Accepted Manuscript Increased short- and long-term mortality among patients with infectious spondylodiscitis compared with a reference population Michala Kehrer, MD., Court Pedersen, MD, professor Dr. Sci., Thøger Gorm Jensen., MD, PhD., Jesper Hallas, MD, Dr. Sci., Annmarie T. Lassen., MD, PhD, Professor Dr. Sci. PII:

S1529-9430(15)00133-3

DOI:

10.1016/j.spinee.2015.02.021

Reference:

SPINEE 56204

To appear in:

The Spine Journal

Received Date: 9 July 2014 Revised Date:

11 January 2015

Accepted Date: 11 February 2015

Please cite this article as: Kehrer M, Pedersen C, Jensen. TG, Hallas J, Lassen. AT, Increased shortand long-term mortality among patients with infectious spondylodiscitis compared with a reference population, The Spine Journal (2015), doi: 10.1016/j.spinee.2015.02.021. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Title: Increased short- and long-term mortality among patients with infectious spondylodiscitis

RI PT

compared with a reference population

Authors: Michala Kehrer, MD. a,b

SC

Court Pedersen, MD, professor Dr. Sci. a,b Thøger Gorm Jensen. MD, PhD. c

M AN U

Jesper Hallas, MD, Dr. Sci.d Annmarie T Lassen. MD, PhD, Professor Dr. Sci.a,e a

Institute of Clinical Research, University of Southern Denmark, Winsløwparken 19/3

5000 Odense C, Denmark b

Department of Infectious Diseases, Odense University Hospital, Sdr. Boulevard 29, 5000 Odense C,

c

TE D

Denmark.

Department of Clinical Microbiology, Odense University Hospital, J.B. Winsløws Vej 21, 2

5000 Odense C , Denmark.

Research Unit of Clinical Pharmacology, University of Southern Denmark, Denmark

EP

d

e

Department of Emergency Medicine, Odense University Hospital. Sdr. Boulevard 29, 5000 Odense C,

AC C

Denmark. Winsløwparken 19, 2., 5000 Odense C, Denmark.

Corresponding author:

Michala Lykkegaard Levin Kehrer. Department of Infectious Diseases, Odense University Hospital. Odense, Sdr. Boulevard 29, 5000 Odense, Denmark. E-mail: [email protected]. Phone: 0045 6541 2651; Fax: 0045 6611 7418.

ACCEPTED MANUSCRIPT

1

Title: Increased short- and long-term mortality among patients with infectious spondylodiscitis

2

compared with a reference population

RI PT

3

Abstract:

5

Background:

6 7

Information on short- and especially long-term mortality among patients with infectious spondylodiscitis is sparse

8

Purpose:

9

We aimed to analyse mortality, factors associated with death, and cause specific mortality rates

M AN U

SC

4

among patients with infectious non-postoperative spondylodiscitis.

11

Study design:

12

Case-cohort study.

13

Patient sample:

14

We identified all patients aged ≥18 years treated for infectious spondylodiscitis January 1994–May

15

2009 at hospitals in Funen County, Denmark.

16

Outcome measures:

17

Overall and cause specific mortality.

18

Methods:

19

Mortality rates among patients were compared with rates among a reference population using

20

Kaplan-Meier plots and mortality rate ratios (MRR). Short-term mortality was defined as deaths

21

within first year after admission and long-term mortality was deaths thereafter. Factors associated

22

with death were determined.

23

The authors have no conflict of interest.

AC C

EP

TE D

10

1

ACCEPTED MANUSCRIPT

Results:

2

Among 298 identified patients 61 (20%) died within the first year. Adjusted MRR was 16.8 (95%

3

Confidence interval: 9.9–28.5) for 0-90 days; 4.2 (2.5–7.0) 91–365 days; 2.2 (1.6–2.9) 1–4 years;

4

and 1.7 (1.2–2.5) 5–14 years. MRR stratified on microbiological aetiology was 8.8 (3.3–22.1) 0-90

5

days; 1.4 (0.3–5.8) 91–365 days; 3.2 (2.0–5.1) 1–4 years; and 1.1 (0.5–2.4) 5–14 years for unknown

6

aetiology and 24.0 (13.0–44.2) 0-90 days; 6.0 (3.1–11.5) 91–365 days; 1.9 (1.1–3.2) 1–4 years; and

7

2.7 (1.5–4.7) 5–14 years among Staphylococcus aureus infections. The main factors associated

8

with short-term mortality were severe neurological deficits at time of admission, epidural abscess

9

and comorbidities. Long-term mortality seemed independent of microbiological aetiology.

M AN U

SC

RI PT

1

Conclusions:

11

Mortality remained high the first year after admission and thereafter decreased with time to a

12

level close to the reference population. Short-term mortality was especially related to infection

13

with abscess formation and neurological deficits and long-term mortality was related to having

14

alcohol dependency.

15

KEY WORDS:

16

Mortality; spine; Bone Diseases, Infectious; cause of death; prognosis; vertebral osteomyelitis;

17

factors associated with death

18

ABBREVIATIONS:

19

S. aureus – Staphylococcus aureus

20

CI – 95% Confidence interval

21

MRR – mortality rate ratio

22

HR – Hazard ratio

23

IQR – interquartile range, 25–75 percentiles

AC C

EP

TE D

10

24

2

ACCEPTED MANUSCRIPT

INTRODUCTION:

2

The incidence of infectious spondylodiscitis seems to be rising.1 2 This is presumably due to a

3

combination of improved diagnostic methods, the increasing elderly population, and a possibly

4

better clinical workup of patients with bacteraemia. Despite many reports concerning the clinical

5

characteristics and the epidemiology of infectious spondylodiscitis, there is still limited knowledge

6

about the prognosis in terms of mortality.

7

Descriptive studies of spondylodiscitis have reported mortality within different time periods. For

8

example the in-hospital mortality has been reported to be 4–6%1 3 and 27% among those aged ≥65

9

years.4 The overall one-year mortality has been estimated to be 11%.5-7 A study of spondylodiscitis

10

cases related to Staphylococcus aureus (S. aureus) bacteremia showed that 16% died within the

11

first months after admission8. Furthermore,another study described that 18% of spondylodiscitis

12

cases with positive culture of blood or biopsy with S. aureus died during the treatment period.9

13

Knowledge of long-term prognosis is still sparse. Only one study has previously compared long-

14

term prognosis with a reference population. They found that patients with S. aureus bacteremia

15

and subsequent spondylodiscitis had a mortality rate ratio of 1.3 compared with the background

16

population during the 11 years after admission, if they survived the first year.10 However, it is not

17

known whether the long-term prognosis of patients with infectious spondylodiscitis is

18

independent of the microbiological aetiology.

19

The aim of the present study was to describe short- and long-term mortality, factors associated

20

with death, and cause specific mortality rates among patients with infectious non-postoperative

21

spondylodiscitis, and to compare these outcome measures with those of a reference population.

AC C

EP

TE D

M AN U

SC

RI PT

1

3

ACCEPTED MANUSCRIPT

MATERIAL AND METHODS:

2

Study design and setting:

3

We conducted a case-cohort study on all incident cases of adult patients treated for infectious

4

non-postoperative spondylodiscitis during January 1994–May 2009 in hospitals in Funen County,

5

Region of Southern Denmark, Denmark (population of 483,123 June 1st, 2008 (Statistics

6

Denmark)).

7

As the county has a tertiary referral centre for patients with spinal infections from the Region of

8

Southern Denmark (population of 1.2 millions), the patients consisted of both residents of Funen

9

County and referred patients.

M AN U

SC

RI PT

1

The clinical characteristics of residents of Funen County diagnosed with spondylodiscitis from

11

1995–2008 have previously been described.2

12

Study population:

13

Spondylodiscitis patients: We used the Funen County Patient Administrative System to identify all

14

in- and out-patients aged ≥18 years with a discharge diagnosis of possible spinal infection (ICD-10

15

codes: A02.2; A18.0; A18.8; M46; B67.2; M49.0–3; M86; M90.0). For these patients we read and

16

evaluated all discharge summaries for information indicating spinal infection and if such

17

information was found, the full medical record was reviewed. The patient was included if the

18

combination of clinical features and results of diagnostic procedures were compatible with

19

infectious spondylodiscitis according to the consulting physicians and the patient was treated with

20

corresponding antibiotics. Cases due to tuberculosis or brucellosis were excluded as were

21

postoperative cases (i.e. registered spinal procedure within 12 months prior to index and believed

22

to be origin of infection).

AC C

EP

TE D

10

4

ACCEPTED MANUSCRIPT

Each case was assigned an index date defined as the admission date to the specific department in

2

which the diagnosis of spondylodiscitis was first suspected.

3

Reference population: Among all residents in Funen County, we selected a reference population

4

frequency matched on sex and 5-year age bands with 10 reference persons per case.11 Ensuring

5

that the longer persons living in the uptake area the higher probability of them becoming part of

6

the reference cohort, each reference person was given a random index date and included if

7

resident of Funen County at this date. Further, reference persons were allowed to become cases

8

but censored at date of becoming a case.

9

Data sources:

M AN U

SC

RI PT

1

Data from the following listed registries were linked using the unique 10-digit personal

11

identification number assigned to all residents of Denmark through The Danish Civil Registration

12

system since 1968.12 The Funen County Patient Administrative System was used in the case

13

identification process and the Danish National Registry of Patients,13 the Danish Cancer Registry,14

14

the National Registry of Drug Abusers undergoing Treatment,15 the National Register of Alcohol

15

Abuse Treatment,16 and Odense Pharmacoepidemiological Database,17 were used in assessment

16

of comorbidities.

17

We further used the Cause of death register18 for information on underlining cause of death. See

18

Appendix A for detailed information on used registries.

19

Furthermore, apart from case defining criteria, the hospital records were reviewed for various

20

variables.

AC C

EP

TE D

10

5

ACCEPTED MANUSCRIPT

Exposures and possible confounders:

2

The microbiological aetiology was based on having a positive culture or PCR with a recognized

3

pathogen from blood, pus, bone, intervertebral disc, or spinal fluid. Remaining cases were defined

4

as having unknown microbiological aetiology.

5

Over-all comorbidity was scored according to the Charlson Comorbidity index19 based on discharge

6

diagnoses (Appendix B).

7

Immunosuppression was defined as having been registered with a diagnosis indicating having an

8

immunosuppressive disorder (e.g. primary immunodeficiency disorder or received an organ

9

transplantation), having redeemed prescription of systemic glucocorticoids, or been diagnosed

10

with a solid cancer within 1 year before index date. Alcohol abuse was defined as a registered

11

diagnosis indicating alcohol dependency, redeemed prescription of disulfiram, or being registered

12

in The National Register of Alcohol Abuse Treatment. Intravenously drug use was defined as

13

having a registered diagnosis indicating intravenous drug use or being registered in The National

14

Registry of Drug Abusers undergoing Treatment. For details on the definition of these three

15

variables see Appendix B.

16

Causes of deaths were grouped according to underlining cause of death based on ICD-10 codes as

17

indicated in Appendix C. The following groups were used: infectious diseases, malignant

18

neoplasms, in situ/benign neoplasm, blood/immune diseases, endocrine diseases, mental

19

diseases/drug abuse, nervous system diseases, cardiovascular diseases, respiratory diseases,

20

digestive system diseases, musculoskeletal / skin disorders, genitourinary diseases; injury/poison,

21

and ill-defined causes.

22

From the Danish Civil Registration System: resident or referred patient, date of death, sex and age.

AC C

EP

TE D

M AN U

SC

RI PT

1

6

ACCEPTED MANUSCRIPT

Furthermore, variables from medical records: presenting region of spinal involvement, presence of

2

severe neurological impairment at time of admission (i.e., signs of cauda equine or spinal cord

3

injury), epidural or intraspinal abscess, and period of admission (predefined as 1994–2002, 2003–

4

2009).

5

Outcome:

6

Over-all mortality was the primary outcome. Cause of death and factors associated with mortality

7

were also assessed.

8

Data management:

9

All data from the medical records were sampled in a database using EpiData (Epidata Association,

10

Denmark). Data were manually checked for logical errors or missing values, and outliers were

11

rechecked. Case identification was performed by author MK and in case of doubt in collaboration

12

with author ATL.

13

Ethics:

14

The study was approved by the Danish Data Protection Agency (2008-41-2527 and 2008-58-0035)

15

and the Danish Health and Medicines Authority (7-604-04-2/118). In accordance with Danish law,

16

observational studies performed in Denmark do not need approval from the Medical Ethics

17

Committee.

18

Follow up period:

19

The cohort was followed from index date until death, emigration, lost to follow-up, becoming a

20

case, completion of 14 years of follow-up, or 1 June 2012, whichever came first.

21

Analysis:

22

Short-term mortality was defined as deaths within first years after admission and long-term

23

mortality was deaths thereafter.

AC C

EP

TE D

M AN U

SC

RI PT

1

7

ACCEPTED MANUSCRIPT

All-cause mortality analysis:

2

All-cause mortality was presented in a Kaplan-Meier plot. Within 90 days, 91–365 days, 1–4 years,

3

and 5–14 years, all-cause mortality rates and unadjusted and adjusted mortality rate ratios (MRR)

4

were calculated. The latter was determined using multivariate Cox regression analysis controlling

5

for predefined variables: sex, age, calendar effect and Charlson comorbidity index.

6

Factors associated with death:

7

First, predefined factors possibly associated with death among cases were evaluated by hazard

8

ratio (HR) using univariate Cox regression analysis. Thereafter factors with p-value ≤0.2 were used

9

in a multivariate Cox regression model. Stepwise backward selection method was used for the

M AN U

SC

RI PT

1

final regression model and p-value for remaining in the model was ≤.05 using likelihood-ratio test.

11

To avoid collinearity, comorbidity was not tested using Charlson score in the multivariate analysis.

12

Sex and age were included in the model as stratifying variables and calendar effect in years as a

13

continuous variable.

14

Furthermore, stratified MRRs on possible factors associated with death comparing rates among

15

cases and reference population were determined and described in a supplementary table.

16

Cause specific mortality:

17

Unadjusted cause-specific mortality rates and rate ratios for deaths from index until 31 December

18

2010 or end of follow-up, whichever came first, were determined.

EP

AC C

19

TE D

10

20

All statistical analyses were performed using Stata® (StataCorp, College Station, TX, USA, version

21

12.1). Proportional hazard assumptions were checked; results are not presented.

22

8

ACCEPTED MANUSCRIPT

RESULTS:

2

We identified 1,386 potential cases from the Funen County Patient Administrative System and 298

3

patients fulfilled the case definition criteria. The reference population consisted of 2,980 persons

4

of whom none later became cases.

5

In the reference population 12 persons emigrated during follow-up. There was no other loss to

6

follow-up. Baseline characteristics of cases and reference population are shown in Table 1. The

7

total time at risk was 1601.8 person years for the cases (median: 4.6; range: 0–14) and 25,939.7

8

person years for the reference population (median: 8.4; range: 0–14).

9

All-cause mortality:

M AN U

SC

RI PT

1

During the entire observation period 141/298 (47%) of the cases died and of these 61/298 (20%)

11

died within the first year. Kaplan-Meier curves are shown in Figure 1 with the overall estimated

12

survival probability for the reference population and stratified by microbiological aetiology for the

13

cases.

14

Table 2 presents the crude mortality rates and the crude and adjusted MRRs in predefined

15

observation periods; the excess mortality for the cases decreased with time.

16

Factors associated with death:

17

Among the cases the main factors associated with short-term mortality were age, comorbidity,

18

and severe neurological deficits at time of admission. These and the rest of the results of the

19

univariate analyses of factors associated with death are presented in Figure 2.

20

In the multivariate analyses of factors associated with short-term mortality the following variables

21

were independently associated with increased risk of death: diabetes: (HR=2.4, 95% confidence

22

interval (CI): 1.3–4.4); immunosuppressive disorder (HR=2.1; CI: 1.1–4.0); epidural abscess

23

(HR=1.9, CI:1.1–3.3); severe neurological deficits at time of admission (HR= 2.3, CI: 1.1–4.8); and

AC C

EP

TE D

10

9

ACCEPTED MANUSCRIPT

being non-referred patient (HR= 2.0, CI: 1.1–3.6). Further, having unknown microbiological

2

aetiology was associated with decreased mortality compared with having S. aureus infection

3

(HR=0.4, CI: 0.2–1.0). There was no difference between having infection with S. aureus and other

4

bacteria (HR= 0.8, CI: 0.4–1.4). Alcohol abuse and immunosuppressive disorders were the only

5

independent factors associated with long-term mortality: HR= 3.5 (CI: 1.7–7.0) and HR=2.4 (CI:

6

1.2–5.0), respectively.

7

MRRs estimating factors associated with death among cases compared with reference population

8

are presented in Appendix Table D.1 and Table D.2.

9

Cause specific mortality:

M AN U

SC

RI PT

1

Table 3 presents the cause specific mortality rates and rate ratios. Within the first year, the most

11

common cause of death was disorders related to spondylodiscitis. Among patients surviving the

12

first year, deaths due to cardiovascular or malignant diseases were the most common.

13

DISCUSSION:

14

In this study describing short- and long-term mortality among adult patients with infectious

15

spondylodiscitis compared with a population-based reference population, we found that patients

16

with spondylodiscitis had significantly increased both short- and long-term mortality.

17

We defined short-term mortality as deaths within the first year of admission and long-term

18

mortality as deaths thereafter. Compared with other studies, we found higher 1-year mortality

19

(20% vs 11%).5-7 Our patients were older than in the other studies (two of the studies included

20

children,6 7 and two studies were from tertiary referral centres5 7). Differences in patient selection

21

probably explain some of the differences in mortality rate. Also lack of complete follow-up in other

22

studies may contribute to the difference6 as loss of follow-up may mask deaths. However, this is

23

difficult to assess since many studies do not account for loss to follow-up. Since increasing age is a

AC C

EP

TE D

10

10

ACCEPTED MANUSCRIPT

strong predictor of mortality and the treatment modalities were presumably similar to the other

2

studies,5 7 we do not believe that differences in treatment modalities have contributed

3

substantially to differences in mortality.

4

In the present study, cases with unknown microbiological aetiology had a better 1-year prognosis

5

than those with known microbiological aetiology. Whether this is due to misdiagnosis by the

6

clinicians is uncertain. It was in accordance with studies finding higher mortality among only S.

7

aureus cases8 than in studies among cases with mixed and unknown aetiology.5 No studies have

8

compared the mortality of culture positive cases with that of cases with unknown microbiological

9

aetiology. Whether or not related to the persons or the type of infection, the high short-term

10

mortality among both cases with known and unknown aetiology revealed the need for timely

11

diagnosing and treatment of all patients with infectious spondylodiscitis. In line with other studies,

12

co-morbidities,1 age,1 and having severe neurological deficits at time of admission6 were all

13

associated with increased risk of death within the first year.

14

We found the mortality remained high the first year after admission and thereafter decreased

15

with time to a level close to the reference population. As studies with long-term prognosis in this

16

field are sparse, few studies are available for direct comparison of our results. Other studies have

17

found increased long-term mortality among patients with spinal cord injury20 as well as among

18

patients S. aureus spondylodiscitis10 or patients with septicaemia.21 Whether this is related to low-

19

grade chronic inflammation, which could increase the risk of illnesses such as diabetes, malignancy

20

or vascular diseases, or whether this is due to residual confounding or unmeasured confounding

21

(e.g., smoking) is unknown.

22

In the present study, the only significant factors associated with long-term mortality were having

23

alcohol dependency or immunosuppressive disorders. This was as expected as excess alcohol

AC C

EP

TE D

M AN U

SC

RI PT

1

11

ACCEPTED MANUSCRIPT

consumption is known to increase long-term mortality.22 We found no indication of long-term

2

prognosis being different for cases with unknown aetiology compared with cases with known

3

microbiological aetiology. We presented data comparing mortality rates among patients with

4

infectious spondylodiscitis with those of a reference population. However, we only presented

5

factors associated with death among cases as we included factors only related to the disease

6

therefore not measureable among the reference population.

7

A possible limitation of this study was that we included both local and referred patients;

8

nevertheless, univariate analysis of risk factors suggested that residents had higher short-term

9

mortality than referred patients. Estimates from tertiary centres with a majority of referred

10

patients may therefore underestimate the short-term mortality. This is presumably because only

11

“likely survivors” are referred to tertiary centres. Another limitation is the retrospective case

12

inclusion, although we attempted to mitigate this by a thorough case review. Furthermore, the

13

number of included cases is limited and type 2 errors due to limited power are possible. High

14

levels of comorbidities are major confounders in mortality analyses.19 As the definitions of the

15

comorbidities were based on registration with discharge diagnoses or intake of prescription drugs,

16

known predisposing diseases diagnosed by the general practitioner may have been missed and

17

contributed to unmeasured confounding. Also, we were not able to adjust for socioeconomical

18

factors. However, this is one of the larger published cohorts, and the thorough case validation and

19

complete and long follow-up are important strengths.

20

Conclusions:

21

In conclusion, we found that patients with infectious spondylodiscitis have both an increased

22

short- and long-term mortality. Short-term mortality was especially related to infection with

AC C

EP

TE D

M AN U

SC

RI PT

1

12

ACCEPTED MANUSCRIPT

1

abscess formation and neurological deficits and long-term mortality was related to having alcohol

2

dependency.

RI PT

3

4

CONFLICTS OF INTEREST:

6

None.

AC C

EP

TE D

M AN U

SC

5

13

ACCEPTED MANUSCRIPT

References:

AC C

EP

TE D

M AN U

SC

RI PT

1. Akiyama T, Chikuda H, Yasunaga H, et al. Incidence and risk factors for mortality of vertebral osteomyelitis: a retrospective analysis using the Japanese diagnosis procedure combination database. BMJ open 2013;3(3). 2. Kehrer M, Pedersen C, Jensen TG, et al. Increasing incidence of pyogenic spondylodiscitis: A 14-year population-based study. The Journal of infection 2014;68(4):313-20. 3. Bhavan KP, Marschall J, Olsen MA, et al. The epidemiology of hematogenous vertebral osteomyelitis: a cohort study in a tertiary care hospital. BMCInfectDis 2010;10:158. 4. Hutchinson C, Hanger C, Wilkinson T, et al. Spontaneous spinal infections in older people. InternMedJ 2009;39(12):845-48. 5. Aagaard T, Roed C, Dragsted C, et al. Microbiological and therapeutic challenges in infectious spondylodiscitis: a cohort study of 100 cases, 2006-2011. Scandinavian journal of infectious diseases 2013;45(6):417-24. 6. McHenry MC, Easley KA, Locker GA. Vertebral osteomyelitis: long-term outcome for 253 patients from 7 Cleveland-area hospitals. ClinInfectDis 2002;34(10):1342-50. 7. Karadimas EJ, Bunger C, Lindblad BE, et al. Spondylodiscitis. A retrospective study of 163 patients. Acta Orthop 2008;79(5):650-59. 8. Jensen AG, Espersen F, Skinhoj P, et al. Bacteremic Staphylococcus aureus spondylitis. ArchInternMed 1998;158(5):509-17. 9. Priest DH, Peacock JE, Jr. Hematogenous vertebral osteomyelitis due to Staphylococcus aureus in the adult: clinical features and therapeutic outcomes. Southern medical journal 2005;98(9):854-62. 10. Aagaard T, Roed C, Larsen AR, et al. Long-term mortality after Staphylococcus aureus spondylodiscitis: A Danish nationwide population-based cohort study. The Journal of infection 2014. 11. Rothman KJ, Greenland S, Lash TL. Modern Epidemiology , 3rd Edition: Lippincott Williams & Wilkins, 2008. 12. Pedersen CB. The Danish Civil Registration System. Scandinavian journal of public health 2011;39(7 Suppl):22-5. 13. Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient Register. Scandinavian journal of public health 2011;39(7 Suppl):30-3. 14. Storm HH, Michelsen EV, Clemmensen IH, et al. The Danish Cancer Registry--history, content, quality and use. Danish medical bulletin 1997;44(5):535-9. 15. The National Registry of Drug Abusers undergoing Treatment. Secondary The National Registry of Drug Abusers undergoing Treatment 19 September 2012. http://www.ssi.dk/Sundhedsdataogit/Indberetning%20og%20patientregistrering/SEI/Register%20o ver%20stofmisbrugere%20i%20behandling%20SIB.aspx. 16. The National Register of Alcohol Abuse Treatment. Secondary The National Register of Alcohol Abuse Treatment September 19 2012. http://www.ssi.dk/Sundhedsdataogit/Indberetning%20og%20patientregistrering/SEI/Det%20Natio nale%20Alkoholbehandlingsregister%20NAB.aspx. 17. Gaist D, Sorensen HT, Hallas J. The Danish prescription registries. Danish medical bulletin 1997;44(4):445-8. 18. Institut SSS. Registerdeklaration for Dødsårsagsregisteret. Secondary Registerdeklaration for Dødsårsagsregisteret January 2013 2013. http://www.ssi.dk/Sundhedsdataogit/Registre/~/media/Indhold/DK%20%20dansk/Sundhedsdata%20og%20it/NSF/Registre/Dodsaarsagsregisteret/Registerdeklaration%20 D%C3%B8ds%C3%A5rsagsregisteret.ashx. 19. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. Journal of chronic diseases 1987;40(5):373-83.

14

ACCEPTED MANUSCRIPT

M AN U

SC

RI PT

20. Strauss DJ, Devivo MJ, Paculdo DR, et al. Trends in life expectancy after spinal cord injury. Archives of physical medicine and rehabilitation 2006;87(8):1079-85. 21. Storgaard M, Hallas J, Gahrn-Hansen B, et al. Short- and long-term mortality in patients with community-acquired severe sepsis and septic shock. Scandinavian journal of infectious diseases 2013;45(8):577-83. 22. Di Castelnuovo A, Costanzo S, Bagnardi V, et al. Alcohol dosing and total mortality in men and women: an updated meta-analysis of 34 prospective studies. Archives of internal medicine 2006;166(22):2437-45.

FIGURE LEGENDS:

Figure 1: Kaplan-Meier curves corresponding to the microbiological aetiology and for the

EP

survival time.

TE D

reference population. Below the curves is listed the number at risk at corresponding intervals in

AC C

Figure 2: Results of univariate Cox regression analysis of prognostic factors for death among patients with infectious spondylodiscitis. Confidence intervals above one favours association. It was not possible to estimate impact of liver cirrhosis after the first year due to limited number events.

Abbrevation: ES: estimates of mortality rate ratio. CI: 95% confidence interval. *Reference group is 1994–2002. **S. aureus infection is reference group. ***Reference group is the lumbar region.

15

ACCEPTED MANUSCRIPT

Table 1: Baseline characteristics of cases and reference population. Reference population

N=298

N=2,980

n (%)

n (%)

Men

182 (61.1)

1,820 (61.1)

Women

116 (38.9)

1,160 (38.9)

Age: median (IQR)

66 (55–75)

66 (55–75)

M AN U

Charlson comorbidity score

SC

Comorbidity:

RI PT

Cases

0

166 (55.7)

2,328 (78.1)

1

47 (15.8)

306 (10.3)

≥2

85 (28.5)

346 (11.6)

6 (2.0)

7 (0.2)

TE D

Liver cirrhosis Diabetes

44 (14.8)

188 (6.3)

32 (10.7)

166 (5.6)

34 (11.4)

79 (2.7)

16 (5.4)

2 (0.1)

1994-2002

114 (38.3)

1,754 (58.9)

2003-2009

184 (61.7)

1,226 (41.1)

Immunosuppression

IDU

AC C

Period of admission:

EP

Alcohol dependency

Severe neurological deficits at

33 (11.1)

time of admission*

Region of bone infection: Cervical

38 (12.8)

95 (31.9)

Lumbar

203 (68.1)

More than one region

35 (11.7)

Epidural or intraspinal abscess

92 (30.9)

Microbiological aetiology: S. aureus

127 (42.6)

Other bacteria**

103 (34.6)

Unknown

68 (22.8)

SC

Thoracic

RI PT

ACCEPTED MANUSCRIPT

M AN U

Abbreviations: IQR interquartile range 25 percentile–75 percentile; IDU intravenously drug user *Severe neurological deficits at time of admission included signs of cauda equina or spinal cord injury. **Other bacteria included 12 cases of coagulase-negative Staphylococci species; 41 Streptococci spp.; 26 Escherichia coli; 15 other Gram-negative bacteria; 3 other Gram-positive bacteria; 6 Enterococcus faecalis; and 8 polybacterial

AC C

EP

TE D

cases.

ACCEPTED MANUSCRIPT

Table 2. Mortality rate and rate ratios at 0–90 days, 91–365 days, 1–4 years, and 5–15 years among patients with infectious

RI PT

spondylodiscitis compared with the reference population. The last three columns are the adjusted mortality rate ratios stratified by microbiological aetiology.

deaths

Person

Overall unadjusted

Overall

years

MR

unadjusted MRR

at risk

(95% CI)

(95% CI)

0–90 days

298

40

17.6

24

Other bacteria

adjusted MRR*

aetiology

Adjusted MRR*

Adjusted MRR*

Adjusted MRR*

(95% CI)

(95% CI)

(95% CI)

(95% CI)

24.0

14.8

(10.6–29.2)

(9.9–28.5)

(3.5–22.1)

(13.0–44.2)

(7.0–31.4)

1

1

1

1

1

114.5

3.9

4.2

1.4

6.0

4.7

(74.7–175.6)

(2.4–6.5)

(2.5–7.0)

(0.3–5.8)

(3.1–11.5)

(2.2–10.0)

1

1

1

1

1

32.9 2,980

S. aureus

8.8

(430.3–799.7)

Reference population

Unknown

16.8

68.2

TE D

586.6 Cases

Overall

SC

at risk

Total

M AN U

Number

730.5

EP

(22.0–49.0)

Cases

258

21

183.4

AC C

91–365 days

29.0

Reference population

2,956

63

2,174.2

(22.6–37.1)

ACCEPTED MANUSCRIPT

1–4 years

237

50

62.5

2.0

2.2

(47.4–82.5)

(1.5–2.7)

(1.6–2.9)

1

1

31.3 Reference population

2,893

332

10,596.6

5–14 years

30

2243

471

1.8

(2.0–5.1)

(1.1–3.2)

(1.1–3.0)

1

1

1

1.1

2.7

1.6

1.5

(38.1–77.9)

(1.0–2.1)

(1.2–2.5)

(0.5–2.4)

(1.5–4.7)

(0.8–2.9)

1

1

1

1

1

37.9 Reference population

1.9

54.5 550.7

12,438.4 (34.6–41.5)

TE D

Abbreviations: MR, mortality rate (deaths/1000 person years); CI, confidence interval. Mortality rate ratios were determined using Cox regression models.

EP

*Adjusted for sex, age, calendar effect in years, and Charlson Comorbidity score (0, 1, ≥2).

AC C

1.7

M AN U

135

SC

(28.1–34.9)

Cases

3.2

799.5

RI PT

Cases

ACCEPTED MANUSCRIPT

Table 3: Cause-specific mortality rates among patients with spondylodiscitis and reference popylation and mortality rate ratios stratified

RI PT

on short-term (within first year) or long-term (from 1 year after index until December 31, 2010 or end of follow-up). Deaths caused by

SC

infections spondylodiscitis may have been recorded as either “Infectious diseases” or “Musculoskeletal / skin disorders”.

M AN U

Short-term mortality – deaths within the first year from admission Patients with spondylodiscitis Deaths

Musculoskeletal / skin disorders^

16

Cardiovascular diseases

9

Infectious diseases

8

Respiratory diseases

6

Nervous system diseases Malignant diseases Injury / poisoning

242.5 (188.7–311.6)

Deaths

Mortality rate ratio** Mortality rate*

(95% CI) (95% CI)

87

30.0 (24.3–37.0)

11.4 (5.6–23.2)***

63.6 (39.0 - 103.8)

0

-

-

35.8 (18.6 - 68.8)

23

7.9 (5.3 - 11.9)

4.4 (2.1 - 9.6)

31.8 (15.9 - 63.6)

3

1.0 (0.3 - 3.2)

2.2 (.5 - 10.2)***

23.8 (10.7 - 53.1)

8

2.8 (1.4 - 5.5)

8.6 (3.0 - 24.8)

6

23.8 (10.7 - 53.1)

2

0.7 (0.2 - 2.8)

0.8 (0.2–4.0)***

4

15.9 (6.0 - 42.4)

25

8.6 (5.8 - 12.7)

1.8 (0.6 - 5.2)

4

15.9 (6.0 - 42.4)

3

1.0 (0.3 - 3.2)

15.3 (3.4 - 68.3)

TE D

61

AC C

Overall

Mortality rate* (95% CI)

EP

Cause of death

Reference population

ACCEPTED MANUSCRIPT

3

11.9 (3.8 - 37.0)

5

1.7 (0.7 - 4.1)

6.8 (1.6 - 28.4)

Blood / immune diseases

1

4.0 (0.6 - 28.2)

0

-

-

Endocrine diseases

1

4.0 (0.6 - 28.2)

5

Ill-defined causes

1

4.0 (0.6 - 28.2)

7

In situ / benign neoplasms

1

4.0 (0.6 - 28.2)

1

Genitourinary diseases

1

4.0 (0.6 - 28.2)

Mental disorders / drug abuse

0

-

RI PT

Digestive system diseases

2.2 (0.3 - 19.1)

2.4 (1.1 - 5.1)

0.0 (0.0 - 0.0)

0.3 (0.0 - 2.4)

11.4 (0.7 - 182.7)

SC

1.7 (0.7 - 4.1)

M AN U

1 4

0.3 (0.0 - 2.4)

11.3 (0.7 - 181.2)

1.4 (0.5 - 3.7)

-

Long-term mortality – recorded deaths after the first year from admission Cause of death

Patients with spondylodiscitis

Reference population

Deaths

(95% CI)

Mortality rate ratio** (95% CI)

57.9 (45.4–73.9)

678

33.1 (30.7–35.7)

1.8 (1.4–2.3)

20.5 (13.6–30.9)

224

10.9 (9.6–12.5)

1.9 (1.3–3.0)

16.0 (10.1–25.5)

188

9.2 (8.0–10.6)

1.8 (1.1–2.9)

6

5.3 (2.4–11.9)

83

4.1 (3.3–5.0)

1.4 (0.6–3.2)

5

4.5 (1.9–10.7)

28

1.4 (0.9–2.0)

3.1 (1.2–8.1)

3

2.7 (0.9–8.3)

26

1.3 (0.9–1.9)

2.2 (0.6–7.2)

65

Cardiovascular diseases

23

Malignant diseases

18

Digestive system diseases Endocrine diseases

AC C

Overall

Respiratory diseases

Mortality rate*

(95% CI)

EP

Deaths

TE D

Mortality rate*

ACCEPTED MANUSCRIPT

2

1.8 (0.4–7.1)

26

1.3 (0.9–1.9)

1.5 (0.3–6.2)

Injury / poisoning

2

1.8 (0.4–7.1)

17

0.8 (0.5–1.3)

1.7 (0.4–7.6)

Musculoskeletal / skin disorders^

1

0.9 (0.1–6.3)

3

Infectious diseases

1

0.9 (0.1–6.3)

9

Genitourinary diseases

1

0.9 (0.1–6.3)

8

Blood / immune diseases

1

0.9 (0.1–6.3)

4

Ill-defined causes

1

0.9 (0.1–6.3)

Nervous system diseases

1

In situ / benign neoplasms

0

RI PT

Mental disorders / drug abuse

5.1 (0.5–49.1)

0.4 (0.2–0.8)

2.2 (0.3–17.3)

0.4 (0.2–0.8)

2.5 (0.3–19.9)

SC

0.1 (0.0–0.5)

4.2 (0.5–38.5)

32

1.6 (1.1–2.2)

0.6 (0.1–4.4)

0.9 (0.1–6.3)

24

1.2 (0.8–1.7)

0.8 (0.1–6.3)

-

6

0.3 (0.1–0.7)

-

M AN U

0.2 (0.1–0.5)

TE D

*Unadjusted mortality rate per 1000 person years.

**Unadjusted mortality rate ratio estimated using a Cox regression model.

***Estimated with time-dependency, due to non-fulfilled proportional hazard assumptions.

EP

^Deaths recorded in the category of musculoskeletal / skin disorders within the first year after admission was only related to infectious or

reference persons.

AC C

inflammatory spine disorders – i.e. spondylodiscitis; thereafter the category included gout among one patient and rheumatic arthritis among three

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

0-365 days level

ES (95% CI)

P-value

Sex

men

1.13 (0.67, 1.91)

0.635

Age

aged above 65 years

2.10 (1.53, 2.88)

0.000

Patient category

residents

2.05 (1.17, 3.59)

0.012

Charlson comorbidity index

equal or more than 2

2.17 (1.31, 3.60)

0.003

Diabetes mellitus

yes

2.30 (1.30, 4.06)

0.004

Liver cirrhosis

yes

4.54 (1.64, 12.57) 0.004

Immunosuppresive disorder

yes

2.75 (1.51, 4.99)

0.001

Alcohol dependency

yes

0.80 (0.35, 1.87)

0.613

Intravenously drug abuse

yes

0.26 (0.04, 1.89)

0.183

Year of admission*

2003-2009

Microbiological etiology**

other bacteria

SC

RI PT

Variable

unknown Region of infection***

thoracic more than one region Severe neurological deficits

yes

Epidural or intra spinal abscess yes

0.64 (0.37, 1.12)

0.120

0.40 (0.19, 0.87)

0.021

0.89 (0.35, 2.27)

0.806

1.05 (0.56, 1.96)

0.879

1.13 (0.53, 2.45)

0.748

1.99 (1.03, 3.82)

0.039

1.55 (0.93, 2.59)

0.094

1

2

3

4

5

TE D

0

0.330

M AN U

cervical

0.78 (0.47, 1.29)

1-14 years

ES (95% CI)

P-value

Variable

level

Sex

men

Age

aged above 65 years

Patient category

residents

1.11 (0.71, 1.73) 0.659

Charlson comorbidity index

equal or more than 2

1.45 (0.89, 2.37) 0.139

Diabetes mellitus

yes

0.97 (0.46, 2.01) 0.927

1.74 (1.06, 2.84) 0.027

AC C

EP

1.59 (1.27, 2.00) 0.000

Immunosuppresive disorder

yes

2.43 (1.20, 4.92) 0.013

Alcohol dependency

yes

2.11 (1.16, 3.83) 0.014

Intravenously drug abuse

yes

1.68 (0.77, 3.65) 0.194

Year of admission*

2003-2009

1.12 (0.68, 1.84) 0.652

other bacteria

1.03 (0.60, 1.75) 0.924

unknown

1.67 (0.98, 2.85) 0.061

cervical

1.19 (0.53, 2.65) 0.670

thoracic

1.06 (0.63, 1.80) 0.819

more than one region

1.18 (0.57, 2.44) 0.648

yes

1.07 (0.51, 2.22) 0.859

Epidural or intra spinal abscess yes

0.80 (0.48, 1.34) 0.396

Microbiological etiology**

Region of infection***

Severe neurological deficits

0

1

2

3

4

5

ACCEPTED MANUSCRIPT

Appendix A: Description of used registries: The Danish civil registration system:

RI PT

The registry contains information on date of birth, sex, vital status, residence, and immigration and emigration status of all persons with a permanent residence in Denmark1. The unique personal

identification number assigned to all persons with a permanent residence in Denmark from this register makes it possible to link data from other registries or medical records.

SC

The Funen County Patient Administrative System:

Since 1977, this database has registered information on all in- and out-patients contacts hospital contacts in Funen County and reports data to the Danish National Registry of Patients. The registration includes date

M AN U

of contact, discharge diagnoses, surgical procedures, and microbiological findings. During 1977–1993 diagnoses were classified according to the ICD-8 system; thereafter according to the ICD-10 system. In this study the registry has been used in the case identification. The Danish National Registry of Patients:

A registry of prospectively collected data on all hospital contacts in Denmark2. As in the Funen County Patient Administrative System, diagnoses were classified according to the ICD-8 system during 1977–1993;

TE D

thereafter according to the ICD-10 system.

The registry has been used in definition of comorbidities. The Danish Cancer Registry:

The registry contains information on all incident cancer diagnoses among Danish residents since 1943.

EP

Since 1978 the cancers has been coded using ICD-103. In this study, the registry was used in the definition of having an immunosuppressive disorder. The National Registry of Drug Abusers undergoing Treatment:

AC C

Run by the National Board of Health and has registered all persons receiving treatment for drug abuse since 19964. In this study the registry was used in the definition of intravenous drug users.

1

Pedersen CB. The Danish Civil Registration System. Scandinavian journal of public health. 2011;39(7 Suppl):22-5 Lynge E et al. The Danish National Patient Register. Scandinavian journal of public health. 2011;39(7):30-3 3 Storm HH et al. The Danish Cancer Registry-history, content, quality and use. Danish medical bulletin. 1997;44(5):535-9 4 The National Registry of Drug Abusers undergoing Treatment Statens Serum Institut: Statenst Serum Institut; [updated 19 September 2012; cited 2013 27 August]. Available from: http://www.ssi.dk/Sundhedsdataogit/Indberetning%20og%20patientregistrering/SEI/Register%20over%20stofmisbru gere%20i%20behandling%20SIB.aspx 2

1

ACCEPTED MANUSCRIPT

The National Register of Alcohol Abuse Treatment: A registry of all persons receiving treatment for alcohol dependence paid by the national health system since 2006. Privately paid treatments may be registered as well5. In this study the registry was used in the definition of persons with alcohol dependency.

RI PT

Odense Pharmacoepidemiological Database:

A research database containing information on redeemed, reimbursed prescription drugs for all residents

comorbidities. Cause of death register

SC

of Funen County since November 19926. In this study, the database was used in the definition of

Since 1970, this register has collected information on all deaths of persons with permanent residence in Denmark. Since 1984, the suspected underlining cause of death has been registered using ICD-107. In this

5

AC C

EP

TE D

M AN U

study, the registry was used in the definition of cause of death.

The National Register of Alcohol Abuse Treatment Statens Serum Institut [updated September 19 2012]. Available from: http://www.ssi.dk/Sundhedsdataogit/Indberetning%20og%20patientregistrering/SEI/Det%20Nationale%20Alkoholbe handlingsregister%20NAB.aspx 6 Gaist D et al. The Danish prescription registries. Danish medical bulletin. 1997;44(4):445-8 7 Institut SSS. Registerdeklaration for Dødsårsagsregisteret: Statens Serum Institut, Denmark; 2013 [updated January 2013; cited 2013 December 13]. Available from: http://www.ssi.dk/Sundhedsdataogit/Registre/~/media/Indhold/DK%20%20dansk/Sundhedsdata%20og%20it/NSF/Registre/Dodsaarsagsregisteret/Registerdeklaration%20D%C3%B8ds%C3% A5rsagsregisteret.ashx

2

ACCEPTED MANUSCRIPT

Appendix B:

Charlson Score:

RI PT

Definition of variables:

A weighted index score assessing comorbidities1 based on the following discharge diagnoses from 10 years before index until 7 days before index: Myocardial infarction:

ICD-8: 410*. ICD-10: I21*, I22*, I23*.

Congestive heart failure:

ICD-8: 42709*, 42710*, 42711*, 42719*, 42899, 78249*. ICD-10: I50, I110*,

SC

I130*, I132*.

Peripheral vascular disease: ICD-8: 440*, 441*, 442*, 443*, 444*, 445*. ICD-10: I70*, I71*, I72*, I73*,

Cerebrovascular disease:

M AN U

I74*, I77*.

ICD-8: 430*, 431*, 432*, 433*, 434*, 435*, 436*, 437*, 438*. ICD-10: I60*, I61*, I62*, I63*, I64*, I65*, I66*, I67*, I68*, I69*, G45*, G46*.

Dementia

ICD-8:290*, 29309*. ICD-10: F00*, F01*, F02*, F03*, F051*, G30*.

Chronic pulmonary disease: ICD-8: 490*, 491*, 492*, 493*, 515*, 516*, 517*, 518*. ICD-10: J40*, J41*, J42*, J43*, J44*, J45*, J46*, J47*, J60*, J61*, J62*, J63*, J64*, J65*, J66*, J67*, J684*, J701*, J703*, J841*, J920*, J961*, J982*, J983*. ICD-8: 712*, 716*, 734*, 446*, 13599*. ICD-10: M05*, M06*, M08*, M09*,

TE D

Connective tissue disease:

M30*, M31*, M32*, M33*, M34*, M35*, M36*, D86*.

Mild liver disease:

ICD-8: 24900*, 24906*, 24907*, 24909*, 25000*, 25006*, 25007*, 25009*. ICD-10: E100*, E101*, E109*, E110*, E111*, E119*. Hemiplegia: ICD-8: 344*. ICD-10: G81*, G82*. Moderate to severe renal disease: ICD-8: 403*, 404*, 580*, 581*, 582*, 583*, 584*, 59009*, 59319*, 75310*, 75311*, 75319*, 792*. ICD-10: I12*, I13*, N00*, N01*, N02*, N03*, N04*, N05*, N07*, N11*, N14*, N17*, N18*, N19*, Q61. Diabetes with end-organ damage: ICD-8: 24901*, 24902*, 24903*, 24904*, 24905*, 24908*, 25001*, 25002*, 25003*, 25004*, 25005*, 25008*. ICD-10: E102*, E103*, E104*, E105*, E106*, E107*, E108*, E112*, E113*, E114*, E115*, E116*, E117*, E118*.

AC C

Diabetes:

ICD-8: 53091*, 53098*, 531*, 532*, 533*, 534*. ICD-10: K221*, K25*, K26*, K27*, K28*. ICD-8: 571*, 57301*, 57304*. ICD-10: B18*, K700*, K701*, K702*, K703*, K709*, K71*, K73*, K74*, K760*.

EP

Ulcer diseases:

1

Charlson ME et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. Journal of chronic diseases. 1987;40(5):373-83

1

ACCEPTED MANUSCRIPT

ICD-8: 140*, 194*. ICD-10: C00*, C75*. ICD-8: 204*, 205*, 206*, 207*. ICD-10: C91*, C92*, C93*, C94*, C95. ICD-8: 200*, 201*, 202*, 203*, 27559*. ICD-10: C81*, C82*, C83*, C84*, C85*, C88*, C90*, C96*. Moderate to severe liver disease: ICD-8: 07000*, 07002*, 07004*, 07006*, 07008*, 57300*, 45600*, 45601*, 45609*. ICD-10: B150*, B160*, B162*, B190*, K704*, K72*, K766*, I85*. Metastatic solid tumour: ICD-8: 195*, 196*, 197*, 198*, 199*. ICD-10: C76*, C77*, C78*, C79*, C80*. AIDS: ICD-8: 07983*. ICD-10: B21*, B22*, B23*, B24*.

Prior immunosuppression Defined as fulfilment of one of the following criteria: •

SC

RI PT

Any tumour: Leukaemia: Lymphoma:

Redeemed prescriptions of prednisolone at a moderate to high intake (cumulative defined daily

•

M AN U

dose ≥1 in the 120 days prior to index date with a compliance rate of 80%; ATC: H02AB*) Redeemed prescriptions of immunosuppressive treatment in 4 months prior to index date (cumulative defined daily dose ≥1 in the 120 days prior to index date with a compliance rate of 80%; ATC: L0*) •

Registered with a discharge diagnosis suggestive of human immunodeficiency virus (ICD-10: B20– B24, F024.0-F024.4, and Z21.0-Z21.4).

Registered with a discharge diagnosis suggestive of a primary immunodeficiency disorder (ICD-10:

TE D

•

D80–D84.2, D89, D61, D70.1-D70.3, and E70.3. G11.3. ICD-8: 2880*) •

Registered with a solid cancer within 1 year prior to index date in the Danish Cancer Register (except the following diagnoses: C44.*, N87.*, O01.*, D05.*, D06.*, D06.*, D076.*, D067.*, D069.*, D091.*, D076A.*, D076B.*, D076T.*, D090.*, D30.*, D330B.*, D330C.*, D330D.*, D331.*, D332.*,

EP

D333.*, D334.*, D337.*, D339.*, D329A.*, D32.*, D33.*, D352.*, D352A.*, D353.*, D354.*, D307.*, D309.*, N879.*, N87.*, O01.*, D45.*, and D46.*). Registered with a discharge diagnosis suggestive of having received an organ transplant (ICD-8:

AC C

•

9977* and Y9509*. ICD-10: Z94* and T86*).

2

ACCEPTED MANUSCRIPT

Prior alcohol abuse Defined as fulfilment of one of the following criteria: •

≥2 admissions with a discharge diagnosis suggestive of acute alcohol intoxication in a nonpsychiatric hospital (ICD-8: 303.09, 303.90, 303.99, 980.99, E8609. ICD-10: F10 (0-1) and DT519.) ≥1 admission with a discharge diagnosis suggestive of chronic alcohol abuse in a non-psychiatric

RI PT

•

hospital (ICD-8: 291, 303.19-29, 303.91, 571.09, 571.10, and 577.10. ICD-10: DG312, DK70, DG312, F10 (2-9), DE244, DE529A, DG621, DG721, DI426, DK292, DK852, DK860, DL278A, DO354, DP043, DT500A, DZ714, DZ721 and DZ502.)

Redeemed prescriptions of disulfiram (ATC: N07BB01) from November, 1992 until index date.

•

Registered in the National Register of Alcohol Abuse Treatment (established in 2006).

M AN U

Prior intravenously drug use:

SC

•

Defined as fulfilment of one of the following criteria: •

Registered with a discharge diagnosis from a non-psyciatric hospital prior to index date suggestive of heroin intoxication (ICD-8: 30419. ICD-10: DT401, DZ2358M11H, DZ2358M21H, DZ239311H, DZ39321H, and T401).

Registered in the National Registry of Drug Abusers undergoing Treatment (established in 1996)

Diabetes:

TE D

•

Defined as fulfilment of one of the following criteria:

Redeemed prescriptions of antidiabetic drugs (ATC: A10*) prior to index.

•

Registered with discharge diagnosis suggestive of diabetes before index (ICD-8: 249*, 250*. ICD-10:

EP

•

E10*, E11*, E12*, E13*, E14).

AC C

Liver cirrhosis:

Defined as fulfilment of one of the following criteria: •

Registered with discharge diagnosis suggestive of liver cirrhosis (ICD-8: 57190*, 57191*, 57192*, 57194*, 57199*, 45600-45609. ICD-10: K702*, K703*, K704*, K740*, K74*, K717*, I85*).

3

ACCEPTED MANUSCRIPT

Appendix D.1:

Short-term mortality 0–90 days

91–365 days

Cases

(95% CI)

Events / PY

Total

40 / 68.2

24 / 730.5

17.6 (10.6–29.2)

21 / 183.4

Men

25 /41.7 15 / 26.5

19 / 445.6 5 / 284.9

13.9 (7.7–25.2) 31.6 (11.5–87.0)

2 / 12.2 2 / 12.4

0 / 125.7 1 / 125.7

8 / 19.0 28 / 24.6

Women Age 18–49 years 50–59 years 60–69 years 70+ years

Reference population Events / PY

Unadjusted MRR*

63 / 2,174.2

3.9 (2.4–8.1)

14 / 72.8 7 / 110.6

38 / 1,325.7 25 / 848.6

4.4 (2.4–8.1) 3.3 (1.4–7.5)

20.1 (1.8–222.2)

3 / 36.0 1 / 36.1

0 / 379.2 1 / 378.6

10.4 (0.7–166.2)

4 / 198.8 19 / 280.4

20.8 (6.3–69.1) 16.3 (9.1–29.2)

4 / 53.4 13 / 57.9

6 / 595.4 56 / 821.0

7.4 (2.1–26.1) 3.3 (1.8–6.0)

14 / 38.8 7 / 11.0

11 / 571.9 3 / 75.0

18.5 (8.4–40.8) 16.2 (4.2–62.6)

7 / 109.9 6 / 27.0

32 / 1,711.0 7 / 222.3

3.4 (1.5–7.7) 6.9 (2.3–20.6)

19 / 18.7 7 / 7.2

10 / 83.6 13 / 692.1

8.6 (3.9–18.0) 2.8 (1.1–7.0)

8 / 46.5 7 / 15.2

24 / 240.9 13 / 109.0

1.7 (0.8–3.8) 3.8 (1.5–9.5)

4 / 8.2 0 / 3.9

0 / 19.5 0 / 0.5

-

2 / 21.8 1 / 11.9

3 / 57.0 0 / 1.5

1.7 (0.3–10.3) -

0 / 2.2 3 / 45.8

19.2 (5.4–68.2)

0 / 1.5 4 / 23.0

0 / 6.7 11 / 133.2

2.1 (0.7–6.5)

1 ≥2 Immunosup. Alcohol IDU Liver cirrhosis Diabetes Period of index: 1994-2002

4 / 1.3 12 / 9.5

EP

Charlson index 0

TE D

Comorbidity:

SC

Unadjusted MRR*

Events / PY

Reference population Events / PY

M AN U

Cases

RI PT

Table D.1: Mortality rate ratios during first year after index stratified on demographics and comorbidities.

AC C

14 / 26.3 15 / 429.7 15.1 (7.3–31.3) 14 / 26.3 15 / 429.7 2003-2009 26 / 41.9 9 / 300.8 20.5 (9.6–43.8) 26 / 41.9 9 / 300.8 Abbreviations: PY – person years; CI – confidence interval; IDU – intravenously drug users *MRR - Mortality rate ratio estimated using Cox regression analyses.

1

(95% CI)

11.0 (5.5–21.9) 1.5 (0.7–3.2)

ACCEPTED MANUSCRIPT

RI PT

Appendix D.2: Table D.2: Mortality rate ratios from 1–14 years after index stratified on demographics and comorbidities.

Events / PY 50 / 799.6

Unadjusted MRR* (95% CI) 2.0 (1.5–2.7)

37 / 476.9

197 / 6,480.2

2.6 ( 1.8–3.7)

Cases

Total Men Women

Events / PY 30 / 550.7

5–14 days Reference population Events / PY 471 / 12,438.4

Unadjusted MRR* (95% CI) 1.5 (1.1–2.1)

21 / 333.7

320 / 7,693.0

1.3 (0.7–2.6)

9 / 217

151 / 4,745.4

1.5 (1.0–2.4)

Cases

SC

Long-term mortality 1–4 years Reference population Events / PY 332 / 10,596.6 135 / 4,116.3

1.2 (0.7–2.2)

7 / 162.7

5 / 1,980.0

17.0 (5.4–53.5)

1 / 169.9

11 / 3,104.8

1.7 / 0.2–13.3)

50–59 years 60–69 years

8 / 166.0 10 / 234.8

20 / 1,949.1 69 / 3,003.5

4.7 (2.1–10.7) 1.9 (1.0–3.7)

6 / 125.4 10 / 147.0

38 / 2,765.7 145 / 3,460.8

3.8 (1.6–8.9) 1.9 (1.0–3.6)

70+ years

13 / 236.1

238 / 3,664.0

1.6 (1.1–2.5)

13 /108.4

277 / 3,107.1

1.4 (0.8–2.5)

25 / 502.7

192 / 8,549.9

2.2 (1.5–3.4)

21 / 393.1

337 / 10,864.6

1.8 (1.2–2.8)

M AN U

13 / 322.6

Age 18–49 years

≥2 Immunosuppression Alcohol dependency IDU Liver cirrhosis Diabetes

2003-2009

58 / 1,021.6 82 / 1,025.0

1.4 (0.7–2.8) 1.1 (0.6–1.9)

3 / 62.4 6 / 95.1

63 / 895.9 71 / 677.9

0.7 (0.2–2.1) 0.6 (0.3–1.4)

7 / 54.0 10 / 77.9

33 / 478.9 17 / 265.6

1.9 (0.8–4.3) 2.0 (0.9–4.4)

2 / 15.2 3 / 33.7

31 / 345.9 14 / 179.8

1.2 (0.3–5.3) 1.2 (0.3–4.2)

6 / 45.1 0 / 6.8

0 / 8.0 2 / 30.1

-

1 / 25.6 0 / 3.6

1 / 4.7 1 / 19.7

-

6 / 93.2

44 / 578.9

0.9 (0.4–2.1)

2 / 46.3

36 / 425.3

0.5 (0.1–2.1)

18 / 315.9

186 / 6,502.7

2.0 (1.2–3.2)

24 / 419.8

410 / 10,957.4

1.6 (1.0–2.3)

61 / 1,481.0

1.1 (0.5–2.6)

AC C

Period of index: 1994-2002

9 / 114.5 16 / 182.4

EP

0 1

TE D

Comorbidity: Charlson index

32 / 483.6 146 / 4,093.8 1.9 (1.3–2.7) 6 / 130.8 Abbreviations: PY – person years; CI – confidence interval; IDU – intravenously drug users *MRR - Mortality rate ratio estimated using Cox regression analyses.

2

ACCEPTED MANUSCRIPT

Appendix C: Infectious diseases

A00-B99

•

Malignant neoplasms

C00-C99

•

In situ/ benign neoplasm

D00-D48

•

Blood/immune diseases

D50-D89

•

Endocrine diseases

E00-E90

•

Mental diseases/drug abuse

F00-F99

•

Nervous system diseases

G00-G99

•

Cardiovascular diseases

I00-I99

•

Respiratory diseases

J00-J99

•

Digestive system diseases

•

Musculoskeletal / skin disorders

•

Genitourinary diseases

•

Injury/poisoning

•

Ill-defined causes

M AN U

SC

•

RI PT

Primary causes of death specified according to the ICD-10 codes:

K00-K93

L00-M99

N00-N99

S00-T98,V,W,X,Y

AC C

EP

TE D

R00-Q99 and no cause of death reported.

ACCEPTED MANUSCRIPT STROBE Statement—Checklist of items that should be included in reports of cohort studies

Title and abstract

Item No 1

Recommendation (a) Indicate the study’s design with a commonly used term in the title or the abstract (b) Provide in the abstract an informative and balanced summary of what was done and what was found

2

Explain the scientific background and rationale for the investigation being reported

Objectives

3

State specific objectives, including any prespecified hypotheses

Methods Study design

4

Present key elements of study design early in the paper

Setting

5

Describe the setting, locations, and relevant dates, including periods of recruitment, exposure, follow-up, and data collection

6

(a) Give the eligibility criteria, and the sources and methods of selection of

SC

Participants

RI PT

Introduction Background/rationale

participants. Describe methods of follow-up

(b) For matched studies, give matching criteria and number of exposed and unexposed 7

Clearly define all outcomes, exposures, predictors, potential confounders, and effect modifiers. Give diagnostic criteria, if applicable

Data sources/

8*

For each variable of interest, give sources of data and details of methods of

M AN U

Variables

measurement

assessment (measurement). Describe comparability of assessment methods if there is more than one group 9

Describe any efforts to address potential sources of bias

Study size

10

Explain how the study size was arrived at

Quantitative variables

11

Explain how quantitative variables were handled in the analyses. If applicable, describe which groupings were chosen and why

Statistical methods

12

TE D

Bias

(a) Describe all statistical methods, including those used to control for confounding (b) Describe any methods used to examine subgroups and interactions (c) Explain how missing data were addressed (d) If applicable, explain how loss to follow-up was addressed

13*

AC C

Results Participants

EP

(e) Describe any sensitivity analyses

Descriptive data

14*

(a) Report numbers of individuals at each stage of study—eg numbers potentially eligible, examined for eligibility, confirmed eligible, included in the study, completing follow-up, and analysed (b) Give reasons for non-participation at each stage (c) Consider use of a flow diagram

(a) Give characteristics of study participants (eg demographic, clinical, social) and information on exposures and potential confounders (b) Indicate number of participants with missing data for each variable of interest (c) Summarise follow-up time (eg, average and total amount)

Outcome data

15*

Report numbers of outcome events or summary measures over time

Main results

16

(a) Give unadjusted estimates and, if applicable, confounder-adjusted estimates and their precision (eg, 95% confidence interval). Make clear which confounders were adjusted for and why they were included (b) Report category boundaries when continuous variables were categorized (c) If relevant, consider translating estimates of relative risk into absolute risk for a meaningful time period 1

ACCEPTED MANUSCRIPT Other analyses

17

Report other analyses done—eg analyses of subgroups and interactions, and sensitivity analyses

Discussion Key results

18

Summarise key results with reference to study objectives

Limitations

19

Discuss limitations of the study, taking into account sources of potential bias or

Interpretation

20

Give a cautious overall interpretation of results considering objectives, limitations,

Generalisability

21

Discuss the generalisability (external validity) of the study results

Other information Funding

22

Give the source of funding and the role of the funders for the present study and, if

imprecision. Discuss both direction and magnitude of any potential bias

RI PT

multiplicity of analyses, results from similar studies, and other relevant evidence

applicable, for the original study on which the present article is based

SC

*Give information separately for exposed and unexposed groups.

Note: An Explanation and Elaboration article discusses each checklist item and gives methodological background and published examples of transparent reporting. The STROBE checklist is best used in conjunction with this article (freely

AC C

EP

TE D

available at http://www.strobe-statement.org.

M AN U

available on the Web sites of PLoS Medicine at http://www.plosmedicine.org/, Annals of Internal Medicine at http://www.annals.org/, and Epidemiology at http://www.epidem.com/). Information on the STROBE Initiative is

2