Journal of Infection (2015) xx, 1e12
www.elsevierhealth.com/journals/jinf
Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study N. Mejer a,*, N. Gotland a,1, M.L. Uhre a,1, H. Westh b,c,2, H.C. Schønheyder d,e,3, A. Petersen f,4, A.G. Jensen f,g,5, A.R. Larsen f,4, R. Skov f,4, T. Benfield a,b,h,1, for the Danish Staphylococcal Bacteremia Study Group a
Department of Infectious Diseases, Hvidovre Hospital, University of Copenhagen, Hvidovre, Denmark Department of Clinical Medicine, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark c Department of Clinical Microbiology, Hvidovre Hospital, University of Copenhagen, Hvidovre, Denmark d Department of Clinical Microbiology, Aalborg University Hospital, Aalborg, Denmark e Department of Clinical Medicine, Aalborg University, Aalborg, Denmark f Staphylococcus Laboratory, Statens Serum Institut, Copenhagen, Denmark g Bristol-Myers Squibb, Virum, Denmark h Clinical Research Centre, Hvidovre Hospital, University of Copenhagen, Hvidovre, Denmark b
Accepted 18 March 2015 Available online - - -
KEYWORDS Staphylococcus aureus; Bacteremia; Acute myocardial infarction; Stroke; Arterial thromboembolism
Summary Objectives: An association between infection and arterial thromboembolic events (ATE) has been suggested. Here we examined the risk of myocardial infarction (MI), stroke and other ATE after Staphylococcus aureus bacteremia (SAB). Methods: Danish register-based nation-wide observational cohort study between 1995 and 2008 with matched control subjects from the general population. Results: Within a year, 278 of 15,669 SAB patients and 2570 of 156,690 controls developed MI, stroke or another ATE. The incidence rates among SAB patients were highest within the first 30 days and decreased over a year. The adjusted relative risk of MI, stroke and other ATE during
* Corresponding author. Ketteg ard Alle ´ 30, 2650 Hvidovre, afd 144, Denmark. Tel.: þ45 29845959. E-mail address: [email protected] (N. Mejer). 1 Ketteg ard Alle ´ 30, 2650 Hvidovre, afd 144, Denmark. 2 Ketteg ard Alle ´ 30, 2650 Hvidovre, afd 445, Denmark. 3 Hobrovej 18-22, 9000 Aalborg, Denmark. 4 Artillerivej 5, Copenhagen, 2300, Denmark. 5 Bristol-Myers Squibb, Hummeltoftevej 49, 2830 Virum, Denmark. http://dx.doi.org/10.1016/j.jinf.2015.03.010 0163-4453/ª 2015 The British Infection Association. Published by Elsevier Ltd. All rights reserved. Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
2
N. Mejer et al. the first 30 days after SAB in patients compared to controls were 2.2 (95% CI: 1.6e3.1), 5.5 (95% CI: 3.8e8.3) and 15.5 (95% CI: 6.9e35), respectively. Compared to controls, the increased adjusted relative risk persisted for 30 days for MI, 180 days for stroke and one year for other ATE. Increasing age, hypertension, atrial flutter/fibrillation, prior ATE and endocarditis in SAB patients were associated with an increased risk of ATE. Conclusions: SAB was associated with a short-term increased risk of ATE that persisted longer dependent on type of event. Studies are warranted to investigate treatment strategies to diminish ATE after SAB. ª 2015 The British Infection Association. Published by Elsevier Ltd. All rights reserved.
Introduction Arterial thromboembolic events (ATE) such as myocardial infarction (MI) and stroke are frequent clinical events with substantial morbidity and mortality.1,2 Inflammation induced by acute or chronic infection is believed to account for a significant proportion of arterial thromboembolism, both as a triggering and participating condition. Recent epidemiological studies have linked infections to acute coronary events3e8 and ischemic stroke.3,9e12 The majority of the studies investigated the risk of ATE after pneumonia, urinary tract infection, or influenza, while few addressed the risk of ATE following bacteremia. A recent Danish study found high rates of MI and stroke after community-acquired bacteremia caused by a variety of etiologies, including SAB and other agents.3 Bacteremia is frequent, at an incidence of 1e2 episodes per 1000 population years and Staphylococcus aureus is one of the leading causes.13e15 Here we sought to estimate the risk of MI, stroke and other ATE within the first year after community- and hospital-acquired SAB and the relative risks compared with population-based matched controls. Further, we identified risk factors that may help to distinguish highrisk groups that could potentially benefit from prevention and treatment strategies.
Materials and methods Design and ethics The study was a register-based matched cohort study, using the unique civil registration number assigned to each Danish resident through the Civil Registration System (CRS).16 The CRS tracks daily changes in vital status, including date of emigration and date of death, for the entire Danish population. The CRS is almost 100% complete so for practical purposes no persons are lost to follow up. The CRS number allowed merging of data from three national databases. The study was approved by the Science Ethics Committee for Copenhagen and Frederiksberg Counties (01-369 ⁄ 93) and the Danish Data Protection Agency (2009-41-4179). Informed consent is not required by Danish legislation for register-based studies.
Study population A continuous registration of SAB in Denmark has been carried out by the Staphylococcus Laboratory at Statens Serum Institut (SSI), Copenhagen since 1957.14,17 The present study defined a case as an individual 16 years or older with a first time SAB with
a registered blood culture date identified via the registry between January 1, 1995 and December 31, 2008. We defined hospital-associated (HA) origin of SAB as an infection acquired after 2 days of hospitalization, an infection in a patient in ambulatory care, or a case residing in a nursing home. Community-acquired (CA) SAB was defined as an infection present or incubating at the time of hospitalization in patients not meeting the definition of HA SAB. If data were inadequate, acquisition of SAB was classified as unknown. For each case, we randomly selected 10 control subjects through the CRS matched by age (year of birth) and sex. We used the risk set sampling technique (i.e. eligible control subjects had to be alive and at risk of a first hospitalization with SAB on the date the corresponding case was admitted18).
Arterial thromboembolism Data on myocardial infarction, stroke and other ATE was collected from the National Patient Registry (NPR).19 Other ATE included occlusion of arteries in the retina, kidney, and intestine (See Supplementary Table 1). The NPR is updated monthly and holds information on all hospital admissions and discharge diagnoses of all patients treated in a Danish non-psychiatric hospital since January 1, 1977, as well as all outpatient contacts since January 1, 1995. Information obtained included date of admission, date of discharge, and diagnoses. Both primary and secondary diagnosis codes were used to identify comorbidity prior to SAB, while only primary codes were used to establish an ATE event. The International Classification of Diseases (ICD) codes were used to determine diagnoses of ATE (acute myocardial infarction, stroke or another ATE); the 8th ICD revision was used from 1977 to 1993 and the 10th revision thereafter. SAB patients and controls were followed until December 31, 2009. In Denmark a short summary on each admission is provided by the discharging doctor. For SAB patients, these reports were reviewed if a primary diagnosis of ATE occurred within 365 days of the blood culture date. To avoid reverse causality (i.e. that ATE precipitated SAB), only cases of ATE occurring on either the same day of SAB or later were included in this study. If ATE occurred during the same admission as SAB, the number of days from SAB to ATE was determined. Otherwise the date of admission was registered as the date of onset of ATE.
Risk factors Information on congestive heart failure, atrial flutter/ fibrillation, hypertension, renal disease, diabetes, cocaine
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
Arterial thromboembolism in S. aureus bacteremia use, human immunodeficiency virus, and obesity were obtained from the NPR (Supplementary Table 1). For ATE and atrial flutter/fibrillation any event prior to SAB was considered, whereas obesity, congestive heart failure, renal disease, hypertension, human immunodeficiency virus, and diabetes mellitus were considered if they occurred within 10 years preceding the SAB index date. Information on endocarditis was extracted from discharge reports for each admission with SAB as described previously.20
Statistics Age was divided into 4 groups (16e35, 36e54, 55e75, >75 years). The variable “Other risk factors” was computed by combining the variables obesity, cocaine use and human immunodeficiency virus. All patients were followed from the date of SAB; data was censored on the date of death or emigration or on December 31, 2009, whichever came first. Thus, the competing risk of death was accounted for in the analysis by means of censoring on the date of death. Categorical variables were compared by Chi-square tests. Incidence rates (IR) were calculated by dividing the observed incidences by the accumulated person-years (PYs) at risk in the respective groups, and were presented with 95% confidence intervals (CIs). The CIs were only calculated when the number of events were at least 7. Unadjusted and adjusted Cox regression analysis with 95% CIs was performed to determine the relative risk (RR) of ATE within the first year of the SAB index date. If more than one myocardial infarction, stroke or other arterial thromboembolic events was registered, only the first observed episode was used for analysis. For the survival analysis, controls were censored along with their matched SAB patients. All variables in the adjusted analysis were performed using the forced entry procedure, and all analyses were explorative. Analyses were performed using IBM SPSS Statistics version 20 (Armonk, New York, United States).
Results Subjects and baseline characteristics From 1995 to 2008, 17,450 patients were registered with a first episode of SAB. A total of 1781 SAB patients were excluded due to a temporary or invalid personal identifier (n Z 652), negative follow up time (n Z 142), no sample date (n Z 18), a duplicate person identifier (n Z 8), or age75 Sex Female Male Risk factors Hypertension Prior ATE Renal disease Obesity Atrial flutter/fibrillation Human immunodeficiency virus Diabetes Congestive heart failure Cocaine abuse Time period 1995e1997 1998e2000 2001e2003 2004e2006 2007e2008 Origin Hospital-associated Community-acquired Unknown acquisition Endocarditis Methicillin-resistant SAB
Controls
No ATE
ATE (%)
No ATE
ATE (%)
15,391
278(1.8%)
154,120
2570(1.6%)
1081 3066 6525 4719
5(.5%) 48(1.5%) 130(2.0%) 95(2.0%)
10,857 31,036 65,506 46,721
3(.0%) 104(.3%) 1044(1.6%) 1419(2.9%)
6056 9335
106(1.7%) 172(1.8%)
60,838 93,282
782(1.3%) 1788(1.9%)
2920 2596 1514 498 1604 102 2348 1960 24
104(3.4%) 87(3.2%) 49(3.1%) 15(2.9%) 49(3.0%) 3(2.9%) 63(2.6%) 48(2.4%) 0(0.0%)
12,138 15,051 741 1768 7305 59 6623 6434 3
383(3.1%) 823(5.2%) 27(3.5%) 42(2.3%) 274(3.6%) 1(1.7%) 269(3.9%) 320(4.7%) 0(0.0%)
3223 3399 3132 3471 2166
32(1.0%) 66(1.9%) 63(2.0%) 67(1.9%) 50(2.3%)
31,804 34,053 31,473 34,884 21,906
746(2.3%) 597(1.7%) 477(1.5%) 496(1.4%) 254(1.1%)
8686 4704 2001 741 127
157(1.8%) 77(1.6%) 44(2.2%) 21(2.8%) 2(1.6%)
SAB: Staphylococcus aureus bacteremia; ATE: Arterial thromboembolic events.
compared to controls (aRR: 3.5 (95% CI: 2.8e4.5, p Z 0.0001)) but not at 31e180 and 181e365 days after SAB (aRR: 1.1 (95% CI: 0.9e1.4, p Z 0.5) and 0.9 (95% CI: 0.7e1.2, p Z 0.5), respectively.
Incidence rates and relative risk of MI Rates of MI after SAB were highest within the first 30 days and in the oldest, in individuals with prior ATE, hypertension, congestive heart failure or with diabetes mellitus (Table 2). The aRR for MI within 30 days among SAB patients compared to controls was 2.2 (95% CI: 1.6e3.1, p Z 0.0001). The aRR of MI 31e180 and 181e365 days after SAB were lower among SAB patients than controls (0.66 (95% CI: 0.47e0.93, p Z 0.02) and 0.66 (95% CI: 0.46e0.96, p Z 0.03), respectively (Fig. 2)).
Incidence rates and relative risk of stroke Rates of stroke were highest in the oldest and in individuals with prior ATE (Table 3). The aRR for stroke within 30 days
from SAB among SAB patients compared to controls was 5.6 (95% CI: 3.8e8.3), p Z 0.0001). The aRR of stroke 31e180 and 181e365 days after SAB remained elevated among SAB patients compared to controls (1.7 (95% CI: 1.2e2.4, p Z 0.007) and 1.5 (95% CI: 0.98e2.3, p Z 0.07), respectively (Fig. 2)). When SAB patients with endocarditis were excluded, the aRR of stroke within 30, 31e180 and 181e365 days after SAB was 5.1(95% CI: 3.4e7.7, p Z 0.0001), 1.6(95% CI: 1.1e2.3, p Z 0.03) and 1.2(95% CI: 0.7e1.9, p Z 0.5), respectively, compared to controls.
Incidence rates and relative risk of other ATE As a consequence of few events of other ATE, stratified IR were not calculated (Table 4). The aRR for other ATE within 30 days from SAB among SAB patients compared to controls was 15.5 (95% CI: 6.9e35), p Z 0.0001). The aRR of other ATE 31e180 and 181e365 days after SAB remained elevated among SAB patients compared to controls (4.4 (95% CI: 2.4e8.0, p Z 0.0001) and 2.3 (95% CI: 1.1e4.9, p Z 0.03), respectively (Fig. 2)). When SAB patients with
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
Arterial thromboembolism in S. aureus bacteremia
5
Table 2 Incidence rates of acute myocardial infarction among 15,669 cases of Staphylococcus aureus bacteremia and 156,690 controls in Denmark from 1995 through 2008. Events per 1000 person-years of follow up SAB patients
All Age, years 16e35 36e55 56e75 >75 Sex Female Male Origin Hospital- associated Community-acquired Unknown acquisition Previous ATE No Yes Prior hypertension No Yes Prior Renal disease No Yes Prior Atrial flutter/fibrillation No Yes Prior Congestive heart failure No Yes Prior Diabetes No Yes
Controls
0e30 days
31e180 days
181e365 days
0e365 days
49.1(37.3e64.6) (n Z 51)
10.2(7.5e13.8) (n Z 41)
8.6(6.2e11.9) (n Z 37)
11.6(11.1e12.2) (n Z 1689)
0.0 (n Z 0) 8.8 (n Z 2) 62.3(43.0e90.3) (n Z 28) 76.1(49.6e116.7) (n Z 21)
0.0 (n Z 0) 3.0 (n Z 3) 13.8(9.3e20.6) (n Z 24) 15.6(9.2e26.3) (n Z 14)
0.0 (n Z 0) 7.3(3.6e14.6) (n Z 8) 10.4(6.6e16.3) (n Z 19) 11.2(6.0e20.8) (n Z 10)
0.1 (n Z 1) 1.9(1.5e2.4) (n Z 58) 11.3(10.5e12.2) (n Z 708) 22.4(21.0e23.9) (n Z 922)
42.2(26.3e67.9) (n Z 17) 53.4(38.2e74.8) (n Z 34)
8.5(4.9e14.7) (n Z 13) 11.2(7.7e16.2) (n Z 28)
8.6(5.1e14.5) (n Z 14) 8.6(5.7e13.0) (n Z 23)
8.3(7.6e9.1) (n Z 480) 13.9(13.1e14.7) (n Z 1209)
59.2(42.5e82.4) (n Z 35) 32.5(17.5e60.4) (n Z 10) 42.8 (n Z 6)
11.0(7.4e16.3) (n Z 25) 7.4(3.8e14.2) (n Z 9) 12.9(6.1e27.0) (n Z 7)
8.8(5.7e13.4) (n Z 21) 4.4 (n Z 6) 18.1(9.7e33.6) (n Z 10)
e
32.2(22.2e46.6) (n Z 28) 136.4(90.6e205.3) (n Z 23)
7.6(5.2e11.2) (n Z 26) 24.5(14.8e40.7) (n Z 15)
6.5(4.4e9.8) (n Z 24) 20.6(11.9e35.4) (n Z 13)
8.5(8.0e9.0) (n Z 1122) 42.2(38.9e45.8) (n Z 567)
33.5(23.1e48.5) (n Z 28) 113.3(75.3e170.5) (n Z 23)
6.5(4.2e9.9) (n Z 21) 25.5(16.4e39.5) (n Z 20)
3.7(2.2e6.4) (n Z 13) 29.2(19.6e43.5) (n Z 24)
10.7(10.2e11.3) (n Z 1436) 22.5(19.9e25.4) (n Z 253)
43.1(31.6e58.8) (n Z 40) 98.4(54.5e177.6) (n Z 11)
9.2(6.5e13.0) (n Z 33) 17.8(8.9e35.6) (n Z 8)
6.0(4.0e9.0) (n Z 23) 30.0(17.7e50.6) (n Z 14)
11.5(11.0e12.1) (n Z 1668) 34.0(22.2e52.2) (n Z 21)
45.8(33.9e61.7) (n Z 43) 80.4(40.2e160.8) (n Z 8)
7.6(5.2e11.0) (n Z 28) 37.7(21.9e64.9) (n Z 13)
7.8(5.5e11.1) (n Z 31) 17.6 (n Z 6)
11.0(10.5e11.6) (n Z 1530) 24.3(20.8e28.4) (n Z 159)
41.2(30.0e56.6) (n Z 38) 111.2(64.6e191.6) (n Z 13)
9.3(6.7e13.1) (n Z 34) 17.7(8.4e37.1) (n Z 7)
7.1(4.9e10.4) (n Z 28) 23.4(12.2e44.9) (n Z 9)
10.3(9.8e10.9) (n Z 1447) 45.1(39.8e51.2) (n Z 242)
39.9(28.6e55.5) (n Z 35) 99.3(60.8e162.0) (n Z 16)
8.5(5.9e12.2) (n Z 29) 19.2(10.9e33.9) (n Z 12)
7.1(4.9e10.5) (n Z 26) 16.8(9.3e30.3) (n Z 11)
10.8(10.3e11.4) (n Z 1507) 30.8(26.6e35.6) (n Z 182)
e e
(continued on next page)
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
6
N. Mejer et al. Table 2 (continued ) Events per 1000 person-years of follow up SAB patients
Other risk factors No Yes
Controls
0e30 days
31e180 days
181e365 days
0e365 days
48.2(36.3e64.0) (n Z 48) 69.2 (n Z 3)
9.9(7.2e13.5) (n Z 38) 17.0 (n Z 3)
8.5(6.1e11.9) (n Z 35) 10.5 (n Z 2)
11.6(11.0e12.2) (n Z 1663) 15.1(10.3e22.2) (n Z 26)
a
ATE: Arterial thromboembolic events. SAB: Staphylococcus aureus bacteremia. a Obesity, Human immunodeficiency virus, cocaine abuse.
endocarditis were excluded, the aRR of other ATE within 30, 31e180 and 181e365 days after SAB was 15.5(95% CI: 6.9e35, p Z 0.0001), 4.4(95% CI: 2.4e8.0, p Z 0.0001) and 2.5(95% CI: 1.2e5.4, p Z 0.02), respectively, compared to controls.
Risk factors among individuals with SAB In adjusted analysis, the 30-day risk of any ATE was associated with increasing age and a prior diagnosis of ATE (Table 5). In multivariate analysis restricted to SAB patients with prior ATE, no other significant risk factors were identified (data not shown), whereas for SAB patients without prior ATE, the 30-day risk of ATE increased with age and endocarditis (aRR: 2.3 (95% CI: 1.1e5.1, p Z 0.04)). The 30-day risk of MI increased with age, prior ATE and hypertension. For stroke, the 30-day risk was increased with increasing age, prior ATE and endocarditis (Table 5). For other ATE, the 30 day risk increased only with atrial flutter/fibrillation (aRR: 4.6 (95% CI: 1.2e17.6, p Z 0.03)). By one year, increasing age, prior ATE, hypertension (p Z 0.0001), atrial flutter/fibrillation (p Z 0.04) and endocarditis (p Z 0.02) were associated with an increased risk of any ATE (Table 6). In stratified analysis, increasing age, prior ATE, hypertension and renal disease but not endocarditis was associated with an increased risk of MI (Table 6). The 1-year risk of stroke was associated with increasing age, prior ATE and endocarditis (Table 6). For other ATE, the 1-year risk was only increased for atrial flutter/fibrillation (aRR: 2.8 (95% CI: 1.2e6.3, p Z 0.02)).
Discussion Here we demonstrated that SAB was associated with a more than three-fold overall increased short-term risk of any ATE compared to population controls. Other ATE representing a variety of manifestations were associated with the highest risk (15-fold), followed by stroke (5-fold) and MI (2-fold). The increased risk of incident stroke and other ATE persisted beyond 30 days. To our knowledge, this is the first large scale cohort study to evaluate the risk and risk factors of any ATE after systemic S. aureus infection. The study suggests that individuals with SAB may benefit from short-term strategies to prevent ATE.
Several studies have evaluated the role of influenza and pneumonia as a triggering factor for incident ATE but few studies have investigated the role of bacteremia as a triggering event.5e8 Many have limited their study to either MI or stroke. In a self-controlled case-series study of 588 cases, Corrales-Medina et al. found an increased risk of MI immediately after SAB.4 In a study of severe sepsis by Levine et al., rates of MI and stroke were around 1% at day 28 and were highest within the first 10 days.21 A recent Danish study of community-acquired bacteremia of any cause showed a similarly increased risk of stroke and MI within 30 days of bacteremia.3 Their study identified 321 cases of community-acquired SAB with 7 (2.1%) cases of MI and 13 (4%) cases of stroke within 30 days. Our study found significantly fewer incident cases of MI and stroke (approximately 0.3% of each). We believe that differences in case ascertainment may in part explain these differences. In the absence of prospective case ascertainment, we relied on administrative data to capture events and some outcome events may have represented previous ATEs. To account for this, we reviewed discharge summaries of all SAB patients with ATE and discovered that a significant proportion of ATE registered within a year from SAB represented thrombotic events that had occurred prior to SAB and for this reason these SAB patients were excluded from our study. On the other hand, our use of administrative data may not have captured all outcome events although a MI or stroke is a serious event that we believe would be coded as such had it occurred. However, we may have underestimated the true incidence of ATE following SAB. Regardless, our study and the study by Dalager-Pedersen et al. have suggested an increased risk of MI and stroke within 30 days of SAB that persisted for stroke. Our findings are further supported by several studies of MI following pneumonia6,22e24 and of stroke following a number of infections10 that each showed an increased short-term risk of MI and stroke. We found that several types of ATE other than MI and stroke carried the highest risk of incident ATE following SAB. These included occlusion of arteries in the retina, kidney, and intestine that have not been considered previously. Although these other ATE contributed just 17% of all ATE, the relative risk was nearly 15-fold higher than among population controls within 30 days of SAB and the risk persisted throughout the study period of one year. Bacteremia is characterized by a state of inflammation and procoagulation with endothelial damage, vascular
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
Arterial thromboembolism in S. aureus bacteremia
7
Table 3 Incidence rates of stroke among 15,669 cases of Staphylococcus aureus bacteremia and 156,690 controls in Denmark from 1995 through 2008. Events per 1000 person years of follow up SAB patients
All Age, years 16e35 36e55 56e75 >75 Sex Female Male Origin Hospital- associated Community-acquired Unknown acquisition Prior ATE No Yes Prior hypertension No Yes Prior Renal disease No Yes Prior Atrial flutter/fibrillation No Yes Prior Congestive heart failure No Yes Prior Diabetes No Yes
Controls
0e30 days
31e180 days
181e365 days
0e365 days
41.2(30.6e55.6) (n Z 43)
9.4(6.8e12.9) (n Z 38)
6.2(4.3e9.1) (n Z 27)
5.2(4.8e5.6) (n Z 757)
0.0 (n Z 0) 39.6(20.6e76.0) (n Z 9) 24.4(13.5e44.0) (n Z 11) 83.0(55.2e124.9) (n Z 23)
2.4 (n Z 1) 6.1 (n Z 6) 12.0(7.8e18.4) (n Z 21) 11.0(5.9e20.5) (n Z 10)
6.2 (n Z 3) 3.6 (n Z 4) 6.5(3.7e11.4) (n Z 12) 8.8(4.4e17.7) (n Z 8)
0.1 (n Z 1) 1.2(0.8e1.6) (n Z 36) 4.6(4.1e5.2) (n Z 290) 10.4(9.5e11.5) (n Z 430)
52.1(34.0e79.9) (n Z 21) 34.4(22.6e52.2) (n Z 22)
9.8(5.9e16.2) (n Z 15) 9.1(6.1e13.7) (n Z 23)
7.3(4.2e12.9) (n Z 12) 5.6(3.3e9.2) (n Z 15)
4.2(3.7e4.8) (n Z 244) 5.9(5.4e6.4) (n Z 513)
37.0(24.4e56.2) (n Z 22) 55.3(34.3e88.9) (n Z 17) 28.5 (n Z 4)
8.3(5.3e13.0) (n Z 19) 9.8(5.6e17.3) (n Z 12) 12.8(6.1e27.0)’ (n Z 7)
4.5(2.5e8.2) (n Z 11) 8.9(5.0e15.6) (n Z 12) 7.2 (n Z 4)
e
32.1(22.2e46.5) (n Z 28) 87.7(52.9e145.5) (n Z 15)
7.9(5.4e11.5) (n Z 27) 17.5(9.7e31.6) (n Z 11)
6.2(4.2e9.4) (n Z 23) 6.1 (n Z 4)
4.1(3.7e4.4) (n Z 535) 16.4(14.4e18.7) (n Z 222)
39.4(28.0e55.4) (n Z 33) 49.0(26.3e91.0) (n Z 10)
7.0(4.7e10.6) (n Z 23) 18.9(11.4e31.3) (n Z 15)
6.0(3.9e9.2) (n Z 21) 7.1 (n Z 6)
4.8(4.4e5.2) (n Z 641) 10.3(8.6e12.3) (n Z 116)
45.1(33.4e61.1) (n Z 42) 8.9 (n Z 1)
8.6(6.0e12.2) (n Z 31) 15.4(7.3e32.3) (n Z 7)
6.5(4.4e9.6) (n Z 25) 4.2 (n Z 2)
5.2(4.8e5.6) (n Z 752) 8.0 (n Z 5)
41.4(30.2e56.6) (n Z 39) 40.0 (n Z 4)
9.2(6.5e12.8) (n Z 34) 11.5 (n Z 4)
5.3(3.4e8.1) (n Z 21) 17.2 (n Z 6)
4.8(4.4e5.2) (n Z 664) 14.2(11.6e17.4) (n Z 93)
41.1(29.9e56.5) (n Z 38) 42.5 (n Z 5)
8.8(6.2e12.4) (n Z 32) 14.9 (n Z 6)
6.6(4.5e9.7) (n Z 26) 2.5 (n Z 1)
5.0(4.6e5.3) (n Z 694) 11.6(9.1e14.9) (n Z 63)
39.7(28.5e55.3) (n Z 35) 49.5(24.7e98.9) (n Z 8)
9.0(6.4e12.8) (n Z 31) 11.2(5.3e23.4) (n Z 7)
6.5(4.4e9.7) (n Z 24) 4.5 (n Z 3)
4.9(4.6e5.3) (n Z 686) 12.0(9.5e15.1) (n Z 71)
e e
(continued on next page)
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
8
N. Mejer et al. Table 3 (continued ) Events per 1000 person years of follow up SAB patients
Other risk factors No
Controls
0e30 days
31e180 days
181e365 days
0e365 days
40.0(29.4e54.6) (n Z 40) 69.3 (n Z 3)
9.0(6.5e12.6) (n Z 35) 17.0 (n Z 3)
6.3(4.3e9.2) (n Z 26) 5.2 (n Z 1)
5.2(4.8e5.5) (n Z 742) 8.7(5.3e14.5) (n Z 15)
a
Yes
ATE: Arterial thromboembolic events. SAB: Staphylococcus aureus bacteremia. a Obesity, Human immunodeficiency virus, cocaine abuse.
Table 4 Incidence rates of other arterial thromboembolism among 15,669 cases of Staphylococcus aureus bacteremia and 156,690 controls in Denmark from 1995 through 2008. Events per 1000 person years of follow up SAB patients
All
Controls
0e30 days
31e180 days
181e365 days
0e365 days
15.3(9.4e25.0) (n Z 16)
4.7(3.0e7.3) (n Z 19)
2.5(1.4e4.6) (n Z 11)
1.1(0.9e1.3) (n Z 156)
leakage, and hypotension,25,26 events that each may induce arterial occlusion. Further, infections may result in atherosclerotic plaque instability and rupture leading to arterial thrombosis.5 In this study the long term risk of stroke and other ATE was elevated compared to controls, but was lower for MI. A possible explanation for this observation may be that the short-term risk of any ATE was determined by plaque instability and rupture induced by the bacteremic event whereas the persistent increased risk of stroke and other ATE was determined by complicating events in the form of new-onset atrial fibrillation and endocarditis. Certainly, new-onset atrial fibrillation is frequently associated with sepsis and may lead to thromboembolic events12 and in our study prior atrial flutter/fibrillation significantly increased the 1-year risk of any ATE. Endocarditis may, of course, be a source of septic embolism. In our study, 762 (5%) of SAB patients had endocarditis based on discharge
reports, and endocarditis significantly increased the risk of stroke. However, our estimate of endocarditis was likely an underestimation because up to a fifth of all patients with SAB have endocarditis if evaluated by echocardiography.27 This further highlights the importance of identifying the infective focus in patients with SAB because the presence of endocarditis was associated with an increased shortand long-term risk of stroke. Importantly, even when SAB patients with registered endocarditis were excluded there remained an excess increased risk of stroke, other ATE and any ATE compared to controls. Optimal preventive treatment strategies are lacking because the mechanism of cardiovascular complications after infection is unclear.9 Possible strategies may include optimal treatment of bacteremia, anticoagulation during and after bacteremia, or vaccination of individuals at risk. We found that risk factors that may help to designate
Figure 1 Adjusted relative risks of any arterial thromboembolic events (ATE) in SAB patients versus controls in the year following Staphylococcus aureus bacteremia.
Figure 2 Adjusted relative risks of myocardial infarction, stroke or other ATE in SAB patients versus controls in the year following Staphylococcus aureus bacteremia.
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
Arterial thromboembolism in S. aureus bacteremia
9
Table 5 Risk factors associated with any arterial thromboembolism, with myocardial infarction or with stroke within 30-days after Staphylococcus aureus bacteremia.
Age 16e35 36e55 56e75 >75 Sex Female Male Origin Hospital- associated Community-acquired Unknown acquisition Prior ATE No Yes Prior Hypertension No Yes Prior Renal disease No Yes Prior Atrial flutter/fibrillation No Yes Prior Congestive heart failure No Yes Prior Diabetes No Yes Other risk factorsa No Yes Endocarditis No Yes
Any arterial thromboembolism aRR (95% CI)
Myocardial infarction aRR (95% CI)
Stroke aRR (95% CI)
n Z 109
n Z 51
n Z 43
0.1(0.0e0.7) 0.6(0.3e1.0) 0.5(0.3e0.8) 1.0
0.0 0.1(0.0e0.6) 0.8(0.4e1.4) 1.0
0.0 0.6(0.3e1.3) 0.3(0.2e0.7) 1.0
1.0 0.9(0.6e1.3)
1.0 1.1(0.6e2.0)
1.0 0.7(0.4e1.2)
1.0 1.1(0.7e1.7) 0.8(0.4e1.5)
1.0 0.7(0.3e1.3) 0.8(0.3e1.8)
1.0 1.2(0.6e2.4) 0.8(0.3e2.4)
1.0 2.4(1.6e3.6)
1.0 2.5(1.4e4.5)
1.0 2.8(1.4e5.4)
1.0 1.5(1.0e2.3)
1.0 2.0(1.1e3.7)
1.0 1.2(0.6e2.6)
1.0 0.9(0.5e1.7)
1.0 1.4(0.7e2.8)
1.0 0.2(0.0e1.5)
1.0 1.0(0.6e1.7)
1.0 0.8(0.4e1.9)
1.0 0.6(0.2e1.9)
1.0 1.0(0.6e1.7)
1.0 1.2(0.6e2.4)
1.0 0.7(0.2e1.8)
1.0 1.3(0.8e2.1)
1.0 1.6(0.9e3.0)
1.0 1.2(0.6e2.8)
1.0 1.4(0.6e3.2)
1.0 1.1(0.3e3.6)
1.0 2.0(0.6e6.8)
1.0 1.4(0.6e3.0)
1.0 0.5(0.1e3.3)
1.0 3.3(1.3e7.9)
a Obesity, Human immunodeficiency virus, cocaine abuse. ATE: Arterial thromboembolic events. aRR: adjusted relative risk, CI: confidence interval.
high-risk individuals of ATE after bacteremia include old age, prior ATE, atrial flutter/fibrillation, hypertension and endocarditis. The strengths of this study lie in its size, the longitudinal and population-based design, the inclusion based on the exposure to infection rather than the outcome of interest, complete follow-up on both mortality and ATE, the review of discharge reports, and the use of randomly selected age- and sex matched population controls for comparative analysis. Still, there are potential limitations. First, because disease criteria relied on diagnostic codes from the NPR, it is likely that a degree of misclassification may have occurred. However, validation
studies have demonstrated high validity of MI and stroke diagnoses with positive predictive values of 86e97% for MI and 76e95% for stroke, respectively,28e34 while the negative predictive values have been estimated to be 53% for MI and 79% for stroke.31 To further improve the diagnostic accuracy, we performed a review of all discharge notes to ensure that a case of ATE in fact had occurred after a diagnosis of SAB. The validity of other ATE has not been assessed, but these contributed less than 20% of the total number of cases and we performed stratified analysis based on the three different endpoints. Second, information on use of anticoagulants was not available, and we may have underestimated the true risk of ATE following
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
10
N. Mejer et al.
Table 6 Risk factors associated with any arterial thromboembolism, with myocardial infarction or with stroke within a year after Staphylococcus aureus bacteremia.
Age 16e35 36e55 56e75 >75 Sex Female Male Origin Hospital- associated Community-acquired Unknown acquisition Prior ATE No Yes Prior Hypertension No Yes Prior Renal disease No Yes Prior Atrial flutter/fibrillation No Yes Prior Congestive heart failure No Yes Prior Diabetes No Yes Other risk factorsa No Yes Endocarditis No Yes
Any arterial thromboembolism aRR (95% CI)
Myocardial infarction aRR (95% CI)
Stroke aRR (95% CI)
n Z 278
n Z 129
n Z 108
0.2(0.1e0.4) 0.6(0.4e0.9) 0.8(0.6e1.0) 1.0
0.0 0.3(0.2e0.7) 0.8(0.6e1.2) 1.0
0.2(0.1e0.7) 0.5(0.3e0.9) 0.6(0.4e1.0) 1.0
1.0 0.9(0.7e1.2)
1.0 1.1(0.7e1.6)
1.0 0.8(0.5e1.1)
1.0 1.0(0.8e1.3) 1.2(0.8e1.7)
1.0 0.7(0.5e1.1) 1.2(0.8e1.9)
1.0 1.4(0.9e2.1) 1.2(0.7e2.2)
1.0 1.8(1.4e2.3)
1.0 2.0(1.4e2.9)
1.0 1.8(1.2e2.9)
1.0 1.8(1.3e2.3)
1.0 2.6(1.8e3.9)
1.0 1.5(1.0e2.4)
1.0 1.2(0.9e1.8)
1.0 1.5(1.0e2.3)
1.0 0.8(0.4e1.6)
1.0 1.4(1.0e2.0)
1.0 1.4(0.9e2.2)
1.0 1.1(0.6e2.0)
1.0 1.0(0.7e1.4)
1.0 1.0(0.7e1.6)
1.0 0.7(0.4e1.3)
1.0 1.1(0.8e1.5)
1.0 1.4(0.9e2.1)
1.0 0.9(0.5e1.6)
1.0 1.3(0.8e2.2)
1.0 1.1(0.5e2.2)
1.0 1.6(0.7e3.6)
1.0 1.7(1.1e2.7)
1.0 1.3(0.6e2.8)
1.0 3.1(1.8e5.4)
a Obesity, Human immunodeficiency virus, cocaine abuse. ATE: Arterial thromboembolic events. aRR: adjusted relative risk, CI: confidence interval.
SAB. Third, in this study the primary outcome was based on registrations of primary diagnosis codes of ATE within 365 days from SAB. Because we did not use secondary codes we may have underestimated the rate of ATE after SAB. Fourth, due to lack of data we were unable to adjust for several well-known cardiovascular risk factors including smoking, hypercholesterolemia, or a family history of ATE. In conclusion, SAB was associated with an increased risk of MI, stroke and ‘other’ ATE within 30 days of SAB that persisted for stroke and other ATE. Further studies are required to determine the efficacy of prevention and treatment strategies to lower the risk ATE in patients with SAB.
Competing interests A.G. Jensen has been employed at Pfizer Denmark since 2005 and Bristol-Myers Squibb Denmark since 2014 but neither Pfizer Denmark nor Bristol-Myers Squibb Denmark has financial interests in- or support to the present study. All other authors declare that they have no competing interests.
Financial support This work was supported by Preben and Anna Simonsen’s Foundation (021892-0003).
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
Arterial thromboembolism in S. aureus bacteremia
11
Acknowledgments The authors thank the staff at the participating departments of clinical microbiology for their contribution, continuous support and enthusiasm.
16.
17.
Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jinf.2015.03.010
18.
References
19.
1. Feigin VL, Lawes CM, Bennett DA, Barker-Collo SL, Parag V. Worldwide stroke incidence and early case fatality reported in 56 population-based studies: a systematic review. Lancet Neurol 2009 Apr;8(4):355e69. 2. Yeh RW, Sidney S, Chandra M, Sorel M, Selby JV, Go AS. Population trends in the incidence and outcomes of acute myocardial infarction. N Engl J Med 2010 Jun 10;362(23):2155e65. 3. Dalager-Pedersen M, Sogaard M, Schonheyder HC, Nielsen H, Wernich TR. Risk for myocardial infarction and stroke after community-acquired bacteremia: a 20-Year Population-Based Cohort Study. Circulation 2014 Apr 1;129(13):1387e96. 4. Corrales-Medina VF, Fatemi O, Serpa J, Valayam J, Bozkurt B, Madjid M, et al. The association between Staphylococcus aureus bacteremia and acute myocardial infarction. Scand J Infect Dis 2009;41(6e7):511e4. 5. Corrales-Medina VF, Madjid M, Musher DM. Role of acute infection in triggering acute coronary syndromes. Lancet Infect Dis 2010 Feb;10(2):83e92. 6. Corrales-Medina VF, Musher DM, Shachkina S, Chirinos JA. Acute pneumonia and the cardiovascular system. Lancet 2013 Feb 9;381(9865):496e505. 7. Smeeth L, Thomas SL, Hall AJ, Hubbard R, Farrington P, Vallance P. Risk of myocardial infarction and stroke after acute infection or vaccination. N Engl J Med 2004 Dec 16;351(25): 2611e8. 8. Warren-Gash C, Smeeth L, Hayward AC. Influenza as a trigger for acute myocardial infarction or death from cardiovascular disease: a systematic review. Lancet Infect Dis 2009 Oct; 9(10):601e10. 9. Walkey AJ. Preventing cardiovascular complications of acute infection: a missed opportunity? Circulation 2014 Apr 1; 129(13):1375e7. 10. Guiraud V, Amor MB, Mas JL, Touze E. Triggers of ischemic stroke: a systematic review. Stroke 2010 Nov;41(11):2669e77. 11. Paganini-Hill A, Lozano E, Fischberg G, Perez BM, Rajamani K, Ameriso SF, et al. Infection and risk of ischemic stroke: differences among stroke subtypes. Stroke 2003 Feb;34(2):452e7. 12. Lee-Iannotti JK, Capampangan DJ, Hoffman-Snyder C, Wellik KE, Patel B, Tondato F, et al. New-onset atrial fibrillation in severe sepsis and risk of stroke and death: a critically appraised topic. Neurologist 2012 Jul;18(4):239e43. 13. Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology of sepsis in the United States from 1979 through 2000. N Engl J Med 2003 Apr 17;348(16):1546e54. 14. Mejer N, Westh H, Schonheyder HC, Jensen AG, Larsen AR, Skov R, et al. Stable incidence and continued improvement in short term mortality of Staphylococcus aureus bacteraemia between 1995 and 2008. BMC Infect Dis 2012 Oct 17;12(1):260. 15. Sogaard M, Norgaard M, Dethlefsen C, Schonheyder HC. Temporal changes in the incidence and 30-day mortality associated
20.
21.
22.
23.
24.
25. 26. 27.
28.
29.
30.
31.
32.
with bacteremia in hospitalized patients from 1992 through 2006: a population-based cohort study. Clin Infect Dis 2011 Jan 1;52(1):61e9. Pedersen CB, Gotzsche H, Moller JO, Mortensen PB. The Danish Civil Registration System. A cohort of eight million persons. Dan Med Bull 2006 Nov;53(4):441e9. Jessen O, Rosendal K, Bulow P, Faber V, Eriksen KR. Changing staphylococci and staphylococcal infections. A ten-year study of bacteria and cases of bacteremia. N Engl J Med 1969 Sep 18;281(12):627e35. Wacholder S, McLaughlin JK, Silverman DT, Mandel JS. Selection of controls in case-control studies. I. Principles. Am J Epidemiol 1992 May 1;135(9):1019e28. Andersen TF, Madsen M, Jorgensen J, Mellemkjær L, Olsen JH. The Danish National Hospital Register. A valuable source of data for modern health sciences. Dan Med Bull 1999 Jun; 46(3):263e8. Benfield T, Espersen F, Frimodt-Moller N, Jensen AG, Larsen AR, Pallesen LV, et al. Increasing incidence but decreasing in-hospital mortality of adult Staphylococcus aureus bacteraemia between 1981 and 2000. Clin Microbiol Infect 2007 Mar;13(3):257e63. Levine RL, LeClerc JR, Bailey JE, Monberg MJ, Sarwat S. Venous and arterial thromboembolism in severe sepsis. Thromb Haemost 2008 May;99(5):892e8. Corrales-Medina VF, Musher DM, Wells GA, Chirinos JA, Chen L, Fine MJ. Cardiac complications in patients with communityacquired pneumonia: incidence, timing, risk factors, and association with short-term mortality. Circulation 2012 Feb 14; 125(6):773e81. Musher DM, Rueda AM, Kaka AS, Mapara SM. The association between pneumococcal pneumonia and acute cardiac events. Clin Infect Dis 2007 Jul 15;45(2):158e65. Corrales-Medina VF, Suh KN, Rose G, Chirinos JA, Doucette S, Cameron DW, et al. Cardiac complications in patients with community-acquired pneumonia: a systematic review and meta-analysis of observational studies. PLoS Med 2011 Jun; 8(6):e1001048. Dellinger RP. Inflammation and coagulation: implications for the septic patient. Clin Infect Dis 2003 May 15;36(10):1259e65. Libby P, Simon DI. Inflammation and thrombosis: the clot thickens. Circulation 2001 Apr 3;103(13):1718e20. Rasmussen RV, Host U, Arpi M, Hassager C, Johansen HK, Korup E, et al. Prevalence of infective endocarditis in patients with Staphylococcus aureus bacteraemia: the value of screening with echocardiography. Eur J Echocardiogr 2011 Jun;12(6):414e20. Petersen LA, Wright S, Normand SL, Daley J. Positive predictive value of the diagnosis of acute myocardial infarction in an administrative database. J Gen Intern Med 1999 Sep;14(9):555e8. Kiyota Y, Schneeweiss S, Glynn RJ, Cannuscio CC, Avorn J, Solomon DH. Accuracy of medicare claims-based diagnosis of acute myocardial infarction: estimating positive predictive value on the basis of review of hospital records. Am Heart J 2004 Jul;148(1):99e104. Rinaldi R, Vignatelli L, Galeotti M, Azzimondi G, de CP. Accuracy of ICD-9 codes in identifying ischemic stroke in the General Hospital of Lugo di Romagna (Italy). Neurol Sci 2003 Jun; 24(2):65e9. Birman-Deych E, Waterman AD, Yan Y, Nilasena DS, Radford MJ, Gage BF. Accuracy of ICD-9-CM codes for identifying cardiovascular and stroke risk factors. Med Care 2005 May;43(5):480e5. Tolonen H, Salomaa V, Torppa J, Sivenius J, Immonen-Raiha P, Lehtonen A. The validation of the Finnish Hospital Discharge
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
12 Register and Causes of Death Register data on stroke diagnoses. Eur J Cardiovasc Prev Rehabil 2007 Jun;14(3):380e5. 33. Rosamond WD, Chambless LE, Sorlie PD, Bell EM, Weitzman S, Smith JC, et al. Trends in the sensitivity, positive predictive value, false-positive rate, and comparability ratio of hospital discharge diagnosis codes for acute myocardial infarction in
N. Mejer et al. four US communities, 1987-2000. Am J Epidemiol 2004 Dec 15;160(12):1137e46. 34. Cutrona SL, Toh S, Iyer A, Foy S, Daniel GW, Nair VP, et al. Validation of acute myocardial infarction in the Food and Drug Administration’s Mini-Sentinel program. Pharmacoepidemiol Drug Saf 2013 Jan;22(1):40e54.
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
www.elsevierhealth.com/journals/jinf
Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study N. Mejer a,*, N. Gotland a,1, M.L. Uhre a,1, H. Westh b,c,2, H.C. Schønheyder d,e,3, A. Petersen f,4, A.G. Jensen f,g,5, A.R. Larsen f,4, R. Skov f,4, T. Benfield a,b,h,1, for the Danish Staphylococcal Bacteremia Study Group a
Department of Infectious Diseases, Hvidovre Hospital, University of Copenhagen, Hvidovre, Denmark Department of Clinical Medicine, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark c Department of Clinical Microbiology, Hvidovre Hospital, University of Copenhagen, Hvidovre, Denmark d Department of Clinical Microbiology, Aalborg University Hospital, Aalborg, Denmark e Department of Clinical Medicine, Aalborg University, Aalborg, Denmark f Staphylococcus Laboratory, Statens Serum Institut, Copenhagen, Denmark g Bristol-Myers Squibb, Virum, Denmark h Clinical Research Centre, Hvidovre Hospital, University of Copenhagen, Hvidovre, Denmark b
Accepted 18 March 2015 Available online - - -
KEYWORDS Staphylococcus aureus; Bacteremia; Acute myocardial infarction; Stroke; Arterial thromboembolism
Summary Objectives: An association between infection and arterial thromboembolic events (ATE) has been suggested. Here we examined the risk of myocardial infarction (MI), stroke and other ATE after Staphylococcus aureus bacteremia (SAB). Methods: Danish register-based nation-wide observational cohort study between 1995 and 2008 with matched control subjects from the general population. Results: Within a year, 278 of 15,669 SAB patients and 2570 of 156,690 controls developed MI, stroke or another ATE. The incidence rates among SAB patients were highest within the first 30 days and decreased over a year. The adjusted relative risk of MI, stroke and other ATE during
* Corresponding author. Ketteg ard Alle ´ 30, 2650 Hvidovre, afd 144, Denmark. Tel.: þ45 29845959. E-mail address: [email protected] (N. Mejer). 1 Ketteg ard Alle ´ 30, 2650 Hvidovre, afd 144, Denmark. 2 Ketteg ard Alle ´ 30, 2650 Hvidovre, afd 445, Denmark. 3 Hobrovej 18-22, 9000 Aalborg, Denmark. 4 Artillerivej 5, Copenhagen, 2300, Denmark. 5 Bristol-Myers Squibb, Hummeltoftevej 49, 2830 Virum, Denmark. http://dx.doi.org/10.1016/j.jinf.2015.03.010 0163-4453/ª 2015 The British Infection Association. Published by Elsevier Ltd. All rights reserved. Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
2
N. Mejer et al. the first 30 days after SAB in patients compared to controls were 2.2 (95% CI: 1.6e3.1), 5.5 (95% CI: 3.8e8.3) and 15.5 (95% CI: 6.9e35), respectively. Compared to controls, the increased adjusted relative risk persisted for 30 days for MI, 180 days for stroke and one year for other ATE. Increasing age, hypertension, atrial flutter/fibrillation, prior ATE and endocarditis in SAB patients were associated with an increased risk of ATE. Conclusions: SAB was associated with a short-term increased risk of ATE that persisted longer dependent on type of event. Studies are warranted to investigate treatment strategies to diminish ATE after SAB. ª 2015 The British Infection Association. Published by Elsevier Ltd. All rights reserved.
Introduction Arterial thromboembolic events (ATE) such as myocardial infarction (MI) and stroke are frequent clinical events with substantial morbidity and mortality.1,2 Inflammation induced by acute or chronic infection is believed to account for a significant proportion of arterial thromboembolism, both as a triggering and participating condition. Recent epidemiological studies have linked infections to acute coronary events3e8 and ischemic stroke.3,9e12 The majority of the studies investigated the risk of ATE after pneumonia, urinary tract infection, or influenza, while few addressed the risk of ATE following bacteremia. A recent Danish study found high rates of MI and stroke after community-acquired bacteremia caused by a variety of etiologies, including SAB and other agents.3 Bacteremia is frequent, at an incidence of 1e2 episodes per 1000 population years and Staphylococcus aureus is one of the leading causes.13e15 Here we sought to estimate the risk of MI, stroke and other ATE within the first year after community- and hospital-acquired SAB and the relative risks compared with population-based matched controls. Further, we identified risk factors that may help to distinguish highrisk groups that could potentially benefit from prevention and treatment strategies.
Materials and methods Design and ethics The study was a register-based matched cohort study, using the unique civil registration number assigned to each Danish resident through the Civil Registration System (CRS).16 The CRS tracks daily changes in vital status, including date of emigration and date of death, for the entire Danish population. The CRS is almost 100% complete so for practical purposes no persons are lost to follow up. The CRS number allowed merging of data from three national databases. The study was approved by the Science Ethics Committee for Copenhagen and Frederiksberg Counties (01-369 ⁄ 93) and the Danish Data Protection Agency (2009-41-4179). Informed consent is not required by Danish legislation for register-based studies.
Study population A continuous registration of SAB in Denmark has been carried out by the Staphylococcus Laboratory at Statens Serum Institut (SSI), Copenhagen since 1957.14,17 The present study defined a case as an individual 16 years or older with a first time SAB with
a registered blood culture date identified via the registry between January 1, 1995 and December 31, 2008. We defined hospital-associated (HA) origin of SAB as an infection acquired after 2 days of hospitalization, an infection in a patient in ambulatory care, or a case residing in a nursing home. Community-acquired (CA) SAB was defined as an infection present or incubating at the time of hospitalization in patients not meeting the definition of HA SAB. If data were inadequate, acquisition of SAB was classified as unknown. For each case, we randomly selected 10 control subjects through the CRS matched by age (year of birth) and sex. We used the risk set sampling technique (i.e. eligible control subjects had to be alive and at risk of a first hospitalization with SAB on the date the corresponding case was admitted18).
Arterial thromboembolism Data on myocardial infarction, stroke and other ATE was collected from the National Patient Registry (NPR).19 Other ATE included occlusion of arteries in the retina, kidney, and intestine (See Supplementary Table 1). The NPR is updated monthly and holds information on all hospital admissions and discharge diagnoses of all patients treated in a Danish non-psychiatric hospital since January 1, 1977, as well as all outpatient contacts since January 1, 1995. Information obtained included date of admission, date of discharge, and diagnoses. Both primary and secondary diagnosis codes were used to identify comorbidity prior to SAB, while only primary codes were used to establish an ATE event. The International Classification of Diseases (ICD) codes were used to determine diagnoses of ATE (acute myocardial infarction, stroke or another ATE); the 8th ICD revision was used from 1977 to 1993 and the 10th revision thereafter. SAB patients and controls were followed until December 31, 2009. In Denmark a short summary on each admission is provided by the discharging doctor. For SAB patients, these reports were reviewed if a primary diagnosis of ATE occurred within 365 days of the blood culture date. To avoid reverse causality (i.e. that ATE precipitated SAB), only cases of ATE occurring on either the same day of SAB or later were included in this study. If ATE occurred during the same admission as SAB, the number of days from SAB to ATE was determined. Otherwise the date of admission was registered as the date of onset of ATE.
Risk factors Information on congestive heart failure, atrial flutter/ fibrillation, hypertension, renal disease, diabetes, cocaine
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
Arterial thromboembolism in S. aureus bacteremia use, human immunodeficiency virus, and obesity were obtained from the NPR (Supplementary Table 1). For ATE and atrial flutter/fibrillation any event prior to SAB was considered, whereas obesity, congestive heart failure, renal disease, hypertension, human immunodeficiency virus, and diabetes mellitus were considered if they occurred within 10 years preceding the SAB index date. Information on endocarditis was extracted from discharge reports for each admission with SAB as described previously.20
Statistics Age was divided into 4 groups (16e35, 36e54, 55e75, >75 years). The variable “Other risk factors” was computed by combining the variables obesity, cocaine use and human immunodeficiency virus. All patients were followed from the date of SAB; data was censored on the date of death or emigration or on December 31, 2009, whichever came first. Thus, the competing risk of death was accounted for in the analysis by means of censoring on the date of death. Categorical variables were compared by Chi-square tests. Incidence rates (IR) were calculated by dividing the observed incidences by the accumulated person-years (PYs) at risk in the respective groups, and were presented with 95% confidence intervals (CIs). The CIs were only calculated when the number of events were at least 7. Unadjusted and adjusted Cox regression analysis with 95% CIs was performed to determine the relative risk (RR) of ATE within the first year of the SAB index date. If more than one myocardial infarction, stroke or other arterial thromboembolic events was registered, only the first observed episode was used for analysis. For the survival analysis, controls were censored along with their matched SAB patients. All variables in the adjusted analysis were performed using the forced entry procedure, and all analyses were explorative. Analyses were performed using IBM SPSS Statistics version 20 (Armonk, New York, United States).
Results Subjects and baseline characteristics From 1995 to 2008, 17,450 patients were registered with a first episode of SAB. A total of 1781 SAB patients were excluded due to a temporary or invalid personal identifier (n Z 652), negative follow up time (n Z 142), no sample date (n Z 18), a duplicate person identifier (n Z 8), or age75 Sex Female Male Risk factors Hypertension Prior ATE Renal disease Obesity Atrial flutter/fibrillation Human immunodeficiency virus Diabetes Congestive heart failure Cocaine abuse Time period 1995e1997 1998e2000 2001e2003 2004e2006 2007e2008 Origin Hospital-associated Community-acquired Unknown acquisition Endocarditis Methicillin-resistant SAB
Controls
No ATE
ATE (%)
No ATE
ATE (%)
15,391
278(1.8%)
154,120
2570(1.6%)
1081 3066 6525 4719
5(.5%) 48(1.5%) 130(2.0%) 95(2.0%)
10,857 31,036 65,506 46,721
3(.0%) 104(.3%) 1044(1.6%) 1419(2.9%)
6056 9335
106(1.7%) 172(1.8%)
60,838 93,282
782(1.3%) 1788(1.9%)
2920 2596 1514 498 1604 102 2348 1960 24
104(3.4%) 87(3.2%) 49(3.1%) 15(2.9%) 49(3.0%) 3(2.9%) 63(2.6%) 48(2.4%) 0(0.0%)
12,138 15,051 741 1768 7305 59 6623 6434 3
383(3.1%) 823(5.2%) 27(3.5%) 42(2.3%) 274(3.6%) 1(1.7%) 269(3.9%) 320(4.7%) 0(0.0%)
3223 3399 3132 3471 2166
32(1.0%) 66(1.9%) 63(2.0%) 67(1.9%) 50(2.3%)
31,804 34,053 31,473 34,884 21,906
746(2.3%) 597(1.7%) 477(1.5%) 496(1.4%) 254(1.1%)
8686 4704 2001 741 127
157(1.8%) 77(1.6%) 44(2.2%) 21(2.8%) 2(1.6%)
SAB: Staphylococcus aureus bacteremia; ATE: Arterial thromboembolic events.
compared to controls (aRR: 3.5 (95% CI: 2.8e4.5, p Z 0.0001)) but not at 31e180 and 181e365 days after SAB (aRR: 1.1 (95% CI: 0.9e1.4, p Z 0.5) and 0.9 (95% CI: 0.7e1.2, p Z 0.5), respectively.
Incidence rates and relative risk of MI Rates of MI after SAB were highest within the first 30 days and in the oldest, in individuals with prior ATE, hypertension, congestive heart failure or with diabetes mellitus (Table 2). The aRR for MI within 30 days among SAB patients compared to controls was 2.2 (95% CI: 1.6e3.1, p Z 0.0001). The aRR of MI 31e180 and 181e365 days after SAB were lower among SAB patients than controls (0.66 (95% CI: 0.47e0.93, p Z 0.02) and 0.66 (95% CI: 0.46e0.96, p Z 0.03), respectively (Fig. 2)).
Incidence rates and relative risk of stroke Rates of stroke were highest in the oldest and in individuals with prior ATE (Table 3). The aRR for stroke within 30 days
from SAB among SAB patients compared to controls was 5.6 (95% CI: 3.8e8.3), p Z 0.0001). The aRR of stroke 31e180 and 181e365 days after SAB remained elevated among SAB patients compared to controls (1.7 (95% CI: 1.2e2.4, p Z 0.007) and 1.5 (95% CI: 0.98e2.3, p Z 0.07), respectively (Fig. 2)). When SAB patients with endocarditis were excluded, the aRR of stroke within 30, 31e180 and 181e365 days after SAB was 5.1(95% CI: 3.4e7.7, p Z 0.0001), 1.6(95% CI: 1.1e2.3, p Z 0.03) and 1.2(95% CI: 0.7e1.9, p Z 0.5), respectively, compared to controls.
Incidence rates and relative risk of other ATE As a consequence of few events of other ATE, stratified IR were not calculated (Table 4). The aRR for other ATE within 30 days from SAB among SAB patients compared to controls was 15.5 (95% CI: 6.9e35), p Z 0.0001). The aRR of other ATE 31e180 and 181e365 days after SAB remained elevated among SAB patients compared to controls (4.4 (95% CI: 2.4e8.0, p Z 0.0001) and 2.3 (95% CI: 1.1e4.9, p Z 0.03), respectively (Fig. 2)). When SAB patients with
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
Arterial thromboembolism in S. aureus bacteremia
5
Table 2 Incidence rates of acute myocardial infarction among 15,669 cases of Staphylococcus aureus bacteremia and 156,690 controls in Denmark from 1995 through 2008. Events per 1000 person-years of follow up SAB patients
All Age, years 16e35 36e55 56e75 >75 Sex Female Male Origin Hospital- associated Community-acquired Unknown acquisition Previous ATE No Yes Prior hypertension No Yes Prior Renal disease No Yes Prior Atrial flutter/fibrillation No Yes Prior Congestive heart failure No Yes Prior Diabetes No Yes
Controls
0e30 days
31e180 days
181e365 days
0e365 days
49.1(37.3e64.6) (n Z 51)
10.2(7.5e13.8) (n Z 41)
8.6(6.2e11.9) (n Z 37)
11.6(11.1e12.2) (n Z 1689)
0.0 (n Z 0) 8.8 (n Z 2) 62.3(43.0e90.3) (n Z 28) 76.1(49.6e116.7) (n Z 21)
0.0 (n Z 0) 3.0 (n Z 3) 13.8(9.3e20.6) (n Z 24) 15.6(9.2e26.3) (n Z 14)
0.0 (n Z 0) 7.3(3.6e14.6) (n Z 8) 10.4(6.6e16.3) (n Z 19) 11.2(6.0e20.8) (n Z 10)
0.1 (n Z 1) 1.9(1.5e2.4) (n Z 58) 11.3(10.5e12.2) (n Z 708) 22.4(21.0e23.9) (n Z 922)
42.2(26.3e67.9) (n Z 17) 53.4(38.2e74.8) (n Z 34)
8.5(4.9e14.7) (n Z 13) 11.2(7.7e16.2) (n Z 28)
8.6(5.1e14.5) (n Z 14) 8.6(5.7e13.0) (n Z 23)
8.3(7.6e9.1) (n Z 480) 13.9(13.1e14.7) (n Z 1209)
59.2(42.5e82.4) (n Z 35) 32.5(17.5e60.4) (n Z 10) 42.8 (n Z 6)
11.0(7.4e16.3) (n Z 25) 7.4(3.8e14.2) (n Z 9) 12.9(6.1e27.0) (n Z 7)
8.8(5.7e13.4) (n Z 21) 4.4 (n Z 6) 18.1(9.7e33.6) (n Z 10)
e
32.2(22.2e46.6) (n Z 28) 136.4(90.6e205.3) (n Z 23)
7.6(5.2e11.2) (n Z 26) 24.5(14.8e40.7) (n Z 15)
6.5(4.4e9.8) (n Z 24) 20.6(11.9e35.4) (n Z 13)
8.5(8.0e9.0) (n Z 1122) 42.2(38.9e45.8) (n Z 567)
33.5(23.1e48.5) (n Z 28) 113.3(75.3e170.5) (n Z 23)
6.5(4.2e9.9) (n Z 21) 25.5(16.4e39.5) (n Z 20)
3.7(2.2e6.4) (n Z 13) 29.2(19.6e43.5) (n Z 24)
10.7(10.2e11.3) (n Z 1436) 22.5(19.9e25.4) (n Z 253)
43.1(31.6e58.8) (n Z 40) 98.4(54.5e177.6) (n Z 11)
9.2(6.5e13.0) (n Z 33) 17.8(8.9e35.6) (n Z 8)
6.0(4.0e9.0) (n Z 23) 30.0(17.7e50.6) (n Z 14)
11.5(11.0e12.1) (n Z 1668) 34.0(22.2e52.2) (n Z 21)
45.8(33.9e61.7) (n Z 43) 80.4(40.2e160.8) (n Z 8)
7.6(5.2e11.0) (n Z 28) 37.7(21.9e64.9) (n Z 13)
7.8(5.5e11.1) (n Z 31) 17.6 (n Z 6)
11.0(10.5e11.6) (n Z 1530) 24.3(20.8e28.4) (n Z 159)
41.2(30.0e56.6) (n Z 38) 111.2(64.6e191.6) (n Z 13)
9.3(6.7e13.1) (n Z 34) 17.7(8.4e37.1) (n Z 7)
7.1(4.9e10.4) (n Z 28) 23.4(12.2e44.9) (n Z 9)
10.3(9.8e10.9) (n Z 1447) 45.1(39.8e51.2) (n Z 242)
39.9(28.6e55.5) (n Z 35) 99.3(60.8e162.0) (n Z 16)
8.5(5.9e12.2) (n Z 29) 19.2(10.9e33.9) (n Z 12)
7.1(4.9e10.5) (n Z 26) 16.8(9.3e30.3) (n Z 11)
10.8(10.3e11.4) (n Z 1507) 30.8(26.6e35.6) (n Z 182)
e e
(continued on next page)
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
6
N. Mejer et al. Table 2 (continued ) Events per 1000 person-years of follow up SAB patients
Other risk factors No Yes
Controls
0e30 days
31e180 days
181e365 days
0e365 days
48.2(36.3e64.0) (n Z 48) 69.2 (n Z 3)
9.9(7.2e13.5) (n Z 38) 17.0 (n Z 3)
8.5(6.1e11.9) (n Z 35) 10.5 (n Z 2)
11.6(11.0e12.2) (n Z 1663) 15.1(10.3e22.2) (n Z 26)
a
ATE: Arterial thromboembolic events. SAB: Staphylococcus aureus bacteremia. a Obesity, Human immunodeficiency virus, cocaine abuse.
endocarditis were excluded, the aRR of other ATE within 30, 31e180 and 181e365 days after SAB was 15.5(95% CI: 6.9e35, p Z 0.0001), 4.4(95% CI: 2.4e8.0, p Z 0.0001) and 2.5(95% CI: 1.2e5.4, p Z 0.02), respectively, compared to controls.
Risk factors among individuals with SAB In adjusted analysis, the 30-day risk of any ATE was associated with increasing age and a prior diagnosis of ATE (Table 5). In multivariate analysis restricted to SAB patients with prior ATE, no other significant risk factors were identified (data not shown), whereas for SAB patients without prior ATE, the 30-day risk of ATE increased with age and endocarditis (aRR: 2.3 (95% CI: 1.1e5.1, p Z 0.04)). The 30-day risk of MI increased with age, prior ATE and hypertension. For stroke, the 30-day risk was increased with increasing age, prior ATE and endocarditis (Table 5). For other ATE, the 30 day risk increased only with atrial flutter/fibrillation (aRR: 4.6 (95% CI: 1.2e17.6, p Z 0.03)). By one year, increasing age, prior ATE, hypertension (p Z 0.0001), atrial flutter/fibrillation (p Z 0.04) and endocarditis (p Z 0.02) were associated with an increased risk of any ATE (Table 6). In stratified analysis, increasing age, prior ATE, hypertension and renal disease but not endocarditis was associated with an increased risk of MI (Table 6). The 1-year risk of stroke was associated with increasing age, prior ATE and endocarditis (Table 6). For other ATE, the 1-year risk was only increased for atrial flutter/fibrillation (aRR: 2.8 (95% CI: 1.2e6.3, p Z 0.02)).
Discussion Here we demonstrated that SAB was associated with a more than three-fold overall increased short-term risk of any ATE compared to population controls. Other ATE representing a variety of manifestations were associated with the highest risk (15-fold), followed by stroke (5-fold) and MI (2-fold). The increased risk of incident stroke and other ATE persisted beyond 30 days. To our knowledge, this is the first large scale cohort study to evaluate the risk and risk factors of any ATE after systemic S. aureus infection. The study suggests that individuals with SAB may benefit from short-term strategies to prevent ATE.
Several studies have evaluated the role of influenza and pneumonia as a triggering factor for incident ATE but few studies have investigated the role of bacteremia as a triggering event.5e8 Many have limited their study to either MI or stroke. In a self-controlled case-series study of 588 cases, Corrales-Medina et al. found an increased risk of MI immediately after SAB.4 In a study of severe sepsis by Levine et al., rates of MI and stroke were around 1% at day 28 and were highest within the first 10 days.21 A recent Danish study of community-acquired bacteremia of any cause showed a similarly increased risk of stroke and MI within 30 days of bacteremia.3 Their study identified 321 cases of community-acquired SAB with 7 (2.1%) cases of MI and 13 (4%) cases of stroke within 30 days. Our study found significantly fewer incident cases of MI and stroke (approximately 0.3% of each). We believe that differences in case ascertainment may in part explain these differences. In the absence of prospective case ascertainment, we relied on administrative data to capture events and some outcome events may have represented previous ATEs. To account for this, we reviewed discharge summaries of all SAB patients with ATE and discovered that a significant proportion of ATE registered within a year from SAB represented thrombotic events that had occurred prior to SAB and for this reason these SAB patients were excluded from our study. On the other hand, our use of administrative data may not have captured all outcome events although a MI or stroke is a serious event that we believe would be coded as such had it occurred. However, we may have underestimated the true incidence of ATE following SAB. Regardless, our study and the study by Dalager-Pedersen et al. have suggested an increased risk of MI and stroke within 30 days of SAB that persisted for stroke. Our findings are further supported by several studies of MI following pneumonia6,22e24 and of stroke following a number of infections10 that each showed an increased short-term risk of MI and stroke. We found that several types of ATE other than MI and stroke carried the highest risk of incident ATE following SAB. These included occlusion of arteries in the retina, kidney, and intestine that have not been considered previously. Although these other ATE contributed just 17% of all ATE, the relative risk was nearly 15-fold higher than among population controls within 30 days of SAB and the risk persisted throughout the study period of one year. Bacteremia is characterized by a state of inflammation and procoagulation with endothelial damage, vascular
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
Arterial thromboembolism in S. aureus bacteremia
7
Table 3 Incidence rates of stroke among 15,669 cases of Staphylococcus aureus bacteremia and 156,690 controls in Denmark from 1995 through 2008. Events per 1000 person years of follow up SAB patients
All Age, years 16e35 36e55 56e75 >75 Sex Female Male Origin Hospital- associated Community-acquired Unknown acquisition Prior ATE No Yes Prior hypertension No Yes Prior Renal disease No Yes Prior Atrial flutter/fibrillation No Yes Prior Congestive heart failure No Yes Prior Diabetes No Yes
Controls
0e30 days
31e180 days
181e365 days
0e365 days
41.2(30.6e55.6) (n Z 43)
9.4(6.8e12.9) (n Z 38)
6.2(4.3e9.1) (n Z 27)
5.2(4.8e5.6) (n Z 757)
0.0 (n Z 0) 39.6(20.6e76.0) (n Z 9) 24.4(13.5e44.0) (n Z 11) 83.0(55.2e124.9) (n Z 23)
2.4 (n Z 1) 6.1 (n Z 6) 12.0(7.8e18.4) (n Z 21) 11.0(5.9e20.5) (n Z 10)
6.2 (n Z 3) 3.6 (n Z 4) 6.5(3.7e11.4) (n Z 12) 8.8(4.4e17.7) (n Z 8)
0.1 (n Z 1) 1.2(0.8e1.6) (n Z 36) 4.6(4.1e5.2) (n Z 290) 10.4(9.5e11.5) (n Z 430)
52.1(34.0e79.9) (n Z 21) 34.4(22.6e52.2) (n Z 22)
9.8(5.9e16.2) (n Z 15) 9.1(6.1e13.7) (n Z 23)
7.3(4.2e12.9) (n Z 12) 5.6(3.3e9.2) (n Z 15)
4.2(3.7e4.8) (n Z 244) 5.9(5.4e6.4) (n Z 513)
37.0(24.4e56.2) (n Z 22) 55.3(34.3e88.9) (n Z 17) 28.5 (n Z 4)
8.3(5.3e13.0) (n Z 19) 9.8(5.6e17.3) (n Z 12) 12.8(6.1e27.0)’ (n Z 7)
4.5(2.5e8.2) (n Z 11) 8.9(5.0e15.6) (n Z 12) 7.2 (n Z 4)
e
32.1(22.2e46.5) (n Z 28) 87.7(52.9e145.5) (n Z 15)
7.9(5.4e11.5) (n Z 27) 17.5(9.7e31.6) (n Z 11)
6.2(4.2e9.4) (n Z 23) 6.1 (n Z 4)
4.1(3.7e4.4) (n Z 535) 16.4(14.4e18.7) (n Z 222)
39.4(28.0e55.4) (n Z 33) 49.0(26.3e91.0) (n Z 10)
7.0(4.7e10.6) (n Z 23) 18.9(11.4e31.3) (n Z 15)
6.0(3.9e9.2) (n Z 21) 7.1 (n Z 6)
4.8(4.4e5.2) (n Z 641) 10.3(8.6e12.3) (n Z 116)
45.1(33.4e61.1) (n Z 42) 8.9 (n Z 1)
8.6(6.0e12.2) (n Z 31) 15.4(7.3e32.3) (n Z 7)
6.5(4.4e9.6) (n Z 25) 4.2 (n Z 2)
5.2(4.8e5.6) (n Z 752) 8.0 (n Z 5)
41.4(30.2e56.6) (n Z 39) 40.0 (n Z 4)
9.2(6.5e12.8) (n Z 34) 11.5 (n Z 4)
5.3(3.4e8.1) (n Z 21) 17.2 (n Z 6)
4.8(4.4e5.2) (n Z 664) 14.2(11.6e17.4) (n Z 93)
41.1(29.9e56.5) (n Z 38) 42.5 (n Z 5)
8.8(6.2e12.4) (n Z 32) 14.9 (n Z 6)
6.6(4.5e9.7) (n Z 26) 2.5 (n Z 1)
5.0(4.6e5.3) (n Z 694) 11.6(9.1e14.9) (n Z 63)
39.7(28.5e55.3) (n Z 35) 49.5(24.7e98.9) (n Z 8)
9.0(6.4e12.8) (n Z 31) 11.2(5.3e23.4) (n Z 7)
6.5(4.4e9.7) (n Z 24) 4.5 (n Z 3)
4.9(4.6e5.3) (n Z 686) 12.0(9.5e15.1) (n Z 71)
e e
(continued on next page)
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
8
N. Mejer et al. Table 3 (continued ) Events per 1000 person years of follow up SAB patients
Other risk factors No
Controls
0e30 days
31e180 days
181e365 days
0e365 days
40.0(29.4e54.6) (n Z 40) 69.3 (n Z 3)
9.0(6.5e12.6) (n Z 35) 17.0 (n Z 3)
6.3(4.3e9.2) (n Z 26) 5.2 (n Z 1)
5.2(4.8e5.5) (n Z 742) 8.7(5.3e14.5) (n Z 15)
a
Yes
ATE: Arterial thromboembolic events. SAB: Staphylococcus aureus bacteremia. a Obesity, Human immunodeficiency virus, cocaine abuse.
Table 4 Incidence rates of other arterial thromboembolism among 15,669 cases of Staphylococcus aureus bacteremia and 156,690 controls in Denmark from 1995 through 2008. Events per 1000 person years of follow up SAB patients
All
Controls
0e30 days
31e180 days
181e365 days
0e365 days
15.3(9.4e25.0) (n Z 16)
4.7(3.0e7.3) (n Z 19)
2.5(1.4e4.6) (n Z 11)
1.1(0.9e1.3) (n Z 156)
leakage, and hypotension,25,26 events that each may induce arterial occlusion. Further, infections may result in atherosclerotic plaque instability and rupture leading to arterial thrombosis.5 In this study the long term risk of stroke and other ATE was elevated compared to controls, but was lower for MI. A possible explanation for this observation may be that the short-term risk of any ATE was determined by plaque instability and rupture induced by the bacteremic event whereas the persistent increased risk of stroke and other ATE was determined by complicating events in the form of new-onset atrial fibrillation and endocarditis. Certainly, new-onset atrial fibrillation is frequently associated with sepsis and may lead to thromboembolic events12 and in our study prior atrial flutter/fibrillation significantly increased the 1-year risk of any ATE. Endocarditis may, of course, be a source of septic embolism. In our study, 762 (5%) of SAB patients had endocarditis based on discharge
reports, and endocarditis significantly increased the risk of stroke. However, our estimate of endocarditis was likely an underestimation because up to a fifth of all patients with SAB have endocarditis if evaluated by echocardiography.27 This further highlights the importance of identifying the infective focus in patients with SAB because the presence of endocarditis was associated with an increased shortand long-term risk of stroke. Importantly, even when SAB patients with registered endocarditis were excluded there remained an excess increased risk of stroke, other ATE and any ATE compared to controls. Optimal preventive treatment strategies are lacking because the mechanism of cardiovascular complications after infection is unclear.9 Possible strategies may include optimal treatment of bacteremia, anticoagulation during and after bacteremia, or vaccination of individuals at risk. We found that risk factors that may help to designate
Figure 1 Adjusted relative risks of any arterial thromboembolic events (ATE) in SAB patients versus controls in the year following Staphylococcus aureus bacteremia.
Figure 2 Adjusted relative risks of myocardial infarction, stroke or other ATE in SAB patients versus controls in the year following Staphylococcus aureus bacteremia.
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
Arterial thromboembolism in S. aureus bacteremia
9
Table 5 Risk factors associated with any arterial thromboembolism, with myocardial infarction or with stroke within 30-days after Staphylococcus aureus bacteremia.
Age 16e35 36e55 56e75 >75 Sex Female Male Origin Hospital- associated Community-acquired Unknown acquisition Prior ATE No Yes Prior Hypertension No Yes Prior Renal disease No Yes Prior Atrial flutter/fibrillation No Yes Prior Congestive heart failure No Yes Prior Diabetes No Yes Other risk factorsa No Yes Endocarditis No Yes
Any arterial thromboembolism aRR (95% CI)
Myocardial infarction aRR (95% CI)
Stroke aRR (95% CI)
n Z 109
n Z 51
n Z 43
0.1(0.0e0.7) 0.6(0.3e1.0) 0.5(0.3e0.8) 1.0
0.0 0.1(0.0e0.6) 0.8(0.4e1.4) 1.0
0.0 0.6(0.3e1.3) 0.3(0.2e0.7) 1.0
1.0 0.9(0.6e1.3)
1.0 1.1(0.6e2.0)
1.0 0.7(0.4e1.2)
1.0 1.1(0.7e1.7) 0.8(0.4e1.5)
1.0 0.7(0.3e1.3) 0.8(0.3e1.8)
1.0 1.2(0.6e2.4) 0.8(0.3e2.4)
1.0 2.4(1.6e3.6)
1.0 2.5(1.4e4.5)
1.0 2.8(1.4e5.4)
1.0 1.5(1.0e2.3)
1.0 2.0(1.1e3.7)
1.0 1.2(0.6e2.6)
1.0 0.9(0.5e1.7)
1.0 1.4(0.7e2.8)
1.0 0.2(0.0e1.5)
1.0 1.0(0.6e1.7)
1.0 0.8(0.4e1.9)
1.0 0.6(0.2e1.9)
1.0 1.0(0.6e1.7)
1.0 1.2(0.6e2.4)
1.0 0.7(0.2e1.8)
1.0 1.3(0.8e2.1)
1.0 1.6(0.9e3.0)
1.0 1.2(0.6e2.8)
1.0 1.4(0.6e3.2)
1.0 1.1(0.3e3.6)
1.0 2.0(0.6e6.8)
1.0 1.4(0.6e3.0)
1.0 0.5(0.1e3.3)
1.0 3.3(1.3e7.9)
a Obesity, Human immunodeficiency virus, cocaine abuse. ATE: Arterial thromboembolic events. aRR: adjusted relative risk, CI: confidence interval.
high-risk individuals of ATE after bacteremia include old age, prior ATE, atrial flutter/fibrillation, hypertension and endocarditis. The strengths of this study lie in its size, the longitudinal and population-based design, the inclusion based on the exposure to infection rather than the outcome of interest, complete follow-up on both mortality and ATE, the review of discharge reports, and the use of randomly selected age- and sex matched population controls for comparative analysis. Still, there are potential limitations. First, because disease criteria relied on diagnostic codes from the NPR, it is likely that a degree of misclassification may have occurred. However, validation
studies have demonstrated high validity of MI and stroke diagnoses with positive predictive values of 86e97% for MI and 76e95% for stroke, respectively,28e34 while the negative predictive values have been estimated to be 53% for MI and 79% for stroke.31 To further improve the diagnostic accuracy, we performed a review of all discharge notes to ensure that a case of ATE in fact had occurred after a diagnosis of SAB. The validity of other ATE has not been assessed, but these contributed less than 20% of the total number of cases and we performed stratified analysis based on the three different endpoints. Second, information on use of anticoagulants was not available, and we may have underestimated the true risk of ATE following
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
10
N. Mejer et al.
Table 6 Risk factors associated with any arterial thromboembolism, with myocardial infarction or with stroke within a year after Staphylococcus aureus bacteremia.
Age 16e35 36e55 56e75 >75 Sex Female Male Origin Hospital- associated Community-acquired Unknown acquisition Prior ATE No Yes Prior Hypertension No Yes Prior Renal disease No Yes Prior Atrial flutter/fibrillation No Yes Prior Congestive heart failure No Yes Prior Diabetes No Yes Other risk factorsa No Yes Endocarditis No Yes
Any arterial thromboembolism aRR (95% CI)
Myocardial infarction aRR (95% CI)
Stroke aRR (95% CI)
n Z 278
n Z 129
n Z 108
0.2(0.1e0.4) 0.6(0.4e0.9) 0.8(0.6e1.0) 1.0
0.0 0.3(0.2e0.7) 0.8(0.6e1.2) 1.0
0.2(0.1e0.7) 0.5(0.3e0.9) 0.6(0.4e1.0) 1.0
1.0 0.9(0.7e1.2)
1.0 1.1(0.7e1.6)
1.0 0.8(0.5e1.1)
1.0 1.0(0.8e1.3) 1.2(0.8e1.7)
1.0 0.7(0.5e1.1) 1.2(0.8e1.9)
1.0 1.4(0.9e2.1) 1.2(0.7e2.2)
1.0 1.8(1.4e2.3)
1.0 2.0(1.4e2.9)
1.0 1.8(1.2e2.9)
1.0 1.8(1.3e2.3)
1.0 2.6(1.8e3.9)
1.0 1.5(1.0e2.4)
1.0 1.2(0.9e1.8)
1.0 1.5(1.0e2.3)
1.0 0.8(0.4e1.6)
1.0 1.4(1.0e2.0)
1.0 1.4(0.9e2.2)
1.0 1.1(0.6e2.0)
1.0 1.0(0.7e1.4)
1.0 1.0(0.7e1.6)
1.0 0.7(0.4e1.3)
1.0 1.1(0.8e1.5)
1.0 1.4(0.9e2.1)
1.0 0.9(0.5e1.6)
1.0 1.3(0.8e2.2)
1.0 1.1(0.5e2.2)
1.0 1.6(0.7e3.6)
1.0 1.7(1.1e2.7)
1.0 1.3(0.6e2.8)
1.0 3.1(1.8e5.4)
a Obesity, Human immunodeficiency virus, cocaine abuse. ATE: Arterial thromboembolic events. aRR: adjusted relative risk, CI: confidence interval.
SAB. Third, in this study the primary outcome was based on registrations of primary diagnosis codes of ATE within 365 days from SAB. Because we did not use secondary codes we may have underestimated the rate of ATE after SAB. Fourth, due to lack of data we were unable to adjust for several well-known cardiovascular risk factors including smoking, hypercholesterolemia, or a family history of ATE. In conclusion, SAB was associated with an increased risk of MI, stroke and ‘other’ ATE within 30 days of SAB that persisted for stroke and other ATE. Further studies are required to determine the efficacy of prevention and treatment strategies to lower the risk ATE in patients with SAB.
Competing interests A.G. Jensen has been employed at Pfizer Denmark since 2005 and Bristol-Myers Squibb Denmark since 2014 but neither Pfizer Denmark nor Bristol-Myers Squibb Denmark has financial interests in- or support to the present study. All other authors declare that they have no competing interests.
Financial support This work was supported by Preben and Anna Simonsen’s Foundation (021892-0003).
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
Arterial thromboembolism in S. aureus bacteremia
11
Acknowledgments The authors thank the staff at the participating departments of clinical microbiology for their contribution, continuous support and enthusiasm.
16.
17.
Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jinf.2015.03.010
18.
References
19.
1. Feigin VL, Lawes CM, Bennett DA, Barker-Collo SL, Parag V. Worldwide stroke incidence and early case fatality reported in 56 population-based studies: a systematic review. Lancet Neurol 2009 Apr;8(4):355e69. 2. Yeh RW, Sidney S, Chandra M, Sorel M, Selby JV, Go AS. Population trends in the incidence and outcomes of acute myocardial infarction. N Engl J Med 2010 Jun 10;362(23):2155e65. 3. Dalager-Pedersen M, Sogaard M, Schonheyder HC, Nielsen H, Wernich TR. Risk for myocardial infarction and stroke after community-acquired bacteremia: a 20-Year Population-Based Cohort Study. Circulation 2014 Apr 1;129(13):1387e96. 4. Corrales-Medina VF, Fatemi O, Serpa J, Valayam J, Bozkurt B, Madjid M, et al. The association between Staphylococcus aureus bacteremia and acute myocardial infarction. Scand J Infect Dis 2009;41(6e7):511e4. 5. Corrales-Medina VF, Madjid M, Musher DM. Role of acute infection in triggering acute coronary syndromes. Lancet Infect Dis 2010 Feb;10(2):83e92. 6. Corrales-Medina VF, Musher DM, Shachkina S, Chirinos JA. Acute pneumonia and the cardiovascular system. Lancet 2013 Feb 9;381(9865):496e505. 7. Smeeth L, Thomas SL, Hall AJ, Hubbard R, Farrington P, Vallance P. Risk of myocardial infarction and stroke after acute infection or vaccination. N Engl J Med 2004 Dec 16;351(25): 2611e8. 8. Warren-Gash C, Smeeth L, Hayward AC. Influenza as a trigger for acute myocardial infarction or death from cardiovascular disease: a systematic review. Lancet Infect Dis 2009 Oct; 9(10):601e10. 9. Walkey AJ. Preventing cardiovascular complications of acute infection: a missed opportunity? Circulation 2014 Apr 1; 129(13):1375e7. 10. Guiraud V, Amor MB, Mas JL, Touze E. Triggers of ischemic stroke: a systematic review. Stroke 2010 Nov;41(11):2669e77. 11. Paganini-Hill A, Lozano E, Fischberg G, Perez BM, Rajamani K, Ameriso SF, et al. Infection and risk of ischemic stroke: differences among stroke subtypes. Stroke 2003 Feb;34(2):452e7. 12. Lee-Iannotti JK, Capampangan DJ, Hoffman-Snyder C, Wellik KE, Patel B, Tondato F, et al. New-onset atrial fibrillation in severe sepsis and risk of stroke and death: a critically appraised topic. Neurologist 2012 Jul;18(4):239e43. 13. Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology of sepsis in the United States from 1979 through 2000. N Engl J Med 2003 Apr 17;348(16):1546e54. 14. Mejer N, Westh H, Schonheyder HC, Jensen AG, Larsen AR, Skov R, et al. Stable incidence and continued improvement in short term mortality of Staphylococcus aureus bacteraemia between 1995 and 2008. BMC Infect Dis 2012 Oct 17;12(1):260. 15. Sogaard M, Norgaard M, Dethlefsen C, Schonheyder HC. Temporal changes in the incidence and 30-day mortality associated
20.
21.
22.
23.
24.
25. 26. 27.
28.
29.
30.
31.
32.
with bacteremia in hospitalized patients from 1992 through 2006: a population-based cohort study. Clin Infect Dis 2011 Jan 1;52(1):61e9. Pedersen CB, Gotzsche H, Moller JO, Mortensen PB. The Danish Civil Registration System. A cohort of eight million persons. Dan Med Bull 2006 Nov;53(4):441e9. Jessen O, Rosendal K, Bulow P, Faber V, Eriksen KR. Changing staphylococci and staphylococcal infections. A ten-year study of bacteria and cases of bacteremia. N Engl J Med 1969 Sep 18;281(12):627e35. Wacholder S, McLaughlin JK, Silverman DT, Mandel JS. Selection of controls in case-control studies. I. Principles. Am J Epidemiol 1992 May 1;135(9):1019e28. Andersen TF, Madsen M, Jorgensen J, Mellemkjær L, Olsen JH. The Danish National Hospital Register. A valuable source of data for modern health sciences. Dan Med Bull 1999 Jun; 46(3):263e8. Benfield T, Espersen F, Frimodt-Moller N, Jensen AG, Larsen AR, Pallesen LV, et al. Increasing incidence but decreasing in-hospital mortality of adult Staphylococcus aureus bacteraemia between 1981 and 2000. Clin Microbiol Infect 2007 Mar;13(3):257e63. Levine RL, LeClerc JR, Bailey JE, Monberg MJ, Sarwat S. Venous and arterial thromboembolism in severe sepsis. Thromb Haemost 2008 May;99(5):892e8. Corrales-Medina VF, Musher DM, Wells GA, Chirinos JA, Chen L, Fine MJ. Cardiac complications in patients with communityacquired pneumonia: incidence, timing, risk factors, and association with short-term mortality. Circulation 2012 Feb 14; 125(6):773e81. Musher DM, Rueda AM, Kaka AS, Mapara SM. The association between pneumococcal pneumonia and acute cardiac events. Clin Infect Dis 2007 Jul 15;45(2):158e65. Corrales-Medina VF, Suh KN, Rose G, Chirinos JA, Doucette S, Cameron DW, et al. Cardiac complications in patients with community-acquired pneumonia: a systematic review and meta-analysis of observational studies. PLoS Med 2011 Jun; 8(6):e1001048. Dellinger RP. Inflammation and coagulation: implications for the septic patient. Clin Infect Dis 2003 May 15;36(10):1259e65. Libby P, Simon DI. Inflammation and thrombosis: the clot thickens. Circulation 2001 Apr 3;103(13):1718e20. Rasmussen RV, Host U, Arpi M, Hassager C, Johansen HK, Korup E, et al. Prevalence of infective endocarditis in patients with Staphylococcus aureus bacteraemia: the value of screening with echocardiography. Eur J Echocardiogr 2011 Jun;12(6):414e20. Petersen LA, Wright S, Normand SL, Daley J. Positive predictive value of the diagnosis of acute myocardial infarction in an administrative database. J Gen Intern Med 1999 Sep;14(9):555e8. Kiyota Y, Schneeweiss S, Glynn RJ, Cannuscio CC, Avorn J, Solomon DH. Accuracy of medicare claims-based diagnosis of acute myocardial infarction: estimating positive predictive value on the basis of review of hospital records. Am Heart J 2004 Jul;148(1):99e104. Rinaldi R, Vignatelli L, Galeotti M, Azzimondi G, de CP. Accuracy of ICD-9 codes in identifying ischemic stroke in the General Hospital of Lugo di Romagna (Italy). Neurol Sci 2003 Jun; 24(2):65e9. Birman-Deych E, Waterman AD, Yan Y, Nilasena DS, Radford MJ, Gage BF. Accuracy of ICD-9-CM codes for identifying cardiovascular and stroke risk factors. Med Care 2005 May;43(5):480e5. Tolonen H, Salomaa V, Torppa J, Sivenius J, Immonen-Raiha P, Lehtonen A. The validation of the Finnish Hospital Discharge
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
12 Register and Causes of Death Register data on stroke diagnoses. Eur J Cardiovasc Prev Rehabil 2007 Jun;14(3):380e5. 33. Rosamond WD, Chambless LE, Sorlie PD, Bell EM, Weitzman S, Smith JC, et al. Trends in the sensitivity, positive predictive value, false-positive rate, and comparability ratio of hospital discharge diagnosis codes for acute myocardial infarction in
N. Mejer et al. four US communities, 1987-2000. Am J Epidemiol 2004 Dec 15;160(12):1137e46. 34. Cutrona SL, Toh S, Iyer A, Foy S, Daniel GW, Nair VP, et al. Validation of acute myocardial infarction in the Food and Drug Administration’s Mini-Sentinel program. Pharmacoepidemiol Drug Saf 2013 Jan;22(1):40e54.
Please cite this article in press as: Mejer N, et al., Increased risk of arterial thromboembolic events after Staphylococcus aureus bacteremia: A matched cohort study, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.03.010
