C Presenting with Vital Signs. High Altitude 24 Medicine Biology. 15, 58-63 (2014). 25 26 6. Thygesen SK, Christiansen CF, Christensen S, Lash TL, Sørensen HT. The 27 28 predictive value of ICD-10 diagnostic coding used to assess Charlson 29 comorbidity index conditions in the population-based Danish National 30 Registry of Patients. BMC Medical Research Methodology. 11, 1-6 (2011). 31 32 7. Bierens JJ, Uitslager R, Swenne-van Ingen MM, van Stiphout WA, Knape JT. 33 Accidental hypothermia: incidence, risk factors and clinical course of 34 patients admitted to hospital. Eur J Emerg Med. 2, (1995). 35

36 http://bmjopen.bmj.com/ 37 8. Brändström H, Johansson G, Giesbrecht G, Ängquist K-A, Haney MF. 38 Accidental cold-related injury leading to hospitalization in northern 39 Sweden: an eight-year retrospective analysis. Scandinavian Journal of 40 Trauma, Resuscitation and Emergency Medicine. 22, 1-7 (2014). 41 42 9. Taylor NA, Griffiths RF, Cotter JD. Epidemiology of hypothermia: fatalities 43 and hospitalisations in New Zealand. Australian and New Zealand Journal

44 on September 25, 2021 by guest. Protected copyright. 45 of Medicine. 24, 705-710 (1994). 46 47 10. Baumgartner EA, Belson M, Rubin C, Patel M. Hypothermia and Other Cold- 48 Related Morbidity Emergency Department Visits: United States, 1995– 49 2004. Wilderness Environ Med. 19, 233-236 (2008). 50 51 11. Zhang P, Wiens K, Wang R, et al. Cold Weather Conditions and Risk of 52 53 Hypothermia Among People Experiencing Homelessness: Implications for 54 Prevention Strategies. IJERPH. 16, 3259-3259 (2019). 55 56 12. Danzl DF, Pozos RS, Auerbach PS, et al. Multicenter Hypothermia Survey. 57 Annals of Emergency Medicine. 16, 1042-1055 (1987). 58 59 13. The Eurowinter Group. Cold exposure and winter mortality from 60

12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 22

1 2 3 ischaemic heart disease, cerebrovascular disease, respiratory disease, and 4 all causes in warm and cold regions of Europe. The Lancet. 349, 1341- BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 6 1346 (1997). 7 8 14. Rango N. Exposure-Related Hypothermia Mortality in the United States, 9 1970-79. American Journal of Public Health. 74, 1159-1160 (1984). 10 11 15. Hsieh T-M, Kuo P-J, Hsu S-Y, Chien P-C, Hsieh H-Y, Hsieh C-H. Effect of 12 Hypothermia in the Emergency Department on the Outcome of Trauma 13 14 Patients: A Cross-Sectional Analysis. IJERPH. 15, 1769-1811 (2018). 15 16 16. Svendsen ØS, Grong K, Andersen SK, Husby P. Outcome After Rewarming 17 From Accidental Hypothermia by Use of Extracorporeal Circulation. ATS. 18 103, 920-925For (2016). peer review only 19 20 17. Debaty G, Moustapha I, Bouzat P, et al. Outcome after severe accidental 21 hypothermia in the French Alps: A 10-year review. Resuscitation. 93, 118- 22 23 123 (2015). 24 25 18. Ohbe H, Isogai S, Jo T, Matsui H, Fushimi K, Yasunaga H. Extracorporeal 26 membrane oxygenation improves outcomes of accidental hypothermia 27 without vital signs: A nationwide observational study. Resuscitation. 144, 28 27-32 (2019). 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 25, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 22 BMJ Open

1 2 3 Table 1. List of associated ICD-10 diagnoses with a prevalence higher than 1% in 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 patients with accidental hypothermia 6 7 List of secondary ICD-10 diagnoses in patients with List of primary ICD-10 diagnoses in patients with 8 secondary AH, type I (n= 1336) secondary AH, type II (n= 1676) 9 10 Diagnosis ICD-10 n (%) Diagnosis ICD-10 n (%) 11 Acute alcohol intoxication DF100 198 (15) Acute alcohol intoxication DF100 153 (9.1) 12 Alcohol dependence syndrome DF102 118 (8.9) Respiratory insufficiency DJ969 74 (4.4) 13 14 Harmful use of alcohol DF101 96 (7.2) Sepsis DA419 61 (3.6) 15 Pneumonia DJ189 87 (6.5) Acute respiratory insufficiency DJ960 58 (3.5) 16 Dementia DF039 74 (5.5) Hypoglycemia DE162 55 (3.3) 17 18 Hypertension For peerDI109 review59 (4.4) Pneumonia only DJ189 48 (2.9) 19 Atrial fibrillation or flutter DI489 57 (4.3) Dehydration DE869 33 (2.0) 20 Dehydration DE869 52 (3.9) Drowning / nonfatal submersion DT751 32 (1.9) 21 22 Traumatic ischemia of muscle DT796 29 (2.2) Cardiac arrest DI469 28 (1.7) 23 Type II diabetes DE119 27 (2.0) Concussion DS060 28 (1.7) 24 COPD DJ449 24 (1.8) Multiple lesions DT079 29 (1.7) 25 26 Hypoglycemia DE162 23 (1.7) Stroke DI649 27 (1.6) 27 Anemia DD649 21 (1.6) Traumatic ischemia of muscle DT796 25 (1.5) 28 Hypokalemia DE876 21 (1.6) Alcohol dependence syndrome DF102 20 (1.2) 29 30 Cardiac arrest DI469 22 (1.6) Harmful use of alcohol DF101 17 (1.0) 31 Complication after stroke DI694 22 (1.6) Septic shock DR572 17 (1.0) 32 Bakterial pneumonia DJ159 20 (1.5) - - - 33 34 Urinary tract infaction DN390 20 (1.5) - - - 35 Sepsis DA419 16 (1.2) - - -

36 Type I diabetes DE109 16 (1.2) - - - http://bmjopen.bmj.com/ 37 38 Epilepsia DG409 16 (1.2) - - - 39 Hypotension DI959 15 (1.1) - - - 40 Skizofrenia DF209 14 (1.0) - - - 41 42 Secondary AH, type I: Having AH recorded as primary diagnosis with one 43 or more recorded secondary diagnoses

44 on September 25, 2021 by guest. Protected copyright. 45 Secondary AH type II: Having AH recorded as a secondary diagnosis with 46 47 another recorded primary diagnosis 48 49 50 51 52 53 54 55 56 57 58 59 60

14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 22

1 2 3 Table 2. Associations between covariates and time to death within 365 days, 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 analyzed by Cox proportional hazard models. 6 7 Univariable models Multivariable model1 8 Hazard ratios p p Hazard ratios (95%CI) 9 (95%CI) 10 2 11 AH type secondary type I vs. primary 1.2 (1.1 - 1.4) < 0.001 1.2 (1.0 - 1.3) 0.04 12 secondary type II vs. primary 1.7 (1.5 - 2.0) < 0.001 1.7 (1.5 - 2.0) < 0.001 13 14 15 Sex female vs. male 1.3 (1.2 - 1.4) < 0.001 0.89 (0.80 - 0.99) 0.04 16 17 Age at ≤ 20 years ref. ref. < 0.001 18 admission For peer review only 19 21-40 years 3.5 (1.1 - 11) 0.03 3.4 (1.0 - 11) 0.04 20 21 41-60 years 15 (4.7 - 45) < 0.001 14 (4.4 - 43) < 0.001 22 61-80 years 24 (7.6 - 74) <0.001 24 (7.6 - 73) < 0.001 23 > 80 years 36 (12 - 113) <0.001 38 (12 - 117) < 0.001 24 25 26 Calendar year per year 1.0 (0.99 - 1.0) 0.18 0.98 (0.98 - 0.99) < 0.001 27 1: Adjusted for AH type, sex, age, and calendar year. 28 29 2: Accidental hypothermia (AH) 30 31 Primary AH: Having AH recorded as primary diagnosis without any 32 33 recorded secondary diagnoses 34 35 Secondary AH type I: Having AH recorded as primary diagnosis with one

36 or more recorded secondary diagnoses http://bmjopen.bmj.com/ 37 38 Secondary AH type II: Having AH recorded as a secondary diagnosis with 39 40 another recorded primary diagnosis 41 42 43

44 on September 25, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 22 BMJ Open

1 2 3 Figure Legends 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 6 7 Figure 1. incidence of accidental hypothermia per 100,000 inhabitants per 8 9 calendar year. 10 11 12 Accidental hypothermia (AH) 13 14 Primary AH: Having AH recorded as primary diagnosis without any recorded 15 16 secondary diagnoses 17 Secondary AH, type I: Having AH recorded as primary diagnosis with one or 18 For peer review only 19 more recorded secondary diagnoses 20 21 Secondary AH, type II: Having Ah recorded as a secondary diagnosis with 22 23 another recorded primary diagnosis 24 25 26 Figure 2. Distribution of accidental hypothermia diagnoses per calendar month 27 28 per percent. 29 30 31 32 Figure 3. One-year survival in accidental hypothermia patients stratified by AH 33 34 type (a.), male versus female (b.), and by age at admission (c.). Numbers at risk 35 presented for 0, 6, and 12 months.

36 http://bmjopen.bmj.com/ 37 38 39 Accidental hypothermia (AH) 40 41 Primary AH: Having AH recorded as primary diagnosis without any recorded 42 secondary diagnoses 43

44 Secondary AH, type I: Having AH recorded as primary diagnosis with one or on September 25, 2021 by guest. Protected copyright. 45 46 more recorded secondary diagnoses 47 48 Secondary AH, type II: Having Ah recorded as a secondary diagnosis with 49 another recorded primary diagnosis 50 51 52 53 54 55 56 57 58 59 60

16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 22 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 Figure 1. incidence of accidental hypothermia per 100,000 inhabitants per calendar year. 27 28 Accidental hypothermia (AH) 29 Primary AH: Having AH recorded as primary diagnosis without any recorded secondary diagnoses 30 Secondary AH, type I: Having AH recorded as primary diagnosis with one or more recorded secondary diagnoses 31 Secondary AH, type II: Having Ah recorded as a secondary diagnosis with another recorded primary

32 diagnosis http://bmjopen.bmj.com/ 33 34 35 158x95mm (300 x 300 DPI) 36 37 38 39

40 on September 25, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 22 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 Figure 2. Distribution of accidental hypothermia diagnoses per calendar month per percent. 27 28 136x81mm (300 x 300 DPI) 29 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 25, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 22 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 25, 2021 by guest. Protected copyright. 41 42 43 44 45 Figure 3. One-year survival in accidental hypothermia patients stratified by AH type (a.), male versus female 46 (b.), and by age at admission (c.). Numbers at risk presented for 0, 6, and 12 months. 47 Accidental hypothermia (AH) 48 Primary AH: Having AH recorded as primary diagnosis without any recorded secondary diagnoses 49 Secondary AH, type I: Having AH recorded as primary diagnosis with one or more recorded secondary 50 diagnoses 51 Secondary AH, type II: Having Ah recorded as a secondary diagnosis with another recorded primary 52 diagnosis 53 54 187x292mm (300 x 300 DPI) 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from

Page 21 of 22 BMJ Open

1 2 Competing interests: None 3 4 Funding: The study was funded by Trygfonden - Accidental Hypothermia (ID: 117579). The sponsor had no involvement in study design, 5 6 collection, analysis and interpretation of data, writing of the manuscript or decision to submit the manuscript for publication. 7 STROBE 2007 (v4) Statement—Checklist of items that should be included in reports of cohort studies 8 9 10 Item Section/Topic Recommendation Reported on page # 11 # 12 Title and abstract 1 (a) IndicateFor the study’s design peer with a commonly review used term in the title or the onlyabstract 1, 2 13 14 (b) Provide in the abstract an informative and balanced summary of what was done and what was found 2 15 16 Introduction http://bmjopen.bmj.com/ 17 Background/rationale 2 Explain the scientific background and rationale for the investigation being reported 4 18 19 Objectives 3 State specific objectives, including any prespecified hypotheses 4 20 Methods 21 Study design 4 Present key elements of study design early in the paper 4,5 22 23 Setting 5 Describe the setting, locations, and relevant dates, including periods of recruitment, exposure, follow-up, and data 4,5 24 collection on September 25, 2021 by guest. Protected copyright. 25 Participants 6 (a) Give the eligibility criteria, and the sources and methods of selection of participants. Describe methods of follow-up 4,5 26 27 (b) For matched studies, give matching criteria and number of exposed and unexposed 28 29 Variables 7 Clearly define all outcomes, exposures, predictors, potential confounders, and effect modifiers. Give diagnostic criteria, if 4,5 30 applicable 31 Data sources/ 8* For each variable of interest, give sources of data and details of methods of assessment (measurement). Describe 4,5 32 33 measurement comparability of assessment methods if there is more than one group 34 Bias 9 Describe any efforts to address potential sources of bias 10 35 Study size 10 Explain how the study size was arrived at 4 36 Quantitative variables 11 Explain how quantitative variables were handled in the analyses. If applicable, describe which groupings were chosen and 6 37 38 why 39 Statistical methods 12 (a) Describe all statistical methods, including those used to control for confounding 6 40 41 (b) Describe any methods used to examine subgroups and interactions 6 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from

BMJ Open Page 22 of 22

1 2 (c) Explain how missing data were addressed 6 3 (d) If applicable, explain how loss to follow-up was addressed 6 4 5 (e) Describe any sensitivity analyses N/A 6 Results 7 8 Participants 13* (a) Report numbers of individuals at each stage of study—eg numbers potentially eligible, examined for eligibility, confirmed 6 9 eligible, included in the study, completing follow-up, and analysed 10 (b) Give reasons for non-participation at each stage N/A 11 (c) Consider use of a flow diagram N/A 12 For peer review only 13 Descriptive data 14* (a) Give characteristics of study participants (eg demographic, clinical, social) and information on exposures and potential 6 14 confounders 15 (b) Indicate number of participants with missing data for each variable of interest 6 16 http://bmjopen.bmj.com/ 17 (c) Summarise follow-up time (eg, average and total amount) 6 18 Outcome data 15* Report numbers of outcome events or summary measures over time 7 19 Main results 16 (a) Give unadjusted estimates and, if applicable, confounder-adjusted estimates and their precision (eg, 95% confidence 7,8 20 interval). Make clear which confounders were adjusted for and why they were included 21 22 (b) Report category boundaries when continuous variables were categorized 8 23 (c) If relevant, consider translating estimates of relative risk into absolute risk for a meaningful time period N/A on September 25, 2021 by guest. Protected copyright. 24 Other analyses 17 Report other analyses done—eg analyses of subgroups and interactions, and sensitivity analyses 8 25 26 Discussion 27 Key results 18 Summarise key results with reference to study objectives 8 28 Limitations 29 30 Interpretation 20 Give a cautious overall interpretation of results considering objectives, limitations, multiplicity of analyses, results from 9 31 similar studies, and other relevant evidence 32 Generalisability 21 Discuss the generalisability (external validity) of the study results 9 33 34 Other information 35 Funding 22 Give the source of funding and the role of the funders for the present study and, if applicable, for the original study on 3 36 which the present article is based 37 38 39 *Give information separately for cases and controls in case-control studies and, if applicable, for exposed and unexposed groups in cohort and cross-sectional studies. 40 41 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from

Page 23 of 22 BMJ Open

1 2 Note: An Explanation and Elaboration article discusses each checklist item and gives methodological background and published examples of transparent reporting. The STROBE 3 checklist is best used in conjunction with this article (freely available on the Web sites of PLoS Medicine at http://www.plosmedicine.org/, Annals of Internal Medicine at 4 5 http://www.annals.org/, and Epidemiology at http://www.epidem.com/). Information on the STROBE Initiative is available at www.strobe-statement.org. 6 7 8 9 10 11 12 For peer review only 13 14 15 16 http://bmjopen.bmj.com/ 17 18 19 20 21 22 23 24 on September 25, 2021 by guest. Protected copyright. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from

Investigating the Incidence of Accidental Hypothermia and the Associations with Outcomes in Denmark – a Nationwide Cohort Study ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2020-046806.R1

Article Type: Original research

Date Submitted by the 18-Mar-2021 Author:

Primary Subject Emergency medicine Heading:

Secondary Subject Heading: Epidemiology http://bmjopen.bmj.com/

ACCIDENT & EMERGENCY MEDICINE, EPIDEMIOLOGY, GENERAL Keywords: MEDICINE (see Internal Medicine)

on September 25, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 24 BMJ Open

1 2 3

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 25, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 24

1 2 3 1 Investigating the Incidence of Accidental Hypothermia and the Associations 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 with Outcomes in Denmark – a Nationwide Cohort Study 6 7 3 8 9 4 Authors 10 5 Sebastian Wiberg* MD PhD1, Asmus Friborg Mortensen* MD2, Jesper Kjaergaard, 11 12 6 MD PhD DMSc1, Christian Hassager MD DMSc1, Michael Wanscher MD PhD3. 13 14 7 15 16 8 Affiliations 17 9 1) Department of Cardiology, The Heart Centre, Copenhagen University 18 For peer review only 19 10 Hospital Rigshospitalet 20 21 11 2) Department of Anaesthesia and Intensive Care, Bispebjerg and 22 23 12 Frederiksberg Hospital, University of Copenhagen, Denmark 24 13 3) Department of Cardiothoracic Anaesthesiology, The Heart Centre, 25 26 14 Copenhagen University Hospital Rigshospitalet, Denmark 27 28 15 29 30 16 Corresponding Author 31 17 Sebastian Wiberg, MD, PhD 32 33 18 Department of Cardiology 34 35 19 The Heart Centre

36 http://bmjopen.bmj.com/ 37 20 Copenhagen University Hospital Rigshospitalet 38 21 [email protected] 39 40 22 41 42 23 *) Equal author contribution 43 24

44 on September 25, 2021 by guest. Protected copyright. 45 46 25 Research Question 47 26 To investigate the reported incidence of accidental hypothermia in a nationwide 48 49 27 registry, and to further investigate the associations between accidental 50 51 28 hypothermia and the associated outcomes. 52 53 29 54 30 Study Design 55 56 31 Retrospective cohort study 57 58 32 59 60 33

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 24 BMJ Open

1 2 3 1 Abstract 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 Objectives 6 7 3 To investigate the reported incidence of accidental hypothermia (AH) in a 8 9 4 nationwide registry, and to further investigate the associations between 10 5 accidental hypothermia and the associated outcomes. 11 12 6 13 14 7 Design 15 16 8 Nationwide retrospective cohort study 17 9 18 For peer review only 19 10 Participants and settings 20 21 11 All patients above 18 years admitted to hospitals in Denmark with a diagnosis of 22 23 12 AH from January 1996 to November 2016. Patients were identified as all patients 24 13 admitted with an ICD-10 code of T689. Other recorded diagnoses were included 25 26 14 in the analyses. 27 28 15 29 30 16 Primary and secondary outcome measures 31 17 The primary outcome was one-year mortality. 32 33 18 34 35 19 Results

36 http://bmjopen.bmj.com/ 37 20 Throughout the inclusion period, 5242 patients were admitted with a diagnosis 38 39 21 of AH, corresponding to a mean annual incidence of 4.4  1.2 (low to high annual 40 22 incidence per calendar year: 2.9 – 6.4) per 100,000 inhabitants. A total of 2230 41 42 23 (43%) had AH recorded as the primary diagnosis without any recorded 43 24 secondary diagnoses (Primary AH), 1336 (25%) had AH recorded as the primary

44 on September 25, 2021 by guest. Protected copyright. 45 46 25 diagnosis with other recorded secondary diagnoses (Secondary AH, type I), and 47 26 1676 (32%) had AH recorded as a secondary diagnosis with another recorded 48 49 27 primary diagnosis (Secondary AH, type II). Alcohol intoxication was the most 50 51 28 common diagnosis associated with AH. Overall one-year mortality was 27%. In 52 53 29 patients with primary AH, one-year mortality was 22%, compared to 26% in 54 30 patients with secondary AH type I, and 35% in patients with secondary AH type 55 56 31 II (plog-rank < 0.001). 57 58 32 59 60 33 Conclusions

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 24

1 2 3 1 The present study investigated the incidence of AH, associated comorbidities, 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 and mortality after AH in Denmark from 1995 to 2016. The diagnosis was 6 7 3 associated with a high comorbidity burden; the most common being alcohol 8 9 4 intoxication. Further, the diagnosis was associated with a high one-year 10 5 mortality of 27%. Future prospective studies are needed to identify modifiable 11 12 6 factors associated with both incidence of AH and outcome after AH. 13 14 7 15 16 8 Article Summary 17 9 Strengths and Limitations of this study 18 For peer review only 19 10 - Large contemporary study to investigate the incidence, comorbidity 20 21 11 burden and outcome in patients diagnosed with AH. 22 23 12 - Data originated from a national cohort of patients. 24 13 - The registry data limited the data granularity to what has been presented 25 26 14 in the present paper. 27 28 15 - The presented incidences are likely to be significantly different in other 29 30 16 countries, depending on climate, demography, etc. 31 17 32 33 18 Funding 34 35 19 The study was funded by Trygfonden - Accidental Hypothermia (ID: 117579).

36 http://bmjopen.bmj.com/ 37 20 The sponsor had no involvement in study design, collection, analysis and 38 21 interpretation of data, writing of the manuscript or decision to submit the 39 40 22 manuscript for publication. 41 42 23 43 24 Competing interests

44 on September 25, 2021 by guest. Protected copyright. 45 46 25 None 47 26 48 49 27 50 51 28 52 53 29 54 30 55 56 31 57 58 32 59 60 33

3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 24 BMJ Open

1 2 3 1 Introduction 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 Accidental hypothermia (AH), defined as an involuntary drop in core 6 7 3 temperature below 35C, is a serious condition with considerable mortality [1,2]. 8 9 4 AH can result from isolated cold exposure (primary hypothermia) or in relation 10 5 to acute or chronic illness (secondary hypothermia) [2]. The severity of AH can 11 12 6 be graded into four stages based on core temperature: 32°C -35°C (Stage I, mild 13 14 7 AH), 28°C - < 32°C (Stage II, moderate AH), and < 28°C (Stage III, severe AH). 15 16 8 Stage IV is characterized by an accentuated drop of core temperature and 17 9 absence of vital signs [3–5]. In-hospital mortality after severe hypothermia has 18 For peer review only 19 10 been reported to be as high as 30%. It is estimated that approximately 1500 20 21 11 people die from hypothermia every year in the United States [1,4]. Few previous 22 23 12 studies have investigated the incidence and outcome after AH, however, the 24 13 25 studies represent heterogenous populations from different climate zones and 26 14 different time periods[6–9]. Accordingly, our knowledge of the contemporary 27 28 15 incidence, demography and outcomes after AH remains sparse. 29 30 16 The aim of this study was to investigate the reported incidence of AH in a 31 32 17 Danish nationwide registry, and to further investigate the associations between 33 18 AH and comorbidities, and the associated outcomes in adult patients admitted 34 35 19 with AH.

36 http://bmjopen.bmj.com/ 37 20 Geography 38 2 39 21 Denmark covers a relatively small area of 43,000m but has a comparatively 40 22 long coastline of approximately 7300km. The highest point is 173 meters above 41 42 23 sea level. It is located on the 55th to 57th parallel north in the temperate climate 43 24 zone. Mean temperatures ranges from 0C in January and February to 15.7C in 44 on September 25, 2021 by guest. Protected copyright. 45 st 46 25 August. The population was about 5.8 million as of January 1 , 2016. About 1.8 47 26 million (about 31% of the population) lived in the capital,Copenhagen as of 48 49 27 January 1st, 2016. 50 51 28 52 53 29 Methods 54 30 The present study is a nationwide retrospective cohort study including all adult 55 56 31 patients admitted to hospitals in Denmark with a diagnosis of AH from January 57 58 32 1996 to November 2016. The primary outcome was one-year mortality after 59 60 33 diagnosis. We chose a priori to include only those cases with a first-time

4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 24

1 2 3 1 diagnosis of AH. The rationale behind this decision was to ensure that a patient 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 with multiple hospital admissions for AH only counted once in the analysis. 6 7 3 All Danish residents are given a personal identification number at the time of 8 9 4 birth or immigration. This number is used in all contacts with the health care 10 5 system, and it is used as identifier in national registries. We extracted data from 11 12 6 the Danish National Patient Register (NPR) to identify all patients admitted with 13 14 7 a discharge diagnosis of AH, defined by the International Classification of 15 th 16 8 Diseases, 10 edition (ICD-10, which has been used in Denmark since 1993) code 17 9 T689. 18 For peer review only 19 10 We stratified patients into three groups based on whether AH was recorded as 20 21 11 the only diagnosis (Primary AH), whether AH was recorded as the primary 22 23 12 diagnosis with other recorded secondary diagnoses (‘AH + 2 diagnosis’), or 24 25 13 whether AH was recorded as a secondary diagnosis (‘1 diagnosis + AH’) for a 26 14 given admission. 27 28 15 The primary diagnosis is defined as the diagnosis primarily causing the need 29 30 16 for the admission, while the secondary diagnoses are defined as other conditions 31 32 17 that required attention during a given admission. Hospitals are reimbursed 33 18 based on correct reporting of diagnoses, and accordingly, the registry has a high 34 35 19 accuracy for most diagnoses[10]. Main treatment categories, such as ‘intensive

36 http://bmjopen.bmj.com/ 37 20 care admission’, and ‘respirator treatment’ were also extracted from the NPR. 38 39 21 The AH diagnosis was linked to survival (i.e. the primary outcome) using the 40 22 nationwide Danish Register of Causes of Death (RCD). The study was approved 41 42 23 by the Danish Patient Safety Authority Authority (Institutional review board, ref. 43

44 24 no 3-3013-1906/1/) and was conducted in accordance with the Helsinki on September 25, 2021 by guest. Protected copyright. 45 46 25 Declaration. 47 26 48 49 27 Patient and Public Involvement 50 51 28 Patients or the public were not involved in the design of the present study. 52 53 29 54 30 Statistical analyses 55 56 31 Throughout, categorical variables are presented as counts (%), normally 57 58 32 distributed continuous variables are presented as mean ± standard deviation 59 th 60 33 (SD) and skewed continuous variables are presented as median (25 percentile -

5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 24 BMJ Open

1 2 3 1 75th percentile). Normality was assessed visually by QQ plots prior to analyses. 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 Associations between baseline variables and AH types were tested by the Chi- 6 7 3 Square test for categorical variables and with the one-way analysis of variance 8 9 4 (ANOVA) for continuous variable. 10 5 To give an overview of common comorbidity, we presented a list of associated 11 12 6 ICD-10 diagnoses with a prevalence higher than 1% in patients with accidental 13 14 7 hypothermia. This overview was stratified by whether AH was recorded as the 15 16 8 primary diagnosis (‘AH + 2 diagnosis’) or whether AH was recorded as a 17 9 secondary diagnosis (‘1 diagnosis + AH’) (Table 1). 18 For peer review only 19 10 The annual incidence of AH per calendar year was plotted for the total cohort 20 21 11 as well as after stratification by AH type (Figure 1). We presented the mean 22 23 12 annual incidence  SD with range. To assess if the incidence of AH was changing 24 25 13 over time, we applied linear models with calendar year being included as a 26 14 continuous covariate assuming a linear trend, and we presented the absolute 27 28 15 change in AH incidence per calendar year  standard deviation. The distribution 29 30 16 of AH diagnoses per calendar months were presented along with mean outside 31 32 17 temperatures (Figure 2). Furthermore, we presented the mean annual incidence 33 34 18  SD of AH after stratification by age group. 35 19 We presented crude 365-days mortality rates for the total cohort, as well as

36 http://bmjopen.bmj.com/ 37 20 after stratification by AH type, by sex, and by age group (Table 2). Furthermore, 38 39 21 we presented 365-days mortality after stratification by sex and age group. 40 41 22 Differences in mortality between females and males, stratified by age group, 42 23 were analyzed by application of the Chi-Square test (Table 3). 43

44 24 For each patient in the cohort, the date of entry (i.e. the date of AH diagnosis) on September 25, 2021 by guest. Protected copyright. 45 46 25 as well as the date of death was recorded in the registries. The Kaplan-Meier 47 48 26 estimator was applied to visualize time to death within 365 days from admission, 49 27 stratified by AH type, sex, and age group. The log-rank test was applied to assess 50 51 28 possible differences between strata (Figure 3). 52 53 29 To analyze time to death between groups, we applied Cox proportional hazard 54 55 30 models adjusting for AH type, sex, age at admission, and calendar year. We 56 31 presented hazard ratios with 95% confidence intervals (Table 2). A significance 57 58 32 level of < 0.05 was applied throughout. SAS software, version 9.4 (SAS Institute, 59 60 33 Cary, North Carolina, USA) was used for all statistical analysis.

6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 24

1 2 3 1 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 Results 6 7 3 Over the inclusion period of almost 22 years, the Danish Health Care system 8 9 4 received a total of 5242 adult patients admitted with a diagnosis of AH. A total of 10 5 2230 (43%) had AH recorded as the final primary diagnosis without any 11 12 6 recorded secondary diagnoses (Primary AH), 1336 (25%) had AH recorded as 13 14 7 the final primary diagnosis with other recorded secondary diagnoses (‘AH + 2 15 16 8 diagnosis’), and 1676 (32%) had AH recorded as a final secondary diagnosis with 17 9 another recorded primary diagnosis (‘1 diagnosis + AH’). 18 For peer review only 19 10 A total of 1981 (38%) of the patients admitted with AH were female, and 20 21 11 mean age was 6121 years. Female sex was less prevalent among patients with 22 23 12 primary AH compared to patients with ‘AH + 2 diagnosis’ and patients with ‘1 24 25 13 diagnosis + AH’ (35% vs. 39% vs. 40%, p< 0.001). Patients with primary AH were 26 27 14 younger compared to patients with ‘AH + 2 diagnosis’ and patients with ‘1 28 15 diagnosis + AH’ (5922 years vs. 6220 years vs. 6220 years, p < 0.001). 29 30 16 31 32 17 Comorbidity 33 34 18 Both ‘AH + 2 diagnosis’ and ‘1 diagnosis + AH’ were associated with a 35 19 significant comorbidity burden. A list of associated diagnoses with a proportion

36 http://bmjopen.bmj.com/ 37 20 higher than 1% is provided in Table 1. The most frequent diagnosis associated 38 39 21 with AH was alcohol intoxication (Table 1). 40 41 22 A total of 816 (16%) patients with AH received care in an intensive care unit. 42 43 23 A total of 305 (5.8%) were treated with vasopressors and/or inotropic agents,

44 24 397 (7.6%) received ventilator treatment, and 479 (9.1) received dialysis. A total on September 25, 2021 by guest. Protected copyright. 45 46 25 of 21 (0.4%) patients were treated with extracorporeal circulation. 47 48 26 49 50 27 Incidence 51 28 The mean annual incidence of AH was 4.4  1.2 per 100,000 inhabitants with a 52 53 29 range from 2.9 to 6.4 per 100,000 inhabitants (Figure 1). The overall incidence of 54 55 30 AH increased from 2000 to 2016 (absolute increase 0.160.02 per 100,000 56 57 31 inhabitants per calendar year, p < 0.001). 58 59 60

7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 24 BMJ Open

1 2 3 1 The annual incidence of AH was 4.3  1.6 per 100,000 in patients from 18 to 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 20 years, 2.3  0.55 per 100,000 in patients from 21 to 40 years, 4.8  1.1 per 6 7 3 100,000 in patients from 41 to 60 years, 7.2  2.6 in patients from 61 to 80 years, 8 9 4 and 28  11 in patients older than 80 years. 10 11 5 The incidence of AH followed a seasonal trend, with the highest percent of AH 12 6 occurring in January (16%) and the lowest percent of AH occurring in August 13 14 7 (3.0%, Figure 2). Of all AH diagnoses, 29% were recorded in the Danish capital 15 16 8 region inhabited by 31% of the total Danish population. 17 18 9 For peer review only 19 10 Outcome 20 21 11 Overall mortality within the first 7 days of admission was 11%, increasing to 22 23 12 16% at 30 days, and 27% at one year. One-year mortality was 22% in patients 24 25 13 with primary AH, 26% in patients with ‘AH + 2 diagnosis’, and 35% in patients 26 27 14 with ‘1 diagnosis + AH’ (p < 0.001, Figure 3). The higher mortality in patients 28 15 with comorbidity remained unchanged after adjustment for confounding factors 29 30 16 (Table 2). 31 32 17 Increasing age was significantly associated with increased one-year mortality 33 34 18 (Table 2). One-year mortality increased from 1.6% in patients between 18 and 35 19 20 years up to 45% in patients older than 80 years (p < 0.001, Table 2).

36 http://bmjopen.bmj.com/ 37 20 One-year mortality was higher in females compared to males (31% vs. 25%, p 38 39 21 < 0.001, Table 2), corresponding to a HR of 1.3 (1.2 – 1.4). In contrast, we found a 40 41 22 significantly lower mortality in females older than 80 years compared to men 42 23 older than 80 years (41% vs. 52%, p < 0.001), but no significant differences 43

44 24 between the sexes in other age groups (Table 3). A significantly higher on September 25, 2021 by guest. Protected copyright. 45 46 25 proportion of females were older than 80 years compared to men (39% vs. 14%, 47 48 26 p < 0.001, Table 3). Accordingly, the higher overall mortality in women was 49 50 27 caused by a higher proportion of women compared to men being older than 80 51 28 years. In accordance, female sex was associated with a lower mortality compared 52 53 29 to male sex after adjustment for confounding factors including age (HR 0.89 54 55 30 95%CI 0.80 – 0.99, Table 2). 56 57 31 Calendar year was not associated with outcome in univariate analysis, 58 32 however, after adjustment, calendar year was associated with decreased 59 60 33 mortality with a HR of 0.98 (0.98 – 0.99, p < 0.001) per year (Table 2).

8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 24

1 2 3 1 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 Discussion 6 7 3 This retrospective cohort study described the annual incidence of AH, associated 8 9 4 comorbidity and mortality in a single country. In the present study, we chose to 10 5 stratify patients into three groups (Primary AH, ‘AH + 2 diagnosis’, and ‘1 11 12 6 diagnosis + AH’). The rationale behind this stratification was to distinguish 13 14 7 patients with isolated AH and no comorbidity versus patients with comorbidity, 15 16 8 when the admitting physician believed that the AH diagnosis was the primary 17 9 cause of the admission versus patients with comorbidity, when the admitting 18 For peer review only 19 10 physician believed that the AH diagnosis was a secondary cause of admission. 20 21 11 The annual incidence of AH was 4.4 per 100,000 in this study. Other studies 22 23 12 have previously reported incidences of AH. A study from the Netherlands 24 13 25 reported an incidence of AH of 1.1 cases per 100,000 per year from 1987-1990 26 14 [7]. However, in this cohort, trauma was the primary cause of AH suggesting a 27 28 15 significantly different population compared to the present study [7]. A study 29 30 16 from four counties (900,000 inhabitants) in the northern part of Sweden 31 32 17 reported 3.4 cases of AH per 100,000 per year in from 2000-2007 [8]. In contrast 33 18 to the present study being conducted in a temperate climate zone, the Swedish 34 35 19 study was conducted in a sub-artic region, but nonetheless reported a lower

36 http://bmjopen.bmj.com/ 37 20 incidence of AH. However, the Swedish study showed an approximate two-fold 38 39 21 increase in AH incidence throughout the study period and the authors suggest 40 22 that this may be caused by increased physician awareness and increased 41 42 23 reporting [8]. A study from New Zealand reported an AH incidence of 6.9 per 43 24 100,000 per year from 1977-1986 [6]. While the annual incidence of AH was

44 on September 25, 2021 by guest. Protected copyright. 45 46 25 significantly increasing throughout our study period from 1994 to 2016 from a 47 26 statistical standpoint, the increase was small and there was no meaningful 48 49 27 increase in AH incidence throughout the study period. In accordance with 50 51 28 previous findings, more males were admitted with AH compared to females [11– 52 53 29 15]. 54 30 In our cohort, the incidence of AH was low during summer and increased 55 56 31 during the winter months. The likely explanation is that the country is located in 57 58 32 a temperate climate zone with a cold fall and winter. The increased incidence of 59 60 33 AH during the winter is in accordance with findings from other studies of AH in

9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 24 BMJ Open

1 2 3 1 cold climates [8,9,11,16–18]. In contrast to AH cohorts from mountainous 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 countries, avalanches do not occur in Denmark because of the lack of mountains. 6 7 3 This study found an overall 30-day mortality of 16% and one-year mortality 8 9 4 of 27%. A study from Japan previously reported a high in-hospital mortality of 10 5 24% in a population with AH from 2011-2016[9]. One difference between the 11 12 6 two studies was that a significant higher proportion of patients in the Japanese 13 14 7 study were admitted to an ICU (49% versus 16%). The difference in short-term 15 16 8 mortality may well be explained by different patient characteristics. 17 9 In the present study, older age was associated with both increased incidence 18 For peer review only 19 10 of AH and with increased mortality, which is in accordance with previously 20 21 11 reported data [7,19,20]. In this study, female sex compared to male sex was 22 23 12 associated with a higher overall mortality. However, the higher mortality in 24 13 females was accounted for by a significantly higher proportion of females with 25 26 14 AH being older than 80 years. The older age of females with AH has previously 27 28 15 been observed by Brändström et al [8]. Stratification into age groups revealed 29 30 16 that females and males had comparable mortality in all age groups, with the 31 17 exception of patients older than 80 years. In this age group, females had a 32 33 18 significantly lower mortality than males. This may well be accounted for by the 34 35 19 long follow-up and the longer life expectancy of females.

36 http://bmjopen.bmj.com/ 37 20 In this study, the diagnosis of AH was associated with a high comorbidity 38 21 burden. Having other recorded diagnoses was associated with a significantly 39 40 22 higher mortality compared to having AH as the only recorded diagnosis (i.e. 41 42 23 primary AH). The association between comorbidity and mortality is in 43 24 accordance with previous studies in AH populations [9,19].

44 on September 25, 2021 by guest. Protected copyright. 45 46 25 This study sought to investigate the incidence of AH, as well as the 47 26 associations with comorbidity and outcomes, using data from a national registry. 48 49 27 Previous studies have sought to address some of the same research questions 50 51 28 using registry data. Naturally, the incidence and outcomes of AH will depend on 52 53 29 both climate, geography, and socio-economic factors. AH is a relatively rare 54 30 diagnosis, and may be overlooked; especially mild AH presenting with competing 55 56 31 diagnoses such as sepsis. Furthermore, patients with AH constitutes a highly 57 58 32 heterogenous group depending on the cause of AH. While all of these factors 59 60 33 make it difficult to conduct reproducible studies and trials, future studies should

10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 24

1 2 3 1 ideally be prospective, with highly specific eligibility criteria, pre-defined data 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 collection and endpoints. 6 7 3 8 9 4 Limitations 10 5 A primary limitation of the present study was caused by data originating from 11 12 6 national registries. This limited the data granularity to what has been presented 13 14 7 in the present paper. While the diagnosis of AH likely has a high specificity, a 15 16 8 lower sensitivity may have resulted in underreporting of patients with 17 9 hypothermia, especially in presence of competing diagnoses. Accordingly, the 18 For peer review only 19 10 incidence of secondary AH may be higher than we reported. While diagnoses 20 21 11 were classified by treating physicians as primary or secondary, it is unknown to 22 23 12 which extent this classification corresponds to the definition of primary AH (AH 24 13 from isolated cold exposure) and secondary AH (AH in relation to acute or 25 26 14 chronic illness)[2]. However, the group of patients with AH recorded as the only 27 28 15 diagnosis should correspond to having primary hypothermia. The exclusions of 29 30 16 patients with more than one admission with AH recorded as the diagnosis could 31 17 result in underestimation of the incidence during the study period. 32 33 18 34 35 19 Conclusions

36 http://bmjopen.bmj.com/ 37 20 The present study investigated the incidence of AH, associated comorbidities, 38 21 and mortality after AH in Denmark from 1995 to 2016. The diagnosis was 39 40 22 associated with a high comorbidity burden; the most common being alcohol 41 42 23 intoxication. Further, the diagnosis was associated with a high one-year 43 24 mortality of 27%. Future prospective studies are needed to identify modifiable

44 on September 25, 2021 by guest. Protected copyright. 45 46 25 factors associated with both incidence of AH and outcome after AH. 47 26 48 49 27 Ethics approval and consent to participate 50 51 28 The study was approved by the Danish Patient Safety Authority and was 52 53 29 conducted in accordance with the Helsinki Declaration. As we conducted a 54 30 national retrospective study over a time period of 21 years, the Danish Patient 55 56 31 Safety Authority approved the study without need for obtaining consent from 57 58 32 each individual. 59 60 33

11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 24 BMJ Open

1 2 3 1 Availability of data and materials 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 Data was leveraged from national Danish registries. Accordingly, assess to data 6 7 3 can be obtained by application to the relevant Danish authorities. 8 9 4 10 5 Authors contributions 11 12 6 Author MW initiated the project. SW, AFM, JK, CH, and MW were involved in 13 14 7 planning the study including the statistical analyses. Authors SW and AFM wrote 15 16 8 the first draft of the manuscript. SW, AFM, JK, CH, and MW were all involved in 17 9 the production of the final manuscript. 18 For peer review only 19 10 20 21 11 22 23 12 24 13 25 26 14 27 28 15 29 16 30 31 17 32 33 18 34 35 19

36 20 http://bmjopen.bmj.com/ 37 38 21 39 40 22 41 42 23 43 24

44 on September 25, 2021 by guest. Protected copyright. 45 25 46 47 26 48 27 49 50 28 51 52 29 53 54 30 55 31 56 57 32 58 59 33 60

12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 24

1 2 3 1 References 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 1 van der Ploeg G-J, Goslings JC, Walpoth BH, et al. Accidental hypothermia: 6 7 3 Rewarming treatments, complications and outcomes from one university 8 4 medical centre. Resuscitation 2010;81:1550–5. 9 10 5 doi:10.1016/j.resuscitation.2010.05.023 11 12 6 2 Brown DJA, Brugger H, Boyd J, et al. Accidental Hypothermia. N Engl J Med 13 14 7 2012;367:1930–8. doi:10.1056/NEJMra1114208 15 8 3 Durrer B, Brugger H, Syme D. The Medical On-site Treatment of 16 17 9 Hypothermia* ICAR-MEDCOM Recommendation. High Alt Med Biol 18 For peer review only 19 10 2003;4:99–103. 20 21 11 4 Truhlář A, Deakin CD, Soar J, et al. European Resuscitation Council 22 12 Guidelines for Resuscitation 2015. Section 4. Cardiac arrest in special 23 24 13 circumstances. Resuscitation 2015;95:148–201. 25 26 14 doi:10.1016/j.resuscitation.2015.07.017 27 15 5 Pasquier M, Zurron N, Weith B, et al. Deep accidental hypothermia with core 28 29 16 temperature below 24°c presenting with vital signs. High Alt Med Biol 30 31 17 2014;15:58–63. doi:10.1089/ham.2013.1085 32 33 18 6 Taylor NAS, Griffiths RF, Cotter JD. Epidemiology of hypothermia: fatalities 34 19 and hospitalisations in New Zealand. Aust N Z J Med 1994;24. 35

36 20 doi:10.1111/j.1445-5994.1994.tb01788.x http://bmjopen.bmj.com/ 37 38 21 7 Bierens JJLM, Uitslager R, Swenne-Van Ingen MME, et al. Accidental 39 40 22 hypothermia: incidence, risk factors and clinical course of patients admitted to 41 23 hospital. Eur J Emerg Med 1995;2. doi:10.1097/00063110-199503000-00009 42 43 24 8 Brändström H, Johansson G, Giesbrecht GG, et al. Accidental cold-related

44 on September 25, 2021 by guest. Protected copyright. 45 25 injury leading to hospitalization in northern Sweden: An eight-year 46 26 retrospective analysis. Scand J Trauma Resusc Emerg Med 2014;22. 47 48 27 doi:10.1186/1757-7241-22-6 49 50 28 9 Matsuyama T, Morita S, Ehara N, et al. Characteristics and outcomes of 51 52 29 accidental hypothermia in Japan: The J-Point registry. Emerg Med J 53 30 2018;35:659–66. doi:10.1136/emermed-2017-207238 54 55 31 10 Thygesen SK, Christiansen CF, Christensen S, et al. The predictive value of 56 57 32 ICD-10 diagnostic coding used to assess Charlson comorbidity index 58 33 conditions in the population-based Danish National Registry of Patients. BMC 59 60 34 Med Res Methodol 2011;11. doi:10.1186/1471-2288-11-83

13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 24 BMJ Open

1 2 3 1 11 Baumgartner EA, Belson M, Rubin C, et al. Hypothermia and Other Cold- 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 Related Morbidity Emergency Department Visits: United States, 1995–2004. 6 7 3 Wilderness Environ Med 2008;19:233–6. doi:10.1580/07-WEME-OR-104.1 8 4 12 Rango N. Exposure-related hypothermia mortality in the United States, 1970- 9 10 5 79. Am J Public Health 1984;74:1159–60. 11 12 6 13 Hsieh TM, Kuo PJ, Hsu SY, et al. Effect of hypothermia in the emergency 13 14 7 department on the outcome of trauma patients: A cross-sectional analysis. Int J 15 8 Environ Res Public Health 2018;15. doi:10.3390/ijerph15081769 16 17 9 14 Svendsen ØS, Grong K, Andersen KS, et al. Outcome After Rewarming From 18 For peer review only 19 10 Accidental Hypothermia by Use of Extracorporeal Circulation. Ann Thorac 20 21 11 Surg 2017;103:920–5. doi:10.1016/j.athoracsur.2016.06.093 22 12 15 Debaty G, Moustapha I, Bouzat P, et al. Outcome after severe accidental 23 24 13 hypothermia in the French Alps: A 10-year review. Resuscitation 2015;93:118– 25 26 14 23. doi:10.1016/j.resuscitation.2015.06.013 27 15 16 Zhang P, Wiens K, Wang R, et al. Cold weather conditions and risk of 28 29 16 hypothermia among people experiencing homelessness: Implications for 30 31 17 prevention strategies. Int J Environ Res Public Health 2019;16. 32 33 18 doi:10.3390/ijerph16183259 34 19 17 Danzl DF, Pozos RS, Auerbach PS, et al. Multicenter hypothermia survey. Ann 35

36 20 Emerg Med 1987;16. doi:10.1016/S0196-0644(87)80757-6 http://bmjopen.bmj.com/ 37 38 21 18 The Eurowinter Group. Cold exposure and winter mortality from ischaemic 39 40 22 heart disease, cerebrovascular disease, respiratory disease, and all causes in 41 23 warm and cold regions of Europe. Lancet 1997;349:1341–6. 42 43 24 doi:10.1016/S0140-6736(96)12338-2

14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 24

1 2 3 1 Table 1. List of associated ICD-10 diagnoses with a prevalence higher than 1% in 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 patients with accidental hypothermia 6 7 List of secondary ICD-10 diagnoses in patients with List of primary ICD-10 diagnoses in patients with 8 ‘AH + 2 diagnosis’ (n= 1336) ‘1 diagnosis + AH’ (n= 1676) 9 10 Diagnosis ICD-10 n (%) Diagnosis ICD-10 n (%) 11 12 Acute alcohol intoxication DF100 198 (15) Acute alcohol intoxication DF100 153 (9.1) 13 14 Alcohol dependence syndrome DF102 118 (8.9) Respiratory insufficiency DJ969 74 (4.4) 15 Harmful use of alcohol DF101 96 (7.2) Sepsis DA419 61 (3.6) 16 17 Pneumonia DJ189 87 (6.5) Acute respiratory insufficiency DJ960 58 (3.5) 18 For peer review only 19 Dementia DF039 74 (5.5) Hypoglycaemia DE162 55 (3.3) 20 Hypertension DI109 59 (4.4) Pneumonia DJ189 48 (2.9) 21 22 Atrial fibrillation or flutter DI489 57 (4.3) Dehydration DE869 33 (2.0) 23 24 Dehydration DE869 52 (3.9) Drowning / nonfatal submersion DT751 32 (1.9) 25 Traumatic ischaemia of muscle DT796 29 (2.2) Cardiac arrest DI469 28 (1.7) 26 27 Type II diabetes DE119 27 (2.0) Concussion DS060 28 (1.7) 28 29 COPD DJ449 24 (1.8) Multiple lesions DT079 29 (1.7) 30 Hypoglycaemia DE162 23 (1.7) Stroke DI649 27 (1.6) 31 32 Anaemia DD649 21 (1.6) Traumatic ischaemia of muscle DT796 25 (1.5) 33 34 Hypokalaemia DE876 21 (1.6) Alcohol dependence syndrome DF102 20 (1.2) 35 Cardiac arrest DI469 22 (1.6) Harmful use of alcohol DF101 17 (1.0)

36 http://bmjopen.bmj.com/ 37 Complication after stroke DI694 22 (1.6) Septic shock DR572 17 (1.0) 38 39 Bacterial pneumonia DJ159 20 (1.5) - - - 40 Urinary tract infection DN390 20 (1.5) - - - 41 42 Sepsis DA419 16 (1.2) - - - 43 Type I diabetes DE109 16 (1.2) - - -

44 on September 25, 2021 by guest. Protected copyright. 45 Epilepsy DG409 16 (1.2) - - - 46 47 Hypotension DI959 15 (1.1) - - - 48 49 Schiizophrenia DF209 14 (1.0) - - - 50 51 3 ‘AH + 2 diagnosis’: Having AH recorded as primary diagnosis with one or more 52 4 recorded secondary diagnoses 53 54 5 ‘1 diagnosis + AH’: Having AH recorded as a secondary diagnosis with another 55 56 6 recorded primary diagnosis 57 58 59 60

15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from

Page 17 of 24 BMJ Open

1 2 3 4 5 1 Table 2. Associations between covariates and mortality within 365 days. Results presented as crude mortality rates as well as hazard 6 7 2 ratios after application of Cox proportional hazard models, including time to death within 365 days as the outcome variable. 8 Patients per Mortality Univariate models Multivariate model1 9 p p 10 group n (%) Hazard ratios (95%CI) Hazard ratios (95%CI) 11 AH type2 primary AH 2230 496 (22) ref. ref. 12 ‘AH + 2 diagnosis’ For1336 peer352 (26) review1.2 (1.1 - 1.4) < 0.001 only1.2 (1.0 - 1.3) 0.04 13 14 ‘1 diagnosis + AH’ 1676 582 (35) 1.7 (1.5 - 2.0) < 0.001 1.7 (1.5 - 2.0) < 0.001 15 16 http://bmjopen.bmj.com/ 17 Sex female 1981 615 (31) ref. ref. 18 male 3261 815 (25) 1.3 (1.2 - 1.4) < 0.001 0.89 (0.80 - 0.99) 0.04 19 20 Age at 21 18-20 years 181 3 (1.7) ref. ref. 22 admission 21-40 years 753 43 (5.7) 3.5 (1.1 - 11) 0.03 3.4 (1.0 - 11) 0.04 23 41-60 years 1571 337 (21) 15 (4.7 - 45) < 0.001 14 (4.4 - 43) < 0.001 24 on September 25, 2021 by guest. Protected copyright. 25 61-80 years 1512 490 (32) 24 (7.6 - 74) <0.001 24 (7.6 - 73) < 0.001 26 > 80 years 1225 557 (45) 36 (12 - 113) <0.001 38 (12 - 117) < 0.001 27 3 1: Adjusted for AH type, sex, age, and calendar year. 28 29 4 2: Accidental hypothermia (AH) 30 31 5 Primary AH: Having AH recorded as primary diagnosis without any recorded secondary diagnoses 32 6 ‘AH + 2 diagnosis’: Having AH recorded as primary diagnosis with one or more recorded secondary diagnoses 33 34 7 ‘1 diagnosis + AH’: Having AH recorded as a secondary diagnosis with another recorded primary diagnosis 35 36 37 38 39 40 41 42 16 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 18 of 24

1 2 3 1 Table 3. Mortality within 365 days stratified by age group and sex. 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 Total number in Female sex Male sex 6 Age group p each age group 7 1981 3261 8 female, n = 40 18-20 years 0 (0) 3 (2.1) 1.0 9 male, n = 141 10 female, n = 159 11 21-40 years 10 (6.3) 33 (5.6) 0.72 12 male, n = 594 13 female, n = 419 14 41-60 years 87 (21) 250 (22) 0.69 15 male, n = 1152 16 female, n = 597 61-80 years 201 (34) 289 (32) 0.40 17 male, n = 915 18 For peer review only female, n = 766 19 > 80 years 317 (41) 240 (52) 0.0002 20 male, n = 459 21 2 p-values for Chi-square test (or Fisher’s exact test) for difference in mortality 22 23 3 between sex. 24 25 26 27 28 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 25, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 24 BMJ Open

1 2 3 1 Figure Legends 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 6 7 3 Figure 1. incidence of accidental hypothermia per 100,000 inhabitants per 8 9 4 calendar year. 10 5 11 12 6 Accidental hypothermia (AH) 13 14 7 Primary AH: Having AH recorded as primary diagnosis without any recorded 15 16 8 secondary diagnoses 17 9 'AH + 2 diagnosis’: Having AH recorded as primary diagnosis with one or more 18 For peer review only 19 10 recorded secondary diagnoses 20 21 11 ‘1 diagnosis + AH’: Having AH recorded as a secondary diagnosis with another 22 23 12 recorded primary diagnosis 24 25 13 26 14 Figure 2. Distribution of accidental hypothermia diagnoses per calendar month. 27 28 15 29 30 16 Figure 3. One-year survival in accidental hypothermia patients stratified by AH 31 32 17 type (a.), male versus female (b.), and by age at admission (c.). Numbers at risk 33 18 presented for 0, 6, and 12 months. 34 35 19

36 http://bmjopen.bmj.com/ 37 20 Accidental hypothermia (AH) 38 39 21 Primary AH: Having AH recorded as primary diagnosis without any recorded 40 22 secondary diagnoses 41 42 23 'AH + 2 diagnosis’: Having AH recorded as primary diagnosis with one or more 43

44 24 recorded secondary diagnoses on September 25, 2021 by guest. Protected copyright. 45 46 25 ‘1 diagnosis + AH’: Having Ah recorded as a secondary diagnosis with another 47 48 26 recorded primary diagnosis 49 50 27 51 28 52 53 54 55 56 57 58 59 60

18 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 24 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 Figure 1. incidence of accidental hypothermia per 100,000 inhabitants per calendar year. Accidental 27 hypothermia (AH)Primary AH: Having AH recorded as primary diagnosis without any recorded secondary 28 diagnoses'AH + 2nd diagnosis’: Having AH recorded as primary diagnosis with one or more recorded 29 secondary diagnoses‘1st diagnosis + AH’: Having AH recorded as a secondary diagnosis with another 30 recorded primary diagnosis 31 160x96mm (300 x 300 DPI)

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 25, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 24 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 Figure 2. Distribution of accidental hypothermia diagnoses per calendar month per percent. 28 29 149x96mm (300 x 300 DPI) 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 25, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 24 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 25, 2021 by guest. Protected copyright. 41 42 43 44 45 Figure 3. One-year survival in accidental hypothermia patients stratified by AH type (a.), male versus female 46 (b.), and by age at admission (c.). Numbers at risk presented for 0, 6, and 12 months. 47 Accidental hypothermia (AH) 48 Primary AH: Having AH recorded as primary diagnosis without any recorded secondary diagnoses 49 'AH + 2nd diagnosis’: Having AH recorded as primary diagnosis with one or more recorded secondary 50 diagnoses 51 ‘1st diagnosis + AH’: Having Ah recorded as a secondary diagnosis with another recorded primary diagnosis 52 53 150x233mm (300 x 300 DPI) 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from

Page 23 of 24 BMJ Open

BMJ Open Page 24 of 24

Page 25 of 24 BMJ Open

Accidental Hypothermia in Denmark: Incidence and Outcomes

Journal: BMJ Open ManuscriptFor ID peerbmjopen-2020-046806.R2 review only Article Type: Original research

Date Submitted by the 02-May-2021 Author:

Primary Subject Emergency medicine Heading:

Secondary Subject Heading: Epidemiology

ACCIDENT & EMERGENCY MEDICINE, EPIDEMIOLOGY, GENERAL http://bmjopen.bmj.com/ Keywords: MEDICINE (see Internal Medicine)

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 24 BMJ Open

1 2 3

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 25, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 24

1 2 3 1 Accidental Hypothermia in Denmark: Incidence and Outcomes 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 Authors 6 7 3 Sebastian Wiberg* MD PhD1, Asmus Friborg Mortensen* MD2, Jesper Kjaergaard, 8 1 1 3 9 4 MD PhD DMSc , Christian Hassager MD DMSc , Michael Wanscher MD PhD . 10 5 11 12 6 Affiliations 13 14 7 1) Department of Cardiology, The Heart Centre, Copenhagen University 15 16 8 Hospital Rigshospitalet 17 9 2) Department of Anaesthesia and Intensive Care, Bispebjerg and 18 For peer review only 19 10 Frederiksberg Hospital, University of Copenhagen, Denmark 20 21 11 3) Department of Cardiothoracic Anaesthesiology, The Heart Centre, 22 23 12 Copenhagen University Hospital Rigshospitalet, Denmark 24 13 25 26 14 Corresponding Author 27 28 15 Sebastian Wiberg, MD, PhD 29 30 16 Department of Cardiology 31 17 The Heart Centre 32 33 18 Copenhagen University Hospital Rigshospitalet 34 35 19 [email protected]

36 http://bmjopen.bmj.com/ 37 20 38 21 *) Equal author contribution 39 40 22 41 42 23 Research Question 43 24 To investigate the reported incidence of accidental hypothermia in a nationwide

44 on September 25, 2021 by guest. Protected copyright. 45 46 25 registry, and to further investigate the associations between accidental 47 26 hypothermia and the associated outcomes. 48 49 27 50 51 28 Study Design 52 53 29 Retrospective cohort study 54 30 55 56 31 57 58 32 59 60 33

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 24 BMJ Open

1 2 3 1 Abstract 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 Objectives 6 7 3 To investigate the incidence of accidental hypothermia (AH) in a nationwide 8 9 4 registry and the associated outcomes. 10 5 11 12 6 Design 13 14 7 Nationwide retrospective cohort study 15 16 8 17 9 Participants and settings 18 For peer review only 19 10 All patients at least 18 years old, admitted to hospitals in Denmark with a 20 21 11 diagnosis of AH, with an ICD-10 code of T689, from January 1996 to November 22 23 12 2016. Other recorded diagnoses were included in the analyses. 24 13 25 26 14 Primary and secondary outcome measures 27 28 15 The primary outcome was one-year mortality. 29 30 16 31 17 Results 32 33 18 During the inclusion period, 5242 patients were admitted with a diagnosis of AH, 34 35 19 corresponding to a mean annual incidence of 4.4  1.2 (range by calendar year:

36 http://bmjopen.bmj.com/ 37 20 2.9 – 6.4) per 100,000 inhabitants. A total of 2230 (43%) had AH recorded as the 38 39 21 primary diagnosis without any recorded secondary diagnoses (Primary AH), 40 22 1336 (25%) had AH recorded as the primary diagnosis with other recorded 41 42 23 secondary diagnoses (AH + 2 diagnosis), and 1676 (32%) had AH recorded as a 43

44 24 secondary diagnosis with another recorded primary diagnosis (1 diagnosis + on September 25, 2021 by guest. Protected copyright. 45 46 25 AH). Alcohol intoxication was the most common diagnosis associated with AH. 47 48 26 Overall one-year mortality was 27%. In patients with primary AH, one-year 49 27 mortality was 22%, compared to 26% in patients with secondary AH type I, and 50 51 28 35% in patients with secondary AH type II (plog-rank < 0.001). 52 53 29 54 55 30 Conclusions 56 31 The present study investigated the incidence of AH, associated comorbidities, 57 58 32 and mortality after AH in Denmark from 1995 to 2016. The diagnosis is 59 60 33 associated with a high comorbidity burden and a considerable one-year

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 24

1 2 3 1 mortality. In the high proportion of patients with associated comorbidities, 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 establishing whether AH or the comorbidities are the drivers of mortality 6 7 3 remains difficult. This complicates our understanding of AH and makes it difficult 8 9 4 to find modifiable factors associated with both AH and outcomes. Future 10 5 prospective studies are needed elucidate the causal relationship between AH and 11 12 6 associated comorbidities. 13 14 7 15 16 8 Article Summary 17 9 Strengths and Limitations of this study 18 For peer review only 19 10 - Large contemporary study to investigate the incidence, comorbidity 20 21 11 burden and outcome in patients diagnosed with AH. 22 23 12 - Data originated from a national cohort of patients. 24 13 - The registry data limited the data granularity to what has been presented 25 26 14 in the present paper. 27 28 15 - The presented incidences are likely to be significantly different in other 29 30 16 countries, depending on climate, demography, etc. 31 17 32 33 18 Funding 34 35 19 The study was funded by Trygfonden - Accidental Hypothermia (ID: 117579).

44 on September 25, 2021 by guest. Protected copyright. 45 46 25 None 47 26 48 49 27 50 51 28 52 53 29 54 30 55 56 31 57 58 32 59 60 33

3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 24 BMJ Open

1 2 3 1 Introduction 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 Accidental hypothermia (AH), defined as an involuntary drop in core 6 7 3 temperature below 35C, is a serious condition with considerable mortality [1,2]. 8 9 4 AH can result from isolated cold exposure (primary hypothermia) or in 10 5 association with acute or chronic illness (secondary hypothermia) [2]. The 11 12 6 severity of hypothermia can be graded into three stages based on core 13 14 7 temperature: 32°C -35°C (mild AH), 28°C - 32°C (moderate AH), and < 28°C 15 16 8 (severe AH) [3–6]. In-hospital mortality after severe hypothermia has been 17 9 reported to be as high as 30%. It is estimated that approximately 1500 people 18 For peer review only 19 10 die from hypothermia every year in the United States [1,4]. Few previous studies 20 21 11 have investigated the incidence and outcome after AH, however, the studies 22 23 12 represent heterogenous populations from different climate zones and different 24 13 25 time periods[7–10]. Accordingly, our knowledge of the contemporary incidence, 26 14 demographics and outcomes after AH remains sparse. 27 28 15 The aim of this study was to investigate the incidence of AH in a Danish 29 30 16 nationwide registry, and the associated outcomes in adult patients. 31 32 17 33 18 Geography 34 35 19 Denmark covers a relatively small area of 4,300,000 hectares but has a

36 http://bmjopen.bmj.com/ 37 20 comparatively long coastline. The highest point is 173 meters above sea level. 38 th th 39 21 Denmark is located on the 55 to 57 parallel north in the temperate climate 40 22 zone. Mean temperatures range from 0C in January and February to 15.7C in 41 42 23 August. The population was about 5.8 million as of January 1st, 2016. About 1.8 43

44 24 million (about 31% of the population) lived in the capital, Copenhagen. on September 25, 2021 by guest. Protected copyright. 45 46 25 47 26 Methods 48 49 27 The present study is a nationwide retrospective cohort study including all adult 50 51 28 patients admitted to hospitals in Denmark with a diagnosis of AH from January 52 53 29 1996 to November 2016. The primary outcome was one-year mortality after 54 30 diagnosis. We chose a priori to include only those cases with a first-time 55 56 31 diagnosis of AH. The rationale behind this decision was to ensure that a patient 57 58 32 with multiple hospital admissions for AH only counted once in the analysis. 59 60

4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 24

1 2 3 1 All Danish residents are given a personal identification number at the time of 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 birth or immigration. This number is used in all contacts with the health care 6 7 3 system, and it is used as identifier in national registries. Data were extracted 8 9 4 from the Danish National Patient Register (NPR) to identify all patients admitted 10 5 with a discharge diagnosis of AH, defined by the International Classification of 11 12 6 Diseases, 10th edition (ICD-10, which has been used in Denmark since 1993) code 13 14 7 T689. 15 16 8 Patients were stratified into three groups based on whether AH was recorded 17 9 as the only diagnosis (Primary AH), whether AH was recorded as the primary 18 For peer review only 19 10 diagnosis with other recorded secondary diagnoses (AH + 2 diagnosis), or 20 21 11 whether AH was recorded as a secondary diagnosis (1 diagnosis + AH) for a 22 23 12 given admission. 24 25 13 The primary diagnosis is defined as the diagnosis primarily causing the need 26 14 for the admission, while the secondary diagnoses are defined as other conditions 27 28 15 that required attention during a given admission. Main treatment categories, 29 30 16 such as ‘intensive care admission’, and ‘respirator treatment’ were also extracted 31 32 17 from the NPR. The AH diagnosis was linked to survival (i.e. the primary outcome) 33 18 using the nationwide Danish Register of Causes of Death (RCD). The study was 34 35 19 approved by the Danish Patient Safety Authority Authority (Institutional review

36 http://bmjopen.bmj.com/ 37 20 board, ref. no 3-3013-1906/1/) and was conducted consistent with the Helsinki 38 39 21 Declaration. 40 22 41 42 23 Patient and Public Involvement 43

44 24 Patients or the public were not involved in the design of the present study. on September 25, 2021 by guest. Protected copyright. 45 46 25 47 26 Statistical analyses 48 49 27 Throughout, categorical variables are presented as counts (%), normally 50 51 28 distributed continuous variables are presented as mean ± standard deviation 52 th 53 29 (SD) and skewed continuous variables are presented as median (25 percentile - 54 30 75th percentile). Normality was assessed visually by QQ plots prior to analyses. 55 56 31 Associations between baseline variables and AH types were tested by the Chi- 57 58 32 Square test for categorical variables and with the one-way analysis of variance 59 60 33 (ANOVA) for continuous variable.

5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 24 BMJ Open

1 2 3 1 To give an overview of common comorbidity, a list of associated ICD-10 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 diagnoses with a prevalence higher than 1% in patients with AH was presented. 6 7 3 This overview was stratified by whether AH was recorded as the primary 8 9 4 diagnosis (AH + 2 diagnosis) or whether AH was recorded as a secondary 10 5 diagnosis (1 diagnosis + AH). 11 12 6 The annual incidence of AH per calendar year was plotted for the total cohort 13 14 7 as well as after stratification by AH type. The mean annual incidence  SD with 15 16 8 range was presented. To assess if the incidence of AH was changing over time, 17 18 9 linear models wereFor applied peer with calendar review year being only included as a continuous 19 10 covariate assuming a linear trend. The absolute change in AH incidence per 20 21 11 calendar year  standard deviation was presented. The distribution of AH 22 23 12 diagnoses per calendar months were displayed along with mean outside 24 25 13 temperatures. The mean annual incidence  SD of AH after stratification by age 26 27 14 group was presented. 28 15 We presented crude 1-year mortality rates for the total cohort, as well as after 29 30 16 stratification by AH type, by sex, and by age group. We presented 1-year 31 32 17 mortality after stratification by sex and age group. Differences in mortality 33 34 18 between females and males, stratified by age group, were analyzed by 35 19 application of the Chi-Square test.

36 http://bmjopen.bmj.com/ 37 20 For each patient in the cohort, the date of entry (i.e. the date of AH diagnosis) 38 39 21 as well as the date of death was recorded in the registries. The Kaplan-Meier 40 41 22 estimator was applied to visualize time to death within 1-year from admission, 42 23 stratified by AH type, sex, and age group. The log-rank test was applied to assess 43

44 24 possible differences between strata. on September 25, 2021 by guest. Protected copyright. 45 46 25 To analyze time to death between groups, we applied Cox proportional hazard 47 48 26 models adjusting for AH type, sex, age at admission, and calendar year. We 49 27 presented hazard ratios with 95% confidence intervals. A significance level of < 50 51 28 0.05 was applied throughout. SAS software, version 9.4 (SAS Institute, Cary, 52 53 29 North Carolina, USA) was used for all statistical analysis. 54 55 30 56 31 Results 57 58 32 Over the inclusion period of almost 22 years, the Danish Health Care system 59 60 33 received a total of 5242 adult patients admitted with a diagnosis of AH. A total of

6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 24

1 2 3 1 2230 (43%) had AH recorded as the final primary diagnosis without any 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 recorded secondary diagnoses (Primary AH), 1336 (25%) had AH recorded as 6 7 3 the final primary diagnosis with other recorded secondary diagnoses (AH + 2 8 9 4 diagnosis), and 1676 (32%) had AH recorded as a final secondary diagnosis with 10 5 another recorded primary diagnosis (1 diagnosis + AH). 11 12 6 A total of 1981 (38%) of the patients admitted with AH were female. The 13 14 7 mean age was 6121 years. Female sex was slightly less prevalent among 15 16 8 patients with primary AH compared to patients with AH + 2 diagnosis and 17 18 9 patients with 1For diagnosis peer + AH (35% review vs. 39% vs. 40%, only p< 0.001). Patients with 19 20 10 primary AH were slightly younger compared to patients with AH + 2 diagnosis 21 11 and patients with 1 diagnosis + AH (5922 years vs. 6220 years vs. 6220 22 23 12 years, p < 0.001). 24 25 13 26 27 14 Comorbidity 28 29 15 Both AH + 2 diagnosis and 1 diagnosis + AH were associated with a significant 30 16 comorbidity burden. A list of associated diagnoses with a proportion higher than 31 32 17 1% is provided in Table 1. The most frequent diagnosis associated with AH was 33 34 18 alcohol intoxication (Table 1). 35 19 A total of 816 (16%) patients with AH received care in an intensive care unit.

36 http://bmjopen.bmj.com/ 37 20 A total of 305 (5.8%) were treated with vasopressors and/or inotropic agents, 38 39 21 397 (7.6%) received ventilator treatment, and 479 (9.1) received dialysis. A total 40 41 22 of 21 (0.4%) patients were treated with extracorporeal circulation. 42 43 23

44 24 Incidence on September 25, 2021 by guest. Protected copyright. 45 46 25 The mean annual incidence of AH was 4.4  1.2 per 100,000 inhabitants with a 47 48 26 range from 2.9 to 6.4 per 100,000 inhabitants (Figure 1). The overall incidence of 49 50 27 AH increased from 2000 to 2016 (absolute increase 0.160.02 per 100,000 51 52 28 inhabitants per calendar year, p < 0.001). 53 29 The annual incidence of AH was 4.3  1.6 per 100,000 in inhabitants from 18 54 55 30 to 20 years, 2.3  0.55 per 100,000 in inhabitants from 21 to 40 years, 4.8  1.1 56 57 31 per 100,000 in inhabitants from 41 to 60 years, 7.2  2.6 in inhabitants from 61 58 59 32 to 80 years, and 28  11 in inhabitants older than 80 years. 60

7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 24 BMJ Open

1 2 3 1 The incidence of AH followed a seasonal trend, with the highest percent of AH 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 occurring in January (16%) and the lowest percent of AH occurring in August 6 7 3 (3.0%, Figure 2). Of all AH diagnoses, 29% were recorded in the Danish capital 8 9 4 region inhabited by 31% of the total Danish population. 10 5 11 12 6 Outcome 13 14 7 Overall mortality within the first 7 days of admission was 11%, increasing to 15 16 8 16% at 30 days, and 27% at one year. One-year mortality was 22% in patients 17 9 with primary AH, 26% in patients with AH + 2 diagnosis, and 35% in patients 18 For peer review only 19 10 with 1 diagnosis + AH (p < 0.001, Figure 3). The higher mortality in patients 20 21 11 with comorbidities remained unchanged after adjustment for confounding 22 23 12 factors (Table 2). 24 25 13 Increasing age was significantly associated with increased one-year mortality 26 14 (Table 2). One-year mortality increased from 1.6% in patients between 18 and 27 28 15 20 years to 45% in patients older than 80 years (p < 0.001, Table 2). 29 30 16 One-year mortality was higher in females compared to males (31% vs. 25%, p 31 32 17 < 0.001, Table 2), corresponding to a HR of 1.3 (1.2 – 1.4). In contrast, we found a 33 18 significantly lower mortality in females older than 80 years compared to men 34 35 19 older than 80 years (41% vs. 52%, p < 0.001), but no significant differences

36 http://bmjopen.bmj.com/ 37 20 between the sexes in other age groups (Table 3). A significantly higher 38 39 21 proportion of females were older than 80 years compared to men (39% vs. 14%, 40 22 p < 0.001, Table 3). The higher overall mortality in women was caused by a 41 42 23 higher proportion of women compared to men being older than 80 years. Female 43

44 24 sex was associated with a lower mortality compared to male sex after on September 25, 2021 by guest. Protected copyright. 45 46 25 adjustment for confounding factors including age (HR 0.89 95%CI 0.80 – 0.99, 47 26 Table 2). 48 49 27 Calendar year was not associated with outcome in univariate analysis, 50 51 28 however, after adjustment, calendar year was associated with slightly decreased 52 53 29 mortality with a HR of 0.98 (0.98 – 0.99, p < 0.001) per year (Table 2). 54 30 55 56 31 Discussion 57 58 32 We chose to stratify patients into three groups (Primary AH, AH + 2 diagnosis, 59 60 33 and 1 diagnosis + AH). The rationale behind this stratification was to distinguish

8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 24

1 2 3 1 patients with isolated AH and no comorbidity versus patients with 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 comorbidities, when the admitting physician believed that the AH diagnosis was 6 7 3 the primary cause of the admission versus patients with comorbidities, when the 8 9 4 admitting physician believed that the AH diagnosis was a secondary cause of 10 5 admission. 11 12 6 The annual incidence of AH was 4.4 per 100,000 in this study. A study from 13 14 7 the Netherlands reported an incidence of AH of 1.1 cases per 100,000 per year 15 16 8 from 1987-1990 [8]. However, in this cohort, trauma was the primary cause of 17 9 AH suggesting a significantly different population compared to the present study 18 For peer review only 19 10 [8]. A study from four counties with 900,000 inhabitants in the northern part of 20 21 11 Sweden reported 3.4 cases of AH per 100,000 per year in from 2000-2007 [9]. In 22 23 12 contrast to the present study from a temperate climate zone, the Swedish study 24 13 was conducted in a sub-artic region. However, the Swedish study reported a 25 26 14 lower incidence of AH. The Swedish study found an approximate two-fold 27 28 15 increase in the incidence of AH during the study perio. The authors suggest that 29 30 16 this may be caused by increased physician awareness and increased reporting 31 17 [9]. A study from New Zealand reported an AH incidence of 6.9 per 100,000 per 32 33 18 year from 1977-1986 [7]. 34 35 19 There was no clinically significant increase in the incidence of AH during our

36 http://bmjopen.bmj.com/ 37 20 study period. More males were admitted with AH compared to females. This is 38 21 consistent with the results of other studies [11–15]. 39 40 22 The incidence of AH was low during summer and increased during the winter 41 42 23 months. The likely explanation is that the Denmark has a temperate climate with 43 24 a cold fall and winter. The increased incidence of AH during the winter is

44 on September 25, 2021 by guest. Protected copyright. 45 46 25 consistent with findings from other studies of AH in cold climates [9–11,16–18]. 47 26 We found an overall 30-day mortality of 16% and one-year mortality of 27%. 48 49 27 A study from Japan previously reported a high in-hospital mortality of 24% in a 50 51 28 population with AH from 2011-2016[10]. One difference between the two 52 53 29 studies was that a significantly higher proportion of patients in the Japanese 54 30 study were admitted to an ICU (49% versus 16%). The difference in short-term 55 56 31 mortality may be explained by different patient characteristics. 57 58 32 In the present study, older age was associated with both increased incidence 59 60 33 of AH and with increased mortality consistent with previously reported data

9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 24 BMJ Open

1 2 3 1 [8,19,20]. The Swedish study also found a higher proportion of older females 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 with AH [9]. In our study, females and males had comparable mortality in all age 6 7 3 groups with the exception of patients older than 80 years. Females older than 80 8 9 4 years had a significantly lower mortality than males. This may be explained by 10 5 the long follow-up and the longer life expectancy of females. 11 12 6 We found that the diagnosis of AH was associated with a burden of 13 14 7 comorbidities. Having other recorded diagnoses was associated with a 15 16 8 significantly higher mortality compared to having AH as the only recorded 17 9 diagnosis (i.e. primary AH). The association between comorbidity and mortality 18 For peer review only 19 10 is consistent with previous studies in AH populations [10,19]. 20 21 11 As with our study, previous studies have also used registry data to address 22 23 12 some of the same research questions. The incidence and outcomes of AH depend 24 13 on climate, geography and socioeconomic factors. 25 26 14 AH is a relatively rare diagnosis, and may be overlooked; especially in the case 27 28 15 of mild AH presenting with associated diagnoses such as sepsis. Patients with AH 29 30 16 constitute a highly heterogenous group depending on the cause of AH. Whether 31 17 AH or associated diagnoses are the drivers of ensuing morbidity or mortality 32 33 18 remains poorly understood. Future studies on AH should ideally be prospective. 34 35 19 Eligibility criteria should be based on body temperature measured upon

36 http://bmjopen.bmj.com/ 37 20 admission. Clinical signs on admission should be reported to validate current 38 21 classification methods of AH stages[6]. Data collection should include a brief 39 40 22 description of the circumstances associated with the event, for example water 41 42 23 accident. Associated diagnoses should be reported, and the treating clinician 43 24 should ideally record, whether AH was the driver of any associated diagnoses,

44 on September 25, 2021 by guest. Protected copyright. 45 46 25 whether an associated diagnosis was the driver of AH, or whether AH and 47 26 associated diagnoses were considered unrelated. Endpoints should include 48 49 27 readmissions after index-admission and all-cause mortality. 50 51 28 52 53 29 Limitations 54 30 A primary limitation of the present study was caused by data originating from 55 56 31 national registries. This limited the data granularity to what has been presented 57 58 32 in the present paper. While the diagnosis of AH likely has a high specificity, a 59 60 33 lower sensitivity may have resulted in underreporting of patients with

10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 24

1 2 3 1 hypothermia, especially in presence of competing diagnoses. The incidence of 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 secondary AH may be higher than we reported. However, hospitals are 6 7 3 reimbursed based on correct reporting of diagnoses, and accordingly, the 8 9 4 registry has a high accuracy for most diagnoses[21]. While diagnoses were 10 5 classified by treating physicians as primary or secondary, it is unknown to what 11 12 6 extent this classification corresponds to primary AH (AH from isolated cold 13 14 7 exposure) and secondary AH (AH in relation to acute or chronic illness)[2]. 15 16 8 However, patients whose only diagnosis was AH most likely had primary 17 9 hypothermia. Excluding additional admissions of patients admitted more than 18 For peer review only 19 10 once with a diagnosis of AH might have caused underestimation of the true 20 21 11 incidence of AH. 22 23 12 24 13 Conclusions 25 26 14 The present study investigated the incidence of AH, associated comorbidities, 27 28 15 and mortality after AH in Denmark from 1995 to 2016. The diagnosis is 29 30 16 associated with a high comorbidity burden and a considerable one-year 31 17 mortality. In the high proportion of patients with associated comorbidities, 32 33 18 establishing whether AH or the comorbidities are the drivers of mortality 34 35 19 remains difficult. This complicates our understanding of AH and makes it difficult

36 http://bmjopen.bmj.com/ 37 20 to find modifiable factors associated with both AH and outcomes. Future 38 21 prospective studies are needed elucidate the causal relationship between AH and 39 40 22 associated comorbidities. 41 42 23 43 24 Ethics approval and consent to participate

44 on September 25, 2021 by guest. Protected copyright. 45 46 25 The study was approved by the Danish Patient Safety Authority and was 47 26 conducted consistent with the Helsinki Declaration. As we conducted a national 48 49 27 retrospective study over a time period of 21 years, the Danish Patient Safety 50 51 28 Authority approved the study without need for obtaining consent from each 52 53 29 individual. 54 30 55 56 31 Availability of data and materials 57 58 32 Data was extracted from national Danish registries. Accordingly, assess to data 59 60 33 can be obtained by application to the relevant Danish authorities.

11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 24 BMJ Open

1 2 3 1 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 Authors contributions 6 7 3 Author MW initiated the project. SW, AFM, JK, CH, and MW were involved in 8 9 4 planning the study including the statistical analyses. Authors SW and AFM wrote 10 5 the first draft of the manuscript. SW, AFM, JK, CH, and MW were all involved in 11 12 6 the production of the final manuscript. 13 14 7 15 16 8 17 9 18 For peer review only 19 10 20 21 11 22 12 23 24 13 25 26 14 27 28 15 29 16 30 31 17 32 33 18 34 35 19

36 20 http://bmjopen.bmj.com/ 37 38 21 39 40 22 41 23 42 43 24

44 on September 25, 2021 by guest. Protected copyright. 45 25 46 47 26 48 27 49 50 28 51 52 29 53 54 30 55 31 56 57 32 58 59 33 60

12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 24

1 2 3 1 References 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 1 van der Ploeg G-J, Goslings JC, Walpoth BH, et al. Accidental hypothermia: 6 7 3 Rewarming treatments, complications and outcomes from one university 8 4 medical centre. Resuscitation 2010;81:1550–5. 9 10 5 doi:10.1016/j.resuscitation.2010.05.023 11 12 6 2 Brown DJA, Brugger H, Boyd J, et al. Accidental Hypothermia. N Engl J Med 13 14 7 2012;367:1930–8. doi:10.1056/NEJMra1114208 15 8 3 Durrer B, Brugger H, Syme D. The Medical On-site Treatment of 16 17 9 Hypothermia* ICAR-MEDCOM Recommendation. High Alt Med Biol 18 For peer review only 19 10 2003;4:99–103. 20 21 11 4 Truhlář A, Deakin CD, Soar J, et al. European Resuscitation Council 22 12 Guidelines for Resuscitation 2015. Section 4. Cardiac arrest in special 23 24 13 circumstances. Resuscitation 2015;95:148–201. 25 26 14 doi:10.1016/j.resuscitation.2015.07.017 27 15 5 Pasquier M, Zurron N, Weith B, et al. Deep accidental hypothermia with core 28 29 16 temperature below 24°c presenting with vital signs. High Alt Med Biol 30 31 17 2014;15:58–63. doi:10.1089/ham.2013.1085 32 33 18 6 Musi ME, Sheets A, Zafren K, et al. Clinical staging of accidental 34 19 hypothermia: The Revised Swiss System: Recommendation of the International 35

36 20 Commission for Mountain Emergency Medicine (ICAR MedCom). http://bmjopen.bmj.com/ 37 38 21 Resuscitation 2021;162:182–7. doi:10.1016/j.resuscitation.2021.02.038 39 40 22 7 Taylor NAS, Griffiths RF, Cotter JD. Epidemiology of hypothermia: fatalities 41 23 and hospitalisations in New Zealand. Aust N Z J Med 1994;24. 42 43 24 doi:10.1111/j.1445-5994.1994.tb01788.x

44 on September 25, 2021 by guest. Protected copyright. 45 25 8 Bierens JJLM, Uitslager R, Swenne-Van Ingen MME, et al. Accidental 46 26 hypothermia: incidence, risk factors and clinical course of patients admitted to 47 48 27 hospital. Eur J Emerg Med 1995;2. doi:10.1097/00063110-199503000-00009 49 50 28 9 Brändström H, Johansson G, Giesbrecht GG, et al. Accidental cold-related 51 52 29 injury leading to hospitalization in northern Sweden: An eight-year 53 30 retrospective analysis. Scand J Trauma Resusc Emerg Med 2014;22. 54 55 31 doi:10.1186/1757-7241-22-6 56 57 32 10 Matsuyama T, Morita S, Ehara N, et al. Characteristics and outcomes of 58 33 accidental hypothermia in Japan: The J-Point registry. Emerg Med J 59 60 34 2018;35:659–66. doi:10.1136/emermed-2017-207238

13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 24 BMJ Open

44 on September 25, 2021 by guest. Protected copyright. 45 25 19 Okada Y, Matsuyama T, Morita S, et al. Prognostic factors for patients with 46 26 accidental hypothermia: A multi-institutional retrospective cohort study. Am J 47 48 27 Emerg Med 2019;37:565–70. doi:10.1016/j.ajem.2018.06.025 49 50 28 20 Morita S, Matsuyama T, Ehara N, et al. Prevalence and outcomes of accidental 51 52 29 hypothermia among elderly patients in Japan: Data from the J-Point registry. 53 30 Geriatr Gerontol Int 2018;18:1427–32. doi:10.1111/ggi.13502 54 55 31 21 Thygesen SK, Christiansen CF, Christensen S, et al. The predictive value of 56 57 32 ICD-10 diagnostic coding used to assess Charlson comorbidity index 58 33 conditions in the population-based Danish National Registry of Patients. BMC 59 60 34 Med Res Methodol 2011;11. doi:10.1186/1471-2288-11-83

14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 24

1 2 3 1 Table 1. List of associated ICD-10 diagnoses with a prevalence higher than 1% in 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 patients with accidental hypothermia 6 7 List of secondary ICD-10 diagnoses in patients with List of primary ICD-10 diagnoses in patients with 8 AH + 2 diagnosis (n= 1336) 1 diagnosis + AH (n= 1676) 9 10 Diagnosis ICD-10 n (%) Diagnosis ICD-10 n (%) 11 12 Acute alcohol intoxication DF100 198 (15) Acute alcohol intoxication DF100 153 (9.1) 13 14 Alcohol dependence syndrome DF102 118 (8.9) Respiratory insufficiency DJ969 74 (4.4) 15 Harmful use of alcohol DF101 96 (7.2) Sepsis DA419 61 (3.6) 16 17 Pneumonia DJ189 87 (6.5) Acute respiratory insufficiency DJ960 58 (3.5) 18 For peer review only 19 Dementia DF039 74 (5.5) Hypoglycaemia DE162 55 (3.3) 20 Hypertension DI109 59 (4.4) Pneumonia DJ189 48 (2.9) 21 22 Atrial fibrillation or flutter DI489 57 (4.3) Dehydration DE869 33 (2.0) 23 24 Dehydration DE869 52 (3.9) Drowning / nonfatal submersion DT751 32 (1.9) 25 Traumatic ischaemia of muscle DT796 29 (2.2) Cardiac arrest DI469 28 (1.7) 26 27 Type II diabetes DE119 27 (2.0) Concussion DS060 28 (1.7) 28 29 COPD DJ449 24 (1.8) Multiple lesions DT079 29 (1.7) 30 Hypoglycaemia DE162 23 (1.7) Stroke DI649 27 (1.6) 31 32 Anaemia DD649 21 (1.6) Traumatic ischaemia of muscle DT796 25 (1.5) 33 34 Hypokalaemia DE876 21 (1.6) Alcohol dependence syndrome DF102 20 (1.2) 35 Cardiac arrest DI469 22 (1.6) Harmful use of alcohol DF101 17 (1.0)

44 on September 25, 2021 by guest. Protected copyright. 45 Epilepsy DG409 16 (1.2) - - - 46 47 Hypotension DI959 15 (1.1) - - - 48 49 Schiizophrenia DF209 14 (1.0) - - - 50 51 3 AH + 2 diagnosis: AH recorded as primary diagnosis with one or more recorded 52 4 secondary diagnoses 53 54 5 1 diagnosis + AH: AH recorded as a secondary diagnosis with another recorded 55 56 6 primary diagnosis 57 58 59 60

15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from

Page 17 of 24 BMJ Open

1 2 3 4 5 1 Table 2. Associations between covariates and mortality within 1 year. Results presented as crude mortality rates as well as hazard 6 7 2 ratios after application of Cox proportional hazard models, including time to death within 1 year as the outcome variable. 8 Patients per Mortality Univariate models Multivariate model1 9 p p 10 group n (%) Hazard ratios (95%CI) Hazard ratios (95%CI) 11 AH type2 primary AH 2230 496 (22) ref. ref. 12 AH + 2 diagnosis For1336 peer352 (26) review1.2 (1.1 - 1.4) < 0.001 only1.2 (1.0 - 1.3) 0.04 13 14 1 diagnosis + AH 1676 582 (35) 1.7 (1.5 - 2.0) < 0.001 1.7 (1.5 - 2.0) < 0.001 15 16 http://bmjopen.bmj.com/ 17 Sex female 1981 615 (31) ref. ref. 18 male 3261 815 (25) 1.3 (1.2 - 1.4) < 0.001 0.89 (0.80 - 0.99) 0.04 19 20 Age at 21 18-20 years 181 3 (1.7) ref. ref. 22 admission 21-40 years 753 43 (5.7) 3.5 (1.1 - 11) 0.03 3.4 (1.0 - 11) 0.04 23 41-60 years 1571 337 (21) 15 (4.7 - 45) < 0.001 14 (4.4 - 43) < 0.001 24 on September 25, 2021 by guest. Protected copyright. 25 61-80 years 1512 490 (32) 24 (7.6 - 74) <0.001 24 (7.6 - 73) < 0.001 26 > 80 years 1225 557 (45) 36 (12 - 113) <0.001 38 (12 - 117) < 0.001 27 3 1: Adjusted for AH type, sex, age, and calendar year. 28 29 4 2: Accidental hypothermia (AH) 30 31 5 Primary AH: AH recorded as primary diagnosis without any recorded secondary diagnoses 32 6 AH + 2 diagnosis: AH recorded as primary diagnosis with one or more recorded secondary diagnoses 33 34 7 1 diagnosis + AH: AH recorded as a secondary diagnosis with another recorded primary diagnosis 35 36 37 38 39 40 41 42 16 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 18 of 24

1 2 3 1 Table 3. Mortality within 1 year stratified by age group and sex. 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 Total number in Female sex Male sex 6 Age group p each age group 7 1981 3261 8 female, n = 40 18-20 years 0 (0) 3 (2.1) 1.0 9 male, n = 141 10 female, n = 159 11 21-40 years 10 (6.3) 33 (5.6) 0.72 12 male, n = 594 13 female, n = 419 14 41-60 years 87 (21) 250 (22) 0.69 15 male, n = 1152 16 female, n = 597 61-80 years 201 (34) 289 (32) 0.40 17 male, n = 915 18 For peer review only female, n = 766 19 > 80 years 317 (41) 240 (52) 0.0002 20 male, n = 459 21 2 p-values for Chi-square test (or Fisher’s exact test) for difference in mortality 22 23 3 between sexes. 24 25 26 27 28 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 25, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 24 BMJ Open

1 2 3 1 Figure Legends 4 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 5 2 6 7 3 Figure 1. Incidence of accidental hypothermia per 100,000 inhabitants per 8 9 4 calendar year. 10 5 11 12 6 Accidental hypothermia (AH) 13 14 7 Primary AH: AH recorded as primary diagnosis without any recorded secondary 15 16 8 diagnoses 17 9 AH + 2 diagnosis: AH recorded as primary diagnosis with one or more recorded 18 For peer review only 19 10 secondary diagnoses 20 21 11 1 diagnosis + AH: AH recorded as a secondary diagnosis with another recorded 22 23 12 primary diagnosis 24 25 13 26 14 Figure 2. Distribution of accidental hypothermia diagnoses per calendar month. 27 28 15 29 30 16 Figure 3. One-year survival in accidental hypothermia patients stratified by AH 31 32 17 type (a.), male versus female (b.), and by age at admission (c.). 33 18 34 35 19 Accidental hypothermia (AH)

36 http://bmjopen.bmj.com/ 37 20 Primary AH: AH recorded as primary diagnosis without any recorded secondary 38 39 21 diagnoses 40 22 AH + 2 diagnosis: AH recorded as primary diagnosis with one or more recorded 41 42 23 secondary diagnoses 43

44 24 1 diagnosis + AH: AH recorded as a secondary diagnosis with another recorded on September 25, 2021 by guest. Protected copyright. 45 46 25 primary diagnosis 47 48 26 49 50 27 51 52 53 54 55 56 57 58 59 60

18 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 24 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 Figure 1. Incidence of accidental hypothermia per 100,000 inhabitants per calendar year. 27 Accidental hypothermia (AH)Primary AH: AH recorded as primary diagnosis without any recorded secondary 28 diagnoses 29 AH + 2○ diagnosis: AH recorded as primary diagnosis with one or more recorded secondary diagnoses 30 1○ diagnosis + AH: AH recorded as a secondary diagnosis with another recorded primary diagnosis 31 160x96mm (600 x 600 DPI)

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 25, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 24 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 Figure 2. Distribution of accidental hypothermia diagnoses per calendar month. 28 29 150x96mm (600 x 600 DPI) 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 25, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 24 BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 25, 2021 by guest. Protected copyright. 41 42 43 44 45 Figure 3. One-year survival in accidental hypothermia patients stratified by AH type (a.), male versus female 46 (b.), and by age at admission (c.). 47 Accidental hypothermia (AH)Primary AH: AH recorded as primary diagnosis without any recorded secondary diagnoses 48 AH + 2○ diagnosis: AH recorded as primary diagnosis with one or more recorded secondary diagnoses 49 1○ diagnosis + AH: AH recorded as a secondary diagnosis with another recorded primary diagnosis 50 51 150x233mm (600 x 600 DPI) 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2020-046806 on 31 May 2021. Downloaded from

Page 23 of 24 BMJ Open

BMJ Open Page 24 of 24

Page 25 of 24 BMJ Open