Accepted Manuscript The role of Tigecycline in the management of Clostridium difficile infection: a retrospective cohort study Eliza Manea, Jesús Sojo-Dorado, Raluca E. Jipa, Serban N. Benea, Jesús Rodríguez-Baño, Adriana Hristea PII:
S1198-743X(17)30324-5
DOI:
10.1016/j.cmi.2017.06.005
Reference:
CMI 972
To appear in:
Clinical Microbiology and Infection
Received Date: 6 February 2017 Revised Date:
2 June 2017
Accepted Date: 5 June 2017
Please cite this article as: Manea E, Sojo-Dorado J, Jipa RE, Benea SN, Rodríguez-Baño J, Hristea A, The role of Tigecycline in the management of Clostridium difficile infection: a retrospective cohort study, Clinical Microbiology and Infection (2017), doi: 10.1016/j.cmi.2017.06.005. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
1
Article category: Original Article
2
Title: The role of Tigecycline in the management of Clostridium difficile infection: a retrospective cohort study
5
RI PT
3 4
6
Authors: Eliza Manea1,2, Jesús Sojo-Dorado3, Raluca E. Jipa1,2, Serban N.
7
Benea1,2, Jesús Rodríguez-Baño3, Adriana Hristea1,2
8
Affiliations:
SC
9
1 National Institute for Infectious Diseases “Prof. Dr. Matei Bals”, Bucharest,
11
Romania
12
2 University of Medicine and Pharmacy “Carol Davila”, Bucharest, Romania
13
3 Unidad de Enfermedades Infecciosas, Microbiología y Medicina Preventiva,
14
Hospital Universitario Virgen Macarena / Universidad de Sevilla / Instituto de
15
Biomedicina de Sevilla, Seville, Spain.
M AN U
10
16
The study was conducted at the National Institute for Infectious Diseases “Prof.
18
Dr. Matei Bals” – Address: no1 Calistrat Grozovici Street, sect 2, Bucharest,
19
021105, Romania.
20
Corresponding author:
21
Adriana Hristea: e-mail:
[email protected]
22
Address: National Institute of Infectious Diseases “Prof Dr Matei Bals”; no 1
23
Calistrat Grozovici, street, sect 2, postal code: 021105, Bucharest, Romania
25 26 27
EP
AC C
24
TE D
17
Phone: +40213186100 ext 5204; Mobile phone: +40723562370; Fax: +40213186090 Conflict of interest: the authors have no conflicts of interest to declare. Funding: this study had no funding. Keywords: Clostridium difficile, tigecycline, outcome, propensity score,
28
recurrence
29
Word count of summary: 249
30
Word count of text: 2341
31
Number of references: 16; number of tables: 3; number of figures: 1 1
ACCEPTED MANUSCRIPT
32
Objective
34
We aimed to compare the outcomes of patients with C.difficile infection (CDI)
35
treated either with tigecycline associated with vancomycin, or with vancomycin
36
alone.
37
Methods
38
This single-centre retrospective cohort study included all adults hospitalized from
39
September 2014 through August 2015 for symptomatic, incident CDI confirmed
40
by polymerase chain reaction for C.difficile toxin in stools. The primary outcome
41
was the rate of favourable outcome, defined as a composite of clinical response
42
(resolution of symptoms without need for additional CDI therapy) and achieving
43
discharge without CDI-related surgery or intensive care; a secondary outcome
44
was CDI recurrence. We constructed a non-parsimonious logistic regression
45
model to calculate a propensity score (PS) for those receiving tigecycline.
46
Results
47
In all, 266 patients were included: 62 patients received both vancomycin and
48
tigecycline, and 204 patients received vancomycin alone. The patients from the
49
two groups were similar regarding demographics and comorbidities but patients in
50
the tigecycline group had a more severe CDI A favourable outcome in the
51
tigecycline group vs the vancomycin group was found in 50/62 (81%) vs 193/204
52
(95%). We matched patients receiving tigecycline or not according to the PS and
54 55
SC
M AN U
TE D
EP
AC C
53
RI PT
33
86 patients (43 pairs) could be matched. The OR for favourable outcome with tigecycline in the matched analysis was 0.92 (95% CI: 0.60-1.44; p=0.74). The rate of CDI recurrences was 8/62 (13%) in the tigecycline group versus 39/204
56
(19%) in the vancomycin group (p=0.2).
57
Conclusion
2
ACCEPTED MANUSCRIPT
58
Adding tigecycline to CDI standard therapy did not increase the clinical cure nor
59
reduced the rate of CDI recurrences.
60
Introduction
62
In the past 10-15 years, Clostridium difficile infection (CDI) has re-emerged as a
63
large epidemiological problem for hospitalized patients across the world [1,2].
64
Despite the increasing incidence and worsening outcomes associated with CDI,
65
there are few treatment options. Metronidazole and oral vancomycin are the most
66
frequently drugs used for the treatment of CDI and are recommended by current
67
guidelines [3-5].
68
Tigecycline is a broad spectrum glycylcycline antibiotic, approved in 2005 for the
69
treatment of complicated skin and soft tissue infections, intra-abdominal
70
infections and subsequently for community-acquired pneumonia. Additionally, it
71
is active in vitro against C. difficile isolates [6]. Faecal concentrations of
72
tigecycline are generally much higher than the MIC for C. difficile (range of 3.0-
73
14.1 µg/g faeces) due to the biliary elimination of the drug [7]. Also, some in vitro
74
studies suggested that sub-MIC concentrations of tigecycline have inhibitory
75
effects on both toxin production and sporulation of most C. difficile isolates,
77 78
SC
M AN U
TE D
EP
AC C
76
RI PT
61
probably by inhibiting protein synthesis [8,9]. The biological significance of the inhibition of sporulation by sub-inhibitory concentrations of antibiotics in vitro seemed promising for CDI treatment but it remains to be determined in vivo.
79
Nevertheless, a murine model showed that tigecycline may alter the microbiota
80
leading to susceptibility to CDI [10]. Because many patients with CDI need to
81
continue receiving antibiotics for other concomitant infections, tigecycline might 3
ACCEPTED MANUSCRIPT
be an interesting option in such special situations. This drug is mentioned in
83
European Society of Clinical Microbiology and Infectious Diseases (ESCMID)
84
guidelines for CDI treatment with a grade C-III recommendation (“marginally
85
supported recommendation for use – expert opinion evidence”) and reserved for
86
severe CDI cases when oral therapy is not possible [3]. A recently published
87
retrospective study, comparing the administration of tigecycline monotherapy in
88
severe CDI with vancomycin plus metronidazole suggested a better outcome,
89
regarding clinical cure, in patients treated with tigecycline [11].
90
The aim of this study was to compare the outcomes of patients with CDI treated
91
with tigecycline in combination with vancomycin with those treated with
92
vancomycin alone.
93
Methods
94
Study design and patients
95
A retrospective cohort study was conducted in a tertiary care hospital in Romania
96
with 280 beds in adult departments. All adult patients (age ≥18 years) hospitalized
97
for incident CDI from September 2014 through August 2015 were eligible for
98
inclusion in this study. Patients were identified by reviewing the microbiology
99
records. Diagnosis of CDI was based on clinical symptoms (presence of
101 102
SC
M AN U
TE D
EP
AC C
100
RI PT
82
diarrhoea) and detection of the gene coding for C. difficile toxin in stool using polymerase chain reaction. Diarrhoea was defined according to ESCMID guidelines as loose stools corresponding to Bristol stool chart types 5–7, plus a
103
stool frequency of at least three loose stools within 24 hours [12]. We included
104
patients treated with intravenous tigecycline plus oral vancomycin for at least 48
4
ACCEPTED MANUSCRIPT
hours or with oral vancomycin alone. Patients who received treatment with
106
metronidazole, fidaxomicin, tigecycline in monotherapy or faecal microbiota
107
transplantation were excluded. We also excluded the patients for whom the data
108
to calculate the ATLAS severity score were not available, the patients with
109
history of previous CDI and patients who died in the first 48 hours after the
110
diagnosis of CDI [13]. The patients were followed for 8 weeks after the resolution
111
of CDI episode. Hospital ethics committee approved the study.
112
Study variables and definitions
113
The primary outcome of this study was the proportion of patients with favourable
114
outcome, defined as a composite including clinical response (see below), being
115
discharged from hospital and no requirement for surgery or transfer to the
116
intensive care unit due to the severity of CDI. Clinical response was classified as
117
clinical cure if there was a resolution of diarrhoea, abdominal pain, fever and
118
leucocytosis, and no need for addition of another anti-CDI therapy; CDI was
119
considered refractory to standard therapy (vancomycin) when either the stool
120
frequency did not decrease or stool consistency did not improve and/or
121
parameters of disease severity (clinical, laboratory, radiological) did not improve,
122
after at least 3-5 days of therapy, according to ESCMID guidelines [3]. The
124 125
SC
M AN U
TE D
EP
AC C
123
RI PT
105
secondary outcome was CDI recurrence, defined as another episode of CDI within 8 weeks after the onset of a previous episode, provided the symptoms from the previous episode resolved after completion of initial treatment [14]. We routinely
126
followed patients only until discharge, but CDI recurrences are unlikely to be
127
missed given the setup of our healthcare system, with close communication with
5
ACCEPTED MANUSCRIPT
primary care physicians. The primary exposure was treatment with tigecycline
129
plus vancomycin or with vancomycin alone.
130
All the patients included in the final analysis were classified according to
131
Charlson comorbidity score and ATLAS severity score (based on age, treatment
132
with systemic antibiotics, leucocyte count, albumin and temperature) [13, 15].
133
According to ESCMID guidelines, the criteria for severe CDI used in the analysis
134
included were: age over 65 years, fever (core body temperature >38.5°C), marked
135
leucocytosis (leukocyte count >15x109/L), markedly reduced serum albumin (<30
136
g/L), rise in serum creatinine (>50% above the baseline), presence of toxic
137
megacolon or sepsis [3]. We also classified patients according to the Society for
138
Healthcare Epidemiology of America (SHEA) in severe (the presence of
139
leukocytosis of 15x109/L or higher, or serum creatinine level greater than or equal
140
to 1.5 times the premorbid level) and severe complicated (hypotension or shock,
141
ileus, megacolon) CDI [4]. The definition used for toxic megacolon was: presence
142
of radiological signs of distension of the colon (>6 cm in transverse width of
143
colon) and signs of a severe systemic inflammatory response. Sepsis was defined,
144
according to the previously accepted definition, as more than one of the following
145
clinical findings: body temperature higher than 38°C or lower than 36°C, heart
147 148 149
SC
M AN U
TE D
EP
AC C
146
RI PT
128
rate higher than 90/min, hyperventilation evidenced by respiratory rate higher than 20/min or PaCO2 lower than 32 mmHg and white blood cell count higher than 12x109/L or lower than 4x109/L [16]. Statistical analysis
6
ACCEPTED MANUSCRIPT
The data collected were analysed using the statistical analysis software package
151
SPSS 18.0 (IBM SPSS, Chicago, IL, USA). We performed an analysis of all
152
included (AI) patients in which all patients with the above criteria were
153
considered. Differences between groups were analysed using the Mann-Whitney
154
U test for continuous variables and the chi-square test or Fisher's exact test for
155
dichotomous variables as appropriate. Multivariate analysis was performed using
156
binary logistic regression, including the variables with a p value less than 0.1 in
157
the univariate analysis.
158
We constructed a non-parsimonious logistic regression model to calculate a
159
propensity score (PS) for receiving tigecycline. For the AI analysis, the following
160
covariates were included for calculating the PS: sex, age, Charlson's index,
161
albumin, fever, creatinine, leukocytes, other concomitant infection, , ATLAS
162
score, toxic megacolon and sepsis criteria. The PS model showed an area under
163
the receiver operating characteristic curves (AUROC) of 0.84, showing a good
164
predictive ability. The PS was used first as a covariate in the multivariate analysis
165
and to match patients receiving tigecycline and vancomycin with those receiving
166
vancomycin alone using calipers of 0.2 width of the standard deviation of the logit
167
of the propensity score.
169 170
SC
M AN U
TE D
EP
AC C
168
RI PT
150
Crude and multivariate analyses in the matched cohorts were performed by conditional logistic regression. Multivariate analyses of factors associated with a favourable outcome were performed using logistic regression; variables with a p
171
value <0.1 in the univariate analysis and the propensity score for receiving
172
tigecycline were introduced as predictors. Potential interactions of interested were
7
ACCEPTED MANUSCRIPT
also studied (i.e. treatment with tigecycline and presence of other infection or
174
treatment with tigecycline and presence of sepsis). The goodness of fit of the final
175
model was checked with the Hosmer-Lemeshow test, and the prediction ability by
176
calculating the area under the receiver operating curve (ROC). The analyses were
177
also performed for patients treated for seven days or more (T7D) with either
178
tigecycline or vancomycin alone; in this analysis, patients who died in <7 days
179
were also included and considered as failures.
180
Results
181
A total of 660 CDI episodes, recorded in 549 patients, were screened for inclusion
182
in the study period. Of those, 266 patients with CDI were eligible and were
183
included in the AI analysis; the reasons for exclusion are shown in Figure 1.
184
Tigecycline was used to treat 62 episodes and vancomycin alone was used in 204
185
episodes. The baseline characteristics of the patients are summarized in Table 1.
186
Overall, the patients treated with tigecycline were more likely to have had severe
187
CDI according to ESCMID guidelines and a concomitant infection. The median
188
duration of tigecycline administration in the AI analysis was 9 days (interquartile
189
range, IQR: 7-14 days). In 36 (58%) cases in the tigecycline group, this was
190
added to vancomycin from the first day of CDI treatment. For the rest 26 (42%)
192 193
SC
M AN U
TE D
EP
AC C
191
RI PT
173
cases, tigecycline was initiated after a median of 1 day (IQR: 1-4 days). The reasons for adding tigecycline to vancomycin were: CDI severity in 33 (53%) cases, the presence of a concomitant infection in 19 (31%) cases, refractory CDI
194
in 5 (8%) cases, and unknown reason in 5 (8%) cases. Concerning the outcomes
195
(Table 1), 50 (81%) of cases in the tigecycline group and 193 (95%) of those in
8
ACCEPTED MANUSCRIPT
the vancomycin group met the criteria for clinical cure in the AI analysis
197
(p<0.001). The rate of CDI recurrences was 8 (13%) in the tigecycline group
198
versus 39 (19%) in the vancomycin group, respectively (p=0.2).
199
We analysed the CDI episodes according to SHEA classification, in severe CDI
200
and severe complicated CDI. For severe CDI, a favourable outcome was seen in
201
31/39 (80%) patients receiving tigecycline and vancomycin versus 67/73 (92%) in
202
patients receiving vancomycin only (p=0.06). For severe complicated CDI, a
203
favourable outcome was noted in 3/10 (30%) patients receiving tigecycline and
204
vancomycin versus 2/2 (100%) patients receiving vancomycin only (p=0.3).
205
The univariate analysis of factors associated with unfavourable outcome is shown
206
in table 2. We then performed a multivariate analysis, adjusted by the PS (Table
207
3). Interactions between tigecycline and concomitant infection and sepsis were
208
investigated but were not significant, the p value for the Hosmer-Lemeshow test
209
was 0.25, and the AUROC for observed data was 0.81 (95% CI: 0.70-0.92),
210
showing good predictive ability. The negative association of tigecycline and
211
favourable outcome found in the univariate analysis disappeared but we could not
212
find a trend towards higher probability of favourable outcome with tigecycline.
213
The analysis was repeated excludingthe other variables or including several
215 216
SC
M AN U
TE D
EP
AC C
214
RI PT
196
interactions, but no significant changes in the lack of association between therapy with tigecycline and favourable outcome were found. We then matched patients receiving tigecycline and vancomycin with patients
217
receiving vancomycin alone, according to the PS; 86 patients (43 pairs) could be
218
matched. In the cohort of matched patients by the PS we found no significant
9
ACCEPTED MANUSCRIPT
difference between patients treated with vancomycin alone or those treated with
220
the combination of vancomycin and tigecycline, suggesting that the PS worked
221
(data not shown). The OR for favourable outcome with tigecycline in the matched
222
analysis was 0.92 (95% CI: 0.60-1.44; p=0.74). The rate of CDI recurrences was
223
6/43 (14%) in the tigecycline group versus 12/28 (28%) in the vancomycin group
224
(p=0.27).
225
In the T7D analysis, 251 patients were included; 47 were treated with tigecycline
226
and 204 cases with vancomycin monotherapy. The features of patients included in
227
the T7D analysis are shown in Table 1. The adjusted OR for favourable outcome
228
of patient treated with tigecycline in this population was 0.63 (95% CI: 0.15-2.61;
229
p=0.52).
230
Discussion
231
In this observational study, we could not find that treatment of CDI with
232
tigecycline plus oral vancomycin in combination was associated with better
233
outcomes than treatment with vancomycin alone after attempting to control for
234
confounders and prescription bias.
235
Tigecycline is frequently considered whenever the patients need an antibiotic to
236
treat a concomitant infection as this drug could also contribute to the cure of CDI.
238 239 240
SC
M AN U
TE D
EP
AC C
237
RI PT
219
In our setting tigecycline has been frequently used in severely ill patients, especially when an additional source of sepsis beyond CDI cannot be excluded and in patients who require antibiotics for a concomitant infection; in the vast majority of cases, tigecycline is used in association with oral vancomycin.
10
ACCEPTED MANUSCRIPT
In our study, patients in the tigecycline group had a more severe CDI and more
242
frequent concomitant infections. We attempted to control for the obvious
243
prescription bias with propensity score matching; analyses of this subgroup did
244
not show an association between addition of tigecycline and improved outcomes.
245
Furthermore, the rate of CDI recurrences was not significantly lowered after the
246
treatment with tigecycline, both in patients with history of previous CDI and in
247
those with primary episodes. In contrast to our findings, a recently published
248
retrospective study compared the outcomes of patients with severe CDI (median
249
ATLAS score, 8) treated with tigecycline as monotherapy with those treated with
250
oral vancomycin plus intravenous metronidazole. The cure rate with tigecycline
251
was 75.6%, which was significantly higher than in the standard therapy group
252
(53.3%) [11]. However, the results of the multivariate analyses show a very wide
253
95% confidence interval making the results difficult to interpret.
254
The main limitations of our study are its observational and retrospective nature.
255
Despite the fact that we tried to control for confounders using multivariate
256
analysis and propensity score-based matching, we cannot exclude the influence of
257
unmeasured confounders or residual confounding, such as confounding by
258
indication (prescription bias). The propensity score is limited in its ability to
260
SC
M AN U
TE D
EP
AC C
259
RI PT
241
mitigate prescription bias, because it mainly serves to reduce power and thus allows for a type II error (assuming there is no difference when there is one).
261
Also, we excluded a high number of patients due to missing data, but the rate of
262
favourable outcomes among these was similar to those included. Another
263
limitation is that we could not analyse the duration and types of concomitant
11
ACCEPTED MANUSCRIPT
antimicrobial therapy other than tigecycline due to a large variety of
265
antimicrobials administered.
266
In conclusion, adding tigecycline to the standard therapy of CDI did not
267
significantly increase the clinical cure nor reduced the rate of CDI recurrences.
268
Randomised controlled trials assessing tigecycline as adjuvant therapy are needed.
269
In the meantime, the clinical value of tigecycline as additional therapy remains
270
undefined, and we can only stress the importance of limiting the unnecessary use
271
of a broad spectrum antimicrobial in the context of an increasing antimicrobial
272
resistance.
M AN U
SC
RI PT
264
273
Acknowledgments
275
JRB received funding for research from Ministerio de Economía y
276
Competitividad, Instituto de Salud Carlos III - co-financed by European
277
Development Regional Fund "A way to achieve Europe" ERDF, Spanish Network
278
for the Research in Infectious Diseases (REIPI RD16/0016).
TE D
274
281 282 2831 284 285
The authors have nothing to disclose.
AC C
280
EP
279
References
1. Reveles KR, Lee GC, Boyd NK, Frei CR. The rise in Clostridium
difficile infection incidence among hospitalized adults in the United States: 20012010. Am J Infect Control 2014;42:1028–32.
12
ACCEPTED MANUSCRIPT
2. Vindigni SM, Surawicz CM. C. difficile Infection: Changing Epidemiology and
287
Management Paradigms. Clin Transl Gatroenterol 2015;6:e99.
2883
3. Debast SB, Bauer MP, Kuijper EJ. European Society of Clinical Microbiology
289
and Infectious Diseases: Update of the Treatment Guidance Document for
290
Clostridium difficile Infection. Clin Microbiol Infect 2014;20:1–26.
2914
4. Stuart HC, Dale NG, Stuart J, Ciaran PK, Vivian GL, McDonald LC, et al.
292
Clinical Practice Guidelines for Clostridium difficile Infection in Adults: 2010
293
Update by the Society for Healthcare Epidemiology of America (SHEA) and the
294
Infectious Diseases Society of America (IDSA). Infect Control Hosp Epidemiol
295
2010;31:431-55.
2965
5. Surawicz CM, Brandt LJ, Binion DG, Ananthakrishnan AN, Curry SR,
297
Gilligan PH, et al. Guidelines for Diagnosis, Treatment, and Prevention of
298
Clostridium difficile Infections, Am J Gastroenterol 2013;108:478–98. quiz499.
2996
6. Britt NS, Steed ME, Potter EM, Clough LA. Tigecycline for the Treatment of
300
Severe and Severe Complicated Clostridium difficile Infection. Infect Dis Ther
301
2014;3:321–31.
3027
7. Nord CE, Sillerstrom E, Wahlund E. Effect of Tigecycline on Normal
303
Oropharyngeal and Intestinal Microflora. Antimicrob Agents Chemother
3058 306 307
SC
M AN U
TE D
EP
AC C
304
RI PT
2862
2006;50:3375–80.
8. Aldape MJ, Heeney DD, Bryant AE, Stevens DL. Tigecycline suppresses toxin
A and B production and sporulation in Clostridium difficile. J Antimicrob Chemother 2015;70:153–9.
13
ACCEPTED MANUSCRIPT
9. Garneau JR, Valiquette L, Fortier LC. Prevention of Clostridium difficile spore
309
formation by sub-inhibitory concentrations of tigecycline and
310
piperacillin/tazobactam. BMC Infect Dis 2014;14:29.
311
10. Bassis CM, Theriot CM, Younga VB. Alteration of the murine gastrointestinal
312
microbiota by tigecycline leads to increased susceptibility to Clostridium difficile
313
infection. Antimicrob Agents Chemother 2014; 58(5): 2767-2774.
314
11. Gergely SB BG, Kadar B, Lenart KS, Dezsenyi B, Kunovszki P, Fried K, et
315
al. Use of intravenous tigecycline in patients with severe Clostridium difficile
316
infection: a retrospective observational cohort study. Clin Microbiol Infect
317
2016;22:990-5.
318
12. Crobach MJT, Planche T, Eckert C, Barbut F, Terveer EM, Dekkers OM,
319
Wilcox MH, Kuijper EJ. European Society of Clinical Microbiology and
320
infectious Diseases: update of the diagnostic guidance document for Clostridium
321
difficile infection. Clin Microbiol Infect 2016; 22(s4): s63-s81.
322
13. Miller MA, Louie T, Mullane K. Derivation and validation of a simple clinical
323
bedside score (ATLAS) for Clostridium difficile infection, which predicts
324
response to therapy. BMC Infect Dis 2013;13:148.
325
14. McDonald LC, Coignard B, Dubberke E, Song X, Horan T, Kutty PK; Ad
327 328
SC
M AN U
TE D
EP
AC C
326
RI PT
3089
Hoc Clostridium difficile Surveillance Working Group. Recommendations for surveillance of Clostridium difficile-associated disease. Infect Control Hosp Epidemiol 2007; 2:140-145.
14
ACCEPTED MANUSCRIPT
15. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of
330
classifying prognostic comorbidity in longitudinal studies: development and
331
validation. J Chronic Dis. 1987;40(5):373-83.
332
16. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, et al. Definitions for
333
sepsis and organ failure and guidelines for the use of innovative therapies in
334
sepsis. The ACCP/SCCM Consensus Conference Committee. American College
335
of Chest Physicians/Society of Critical Care Medicine. Chest. 1992
336
Jun;101(6):1644-55.
AC C
EP
TE D
M AN U
SC
RI PT
329
15
ACCEPTED MANUSCRIPT
Table 1 – Characteristics and outcomes of patients. Data are expressed as number of patients (percentage) except where specified.
89 (44) 68 (57-78) 4 (3-5) 13 (6.07)
22 (36)
81 (40)
9 (15) 2 (3) 5 (8) 3 (5) 3 (5) 2 (3) 2 (3) 30 (48)
18 (9) 8 (4) 25 (12) 9 (4) 4 (2) 7 (3) 7 (3) 55 (27)
12 (40) 11 (37) 7 (23)
EP
AC C
p*
0.9 0.1 0.3 <0.001
20 (43) 69 (61-81) 4 (3-5) 12 (4.2)
SC
27 (44) 70 (61-81) 4 (3-5) 10 (4.8)
89 (44) 68 (57-78) 4 (3-5) 13 (6.07)
0.9 0.1 0.2 0.09
19 (40)
81 (40)
0.9
0.3 0.8 0.5 0.8 0.4 0.9 0.9 0.002
7 (15) 4 (9) 4 (9) 2 (4) 1 (2) 2 (4) 2 (4) 23 (49)
18 (9) 8 (4) 25 (12) 9 (4) 4 (2) 6 (3) 7 (3) 55 (27)
0.3 0.3 0.6 0.9 0.9 0.9 0.7 0.006
19 (35) 31 (56) 5 (9)
0.8 0.1 0.1
8 (35) 9 (39) 6 (26)
19 (35) 31 (56) 5 (9)
0.9 0.2 0.1
2 (7) 7 (23) 9 (30) 8 (27) 1 (3) 3 (10)
5 (9) 4 (7) 9 (16) 22 (40) 1 (2) 12 (22)
0.7 0.07 0.2 0.3 0.6 0.3
2 (9) 6 (26) 7 (30) 5 (22) 1 (4) 2 (9)
5 (9) 4 (7) 9 (16) 22 (40) 1 (2) 12 (22)
0.9 0.05 0.2 0.2 0.5 0.3
6 (10) 19 (31) 39 (63) 32 (52) 18 (29) 36 (58) 65 (74) 5 (4-6) 39 (63)
1 (0.5) 7 (3) 119 (58) 71 (35) 31 (15) 62 (30) 100 (49) 3 (2-4) 73 (36)
<0.001 <0.001 0.5 0.001 0.01 <0.001 <0.001 <0.001 <0.001
4 (9) 16 (34) 30 (64) 26 (55) 14 (30) 28 (60) 36 (77) 5 (4-6) 30 (64)
1 (0.5) 7 (3) 119 (58) 71 (35) 31 (15) 62 (30) 100 (49) 3 (2-4) 73 (36)
0.003 <0.001 0.6 0.01 0.03 <0.001 0.001 0.005 <0.001
10 (16)
2 (1)
<0.001
6 (13)
2 (1)
<0.001
48 (77) 8 (13)
193 (95) 39 (19)
<0.001 0.2
40 (85) 7 (15)
193 (95) 39 (19)
0.05 0.6
M AN U
0.8
TE D
Male sex Age in years (median, IQR) Median Charlson index (IQR) Length of CDI therapy in days (mean, SD) Underlying diseases Cardiovascular disease and/or hypertension Solid malignancy Hematologic malignancy Diabetes mellitus Renal failure Chronic respiratory disease Liver failure HIV infection Concomitant infection Aetiology of concomitant infection Unknown Monomicrobial Polymicrobial Site of concomitant infection Skin Intraabdominal Respiratory Urinary Osteoarticular Other Severity criteria Toxic megacolon Sepsis Age over 65 years Fever Acute renal insufficiency WBC >15x109/L Albumin <3 g/L Median ATLAS score (IQR) Severe CDI (according to SHEA guidelines) Severe complicated CDI (according to SHEA guidelines) Outcomes Favourable outcome Recurrence
Treated for seven days or more Tigecycline + Vancomycin Vancomycin p* (n=204) (n=47)
RI PT
All included patients Tigecycline + Vancomycin Vancomycin (n=204) (n=62)
ACCEPTED MANUSCRIPT
Transfer to another unit In-hospital mortality
9 (15) 5 (8)
6 (3) 5 (2)
<0.001 0.09
3 (6) 4 (9)
IQR: Interquartile range. SD: standard deviation. CDI: Clostridium difficile infection. WBC: White blood cell count.
AC C
EP
TE D
M AN U
SC
RI PT
(*) Chi-square (for categorical features) or U Mann-Whitney for continuous variables
6 (3) 5 (2)
0.4 0.1
ACCEPTED MANUSCRIPT
Table 2. Univariate analysis of variables associated with favourable outcome in all
RI PT
included patients. Data are number of patients (percentage) except where specified.
Favourable outcome (n=244)
OR (95% CI)
Male sex
10 (46)
106 (43)
1.08 (0.4-2.6)
Age over 65 years
14 (64)
149 (61)
0.9 (0.3-2.2)
79 (46-82)
69 (57-79)
-
5 (3-6)
4 (3-5)
-
Median Charlson index (IQR) Concomitant infection
M AN U
Median age in years (IQR)
SC
Unfavourable outcome (n=22)
6 (27)
79 (32)
1.2 (0.4-3.3)
6 (27)
1 (0.4)
0.01 (0.001-0.09)
4 (18)
22 (9)
0.4 (0.1-1.4)
8 (36)
95 (39)
1.1 (0.4-2.7)
1.3 (1.0-2.8)
0.9 (0.7-1.5)
-
9 (41)
49 (20)
0.3 (0.1-0.8)
15.8 (4.8-24.0)
11.7 (8.2-17.6)
-
12 (55)
86 (35)
0.4 (0.1-1.09)
2.6 (1.9-2.9)
2.9 (2.5-3.6)
-
Albumin <3 g/L
18 (82)
128 (53)
0.2 (0.08-0.7)
Median ATLAS score (IQR)
4 (3-6)
3 (2-5)
-
Severe CDI (according to SHEA guidelines)
100 (88)
14 (12)
2.5 (1.1-6.2)
Other antibiotic
62 (90)
181 (92)
1.4 (0.5-3.5)
Treatment with tigecycline
12 (55)
50 (21)
0.2 (0.08-0.5)
Toxic megacolon Sepsis Fever
TE D
Creatinine (mg/dL) (median, IQR) Acute renal insufficiency
WBC (x109) (median, IQR)
EP
WBC >15x109/L
AC C
Albumin (g/L) (median, IQR)
ACCEPTED MANUSCRIPT
IQR: Interquartile range. CDI: Clostridium difficile infection. WBC: White blood cell count.
AC C
EP
TE D
M AN U
SC
RI PT
(*) Chi-square (for categorical features) or U Mann-Whitney for continuous variables
ACCEPTED MANUSCRIPT
Table 3 – Multivariate analysis for characteristics associated with favourable
All included
PS-adjusted
PS matched
patients (n=266)
(n=266)
cohort (n=86)
OR (95%
p
0.8(0.6-
one point increase)
1.0)
to SHEA guidelines)
1x10 -3
0.9
<0.01
(0.0-0.1)
AC C
ATLAS score (per one point increase)
Treatment with tigecycline
0.8 (0.6-
OR
(95%
(95%
CI)
CI)
0.8 (0.6-
0.1
1.1)
3.1)
EP
Toxic megacolon
0.9(0.3-
0.08
TE D
Severe CDI (according
p
M AN U
CI)
Charlson index (per
OR
SC
Variables
RI PT
outcome in all included population, adjusted by the propensity score.
1.9 (0.3-
0.9
3.2)
1.1)
0.6
(0.91.3) 1.1
0.7
(0.52.3)
6.0 x10-3
<0.01
0.8 (0.6-
2.1x10-
0.9
6
(0.0-0.1) 0.2
1.0
p
0.1
1.1)
0.9
0.5
(0.71.2)
0.5 (0.21.5)
0.2
0.4 (0.11.3)
0.1
1.0 (0.61.5)
0.8
ACCEPTED MANUSCRIPT
Included all variables associated with favorable outcome, with a p-value < 0.1, in the univariate analysis (table 2) PS: propensity score
AC C
EP
TE D
M AN U
SC
RI PT
CDI: Clostridium difficile infection
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT