- Email: [email protected]
Surgical outcomes of infective endocarditis among intravenous drug users
Accepted Manuscript Surgical Outcomes of Infective Endocarditis among Intravenous Drug Users Joon Bum Kim, MD, PhD, Julius I. Ejiofor, MD, Maroun Yammine, MD, Masahiko Ando, MD, Janice M. Camuso, RN, Ilan Youngster, MD, Sandra B. Nelson, MD, Arthur Y. Kim, MD, Serguei I. Melnitchouk, MD, James D. Rawn, MD, Thomas E. MacGillivray, MD, Lawrence H. Cohn, MD, John G. Byrne, MD, Thoralf M. Sundt, MD PII:
S0022-5223(16)00432-3
DOI:
10.1016/j.jtcvs.2016.02.072
Reference:
YMTC 10421
To appear in:
The Journal of Thoracic and Cardiovascular Surgery
Received Date: 12 October 2015 Revised Date:
24 January 2016
Accepted Date: 14 February 2016
Please cite this article as: Kim JB, Ejiofor JI, Yammine M, Ando M, Camuso JM, Youngster I, Nelson SB, Kim AY, Melnitchouk SI, Rawn JD, MacGillivray TE, Cohn LH, Byrne JG, Sundt TM, Surgical Outcomes of Infective Endocarditis among Intravenous Drug Users, The Journal of Thoracic and Cardiovascular Surgery (2016), doi: 10.1016/j.jtcvs.2016.02.072. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
1
Surgical Outcomes of Infective Endocarditis among Intravenous Drug Users
2 Joon Bum Kim, MD, PhD1,4*; Julius I. Ejiofor,MD2*; Maroun Yammine, MD2; Masahiko Ando, MD1
4
Janice M. Camuso, RN1; Ilan Youngster, MD3; Sandra B. Nelson, MD3; Arthur Y. Kim, MD3; Serguei I.
5
Melnitchouk, MD1; James D. Rawn, MD2; Thomas E. MacGillivray, MD1; Lawrence H. Cohn, MD2;
6
John G. Byrne, MD2 and Thoralf M. Sundt, MD1.
7
*Drs. Kim and Ejiofor contributed equally as the first authors to this work.
SC
RI PT
3
9
M AN U
8 Institutional affiliation:
10
1
Division of Cardiac Surgery and 3Division of Infectious Disease, Massachusetts General Hospital,
11
Harvard Medical School, Boston, Massachusetts, USA;
12
2
13
Massachusetts, USA;
14
4
15
of Medicine, Seoul, South Korea
TE D
Division of Cardiac Surgery, Brigham and Women’s Hospital, Harvard Medical School, Boston,
EP
Department of Thoracic and Cardiovascular Surgery, Asan Medical Center, University of Ulsan College
AC C
16 17
Corresponding author:
18
Thoralf M. Sundt III
19
Address: Massachusetts General Hospital, Cox 652, 55 Fruit St, Boston, MA 02114
20
Tel: 617-643-9745; Fax: 617-726-5804 ; E-mail: [email protected]
21 1
ACCEPTED MANUSCRIPT
22
Presented at the 95th Annual Meeting of The American Association for Thoracic Surgery, Seattle, WA,
23
April 25-29, 2015.
24 Total word count: 3403
26 Conflict of interest:
28
Dr. Thoralf M Sundt is a consultant for Thrasos Therapeutics
SC
27
RI PT
25
29 Funding source:
31
Dr. Joon Bum Kim was sponsored by AATS Graham Memorial Fellowship for this research
M AN U
30
AC C
EP
TE D
32
2
34
BWH = Brigham and Women’s Hospital
35
CI = confidence interval
36
HR = Hazard ratio
37
IE = infective endocarditis
38
IVDU = intravenous drug user
39
IV = intravenous
40
IQR = interquartile range
41
MGH = Massachusetts General Hospital
42
STS = Society for Thoracic Surgeons
SC
Abbreviations:
M AN U
33
RI PT
ACCEPTED MANUSCRIPT
EP AC C
44
TE D
43
3
ACCEPTED MANUSCRIPT
Abstracts
46
Background: With increasing prevalence of injection drug use in the United States, a growing number of
47
intravenous drug users (IVDU) are at risk for infective endocarditis (IE) that may require surgical
48
intervention, however, little data exist about clinical outcomes of these individuals.
49
Methods: We evaluated consecutive adult patients undergoing surgery for active IE between 2002 and
50
2014 pooled from two prospective institutional databases. Death and valve-related events including
51
reinfection or heart valve reoperation, thromboembolism, and anticoagulation-related hemorrhage were
52
evaluated.
53
Results: Of the 436 patients identified, 78 (17.9%) were current IVDU. The proportion of IVDU
54
increased from 14.8% in 2002-2004 to 26.1% in 2012-2014. IVDU were younger (35.9±9.9 vs.
55
59.3±14.1yrs) and had fewer cardiovascular risk factors than non- IVDU. During follow-up (median, 29.4
56
mo; Quartile 1-3, 4.7-72.6 mo), adverse events among all patients included death in 92, reinfection in 42,
57
valve-reoperation in 35, thromboembolism in 17 and hemorrhage in 16. Operative mortality was lower
58
among IVDU (OR, 0.25; 95% CI, 0.06-0.71), but overall mortality was not significantly different (HR,
59
0.78; 95% CI, 0.44-1.37). When baseline profiles were adjusted by propensity score, IVDU had higher
60
risk of valve-related complications (HR, 3.82; 95% CI, 1.95-7.49), P<0.001) principally attributable to
61
higher rates of reinfection (HR, 6.20, 95% CI, 2.56-15.00, P<0.001).
62
Conclusions: The proportion of IVDU among surgically treated IE patients is increasing. Although
63
IVDU have lower operative risk, long-term outcomes are compromised by reinfection.
SC
M AN U
TE D
EP
AC C
64
RI PT
45
65
Abstract word count: 230
66
Keywords: Intravenous drug use, infective endocarditis, valvular heart disease, operative risk
4
ACCEPTED MANUSCRIPT
Ultramini abstract
68
Among 436 adult patients with active endocarditis undergoing surgery, 78 (17.9%) were current
69
intravenous drug users. Although intravenous drug users had lower operative risk, reinfection (HR, 6.20,
70
95% CI, 2.56-15.00) and valve-related complications (HR, 3.82; 95% CI, 1.95-7.49) were significantly
71
higher compared with non-drug users.
RI PT
67
72 (Word count: 45)
SC
73
AC C
EP
TE D
M AN U
74
5
ACCEPTED MANUSCRIPT
1. Introduction
76
Intravenous drug abuse is increasing rapidly in the United States. Use of heroin, the most common illicit
77
intravenous (IV) drug worldwide, almost doubled in the United States between 2006 and 2013 to 681,000
78
active users, with an estimated 169,000 people starting the drug in 2013.1 The majority of intravenous
79
drug users (IVDU) are young adults who find heroin less expensive and easier to obtain than prescription
80
opioids.1-3 The incidence of heroin-related deaths has surged during this same period as well, with nearly
81
6000 deaths recognized to be caused by heroin intoxication in 2013, a figure almost tripled from that of
82
2006.4 Consequently, this has become a major public health issue in the United States.
83
Of the many medical complications of drug abuse, infective endocarditis (IE) is particularly challenging
84
given the significant risk of acute mortality as well as social factors such as late recidivism and
85
reinfection. Overall, IE carries in-hospital and 1–year mortality rates of 10-20% and a nearly 40%,
86
respectively, even in the current era.5,6 Of those with IE, 60-70% will ultimately require heart valve
87
surgery, particularly those with resistant or highly virulent organisms as is common among IVDU-
88
associated IE.5-8 The decision to proceed with surgical treatment is further complicated among IVDU, as
89
reinfection of a prosthetic valve increases subsequent operative risk and diminishes the likelihood of
90
medical cure.9-12 Given an estimated incidence of IE among this population of 100–400 per 100,000
91
person-years, which is 50-100 times that of the population at large, and the increased prevalence of IDU
92
in the U.S., this is likely to be an increasingly common clinical conundrum.6, 7, 13-15
93
Enthusiasm for operative management of IE in IVDU in the surgical community is mixed, given
94
anecdotal experience of repeated recurrences attributable to continued use of IV drugs.16 Furthermore,
95
independent of reinfection, survival of IVDU is threatened by competing causes of mortality including
96
drug intoxication, trauma and suicide.8, 17-19 These issues raise questions about the utility versus futility of
97
aggressive, resource-intensive surgical procedures particularly among patients who are actively injecting
98
drugs.16,20 There are, however, limited data concerning the clinical characteristics and prognosis after
AC C
EP
TE D
M AN U
SC
RI PT
75
6
ACCEPTED MANUSCRIPT
surgical therapy of these patients to inform clinical decision making and postoperative care.17,36 We
100
therefore sought to examine short- and long-term outcomes in the surgical management of IE among
101
IVDU patients in comparison to non-IVDU associated IE. In the interest of capturing an adequate sized
102
cohort, data were pooled from two closely associated tertiary academic centers in Boston, Massachusetts.
RI PT
99
103
2. Methods
105
-Study Population and Outcome Measures
106
The prospectively collected cardiac surgery databases of the Massachusetts General Hospital (MGH) and
107
the Brigham and Women’s Hospital (BWH) were queried to identify patients aged 17 years or older
108
undergoing heart valve operations due to active IE from January 2002 through August 2014. The disease
109
was considered active according to the Society of Thoracic Surgeons (STS) definitions if the patient was
110
currently being treated for endocarditis.21 These two databases uniformly record baseline patient
111
characteristics, information on cardiac surgical procedures and perioperative outcomes according to STS
112
definitions. In addition, retrospective chart reviews were conducted to collect detailed information
113
regarding IVDU status and profiles on IE such as specific pathogens, characteristics of valvular lesions
114
and treatment details. IVDU currently engaging in IDU was defined as those who had injected in the past
115
12 months as suggested by prior studies.1,4,35
116
Follow-up information on individual patients was obtained through January 2015 by review of
117
longitudinal data from Partners Health Care system, the largest health care system in Massachusetts,
118
which maintains a centralized clinical data registry of all patient encounters within the system including
119
BWH and MGH.19 This data warehouse has been used for other clinical research and its coding data were
120
validated in several prior studies.23,24
AC C
EP
TE D
M AN U
SC
104
7
ACCEPTED MANUSCRIPT
The primary outcome of interest was death and valve-related complications, the latter including
122
reinfection or reoperation of the heart valves, thromboembolic events, and anticoagulation-related
123
hemorrhage as defined by STS.25 Early mortality was defined as those occurring in-hospital or within 30
124
days of surgery. The diagnosis of valvular reinfection was based on one of the following criteria as
125
suggested by STS guidelines: (1) findings on reoperation with evidence of abscess, paravalvular leak,
126
frank pus, or vegetation confirmed as secondary to infection by histologic or bacteriologic studies; or (2)
127
meeting of the Duke Criteria for endocarditis in the absence of reoperation.25 Thromboembolic
128
complications including neurological (stroke or transient ischemic attack) or peripheral embolic events
129
producing signs or symptoms caused by the obstruction of a peripheral artery were diagnosed clinically.
130
Diagnosis of septic pulmonary embolism was made by the findings on enhanced computed tomography
131
of the chest when clinically suspected particularly in cases with IE involving right-sided heart valve.
132
Anticoagulation-related hemorrhage was defined as requiring transfusion, unplanned hospital admission,
133
or a hemostatic intervention.25 Early-, intermediate- and late-onset of recurrent IE were defined as those
134
occurred within 30 days, between 30-and 360 days, and beyond 360 days of surgery based on the current
135
practice guidelines.7
136
The study protocol was approved by the Institutional Review Board of Partners Healthcare, which
137
encompasses the MGH and the BWH. The requirement for informed consent from individual patients was
138
waved due to the retrospective nature of the study design.
139
-Statistical Analysis
140
Categorical variables, presented as frequencies and percentages, were compared using the χ2 test or
141
Fisher’s exact test. Continuous variables, expressed as mean±SD or median with range (or Quartile 1-3),
142
were compared using Student’s unpaired t test or the Mann–Whitney U test. Annual trends in proportions
143
of IVDU among surgical patients with infective endocarditis were tested by Chi-square test for linearity.
144
Kaplan–Meier curves were used to plot the survival and freedom from adverse events. Log-rank tests
AC C
EP
TE D
M AN U
SC
RI PT
121
8
ACCEPTED MANUSCRIPT
were used to compare the differences in the time-related outcomes between groups. Given the significant
146
differences in demographic and clinical risk profiles between the two groups of patients in favor of IVDU
147
group, we assumed that relative risks of deaths may differ across early and late postoperative periods. To
148
address this issue, landmark mortality analyses were performed split into two intervals: 90 days or less
149
and more than 90 days following surgery. 26 To adjust for baseline risk profiles between IVDU and non-
150
IVDU in the evaluations of outcomes, we performed propensity score analysis. The propensity scores
151
were calculated by multiple logistic regression analysis that included baseline variables shown in Table 1
152
and 2. Model discrimination and calibration were evaluated with C statistics (C=0.936) and Hosmer-
153
Lemeshow statistics (P=0.275), respectively. Propensity score matching yielded only 33 pairs of patients,
154
therefore we conducted a covariate adjustment with propensity score rather than using the propensity
155
score as a matching tool. With this technique, propensity scores were incorporated into Cox proportional
156
hazard models as a covariate as well as IVDU status to calculate the propensity-adjusted hazard ratio
157
(HR). Cox-proportional hazard models were used with robust standard errors that accounted for the
158
clustering of center. The proportional hazards assumption was confirmed by examination of log (-
159
log[survival]) curves and by testing of partial (Schoenfeld) residuals and no relevant violations were
160
found.
161
Since defining IVDU depended heavily on patients’ own reports, we assumed that there was potential
162
uncertainty in classifying IVDU status. In order to account for potential impacts of underestimation of
163
current IVDU, a sensitivity analysis was performed. For this analysis, we randomly sampled 3% of
164
patients in the non-IVDU to reclassify them as IVDU and recalculated HRs for clinical endpoints. This
165
procedure was repeated for 100 times to obtain averaged HRs and 95% CIs. We gradually increase the
166
proportions of reclassified patients from 3% to 70% of all non-IVDU, and evaluated how this
167
reclassification influenced on overall HRs.
168
A 2-tailed P value <0.05 was considered to be statistically significant. Analyses were performed with the
169
use of R statistical software, version 3.1.2.
AC C
EP
TE D
M AN U
SC
RI PT
145
9
ACCEPTED MANUSCRIPT
170
3. Results
172
-Baseline Characteristics
173
There were 436 patients meeting the enrollment criteria from the two centers (MGH, n=192; BWH,
174
n=244). Overall, 78 patients (17.9%) were IVDU currently engaging in IDU with the proportion
175
significantly increasing during the study period (P=0.048, Figure 1), particularly after 2012 (14.8% to
176
26.1%, P=0.006). Documented drugs of use included heroin only in 33, cocaine only in 18, combined
177
heroin and cocaine in 19, heroin or cocaine combined with other IV drugs in 5, amphetamine in 2 and
178
polysubstance in 1 with others unknown.
179
The baseline profiles of patients according to current IVDU status are summarized in Table 1. Overall,
180
compared with non-IVDU patients, IVDU were younger and had fewer cardiovascular risk factors (e.g.
181
diabetes, hypertension, renal dysfunction, EuroSCORE II). They presented more frequently, however,
182
with current cigarette smoking, embolic events, large vegetations, and right-sided and multiple valvular
183
involvement. There were no significant differences in gender and race/ethnicity. Causative pathogens
184
including the rates of methicillin-resistant Staphylococcus aureus infection were similar (11.5% in IVDU
185
[9/78] and 8.4% in non-IVDU [30/358], P=0.38) between the two groups.
186
Surgical profiles of patients are detailed in Table 2. Overall, 99 patients (22.7%) underwent emergent
187
surgery and 21 (4.8%) had intra-aortic balloon pumps placed before surgery. Valves were repaired in 45
188
(10.3%). Replacements were with mechanical prostheses in 99 (22.7%), bioprostheses in 206 (47.2%) or
189
human allografts (homografts) in 86 (19.7%). There were no significant differences between IVDU and
190
non-IVDU in operative profiles. All patients received at least 4-6 weeks of IV antibiotic therapy
191
regardless of IVDU status
192
-Outcome: Overall Cohort
AC C
EP
TE D
M AN U
SC
RI PT
171
10
ACCEPTED MANUSCRIPT
Thirty-day mortality rates were lower in IVDU than non-injectors (3.8% vs. 13.7%, P=0.012).
194
Late Follow-up data derived from the Partners database was complete in 76.4% of the entire cohort with a
195
median duration of 29.4 months (Quartile 1-3; 570 patient-years). Completeness of follow-up was lower
196
in IVDU than in the non-injector group (73.5% vs. 77.1%), but without statistical significance (P=0.096).
197
Major adverse events for the entire cohort in the form of valve-related complications were known to occur
198
in 69, including recurrence of IE in 42, valve reoperation in 35, thromboembolism in 17 and
199
anticoagulation-related hemorrhage in 16. (Table 3) Kaplan-Meier curves did not show a significant
200
difference in overall survival rates between IVDU and non-injectors; however, freedom from valve-
201
related complications were significantly lower among IVDU, mainly attributed to higher rates of
202
reinfection and reoperation compared with non-injectors (Figure 3).
203
Overall 5- and 10-year-survival rates were 78.9±6.0% and 69.5±8.3% in IVDU, respectively; and
204
76.1±2.6% and 68.7±3.9% in non-injectors, respectively (P=0.39). When adjusted by propensity score,
205
IVDU tended to have lower risk of early mortality (HR, 0.32; 95% CI, 0.08-1.36; P=0.13 for 90-day death)
206
but with higher risk of late mortality (HR, 2.07; 95% CI, 0.78-5.48; P=0.14) compared with non-IVDU,
207
however, they were statistically insignificant.
208
Five-and 10-year-survival rates free of valve-related events were 40.9±7.5% and 6.9±6.2% among IVDU,
209
respectively; and 63.9±3.3% and 50.8±4.9% in non-injectors (P=0.006). After propensity score
210
adjustments, risks of valve related complications were significantly higher in IVDU compared with non-
211
injectors (HR, 3.82; 95% CI, 1.95-7.49; P<0.001), which was mainly stemmed from a higher risk of
212
reinfection and consequent reoperations (Table 3).
213
-Outcomes: IVDU Subgroup
214
Among early IVDU IE survivors (n=75), 28 patients (37.3%) had known reinfection of the valves at a
215
median of 18.1 months (range 0.2-54.3 months) following valve surgery. Among these, 1 (3.6%) occurred
216
within 30 days, 9 (32.1%) between 30 days and 1 year, and 18 (64.3%) more than 1 year after surgery.
AC C
EP
TE D
M AN U
SC
RI PT
193
11
ACCEPTED MANUSCRIPT
Causative organisms of the reinfections were viridans group Streptococci in 6 (21.4%), methicillin-
218
susceptible Staphylococci in 5 (17.9%), methicillin-resistant Staphylococci in 4 (14.3%), Enterococci in 2
219
(7.1%) and fungus in 2 (7.1%) while 9 patients (32.1%) showed culture-negative results. Overall, only 6
220
patients (21.4%) were found to have the same causative organisms as at initial presentation.
221
Of the 28 patients with reinfection, 14 underwent surgery and 14 were treated with antimicrobial therapy
222
only. The reasons for medical therapy in the latter 14 patients were presence of intracranial hemorrhage in
223
3 (31.4%), prohibitive surgical risk due to severity of illness in 2 (14.3%), active in-hospital use of IV
224
drugs in 3 (31.4%) and cardiac arrest during preparation for operation in 1 (7.1%). Another 5 patients
225
(35.7%) were successfully treated medically with complete resolution of IE. Among those treated
226
surgically there were no early surgical deaths although one patient died of intracranial hemorrhage at 46.7
227
months postoperative. The median duration of antimicrobial therapy in the 14 patients who had only
228
medical therapy was 6.0 weeks (range 3-12 weeks) with 3 in-hospital deaths (21.4%) and 3 late deaths
229
(21.4%). Causes of the early deaths were sepsis in 2 and ventricular arrhythmia in 1, and the causes of late
230
deaths were repeated recurrence of IE in 2 and sudden death in 1. The 5-year survival rates among
231
surgical and medical therapy patients for recurrent IE in IVDU were 91.7±8.0% and 38.9±16.7%,
232
respectively (P=0.066, Supplemental Figure 1). Reoperative valve operations were undertaken for reasons
233
other than reinfection in 6 other patients (7.7%) for insufficiency of a repaired valve (n= 2) or structural
234
deterioration of a prosthetic valve (homograft [n=3] or xenograft bioprostheses [n=1]).
235
Overall, 11 late deaths (14.1%) occurred in IVDU. Causes of deaths were recurrence of IE in 6,
236
intracranial hemorrhage in 1 and metastatic lung cancer in 1 while causes could not be specified in 3
237
patients: one died suddenly and two died of unknown causes, one was found dead while actively using
238
heroin.
239
-Sensitivity Analysis
AC C
EP
TE D
M AN U
SC
RI PT
217
12
ACCEPTED MANUSCRIPT
Because of the challenges associated with determining with certainty the IVDU status of all patients in the
241
study, we undertook a sensitivity analysis of the impact of misclassification as illustrated in Figure 3. The
242
forest plots indicate adjusted HRs for valve-reinfection according to the increasing proportion of
243
reclassified patients from non-IVDU to IVDU (3% to 70%). For valve-reinfection, significantly increased
244
risks are observed up to 50%, while the proportion of reclassification that would be required to account
245
for the observed increased risk of reinfection would be at least 60%.
246
4. Discussion
247
Our study confirms an increase in the proportion of IVDU amongst surgical patients with active IE,
248
almost doubling during the study interval. This is consistent both with our clinical impression and the
249
virtual epidemic of increasing IV drug use in the general population as reported by the Center for Disease
250
Control.1 Perhaps more unexpectedly to us, the data also showed a markedly lower operative risk among
251
IVDU than non-IVDU. This is likely in part because the former were typically 25 years younger and had
252
fewer comorbidities compared to the non-injectors. Another potential factor may be surgical selection
253
bias, as we did not capture all patients seen with IE and cannot therefore account for those who were not
254
accepted for surgical intervention. It is possible that a subset of extremely ill individuals were not offered
255
surgery.
256
Despite superior early survival, late outcomes when corrected for age and other factors by propensity
257
score were worse for IVDU than non-IVDU, with higher rates of reinfection and reoperation. This is
258
likely a reflection of the challenge of recidivism among IVDU, although actual data concerning recurrent
259
drug use are hard to come by. Individuals may well not be forthright in responding to this sensitive
260
question when presenting with IE after valve surgery for IVDU-related IE out of concern that this will
261
impact the enthusiasm of clinicians to reoperation once again. The same may be true of family members.
262
Furthermore, unless a drug screen returns positive, the responses cannot be readily verified.
AC C
EP
TE D
M AN U
SC
RI PT
240
13
ACCEPTED MANUSCRIPT
Our findings of a rising incidence of IVDU-associated IE are consistent with a recent Swedish nationwide
264
cohort study of 7603 patients with IE, 990 of whom were treated surgically.6 In that study, the 30-day
265
mortality remained constant at around 10% over the decade of the study indicating that despite general
266
medical progress in the current era, IE remains a highly lethal condition. Overall, in contrast to our study,
267
the absolute mortality at 5 years follow-up was lower for the 355 IVDU than for non-injectors, when
268
compared to age- and gender-matched general population, however, IVDU showed a Standardized
269
Mortality Ratio of 19.1 (95% CI, 13.5–27.1) which was much higher than other subgroups (Standardized
270
Mortality Ratio range, 1.8-6.3). Similarly, Rabkin and colleagues observed markedly lower long-term
271
survival among IVDU than non-IVDU patients and, conversely, less frequent re-infection .17 The reasons
272
for this difference is a matter of speculation, however it is clear that the socioeconomic aspects of IVDU
273
are rapidly changing in this country, with heroin use markedly up among the middle class or wealthy.37
274
This may translate into greater access to healthcare should use recur. It should be noted that this
275
disappointing long-term survival of IVDU has been demonstrated independent of IE. Many IV drug
276
initiators become chronic users, and survival is significantly compromised should they not resolve IDU.
277
27-30
278
population with an average loss of life of 18.3 years.27,29 Although one study indicated IE as the most
279
common cause of death among IVDU,28 other studies showed drug overdose, accident, violence and other
280
chronic disease such as AIDS and liver cirrhosis as competing leading causes of death in these
281
individuals.27,29-31
282
The combination of anticipated short life expectancy despite young chronologic age and the perceived
283
high risks of treatment failure and recurrence of the disease after surgery challenge many front line health
284
care providers as an ethical controversy .16,20 Despite the growing incidence of IDU, current practice
285
guidelines do not address surgical indications or management strategies for active IVDU specifically and
286
treating physicians are left to formulate therapeutic plans on an individual basis without much hard
287
evidence in the literature. The results of this present study help to fill that evidence gap. These data
TE D
M AN U
SC
RI PT
263
AC C
EP
The mortality rate among long-term IVDU has been shown to be 6.9 times greater than the general
14
ACCEPTED MANUSCRIPT
suggest that for IVDU actively injecting, the technical exercise of replacing the infected valve(s)
289
combined with antimicrobial therapy can be performed at low perioperative risk, but may be incomplete
290
therapy without the addition of effective addiction treatment. It is apparent that resolution or minimization
291
of post-operative IDU would likely improve outcomes for this population via prevention of late re-
292
infection just as discontinuation of tobacco use may improve outcomes after coronary revascularization.
293
From a surgical standpoint, failure to resolve IDU is comparable to leaving the nidus of infection behind
294
as shown by a 10-year event-free survival rate of only 6.9±6.2% among IVDU. Given the very high
295
recidivism rate of IDU, insistence by treating clinicians on agreement by the subject to participation in a
296
drug rehabilitation program might be a reasonable consideration. Consultation to institutional ethics
297
committee may also be appropriate when making a decision to undertake surgery in active IVDU who
298
refuse to enter a drug rehabilitation program and are likely to remain at high risk to re-infect their valve.
299
In one of our institutions (MGH) consultation with an addiction counselor has become standard routine at
300
the time of surgical consultation for IE in IVDU.
301
Finally, we observed Staphylococcus aureus as the most common pathogen of IE among IVDU and right-
302
side valve involvement to be more common in IVDU than non-IVDU as has been reported previously.32-34
303
Unlike earlier prior reports we found left-sided valve involvement predominant among IVDU who
304
undergo surgery.34 This is consistent with the findings of Rabkin and colleagues and may be a reflection
305
of a shift in surgical philosophy away from intervention on right sided lesions in favor of antimicrobial
306
therapy and the implantation of a prosthetic makes any subsequent infection an even more complicated
307
problem.17 The poor late outcomes in IVDU could not be attributed to differences in the virulence or type
308
of organism.
310
SC
M AN U
TE D
EP
AC C
309
RI PT
288
-Limitations
15
ACCEPTED MANUSCRIPT
This study has inherent limitations in the retrospective nature of the research. Study subjects are affected
312
by surgical selection bias since only patients accepted for surgery were included and those who received
313
medical therapy only were not, nor was there a medical treatment only control group for comparison.
314
Obtaining reliable follow-up data concerning reinfection and valve-related complications in this patient
315
population was also challenging. As this group is the most challenging in follow-up, the data are likely
316
biased in favor of lower event rates in this subgroup. Finally, the data set represents the experience of
317
urban tertiary academic referral centers, and therefore, it is subject to referral bias from the community.
318
- Conclusions
319
The proportion of IVDU among surgical patients with IE is increasing. Although IVDU are younger, with
320
lower baseline cardiovascular risk burdens, long-term clinical outcomes are discouraging with a
321
significant rate of reinfection and a trend toward higher risk of late mortality. These data highlight the
322
need to aggressively address the expanding drug use epidemic that is driving the increased rate of IE in
323
this population and provide effective addiction treatment post-operatively to prevent re-infection.
M AN U
References
326
1.
AC C
Publication No. (SMA) 14-4863. Rockville, MD: Substance Abuse and Mental Health Services
328
Administration, 2014.
329 2.
Hooten WM, Bruce BK. Beliefs and attitudes about prescribing opioids among healthcare
providers seeking continuing medical education. J Opioid Manag.;7(6):417-424.
331
333
Substance Abuse and Mental Health Services Administration, Results from the 2013 National Survey on Drug Use and Health: Summary of National Findings, NSDUH Series H-48, HHS
327
332
EP
325
TE D
324
330
SC
RI PT
311
3.
Slevin KA, Ashburn MA. Primary care physician opinion survey on FDA opioid risk evaluation and mitigation strategies. J Opioid Manag.;7(2):109-115. 16
ACCEPTED MANUSCRIPT
334
4.
Trends in Drug-poisoning Deaths Involving Opioid Analgesics and Heroin: United States, 1999–
335
2012. Centers for Disease Control and Prevention
336
(http://www.cdc.gov/nchs/data/hestat/drug_poisoning/drug_poisoning.htm). 5.
Chu VH, Park LP, Athan E, et al. Association between surgical indications, operative risk, and
RI PT
337 338
clinical outcome in infective endocarditis: a prospective study from the international
339
collaboration on endocarditis. Circulation. ;131(2):131-140.
and long-term prognosis in infective endocarditis in Sweden. PLoS One.8(7):e67519.
341
342
Ternhag A, Cederstrom A, Torner A, Westling K. A nationwide cohort study of mortality risk
SC
6.
7.
M AN U
340
Nishimura RA, Otto CM, Bonow RO, et al. 2014 AHA/ACC Guideline for the Management of
343
Patients With Valvular Heart Disease: a report of the American College of Cardiology/American
344
Heart Association Task Force on Practice Guidelines. Circulation.;129(23):e521-643.
345
8.
Shih CJ, Chu H, Chao PW, et al. Long-term clinical outcome of major adverse cardiac events in survivors of infective endocarditis: a nationwide population-based study.
347
Circulation.;130(19):1684-1691. 9.
Grubitzsch H, Schaefer A, Melzer C, Wernecke KD, Gabbieri D, Konertz W. Outcome after
EP
348
TE D
346
surgery for prosthetic valve endocarditis and the impact of preoperative treatment. J Thorac
350
Cardiovasc Surg. 2014;148(5):2052-2059.
351
10.
endocarditis. Ann Thorac Surg. 2002;74(5):S1754-1757.
353
355
Lytle BW, Sabik JF, Blackstone EH, Svensson LG, Pettersson GB, Cosgrove DM, 3rd.
Reoperative cryopreserved root and ascending aorta replacement for acute aortic prosthetic valve
352
354
AC C
349
11.
Meszaros K, Nujic S, Sodeck GH, et al. Long-term results after operations for active infective endocarditis in native and prosthetic valves. Ann Thorac Surg. 2012;94(4):1204-1210.
17
ACCEPTED MANUSCRIPT
356
12.
Musci M, Weng Y, Hubler M, et al. Homograft aortic root replacement in native or prosthetic
357
active infective endocarditis: twenty-year single-center experience. J Thorac Cardiovasc Surg.
358
2010;139(3):665-673. 13.
1988-1990. Am J Cardiol. 1995;76(12):933-936.
360
361
Berlin JA, Abrutyn E, Strom BL, et al. Incidence of infective endocarditis in the Delaware Valley,
14.
RI PT
359
Hogevik H, Olaison L, Andersson R, Lindberg J, Alestig K. Epidemiologic aspects of infective endocarditis in an urban population. A 5-year prospective study. Medicine (Baltimore).
363
1995;74(6):324-339. 15.
M AN U
364
SC
362
Spijkerman IJ, van Ameijden EJ, Mientjes GH, Coutinho RA, van den Hoek A. Human
365
immunodeficiency virus infection and other risk factors for skin abscesses and endocarditis
366
among injection drug users. J Clin Epidemiol. 1996;49(10):1149-1154.
367
16.
DiMaio JM, Salerno TA, Bernstein R, Araujo K, Ricci M, Sade RM. Ethical obligation of surgeons to noncompliant patients: can a surgeon refuse to operate on an intravenous drug-
369
abusing patient with recurrent aortic valve prosthesis infection? Ann Thorac Surg. 2009;88(1):1-8. 17.
Rabkin DG, Mokadam NA, Miller DW, Goetz RR, Verrier ED, Aldea GS. Long-term outcome
EP
370
TE D
368
for the surgical treatment of infective endocarditis with a focus on intravenous drug users. Ann
372
Thorac Surg. 2012;93(1):51-57.
373
18.
Infect Dis. 2007;39(3):197-204.
375
377
Thalme A, Westling K, Julander I. In-hospital and long-term mortality in infective endocarditis in
injecting drug users compared to non-drug users: a retrospective study of 192 episodes. Scand J
374
376
AC C
371
19.
Hubbell G, Cheitlin MD, Rapaport E. Presentation, management, and follow-up evaluation of infective endocarditis in drug addicts. Am Heart J. 1981;102(1):85-94.
18
ACCEPTED MANUSCRIPT
378
20.
Hull SC, Jadbabaie F. When is enough enough? The dilemma of valve replacement in a recidivist intravenous drug user. Ann Thorac Surg. ;97(5):1486-1487.
379
21.
www.riskcalc.sts.org
381
22.
Murphy SN, Chueh HC. A security architecture for query tools used to access large biomedical databases. Proc AMIA Symp. 2002:552-556.
23.
history of asymptomatic severe carotid artery stenosis. J Vasc Surg. 2014;60(5):1218-1225. 24.
Horkan CM, Purtle SW, Mendu ML, Moromizato T, Gibbons FK, Christopher KB. The
M AN U
384
385
Conrad MF, Michalczyk MJ, Opalacz A, Patel VI, LaMuraglia GM, Cambria RP. The natural
SC
382
383
RI PT
380
386
association of acute kidney injury in the critically ill and postdischarge outcomes: a cohort study*.
387
Crit Care Med. 2015;43(2):354-364. 25.
26.
27.
Health. 1987;77(3):347-348.
393
394
28.
29.
399
Smyth B, Fan J, Hser YI. Life expectancy and productivity loss among narcotics addicts thirty-
three years after index treatment. J Addict Dis. 2006;25(4):37-47.
397
398
Perucci CA, Forastiere F, Rapiti E, Davoli M, Abeni DD. The impact of intravenous drug use on
mortality of young adults in Rome, Italy. Br J Addict. 1992;87(12):1637-1641.
395
396
Joe GW, Simpson DD. Mortality rates among opioid addicts in a longitudinal study. Am J Public
EP
392
Therneau TM, Grambsch PM. Modeling survival data : extending the Cox model. New York: Springer; 2000.
391
TE D
cardiac valve interventions. J Thorac Cardiovasc Surg. 2008;135(4):732-738.
389
390
Akins CW, Miller DC, Turina MI, et al. Guidelines for reporting mortality and morbidity after
AC C
388
30.
Smyth B, Hoffman V, Fan J, Hser YI. Years of potential life lost among heroin addicts 33 years after treatment. Prev Med. 2007;44(4):369-374. 19
ACCEPTED MANUSCRIPT
31.
Drug Abuse Warning Network 2003 (09/2004 update).
401 32.
drug users: the time has come? Eur J Clin Microbiol Infect Dis. 1994;13(7):533-534.
403
404
DiNubile MJ. Abbreviated therapy for right-sided Staphylococcus aureus endocarditis in injecting
33.
Hecht SR, Berger M. Right-sided endocarditis in intravenous drug users. Prognostic features in 102 episodes. Ann Intern Med. 1992;117(7):560-566.
405
RI PT
402
DAWN 2003: Area profiles of drug-related mortality. Office of Applied Studies, SAMHSA.
SC
400
34.
407
involvement, bacteriologic findings, and outcome of infective endocarditis in intravenous drug users.
408
Arch Intern Med. 1995;155(15):1641-1648.
409
35.
Mathers BM, Degenhardt L, Phillips B, et al. Global epidemiology of injecting drug use and HIV among people who inject drugs: a systematic review. Lancet. 2008;372(9651):1733-1745.
36.
Dawood MY, Cheema FH, Ghoreishi M, et al. Contemporary outcomes of operations for tricuspid
TE D
410
411
Mathew J, Addai T, Anand A, Morrobel A, Maheshwari P, Freels S. Clinical features, site of
M AN U
406
valve infective endocarditis. Ann Thorac Surg. 2015;99(2):539-546.
412
37. http://www.nytimes.com/interactive/2015/10/30/us/31heroin-deaths.html?_r=0 (accessed Jan 17,
414
2016)
AC C
EP
413
20
ACCEPTED MANUSCRIPT
Table 1. Baseline Characteristics between IVDU and Non-IVDU Non-IVDU
(n=78)
(n=358)
Female gender, n (%)
30 (38.5)
Body mass index, kg/m2
26.7±7.9
M AN U
Ethnicity, n (%) White Black Hispanic
TE D
Asian Others
Hypertension
EP
Diabetes mellitus
59.3±14.1
<0.001
111 (31.0)
0.20
27.8±7.0
0.24
SC
35.9±9.9
Cigarette smoking history
Current cigarette smoking
NYHA functional class III or IV, n (%) On dialysis, n (%)
21
P value
RI PT
IVDU
Age, yr
AC C
415
0.83
68 (87.2)
319 (89.1)
5 (6.4)
14 (3.9)
1 (1.3)
9 (2.5)
1 (1.3)
5 (1.4)
3 (3.8)
11 (3.1)
6 (7.6)
74 (20.7)
0.027
16 (20.5)
225 (62.8)
<0.001
53 (67.9)
142 (39.7)
<0.001
28 (35.9)
29 (8.1)
<0.001
35 (44.9)
177 (49.4)
0.46
3 (3.8)
31 (8.7)
0.24
ACCEPTED MANUSCRIPT
Creatinine clearance, mL/min/1.73m2
73.1±44.1
<0.001
Embolic events
36 (46.2)
107 (29.9)
0.006
Cerebral
21 (26.9)
93 (26.0)
0.86
15
Peripheral
RI PT
103.5±53.6
14
0.14
Causative pathogens, n (%) 16 (20.5)
89 (24.9)
SC
Viridans group streptococci
2 (2.6)
M AN U
Other streptococci
33 (9.2)
30 (38.5)
97 (27.1)
9 (11.5)
30 (8.4)
10 (12.8)
65 (18.2)
1 (1.3)
35 (9.8)
12 (15.4)
48 (13.4)
9 (11.5)
24 (6.7)
9 (11.5)
25 (7.0)
57 (73.1)
294 (82.1)
Aortic valve
32 (41.0)
200 (55.9)
Mitral valve
13 (16.7)
86 (24.0)
Tricuspid valve
12 (15.4)
8 (2.2)
Staphylococcus aureus methicillin-resistant
methicillin-resistant Entercococci
AC C
Negative culture
EP
Others
TE D
Staphylococci, coagulase-negative
Single valve affected, n (%)
22
ACCEPTED MANUSCRIPT
64 (17.9)
13 (16.7)
54 (15.1)
Aortic + Tricuspid
3 (3.8)
6 (1.7)
Mitral + Tricuspid
5 (6.4)
Aortic + Mitral
0
Aortic +Mitral + Tricuspid
18 (5.0)
<0.001
Vegetation diameter ≥10mm, n (%)
55 (70.5)
178 (49.7)
<0.001
18 (23.1)
121 (33.8)
0.066
17 (21.8)
110 (30.7)
0.12
Severe dysfunction of affected valves, n (%)
59 (75.6)
245 (68.4)
0.21
Left ventricular ejection fraction, %
59.5±10.4
60.1±10.9
0.64
9.5±8.3
13.4±12.6
0.001
(1.8-42.0)
(1.8-82.9)
M AN U
TE D
AC C
EP
EuroSCORE II, % (range)
SC
20 (25.6)
Prosthetic endocarditis, n (%)
417
2 (0.6) 2 (0.6)
Right-side valve involvement
Abscess formation, n (%)
416
0.068
RI PT
21 (26.9)
Multiple valves affected, n (%)
23
ACCEPTED MANUSCRIPT
Table 2. Surgical Profiles Non- IVDU
(n=78)
(n=358)
Emergent surgery
16 (20.5) 3 (3.8)
On IABP
M AN U
Primary procedure
6 (0-140)
0.098
83 (23.2)
0.61
18 (5.0)
>0.99
SC
7 (0-76)
P value
RI PT
IVDU
Interval from admission to surgery, days
0.87
10 (12.8)
35 (9.8)
18 (23.1)
81 (22.6)
35 (44.9)
171 (47.8)
15 (19.2)
71 (19.8)
2 (2.6)
20 (5.6)
0.27
10 (12.8)
56 (15.6)
0.60
Aortic root replacement
19 (24.4)
96 (26.8)
0.78
Cardiopulmonary bypass time, min
191.3±103.8
215.5±142.0
0.34
Cardiac ischemic time, min
145.9±85.3
159.3±97.8
0.30
Valve repair Valve replacement, mechanical Valve replacement, bioprostheses
Associated procedures
TE D
Valve replacement, allograft
CABG
EP
Ascending Aorta replacement
AC C
418
24
ACCEPTED MANUSCRIPT
419
Abbreviations: IABP, intra-aortic balloon pumping; CABG, coronary artery bypass grafting; ASD, atrial
420
septal defect; PFO, patent foramen ovale; VSD, ventricular septal defect; MAIVF, mitral-aortic inter-
421
valvular fibrosa.
AC C
EP
TE D
M AN U
SC
RI PT
422
25
ACCEPTED MANUSCRIPT
Table 3. Unadjusted and Adjusted Clinical Outcomes According to Current Status of IV Drug Abuses Crude
Propensity-Adjusted
RI PT
No. of Events Non-IVDU
HR
95% CI
P value
HR
95% CI
P value
30-day mortality
1
40
0.10
0.01-0.48
0.026
0.15
0.01-1.01
0.10
90-day mortality
3
49
0.25
0.024
0.32
0.08-1.36
0.13
Late (> 90 days) mortality
10
30
Valve related complications
32
37
Valve Re-infection
28
14
Valve reoperation
18
Thromboembolism
7
Hemorrhage
5
M AN U
0.06-0.71 0.66-2.63
0.45
2.07
0.78-5.48
0.14
4.07
2.53-6.55
0.001
3.82
1.95-7.49
<0.001
9.28
4.87-17.68
<0.001
6.20
2.56-15.00
<0.001
17
5.03
2.58-9.82
0.005
3.66
1.44-9.30
0.006
10
2.92
1.11-7.68
0.030
2.54
0.62-10.47
0.20
11
1.75
0.60-5.05
0.30
4.33
1.03-18.17
0.045
TE D
1.32
EP
AC C
SC
IVDU
Abbreviations: IVDU, intravenous drug users; HR, hazard ratio; CI, confidence interval.
26
ACCEPTED MANUSCRIPT
Figure Legends Figure 1. Trends in proportions of IVDU among surgical patients with infective endocarditis over time.
RI PT
Abbreviation: IVDU, intravenous drug users. Figure 2. Kaplan-Meier curves for clinical outcomes between IVDU and non-IVDU: Overall death (A), valve reinfection (B), valve reoperation (C) and composite of valve –related complications (D).
Transparent bands indicate area within 95% confidence interval. Abbreviation: IVDU, intravenous drug
SC
users.
M AN U
Figure 3. Effects of reclassification of non-IVDU patients to IVDU on hazard ratios for the risk of reinfection. Dots and horizontal lines indicate hazard ratios and 95% confidence interval, respectively. Abbreviation: HR, hazard ratio.
Supplementary Figure 1. Survival after surgical versus medical therapy for recurrence of infective
AC C
EP
TE D
endocarditis in intravenous drug users
27
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
S0022-5223(16)00432-3
DOI:
10.1016/j.jtcvs.2016.02.072
Reference:
YMTC 10421
To appear in:
The Journal of Thoracic and Cardiovascular Surgery
Received Date: 12 October 2015 Revised Date:
24 January 2016
Accepted Date: 14 February 2016
Please cite this article as: Kim JB, Ejiofor JI, Yammine M, Ando M, Camuso JM, Youngster I, Nelson SB, Kim AY, Melnitchouk SI, Rawn JD, MacGillivray TE, Cohn LH, Byrne JG, Sundt TM, Surgical Outcomes of Infective Endocarditis among Intravenous Drug Users, The Journal of Thoracic and Cardiovascular Surgery (2016), doi: 10.1016/j.jtcvs.2016.02.072. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
1
Surgical Outcomes of Infective Endocarditis among Intravenous Drug Users
2 Joon Bum Kim, MD, PhD1,4*; Julius I. Ejiofor,MD2*; Maroun Yammine, MD2; Masahiko Ando, MD1
4
Janice M. Camuso, RN1; Ilan Youngster, MD3; Sandra B. Nelson, MD3; Arthur Y. Kim, MD3; Serguei I.
5
Melnitchouk, MD1; James D. Rawn, MD2; Thomas E. MacGillivray, MD1; Lawrence H. Cohn, MD2;
6
John G. Byrne, MD2 and Thoralf M. Sundt, MD1.
7
*Drs. Kim and Ejiofor contributed equally as the first authors to this work.
SC
RI PT
3
9
M AN U
8 Institutional affiliation:
10
1
Division of Cardiac Surgery and 3Division of Infectious Disease, Massachusetts General Hospital,
11
Harvard Medical School, Boston, Massachusetts, USA;
12
2
13
Massachusetts, USA;
14
4
15
of Medicine, Seoul, South Korea
TE D
Division of Cardiac Surgery, Brigham and Women’s Hospital, Harvard Medical School, Boston,
EP
Department of Thoracic and Cardiovascular Surgery, Asan Medical Center, University of Ulsan College
AC C
16 17
Corresponding author:
18
Thoralf M. Sundt III
19
Address: Massachusetts General Hospital, Cox 652, 55 Fruit St, Boston, MA 02114
20
Tel: 617-643-9745; Fax: 617-726-5804 ; E-mail: [email protected]
21 1
ACCEPTED MANUSCRIPT
22
Presented at the 95th Annual Meeting of The American Association for Thoracic Surgery, Seattle, WA,
23
April 25-29, 2015.
24 Total word count: 3403
26 Conflict of interest:
28
Dr. Thoralf M Sundt is a consultant for Thrasos Therapeutics
SC
27
RI PT
25
29 Funding source:
31
Dr. Joon Bum Kim was sponsored by AATS Graham Memorial Fellowship for this research
M AN U
30
AC C
EP
TE D
32
2
34
BWH = Brigham and Women’s Hospital
35
CI = confidence interval
36
HR = Hazard ratio
37
IE = infective endocarditis
38
IVDU = intravenous drug user
39
IV = intravenous
40
IQR = interquartile range
41
MGH = Massachusetts General Hospital
42
STS = Society for Thoracic Surgeons
SC
Abbreviations:
M AN U
33
RI PT
ACCEPTED MANUSCRIPT
EP AC C
44
TE D
43
3
ACCEPTED MANUSCRIPT
Abstracts
46
Background: With increasing prevalence of injection drug use in the United States, a growing number of
47
intravenous drug users (IVDU) are at risk for infective endocarditis (IE) that may require surgical
48
intervention, however, little data exist about clinical outcomes of these individuals.
49
Methods: We evaluated consecutive adult patients undergoing surgery for active IE between 2002 and
50
2014 pooled from two prospective institutional databases. Death and valve-related events including
51
reinfection or heart valve reoperation, thromboembolism, and anticoagulation-related hemorrhage were
52
evaluated.
53
Results: Of the 436 patients identified, 78 (17.9%) were current IVDU. The proportion of IVDU
54
increased from 14.8% in 2002-2004 to 26.1% in 2012-2014. IVDU were younger (35.9±9.9 vs.
55
59.3±14.1yrs) and had fewer cardiovascular risk factors than non- IVDU. During follow-up (median, 29.4
56
mo; Quartile 1-3, 4.7-72.6 mo), adverse events among all patients included death in 92, reinfection in 42,
57
valve-reoperation in 35, thromboembolism in 17 and hemorrhage in 16. Operative mortality was lower
58
among IVDU (OR, 0.25; 95% CI, 0.06-0.71), but overall mortality was not significantly different (HR,
59
0.78; 95% CI, 0.44-1.37). When baseline profiles were adjusted by propensity score, IVDU had higher
60
risk of valve-related complications (HR, 3.82; 95% CI, 1.95-7.49), P<0.001) principally attributable to
61
higher rates of reinfection (HR, 6.20, 95% CI, 2.56-15.00, P<0.001).
62
Conclusions: The proportion of IVDU among surgically treated IE patients is increasing. Although
63
IVDU have lower operative risk, long-term outcomes are compromised by reinfection.
SC
M AN U
TE D
EP
AC C
64
RI PT
45
65
Abstract word count: 230
66
Keywords: Intravenous drug use, infective endocarditis, valvular heart disease, operative risk
4
ACCEPTED MANUSCRIPT
Ultramini abstract
68
Among 436 adult patients with active endocarditis undergoing surgery, 78 (17.9%) were current
69
intravenous drug users. Although intravenous drug users had lower operative risk, reinfection (HR, 6.20,
70
95% CI, 2.56-15.00) and valve-related complications (HR, 3.82; 95% CI, 1.95-7.49) were significantly
71
higher compared with non-drug users.
RI PT
67
72 (Word count: 45)
SC
73
AC C
EP
TE D
M AN U
74
5
ACCEPTED MANUSCRIPT
1. Introduction
76
Intravenous drug abuse is increasing rapidly in the United States. Use of heroin, the most common illicit
77
intravenous (IV) drug worldwide, almost doubled in the United States between 2006 and 2013 to 681,000
78
active users, with an estimated 169,000 people starting the drug in 2013.1 The majority of intravenous
79
drug users (IVDU) are young adults who find heroin less expensive and easier to obtain than prescription
80
opioids.1-3 The incidence of heroin-related deaths has surged during this same period as well, with nearly
81
6000 deaths recognized to be caused by heroin intoxication in 2013, a figure almost tripled from that of
82
2006.4 Consequently, this has become a major public health issue in the United States.
83
Of the many medical complications of drug abuse, infective endocarditis (IE) is particularly challenging
84
given the significant risk of acute mortality as well as social factors such as late recidivism and
85
reinfection. Overall, IE carries in-hospital and 1–year mortality rates of 10-20% and a nearly 40%,
86
respectively, even in the current era.5,6 Of those with IE, 60-70% will ultimately require heart valve
87
surgery, particularly those with resistant or highly virulent organisms as is common among IVDU-
88
associated IE.5-8 The decision to proceed with surgical treatment is further complicated among IVDU, as
89
reinfection of a prosthetic valve increases subsequent operative risk and diminishes the likelihood of
90
medical cure.9-12 Given an estimated incidence of IE among this population of 100–400 per 100,000
91
person-years, which is 50-100 times that of the population at large, and the increased prevalence of IDU
92
in the U.S., this is likely to be an increasingly common clinical conundrum.6, 7, 13-15
93
Enthusiasm for operative management of IE in IVDU in the surgical community is mixed, given
94
anecdotal experience of repeated recurrences attributable to continued use of IV drugs.16 Furthermore,
95
independent of reinfection, survival of IVDU is threatened by competing causes of mortality including
96
drug intoxication, trauma and suicide.8, 17-19 These issues raise questions about the utility versus futility of
97
aggressive, resource-intensive surgical procedures particularly among patients who are actively injecting
98
drugs.16,20 There are, however, limited data concerning the clinical characteristics and prognosis after
AC C
EP
TE D
M AN U
SC
RI PT
75
6
ACCEPTED MANUSCRIPT
surgical therapy of these patients to inform clinical decision making and postoperative care.17,36 We
100
therefore sought to examine short- and long-term outcomes in the surgical management of IE among
101
IVDU patients in comparison to non-IVDU associated IE. In the interest of capturing an adequate sized
102
cohort, data were pooled from two closely associated tertiary academic centers in Boston, Massachusetts.
RI PT
99
103
2. Methods
105
-Study Population and Outcome Measures
106
The prospectively collected cardiac surgery databases of the Massachusetts General Hospital (MGH) and
107
the Brigham and Women’s Hospital (BWH) were queried to identify patients aged 17 years or older
108
undergoing heart valve operations due to active IE from January 2002 through August 2014. The disease
109
was considered active according to the Society of Thoracic Surgeons (STS) definitions if the patient was
110
currently being treated for endocarditis.21 These two databases uniformly record baseline patient
111
characteristics, information on cardiac surgical procedures and perioperative outcomes according to STS
112
definitions. In addition, retrospective chart reviews were conducted to collect detailed information
113
regarding IVDU status and profiles on IE such as specific pathogens, characteristics of valvular lesions
114
and treatment details. IVDU currently engaging in IDU was defined as those who had injected in the past
115
12 months as suggested by prior studies.1,4,35
116
Follow-up information on individual patients was obtained through January 2015 by review of
117
longitudinal data from Partners Health Care system, the largest health care system in Massachusetts,
118
which maintains a centralized clinical data registry of all patient encounters within the system including
119
BWH and MGH.19 This data warehouse has been used for other clinical research and its coding data were
120
validated in several prior studies.23,24
AC C
EP
TE D
M AN U
SC
104
7
ACCEPTED MANUSCRIPT
The primary outcome of interest was death and valve-related complications, the latter including
122
reinfection or reoperation of the heart valves, thromboembolic events, and anticoagulation-related
123
hemorrhage as defined by STS.25 Early mortality was defined as those occurring in-hospital or within 30
124
days of surgery. The diagnosis of valvular reinfection was based on one of the following criteria as
125
suggested by STS guidelines: (1) findings on reoperation with evidence of abscess, paravalvular leak,
126
frank pus, or vegetation confirmed as secondary to infection by histologic or bacteriologic studies; or (2)
127
meeting of the Duke Criteria for endocarditis in the absence of reoperation.25 Thromboembolic
128
complications including neurological (stroke or transient ischemic attack) or peripheral embolic events
129
producing signs or symptoms caused by the obstruction of a peripheral artery were diagnosed clinically.
130
Diagnosis of septic pulmonary embolism was made by the findings on enhanced computed tomography
131
of the chest when clinically suspected particularly in cases with IE involving right-sided heart valve.
132
Anticoagulation-related hemorrhage was defined as requiring transfusion, unplanned hospital admission,
133
or a hemostatic intervention.25 Early-, intermediate- and late-onset of recurrent IE were defined as those
134
occurred within 30 days, between 30-and 360 days, and beyond 360 days of surgery based on the current
135
practice guidelines.7
136
The study protocol was approved by the Institutional Review Board of Partners Healthcare, which
137
encompasses the MGH and the BWH. The requirement for informed consent from individual patients was
138
waved due to the retrospective nature of the study design.
139
-Statistical Analysis
140
Categorical variables, presented as frequencies and percentages, were compared using the χ2 test or
141
Fisher’s exact test. Continuous variables, expressed as mean±SD or median with range (or Quartile 1-3),
142
were compared using Student’s unpaired t test or the Mann–Whitney U test. Annual trends in proportions
143
of IVDU among surgical patients with infective endocarditis were tested by Chi-square test for linearity.
144
Kaplan–Meier curves were used to plot the survival and freedom from adverse events. Log-rank tests
AC C
EP
TE D
M AN U
SC
RI PT
121
8
ACCEPTED MANUSCRIPT
were used to compare the differences in the time-related outcomes between groups. Given the significant
146
differences in demographic and clinical risk profiles between the two groups of patients in favor of IVDU
147
group, we assumed that relative risks of deaths may differ across early and late postoperative periods. To
148
address this issue, landmark mortality analyses were performed split into two intervals: 90 days or less
149
and more than 90 days following surgery. 26 To adjust for baseline risk profiles between IVDU and non-
150
IVDU in the evaluations of outcomes, we performed propensity score analysis. The propensity scores
151
were calculated by multiple logistic regression analysis that included baseline variables shown in Table 1
152
and 2. Model discrimination and calibration were evaluated with C statistics (C=0.936) and Hosmer-
153
Lemeshow statistics (P=0.275), respectively. Propensity score matching yielded only 33 pairs of patients,
154
therefore we conducted a covariate adjustment with propensity score rather than using the propensity
155
score as a matching tool. With this technique, propensity scores were incorporated into Cox proportional
156
hazard models as a covariate as well as IVDU status to calculate the propensity-adjusted hazard ratio
157
(HR). Cox-proportional hazard models were used with robust standard errors that accounted for the
158
clustering of center. The proportional hazards assumption was confirmed by examination of log (-
159
log[survival]) curves and by testing of partial (Schoenfeld) residuals and no relevant violations were
160
found.
161
Since defining IVDU depended heavily on patients’ own reports, we assumed that there was potential
162
uncertainty in classifying IVDU status. In order to account for potential impacts of underestimation of
163
current IVDU, a sensitivity analysis was performed. For this analysis, we randomly sampled 3% of
164
patients in the non-IVDU to reclassify them as IVDU and recalculated HRs for clinical endpoints. This
165
procedure was repeated for 100 times to obtain averaged HRs and 95% CIs. We gradually increase the
166
proportions of reclassified patients from 3% to 70% of all non-IVDU, and evaluated how this
167
reclassification influenced on overall HRs.
168
A 2-tailed P value <0.05 was considered to be statistically significant. Analyses were performed with the
169
use of R statistical software, version 3.1.2.
AC C
EP
TE D
M AN U
SC
RI PT
145
9
ACCEPTED MANUSCRIPT
170
3. Results
172
-Baseline Characteristics
173
There were 436 patients meeting the enrollment criteria from the two centers (MGH, n=192; BWH,
174
n=244). Overall, 78 patients (17.9%) were IVDU currently engaging in IDU with the proportion
175
significantly increasing during the study period (P=0.048, Figure 1), particularly after 2012 (14.8% to
176
26.1%, P=0.006). Documented drugs of use included heroin only in 33, cocaine only in 18, combined
177
heroin and cocaine in 19, heroin or cocaine combined with other IV drugs in 5, amphetamine in 2 and
178
polysubstance in 1 with others unknown.
179
The baseline profiles of patients according to current IVDU status are summarized in Table 1. Overall,
180
compared with non-IVDU patients, IVDU were younger and had fewer cardiovascular risk factors (e.g.
181
diabetes, hypertension, renal dysfunction, EuroSCORE II). They presented more frequently, however,
182
with current cigarette smoking, embolic events, large vegetations, and right-sided and multiple valvular
183
involvement. There were no significant differences in gender and race/ethnicity. Causative pathogens
184
including the rates of methicillin-resistant Staphylococcus aureus infection were similar (11.5% in IVDU
185
[9/78] and 8.4% in non-IVDU [30/358], P=0.38) between the two groups.
186
Surgical profiles of patients are detailed in Table 2. Overall, 99 patients (22.7%) underwent emergent
187
surgery and 21 (4.8%) had intra-aortic balloon pumps placed before surgery. Valves were repaired in 45
188
(10.3%). Replacements were with mechanical prostheses in 99 (22.7%), bioprostheses in 206 (47.2%) or
189
human allografts (homografts) in 86 (19.7%). There were no significant differences between IVDU and
190
non-IVDU in operative profiles. All patients received at least 4-6 weeks of IV antibiotic therapy
191
regardless of IVDU status
192
-Outcome: Overall Cohort
AC C
EP
TE D
M AN U
SC
RI PT
171
10
ACCEPTED MANUSCRIPT
Thirty-day mortality rates were lower in IVDU than non-injectors (3.8% vs. 13.7%, P=0.012).
194
Late Follow-up data derived from the Partners database was complete in 76.4% of the entire cohort with a
195
median duration of 29.4 months (Quartile 1-3; 570 patient-years). Completeness of follow-up was lower
196
in IVDU than in the non-injector group (73.5% vs. 77.1%), but without statistical significance (P=0.096).
197
Major adverse events for the entire cohort in the form of valve-related complications were known to occur
198
in 69, including recurrence of IE in 42, valve reoperation in 35, thromboembolism in 17 and
199
anticoagulation-related hemorrhage in 16. (Table 3) Kaplan-Meier curves did not show a significant
200
difference in overall survival rates between IVDU and non-injectors; however, freedom from valve-
201
related complications were significantly lower among IVDU, mainly attributed to higher rates of
202
reinfection and reoperation compared with non-injectors (Figure 3).
203
Overall 5- and 10-year-survival rates were 78.9±6.0% and 69.5±8.3% in IVDU, respectively; and
204
76.1±2.6% and 68.7±3.9% in non-injectors, respectively (P=0.39). When adjusted by propensity score,
205
IVDU tended to have lower risk of early mortality (HR, 0.32; 95% CI, 0.08-1.36; P=0.13 for 90-day death)
206
but with higher risk of late mortality (HR, 2.07; 95% CI, 0.78-5.48; P=0.14) compared with non-IVDU,
207
however, they were statistically insignificant.
208
Five-and 10-year-survival rates free of valve-related events were 40.9±7.5% and 6.9±6.2% among IVDU,
209
respectively; and 63.9±3.3% and 50.8±4.9% in non-injectors (P=0.006). After propensity score
210
adjustments, risks of valve related complications were significantly higher in IVDU compared with non-
211
injectors (HR, 3.82; 95% CI, 1.95-7.49; P<0.001), which was mainly stemmed from a higher risk of
212
reinfection and consequent reoperations (Table 3).
213
-Outcomes: IVDU Subgroup
214
Among early IVDU IE survivors (n=75), 28 patients (37.3%) had known reinfection of the valves at a
215
median of 18.1 months (range 0.2-54.3 months) following valve surgery. Among these, 1 (3.6%) occurred
216
within 30 days, 9 (32.1%) between 30 days and 1 year, and 18 (64.3%) more than 1 year after surgery.
AC C
EP
TE D
M AN U
SC
RI PT
193
11
ACCEPTED MANUSCRIPT
Causative organisms of the reinfections were viridans group Streptococci in 6 (21.4%), methicillin-
218
susceptible Staphylococci in 5 (17.9%), methicillin-resistant Staphylococci in 4 (14.3%), Enterococci in 2
219
(7.1%) and fungus in 2 (7.1%) while 9 patients (32.1%) showed culture-negative results. Overall, only 6
220
patients (21.4%) were found to have the same causative organisms as at initial presentation.
221
Of the 28 patients with reinfection, 14 underwent surgery and 14 were treated with antimicrobial therapy
222
only. The reasons for medical therapy in the latter 14 patients were presence of intracranial hemorrhage in
223
3 (31.4%), prohibitive surgical risk due to severity of illness in 2 (14.3%), active in-hospital use of IV
224
drugs in 3 (31.4%) and cardiac arrest during preparation for operation in 1 (7.1%). Another 5 patients
225
(35.7%) were successfully treated medically with complete resolution of IE. Among those treated
226
surgically there were no early surgical deaths although one patient died of intracranial hemorrhage at 46.7
227
months postoperative. The median duration of antimicrobial therapy in the 14 patients who had only
228
medical therapy was 6.0 weeks (range 3-12 weeks) with 3 in-hospital deaths (21.4%) and 3 late deaths
229
(21.4%). Causes of the early deaths were sepsis in 2 and ventricular arrhythmia in 1, and the causes of late
230
deaths were repeated recurrence of IE in 2 and sudden death in 1. The 5-year survival rates among
231
surgical and medical therapy patients for recurrent IE in IVDU were 91.7±8.0% and 38.9±16.7%,
232
respectively (P=0.066, Supplemental Figure 1). Reoperative valve operations were undertaken for reasons
233
other than reinfection in 6 other patients (7.7%) for insufficiency of a repaired valve (n= 2) or structural
234
deterioration of a prosthetic valve (homograft [n=3] or xenograft bioprostheses [n=1]).
235
Overall, 11 late deaths (14.1%) occurred in IVDU. Causes of deaths were recurrence of IE in 6,
236
intracranial hemorrhage in 1 and metastatic lung cancer in 1 while causes could not be specified in 3
237
patients: one died suddenly and two died of unknown causes, one was found dead while actively using
238
heroin.
239
-Sensitivity Analysis
AC C
EP
TE D
M AN U
SC
RI PT
217
12
ACCEPTED MANUSCRIPT
Because of the challenges associated with determining with certainty the IVDU status of all patients in the
241
study, we undertook a sensitivity analysis of the impact of misclassification as illustrated in Figure 3. The
242
forest plots indicate adjusted HRs for valve-reinfection according to the increasing proportion of
243
reclassified patients from non-IVDU to IVDU (3% to 70%). For valve-reinfection, significantly increased
244
risks are observed up to 50%, while the proportion of reclassification that would be required to account
245
for the observed increased risk of reinfection would be at least 60%.
246
4. Discussion
247
Our study confirms an increase in the proportion of IVDU amongst surgical patients with active IE,
248
almost doubling during the study interval. This is consistent both with our clinical impression and the
249
virtual epidemic of increasing IV drug use in the general population as reported by the Center for Disease
250
Control.1 Perhaps more unexpectedly to us, the data also showed a markedly lower operative risk among
251
IVDU than non-IVDU. This is likely in part because the former were typically 25 years younger and had
252
fewer comorbidities compared to the non-injectors. Another potential factor may be surgical selection
253
bias, as we did not capture all patients seen with IE and cannot therefore account for those who were not
254
accepted for surgical intervention. It is possible that a subset of extremely ill individuals were not offered
255
surgery.
256
Despite superior early survival, late outcomes when corrected for age and other factors by propensity
257
score were worse for IVDU than non-IVDU, with higher rates of reinfection and reoperation. This is
258
likely a reflection of the challenge of recidivism among IVDU, although actual data concerning recurrent
259
drug use are hard to come by. Individuals may well not be forthright in responding to this sensitive
260
question when presenting with IE after valve surgery for IVDU-related IE out of concern that this will
261
impact the enthusiasm of clinicians to reoperation once again. The same may be true of family members.
262
Furthermore, unless a drug screen returns positive, the responses cannot be readily verified.
AC C
EP
TE D
M AN U
SC
RI PT
240
13
ACCEPTED MANUSCRIPT
Our findings of a rising incidence of IVDU-associated IE are consistent with a recent Swedish nationwide
264
cohort study of 7603 patients with IE, 990 of whom were treated surgically.6 In that study, the 30-day
265
mortality remained constant at around 10% over the decade of the study indicating that despite general
266
medical progress in the current era, IE remains a highly lethal condition. Overall, in contrast to our study,
267
the absolute mortality at 5 years follow-up was lower for the 355 IVDU than for non-injectors, when
268
compared to age- and gender-matched general population, however, IVDU showed a Standardized
269
Mortality Ratio of 19.1 (95% CI, 13.5–27.1) which was much higher than other subgroups (Standardized
270
Mortality Ratio range, 1.8-6.3). Similarly, Rabkin and colleagues observed markedly lower long-term
271
survival among IVDU than non-IVDU patients and, conversely, less frequent re-infection .17 The reasons
272
for this difference is a matter of speculation, however it is clear that the socioeconomic aspects of IVDU
273
are rapidly changing in this country, with heroin use markedly up among the middle class or wealthy.37
274
This may translate into greater access to healthcare should use recur. It should be noted that this
275
disappointing long-term survival of IVDU has been demonstrated independent of IE. Many IV drug
276
initiators become chronic users, and survival is significantly compromised should they not resolve IDU.
277
27-30
278
population with an average loss of life of 18.3 years.27,29 Although one study indicated IE as the most
279
common cause of death among IVDU,28 other studies showed drug overdose, accident, violence and other
280
chronic disease such as AIDS and liver cirrhosis as competing leading causes of death in these
281
individuals.27,29-31
282
The combination of anticipated short life expectancy despite young chronologic age and the perceived
283
high risks of treatment failure and recurrence of the disease after surgery challenge many front line health
284
care providers as an ethical controversy .16,20 Despite the growing incidence of IDU, current practice
285
guidelines do not address surgical indications or management strategies for active IVDU specifically and
286
treating physicians are left to formulate therapeutic plans on an individual basis without much hard
287
evidence in the literature. The results of this present study help to fill that evidence gap. These data
TE D
M AN U
SC
RI PT
263
AC C
EP
The mortality rate among long-term IVDU has been shown to be 6.9 times greater than the general
14
ACCEPTED MANUSCRIPT
suggest that for IVDU actively injecting, the technical exercise of replacing the infected valve(s)
289
combined with antimicrobial therapy can be performed at low perioperative risk, but may be incomplete
290
therapy without the addition of effective addiction treatment. It is apparent that resolution or minimization
291
of post-operative IDU would likely improve outcomes for this population via prevention of late re-
292
infection just as discontinuation of tobacco use may improve outcomes after coronary revascularization.
293
From a surgical standpoint, failure to resolve IDU is comparable to leaving the nidus of infection behind
294
as shown by a 10-year event-free survival rate of only 6.9±6.2% among IVDU. Given the very high
295
recidivism rate of IDU, insistence by treating clinicians on agreement by the subject to participation in a
296
drug rehabilitation program might be a reasonable consideration. Consultation to institutional ethics
297
committee may also be appropriate when making a decision to undertake surgery in active IVDU who
298
refuse to enter a drug rehabilitation program and are likely to remain at high risk to re-infect their valve.
299
In one of our institutions (MGH) consultation with an addiction counselor has become standard routine at
300
the time of surgical consultation for IE in IVDU.
301
Finally, we observed Staphylococcus aureus as the most common pathogen of IE among IVDU and right-
302
side valve involvement to be more common in IVDU than non-IVDU as has been reported previously.32-34
303
Unlike earlier prior reports we found left-sided valve involvement predominant among IVDU who
304
undergo surgery.34 This is consistent with the findings of Rabkin and colleagues and may be a reflection
305
of a shift in surgical philosophy away from intervention on right sided lesions in favor of antimicrobial
306
therapy and the implantation of a prosthetic makes any subsequent infection an even more complicated
307
problem.17 The poor late outcomes in IVDU could not be attributed to differences in the virulence or type
308
of organism.
310
SC
M AN U
TE D
EP
AC C
309
RI PT
288
-Limitations
15
ACCEPTED MANUSCRIPT
This study has inherent limitations in the retrospective nature of the research. Study subjects are affected
312
by surgical selection bias since only patients accepted for surgery were included and those who received
313
medical therapy only were not, nor was there a medical treatment only control group for comparison.
314
Obtaining reliable follow-up data concerning reinfection and valve-related complications in this patient
315
population was also challenging. As this group is the most challenging in follow-up, the data are likely
316
biased in favor of lower event rates in this subgroup. Finally, the data set represents the experience of
317
urban tertiary academic referral centers, and therefore, it is subject to referral bias from the community.
318
- Conclusions
319
The proportion of IVDU among surgical patients with IE is increasing. Although IVDU are younger, with
320
lower baseline cardiovascular risk burdens, long-term clinical outcomes are discouraging with a
321
significant rate of reinfection and a trend toward higher risk of late mortality. These data highlight the
322
need to aggressively address the expanding drug use epidemic that is driving the increased rate of IE in
323
this population and provide effective addiction treatment post-operatively to prevent re-infection.
M AN U
References
326
1.
AC C
Publication No. (SMA) 14-4863. Rockville, MD: Substance Abuse and Mental Health Services
328
Administration, 2014.
329 2.
Hooten WM, Bruce BK. Beliefs and attitudes about prescribing opioids among healthcare
providers seeking continuing medical education. J Opioid Manag.;7(6):417-424.
331
333
Substance Abuse and Mental Health Services Administration, Results from the 2013 National Survey on Drug Use and Health: Summary of National Findings, NSDUH Series H-48, HHS
327
332
EP
325
TE D
324
330
SC
RI PT
311
3.
Slevin KA, Ashburn MA. Primary care physician opinion survey on FDA opioid risk evaluation and mitigation strategies. J Opioid Manag.;7(2):109-115. 16
ACCEPTED MANUSCRIPT
334
4.
Trends in Drug-poisoning Deaths Involving Opioid Analgesics and Heroin: United States, 1999–
335
2012. Centers for Disease Control and Prevention
336
(http://www.cdc.gov/nchs/data/hestat/drug_poisoning/drug_poisoning.htm). 5.
Chu VH, Park LP, Athan E, et al. Association between surgical indications, operative risk, and
RI PT
337 338
clinical outcome in infective endocarditis: a prospective study from the international
339
collaboration on endocarditis. Circulation. ;131(2):131-140.
and long-term prognosis in infective endocarditis in Sweden. PLoS One.8(7):e67519.
341
342
Ternhag A, Cederstrom A, Torner A, Westling K. A nationwide cohort study of mortality risk
SC
6.
7.
M AN U
340
Nishimura RA, Otto CM, Bonow RO, et al. 2014 AHA/ACC Guideline for the Management of
343
Patients With Valvular Heart Disease: a report of the American College of Cardiology/American
344
Heart Association Task Force on Practice Guidelines. Circulation.;129(23):e521-643.
345
8.
Shih CJ, Chu H, Chao PW, et al. Long-term clinical outcome of major adverse cardiac events in survivors of infective endocarditis: a nationwide population-based study.
347
Circulation.;130(19):1684-1691. 9.
Grubitzsch H, Schaefer A, Melzer C, Wernecke KD, Gabbieri D, Konertz W. Outcome after
EP
348
TE D
346
surgery for prosthetic valve endocarditis and the impact of preoperative treatment. J Thorac
350
Cardiovasc Surg. 2014;148(5):2052-2059.
351
10.
endocarditis. Ann Thorac Surg. 2002;74(5):S1754-1757.
353
355
Lytle BW, Sabik JF, Blackstone EH, Svensson LG, Pettersson GB, Cosgrove DM, 3rd.
Reoperative cryopreserved root and ascending aorta replacement for acute aortic prosthetic valve
352
354
AC C
349
11.
Meszaros K, Nujic S, Sodeck GH, et al. Long-term results after operations for active infective endocarditis in native and prosthetic valves. Ann Thorac Surg. 2012;94(4):1204-1210.
17
ACCEPTED MANUSCRIPT
356
12.
Musci M, Weng Y, Hubler M, et al. Homograft aortic root replacement in native or prosthetic
357
active infective endocarditis: twenty-year single-center experience. J Thorac Cardiovasc Surg.
358
2010;139(3):665-673. 13.
1988-1990. Am J Cardiol. 1995;76(12):933-936.
360
361
Berlin JA, Abrutyn E, Strom BL, et al. Incidence of infective endocarditis in the Delaware Valley,
14.
RI PT
359
Hogevik H, Olaison L, Andersson R, Lindberg J, Alestig K. Epidemiologic aspects of infective endocarditis in an urban population. A 5-year prospective study. Medicine (Baltimore).
363
1995;74(6):324-339. 15.
M AN U
364
SC
362
Spijkerman IJ, van Ameijden EJ, Mientjes GH, Coutinho RA, van den Hoek A. Human
365
immunodeficiency virus infection and other risk factors for skin abscesses and endocarditis
366
among injection drug users. J Clin Epidemiol. 1996;49(10):1149-1154.
367
16.
DiMaio JM, Salerno TA, Bernstein R, Araujo K, Ricci M, Sade RM. Ethical obligation of surgeons to noncompliant patients: can a surgeon refuse to operate on an intravenous drug-
369
abusing patient with recurrent aortic valve prosthesis infection? Ann Thorac Surg. 2009;88(1):1-8. 17.
Rabkin DG, Mokadam NA, Miller DW, Goetz RR, Verrier ED, Aldea GS. Long-term outcome
EP
370
TE D
368
for the surgical treatment of infective endocarditis with a focus on intravenous drug users. Ann
372
Thorac Surg. 2012;93(1):51-57.
373
18.
Infect Dis. 2007;39(3):197-204.
375
377
Thalme A, Westling K, Julander I. In-hospital and long-term mortality in infective endocarditis in
injecting drug users compared to non-drug users: a retrospective study of 192 episodes. Scand J
374
376
AC C
371
19.
Hubbell G, Cheitlin MD, Rapaport E. Presentation, management, and follow-up evaluation of infective endocarditis in drug addicts. Am Heart J. 1981;102(1):85-94.
18
ACCEPTED MANUSCRIPT
378
20.
Hull SC, Jadbabaie F. When is enough enough? The dilemma of valve replacement in a recidivist intravenous drug user. Ann Thorac Surg. ;97(5):1486-1487.
379
21.
www.riskcalc.sts.org
381
22.
Murphy SN, Chueh HC. A security architecture for query tools used to access large biomedical databases. Proc AMIA Symp. 2002:552-556.
23.
history of asymptomatic severe carotid artery stenosis. J Vasc Surg. 2014;60(5):1218-1225. 24.
Horkan CM, Purtle SW, Mendu ML, Moromizato T, Gibbons FK, Christopher KB. The
M AN U
384
385
Conrad MF, Michalczyk MJ, Opalacz A, Patel VI, LaMuraglia GM, Cambria RP. The natural
SC
382
383
RI PT
380
386
association of acute kidney injury in the critically ill and postdischarge outcomes: a cohort study*.
387
Crit Care Med. 2015;43(2):354-364. 25.
26.
27.
Health. 1987;77(3):347-348.
393
394
28.
29.
399
Smyth B, Fan J, Hser YI. Life expectancy and productivity loss among narcotics addicts thirty-
three years after index treatment. J Addict Dis. 2006;25(4):37-47.
397
398
Perucci CA, Forastiere F, Rapiti E, Davoli M, Abeni DD. The impact of intravenous drug use on
mortality of young adults in Rome, Italy. Br J Addict. 1992;87(12):1637-1641.
395
396
Joe GW, Simpson DD. Mortality rates among opioid addicts in a longitudinal study. Am J Public
EP
392
Therneau TM, Grambsch PM. Modeling survival data : extending the Cox model. New York: Springer; 2000.
391
TE D
cardiac valve interventions. J Thorac Cardiovasc Surg. 2008;135(4):732-738.
389
390
Akins CW, Miller DC, Turina MI, et al. Guidelines for reporting mortality and morbidity after
AC C
388
30.
Smyth B, Hoffman V, Fan J, Hser YI. Years of potential life lost among heroin addicts 33 years after treatment. Prev Med. 2007;44(4):369-374. 19
ACCEPTED MANUSCRIPT
31.
Drug Abuse Warning Network 2003 (09/2004 update).
401 32.
drug users: the time has come? Eur J Clin Microbiol Infect Dis. 1994;13(7):533-534.
403
404
DiNubile MJ. Abbreviated therapy for right-sided Staphylococcus aureus endocarditis in injecting
33.
Hecht SR, Berger M. Right-sided endocarditis in intravenous drug users. Prognostic features in 102 episodes. Ann Intern Med. 1992;117(7):560-566.
405
RI PT
402
DAWN 2003: Area profiles of drug-related mortality. Office of Applied Studies, SAMHSA.
SC
400
34.
407
involvement, bacteriologic findings, and outcome of infective endocarditis in intravenous drug users.
408
Arch Intern Med. 1995;155(15):1641-1648.
409
35.
Mathers BM, Degenhardt L, Phillips B, et al. Global epidemiology of injecting drug use and HIV among people who inject drugs: a systematic review. Lancet. 2008;372(9651):1733-1745.
36.
Dawood MY, Cheema FH, Ghoreishi M, et al. Contemporary outcomes of operations for tricuspid
TE D
410
411
Mathew J, Addai T, Anand A, Morrobel A, Maheshwari P, Freels S. Clinical features, site of
M AN U
406
valve infective endocarditis. Ann Thorac Surg. 2015;99(2):539-546.
412
37. http://www.nytimes.com/interactive/2015/10/30/us/31heroin-deaths.html?_r=0 (accessed Jan 17,
414
2016)
AC C
EP
413
20
ACCEPTED MANUSCRIPT
Table 1. Baseline Characteristics between IVDU and Non-IVDU Non-IVDU
(n=78)
(n=358)
Female gender, n (%)
30 (38.5)
Body mass index, kg/m2
26.7±7.9
M AN U
Ethnicity, n (%) White Black Hispanic
TE D
Asian Others
Hypertension
EP
Diabetes mellitus
59.3±14.1
<0.001
111 (31.0)
0.20
27.8±7.0
0.24
SC
35.9±9.9
Cigarette smoking history
Current cigarette smoking
NYHA functional class III or IV, n (%) On dialysis, n (%)
21
P value
RI PT
IVDU
Age, yr
AC C
415
0.83
68 (87.2)
319 (89.1)
5 (6.4)
14 (3.9)
1 (1.3)
9 (2.5)
1 (1.3)
5 (1.4)
3 (3.8)
11 (3.1)
6 (7.6)
74 (20.7)
0.027
16 (20.5)
225 (62.8)
<0.001
53 (67.9)
142 (39.7)
<0.001
28 (35.9)
29 (8.1)
<0.001
35 (44.9)
177 (49.4)
0.46
3 (3.8)
31 (8.7)
0.24
ACCEPTED MANUSCRIPT
Creatinine clearance, mL/min/1.73m2
73.1±44.1
<0.001
Embolic events
36 (46.2)
107 (29.9)
0.006
Cerebral
21 (26.9)
93 (26.0)
0.86
15
Peripheral
RI PT
103.5±53.6
14
0.14
Causative pathogens, n (%) 16 (20.5)
89 (24.9)
SC
Viridans group streptococci
2 (2.6)
M AN U
Other streptococci
33 (9.2)
30 (38.5)
97 (27.1)
9 (11.5)
30 (8.4)
10 (12.8)
65 (18.2)
1 (1.3)
35 (9.8)
12 (15.4)
48 (13.4)
9 (11.5)
24 (6.7)
9 (11.5)
25 (7.0)
57 (73.1)
294 (82.1)
Aortic valve
32 (41.0)
200 (55.9)
Mitral valve
13 (16.7)
86 (24.0)
Tricuspid valve
12 (15.4)
8 (2.2)
Staphylococcus aureus methicillin-resistant
methicillin-resistant Entercococci
AC C
Negative culture
EP
Others
TE D
Staphylococci, coagulase-negative
Single valve affected, n (%)
22
ACCEPTED MANUSCRIPT
64 (17.9)
13 (16.7)
54 (15.1)
Aortic + Tricuspid
3 (3.8)
6 (1.7)
Mitral + Tricuspid
5 (6.4)
Aortic + Mitral
0
Aortic +Mitral + Tricuspid
18 (5.0)
<0.001
Vegetation diameter ≥10mm, n (%)
55 (70.5)
178 (49.7)
<0.001
18 (23.1)
121 (33.8)
0.066
17 (21.8)
110 (30.7)
0.12
Severe dysfunction of affected valves, n (%)
59 (75.6)
245 (68.4)
0.21
Left ventricular ejection fraction, %
59.5±10.4
60.1±10.9
0.64
9.5±8.3
13.4±12.6
0.001
(1.8-42.0)
(1.8-82.9)
M AN U
TE D
AC C
EP
EuroSCORE II, % (range)
SC
20 (25.6)
Prosthetic endocarditis, n (%)
417
2 (0.6) 2 (0.6)
Right-side valve involvement
Abscess formation, n (%)
416
0.068
RI PT
21 (26.9)
Multiple valves affected, n (%)
23
ACCEPTED MANUSCRIPT
Table 2. Surgical Profiles Non- IVDU
(n=78)
(n=358)
Emergent surgery
16 (20.5) 3 (3.8)
On IABP
M AN U
Primary procedure
6 (0-140)
0.098
83 (23.2)
0.61
18 (5.0)
>0.99
SC
7 (0-76)
P value
RI PT
IVDU
Interval from admission to surgery, days
0.87
10 (12.8)
35 (9.8)
18 (23.1)
81 (22.6)
35 (44.9)
171 (47.8)
15 (19.2)
71 (19.8)
2 (2.6)
20 (5.6)
0.27
10 (12.8)
56 (15.6)
0.60
Aortic root replacement
19 (24.4)
96 (26.8)
0.78
Cardiopulmonary bypass time, min
191.3±103.8
215.5±142.0
0.34
Cardiac ischemic time, min
145.9±85.3
159.3±97.8
0.30
Valve repair Valve replacement, mechanical Valve replacement, bioprostheses
Associated procedures
TE D
Valve replacement, allograft
CABG
EP
Ascending Aorta replacement
AC C
418
24
ACCEPTED MANUSCRIPT
419
Abbreviations: IABP, intra-aortic balloon pumping; CABG, coronary artery bypass grafting; ASD, atrial
420
septal defect; PFO, patent foramen ovale; VSD, ventricular septal defect; MAIVF, mitral-aortic inter-
421
valvular fibrosa.
AC C
EP
TE D
M AN U
SC
RI PT
422
25
ACCEPTED MANUSCRIPT
Table 3. Unadjusted and Adjusted Clinical Outcomes According to Current Status of IV Drug Abuses Crude
Propensity-Adjusted
RI PT
No. of Events Non-IVDU
HR
95% CI
P value
HR
95% CI
P value
30-day mortality
1
40
0.10
0.01-0.48
0.026
0.15
0.01-1.01
0.10
90-day mortality
3
49
0.25
0.024
0.32
0.08-1.36
0.13
Late (> 90 days) mortality
10
30
Valve related complications
32
37
Valve Re-infection
28
14
Valve reoperation
18
Thromboembolism
7
Hemorrhage
5
M AN U
0.06-0.71 0.66-2.63
0.45
2.07
0.78-5.48
0.14
4.07
2.53-6.55
0.001
3.82
1.95-7.49
<0.001
9.28
4.87-17.68
<0.001
6.20
2.56-15.00
<0.001
17
5.03
2.58-9.82
0.005
3.66
1.44-9.30
0.006
10
2.92
1.11-7.68
0.030
2.54
0.62-10.47
0.20
11
1.75
0.60-5.05
0.30
4.33
1.03-18.17
0.045
TE D
1.32
EP
AC C
SC
IVDU
Abbreviations: IVDU, intravenous drug users; HR, hazard ratio; CI, confidence interval.
26
ACCEPTED MANUSCRIPT
Figure Legends Figure 1. Trends in proportions of IVDU among surgical patients with infective endocarditis over time.
RI PT
Abbreviation: IVDU, intravenous drug users. Figure 2. Kaplan-Meier curves for clinical outcomes between IVDU and non-IVDU: Overall death (A), valve reinfection (B), valve reoperation (C) and composite of valve –related complications (D).
Transparent bands indicate area within 95% confidence interval. Abbreviation: IVDU, intravenous drug
SC
users.
M AN U
Figure 3. Effects of reclassification of non-IVDU patients to IVDU on hazard ratios for the risk of reinfection. Dots and horizontal lines indicate hazard ratios and 95% confidence interval, respectively. Abbreviation: HR, hazard ratio.
Supplementary Figure 1. Survival after surgical versus medical therapy for recurrence of infective
AC C
EP
TE D
endocarditis in intravenous drug users
27
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT