Long-term outcome after ischemic stroke in relation to comorbidity – An observational study from the Swedish Stroke Register (Riksstroke)

Abstract

Purpose

Comorbidity in stroke is common, but comprehensive reports are sparse. We describe prevalence of comorbidity and the prognostic impact on mortality and functional outcome in a large national ischemic stroke cohort.

Methods

We used outcome data from a long-term follow-up survey conducted in 2016 by the Swedish Stroke Register (Riksstroke). Those included in the study were 11 775 pre-stroke functionally independent patients with first-ever ischemic stroke followed up at three months and 12 months (all patients), and three years (2013 cohort) or five years (2011 cohort). Pre-stroke comorbidity data for 16 chronic conditions were obtained from the Swedish National Patient Register, the Swedish Prescribed Drugs Register and the Riksstroke register. Individuals were grouped according to number of conditions: none (0), low (1), moderate (2–3) or high (≥4). Co-occurrence was analysed using hierarchical clustering, and multivariable analyses were used to estimate the prognostic significance of individual conditions.

Results

The proportion of patients without comorbidity was 24.8%; 31.8% had low comorbidity; 33.5% had moderate comorbidity and 9.9% had high comorbidity. At 12 months, the proportion of poor outcome (dead or dependent: mRS ≥3) was 24.8% (no comorbidity), 34.7% (low), 45.2% (moderate) and 59.4% (high). At five years, these proportions were 37.7%, 50.3%, 64.3%, and 81.7%, respectively. There was clustering of cardiovascular conditions and substantial negative effects of dementia, kidney, and heart failure.

Conclusion

Comorbidity is common and has a strong impact on mortality and functional outcome. Our results highlight the need for health systems to shift focus to a comprehensive approach in stroke care that includes multimorbidity as a key component.

Keywords

Ischemic stroke comorbidity cluster prognosis mortality functional outcome

Introduction

Multimorbidity refers to the existence of multiple medical conditions in a single individual while comorbidity is used to describe the co-occurrence of other conditions alongside a primary condition.¹ Multimorbidity is present in the majority of individuals over 65 years^2,3 and is markedly increased in stroke patients⁴ where it is associated with increased mortality^5–8 and poor functional outcome.^6,7,9–11

Multimorbidity has received little attention despite its fundamental clinical and economic impact.¹² Also, scientific reporting is heterogeneous, largely due to methodological differences, making it difficult to correctly gauge and address the problem.² In stroke, comprehensive reports on long-term outcome relative to comorbidity are sparse, particularly on functional status.¹³ We aim to describe both mortality and functional outcome in relation to age and comorbidity burden in pre-stroke functionally independent patients from a large national ischemic stroke cohort. In addition, we aim to provide a detailed description of prevalence and clustering of individual conditions in these patients.

Methods

Stroke and comorbidity data

Data on pre-stroke functional status, index stroke, patient characteristics, and outcome (mortality and functional status) were obtained from the Swedish Stroke Register (Riksstroke). Riksstroke is the Swedish quality register for stroke care with an estimated coverage of >90% of stroke patients admitted to hospital.^14,15 In addition to the customary surveys at three and 12 months, a long-term follow-up postal survey was conducted in 2016 on patients who had experienced stroke three years (in 2013) and five years (in 2011) earlier.¹⁶ Thus, three-year follow-up data was only available for the 2013 cohort and five-year follow-up data was only available for the 2011 cohort. Data on mortality status and date of death were obtained continuously from the Swedish Causes of Death register and included in Riksstroke.

Data on comorbidity were primarily obtained from the Swedish National Patient Register (SNPR), which includes data from outpatient and inpatient healthcare contacts. Patients were considered having a specific condition if diagnosed at any time five years prior to index stroke. We used the Charlson Comorbidity Index (CCI) to guide selection of which conditions to include.¹⁷ The original CCI includes 19 chronic conditions and has been used for predicting mortality and functional outcome in stroke.^6,7 In the present study, we did not include ulcer disease, hemiplegia, cerebrovascular disease nor AIDS, and liver disease and diabetes were not subdivided. We also added angina pectoris, atrial fibrillation, and hypertension. This added up to a total of 16 conditions (Supplementary Table I). Riksstroke data were used for atrial fibrillation, diabetes, and hypertension. Also, we obtained data from the Swedish Prescribed Drugs Register (SPDR) to identify additional cases of dementia. The SPDR covers all prescriptions made and filled by any Swedish pharmacy. Patients who had filled their prescriptions for prescription-only drugs specific to dementia (ATC N06D) at any time from 2005 to index stroke were considered having dementia.

We obtained data for highest level of education from Statistics Sweden.

The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statements¹⁸ and the local ethics approval committee (Regional Ethical Review Board, Lund) approved the project in 2017 (Dnr 2017/529). The committee waived the need for patient consent. Requests to access the dataset may be sent to Riksstroke after obtaining the appropriate ethics approval.

Study population

Approximately half of patients registered in Riksstroke in 2011 and 2013 were selected randomly to be included in the study (Figure 1). Since there were no major differences in baseline characteristics, as was demonstrated by us in a previous publication, the two year- cohorts were merged.¹⁶ In all, 11 775 pre-stroke functionally independent patients ≥18 years of age with first-ever ischemic stroke (ICD-10 I63) were included. At subsequent time points, individuals were classified as lost to follow-up upon failure to return the questionnaire or if there was incomplete information in any of the variables needed for estimating a modified Rankin Scale (mRS) score.

Figure 1.

Inclusion flowchart. We included pre-stroke functionally independent patients ≥18 years of age with first-ever ischemic stroke (ICD-10 I63). LTF=loss to follow-up.

Measures and definitions

Comorbidity burden was defined as the sum of individual conditions: none (0), low (1), moderate (2–3), and high (≥4).

Functional status upon follow-up was described using the modified Rankin Scale (mRS).¹⁹ Scores were estimated from information on dependency in specified ADL domains (toileting, dressing, mobility), living conditions, and need of support from next of kin, using a validated and previously specified translation algorithm.²⁰ Independency was defined as mRS ≤ 2. Level of consciousness at admission was used as a proxy for stroke severity²¹ and was registered using the Reactions Level Scale 85 (RLS) with categories of alert (RLS 1), drowsy (RLS 2–3), and comatose (RLS 4–8).

Statistical methods

Bivariate analyses were performed using χ² test for categorical variables and Student’s t-test for normally distributed quantitative variables.

Since data on mortality status were almost complete, patients lost to follow-up were alive but of unknown functional status. In the description of proportions, the distribution of mRS scores in followed-up survivors was extrapolated to all survivors. We did not use imputation methods since we have previously shown in the same cohort that this did not substantially change outcome estimates.¹⁶ As for missing baseline data, we omitted missing data points and presented only valid proportions. In regression models, cases lost to follow-up or with missing data were excluded.

The prognostic impact of comorbidity burden and seven individual conditions were analysed separately through logistic regression models with adjustment for age, sex, level of consciousness at admission and educational level (Table 2). Additionally, a model including all seven conditions was analysed. The effect of age was modeled via restricted cubic splines. For all models, residual deviance per degree of freedom was below 1 indicating a good overall fit and adequacy of the assumption of binomial outcome (no obvious over-dispersion). We checked for separation of data and performed collinearity diagnostics of the predictors. The functional form of the impact of age as well as the appropriateness of the logistic link function was assessed through tests based on cumulative residuals. Finally, we scanned data for highly influential observations by investigating suitable case deletion statistics. Due to its disproportional impact on outcome, patients with known metastatic malignancy (n = 111, 0.9%) were not included in the analysis.

Hierarchical cluster analysis was used to analyse co-occurrence of conditions.²² Pairwise correlation was estimated using Yule’s Q formula (Supplementary Tables II and III) and the “distance” between conditions was defined as 1-Yule’s Q. Subsequently, clusters were identified by ocular inspection of dendrograms illustrating outcomes of the cluster analysis. Due to a prevalence of <1%, liver disease was not included in the analysis.

Two-dimensional multidimensional scaling (MDS) was used to further illustrate co-occurrence of conditions. Each condition was assigned coordinates along two axes in such a way as to make distances in the abstract two-dimensional space as close to corresponding values of 1-Yule’s Q as possible. Note that a priori the dimensions lack clinical meaning. All statistical analyses were conducted using SAS 9.4. A p-value of <0.05 was considered statistically significant.

Findings

Patient characteristics and prevalence of comorbidity

In all, 11 775 patients were included (Figure 1). Patient loss to follow-up was 12.5% at three months, 20.1% at 12 months, 22.4% at three years, and 21.0% at five years. At baseline, there was missing data in <0.5% except for smoking (6.4%) and level of education (1.4%).

Smoking, educational level, and the proportion of males were lower in older patients who also more often presented with severe stroke (Table 1). Six of the 16 chronic conditions had a prevalence higher than 5%. These were, in descending order: hypertension, atrial fibrillation, diabetes, non-metastatic solid tumor, angina pectoris, and congestive heart failure. Prevalence of most conditions increased with age, except for metastatic malignancy, chronic liver disease, chronic obstructive pulmonary disease (COPD), rheumatoid arthritis (RA), and diabetes.

Table 1.

Patient characteristics and prevalence of comorbidity, stratified by age.

		Age group
	Totaln=11 775	≤64 n=2544	65–74n=3412	75–84n=3833	≥85n=1986
Patient characteristics
Sex (male)	56.5%	65.4%	61.7%	53.2%	42.6%
Smoking	16.6%	30.9%	22.0%	9.0%	3.2%
Highest education
Primary	44.6%	24.4%	39.5%	53.1%	62.8%
Secondary	38.7%	52.1%	41.5%	33.5%	26.7%
≥Tertiary	16.7%	23.5%	19.0%	13.5%	10.6%
Index stroke
Level of consciousness at admission
Alert	90.4%	93.6%	92.5%	90.2%	83.0%
Drowsy	7.4%	4.8%	5.8%	7.8%	12.7%
Comatose	2.3%	1.6%	1.8%	2.0%	4.4%
First admitted to stroke unit^a	84.2%	86.7%	85.2%	83.5%	80.6%
Thrombolytic therapy	10.3%	13.7%	12.2%	9.7%	4.4%
Prevalence of individual conditions
Solid tumor, non-metastatic	10.9%	3.9%	10.4%	14.4%	14.1%
Solid tumor, metastatic	0.9%	0.6%	1.3%	0.9%	0.7%
Leukemia/myeloma	0.5%	0.5%	0.4%	0.5%	0.5%
Lymphoma	0.5%	0.4%	0.5%	0.5%	0.6%
Chronic liver disease	0.7%	1.3%	0.8%	0.4%	0.2%
Chronic kidney failure	2.2%	1.3%	1.8%	2.8%	2.6%
COPD	3.4%	1.8%	4.4%	4.0%	2.7%
Rheumatoid arthritis	1.9%	1.3%	2.2%	2.1%	1.5%
Peripheral vascular disease	1.3%	0.9%	1.5%	1.6%	1.1%
Congestive heart failure	6.9%	3.1%	5.1%	8.4%	12.1%
Myocardial infarction	4.4%	2.2%	4.2%	5.0%	6.1%
Diabetes	19.8%	18.5%	22.8%	20.3%	15.4%
Dementia	1.3%	0.2%	0.7%	1.9%	2.4%
Atrial fibrillation	24.4%	9.5%	19.9%	29.5%	41.4%
Angina pectoris	7.5%	3.7%	6.6%	9.6%	10.0%
Hypertension	58.4%	40.5%	58.5%	65.1%	68.3%
Comorbidity burden
None	24.8%	44.9%	25.0%	17.5%	12.7%
Low	31.8%	30.4%	33.9%	31.5%	30.3%
Moderate	33.5%	21.1%	32.6%	38.5%	41.6%
High	9.9%	3.6%	8.5%	12.5%	15.4%

Note: Comorbidity burden was estimated based on number of conditions: none (0), low (1), moderate (2–3), and high (≥4), not including patients with metastatic malignancy (n = 111). Differences between age groups were statistically significant at the p < 0.05 level for all variables except peripheral vascular disease (p = 0.099), leukemia/myeloma (p = 0.68), and lymphoma (p = 0.82).

^aIncluding observational unit and intensive care unit.

Table 2.

Prognostic impact of comorbidity on poor outcome.

	3 months, n = 10 012		12 months, n = 9137		3/5 years, n = 8939
	Simple adjusted model	Extended model	Simple adjusted model	Extended model	Simple adjusted model	Extended model
Comorbidity burden
None (Reference)	–	–	–	–	–	–
Low	1.07 (0.94–1.22)	–	1.19 (1.04–1.37)	–	1.18 (1.03–1.35)	–
Moderate	1.36 (1.20–1.55)	–	1.65 (1.44–1.90)	–	1.86 (1.62–2.13)	–
High	2.18 (1.83–2.61)	–	2.91 (2.40–3.51)	–	4.63 (3.74–5.73)	–
Individual conditions
Heart failure	1.93 (1.61–2.31)	1.70 (1.42–2.05)	2.10 (1.73–2.54)	1.70 (1.40–2.08)	3.32 (2.62–4.2)	2.53 (1.99–3.22)
COPD	1.44 (1.13–1.84)	1.26 (0.98–1.62)	2.01 (1.56–2.60)	1.68 (1.29–2.18)	2.75 (2.06–3.66)	2.18 (1.62–2.93)
Kidneyfailure	2.31 (1.70–3.15)	1.83 (1.34–2.51)	3.07 (2.20–4.30)	2.25 (1.60–3.17)	5.54 (3.56–8.61)	3.54 (2.25–5.55)
Hypertension	1.15 (1.05–1.27)	1.04 (0.94–1.15)	1.24 (1.12–1.37)	1.11 (1.00–1.23)	1.33 (1.2–1.47)	1.13 (1.02–1.26)
Dementia	2.80 (1.83–4.28)	2.79 (1.83–4.28)	3.51 (2.18–5.66)	3.42 (2.11–5.54)	6.85 (3.19–14.73)	6.64 (3.08–14.33)
Malignancy	1.06 (0.93–1.22)	1.03 (0.90–1.18)	1.25 (1.08–1.44)	1.21 (1.05–1.40)	1.41 (1.21–1.64)	1.34 (1.14–1.56)
Diabetes	1.53 (1.36–1.71)	1.45 (1.29–1.64)	1.63 (1.44–1.83)	1.5 (1.33–1.70)	1.94 (1.71–2.21)	1.71 (1.52–1.98)

Note: Poor outcome was defined as dead or dependent (mRS ≥3). Total comorbidity burden as well as individual conditions were analysed using logistic regression (odds ratio, 95% CI) adjusted for age, sex, level of consciousness at admission, and educational level. Additionally, an extended model including all seven conditions was analysed. Patients lost to follow-up and those with known metastatic malignancy were excluded.

As for total comorbidity burden, 24.8% had no comorbidity, 31.8% low comorbidity, 33.5% moderate comorbidity, and 9.9% had high comorbidity. Differences between sexes were minor.

Mortality and functional outcome

A higher comorbidity burden increased the risk of poor outcome (dead or dependent) considerably (Figure 2). At 12 months, the proportion of poor outcome was 24.8% (no comorbidity), 34.7% (low), 45.2% (moderate) and 59.4% (high). At five years, these proportions were 37.7%, 50.3%, 64.3%, and 81.7%, respectively. This was largely driven by mortality, which increased from 7.3% in those without comorbidity to 27.3% in those with high comorbidity at 12 months and from 19.4% to 64.6% at five years.

Figure 2.

(a–d) Proportions of mRS-scores at different time points relative to comorbidity burden. Excluding patients with metastatic malignancy. For five-year analyses, only the 2011 cohort was included and for three-year analyses, only the 2013 cohort was included. No comorbidity = 0 conditions, low comorbidity = 1 condition, moderate comorbidity = 2–3 conditions, high comorbidity ≥ 4 conditions. Only dead and independents were labeled.

When stratifying by age, the results showed that both comorbidity burden and age were key predictors of poor outcome, the effect of increasing comorbidity being proportionally larger in younger patients (Figure 3(a) to (d)). Mortality at five years in individuals ≤65 years of age was 38.5% in high comorbidity compared to 4.4% in low comorbidity. For poor functional outcome in survivors, the corresponding proportions were 34.8% versus 10.6%.

When also stratifying by sex, the analysis revealed large differences in proportion of poor outcome (dead or dependent) at 12 months between men and women (Figure 3(e) to (f)). Women had a significantly worse prognosis in all age and comorbidity strata.

Figure 3.

(a–f) Long-term prognosis relative to comorbidity burden and age. Mortality in all patients (a, c) and proportion of functional dependency (mRS ≥3) in survivors (b, d) stratified by age and comorbidity burden at 12 months and 5 years. Panels (e) and (f) show proportion of poor outcome (dead or dependent) stratified by age, comorbidity burden, and sex at 12 months. For five-year analyses, only the 2011 cohort was included. Patients with metastatic malignancy were excluded in all analyses. No comorbidity = 0 conditions, low comorbidity = 1 condition, moderate comorbidity = 2–3 conditions, high comorbidity ≥4 conditions.

Comorbidity clusters

The cohort was stratified by median age (74) and the two age groups were analysed separately. We identified several distinct clusters that were similar in both groups. There was a large cluster of cardiovascular conditions (angina, chronic kidney failure, heart failure, myocardial infarction, and peripheral vascular disease) that contained several sub-clusters (Figure 4). In older patients, this large cluster also included COPD. Two other vascular conditions, hypertension and diabetes, formed a separate vascular/metabolic cluster. Both malignancy and atrial fibrillation occurred separately and did not cluster with any other condition. The clustering pattern of dementia was inconsistent between age groups, possibly as a result of low prevalence. It was associated with RA in older patients and COPD in younger patients.

Figure 4.

(a–d) Clustering of conditions. In patients ≤74, n = 5956 (a and c); and ≥75, n = 5819 (b and d). (a and b) show dendrograms where a high degree of co-occurrence is illustrated as late divergence and the dashed lines indicate cut-off points for separating clusters. (c and d) use MDS where each condition was assigned to a specific location in a two-dimensional space and distinct clusters were labeled in different colors. A short distance represents a high degree of co-occurrence and circle size represents relative prevalence. AF: atrial fibrillation, COPD: chronic obstructive pulmonary disease, CKF: chronic kidney failure, MI: myocardial infarction, PVD: peripheral vascular disease, RA: rheumatoid arthritis.

Determinants of poor outcome

Odds ratio (OR) for poor outcome (dead or dependent) increased with comorbidity burden and progressively with longer follow-up time; OR 2.18 (1.83–2.61) at three months and OR 4.63 (3.74–5.73) at three/five years in those with high comorbidity (Table 2). We selected seven conditions to be analysed individually because of their particular clinical importance and relatively limited correlation to other included conditions to avoid collinearity issues. At the 3-month follow-up, ad-hoc contrast tests revealed that dementia, kidney and heart failure were the strongest predictors of poor outcome; diabetes showed an intermediate level of association, while hypertension, COPD and malignancy showed no association. The three/five-year analysis showed a more complicated pattern but dementia and kidney failure still had the strongest association to poor outcome.

Discussion

This study revealed a high comorbidity burden in a large national cohort of Swedish first-ever ischemic stroke patients. A high comorbidity burden more than doubled the proportion of poor outcome at 12 months and five years after stroke, confirming the negative long-term prognostic influence of comorbidity on both mortality and functional outcome.

Our findings in relation to previous research

In a large (n= 1 424 378) Scottish study from 2014, Gallacher et al. reported comorbidity in 94.2% of stroke patients compared to 48% in a control population.⁴ In comparison, we found a lower proportion of 75.2%, which might partly be explained by the fact that we excluded pre-stroke dependent patients and those with previous stroke. However, this discrepancy is probably also an example of the heterogeneous methodological practices in multimorbidity research. For example, Gallacher et al. used a relatively extensive list of 39 conditions compared to our 16.

It is well established that high comorbidity is associated with poor outcome,^8,23–25 which has also been shown specifically in the setting of stroke, although the literature is limited. There are reports of increased mortality both in the short and long term after stroke.^6–8 In a large (n = 201 691), community-based study, Corraini et al. reported an increase in five-year mortality from 39.7% in stroke patients without comorbidity to 79.5% in those with high comorbidity (CCI ≥ 4).⁵ We showed a similar increase, from 19.4% to 64.6% over five years, the lower rate possibly due to differences in inclusion criteria and study methods.

Functional outcome in stroke in relation to comorbidity is not well described in the literature. Most studies are of limited size and many only include patients enrolled in rehabilitation programs. There are several reports describing substantial short- to mid-term effects of a high comorbidity (CCI ≥ 2) in stroke patients,^6,7,9–11 e.g. 36% greater odds of poor outcome (mRS ≥2) at discharge⁷ and 37.3% greater odds at 6 months.⁶ However, our study is the first to describe this long term in a large national cohort.

Among individual conditions, dementia, kidney and heart failure were the strongest predictors of poor outcome in the present study. Others have previously reported similar results.^5,8,9,23

There is a known tendency of conditions which share a common etiology, such as combinations of cardiovascular and metabolic conditions to co-occur,^2,26 which we confirmed in this study. We have not been able to find any previous large study describing comorbidity clustering in stroke.

Recommendations and policy

Stratification and prognostication in stroke has previously mainly relied on clinical variables from the acute setting and a few other known risk factors. However, important intrinsic factors such as level of comorbidity have largely been overlooked.²⁷ We show a strong influence of total comorbidity burden and particular patterns of disease clustering, which emphasises the need to develop a comprehensive approach that includes comorbidity as a key component. In addition to improving post-stroke care, better knowledge of the occurrence of stroke in relation to other conditions has the potential to guide the development of better prevention strategies.

Our results tie in to a wider discussion of healthcare organisation and optimal use of resources. In multimorbidity there is increasing complexity of care; multiple services/clinicians involved and separate management plans increase the risk of fragmentation, dilution of responsibility, polypharmacy, and increasing costs.^12,28 A gravitation towards a patient-centred approach that takes account of all medical as well as non-medical factors to tailor care to the realities of the patient has been suggested.²⁹ The need for personalised medicine is highlighted by our findings that prognosis (and probably need of health care and support) differs substantially in relation to age, comorbidity burden, and sex.

However, most healthcare systems are designed around the treatment of single conditions and the coordinated care of multimorbid patients may require structural changes strengthening the role of general medicine. There is currently little evidence to support the integration of multimorbidity in guidelines and universal healthcare. This needs to be prioritised.

Limitations

The first limitation is missing data and bias. Riksstroke only includes stroke cases admitted to hospital. Thus, we are missing many cases of minor stroke as well as those solely managed in primary care. The resulting bias is difficult to estimate as these cases may vary greatly in respect to pre-stroke health status and stroke severity. Among included patients, there was loss to follow-up of 12.5–22.4% at the different follow-up time points and the possibility of biased results is a concern. However, we have previously shown in the same cohort that baseline characteristics between responders and non-responders were similar and that the biasing effect was limited.¹⁶ Further, our main source for comorbidity data, the SNPR, does not include data from primary care. This means that conditions not requiring specialist or hospital care may have been under-reported. For this reason, we used Riksstroke data on atrial fibrillation, diabetes and hypertension. However, the use of separate sources for different conditions might have resulted in skewed prevalence data. Also, SPDR data was used in addition to the SNPR to identify additional cases of dementia. Despite this we found a relatively low prevalence. However, this was partly a result of the exclusion of pre-stroke dependent patients, as prevalence was significantly higher in those excluded.

The second limitation is mental conditions. These have been shown to be an important component in comorbidity, interacting with somatic conditions.³ However, with the exception of dementia, we included no mental conditions. Depression is typically registered in administrative data as a depressive episode. This made it problematic to include this as a chronic condition. Prevalence of schizophrenia was very low in our cohort, precluding meaningful analysis.

Third, heterogeneity of reporting practices. There is currently no clear consensus on reporting practices in multimorbidity which impairs comparability between studies. Also, index weights are dictated by the outcome variable of interest and there is no single gold standard index for studying both mortality and functional outcome.^6,7,9,10 We classified comorbidity burden based on number of conditions rather than a weighted index in order to make our results as generalisable as possible. However, this approach does not account for the varying severity of different conditions. Further research is needed to develop standardised reporting practices and appropriate indices.

Fourth, due to incomplete registration in Riksstroke, data on NIHSS were not included in analyses despite its known association to a poor prognosis. Level of consciousness is strongly associated to stroke mortality²¹ and was used as a marker for stroke severity.

Fifth, we did not conduct in-depth analyses of differences in men and women as this was beyond the scope of the current paper. However, we did include an analysis of differences in prognosis which revealed a substantial disparity. Further research exploring the impact of comorbidity in men and women is warranted.

Last is generalisability. As the aim was to study the impact of comorbidity in first-ever stroke, inclusion was relatively restrictive. Pre-stroke functionally dependents, ICH and those with previous stroke were excluded. Further research is needed to explore the impact of comorbidity in these groups.

Conclusion

Comorbidity is common in stroke and has strong prognostic implications both for mortality and function in all ages. Three out of four first-ever stroke patients had at least one comorbidity, and a high comorbidity burden more than doubled the proportion of poor outcome at 12 months and five years after stroke. Our results emphasise the need for developing a comprehensive approach to stroke care that includes comorbidity as a key component. Also, there was a tendency of certain conditions to co-occur in specific patterns, which offers an opportunity to pool resources and increase health care efficiency.

Supplemental Material

ESO883154 Supplemental Material - Supplemental material for Long-term outcome after ischemic stroke in relation to comorbidity – An observational study from the Swedish Stroke Register (Riksstroke)

Supplemental material, ESO883154 Supplemental Material for Long-term outcome after ischemic stroke in relation to comorbidity – An observational study from the Swedish Stroke Register (Riksstroke) by Stefan Sennfält, Mats Pihlsgård, Jesper Petersson, Bo Norrving and Teresa Ullberg in European Stroke Journal

Footnotes

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: BN received honoraria for serving on data monitoring committees from Astra Zeneca (SOCRATES and THALES trials) and Bayer AG (NAVIGATE-ESUS trial).

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was funded by the Swedish Stroke Association, Neuro Sweden, Sparbanken Färs och Frosta and received ALF funding from Region Skåne.

Informed consent

The local ethics committee waived the need for patient/caregiver consent.

Ethical approval

Ethical approval for this study was obtained from the local ethics committee (Regionala Etikprövningsnämnden, Lund) in 2017 (Dnr 2017/529).

Guarantor

SS.

Contributorship

The following are the contributions of the authors: SS: Participated in literature search, study design, data collection, data analysis, and interpretation of results. Wrote the first manuscript draft which was then developed further in collaboration with the other authors. Revised and approved the final version. TU: Participated in the study design, data collection, data interpretation, writing of manuscript in collaboration with the other authors. Revised and approved the final version. JP: Participated in the study design, data interpretation, writing of manuscript in collaboration with the other authors. Revised and approved the final version. BN: Participated in the study design, data interpretation, writing of manuscript in collaboration with the other authors. Revised and approved the final version. MP: Primarily active in statistical methodology and data interpretation, but also in manuscript writing in collaboration with the other authors. Revised and approved the final version.

Acknowledgements

We would like to thank the staff at Riksstroke, in particular statistician Fredrik Jonsson, for preparing the data from Riksstroke used in the study. Also, we thank the external proofreader Lee Nolan for making the text more fluent.

Supplemental Material

Supplemental material for this article is available online.

References

van der Akker

Buntinx

Knottnerus

Comorbidity or multimorbidity: what’s in a name? A systematic review of literature. Eur J General Practice 1996; 2: 65–70.

Violan

Foguet-Boreu

Flores-Mateo

, et al. Prevalence, determinants and patterns of multimorbidity in primary care: a systematic review of observational studies. PLoS One 2014; 9: e102149.

Barnett

Mercer

Norbury

, et al. Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross-sectional study. Lancet 2012; 380: 37–43.

Gallacher

Batty

McLean

, et al. Stroke, multimorbidity and polypharmacy in a nationally representative sample of 1,424,378 patients in Scotland: implications for treatment burden. BMC Med 2014; 12: 151.

Corraini

Szepligeti

Henderson

, et al. Comorbidity and the increased mortality after hospitalization for stroke: a population-based cohort study. J Thromb Haemost 2018; 16: 242–252.

Jimenez Caballero

Lopez Espuela

Portilla Cuenca

, et al. Charlson comorbidity index in ischemic stroke and intracerebral hemorrhage as predictor of mortality and functional outcome after 6 months. J Stroke Cerebrovasc Dis 2013; 22: e214–e218.

Goldstein

Samsa

Matchar

, et al. Charlson Index comorbidity adjustment for ischemic stroke outcome studies. Stroke 2004; 35: 1941–1945.

Gallacher

McQueenie

Nicholl

, et al. Risk factors and mortality associated with multimorbidity in people with stroke or transient ischaemic attack: a study of 8,751 UK Biobank participants. J Comorb 2018; 8: 1–8.

Fischer

Arnold

Nedeltchev

, et al. Impact of comorbidity on ischemic stroke outcome. Acta Neurol Scand 2006; 113: 108–113.

10.

Berlowitz

Hoenig

Cowper

, et al. Impact of comorbidities on stroke rehabilitation outcomes: does the method matter? Arch Phys Med Rehabil 2008; 89: 1903–1906.

11.

Kabboord

van Eijk

Fiocco

, et al. Assessment of comorbidity burden and its association with functional rehabilitation outcome after stroke or hip fracture: a systematic review and meta-analysis. J Am Med Dir Assoc 2016; 17: 1066; e1013–1066 e1021.

12.

Lehnert

Heider

Leicht

, et al. Review: health care utilization and costs of elderly persons with multiple chronic conditions. Med Care Res Rev 2011; 68: 387–420.

13.

Nelson

MLA

McKellar

, et al. Stroke rehabilitation evidence and comorbidity: a systematic scoping review of randomized controlled trials. Top Stroke Rehabil 2017; 24: 374–380.

14.

Riksstroke The riksstroke annual report of 2013. Umeå, Sweden: Västerbottens läns landsting, 2014.

15.

Riksstroke The riksstroke annual report of 2011. Umeå, Sweden: Västerbottens läns landsting, 2012.

16.

Sennfalt

Norrving

Petersson

, et al. Long-term survival and function after stroke. Stroke 2018; STROKEAHA118022913.

17.

Charlson

Pompei

Ales

, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40: 373–383.

18.

von Elm

Altman

Egger

, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet 2007; 370: 1453–1457.

19.

Banks

Marotta

CA.

Outcomes validity and reliability of the modified Rankin scale: implications for stroke clinical trials: a literature review and synthesis. Stroke 2007; 38: 1091–1096.

20.

Eriksson

Appelros

Norrving

, et al. Assessment of functional outcome in a national quality register for acute stroke: can simple self-reported items be transformed into the modified Rankin Scale? Stroke 2007; 38: 1384–1386.

21.

Bray

Campbell

Cloud

, et al. Derivation and external validation of a case mix model for the standardized reporting of 30-day stroke mortality rates. Stroke 2014; 45: 3374–3380.

22.

Roso-Llorach

Violan

Foguet-Boreu

, et al. Comparative analysis of methods for identifying multimorbidity patterns: a study of ‘real-world' data. BMJ Open 2018; 8: e018986.

23.

Hall

Dondo

Yan

, et al. Multimorbidity and survival for patients with acute myocardial infarction in England and Wales: Latent class analysis of a nationwide population-based cohort. PLoS Med 2018; 15: e1002501.

24.

Nunes

Flores

Mielke

, et al. Multimorbidity and mortality in older adults: A systematic review and meta-analysis. Arch Gerontol Geriatr 2016; 67: 130–138.

25.

Ryan

Wallace

O'Hara

, et al. Multimorbidity and functional decline in community-dwelling adults: a systematic review. Health Qual Life Outcomes 2015; 13: 168.

26.

Prados-Torres

Calderon-Larranaga

Hancco-Saavedra

, et al. Multimorbidity patterns: a systematic review. J Clin Epidemiol 2014; 67: 254–266.

27.

Thijs

Prognosis after stroke. In: B

Norrving

(ed.) Oxford textbook of stroke and cerebrovascular disease. 1st ed. Oxford: Oxford University Press, 2014, pp.185–193.

28.

Wallace

Salisbury

Guthrie

, et al. Managing patients with multimorbidity in primary care. BMJ 2015; 350: h176.

29.

Muth

van den Akker

Blom

, et al. The Ariadne principles: how to handle multimorbidity in primary care consultations. BMC Med 2014; 12: 223.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.17 MB