- Email: [email protected]
Risk of Venous Thromboembolism Following Peripherally Inserted Central Catheter Exchange: An Analysis of 23,000 Hospitalized Patients
Accepted Manuscript Title: Risk of Venous Thromboembolism Following Peripherally Inserted Central Catheter Exchange: an Analysis of 23,000 Hospitalized Patients Author: Vineet Chopra, Scott Kaatz, Paul Grant, Lakshmi Swaminathan, Tanya Boldenow, Anna Conlon, Steven J. Bernstein, Scott A. Flanders PII: DOI: Reference:
S0002-9343(18)30089-5 https://doi.org/10.1016/j.amjmed.2018.01.017 AJM 14500
To appear in:
The American Journal of Medicine
Please cite this article as: Vineet Chopra, Scott Kaatz, Paul Grant, Lakshmi Swaminathan, Tanya Boldenow, Anna Conlon, Steven J. Bernstein, Scott A. Flanders, Risk of Venous Thromboembolism Following Peripherally Inserted Central Catheter Exchange: an Analysis of 23,000 Hospitalized Patients, The American Journal of Medicine (2018), https://doi.org/10.1016/j.amjmed.2018.01.017. 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.
1 2 3
Risk of Venous Thromboembolism following Peripherally Inserted Central
4
Catheter Exchange: An Analysis of 23,000 Hospitalized Patients
5
Running Title: PICC Exchange and Deep Vein Thrombosis
6 7
Vineet Chopra MD, MSc (a, b, c)
8
Scott Kaatz DO, MSc (d)
9
Paul Grant MD (a, c)
10
Lakshmi Swaminathan MD (e)
11
Tanya Boldenow MD (f)
12
Anna Conlon, PhD (a, b)
13
Steven J. Bernstein MD, MPH (b, c, g)
14
Scott A. Flanders MD (a, c)
15 16
From: (a) The Division of Hospital Medicine, Department of Internal Medicine,
17
University of Michigan Medical School, Ann Arbor, MI; (b) Patient Safety Enhancement
18
Program and Center for Clinical Management Research, VA Ann Arbor Health Care
19
System, Ann Arbor, MI (c); the Michigan Hospital Medicine Safety Consortium, Ann
20
Arbor, MI (d) Henry Ford Health System, Detroit, MI; (e) Beaumont Hospital, Dearborn,
21
MI; (f) St. Josephs Health Center, Ypsilanti, MI and (g) The Division of General
22
Medicine, Department of Internal Medicine, University of Michigan Medical School, Ann
1
Page 1 of 34
23
Arbor, MI.
24 25
Abstract Word Count: 262
26
Manuscript Word Count: 3100
27 28
Conflicts of Interest: Letters from all authors indicating COIs are attached. All authors
29
had access to the data and a role in writing the manuscript.
30 31
Funding: Support for HMS is provided by Blue Cross and Blue Shield of Michigan and
32
Blue Care Network as part of the BCBSM Value Partnerships program. Although Blue
33
Cross Blue Shield of Michigan and HMS work collaboratively, the opinions, beliefs and
34
viewpoints expressed by the author do not necessarily reflect the opinions, beliefs and
35
viewpoints of BCBSM or any of its employees. Dr. Chopra is supported by a career
36
development award from AHRQ (1-K08HS022835-01).
37 38
Role of Study Sponsor: Blue Cross / Blue Shield of Michigan and Blue Care Network
39
supported data collection at each participating site and funded the data coordinating
40
center but had no role in study concept, interpretation of findings, or in the preparation,
41
final approval or decision to submit the manuscript.
42 43
Keywords: Peripherally inserted central catheter; PICC; exchange, deep vein
44
thrombosis; central venous catheter; thromboembolism.
45
2
Page 2 of 34
46
Address for Correspondence:
47
Vineet Chopra MD, MSc
48
2800 Plymouth Road, Building 16, #432W
49
Ann Arbor, MI 48109
50
[email protected]
51
2800 Plymouth Road Building 16 #432W
52
Ann Arbor, MI 48109
53
[email protected]
54 55 56 57
CLINICAL SIGNIFICANCE
following catheter occlusion or dislodgement.
58 59
62
Patients that underwent a catheter exchange experienced deep vein thrombosis more often and more quickly than those that did not.
60 61
Exchange of peripherally inserted central catheters is often performed
The risk of thrombosis following exchange was similar to that of using a larger catheter with more lumens.
63 64 65
3
Page 3 of 34
66
ABSTRACT
67
BACKGROUND. Catheter exchange over a guidewire is frequently performed for
68
malfunctioning peripherally inserted central catheters (PICCs). Whether such
69
exchanges are associated with venous thromboembolism is not known.
70 71
METHODS.
We performed a retrospective cohort study to assess the association between
72
PICC exchange and risk of thromboembolism. Adult hospitalized patients that received
73
a PICC during clinical care one of 51 hospitals participating in the Michigan Hospital
74
Medicine Safety consortium were included. The primary outcome was hazard of
75
symptomatic venous thromboembolism (radiographically confirmed upper-extremity
76
deep vein thrombosis and pulmonary embolism) in those that underwent PICC
77
exchange vs. those that did not.
78 79
RESULTS. Of 23,010 patients that underwent PICC insertion in the study, 589 patients
80
(2.6%) experienced a PICC exchange. Almost half of all exchanges were performed for
81
catheter dislodgement or occlusion. A total of 480 patients (2.1%) experienced PICC-
82
associated deep vein thrombosis. The incidence of deep vein thrombosis was greater in
83
those that underwent PICC exchange vs. those that did not (3.6% vs. 2.0%, p<0.001).
84
Median time to thrombosis was shorter among those that underwent exchange
85
compared vs. those that did not (5 vs. 11 days, p=0.02). Following adjustment, PICC
86
exchange was independently associated with two-fold greater risk of thrombosis
87
(hazard ratio [HR]=1.98, 95%CI=1.37-2.85) vs. no exchange. The effect size of PICC
88
exchange on thrombosis was second in magnitude to device lumens (HR=2.06
4
Page 4 of 34
89
[95%CI=1.59-2.66] and HR=2.31 [95%CI=1.6-3.33] for double- and triple lumen
90
devices, respectively).
91 92
CONCLUSION. Guidewire exchange of PICCs may be associated with increased risk of
93
thrombosis. As some exchanges may be preventable, consideration of risks and
94
benefits of exchanges in clinical practice is needed.
95 96
5
Page 5 of 34
97 98 99
INTRODUCTION Venous thromboembolism is an important complication associated with central venous catheters. Deep vein thrombosis and pulmonary embolism are especially
100
relevant in the case of peripherally inserted central catheters (PICCs),1 devices
101
associated with 2.5-fold greater risk of thrombosis than central venous catheters.2 Thus,
102
PICCs are associated with a higher baseline risk of thrombosis than traditional central
103
venous catheters.
104
PICCs are placed in smaller peripheral veins of the upper extremity rather than
105
large veins of the neck or chest. This unique route avoids important complications such
106
as pneumothorax or injury to the great vessels,3 and is an important reason behind
107
growing PICC use in clinical practice.4, 5 However, placement of PICCs in the arm also
108
means insertion into veins that are of smaller diameter compared to the chest.
109
Additionally, PICCs are longer and more slender than conventional catheters. Thus,
110
complications such as tip migration, catheter occlusion and device dislodgement occur
111
2-3 times more frequently with PICCs than central venous catheters.6, 7 When these
112
events occur, an over-the-wire guidewire exchange of the PICC is usually performed.
113
While such guidewire exchanges are sometimes necessary, they may not be risk free.
114
For example, one study reported increased risk of infection among neonates following
115
over-the-wire PICC exchanges.8 Similarly, in a study of hemodialysis recipients,
116
increased risk of dialysis catheter dysfunction was reported following guidewire
117
exchange.9 To date, however, no study has examined whether PICC exchanges are
118
associated with thrombosis. Given the known thrombogenicity of PICCs and the
6
Page 6 of 34
119
potential that exchanges may cause venous injury, this question is particularly relevant
120
for clinical practice.
121
Therefore, we used data from an ongoing study of hospitalized patients who
122
received PICCs to assess whether guidewire exchanges of PICCs were independently
123
associated with risk of thrombosis.
124 125
METHODS
126
Study Setting and Participants
127
The study was conducted using data from the Michigan Hospital Medicine Safety
128
(HMS) consortium, a 51-hospital collaborative quality initiative supported by Blue Cross
129
Blue Shield of Michigan and Blue Care Network. The design and setting of this
130
consortium have been previously described.1, 10, 11 In brief, adult patients admitted to a
131
general medicine ward or intensive care unit (ICU) of a participating hospital who
132
receive a PICC for any reason during clinical care are eligible for inclusion. Patients who
133
are (a) under the age of 18; (b) pregnant; (c) admitted to a non-medical service (e.g.,
134
general surgery); or (d) admitted under observation status are excluded from sampling.
135
At each hospital, dedicated, trained medical record abstractors use a defined
136
protocol to collect clinical data directly from medical records of patients. Based on
137
available resources and the time required to collect data from each case, patients with
138
PICCs are sampled on a 14-day cycle, with data from the first 17 cases that meet
139
eligibility criteria within each cycle included. To ensure adequate representation of
140
critically ill patients, at least 7 of the 17 eligible cases include PICCs placed in ICU
141
patients. We chose seven patients again based on the time required to abstract these
7
Page 7 of 34
142
cases (typically twice that of a patient with a PICC placed in a non-ICU setting). If a
143
hospital could not find 7 eligible patients that received PICCs in an ICU setting,
144
additional patients that received PICCs on general wards are used to backfill thus
145
ensuring 17 patients are included every two weeks. Because we were specifically
146
interested in risk of thrombosis following catheter exchange, patients with a history of
147
prior DVT (within six months), presumed VTE, palliative care status or direct ICU
148
admissions were not excluded unlike prior studies.12 All patients are followed until PICC
149
removal, death, or 70-days, whichever occurs first. To ensure completeness and
150
accuracy, staff from the University of Michigan perform annual on-site audits of all
151
participating hospitals to assess veracity and accuracy of abstracted data.
152
PICCs are defined as vascular access devices inserted in veins of the upper
153
extremity that terminate at the cavoatrial junction; thus, midlines, conventional catheters
154
or those placed in lower extremity veins are excluded. However, the presence of a
155
conventional central catheter in the lower extremity, chest or neck at the time of PICC
156
insertion is captured. Data regarding PICC characteristics (e.g., indication, gauge,
157
lumens, tip position verification) are obtained directly from vascular nursing or
158
interventional radiology notes or the order for PICC placement. All testing for thrombosis
159
occurs at the discretion of clinicians at each hospital. Since we were interested in
160
thrombosis following PICC exchange, both patients that underwent exchange of an
161
existing PICC (e.g., those with “index” PICCs) as well as those enrolled in the study
162
following a PICC exchange (e.g., patient admitted from an outside hospital with a non-
163
functioning PICC that required exchange) were included. When patients with an
164
exchange were enrolled, characteristics related to the exchanged PICC were used for
8
Page 8 of 34
165
analysis. For this analysis, data from patients enrolled between January 2013 and
166
October 2016 were included.
167 168
Covariates
169
Detailed medical history including comorbidities, physical examination findings,
170
laboratory and medication data were collected from the medical record. Standardized
171
definitions using ICD-10 and Elixhauser criteria were used to define comorbidities
172
present at hospital admission.13 Variables including age (<64 versus >65 years), sex,
173
race, body mass index, tobacco use (never, former, current), principal admitting
174
diagnosis, prior upper or lower-extremity deep vein thrombosis (within 30-days, beyond
175
30-days, never), inpatient surgery within 30 days of PICC placement, chemotherapy or
176
blood administration during hospitalization, trauma requiring hospitalization within 30
177
days, immobilizing cast at the time of PICC placement, hip or knee replacement within
178
30 days of PICC placement, presence of active infection, existing central venous
179
catheter when PICC was placed (yes/no), diabetes mellitus (uncomplicated vs
180
complicated by micro- or macrovascular complications), history of cerebrovascular
181
accident or transient ischemic attack, history of myocardial infarction, sickle cell
182
disease, venous thromboembolism prophylaxis (i.e., receipt of subcutaneous heparin
183
twice or thrice daily regimens or use of enoxaparin at prophylactic doses), receipt of
184
treatment dose anticoagulation for any reason, aspirin, statin, erythropoiesis stimulating
185
agents, and antiplatelet medication administration were abstracted directly from medical
186
records.
9
Page 9 of 34
187
Active cancer was defined as admission to the hospital for a cancer diagnosis
188
made within the past six months (e.g., acute leukemia) or for receipt of scheduled
189
chemotherapy for malignancy. Serious lung disease was defined as receipt of invasive
190
or non-invasive ventilation at any point during hospitalization. Life-threatening illness
191
was defined as care in an ICU-setting at admission or at any point during
192
hospitalization. Laboratory values including white blood cell (WBC) count, hemoglobin,
193
platelet count and international normalized ratio (INR) were collected at the time of
194
PICC insertion.
195 196 197
Ascertainment of Outcomes The primary outcome was acute, symptomatic, radiographically-confirmed,
198
upper-extremity deep vein thrombosis (e.g., compression or duplex ultrasound with
199
visible thrombus or non-compressibility of a deep vein of the arm [brachial, basilic,
200
axillary or subclavian]), or PE (e.g., high probability V/Q scan or positive contrast-
201
enhanced computerized tomography) occurring after initial PICC placement or after
202
catheter exchange. At all hospitals, testing for thromboembolism occurs only in the
203
presence of clinical symptoms (e.g., arm pain, swelling). Clinicians are at liberty of
204
ordering tests for these events based on clinical suspicion or findings.
205 206 207
Statistical Analyses Factors associated with thrombosis following catheter exchange were assessed
208
according to a previously published14 and validated conceptual model.15, 16 In modeling,
209
all factors known to be associated with deep vein thrombosis (e.g., elements from
10
Page 10 of 34
210
standardized risk scores) as well as those associated with risk of thromboembolism
211
following PICC placement from our prior studies were included. Descriptive variables
212
were used to summarize the distribution of patient-, provider- and device-factors
213
between patients with and without PICC exchange. Because PICC exchanges can
214
occur at any point during the dwell of the PICC (and we are interested in time-to-event
215
for deep vein thrombosis), Cox proportional hazards models with robust sandwich
216
covariance matrix estimates to account for hospital-level clustering were employed to
217
account for the time-dependent nature of exchanges.17 In addition to PICC exchange,
218
covariates that could change value during the dwell of the PICC (e.g., presence of
219
another central venous catheter, admission to ICU) were analyzed as time-dependent
220
variables and coded as null until such events occurred. Similarly, any device- or
221
provider-specific variables that could change value at the time of a PICC exchange
222
(e.g., number of PICC lumens), were also treated as time-dependent with values
223
changing if the variable was different for the exchanged device compared to the original
224
PICC. Unadjusted associations of covariates with PICC thrombosis were first expressed
225
as hazard ratios (HR) with corresponding 95% confidence intervals (CI). Missing data
226
were imputed through a 10-fold multiple imputation procedure.18
227
The final multivariable Cox model for the association of PICC exchange with
228
thrombosis was adjusted for factors previously reported as independent predictors of
229
PICC thrombosis,14, 19, 20 as well as potential confounding factors associated with both
230
deep vein thrombosis and probability of PICC exchange on unadjusted analyses. A two-
231
sided p-value of less than 0.05 was used to indicate significance in all analyses. All
232
analyses were performed in SAS, version 9.4 (SAS Institute Inc., Cary, NC) and R
11
Page 11 of 34
233
version 3.2.4.
234 235 236 237
Ethical and Regulatory Oversight The University of Michigan Medical School’s Institutional Review Board reviewed this study and it received a “Not Regulated” status.
238 239
RESULTS
240
Patient Characteristics and Outcomes
241
During the study period, 23,010 patients underwent PICC insertion in 51
242
Michigan hospitals and were included in this analysis. Of these, 589 patients (2.6%)
243
experienced a PICC exchange during the dwell of the device while 22,421 did not
244
(Figure 1). A total of 480 patients (2.1%) experienced venous thromboembolism in the
245
study (460=deep vein thrombosis; 20=pulmonary embolism). Notably, the frequency of
246
PICC thrombosis was greater among patients that underwent exchange vs. those that
247
did not (3.6% vs. 2.0%, p<0.001). Approximately half of all PICC exchanges were
248
performed for catheter dislodgement (30.2%) and occlusion (19.7%). Other indications
249
for exchange included need for more lumens (22.4%) and catheter migration (7.4%).
250
Bivariate comparisons showed that patients that underwent PICC exchanges
251
differed in several ways from those that did not (Table 1). With respect to comorbidities,
252
patients that underwent exchange had greater prevalence of moderate-severe liver
253
disease (5.4% vs. 3.8%, p=0.04) and active cancer (9.0% vs 6.1%, p<0.01) than those
254
that did not. Similarly, patients that underwent PICC exchange more often had a history
255
of deep vein thrombosis within 30 days (6.1% vs. 3.9, p<0.01) and history of central
12
Page 12 of 34
256
line-associated bloodstream infection (3.1% vs. 1.2%, p<0.01). PICC exchanges also
257
occurred more frequently among patients that had a prior central venous catheter or
258
PICC (28.0% vs. 21.1%, p<0.01) or an existing central line at the time of the exchange
259
(25.6% vs. 13.3%, p<0.01).
260
With respect to provider-characteristics, patients that underwent PICC
261
exchanges more often had their PICC placed by interventional radiology than other
262
providers (37.0% vs. 20.1%, p<0.01). Similarly, differences in device characteristics,
263
including catheter gauge, coating, and number of lumens were noted between those
264
that did vs. did not undergo an exchange (p<0.01 for all). Importantly, patients that
265
experienced exchanges more often had double and triple lumen PICCs than those that
266
did not (53.7% and 18.5% vs. 48.4% and 13.9%, p<0.01, respectively).
267 268
Unadjusted Association Between PICC Exchange, Patient-, Provider-, Device-
269
Factors and Thromboembolism
270
Unadjusted associations using Cox models between patient-, provider- and
271
device-characteristics, PICC exchange and thromboembolism are shown in Table 2.
272
Several comorbidities including body mass index (HR=0.99 [95%CI=0.98-1.00]), active
273
cancer (HR=1.88 [95%CI=1.34-2.64]), coagulopathy (HR=1.82 [95%CI=1.20-2.75]),
274
history of venous thrombosis within 30-days (HR=2.57 [95%CI=1.80-3.66]),
275
inflammatory bowel disease (HR=1.77 [95%CI=1.20-2.62]) and ICU stay (HR=1.97
276
[95%CI=1.63-2.38]) were associated with deep vein thrombosis. Additionally, patients
277
that experienced a deep vein thrombosis within 30-days (HR=2.65 [95%CI=1.87-3.77]),
278
had a central venous catheter or PICC in the preceding six months (HR=1.37
13
Page 13 of 34
279
[95%CI=1.08-1.75]) or had another central line at the time of PICC exchange (HR=1.88
280
[95%CI=1.52-2.32]) were more likely to experience thrombosis than those that did not.
281
Importantly, PICC exchange was strongly associated with deep vein thrombosis
282
on unadjusted analysis (HR=2.27 [95%CI=1.62-3.17]). Likewise, multi-lumen PICCs
283
were also associated with thrombosis (HR=2.67 [95%CI=2.12-3.37] and HR=3.38
284
[95%CI=2.40-4.76] for double and triple lumen vs. single lumen PICCs, respectively).
285 286
Adjusted Association Between PICC Exchange, Patient-, Provider-, Device-
287
Factors and Thrombosis
288
Following adjustment for patient-, provider- and device-characteristics, PICC
289
exchange remained associated with a nearly two-fold greater risk of thrombosis
290
(HR=1.98 [95%CI=1.37-2.85]) (Figure 2). The risk of thrombosis following PICC
291
exchange was second in magnitude to the device lumens, with double and triple-lumen
292
devices being associated with marginally greater risk of thrombosis (HR=2.06
293
[95%CI=1.59-2.66] and HR=2.31 [95%CI=1.6-3.33], respectively) than exchange. The
294
effect size of PICC exchange on risk of thrombosis was comparable to that of deep vein
295
thrombosis within 30-days (HR=1.97 [95%CI=1.32-2.96]). Active cancer (HR=1.68
296
[95%CI=1.15-2.46]), presence of another central venous catheter or PICC (HR=1.24
297
[95%CI=1.00-1.53]), and elevated WBC count (HR=1.41 [95%CI=1.17-1.71 for > 12,000
298
vs. < 12,000]) were also associated with deep vein thrombosis; ICU status, however,
299
was not significantly associated with thrombosis in the adjusted Cox model (HR 1.48
300
(95%CI: 0.86-2.53, p=0.16)
301
14
Page 14 of 34
302
DISCUSSION
303
Exchange of central venous catheters and PICCs over guidewires is often
304
necessary in the setting of device dislodgement, malfunction or occlusion. While
305
concerns regarding infection risk with this practice have been raised in neonates and
306
those receiving dialysis,8, 9 to date – no study has examined whether the practice might
307
be associated with thrombosis. In this cohort study, we found that PICC exchanges
308
were strongly and independently associated with thrombosis after adjusting for other
309
putative risk factors. The magnitude of the hazard of thrombosis in patients with
310
exchanges was comparable to that of the most widely recognized risk factors for PICC
311
thrombosis – number of lumens and catheter size.21, 22 This finding has important
312
clinical and policy implications as some factors leading to exchange might be
313
preventable – in turn, lowering risk of thrombosis. Furthermore, as exchanges are
314
clearly not innocuous, closer attention to risks and benefits of exchanging PICCs over
315
guidewires now appears necessary.
316
Why might catheter exchange increase the risk of venous thrombosis? Several
317
explanations exist. For example, dislodged or occluded catheters may injure the vessel
318
intima, predisposing to inflammation and venous stasis. Threading a wire in this setting
319
may further exacerbate injury as the wire is left in direct contact with the vessel during
320
device replacement. Alternatively, the act of removing and threading a PICC through the
321
vascular tree through smaller diameter veins may itself cause vessel injury and
322
thrombosis. Mechanistic studies using biomarkers to measure inflammation or
323
advanced imaging modalities to monitor vessel condition are needed to understand the
324
biology of this phenomenon.
15
Page 15 of 34
325
Although exchanges were relatively infrequent in this study, rates vary from 2%
326
to 20% in the literature.8, 23-25 Notably, PICC occlusion and dislodgement were among
327
the two most common causes that led to catheter exchange in our cohort. These events
328
are relevant because they may be preventable with better catheter care and
329
maintenance. For instance, close attention to measuring exposed catheter length
330
(especially in disoriented or delirious patients) could help address dislodgement before
331
it becomes necessary to exchange the catheter. Alternatively, newer technology such
332
as subcutaneous anchoring devices may help avoid exchange.26, 27 Similarly, preventing
333
catheter occlusion through meticulous attention to flushing and better device selection
334
(greater use of single lumen devices) is important.28 Of note, it is pertinent to emphasize
335
that occlusion itself does not necessarily indicate thrombosis, as occlusion more
336
frequently represents intraluminal clot, precipitation of medications, or fibrin sheath
337
around the catheter tip than deep vein thrombosis.29 Addressing occlusion might help
338
avoid a substantial number of exchanges, which in turn might prevent thrombosis.
339
However, some clinical conditions prompting exchange might not be avoidable.
340
For example, we observed that patients with recent deep vein thrombosis and
341
bloodstream infection experienced more frequent exchanges than those without such
342
diagnoses. While this finding may reflect therapeutic decisions related to managing
343
these conditions, some exchanges in these cases may be inappropriate. For example,
344
removal or exchange of functional PICCs in the setting of thrombosis provided the
345
device is still clinically necessary is not recommended by guidelines,30 yet often occurs
346
in real world settings.4, 31 Similarly, exchanging devices for antimicrobial central venous
347
catheters when treating line infections is gaining popularity as a novel strategy,32 but is
16
Page 16 of 34
348
recommended without regard to thrombotic risk. Understanding and balancing
349
thrombotic and infectious risks from exchange thus appears relevant in these decisions.
350
Our study has limitations. First, this is an observational study and can only
351
assess association, not causality. Relatedly, we did not sample consecutive patients –
352
but rather included those that met eligibility criteria within defined abstractor workloads.
353
Thus, selection bias in our sampling may affect our results. Second, while we adjust for
354
multiple known confounders using a validated conceptual model, we cannot be certain
355
that our findings are not confounded by unmeasured variables. This is particularly
356
relevant when patients entered the database as an exchange with limited data of the
357
prior device available. Furthermore, because patient follow-up was terminated at the
358
time of PICC removal, differential follow-up times and ascertainment of VTE between
359
those with vs. those without an exchange is possible. Third, although we track
360
indication for PICC exchange, we cannot determine whether these were performed for
361
clinically appropriate reasons. Relatedly, alternatives to PICC exchange are not readily
362
apparent through this study. Future studies assessing these aspects are necessary to
363
determine preventability of thrombosis in the setting of catheter exchange. Fourth,
364
because the HMS consortium is focused on medical patients, whether our findings hold
365
true for other patient groups (e.g., surgical and cancer patients) is not known. However,
366
given the increased risk of thrombosis in these subsets, caution in use and exchange of
367
PICCs seems reasonable.
368
Our study also has strengths. First, to our knowledge, this is among the first
369
studies to report an association between PICC exchange and thrombosis. Our findings
370
have important clinical and policy recommendations and suggest, at the very least, that
17
Page 17 of 34
371
more careful consideration of the appropriateness of exchanging a PICC is necessary.
372
Second, as almost half of all exchanges occurred in the setting of catheter malfunction
373
or dislodgement, our study highlights the importance of care, securement and
374
maintenance of PICCs. Third, our findings add to the growing literature regarding
375
increasing thrombogenicity of PICCs. Now, more than ever, use of appropriateness
376
criteria to guide insertion of PICCs is necessary to prevent inappropriate PICC insertion
377
and subsequent complications.33 As more patients that experienced exchanges had
378
double and triple lumen PICCs and as increasing lumens are associated with occlusion,
379
using the fewest number of PICC lumens to prevent thrombosis, occlusion and
380
exchange all appears necessary.
381
In conclusion, exchange of PICCs over guidewires may not be benign and is
382
associated with a nearly two-fold greater risk of venous thrombosis. Continued focus on
383
ensuring PICCs are placed for appropriate reasons, cared for with the highest
384
standards, and not exchanged unless necessary is warranted. A randomized trial of
385
exchange vs. new PICC placement – with longitudinal assessment of biological markers
386
of thrombosis - appears necessary.
387 388 389
18
Page 18 of 34
390 391 392
BIBLIOGRAPHY 1.
Greene MT, Flanders SA, Woller SC, Bernstein SJ, Chopra V. The Association
393
Between PICC Use and Venous Thromboembolism in Upper and Lower
394
Extremities. Am J Med. 2015;128:986-993 e981.
395
2.
Chopra V, Anand S, Hickner A, et al. Risk of venous thromboembolism
396
associated with peripherally inserted central catheters: a systematic review and
397
meta-analysis. Lancet. 2013;382:311-325.
398
3.
Hoshal VL, Jr. Total intravenous nutrition with peripherally inserted silicone
399
elastomer central venous catheters. Archives of surgery (Chicago, Ill. : 1960).
400
1975;110:644-646.
401
4.
Chopra V, Kuhn L, Flanders SA, Saint S, Krein SL. Hospitalist experiences,
402
practice, opinions, and knowledge regarding peripherally inserted central
403
catheters: results of a national survey. J Hosp Med. 2013;8:635-638.
404
5.
Gibson C, Connolly BL, Moineddin R, Mahant S, Filipescu D, Amaral JG.
405
Peripherally inserted central catheters: use at a tertiary care pediatric center.
406
Journal of vascular and interventional radiology : JVIR. 2013;24:1323-1331.
407
6.
central routes for central venous cannulation. Anaesthesia. 2012;67:65-71.
408 409
Pikwer A, Akeson J, Lindgren S. Complications associated with peripheral or
7.
Song L, Li H. Malposition of peripherally inserted central catheter: experience
410
from 3,012 patients with cancer. Experimental and therapeutic medicine.
411
2013;6:891-893.
19
Page 19 of 34
412
8.
McCoy M, Bedwell S, Noori S. Exchange of peripherally inserted central
413
catheters is associated with an increased risk for bloodstream infection. Am J
414
Perinatol. 2011;28:419-424.
415
9.
Guttmann DM, Trerotola SO, Clark TW, et al. Malfunctioning and infected
416
tunneled infusion catheters: over-the-wire catheter exchange versus catheter
417
removal and replacement. Journal of vascular and interventional radiology :
418
JVIR. 2011;22:642-646; quiz 646.
419
10.
Greene MT, Spyropoulos AC, Chopra V, et al. Validation of Risk Assessment
420
Models of Venous Thromboembolism in Hospitalized Medical Patients. Am J
421
Med. 2016;129:1001 e1009-1001 e1018.
422
11.
Grant PJ, Greene MT, Chopra V, Bernstein SJ, Hofer TP, Flanders SA.
423
Assessing the Caprini Score for Risk Assessment of Venous Thromboembolism
424
in Hospitalized Medical Patients. Am J Med. 2016;129:528-535.
425
12.
Herc E, Patel P, Washer LL, Conlon A, Flanders SA, Chopra V. A Model to
426
Predict Central-Line-Associated Bloodstream Infection Among Patients With
427
Peripherally Inserted Central Catheters: The MPC Score. Infect Control Hosp
428
Epidemiol. 2017;38:1155-1166.
429
13.
Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining
430
comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care.
431
2005;43:1130-1139.
432
14.
Chopra V, Anand S, Krein SL, Chenoweth C, Saint S. Bloodstream infection,
433
venous thrombosis, and peripherally inserted central catheters: reappraising the
434
evidence. Am J Med. 2012;125:733-741.
20
Page 20 of 34
435
15.
Chopra V, Ratz D, Kuhn L, Lopus T, Chenoweth C, Krein S. PICC-associated
436
bloodstream infections: prevalence, patterns, and predictors. Am J Med.
437
2014;127:319-328.
438
16.
Pongruangporn M, Ajenjo MC, Russo AJ, et al. Patient- and device-specific risk
439
factors for peripherally inserted central venous catheter-related bloodstream
440
infections. Infect Control Hosp Epidemiol. 2013;34:184-189.
441
17.
Lee EW, Wei LJ, Amato DA, Leurgans S. Cox-Type Regression Analysis for
442
Large Numbers of Small Groups of Correlated Failure Time Observations. .
443
Survival Analysis: State of the Art. Nato Science (Series E: Applied Sciences).
444
1992;211.
445
18.
Rubin D. Multiple imputation in sample surveys-a phenomenological Bayesian
446
approach to nonresponse. Journal of the American Statistical Association, .
447
1978;20:20-34.
448
19.
peripherally inserted central catheters. Chest. 2010;138:803-810.
449 450
Evans RS, Sharp JH, Linford LH, et al. Risk of symptomatic DVT associated with
20.
Chopra V, Ratz D, Kuhn L, Lopus T, Lee A, Krein S. Peripherally inserted central
451
catheter-related deep vein thrombosis: contemporary patterns and predictors. J
452
Thromb Haemost. 2014;12:847-854.
453
21.
O'Brien J, Paquet F, Lindsay R, Valenti D. Insertion of PICCs with minimum
454
number of lumens reduces complications and costs. Journal of the American
455
College of Radiology : JACR. 2013;10:864-868.
456 457
22.
Evans RS, Sharp JH, Linford LH, et al. Reduction of peripherally inserted central catheter-associated DVT. Chest. 2013;143:627-633.
21
Page 21 of 34
458
23.
Gnannt R, Patel P, Temple M, et al. Peripherally Inserted Central Catheters in
459
Pediatric Patients: To Repair or Not Repair. Cardiovasc Intervent Radiol.
460
2017;40:845-851.
461
24.
Trerotola SO, Thompson S, Chittams J, Vierregger KS. Analysis of tip
462
malposition and correction in peripherally inserted central catheters placed at
463
bedside by a dedicated nursing team. Journal of vascular and interventional
464
radiology : JVIR. 2007;18:513-518.
465
25.
Leroyer C, Lasheras A, Marie V, et al. Prospective follow-up of complications
466
related to peripherally inserted central catheters. Med Mal Infect. 2013;43:350-
467
355.
468
26.
Zerla PA, Canelli A, Cerne L, et al. Evaluating safety, efficacy, and cost-
469
effectiveness of PICC securement by subcutaneously anchored stabilization
470
device. J Vasc Access. 2017;18:238-242.
471
27.
Nurs. 2014;23:S12, S14-18.
472 473
Elen Hughes M. Reducing PICC migrations and improving patient outcomes. Br J
28.
Smith SN, Moureau N, Vaughn VM, et al. Patterns and Predictors of Peripherally
474
Inserted Central Catheter Occlusion: The 3P-O Study. Journal of vascular and
475
interventional radiology : JVIR. 2017;28:749-756 e742.
476
29.
Baskin JL, Pui CH, Reiss U, et al. Management of occlusion and thrombosis
477
associated with long-term indwelling central venous catheters. Lancet.
478
2009;374:159-169.
479 480
30.
Kearon C, Akl EA, Comerota AJ, et al. Antithrombotic therapy for VTE disease: Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College
22
Page 22 of 34
481
of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest.
482
2012;141:e419S-e496S.
483
31.
Chopra V, Kuhn L, Coffey CE, Jr., et al. Hospitalist experiences, practice,
484
opinions, and knowledge regarding peripherally inserted central catheters: a
485
Michigan survey. J Hosp Med. 2013;8:309-314.
486
32.
Zakhour R, Chaftari AM, Raad, II. Catheter-related infections in patients with
487
haematological malignancies: novel preventive and therapeutic strategies.
488
Lancet Infect Dis. 2016;16:e241-e250.
489
33.
Chopra V, Flanders SA, Saint S, et al. The Michigan Appropriateness Guide for
490
Intravenous Catheters (MAGIC): Results From a Multispecialty Panel Using the
491
RAND/UCLA Appropriateness Method. Ann Intern Med. 2015;163:S1-40.
492 493 494 495
23
Page 23 of 34
Figure 1. Patient Flow Diagram
Patients identified as having received a PICC in 51 HMS Hospitals N=23,010
Patients without exchange N=22,421 (97.4%)
Patients without image confirmed, symptomatic VTE N=21,962 (98.0%)
Patients with image confirmed, symptomatic VTE N=459 (2.0%)
Patients with exchange N=589 (2.6%)
Patients without image confirmed, symptomatic VTE N=568 (96.4%)
Patients with image confirmed, symptomatic VTE N=21 (3.6%)
Legend: PICC=Peripherally inserted central catheter; VTE=venous thromboembolism
24
Page 24 of 34
Figure 2. Multivariable Cox Model for Time to Venous Thromboembolism following PICC Exchange
25
Page 25 of 34
Table 1. General Characteristics of Patients with and without PICC exchange (n=23,010) Category/Variable
Modifier
Patient Characteristics Male Gender
No exchange (n=22,421)
Exchange (n=589)
p
10985 (49.0%)
320 (54.3%)
0.01
Race
White
16647 (76.1%)
391 (68.6%)
<0.01
Age group
>65 years
11065 (49.4%)
257 (43.6%)
<0.01
Body Mass Index (BMI)
Median (IQR)
28.7 (24.0-35.2)
29.4 (24.2-36.5)
0.16
8282 (36.9%)
197 (33.4%)
0.08
15350 (68.5%)
387 (65.7%)
0.16
3417 (15.2%)
76 (12.9%)
0.12
6448 (28.8%)
175 (29.7%)
0.61
3433 (15.3%)
77 (13.1%)
0.14
3409 (15.2%)
79 (13.4%)
0.23
1917 (8.6%)
51 (8.7%)
0.93
6083 (27.1%)
155 (26.3%)
0.66
1377 (6.1%)
39 (6.6%)
0.63
859 (3.8%)
22 (3.7%)
0.90
4590 (20.5%)
119 (20.2%)
0.87
4387 (19.6%)
132 (22.4%)
0.09
7782 (34.7%)
201 (34.1%)
0.77
128 (0.6%)
6 (1.0%)
0.16
771 (3.4%)
18 (3.1%)
0.61
39 (0.2%)
3 (0.5%)
0.06
1073 (4.8%)
29 (4.9%)
0.88
Mild liver disease
1469 (6.6%)
38 (6.5%)
0.92
Moderate to severe liver diseasea Known HIV or AIDSa
844 (3.8%)
32 (5.4%)
0.04
156 (0.7%)
6 (1.0%)
0.35
History of Cancer
5247 (23.4%)
157 (26.7%)
0.07
Active cancer
1378 (6.1%)
53 (9.0%)
<0.01
799 (3.6%)
19 (3.2%)
0.66
3 (1-5)
3 (2-5)
0.05
Hyperlipidemia
a
a
Hypertension
Myocardial Infarction
b b
Congestive Heart Failure
Peripheral vascular disorders
a
a
Cerebrovascular disease a
Dementia b
COPD
Rheumatoid arthritis
b
Peptic Ulcer Disease
a
Diabetes without complicationsa Diabetes with complicationsa Renal Failure
a
Kidney transplant Hemodialysis
a
Peritoneal Dialysis
a
Hemi- or paraplegia
b
a
a
Coagulopathy
Charlson/Deyo Comorbidity Index
Median (IQR)
26
Page 26 of 34
History of CLABSI
259 (1.2%)
18 (3.1%)
<0.01
647 (2.9%)
28 (4.8%)
0.01
2422 (10.8%)
73 (12.4%)
340 (1.5%)
14 (2.4%)
1294 (5.8%)
37 (6.3%)
874 (3.9%)
36 (6.1%)
2893 (12.9%)
87 (14.8%)
675 (3.0%)
15 (2.5%)
0.51
9044 (40.3%)
257 (43.6%)
0.11
7045 (31.4%)
162 (27.5%)
0.04
6475 (28.9%)
165 (28.0%)
0.65
8158 (36.4%)
207 (35.1%)
0.54
3837 (17.1%)
85 (14.4%)
0.09
1460 (6.5%)
38 (6.5%)
0.95
12976 (57.9%)
337 (57.2%)
0.75
7978 (35.6%)
210 (35.7%)
0.97
Aspirin
7119 (31.8%)
189 (32.1%)
0.86
Other antiplatelet therapy
3044 (13.6%)
76 (12.9%)
0.64
Anticoagulant therapy
15089 (67.3%)
431 (73.2%)
<0.01
History of DVT
Within 30 days Prior history
History of PE
History of any VTE Event
Inflammatory Bowel Diseaseb
Within 30 days Positive history Within 30 days Positive history
b
Serious lung disease
Life-threatening illness
b
b
Pneumonia b
Sepsis
History of Prior CVA/TIA Venous stasis
a
Smoking Status Statin
Current/ Former
0.21
<0.01
eGFR
Median (IQR)
60 (47-90)
60 (54-89)
0.59
WBC
Median (IQR)
9.2 (6.7-12.9)
9.0 (6.6-12.5)
0.16
Hemoglobin
Median (IQR)
10.2 (8.8-11.8)
9.6 (8.3-11.1)
<0.01
Platelets
Median (IQR)
226 (160-309)
234 (157-319)
0.29
INR
Median (IQR)
1.15 (1.03-1.39)
1.20 (1.05-1.40)
0.07
CVC or PICC in prior 6 months
4736 (21.1%)
302 (51.3%)
<0.01
Presence of another CVC
2982 (13.3%)
151 (25.6%)
<0.01
15006 (66.9%)
292 (49.6%)
<0.01
4496 (20.1%)
218 (37.0%)
Provider Characteristics Operator Type
Vascular Access Nurse Interventional Radiologist
27
Page 27 of 34
Physician
237 (1.1%)
4 (0.7%)
2682 (12.0%)
75 (12.7%)
6788 (30.3%)
176 (29.9%)
0.84
>1
2547 (11.6%)
19 (3.4%)
<0.01
Arm selected for insertion
Right Arm
15743 (70.3%)
398 (67.6%)
0.16
Vein selected for insertion
Basilic
13657 (60.9%)
328 (55.7%)
<0.01
Brachial
6910 (30.8%)
171 (29.0%)
Cephalic
1158 (5.2%)
21 (3.6%)
Other
696 (3.1%)
69 (11.7%)
4 French
6647 (31.1%)
134 (23.3%)
5 French
12844 (60.1%)
365 (63.4%)
1885 (8.8%)
77 (13.4%)
20352 (90.8%)
503 (85.4%)
<0.01
Antimicrobial coating
1578 (7.0%)
12 (2.0%)
<0.01
Anti-thrombotic coating
390 (1.7%)
6 (1.0%)
0.18
5567 (24.8%)
124 (21.1%)
0.04
Single
8431 (37.7%)
163 (27.9%)
<0.01
Double
10841 (48.4%)
314 (53.7%)
Triple/Quad
3106 (13.9%)
108 (18.5%)
ICU prior to or at time of exchange Placement attempts
Device characteristics PICC Gauge
Other
> 6 French Power PICC
Valved PICC Number of PICC lumens
<0.01
a
at time of initial PICC placement; bwithin 30 days of PICC placement;
Legend: IQR=interquartile range; COPD=chronic obstructive pulmonary disease; HIV=human immunodeficiency virus; AIDS=acquired immune deficiency syndrome; CLABSI=central line-associated bloodstream infection; DVT=deep vein thrombosis; PE=pulmonary embolism; VTE=venous thromboembolism; CVA=cerebrovascular accident; TIA=transient ischemic attack; eGFR=estimated glomerular filtration rate; LOS=length of stay; CVC=central venous catheter; PICC=peripherally inserted central catheter; TPN=total parenteral nutrition; ICU=intensive care unit. Definitions: Diabetes without complications = diabetes without documented retinopathy, nephropathy, neuropathy, cardio- or cerebrovascular events.
28
Page 28 of 34
29
Page 29 of 34
Table 2. Unadjusted associations of PICC Exchange, Patient-, Provider and Device-Factors with PICC-DVT (n=23,010) Category/Variable
Modifier HR (95% CI)
P-value
PICC Exchange
Yes vs. No
2.27 (1.62, 3.17)
<0.01
Male Gender
Male vs. Female White vs. Other >65 vs. <64
0.98 (0.80, 1.19)
0.81
0.81 (0.60, 1.09)
0.16
1.05 (0.87, 1.26)
0.63
Per unit increase Yes vs. No
0.99 (0.98, 1.00)
0.02
0.94 (0.72, 1.22)
0.63
Yes vs. No
0.96 (0.81, 1.15)
0.67
Yes vs. No
0.75 (0.57, 0.98)
0.04
Yes vs. No
0.83 (0.69, 1.01)
0.07
Yes vs. No
0.71 (0.55, 0.92)
0.01
Yes vs. No
1.10 (0.92, 1.31)
0.28
Yes vs. No
1.04 (0.72, 1.49)
0.85
Yes vs. No
1.05 (0.82, 1.33)
0.70
Yes vs. No
0.89 (0.58, 1.37)
0.60
Yes vs. No
1.46 (1.04, 2.05)
0.03
Yes vs. No
0.87 (0.70, 1.07)
0.19
Yes vs. No
0.87 (0.70, 1.08)
0.22
Yes vs. No
0.90 (0.74, 1.09)
0.29
Yes vs. No
1.11 (0.47, 2.61)
0.81
Yes vs. No
1.50 (0.97, 2.32)
0.07
Yes vs. No
1.14 (0.76, 1.71)
0.54
Yes vs. No
0.77 (0.53, 1.12)
0.18
Yes vs. No
0.78 (0.51, 1.19)
0.25
Known HIV or AIDS
Yes vs. No
0.29 (0.04, 2.11)
0.22
History of Cancer
Yes vs. No
1.27 (1.05, 1.55)
0.012
Active cancer
Yes vs. No
1.88 (1.34, 2.64)
<0.01
Coagulopathy
Yes vs. No
1.82 (1.20, 2.75)
<0.01
Charlson/Deyo Comorbidity Index
Per unit
0.98 (0.95, 1.01)
0.17
Race Age group Body Mass Index (BMI) Hyperlipidemiaa a
Hypertension
Myocardial Infarction
b b
Congestive Heart Failure
Peripheral vascular disorders
a
a
Cerebrovascular disease a
Dementia b
COPD
Rheumatoid arthritis
b
Peptic Ulcer Disease
a
Diabetes without complications Diabetes with complications Renal Failure
a
a
a
Kidney transplant Hemodialysis
a
Hemi- or paraplegia
b
a
Mild liver disease
Moderate to severe liver disease
a
a
a
30
Page 30 of 34
increase Yes vs. No
History of CLABSI
1.07 (0.56, 2.04)
0.85
Within 30 days vs. none Prior history vs. none Within 30 days vs. none Positive history vs. none Within 30 days vs. none Positive history vs. none Yes vs. No
2.65 (1.87, 3.77)
<0.01
1.77 (1.20, 2.62)
<0.01
Yes vs. No
1.35 (1.09, 1.67)
<0.01
Yes vs. No
1.97 (1.63, 2.38)
<0.01
Yes vs. No
1.18 (0.98, 1.44)
0.09
Yes vs. No
1.02 (0.81, 1.28)
0.88
Yes vs. No
1.05 (0.85, 1.31)
0.64
Yes vs. No
0.70 (0.44, 1.11)
0.13
1.01 (0.84, 1.22)
0.88
Statin
Current/former vs. Never Yes vs. No
0.85 (0.70, 1.03)
0.10
Aspirin
Yes vs. No
0.68 (0.57, 0.80)
<0.01
Other antiplatelet therapy
Yes vs. No
0.81 (0.64, 1.03)
0.08
Anticoagulant therapy
Yes vs. No
0.88 (0.74, 1.05)
0.15
EGFR
Per unit increase
1.00 (0.99, 1.00)
0.19
WBC
Per unit increase
1.01 (1.00, 1.01)
<0.01
Hemoglobin
Per unit increase
0.96 (0.92, 1.01)
0.11
Platelets
Per unit increase
1.00 (1.00, 1.00)
0.10
INR
Per unit increase
1.00 (0.99, 1.00)
0.11
CVC or PICC in prior 6 months
Yes vs. No
1.39 (1.10, 1.76)
<0.01
Presence of another CVC
Yes vs. No
1.88 (1.52, 2.32)
<0.01
History of DVT
History of PE
History of any VTE Event
Inflammatory Bowel Diseaseb b
Serious lung disease
Life-threatening illness
b
b
Pneumonia b
Sepsis
History of Prior CVA/TIA Venous stasis
a
Smoking Status
1.84 (1.51, 2.24) 1.56 (0.82, 3.00)
<0.01
1.65 (1.28, 2.14) 2.57 (1.80, 3.66)
<0.01
1.79 (1.47, 2.19)
31
Page 31 of 34
Operator Type
ICU prior to or at time of exchange
Interventional Radiologist vs. Vascular Access Nurse Physician vs. Vascular Access Nurse Advanced Practice Prof vs. Vascular Access Nurse Other vs. Vascular Access Nurse Yes vs. No
Placement attempts
>1 vs. 1
Arm selected for insertion Vein selected for insertion
1.33 (0.84, 2.11)
0.02
0.57 (0.20, 1.67) 1.53 (1.02, 2.30)
0.46 (0.22, 0.95) 2.08 (1.74, 2.49)
<0.01
12.12 (1.63, 89.83)
0.01
Right vs. Left
0.98 (0.79, 1.22)
0.87
Brachial vs. Basilic Cephalic vs. Basilic Other vs. Basilic
1.26 (1.02, 1.55)
<0.01
5 Fr. vs 4 Fr.
2.17 (1.67, 2.82)
<0.01
6Fr vs. 4 Fr.
2.90 (1.83, 4.59)
<0.01
Power PICC
Yes vs. No
0.84 (0.64, 1.12)
0.24
Antimicrobial coating
Yes vs. No
1.85 (1.28, 2.67)
<0.01
Anti-thrombotic coating
Yes vs. No
1.12 (0.53, 2.40)
0.77
Valved PICC
Yes vs. No
0.91 (0.62, 1.34)
0.64
Number of PICC lumens
Double vs. Single Triple/Quad vs. Single vs. Single
2.67 (2.12, 3.37)
<0.01
Device characteristics PICC Gauge
0.63 (0.40, 0.98) 1.13 (0.77, 1.66)
3.38 (2.40, 4.76)
Legend: COPD=chronic obstructive pulmonary disease; HIV=human immunodeficiency virus; AIDS=acquired immune deficiency syndrome; CLABSI=central line-associated bloodstream infection; DVT=deep vein thrombosis; PE=pulmonary embolism; VTE=venous thromboembolism; CVA=cerebrovascular accident; TIA=transient ischemic attack; eGFR=estimated glomerular filtration rate; LOS=length of stay; CVC=central venous catheter; PICC=peripherally inserted central catheter; TPN=total parenteral nutrition; ICU=intensive care unit;
32
Page 32 of 34
Table 3. Multivariable Cox Model for Upper Extremity Deep Vein Thrombosis and Pulmonary Embolism Variable Exchange VTE History
Active cancer Presence of another CVC/PICC Lumens
WBC Life-threatening illness CVC/PICC in previous 6 months Operator Type
ICU Status Placement attempts Antimicrobial coated Insertion vein
HR (95% CI) 1.98 (1.37, 2.85) Positive history vs. No history 1.87 (1.51, 2.33) Within previous 30 days vs. No history 1.97 (1.32, 2.96) Yes vs. No 1.68 (1.15, 2.46) Yes vs. No 1.24 (1.00, 1.53) Double vs. Single 2.06 (1.59, 2.66) Triple/Quad vs. Single 2.31 (1.60, 3.33) >12 vs. <=12 1.41 (1.17, 1.71) Yes vs. No 1.28 (1.00, 1.65) Yes vs. No 1.03 (0.83, 1.28) Interventional Radiologist vs. Vascular 0.70 (0.24, Access Nurse 2.10) Physician vs. Vascular Access Nurse 1.40 (1.11, 1.78) Advance Practice Professional vs. 0.39 (0.20, Vascular Access Nurse 0.78) Other vs. Vascular Access Nurse 0.40 (0.20, 0.78) Yes vs. No 1.48 (0.86 – 2.53) >1 vs. Only 1 0.91 (0.68, 1.24) Yes vs. No 1.54 (1.22, 1.95) Brachial vs. Basilic 1.18 (0.99, 1.40) Cephalic vs. Basilic 0.60 (0.39, 0.92) Other vs. Basilic 0.85 (0.58, 1.24) Yes vs. No
P-value 0.0003 <.0001
0.0077 0.0529 <.0001
0.0004 0.0523 0.7793 0.0003
0.16 0.5641 0.0003 0.0121
Legend: CI=confidence interval; VTE=venous thromboembolism; PICC=peripherally inserted central catheter; CVC=central venous catheter
33
Page 33 of 34
1 2 3 4 5
34
Page 34 of 34
S0002-9343(18)30089-5 https://doi.org/10.1016/j.amjmed.2018.01.017 AJM 14500
To appear in:
The American Journal of Medicine
Please cite this article as: Vineet Chopra, Scott Kaatz, Paul Grant, Lakshmi Swaminathan, Tanya Boldenow, Anna Conlon, Steven J. Bernstein, Scott A. Flanders, Risk of Venous Thromboembolism Following Peripherally Inserted Central Catheter Exchange: an Analysis of 23,000 Hospitalized Patients, The American Journal of Medicine (2018), https://doi.org/10.1016/j.amjmed.2018.01.017. 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.
1 2 3
Risk of Venous Thromboembolism following Peripherally Inserted Central
4
Catheter Exchange: An Analysis of 23,000 Hospitalized Patients
5
Running Title: PICC Exchange and Deep Vein Thrombosis
6 7
Vineet Chopra MD, MSc (a, b, c)
8
Scott Kaatz DO, MSc (d)
9
Paul Grant MD (a, c)
10
Lakshmi Swaminathan MD (e)
11
Tanya Boldenow MD (f)
12
Anna Conlon, PhD (a, b)
13
Steven J. Bernstein MD, MPH (b, c, g)
14
Scott A. Flanders MD (a, c)
15 16
From: (a) The Division of Hospital Medicine, Department of Internal Medicine,
17
University of Michigan Medical School, Ann Arbor, MI; (b) Patient Safety Enhancement
18
Program and Center for Clinical Management Research, VA Ann Arbor Health Care
19
System, Ann Arbor, MI (c); the Michigan Hospital Medicine Safety Consortium, Ann
20
Arbor, MI (d) Henry Ford Health System, Detroit, MI; (e) Beaumont Hospital, Dearborn,
21
MI; (f) St. Josephs Health Center, Ypsilanti, MI and (g) The Division of General
22
Medicine, Department of Internal Medicine, University of Michigan Medical School, Ann
1
Page 1 of 34
23
Arbor, MI.
24 25
Abstract Word Count: 262
26
Manuscript Word Count: 3100
27 28
Conflicts of Interest: Letters from all authors indicating COIs are attached. All authors
29
had access to the data and a role in writing the manuscript.
30 31
Funding: Support for HMS is provided by Blue Cross and Blue Shield of Michigan and
32
Blue Care Network as part of the BCBSM Value Partnerships program. Although Blue
33
Cross Blue Shield of Michigan and HMS work collaboratively, the opinions, beliefs and
34
viewpoints expressed by the author do not necessarily reflect the opinions, beliefs and
35
viewpoints of BCBSM or any of its employees. Dr. Chopra is supported by a career
36
development award from AHRQ (1-K08HS022835-01).
37 38
Role of Study Sponsor: Blue Cross / Blue Shield of Michigan and Blue Care Network
39
supported data collection at each participating site and funded the data coordinating
40
center but had no role in study concept, interpretation of findings, or in the preparation,
41
final approval or decision to submit the manuscript.
42 43
Keywords: Peripherally inserted central catheter; PICC; exchange, deep vein
44
thrombosis; central venous catheter; thromboembolism.
45
2
Page 2 of 34
46
Address for Correspondence:
47
Vineet Chopra MD, MSc
48
2800 Plymouth Road, Building 16, #432W
49
Ann Arbor, MI 48109
50
[email protected]
51
2800 Plymouth Road Building 16 #432W
52
Ann Arbor, MI 48109
53
[email protected]
54 55 56 57
CLINICAL SIGNIFICANCE
following catheter occlusion or dislodgement.
58 59
62
Patients that underwent a catheter exchange experienced deep vein thrombosis more often and more quickly than those that did not.
60 61
Exchange of peripherally inserted central catheters is often performed
The risk of thrombosis following exchange was similar to that of using a larger catheter with more lumens.
63 64 65
3
Page 3 of 34
66
ABSTRACT
67
BACKGROUND. Catheter exchange over a guidewire is frequently performed for
68
malfunctioning peripherally inserted central catheters (PICCs). Whether such
69
exchanges are associated with venous thromboembolism is not known.
70 71
METHODS.
We performed a retrospective cohort study to assess the association between
72
PICC exchange and risk of thromboembolism. Adult hospitalized patients that received
73
a PICC during clinical care one of 51 hospitals participating in the Michigan Hospital
74
Medicine Safety consortium were included. The primary outcome was hazard of
75
symptomatic venous thromboembolism (radiographically confirmed upper-extremity
76
deep vein thrombosis and pulmonary embolism) in those that underwent PICC
77
exchange vs. those that did not.
78 79
RESULTS. Of 23,010 patients that underwent PICC insertion in the study, 589 patients
80
(2.6%) experienced a PICC exchange. Almost half of all exchanges were performed for
81
catheter dislodgement or occlusion. A total of 480 patients (2.1%) experienced PICC-
82
associated deep vein thrombosis. The incidence of deep vein thrombosis was greater in
83
those that underwent PICC exchange vs. those that did not (3.6% vs. 2.0%, p<0.001).
84
Median time to thrombosis was shorter among those that underwent exchange
85
compared vs. those that did not (5 vs. 11 days, p=0.02). Following adjustment, PICC
86
exchange was independently associated with two-fold greater risk of thrombosis
87
(hazard ratio [HR]=1.98, 95%CI=1.37-2.85) vs. no exchange. The effect size of PICC
88
exchange on thrombosis was second in magnitude to device lumens (HR=2.06
4
Page 4 of 34
89
[95%CI=1.59-2.66] and HR=2.31 [95%CI=1.6-3.33] for double- and triple lumen
90
devices, respectively).
91 92
CONCLUSION. Guidewire exchange of PICCs may be associated with increased risk of
93
thrombosis. As some exchanges may be preventable, consideration of risks and
94
benefits of exchanges in clinical practice is needed.
95 96
5
Page 5 of 34
97 98 99
INTRODUCTION Venous thromboembolism is an important complication associated with central venous catheters. Deep vein thrombosis and pulmonary embolism are especially
100
relevant in the case of peripherally inserted central catheters (PICCs),1 devices
101
associated with 2.5-fold greater risk of thrombosis than central venous catheters.2 Thus,
102
PICCs are associated with a higher baseline risk of thrombosis than traditional central
103
venous catheters.
104
PICCs are placed in smaller peripheral veins of the upper extremity rather than
105
large veins of the neck or chest. This unique route avoids important complications such
106
as pneumothorax or injury to the great vessels,3 and is an important reason behind
107
growing PICC use in clinical practice.4, 5 However, placement of PICCs in the arm also
108
means insertion into veins that are of smaller diameter compared to the chest.
109
Additionally, PICCs are longer and more slender than conventional catheters. Thus,
110
complications such as tip migration, catheter occlusion and device dislodgement occur
111
2-3 times more frequently with PICCs than central venous catheters.6, 7 When these
112
events occur, an over-the-wire guidewire exchange of the PICC is usually performed.
113
While such guidewire exchanges are sometimes necessary, they may not be risk free.
114
For example, one study reported increased risk of infection among neonates following
115
over-the-wire PICC exchanges.8 Similarly, in a study of hemodialysis recipients,
116
increased risk of dialysis catheter dysfunction was reported following guidewire
117
exchange.9 To date, however, no study has examined whether PICC exchanges are
118
associated with thrombosis. Given the known thrombogenicity of PICCs and the
6
Page 6 of 34
119
potential that exchanges may cause venous injury, this question is particularly relevant
120
for clinical practice.
121
Therefore, we used data from an ongoing study of hospitalized patients who
122
received PICCs to assess whether guidewire exchanges of PICCs were independently
123
associated with risk of thrombosis.
124 125
METHODS
126
Study Setting and Participants
127
The study was conducted using data from the Michigan Hospital Medicine Safety
128
(HMS) consortium, a 51-hospital collaborative quality initiative supported by Blue Cross
129
Blue Shield of Michigan and Blue Care Network. The design and setting of this
130
consortium have been previously described.1, 10, 11 In brief, adult patients admitted to a
131
general medicine ward or intensive care unit (ICU) of a participating hospital who
132
receive a PICC for any reason during clinical care are eligible for inclusion. Patients who
133
are (a) under the age of 18; (b) pregnant; (c) admitted to a non-medical service (e.g.,
134
general surgery); or (d) admitted under observation status are excluded from sampling.
135
At each hospital, dedicated, trained medical record abstractors use a defined
136
protocol to collect clinical data directly from medical records of patients. Based on
137
available resources and the time required to collect data from each case, patients with
138
PICCs are sampled on a 14-day cycle, with data from the first 17 cases that meet
139
eligibility criteria within each cycle included. To ensure adequate representation of
140
critically ill patients, at least 7 of the 17 eligible cases include PICCs placed in ICU
141
patients. We chose seven patients again based on the time required to abstract these
7
Page 7 of 34
142
cases (typically twice that of a patient with a PICC placed in a non-ICU setting). If a
143
hospital could not find 7 eligible patients that received PICCs in an ICU setting,
144
additional patients that received PICCs on general wards are used to backfill thus
145
ensuring 17 patients are included every two weeks. Because we were specifically
146
interested in risk of thrombosis following catheter exchange, patients with a history of
147
prior DVT (within six months), presumed VTE, palliative care status or direct ICU
148
admissions were not excluded unlike prior studies.12 All patients are followed until PICC
149
removal, death, or 70-days, whichever occurs first. To ensure completeness and
150
accuracy, staff from the University of Michigan perform annual on-site audits of all
151
participating hospitals to assess veracity and accuracy of abstracted data.
152
PICCs are defined as vascular access devices inserted in veins of the upper
153
extremity that terminate at the cavoatrial junction; thus, midlines, conventional catheters
154
or those placed in lower extremity veins are excluded. However, the presence of a
155
conventional central catheter in the lower extremity, chest or neck at the time of PICC
156
insertion is captured. Data regarding PICC characteristics (e.g., indication, gauge,
157
lumens, tip position verification) are obtained directly from vascular nursing or
158
interventional radiology notes or the order for PICC placement. All testing for thrombosis
159
occurs at the discretion of clinicians at each hospital. Since we were interested in
160
thrombosis following PICC exchange, both patients that underwent exchange of an
161
existing PICC (e.g., those with “index” PICCs) as well as those enrolled in the study
162
following a PICC exchange (e.g., patient admitted from an outside hospital with a non-
163
functioning PICC that required exchange) were included. When patients with an
164
exchange were enrolled, characteristics related to the exchanged PICC were used for
8
Page 8 of 34
165
analysis. For this analysis, data from patients enrolled between January 2013 and
166
October 2016 were included.
167 168
Covariates
169
Detailed medical history including comorbidities, physical examination findings,
170
laboratory and medication data were collected from the medical record. Standardized
171
definitions using ICD-10 and Elixhauser criteria were used to define comorbidities
172
present at hospital admission.13 Variables including age (<64 versus >65 years), sex,
173
race, body mass index, tobacco use (never, former, current), principal admitting
174
diagnosis, prior upper or lower-extremity deep vein thrombosis (within 30-days, beyond
175
30-days, never), inpatient surgery within 30 days of PICC placement, chemotherapy or
176
blood administration during hospitalization, trauma requiring hospitalization within 30
177
days, immobilizing cast at the time of PICC placement, hip or knee replacement within
178
30 days of PICC placement, presence of active infection, existing central venous
179
catheter when PICC was placed (yes/no), diabetes mellitus (uncomplicated vs
180
complicated by micro- or macrovascular complications), history of cerebrovascular
181
accident or transient ischemic attack, history of myocardial infarction, sickle cell
182
disease, venous thromboembolism prophylaxis (i.e., receipt of subcutaneous heparin
183
twice or thrice daily regimens or use of enoxaparin at prophylactic doses), receipt of
184
treatment dose anticoagulation for any reason, aspirin, statin, erythropoiesis stimulating
185
agents, and antiplatelet medication administration were abstracted directly from medical
186
records.
9
Page 9 of 34
187
Active cancer was defined as admission to the hospital for a cancer diagnosis
188
made within the past six months (e.g., acute leukemia) or for receipt of scheduled
189
chemotherapy for malignancy. Serious lung disease was defined as receipt of invasive
190
or non-invasive ventilation at any point during hospitalization. Life-threatening illness
191
was defined as care in an ICU-setting at admission or at any point during
192
hospitalization. Laboratory values including white blood cell (WBC) count, hemoglobin,
193
platelet count and international normalized ratio (INR) were collected at the time of
194
PICC insertion.
195 196 197
Ascertainment of Outcomes The primary outcome was acute, symptomatic, radiographically-confirmed,
198
upper-extremity deep vein thrombosis (e.g., compression or duplex ultrasound with
199
visible thrombus or non-compressibility of a deep vein of the arm [brachial, basilic,
200
axillary or subclavian]), or PE (e.g., high probability V/Q scan or positive contrast-
201
enhanced computerized tomography) occurring after initial PICC placement or after
202
catheter exchange. At all hospitals, testing for thromboembolism occurs only in the
203
presence of clinical symptoms (e.g., arm pain, swelling). Clinicians are at liberty of
204
ordering tests for these events based on clinical suspicion or findings.
205 206 207
Statistical Analyses Factors associated with thrombosis following catheter exchange were assessed
208
according to a previously published14 and validated conceptual model.15, 16 In modeling,
209
all factors known to be associated with deep vein thrombosis (e.g., elements from
10
Page 10 of 34
210
standardized risk scores) as well as those associated with risk of thromboembolism
211
following PICC placement from our prior studies were included. Descriptive variables
212
were used to summarize the distribution of patient-, provider- and device-factors
213
between patients with and without PICC exchange. Because PICC exchanges can
214
occur at any point during the dwell of the PICC (and we are interested in time-to-event
215
for deep vein thrombosis), Cox proportional hazards models with robust sandwich
216
covariance matrix estimates to account for hospital-level clustering were employed to
217
account for the time-dependent nature of exchanges.17 In addition to PICC exchange,
218
covariates that could change value during the dwell of the PICC (e.g., presence of
219
another central venous catheter, admission to ICU) were analyzed as time-dependent
220
variables and coded as null until such events occurred. Similarly, any device- or
221
provider-specific variables that could change value at the time of a PICC exchange
222
(e.g., number of PICC lumens), were also treated as time-dependent with values
223
changing if the variable was different for the exchanged device compared to the original
224
PICC. Unadjusted associations of covariates with PICC thrombosis were first expressed
225
as hazard ratios (HR) with corresponding 95% confidence intervals (CI). Missing data
226
were imputed through a 10-fold multiple imputation procedure.18
227
The final multivariable Cox model for the association of PICC exchange with
228
thrombosis was adjusted for factors previously reported as independent predictors of
229
PICC thrombosis,14, 19, 20 as well as potential confounding factors associated with both
230
deep vein thrombosis and probability of PICC exchange on unadjusted analyses. A two-
231
sided p-value of less than 0.05 was used to indicate significance in all analyses. All
232
analyses were performed in SAS, version 9.4 (SAS Institute Inc., Cary, NC) and R
11
Page 11 of 34
233
version 3.2.4.
234 235 236 237
Ethical and Regulatory Oversight The University of Michigan Medical School’s Institutional Review Board reviewed this study and it received a “Not Regulated” status.
238 239
RESULTS
240
Patient Characteristics and Outcomes
241
During the study period, 23,010 patients underwent PICC insertion in 51
242
Michigan hospitals and were included in this analysis. Of these, 589 patients (2.6%)
243
experienced a PICC exchange during the dwell of the device while 22,421 did not
244
(Figure 1). A total of 480 patients (2.1%) experienced venous thromboembolism in the
245
study (460=deep vein thrombosis; 20=pulmonary embolism). Notably, the frequency of
246
PICC thrombosis was greater among patients that underwent exchange vs. those that
247
did not (3.6% vs. 2.0%, p<0.001). Approximately half of all PICC exchanges were
248
performed for catheter dislodgement (30.2%) and occlusion (19.7%). Other indications
249
for exchange included need for more lumens (22.4%) and catheter migration (7.4%).
250
Bivariate comparisons showed that patients that underwent PICC exchanges
251
differed in several ways from those that did not (Table 1). With respect to comorbidities,
252
patients that underwent exchange had greater prevalence of moderate-severe liver
253
disease (5.4% vs. 3.8%, p=0.04) and active cancer (9.0% vs 6.1%, p<0.01) than those
254
that did not. Similarly, patients that underwent PICC exchange more often had a history
255
of deep vein thrombosis within 30 days (6.1% vs. 3.9, p<0.01) and history of central
12
Page 12 of 34
256
line-associated bloodstream infection (3.1% vs. 1.2%, p<0.01). PICC exchanges also
257
occurred more frequently among patients that had a prior central venous catheter or
258
PICC (28.0% vs. 21.1%, p<0.01) or an existing central line at the time of the exchange
259
(25.6% vs. 13.3%, p<0.01).
260
With respect to provider-characteristics, patients that underwent PICC
261
exchanges more often had their PICC placed by interventional radiology than other
262
providers (37.0% vs. 20.1%, p<0.01). Similarly, differences in device characteristics,
263
including catheter gauge, coating, and number of lumens were noted between those
264
that did vs. did not undergo an exchange (p<0.01 for all). Importantly, patients that
265
experienced exchanges more often had double and triple lumen PICCs than those that
266
did not (53.7% and 18.5% vs. 48.4% and 13.9%, p<0.01, respectively).
267 268
Unadjusted Association Between PICC Exchange, Patient-, Provider-, Device-
269
Factors and Thromboembolism
270
Unadjusted associations using Cox models between patient-, provider- and
271
device-characteristics, PICC exchange and thromboembolism are shown in Table 2.
272
Several comorbidities including body mass index (HR=0.99 [95%CI=0.98-1.00]), active
273
cancer (HR=1.88 [95%CI=1.34-2.64]), coagulopathy (HR=1.82 [95%CI=1.20-2.75]),
274
history of venous thrombosis within 30-days (HR=2.57 [95%CI=1.80-3.66]),
275
inflammatory bowel disease (HR=1.77 [95%CI=1.20-2.62]) and ICU stay (HR=1.97
276
[95%CI=1.63-2.38]) were associated with deep vein thrombosis. Additionally, patients
277
that experienced a deep vein thrombosis within 30-days (HR=2.65 [95%CI=1.87-3.77]),
278
had a central venous catheter or PICC in the preceding six months (HR=1.37
13
Page 13 of 34
279
[95%CI=1.08-1.75]) or had another central line at the time of PICC exchange (HR=1.88
280
[95%CI=1.52-2.32]) were more likely to experience thrombosis than those that did not.
281
Importantly, PICC exchange was strongly associated with deep vein thrombosis
282
on unadjusted analysis (HR=2.27 [95%CI=1.62-3.17]). Likewise, multi-lumen PICCs
283
were also associated with thrombosis (HR=2.67 [95%CI=2.12-3.37] and HR=3.38
284
[95%CI=2.40-4.76] for double and triple lumen vs. single lumen PICCs, respectively).
285 286
Adjusted Association Between PICC Exchange, Patient-, Provider-, Device-
287
Factors and Thrombosis
288
Following adjustment for patient-, provider- and device-characteristics, PICC
289
exchange remained associated with a nearly two-fold greater risk of thrombosis
290
(HR=1.98 [95%CI=1.37-2.85]) (Figure 2). The risk of thrombosis following PICC
291
exchange was second in magnitude to the device lumens, with double and triple-lumen
292
devices being associated with marginally greater risk of thrombosis (HR=2.06
293
[95%CI=1.59-2.66] and HR=2.31 [95%CI=1.6-3.33], respectively) than exchange. The
294
effect size of PICC exchange on risk of thrombosis was comparable to that of deep vein
295
thrombosis within 30-days (HR=1.97 [95%CI=1.32-2.96]). Active cancer (HR=1.68
296
[95%CI=1.15-2.46]), presence of another central venous catheter or PICC (HR=1.24
297
[95%CI=1.00-1.53]), and elevated WBC count (HR=1.41 [95%CI=1.17-1.71 for > 12,000
298
vs. < 12,000]) were also associated with deep vein thrombosis; ICU status, however,
299
was not significantly associated with thrombosis in the adjusted Cox model (HR 1.48
300
(95%CI: 0.86-2.53, p=0.16)
301
14
Page 14 of 34
302
DISCUSSION
303
Exchange of central venous catheters and PICCs over guidewires is often
304
necessary in the setting of device dislodgement, malfunction or occlusion. While
305
concerns regarding infection risk with this practice have been raised in neonates and
306
those receiving dialysis,8, 9 to date – no study has examined whether the practice might
307
be associated with thrombosis. In this cohort study, we found that PICC exchanges
308
were strongly and independently associated with thrombosis after adjusting for other
309
putative risk factors. The magnitude of the hazard of thrombosis in patients with
310
exchanges was comparable to that of the most widely recognized risk factors for PICC
311
thrombosis – number of lumens and catheter size.21, 22 This finding has important
312
clinical and policy implications as some factors leading to exchange might be
313
preventable – in turn, lowering risk of thrombosis. Furthermore, as exchanges are
314
clearly not innocuous, closer attention to risks and benefits of exchanging PICCs over
315
guidewires now appears necessary.
316
Why might catheter exchange increase the risk of venous thrombosis? Several
317
explanations exist. For example, dislodged or occluded catheters may injure the vessel
318
intima, predisposing to inflammation and venous stasis. Threading a wire in this setting
319
may further exacerbate injury as the wire is left in direct contact with the vessel during
320
device replacement. Alternatively, the act of removing and threading a PICC through the
321
vascular tree through smaller diameter veins may itself cause vessel injury and
322
thrombosis. Mechanistic studies using biomarkers to measure inflammation or
323
advanced imaging modalities to monitor vessel condition are needed to understand the
324
biology of this phenomenon.
15
Page 15 of 34
325
Although exchanges were relatively infrequent in this study, rates vary from 2%
326
to 20% in the literature.8, 23-25 Notably, PICC occlusion and dislodgement were among
327
the two most common causes that led to catheter exchange in our cohort. These events
328
are relevant because they may be preventable with better catheter care and
329
maintenance. For instance, close attention to measuring exposed catheter length
330
(especially in disoriented or delirious patients) could help address dislodgement before
331
it becomes necessary to exchange the catheter. Alternatively, newer technology such
332
as subcutaneous anchoring devices may help avoid exchange.26, 27 Similarly, preventing
333
catheter occlusion through meticulous attention to flushing and better device selection
334
(greater use of single lumen devices) is important.28 Of note, it is pertinent to emphasize
335
that occlusion itself does not necessarily indicate thrombosis, as occlusion more
336
frequently represents intraluminal clot, precipitation of medications, or fibrin sheath
337
around the catheter tip than deep vein thrombosis.29 Addressing occlusion might help
338
avoid a substantial number of exchanges, which in turn might prevent thrombosis.
339
However, some clinical conditions prompting exchange might not be avoidable.
340
For example, we observed that patients with recent deep vein thrombosis and
341
bloodstream infection experienced more frequent exchanges than those without such
342
diagnoses. While this finding may reflect therapeutic decisions related to managing
343
these conditions, some exchanges in these cases may be inappropriate. For example,
344
removal or exchange of functional PICCs in the setting of thrombosis provided the
345
device is still clinically necessary is not recommended by guidelines,30 yet often occurs
346
in real world settings.4, 31 Similarly, exchanging devices for antimicrobial central venous
347
catheters when treating line infections is gaining popularity as a novel strategy,32 but is
16
Page 16 of 34
348
recommended without regard to thrombotic risk. Understanding and balancing
349
thrombotic and infectious risks from exchange thus appears relevant in these decisions.
350
Our study has limitations. First, this is an observational study and can only
351
assess association, not causality. Relatedly, we did not sample consecutive patients –
352
but rather included those that met eligibility criteria within defined abstractor workloads.
353
Thus, selection bias in our sampling may affect our results. Second, while we adjust for
354
multiple known confounders using a validated conceptual model, we cannot be certain
355
that our findings are not confounded by unmeasured variables. This is particularly
356
relevant when patients entered the database as an exchange with limited data of the
357
prior device available. Furthermore, because patient follow-up was terminated at the
358
time of PICC removal, differential follow-up times and ascertainment of VTE between
359
those with vs. those without an exchange is possible. Third, although we track
360
indication for PICC exchange, we cannot determine whether these were performed for
361
clinically appropriate reasons. Relatedly, alternatives to PICC exchange are not readily
362
apparent through this study. Future studies assessing these aspects are necessary to
363
determine preventability of thrombosis in the setting of catheter exchange. Fourth,
364
because the HMS consortium is focused on medical patients, whether our findings hold
365
true for other patient groups (e.g., surgical and cancer patients) is not known. However,
366
given the increased risk of thrombosis in these subsets, caution in use and exchange of
367
PICCs seems reasonable.
368
Our study also has strengths. First, to our knowledge, this is among the first
369
studies to report an association between PICC exchange and thrombosis. Our findings
370
have important clinical and policy recommendations and suggest, at the very least, that
17
Page 17 of 34
371
more careful consideration of the appropriateness of exchanging a PICC is necessary.
372
Second, as almost half of all exchanges occurred in the setting of catheter malfunction
373
or dislodgement, our study highlights the importance of care, securement and
374
maintenance of PICCs. Third, our findings add to the growing literature regarding
375
increasing thrombogenicity of PICCs. Now, more than ever, use of appropriateness
376
criteria to guide insertion of PICCs is necessary to prevent inappropriate PICC insertion
377
and subsequent complications.33 As more patients that experienced exchanges had
378
double and triple lumen PICCs and as increasing lumens are associated with occlusion,
379
using the fewest number of PICC lumens to prevent thrombosis, occlusion and
380
exchange all appears necessary.
381
In conclusion, exchange of PICCs over guidewires may not be benign and is
382
associated with a nearly two-fold greater risk of venous thrombosis. Continued focus on
383
ensuring PICCs are placed for appropriate reasons, cared for with the highest
384
standards, and not exchanged unless necessary is warranted. A randomized trial of
385
exchange vs. new PICC placement – with longitudinal assessment of biological markers
386
of thrombosis - appears necessary.
387 388 389
18
Page 18 of 34
390 391 392
BIBLIOGRAPHY 1.
Greene MT, Flanders SA, Woller SC, Bernstein SJ, Chopra V. The Association
393
Between PICC Use and Venous Thromboembolism in Upper and Lower
394
Extremities. Am J Med. 2015;128:986-993 e981.
395
2.
Chopra V, Anand S, Hickner A, et al. Risk of venous thromboembolism
396
associated with peripherally inserted central catheters: a systematic review and
397
meta-analysis. Lancet. 2013;382:311-325.
398
3.
Hoshal VL, Jr. Total intravenous nutrition with peripherally inserted silicone
399
elastomer central venous catheters. Archives of surgery (Chicago, Ill. : 1960).
400
1975;110:644-646.
401
4.
Chopra V, Kuhn L, Flanders SA, Saint S, Krein SL. Hospitalist experiences,
402
practice, opinions, and knowledge regarding peripherally inserted central
403
catheters: results of a national survey. J Hosp Med. 2013;8:635-638.
404
5.
Gibson C, Connolly BL, Moineddin R, Mahant S, Filipescu D, Amaral JG.
405
Peripherally inserted central catheters: use at a tertiary care pediatric center.
406
Journal of vascular and interventional radiology : JVIR. 2013;24:1323-1331.
407
6.
central routes for central venous cannulation. Anaesthesia. 2012;67:65-71.
408 409
Pikwer A, Akeson J, Lindgren S. Complications associated with peripheral or
7.
Song L, Li H. Malposition of peripherally inserted central catheter: experience
410
from 3,012 patients with cancer. Experimental and therapeutic medicine.
411
2013;6:891-893.
19
Page 19 of 34
412
8.
McCoy M, Bedwell S, Noori S. Exchange of peripherally inserted central
413
catheters is associated with an increased risk for bloodstream infection. Am J
414
Perinatol. 2011;28:419-424.
415
9.
Guttmann DM, Trerotola SO, Clark TW, et al. Malfunctioning and infected
416
tunneled infusion catheters: over-the-wire catheter exchange versus catheter
417
removal and replacement. Journal of vascular and interventional radiology :
418
JVIR. 2011;22:642-646; quiz 646.
419
10.
Greene MT, Spyropoulos AC, Chopra V, et al. Validation of Risk Assessment
420
Models of Venous Thromboembolism in Hospitalized Medical Patients. Am J
421
Med. 2016;129:1001 e1009-1001 e1018.
422
11.
Grant PJ, Greene MT, Chopra V, Bernstein SJ, Hofer TP, Flanders SA.
423
Assessing the Caprini Score for Risk Assessment of Venous Thromboembolism
424
in Hospitalized Medical Patients. Am J Med. 2016;129:528-535.
425
12.
Herc E, Patel P, Washer LL, Conlon A, Flanders SA, Chopra V. A Model to
426
Predict Central-Line-Associated Bloodstream Infection Among Patients With
427
Peripherally Inserted Central Catheters: The MPC Score. Infect Control Hosp
428
Epidemiol. 2017;38:1155-1166.
429
13.
Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining
430
comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care.
431
2005;43:1130-1139.
432
14.
Chopra V, Anand S, Krein SL, Chenoweth C, Saint S. Bloodstream infection,
433
venous thrombosis, and peripherally inserted central catheters: reappraising the
434
evidence. Am J Med. 2012;125:733-741.
20
Page 20 of 34
435
15.
Chopra V, Ratz D, Kuhn L, Lopus T, Chenoweth C, Krein S. PICC-associated
436
bloodstream infections: prevalence, patterns, and predictors. Am J Med.
437
2014;127:319-328.
438
16.
Pongruangporn M, Ajenjo MC, Russo AJ, et al. Patient- and device-specific risk
439
factors for peripherally inserted central venous catheter-related bloodstream
440
infections. Infect Control Hosp Epidemiol. 2013;34:184-189.
441
17.
Lee EW, Wei LJ, Amato DA, Leurgans S. Cox-Type Regression Analysis for
442
Large Numbers of Small Groups of Correlated Failure Time Observations. .
443
Survival Analysis: State of the Art. Nato Science (Series E: Applied Sciences).
444
1992;211.
445
18.
Rubin D. Multiple imputation in sample surveys-a phenomenological Bayesian
446
approach to nonresponse. Journal of the American Statistical Association, .
447
1978;20:20-34.
448
19.
peripherally inserted central catheters. Chest. 2010;138:803-810.
449 450
Evans RS, Sharp JH, Linford LH, et al. Risk of symptomatic DVT associated with
20.
Chopra V, Ratz D, Kuhn L, Lopus T, Lee A, Krein S. Peripherally inserted central
451
catheter-related deep vein thrombosis: contemporary patterns and predictors. J
452
Thromb Haemost. 2014;12:847-854.
453
21.
O'Brien J, Paquet F, Lindsay R, Valenti D. Insertion of PICCs with minimum
454
number of lumens reduces complications and costs. Journal of the American
455
College of Radiology : JACR. 2013;10:864-868.
456 457
22.
Evans RS, Sharp JH, Linford LH, et al. Reduction of peripherally inserted central catheter-associated DVT. Chest. 2013;143:627-633.
21
Page 21 of 34
458
23.
Gnannt R, Patel P, Temple M, et al. Peripherally Inserted Central Catheters in
459
Pediatric Patients: To Repair or Not Repair. Cardiovasc Intervent Radiol.
460
2017;40:845-851.
461
24.
Trerotola SO, Thompson S, Chittams J, Vierregger KS. Analysis of tip
462
malposition and correction in peripherally inserted central catheters placed at
463
bedside by a dedicated nursing team. Journal of vascular and interventional
464
radiology : JVIR. 2007;18:513-518.
465
25.
Leroyer C, Lasheras A, Marie V, et al. Prospective follow-up of complications
466
related to peripherally inserted central catheters. Med Mal Infect. 2013;43:350-
467
355.
468
26.
Zerla PA, Canelli A, Cerne L, et al. Evaluating safety, efficacy, and cost-
469
effectiveness of PICC securement by subcutaneously anchored stabilization
470
device. J Vasc Access. 2017;18:238-242.
471
27.
Nurs. 2014;23:S12, S14-18.
472 473
Elen Hughes M. Reducing PICC migrations and improving patient outcomes. Br J
28.
Smith SN, Moureau N, Vaughn VM, et al. Patterns and Predictors of Peripherally
474
Inserted Central Catheter Occlusion: The 3P-O Study. Journal of vascular and
475
interventional radiology : JVIR. 2017;28:749-756 e742.
476
29.
Baskin JL, Pui CH, Reiss U, et al. Management of occlusion and thrombosis
477
associated with long-term indwelling central venous catheters. Lancet.
478
2009;374:159-169.
479 480
30.
Kearon C, Akl EA, Comerota AJ, et al. Antithrombotic therapy for VTE disease: Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College
22
Page 22 of 34
481
of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest.
482
2012;141:e419S-e496S.
483
31.
Chopra V, Kuhn L, Coffey CE, Jr., et al. Hospitalist experiences, practice,
484
opinions, and knowledge regarding peripherally inserted central catheters: a
485
Michigan survey. J Hosp Med. 2013;8:309-314.
486
32.
Zakhour R, Chaftari AM, Raad, II. Catheter-related infections in patients with
487
haematological malignancies: novel preventive and therapeutic strategies.
488
Lancet Infect Dis. 2016;16:e241-e250.
489
33.
Chopra V, Flanders SA, Saint S, et al. The Michigan Appropriateness Guide for
490
Intravenous Catheters (MAGIC): Results From a Multispecialty Panel Using the
491
RAND/UCLA Appropriateness Method. Ann Intern Med. 2015;163:S1-40.
492 493 494 495
23
Page 23 of 34
Figure 1. Patient Flow Diagram
Patients identified as having received a PICC in 51 HMS Hospitals N=23,010
Patients without exchange N=22,421 (97.4%)
Patients without image confirmed, symptomatic VTE N=21,962 (98.0%)
Patients with image confirmed, symptomatic VTE N=459 (2.0%)
Patients with exchange N=589 (2.6%)
Patients without image confirmed, symptomatic VTE N=568 (96.4%)
Patients with image confirmed, symptomatic VTE N=21 (3.6%)
Legend: PICC=Peripherally inserted central catheter; VTE=venous thromboembolism
24
Page 24 of 34
Figure 2. Multivariable Cox Model for Time to Venous Thromboembolism following PICC Exchange
25
Page 25 of 34
Table 1. General Characteristics of Patients with and without PICC exchange (n=23,010) Category/Variable
Modifier
Patient Characteristics Male Gender
No exchange (n=22,421)
Exchange (n=589)
p
10985 (49.0%)
320 (54.3%)
0.01
Race
White
16647 (76.1%)
391 (68.6%)
<0.01
Age group
>65 years
11065 (49.4%)
257 (43.6%)
<0.01
Body Mass Index (BMI)
Median (IQR)
28.7 (24.0-35.2)
29.4 (24.2-36.5)
0.16
8282 (36.9%)
197 (33.4%)
0.08
15350 (68.5%)
387 (65.7%)
0.16
3417 (15.2%)
76 (12.9%)
0.12
6448 (28.8%)
175 (29.7%)
0.61
3433 (15.3%)
77 (13.1%)
0.14
3409 (15.2%)
79 (13.4%)
0.23
1917 (8.6%)
51 (8.7%)
0.93
6083 (27.1%)
155 (26.3%)
0.66
1377 (6.1%)
39 (6.6%)
0.63
859 (3.8%)
22 (3.7%)
0.90
4590 (20.5%)
119 (20.2%)
0.87
4387 (19.6%)
132 (22.4%)
0.09
7782 (34.7%)
201 (34.1%)
0.77
128 (0.6%)
6 (1.0%)
0.16
771 (3.4%)
18 (3.1%)
0.61
39 (0.2%)
3 (0.5%)
0.06
1073 (4.8%)
29 (4.9%)
0.88
Mild liver disease
1469 (6.6%)
38 (6.5%)
0.92
Moderate to severe liver diseasea Known HIV or AIDSa
844 (3.8%)
32 (5.4%)
0.04
156 (0.7%)
6 (1.0%)
0.35
History of Cancer
5247 (23.4%)
157 (26.7%)
0.07
Active cancer
1378 (6.1%)
53 (9.0%)
<0.01
799 (3.6%)
19 (3.2%)
0.66
3 (1-5)
3 (2-5)
0.05
Hyperlipidemia
a
a
Hypertension
Myocardial Infarction
b b
Congestive Heart Failure
Peripheral vascular disorders
a
a
Cerebrovascular disease a
Dementia b
COPD
Rheumatoid arthritis
b
Peptic Ulcer Disease
a
Diabetes without complicationsa Diabetes with complicationsa Renal Failure
a
Kidney transplant Hemodialysis
a
Peritoneal Dialysis
a
Hemi- or paraplegia
b
a
a
Coagulopathy
Charlson/Deyo Comorbidity Index
Median (IQR)
26
Page 26 of 34
History of CLABSI
259 (1.2%)
18 (3.1%)
<0.01
647 (2.9%)
28 (4.8%)
0.01
2422 (10.8%)
73 (12.4%)
340 (1.5%)
14 (2.4%)
1294 (5.8%)
37 (6.3%)
874 (3.9%)
36 (6.1%)
2893 (12.9%)
87 (14.8%)
675 (3.0%)
15 (2.5%)
0.51
9044 (40.3%)
257 (43.6%)
0.11
7045 (31.4%)
162 (27.5%)
0.04
6475 (28.9%)
165 (28.0%)
0.65
8158 (36.4%)
207 (35.1%)
0.54
3837 (17.1%)
85 (14.4%)
0.09
1460 (6.5%)
38 (6.5%)
0.95
12976 (57.9%)
337 (57.2%)
0.75
7978 (35.6%)
210 (35.7%)
0.97
Aspirin
7119 (31.8%)
189 (32.1%)
0.86
Other antiplatelet therapy
3044 (13.6%)
76 (12.9%)
0.64
Anticoagulant therapy
15089 (67.3%)
431 (73.2%)
<0.01
History of DVT
Within 30 days Prior history
History of PE
History of any VTE Event
Inflammatory Bowel Diseaseb
Within 30 days Positive history Within 30 days Positive history
b
Serious lung disease
Life-threatening illness
b
b
Pneumonia b
Sepsis
History of Prior CVA/TIA Venous stasis
a
Smoking Status Statin
Current/ Former
0.21
<0.01
eGFR
Median (IQR)
60 (47-90)
60 (54-89)
0.59
WBC
Median (IQR)
9.2 (6.7-12.9)
9.0 (6.6-12.5)
0.16
Hemoglobin
Median (IQR)
10.2 (8.8-11.8)
9.6 (8.3-11.1)
<0.01
Platelets
Median (IQR)
226 (160-309)
234 (157-319)
0.29
INR
Median (IQR)
1.15 (1.03-1.39)
1.20 (1.05-1.40)
0.07
CVC or PICC in prior 6 months
4736 (21.1%)
302 (51.3%)
<0.01
Presence of another CVC
2982 (13.3%)
151 (25.6%)
<0.01
15006 (66.9%)
292 (49.6%)
<0.01
4496 (20.1%)
218 (37.0%)
Provider Characteristics Operator Type
Vascular Access Nurse Interventional Radiologist
27
Page 27 of 34
Physician
237 (1.1%)
4 (0.7%)
2682 (12.0%)
75 (12.7%)
6788 (30.3%)
176 (29.9%)
0.84
>1
2547 (11.6%)
19 (3.4%)
<0.01
Arm selected for insertion
Right Arm
15743 (70.3%)
398 (67.6%)
0.16
Vein selected for insertion
Basilic
13657 (60.9%)
328 (55.7%)
<0.01
Brachial
6910 (30.8%)
171 (29.0%)
Cephalic
1158 (5.2%)
21 (3.6%)
Other
696 (3.1%)
69 (11.7%)
4 French
6647 (31.1%)
134 (23.3%)
5 French
12844 (60.1%)
365 (63.4%)
1885 (8.8%)
77 (13.4%)
20352 (90.8%)
503 (85.4%)
<0.01
Antimicrobial coating
1578 (7.0%)
12 (2.0%)
<0.01
Anti-thrombotic coating
390 (1.7%)
6 (1.0%)
0.18
5567 (24.8%)
124 (21.1%)
0.04
Single
8431 (37.7%)
163 (27.9%)
<0.01
Double
10841 (48.4%)
314 (53.7%)
Triple/Quad
3106 (13.9%)
108 (18.5%)
ICU prior to or at time of exchange Placement attempts
Device characteristics PICC Gauge
Other
> 6 French Power PICC
Valved PICC Number of PICC lumens
<0.01
a
at time of initial PICC placement; bwithin 30 days of PICC placement;
Legend: IQR=interquartile range; COPD=chronic obstructive pulmonary disease; HIV=human immunodeficiency virus; AIDS=acquired immune deficiency syndrome; CLABSI=central line-associated bloodstream infection; DVT=deep vein thrombosis; PE=pulmonary embolism; VTE=venous thromboembolism; CVA=cerebrovascular accident; TIA=transient ischemic attack; eGFR=estimated glomerular filtration rate; LOS=length of stay; CVC=central venous catheter; PICC=peripherally inserted central catheter; TPN=total parenteral nutrition; ICU=intensive care unit. Definitions: Diabetes without complications = diabetes without documented retinopathy, nephropathy, neuropathy, cardio- or cerebrovascular events.
28
Page 28 of 34
29
Page 29 of 34
Table 2. Unadjusted associations of PICC Exchange, Patient-, Provider and Device-Factors with PICC-DVT (n=23,010) Category/Variable
Modifier HR (95% CI)
P-value
PICC Exchange
Yes vs. No
2.27 (1.62, 3.17)
<0.01
Male Gender
Male vs. Female White vs. Other >65 vs. <64
0.98 (0.80, 1.19)
0.81
0.81 (0.60, 1.09)
0.16
1.05 (0.87, 1.26)
0.63
Per unit increase Yes vs. No
0.99 (0.98, 1.00)
0.02
0.94 (0.72, 1.22)
0.63
Yes vs. No
0.96 (0.81, 1.15)
0.67
Yes vs. No
0.75 (0.57, 0.98)
0.04
Yes vs. No
0.83 (0.69, 1.01)
0.07
Yes vs. No
0.71 (0.55, 0.92)
0.01
Yes vs. No
1.10 (0.92, 1.31)
0.28
Yes vs. No
1.04 (0.72, 1.49)
0.85
Yes vs. No
1.05 (0.82, 1.33)
0.70
Yes vs. No
0.89 (0.58, 1.37)
0.60
Yes vs. No
1.46 (1.04, 2.05)
0.03
Yes vs. No
0.87 (0.70, 1.07)
0.19
Yes vs. No
0.87 (0.70, 1.08)
0.22
Yes vs. No
0.90 (0.74, 1.09)
0.29
Yes vs. No
1.11 (0.47, 2.61)
0.81
Yes vs. No
1.50 (0.97, 2.32)
0.07
Yes vs. No
1.14 (0.76, 1.71)
0.54
Yes vs. No
0.77 (0.53, 1.12)
0.18
Yes vs. No
0.78 (0.51, 1.19)
0.25
Known HIV or AIDS
Yes vs. No
0.29 (0.04, 2.11)
0.22
History of Cancer
Yes vs. No
1.27 (1.05, 1.55)
0.012
Active cancer
Yes vs. No
1.88 (1.34, 2.64)
<0.01
Coagulopathy
Yes vs. No
1.82 (1.20, 2.75)
<0.01
Charlson/Deyo Comorbidity Index
Per unit
0.98 (0.95, 1.01)
0.17
Race Age group Body Mass Index (BMI) Hyperlipidemiaa a
Hypertension
Myocardial Infarction
b b
Congestive Heart Failure
Peripheral vascular disorders
a
a
Cerebrovascular disease a
Dementia b
COPD
Rheumatoid arthritis
b
Peptic Ulcer Disease
a
Diabetes without complications Diabetes with complications Renal Failure
a
a
a
Kidney transplant Hemodialysis
a
Hemi- or paraplegia
b
a
Mild liver disease
Moderate to severe liver disease
a
a
a
30
Page 30 of 34
increase Yes vs. No
History of CLABSI
1.07 (0.56, 2.04)
0.85
Within 30 days vs. none Prior history vs. none Within 30 days vs. none Positive history vs. none Within 30 days vs. none Positive history vs. none Yes vs. No
2.65 (1.87, 3.77)
<0.01
1.77 (1.20, 2.62)
<0.01
Yes vs. No
1.35 (1.09, 1.67)
<0.01
Yes vs. No
1.97 (1.63, 2.38)
<0.01
Yes vs. No
1.18 (0.98, 1.44)
0.09
Yes vs. No
1.02 (0.81, 1.28)
0.88
Yes vs. No
1.05 (0.85, 1.31)
0.64
Yes vs. No
0.70 (0.44, 1.11)
0.13
1.01 (0.84, 1.22)
0.88
Statin
Current/former vs. Never Yes vs. No
0.85 (0.70, 1.03)
0.10
Aspirin
Yes vs. No
0.68 (0.57, 0.80)
<0.01
Other antiplatelet therapy
Yes vs. No
0.81 (0.64, 1.03)
0.08
Anticoagulant therapy
Yes vs. No
0.88 (0.74, 1.05)
0.15
EGFR
Per unit increase
1.00 (0.99, 1.00)
0.19
WBC
Per unit increase
1.01 (1.00, 1.01)
<0.01
Hemoglobin
Per unit increase
0.96 (0.92, 1.01)
0.11
Platelets
Per unit increase
1.00 (1.00, 1.00)
0.10
INR
Per unit increase
1.00 (0.99, 1.00)
0.11
CVC or PICC in prior 6 months
Yes vs. No
1.39 (1.10, 1.76)
<0.01
Presence of another CVC
Yes vs. No
1.88 (1.52, 2.32)
<0.01
History of DVT
History of PE
History of any VTE Event
Inflammatory Bowel Diseaseb b
Serious lung disease
Life-threatening illness
b
b
Pneumonia b
Sepsis
History of Prior CVA/TIA Venous stasis
a
Smoking Status
1.84 (1.51, 2.24) 1.56 (0.82, 3.00)
<0.01
1.65 (1.28, 2.14) 2.57 (1.80, 3.66)
<0.01
1.79 (1.47, 2.19)
31
Page 31 of 34
Operator Type
ICU prior to or at time of exchange
Interventional Radiologist vs. Vascular Access Nurse Physician vs. Vascular Access Nurse Advanced Practice Prof vs. Vascular Access Nurse Other vs. Vascular Access Nurse Yes vs. No
Placement attempts
>1 vs. 1
Arm selected for insertion Vein selected for insertion
1.33 (0.84, 2.11)
0.02
0.57 (0.20, 1.67) 1.53 (1.02, 2.30)
0.46 (0.22, 0.95) 2.08 (1.74, 2.49)
<0.01
12.12 (1.63, 89.83)
0.01
Right vs. Left
0.98 (0.79, 1.22)
0.87
Brachial vs. Basilic Cephalic vs. Basilic Other vs. Basilic
1.26 (1.02, 1.55)
<0.01
5 Fr. vs 4 Fr.
2.17 (1.67, 2.82)
<0.01
6Fr vs. 4 Fr.
2.90 (1.83, 4.59)
<0.01
Power PICC
Yes vs. No
0.84 (0.64, 1.12)
0.24
Antimicrobial coating
Yes vs. No
1.85 (1.28, 2.67)
<0.01
Anti-thrombotic coating
Yes vs. No
1.12 (0.53, 2.40)
0.77
Valved PICC
Yes vs. No
0.91 (0.62, 1.34)
0.64
Number of PICC lumens
Double vs. Single Triple/Quad vs. Single vs. Single
2.67 (2.12, 3.37)
<0.01
Device characteristics PICC Gauge
0.63 (0.40, 0.98) 1.13 (0.77, 1.66)
3.38 (2.40, 4.76)
Legend: COPD=chronic obstructive pulmonary disease; HIV=human immunodeficiency virus; AIDS=acquired immune deficiency syndrome; CLABSI=central line-associated bloodstream infection; DVT=deep vein thrombosis; PE=pulmonary embolism; VTE=venous thromboembolism; CVA=cerebrovascular accident; TIA=transient ischemic attack; eGFR=estimated glomerular filtration rate; LOS=length of stay; CVC=central venous catheter; PICC=peripherally inserted central catheter; TPN=total parenteral nutrition; ICU=intensive care unit;
32
Page 32 of 34
Table 3. Multivariable Cox Model for Upper Extremity Deep Vein Thrombosis and Pulmonary Embolism Variable Exchange VTE History
Active cancer Presence of another CVC/PICC Lumens
WBC Life-threatening illness CVC/PICC in previous 6 months Operator Type
ICU Status Placement attempts Antimicrobial coated Insertion vein
HR (95% CI) 1.98 (1.37, 2.85) Positive history vs. No history 1.87 (1.51, 2.33) Within previous 30 days vs. No history 1.97 (1.32, 2.96) Yes vs. No 1.68 (1.15, 2.46) Yes vs. No 1.24 (1.00, 1.53) Double vs. Single 2.06 (1.59, 2.66) Triple/Quad vs. Single 2.31 (1.60, 3.33) >12 vs. <=12 1.41 (1.17, 1.71) Yes vs. No 1.28 (1.00, 1.65) Yes vs. No 1.03 (0.83, 1.28) Interventional Radiologist vs. Vascular 0.70 (0.24, Access Nurse 2.10) Physician vs. Vascular Access Nurse 1.40 (1.11, 1.78) Advance Practice Professional vs. 0.39 (0.20, Vascular Access Nurse 0.78) Other vs. Vascular Access Nurse 0.40 (0.20, 0.78) Yes vs. No 1.48 (0.86 – 2.53) >1 vs. Only 1 0.91 (0.68, 1.24) Yes vs. No 1.54 (1.22, 1.95) Brachial vs. Basilic 1.18 (0.99, 1.40) Cephalic vs. Basilic 0.60 (0.39, 0.92) Other vs. Basilic 0.85 (0.58, 1.24) Yes vs. No
P-value 0.0003 <.0001
0.0077 0.0529 <.0001
0.0004 0.0523 0.7793 0.0003
0.16 0.5641 0.0003 0.0121
Legend: CI=confidence interval; VTE=venous thromboembolism; PICC=peripherally inserted central catheter; CVC=central venous catheter
33
Page 33 of 34
1 2 3 4 5
34
Page 34 of 34