Estimation of adjusted expected excess length-of-stay associated with
ventilation-acquired pneumonia in intensive care: A multistate approach
accounting for time-dependent mechanical ventilation.
Bluhmki T(1), Allignol A(2), Ruckly S(3), Timsit JF(3)(4), Wolkewitz M(5),
Beyersmann J(1).
Author information:
(1)Institute of Statistics, Ulm University, Ulm, Germany.
(2)Merck KGaA, Darmstadt, Germany.
(3)UMR 1137 IAME Inserm/University Paris Diderot, Paris, France.
(4)APHP, Bichat Hospital, Intensive Care Unit, Paris, France.
(5)Institute for Medical Biometry and Statistics, Faculty of Medicine and
Medical Center-University of Freiburg, Freiburg, Germany.
The expected excess length-of-stay is an established concept to assess the
health and economic impact of nosocomial, that is, hospital-acquired infections
such as ventilation-acquired pneumonia in intensive care. Estimation must
account for the timing of infection as in a multistate perspective, because
common retrospective comparisons yield inflated estimates due to time-dependent
bias. Since occurrence of ventilation-acquired pneumonia is closely linked to
ventilation status, we suggest a multistate model incorporating time-dependent
mechanical ventilation as additional states. The appeal is that the expected
excess length-of-stay decomposes into extra days spent under ventilation and not
under ventilation. This is not only highly relevant from a patient's perspective
regarding quality of life, but also from an economic point of view, because
ventilation is a major cost driver. The challenge is that estimation involves
complex functionals of the matrix of transition probabilities, which in turn are
based on the transition hazards. To address heterogeneity between patients,
which is a common phenomenon in observational hospital epidemiology, we apply
pseudovalue regression to adjust the ventilation-specific quantities for
baseline confounding. The performance of our proposal is assessed by simulation
and the methods are illustrated on data provided by 12 French intensive care
units. Preliminary results indicate that the expected excess length-of-stay
associated with ventilation-acquired pneumonia is mainly triggered by extra days
spent under mechanical ventilation, and that the excess is most pronounced for
intensive care patients with fewer comorbidities at baseline. We also find that
such a decomposition is challenging for early times. Example code is provided.
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.
PMID: 30198195 [Indexed for MEDLINE]
Association Between Body Weight Variation and Survival and Other Adverse Events
in Critically Ill Patients With Shock: A Multicenter Cohort Study of the
OUTCOMEREA Network.
Gros A(1), Dupuis C(2)(3), Ruckly S(4), Lautrette A(5), Garrouste-Orgeas
M(2)(6), Gainnier M(7), Forel JM(8), Marcotte G(9), Azoulay E(10), Cohen Y(11),
Schwebel C(12), Argaud L(9), de Montmollin E(2)(13), Siami S(14),
Goldgran-Toledano D(15), Darmon M(16), Timsit JF(2)(3); OUTCOMEREA Study Group.
OBJECTIVES: This study in critically ill patients with shock assessed the
prognostic value of body weight variations occurring each day from day 3 to day
7 on the 30-day outcome in terms of mortality, occurrence of
ventilator-associated pneumonia and of bedsore, and occurrence of length of
stay.
DESIGN: Retrospective analysis of data. Multivariate subdistribution survival
models were used at each day, from day 3 to day 7. The impact of body weight
variations on length of stay was estimated through a multivariate negative
binomial regression model.
SETTING: Prospective multicenter cohort study.
PATIENTS: Critically ill patients admitted in ICU with shock and requiring
mechanical ventilation within 48 hours.
INTERVENTION: None.
MEASUREMENTS AND MAIN RESULTS: Two-thousand three-hundred seventy-four patients
were included. Their median body weight variations increased from 0.4 kg
(interquartile range, 0-4.8 kg) on day 3 to 3 kg (interquartile range, -0.4 to
8.2 kg) on day 7. Categories of body weight variations were defined depending on
body weight variations interquartiles: weight loss, no weight gain, moderate and
severe weight gain. A severe weight gain tended to be associated with death at
days 5 and 6 (day 5: subdistribution hazard ratio, 1.27; 95% CI, 0.99-1.63; p =
0.06 and day 6: subdistribution hazard ratio, 1.43; 95% CI, 1.08-1.89; p =
0.01), a weight loss tended to be associated with bedsore, and a severe gain
between at days 5 and 6 was associated with ventilator-associated pneumonia. Any
body weight variations were associated with an increased length of stay.
CONCLUSIONS: In survivors at day 3, body weight variations during the first days
of ICU stay might be a clinically relevant tool to prevent weight gain but also
for prognostication of 30-day mortality, occurrence of ventilator-associated
pneumonia, and occurrence of prolonged ICU stay.
DOI: 10.1097/CCM.0000000000003338
PMID: 30059364 [Indexed for MEDLINE]
The Impact of Early Adequate Treatment on Extubation and Discharge Alive of
Patients With Pseudomonas aeruginosa-Related Ventilator-Associated Pneumonia.
Sommer H(1), Timsit JF(2)(3), von Cube M(1), Schumacher M(1), Darmon M(4),
Schwebel C(5), Ruckly S(6), Wolkewitz M(1); COMBACTE-MAGNET Consortium.
OBJECTIVES: We aim to examine the effect of early adequate treatment in
comparison with inadequate or delayed treatment on being extubated or discharged
alive over time, in patients with Pseudomonas aeruginosa-related
ventilator-associated pneumonia.
DESIGN: Retrospective analyses of a prospective observational multicenter cohort
study.
SETTING: ICU.
PATIENTS: Patients of the French prospective database (OUTCOMEREA) were included
if they acquired a ventilator-associated pneumonia due to P. aeruginosa between
1997 and 2014 and were mechanically ventilated for more than 48 hours.
INTERVENTIONS: Early adequate treatment in comparison with inadequate or delayed
adequate treatment.
MEASUREMENTS AND MAIN RESULTS: Multistate models were applied to estimate the
time-dependent probability of being extubated or discharged alive, and separate
Cox regression analyses were used to assess the treatment effect on all
important events that influence the outcome of interest. A propensity
score-adjusted innovative regression technique was used for a combined and
comprehensive patient-relevant summary effect measure. No evidence was found for
a difference between adequate and inadequate or delayed treatment on being
extubated or discharged alive. However, for all patients, the probability of
being extubated or discharged alive remains low and does not exceed 50% even 40
days after a P. aeruginosa-related ventilator-associated pneumonia.
CONCLUSIONS: Early adequate treatment does not seem to be associated with an
improved prognosis. Its potential benefit requires further investigation in
larger observational studies.
DOI: 10.1097/CCM.0000000000003305
PMID: 29985212 [Indexed for MEDLINE]
Diabetes was the only comorbid condition associated with mortality of invasive
pneumococcal infection in ICU patients: a multicenter observational study from
the Outcomerea research group.
Garrouste-Orgeas M(1)(2)(3), Azoulay E(4), Ruckly S(5), Schwebel C(6), de
Montmollin E(7), Bedos JP(8), Souweine B(9), Marcotte G(10), Adrie C(11),
Goldgran-Toledano D(12), Dumenil AS(13), Kallel H(14), Jamali S(15), Argaud
L(16), Darmon M(17)(18), Zahar JR(19), Timsit JF(20)(5)(21).
PURPOSES: Streptococcus pneumoniae is a leading pathogen of severe community,
hospital or nursing facility infections. We sought to describe characteristics
of invasive pneumococcal infection (IPI) and pneumonia (due to the high
mortality of intensive care-associated pneumonia) and to report outcomes
according to various types of comorbidity.
METHODS: Multicenter observational cohort study on the prospective Outcomerea
database, including adult patients, with a hospital stay < 48 h before ICU
admission and a documented IPI within the first 72 h of ICU admission. Comorbid
conditions were defined according to the Knaus and Charlson classification.
RESULTS: Of the 20,235 patients, 5310 (26.4%) had an invasive infection,
including 560/5,310 (10.6%) who had an IPI. The ICU 28-day mortality was 109/560
(19.8%). Four factors were independently associated with mortality: SOFA day
1-2: [hazard ratio (HR) 1.21; 95% confidence interval (95% CI) 1.15-1.27,
p < 0.001]; maximum lactate level day 1-2: (HR 1.07, 95% CI 1.02-1.12,
p = 0.006); diabetes mellitus: (HR 1.91, 95% CI 1.23-3.03, p = 0.006) and
appropriate antibiotics (HR 0.28, 95% CI 0.15-0.50, p < 0.001). Comparable
results were obtained when other comorbid conditions were forced into the model.
Diabetes impact was more pronounced in case of micro- or macro-angiopathy (HR
4.17, 95%CI 1.68-10.54, p = 0.003), in patients ≥ 65 years old (HR 2.59, 95% CI
1.56-4.28, < 0.001) and in those with body mass index (BMI) < 25 kg/m2 (HR 2.11,
95% CI 1.10-4.06, p = 0.025).
CONCLUSIONS: Diabetes mellitus was the only comorbid condition which
independently influenced mortality in patients with IPI. Its impact was more
pronounced in patients with complications, aged ≥ 65 years and with
BMI < 25 kg/m2.
DOI: 10.1007/s15010-018-1169-6
PMID: 29974388 [Indexed for MEDLINE]
Management and outcomes of acute respiratory distress syndrome patients with and
without comorbid conditions.
Azoulay E(1), Lemiale V(2), Mourvillier B(3), Garrouste-Orgeas M(4), Schwebel
C(5), Ruckly S(6), Argaud L(7), Cohen Y(8), Souweine B(9), Papazian L(10),
Reignier J(11), Marcotte G(12), Siami S(13), Kallel H(14), Darmon M(2), Timsit
JF(15); OUTCOMEREA Study Group.
RATIONALE: The standard of care for patients with acute respiratory distress
syndrome (ARDS) has been developed based on studies that usually excluded
patients with major comorbidities.
OBJECTIVES: To describe treatments and outcomes according to comorbidities in
patients with ARDS admitted to 19 ICUs (1997-2014).
METHODS: Patients were grouped based on comorbidities. Determinants of day-28
mortality were identified by multivariable Cox analysis stratified on center.
MEASUREMENTS AND MAIN RESULTS: Among 4953 ARDS patients, 2545 (51.4%) had major
comorbidities; the proportion with major comorbidities increased after 2008.
Hematological malignancy was associated with severe ARDS and rescue therapies
for refractory hypoxemia. COPD, HIV infection, and hematological malignancy were
associated with a lower likelihood of invasive mechanical ventilation on the
admission day. Admission-day SOFA score was higher in patients with major
comorbidities, who more often received vasopressors, dialysis, or
treatment-limitation decisions. Day-28 mortality was 33.7% overall, 27.2% in
patients without major comorbidities, and 31.1% (COPD) to 56% (hematological
malignancy) in patients with major comorbidities. By multivariable analysis,
mortality was lower in patients with COPD and higher in those with chronic heart
failure, solid tumors, or hematological malignancies. Mortality was
independently associated with PaO2/FiO2 and PaCO2 on day 1, ARDS of pulmonary
origin, worse SOFA score, and ICU-acquired events.
CONCLUSIONS: Half the patients with ARDS had major comorbidities, which were
associated with severe ARDS, multiple organ dysfunction, and day-28 mortality.
These findings do not support the exclusion of ARDS patients with severe
comorbidities from randomized clinical trials. Trials in ARDS patients with
whatever comorbidities are warranted.
DOI: 10.1007/s00134-018-5209-6
PMCID: PMC7095161
PMID: 29881987 [Indexed for MEDLINE]