Article Text
PDF
Abstract
Objective Relatively little is known about the effect of malignancy on patient outcomes after acute ischemic stroke (AIS) or utilization rates of stroke interventions in this population. We aimed to assess the effect of underlying malignancy on outcomes and treatment of AIS at a population level.
Methods Outcomes after AIS between patients with and without malignancy were compared using a national database of hospital reported outcomes.
Results There were 351 institutions reporting the outcomes of 3 18 127 admissions for AIS. Of these admissions, 16 141 patients carried a pre-existing diagnosis of malignancy at the time of admission. Administration of intravenous tissue plasminogen activator (IV tPA) was less common in patients with malignancy compared with patients without malignancy (7.3% vs 10.7%; P<0.001) but there was no difference in the rate of mechanical thrombectomy (3.1% vs 3.1%; P=0.967). Mortality rates were higher among patients with malignancy (7.1% vs 3.7%; P<0.001), a relationship which persisted when analysis was restricted to patients receiving IV tPA (10.8% vs 6.1%; P<0.001) or thrombectomy (20.3% vs 13.5%; P<0.001). Rates of both IV tPA administration (2.5% vs 10.5%; P<0.001) and mechanical thrombectomy (2.1% vs 5.4%; P<0.001) were lower in patients with brain malignancy relative to patients with malignancy of non-CNS origin.
Conclusion A diagnosis of malignancy on admission for acute stroke was associated with a higher rate of mortality. Malignancy was also associated with a lower rate of IV tPA administration but no difference in mechanical thrombectomy utilization.
- cancer and stroke
- tissue plasminogen activator
- thrombus aspiration
- stroke subject terms: ischemic stroke
- quality and outcomes
- revascularization
Statistics from Altmetric.com
- cancer and stroke
- tissue plasminogen activator
- thrombus aspiration
- stroke subject terms: ischemic stroke
- quality and outcomes
- revascularization
Introduction
A diagnosis of cancer has been associated with an increased risk of ischemic cerebrovascular disease,1–3 and may be associated with worse outcomes relative to patients without underlying malignancy.4 Despite this association, there is a relative paucity of information on the association between malignancy and immediate outcomes after acute ischemic stroke (AIS). In particular, data on patient outcomes after interventions, such as intravenous tissue plasminogen activator (IV tPA) administration and mechanical thrombectomy, as well as the utilization of these interventions, is limited and consists mostly of small case series.5–8 As AISs are not uncommon in this patient population, determining whether a cancer diagnosis significantly impacts stroke outcomes could provide valuable prognostic information. In addition, identifying utilization rates and complications associated with stroke interventions in this population may identify areas for improvement in stroke care. Herein, we reviewed a national database of hospital reported outcomes to determine the effect of underlying malignancy on immediate stroke outcomes and utilization rates of IV tPA and mechanical thrombectomy in this patient population.
Materials and methods
Patient identification
The Vizient clinical database/resource manager (CDB/RM) is a database of hospital reported outcomes that receives monthly updates from affiliated institutions across the country. Participating institutions provide data from all admissions during the monthly period to the CDB/RM, which then tracks various metrics (eg, length of stay) and incidence of inhospital events (eg, mortality) for each institution. Admissions of interest were identified using International Classification of Disease, ninth (ICD-9) and 10th (ICD-10) edition codes. Institutional review board approval was not required for this study. Patients admitted for the treatment of AIS from October 2014 to March 2018 were identified using variations of the ICD-9 and ICD-10 diagnosis codes 433–434, 436–437, and I63. Patients with a diagnosis of malignancy present at the time of admission were identified using variations of the ICD-9 and ICD-10 diagnosis codes 140–165, 170–176, and 179–209, and C00-26, C30-41, C43-C58, C60-C96, C7A, C7B, and D00-D09. To provide a measure of the distribution of malignancy site, malignancies (and respective ICD-9 and ICD-10 codes) were grouped under the following categories: brain (ICD-9: 191; ICD-10: C71), eye (ICD-9: 190; ICD-10: C69), oral cavity and glands (ICD-9: 140–145; ICD-10: C00-09, D00), pharynx and larynx (ICD-9: 146–149, 160, 161; ICD-10: C10-14, C30-32), respiratory system (ICD-9: 162–163, 165; ICD-10: C33-34, C39, D02), mediastinum (ICD-9: 164; ICD-10: C37-38), digestive tract organs and peritoneum (ICD-9: 150–154, 158–159, 197; ICD-10: C15-21, C26, C48, C78, D01), liver, biliary system, and pancreas (ICD-9: 155–157; ICD-10: C22-25), breast (ICD-9: 174–175; ICD-10: C50, D05), endocrine (ICD-9: 193, 194, 209; ICD-10: C73-75, C7A, C7B), urinary tract (ICD-9: 188–189; ICD-10: C64-68), genital and reproductive organs (ICD-9: 179–187; ICD-10: C51-58, C60-63, D06-07), bone, articular cartilage, connective tissue, and peripheral nerves (ICD-9: 170–171, 200; ICD-10: C31, C40-41, C47, C49, C70, C72), hematologic (ICD-9: 201–208, ICD-10: C81-86, C88-96), skin (ICD-9: 172–173, 176; ICD-10: C43-46, D03-04), and undefined (ICD-9: 195, 198, 199; ICD-10: C76, C79, C80, D09). Patients receiving IV tPA either immediately prior or during their admission were identified using ICD-9 and ICD-10 diagnosis codes V4588 and Z9282 or ICD-9 and ICD-10 procedural codes 9910, 3E04317, and 3E03317, respectively. Patients who received mechanical thrombectomy were identified using the ICD-9 code 3974 and variations of the ICD-10 code 03C.
Patient characteristics and outcomes
Information on patient sex and age were collected. Specifically, the proportion of patients treated at each institution who were women and ≥65 years of age was noted. Outcome of interest included absolute and institutional mortality rates, adjusted mortality as indicated by mortality index, incidence of intracerebral hemorrhage, and discharge to home or self-care. Absolute mortality rates for all patients and specific subgroups (eg, patients receiving IV tPA) were calculated by dividing the total number of deaths by the total number of stroke admissions across all institutions. Institutional mortality rates were determined by dividing the total number of deaths by the number of stroke admissions at a given institution. Institutional rates were used for statistical comparisons between patients with and without malignancy. A similar methodology was used to determine absolute and institutional rates of IV tPA and mechanical thrombectomy utilization, intracerebral hemorrhage, and discharge to home. The mortality index denotes the ratio of observed to expected mortality for a given institution. Expected mortality is determined using validated risk prediction models accounting for patient demographics and comorbidities.9–12Patients that suffered intracerebral hemorrhage during admission were identified using the ICD-9 and ICD-10 diagnosis codes 431and I61-1619.
Statistical analysis
Descriptive statistics for continuous variables were reported as mean (SD). Comparisons of institutional rates of specific outcomes (eg, mortality and discharge to home) and interventions (eg, IV tPA and mechanical thrombectomy) between patients with and without an underlying malignancy were performed using the paired Student’s t test. For these comparisons, only institutions reporting outcomes of both patients with and without malignancy were included for analysis. In other words, data from institutions reporting outcomes of admissions consisting only of patients without malignancy were not included in analyses comparing, for example, mortality rates between patients with and without malignancy. To determine whether demographic differences between patients with and without malignancy influenced the effect of malignancy on patient outcomes, multivariate linear regression analysis was performed. The proportion of patients with an AIS and a concomitant underlying malignancy was determined for each institution and related to institutional mortality and discharge to home rates while adjusting for the proportion of patients who were women and ≥65 years of age. All statistical tests were two sided with an alpha level set at 0.05 for statistical significance. Analyses were performed using commercially available software (JMP 10.0.0, ©2012 SAS Institute Inc).
Results
Institutional outcomes
There were 351 institutions that reported the outcomes of 3 18 127 admissions for an acute stroke. The percentage of patients who were women was 48.3%, while 62.0% of patients were >65 years of age. Among these admissions, there were 50 032 (15.7%) and 17 268 (5.4%) patients who were treated with IV tPA and mechanical thrombectomy, respectively. There were 16 141 stroke admissions for patients with a previously diagnosed malignancy, 430 of whom had an underlying brain malignancy. The distribution of malignancy site is presented in the online supplementary table 1; 46.8% of patients were women and 70.6% of patients were >65 years of age. Among patients with a malignancy, 1662 (10.3%) and 857 (5.3%) were treated with IV tPA and mechanical thrombectomy. The observed mortality rate of the overall cohort was 4.6%, with a mean institutional mortality index of 1.0 (SD 1.6). The percentage of patients discharged to home was 37.3%, with respective percentages of patients treated with IV tPA and mechanical thrombectomy of 34.6% and 18.9%, respectively. A complete summary of patient demographics and treatment outcomes is shown in table 1.
Supplemental material
Characteristics of stroke admissions and general outcomes
Comparison of patients with and without malignancy
Patient outcomes and utilization rates of IV tPA and mechanical thrombectomy were subsequently compared between patients with and without pre-existing malignancy. The combined institutional rate of IV tPA administration was lower for the group of patients with an underlying malignancy compared with patients without malignancy (7.3% vs 10.7%; P<0.001) but there was no difference in the institutional rate of mechanical thrombectomy (3.1% vs 3.1%; P=0.967; table 2). Certified stroke centers (CSCs) had higher rates of IV tPA relative to non-certified centers, but at both categories of institution patients without malignancy had a higher rate of IV TPA administration relative to patient with malignancy (CSC: 15.0% vs 9.7%, P<0.001; non-CSC: 7.6% vs 5.5%, P<0.001). There was no difference in utilization of mechanical thrombectomy for patients with and without malignancy regardless of hospital stroke certification (CSC: 5.3% vs 4.9%, P=0.133; non-CSC: 1.5% vs 1.7%, P=0.377).
Outcomes for patients with and without malignancy
For all patients, both mortality (7.1% vs 3.7%; P<0.001) and mortality index (1.1 vs 0.9; P=0.022) were significantly higher in patients with malignancy. Mortality remained more frequent in patients with malignancy when the analysis was restricted to patients receiving IV tPA (10.8% vs 6.1%; P<0.001) or thrombectomy (20.3% vs 13.5%; P<0.001), although only for patients undergoing thrombectomy was there a difference in mortality index (1.4 vs 1.0; P=0.032). Discharge to home was less frequent for patients with malignancy (28.0% vs 38.8%; P<0.001). Discharge to home rates were also lower when analyzed specifically among patients with malignancy who received IV tPA (27.8% vs 36.9%; P<0.001) or underwent mechanical thrombectomy (15.2% vs 19.0%; P=0.045). There were no differences in the rate of intracerebral hemorrhage after IV tPA (11.6% vs 10.1%; P=0.230) or thrombectomy (16.9% vs 20.2%; P=0.072) between patients with and without malignancy. Outcomes for patients with and without malignancy subdivided by stroke intervention are shown in table 2.
Analysis of patients with brain malignancy
Given the paucity of information regarding outcomes of reperfusion therapies for patients with brain malignancies, an analysis of outcomes specifically for this patient subgroup was performed. There were 430 patients admitted for AIS with a pre-existing underlying brain malignancy, of whom 169 patients were women (39.3%) and 171 patients were ≥65 years of age (39.8%). Absolute mortality rate and mean institutional mortality index were 3.0% and 1.1 (SD 8.9), respectively. The percentage of patients discharged to home was 25.6%. Among these 430 patients, 10 (2.3%) received IV tPA while 10 (2.3%) underwent mechanical thrombectomy. Among patients receiving IV tPA, 2 (20.0%) suffered an intracerebral hemorrhage, there were no mortalities, and 2 patients (20.0%) were ultimately discharged home. Among patients receiving mechanical thrombectomy, there was a single intracerebral hemorrhage (10.0%), there were 4 mortalities (40.0%), and no patients were discharged home (table 3).
Characteristics and outcomes of stroke admissions for patients with brain malignancy
Institutional outcome and intervention rates for patients with a brain malignancy were compared with those for patients with a non-brain malignancy. Mortality rates were lower for patients with a brain malignancy (2.9% vs 8.8%; P<0.001), although there was no difference in mortality index (1.1 vs 1.2; P=0.851). Institutional rates of IV tPA administration (2.5% vs 10.5%; P<0.001) and mechanical thrombectomy (2.1% vs 5.4%; P<0.001) were significantly lower in patients with a brain malignancy compared with patients with a malignancy of non-CNS origin (table 4). When compared with patients without any type of malignancy, rates of IV tPA administration (2.3% vs 16.2%; P<0.001) and mechanical thrombectomy (2.1% vs 5.5%; P<0.001) were lower in patients with brain malignancy.
Comparison of institutional outcome and intervention rates for patients with brain and non-brain malignancies
Discussion
In the present study, we investigated the effect of a previously diagnosed malignancy on immediate outcomes after AIS, as well as the effect of malignancy on the utilization of IV tPA and mechanical thrombectomy. Our results suggest that patients with malignancy may have a worse immediate prognosis after AIS, and that they may be less likely to be treated with IV tPA but equally as likely to be treated with mechanical thrombectomy. These findings are important as they highlight the fact that an active malignancy is a poor prognostic marker among acute ischemic stroke patients, even in the immediate post-stroke hospitalization period.
An explanation for the lower rate of IV tPA administration in patients with malignancy is not immediately clear. Underlying systemic malignancy is not a strict contraindication to IV tPA (except for gastrointestinal cancer with recent major hemorrhage), but the presence of a brain neoplasm is considered a contraindication.13 14 Yet patients with brain neoplasms constituted a relative minority of patients with malignancy admitted for AIS within our dataset (table 1), and thus brain malignancies cannot wholly explain the observed differences in IV tPA administration between patients with and without malignancy. Hematological derangements deemed to increase the bleeding risk may be more common in patients with a malignancy,1 which may have been considered a contraindication to the use of IV tPA in certain cases.13 However, we did not observe differences in the rate of intracerebral hemorrhage after IV tPA between patients with and without malignancy, which is consistent with observations from previous studies.6 7 As such, a perceived increased risk of hemorrhage in patients with underlying malignancy is not currently justified by available evidence, and should not discourage the use of IV tPA in patients with malignancy in the absence of other risk factors. Finally, it is likely that in some instances the overall disease burden of patients with advanced malignancy and short life expectancy may have discouraged providers from administering IV tPA given its attendant risks,15 particularly for certain malignancies known to be associated with bleeding complications (eg, hemoptysis in the setting of lung cancer). On the other hand, we did not observe differences in the rate of mechanical thrombectomy between patients with and without malignancy (table 2). Further work is needed to determine the influence of malignancy on treatment decisions for AIS.
We observed higher rates of inhospital mortality and lower rates of discharge to home in patients with underlying malignancy. Outcomes remained worse when the analysis was restricted to patients receiving IV tPA or mechanical thrombectomy (table 3). The detrimental effect of malignancy did not appear to be secondary to hemorrhagic complications, although information on other types of complications was lacking. Previous studies identified worsening of pre-existing medical comorbidities as a significant driver of mortality in patients with concurrent malignancy,6 and thus overall disease burden was likely a contributing factor to worse outcomes in these patients. Regardless of the mechanism, our results indicate that prognosis after AIS is negatively affected by an underlying malignancy, although this effect is present for all patients and not limited to patients receiving acute stroke reperfusion therapies. More data are needed on the efficacy of both IV tPA and thrombectomy in patients with underlying malignancy and whether therapeutic indications should differ in patients with malignancy compared with the general population.
As stated above, patients with an underlying brain malignancy constituted a minority within the overall cohort of patients with a previously diagnosed malignancy. Overall mortality rates were lower in patients with brain malignancy relative to patients with non-CNS malignancy, although mortality indices were similar, perhaps suggesting greater disease burden, and thus higher expected mortality, in patients with malignancy originating outside of the CNS. Of interest was the small minority of patients with brain malignancy treated with IV tPA despite its contraindication in this population,13 14 although the small sample size precluded a meaningful comparison of outcomes between cancer patients with and without a primary brain malignancy. Also of interest was the lower rate of mechanical thrombectomy utilization in patients with CNS malignancy relative to both patients with malignancy of non-CNS origin and patients without malignancy. There is limited evidence to suggest whether a concomitant brain malignancy increases the risks of mechanical thrombectomy, which is reflected in the lack of official guidelines on this topic.14 Ultimately, patients with brain malignancies that would otherwise qualify for reperfusion therapies on the basis of clinical presentation present challenging clinical dilemmas. Further work is needed to determine whether the risk of stroke interventions is low enough to warrant invasive treatment for certain patients within this population.
Limitations
Our study has limitations inherent in the use of national databases. The accurate identification of patients and incidence of complications, and therefore the validity of our results, is contingent on proper ICD-9 and ICD-10 coding. Information on stroke severity—a major determinant of stroke outcome16 —and time from stroke onset to presentation—a major determinant of eligibility for acute reperfusion therapy—were not available. There were also demographic differences between patients with and without malignancy, both of which were associated with differences in outcome, and thus are results were likely affected to some degree by baseline differences in patient groups (see online supplementary appendix I). Information on cancer stage and the pathologic diagnosis of patient malignancies were also not available. Lack of information on survival information after discharge and data on functional outcomes also limit the interpretation of our results. These limitations are hopefully mitigated to some extent by the large sample size of our study. Our aim was to provide a general sense of the effect of malignancy on outcomes after AIS and utilization of stroke interventions. Further work will be needed to confirm and expand our results.
Conclusion
Through the use of a national database, we report that pre-existing malignancy is associated with worse outcomes and lower rates of IV tPA administration but not with any difference in mechanical thrombectomy utilization in patients admitted for management of AIS. Further work is needed to determine optimal management strategies for patients with malignancy suffering AIS.
References
Footnotes
Contributors LR: conception and design, data collection, statistical analysis, drafting manuscript, critical revision of the manuscript, and reviewed final version. HJC: conception and design, critical revision of the manuscript, and reviewed final version. LRC: conception and design, critical revision of the manuscript, and reviewed final version. AAR: conception and design, critical revision of the manuscript, and reviewed final version. WB: conception and design, critical revision of the manuscript, and reviewed final version.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial, or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Patient consent for publication Not required.