Accepted Manuscript

An Analysis of Factors Associated with In-Hospital Mortality in Lung Cancer

Chemotherapy Patients with Neutropenia

Julia Cupp, MD, Eva Culakova, PhD, MS, Marek S. Poniewierski, MD, MS, David C.
Dale, MD, Gary H. Lyman, MD, MPH, Jeffrey Crawford, MD
PII: S1525-7304(17)30307-8
DOI: 10.1016/j.cllc.2017.10.013
Reference: CLLC 709

To appear in: Clinical Lung Cancer

Received Date: 11 May 2017

Revised Date: 11 October 2017
Accepted Date: 18 October 2017

Please cite this article as: Cupp J, Culakova E, Poniewierski MS, Dale DC, Lyman GH, Crawford J, An
Analysis of Factors Associated with In-Hospital Mortality in Lung Cancer Chemotherapy Patients with
Neutropenia, Clinical Lung Cancer (2017), doi: 10.1016/j.cllc.2017.10.013.

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

An Analysis of Factors Associated with In-Hospital Mortality in Lung Cancer Chemotherapy Patients with
Neutropenia

Julia Cupp, MDa, Eva Culakova, PhD, MSb, Marek S. Poniewierski, MD, MSb, David C. Dale, MDc, Gary H.
Lyman, MD, MPHb,c, and Jeffrey Crawford, MDa

a
Duke Department of Medicine and Cancer Institute

PT
DUMC 3476
Durham, NC 27710
USA

RI
jcupp@ghs.org
jeffrey.a.crawford@duke.edu (Corresponding author)

SC
b
Hutchinson Institute for Cancer Outcomes Research
Fred Hutchinson Cancer Research Center
1100 Fairview Ave N

U
Seattle, WA 98109
USA
AN
eculak@fredhutch.org
mponiewi@fredhutch.org
glyman@fredhutch.org
M

c
Department of Medicine
RR-512, Health Sciences Building
D

University of Washington
Box 356420
TE

1959 NE Pacific Street

Seattle, WA 98195-6420
USA
dcdale@uw.edu
C EP

Funding disclosure: The ANC Study Group is supported by a research grant from Amgen to Fred
Hutchinson Cancer Research Center (Dr. Lyman, PI). The funding agency was not involved in study design
AC

or the collection, analysis, interpretation or reporting of data. Additional research support provided by
the Duke Stead Resident Research Grant.

Conflict of interest statement for each author is attached.

Abstract

Page 1 of 19
 ACCEPTED MANUSCRIPT

Introduction: Lung cancer, compared to other solid tumors, is associated with high mortality rates from febrile
neutropenia. Risk factors associated with in-hospital mortality were identified and compared for patients with lung
cancer and patients with other solid tumors.

Methods: Hospitalization data from the University Health Consortium database inclusive of the years 2004-2012
were analyzed. The study population included all adult patients with solid tumors who had neutropenia. Cancer
type, presence of neutropenia, and further subgroups were based on ICD-9-CM codes. The primary study outcome
was in-hospital mortality in lung cancer patients vs. other solid tumors. Further analysis concentrated on

PT
comparisons of the two groups.

Results: The analysis was based on 11,111 lung cancer patients and 49,975 patients with other solid tumors.

RI
Overall 4290 (7.0%) patients died. Lung cancer was associated with highest mortality (11.2%, compared with other
solid tumors, 6.1%; p <0.0001). Lung cancer patients were older and more likely to have multiple comorbidities,
and the risk of mortality was directly related to the number of comorbidities. Four additional risk factors for
mortality were identified: pneumonia, sepsis, any infection, and ICU stay. Pneumonia occurred more commonly in

SC
the lung cancer patients (26.4% vs. 10.3%) and was associated with comorbid pulmonary disease, which occurred
more often in the lung cancer patients (52.1% vs. 24.0%).

Conclusions: Lung cancer patients presenting with febrile neutropenia are older, have more comorbidities, have a

U
higher incidence of comorbid pulmonary disease, and are more likely to have pneumonia. Awareness of these risk
factors for mortality should guide the clinician for more personalized approaches to chemotherapy and supportive
AN
care decisions, both at initial treatment and at the time of hospitalization.
M

Introduction
D

Febrile neutropenia (FN) is considered one of the most severe complications of cancer chemotherapy as
it is associated with high mortality.1 The incidence of FN during lung cancer treatment ranges from 1.9%
TE

(during pemetrexed treatment for recurrent non-small cell lung cancer) to 28% (during topotecan
treatment for recurrent small cell lung cancer).2 Of the patients with solid tumors presenting with FN,
lung cancer patients experience the highest mortality rates; FN mortality for lung cancer is surpassed
only by leukemia and lymphoma.1,3-5
EP

An understanding of FN mortality contributes to both preventive and responsive strategies to improve

chemotherapy-related outcomes. Responsive strategies include adjusting the threshold for hospital
admission and intervening early with broad-spectrum antibiotics6 when high-risk FN is identified.
C

Preventive strategies include risk prediction, risk-adjusted chemotherapy dosing,7 and risk-stratified
growth factor support. Mortality data guide optimization of both curative and palliative treatment
AC

plans.8

Retrospective analyses of morbidity and mortality and related risk factors in FN have been reported in
the past.5 Previously identified risk factors for adverse outcomes of established febrile neutropenia
include the following: older age, sepsis, hypotension, organ dysfunction, admission to ICU, bacteremia,
pneumonia, no prophylactic antibiotics, previous fungal infection, various comorbidities including
cardiovascular and pulmonary disease, severity and duration of neutropenia, visceral organ
involvement, and uncontrolled malignancy.9 However, now with the increasing availability and use of G-
CSF, results based on historical data may not accurately reflect current mortality rates and risk factors.

Page 2 of 19
 ACCEPTED MANUSCRIPT

Some updated studies are now becoming available4,10 and lung cancer continues to be associated with
the highest mortality rate of the solid tumors. There is little research available that further explores
these outcomes and associated risk factors specific to the lung cancer population.

The focus of this analysis was to provide updated information regarding mortality in admissions for
neutropenic fever, to identify risk factors most associated with in-hospital mortality, and to describe
their impact on mortality in patients with lung cancer.

PT
Methods

RI
Hospitalization data from the University Health Consortium (UHC) database5,11 were utilized. UHC is the
consortium of 117 academic medical centers and over 337 of their affiliated hospitals from 43 US states

SC
and District of Columbia. The administrative longitudinal UHC database is built based on the information
from discharge summaries. It contains patient level demographic, clinical, inpatient cost data and
hospitalization related data such as length of stay for all admissions within the consortium. Clinical data
are recorded using the ninth revision of the International Classification of Diseases–Clinical Modification

U
(ICD-9-CM) codes including up to 99 positions for the diagnostic codes and 99 positions for the
procedure coding. For this analysis hospitalization data from years 2004 – 2012 from 239 hospitals were
AN
utilized. Data did not include information on laboratory studies, radiologic studies, pathologic reports,
medication utilization, or cause of death. No patient or institutional identifiers were provided to the
investigators, and therefore the research activity was determined to be exempt from the Institutional
M

Review Board governing research involving human subjects. Cancer type, neutropenia, comorbidities,
and infection types were identified by ICD-9-CM codes.

Study population comprised of all adult (>= 18 years of age) patients with solid tumors (ICD-9 CM 140-
D

199.99, 209.0-209.39, 209.7-209.79) who were hospitalized and diagnosed with neutropenia (288.0-
288.09). Lung cancer patients were identified by the presence of the codes within the range 162-163.99
TE

or 209.21. The codes do not differentiate small cell vs. non-small cell lung cancer and information about
cancer stage was not available. Studied comorbidities included diabetes mellitus, heart, hepatic, lung,
renal, cerebrovascular, peripheral vascular disease, and venous thromboembolism.
EP

The primary outcome was in-hospital mortality. Demographic and clinical characteristics of patients with
lung cancer were compared to patients with other solid tumors, concentrating on comparisons of the
two groups with respect to age, number and type of comorbidities, occurrence of sepsis, pneumonia, or
C

any infection, and intensive care unit (ICU) stay, and the influence of these factors on mortality.
AC

Statistical methods: Patients with more than one hospitalization were identified and if multiple
hospitalizations were detected, only one chronologically first hospitalization per patient was included in
the analysis. Patients’ cohort was described reporting proportion of patients within the demographics
and clinical categories of interest. Proportions of events in groups were reported and differences
between the two groups were compared using chi-square test. Multivariable logistic regression was
employed to identify patients with lung cancer at an increased risk of in-hospital mortality. First the
covariates were screened by a stepwise regression process (p=.05 required for the entry and
elimination). During the screening phase age, gender, ethnicity and year of discharge were included in
the model a-priori and the pool of screened clinical factors consisted of number of comorbidities as well

Page 3 of 19
 ACCEPTED MANUSCRIPT

as individual comorbid diseases, infectious variables, in-hospital chemotherapy and need for ICU stay.
The final model included the demographic variables and clinical variables that were associated with
increased risk during the selection process. Interactions between age and the clinical factors were
evaluated. The standard error and associated confidence intervals (CIs) were calculated using Wald’s
method. In addition, to adjust for correlation between observations from the same hospital, robust
standard errors and CI were calculated using the generalized estimating equation method. P-values
<0.05 were considered statistically significant. Statistical analysis was performed using a software

PT
program (SAS version 9.4; SAS Institute Inc., Cary, North Carolina).

RI
Results

The analysis was based on 61,086 hospitalized patients with solid tumors, of whom 11,111 patients

SC
(18.2%) had lung cancer. The other solid tumors included breast cancer (14.0% of the total study
population), colorectal (6.8%), other gastrointestinal (11.5%), gynecologic (8.9%), head and neck (6.6%),
genitourinary (6.4%), sarcoma (6.1%), brain (1.9%) and additionally 1.6% of the population had multiple

U
solid tumors and 18.1% had other or unknown tumor types. Patient demographics are shown in Table 1.
Lung cancer patients were older: 50% of lung cancer patients were age 65 or older, compared to 31.6%
AN
of patients with other solid tumors (p<0.0001). There were relatively more men making up the lung
cancer group (51.8%) as compared to the group of other solid tumors (40.8%). The study population was
spread across the study time period, and each discharge year was well represented, ranging from 7.4%
of the study population discharged in 2004, to 15.2% of the study population discharged in 2012. Table 2
M

summarizes the number and percentage of major comorbidities and infectious complications as well as
incidence of ICU stay. Major comorbid illnesses in addition to solid tumor and FN were reported in
70.8% of patients, with 41.0% reporting 2 or more major comorbidities. Infection was documented in
D

53.9% of the total population, and 16.2% had an ICU stay.

TE

Overall Mortality:

The overall mortality was 7.0%: 4290 patients hospitalized with FN died. Lung cancer was associated
with the highest mortality (11.2%, compared with other solid tumors, 6.1%; p <0.0001). The mortality
EP

rate for each tumor type is shown in Figure 1: lung cancer was followed by brain and multiple solid
tumors for the highest mortality rates, while breast and testicular cancers had the lowest.

Documented infection, sepsis, pneumonia, and ICU admission were found to be important risk factors
C

for mortality, with associated mortality rates in the total population of 11.0%, 31.1%, 19.1%, and 26.1%,
respectively.
AC

Comorbidities:

The risk of mortality was directly related to the number of comorbidities. For the entire study
population, the mortality rate ranged from 0.9% for those patients with no comorbidities to 35.2% for
those with 5 or more. Figure 2 shows this association between number of comorbidities and mortality
for both the lung cancer population and other solid tumors. Greater mortality with increasing number of
comorbidities persisted after adjusting for cancer type (lung vs. other), presence of infection, age,
gender, ethnicity, and year of hospitalization, (odds ratio [OR] for mortality per +1 comorbidity increase:
2.05; 95% CI: 1.99-2.10].

Page 4 of 19
 ACCEPTED MANUSCRIPT

More of the lung cancer patients had higher numbers of comorbidities than the patients with other solid
tumors, as shown in Figure 3 and Table 2. Lung cancer patients were more likely to have multiple (≥2)
comorbidities than patients with other solid tumors (57.3% vs. 37.3%; p<0.0001). In evaluating the
specific types of comorbidities (Figure 4), comorbid heart and lung diseases occurred considerably more
often in the lung cancer patients (heart 66.8% vs. 49.1%; lung 52.1% vs. 24.0%; p<0.0001).

Lung Disease and Pneumonia:

PT
Even independent of number of comorbidities and age, lung cancer patients had higher mortality (OR
1.38, 95% CI: 1.28-1.48). Lung cancer patients had a higher frequency of documented infection, sepsis,
pneumonia, and ICU admission (Figure 5a). In particular, pneumonia occurred considerably more often

RI
in the lung cancer patients (26.4% vs. 10.3%). Lung cancer patients also had higher mortality rates
associated with these risk factors (Figure 5b).

SC
Comorbid lung disease was strongly associated with the development of pneumonia (OR 4.52, 95% CI:
4.30-4.74). With or without lung disease as a comorbidity, lung cancer patients were more likely to have
pneumonia than patients with other solid tumor types (with – 36.0% vs. 22.8%, p<0.0001) (without –

U
16.1% vs. 6.4%, p<0.0001) (Figure 6a).

Figure 6b explores the mortality associated with pneumonia. Overall, as well as of the subgroup of
AN
those patients without comorbid lung disease, patients with lung cancer had higher mortality from
pneumonia than patients with other solid tumors. Patients with comorbid lung disease also had
increased mortality from pneumonia in both populations, and for this subgroup, mortality was not
M

higher among the lung cancer patients compared with the solid tumor patients.

Lung Cancer-Specific Risk factors for Mortality:

D

Independent risk factors for mortality specific to the patients with lung cancer were further clarified
through a multivariable logistic regression model (Table 3). As with the overall population, sepsis and
TE

ICU stay had the highest risk for mortality (OR 4.41 and 3.13, respectively). Pneumonia increased risk of
mortality by an OR of 1.52 and the presence of 2 or more comorbidities increased risk of mortality by an
OR of 1.62. Specific comorbidities, independent of number of comorbidities, also increased risk for
mortality; in particular, comorbid lung disease increased this risk by an OR of 2.02.
EP

Discussion
C

The overall mortality rate for this population – patients with solid tumors hospitalized with FN – was
AC

7.0%. Previously reported mortality rates for hospitalized FN patients derived from electronic health
record and insurance claims databases within the last decade include: 8.1%,12 9.5%,5 10.6%,13 and 14%.4
These numbers are comparable to our study, given the difference in study populations and the
limitations of administrative databases. The higher mortality for lung cancer compared to other solid
tumors and the association of increasing age with higher mortality in this population also replicates
previous findings.1,3-5,8,9

This study found an increased risk of mortality with increasing number of comorbidities and particularly
focused on the increased mortality associated with comorbid lung disease. This is consistent with prior
studies. Comorbidities have been known to be associated with worse outcomes, and of the common

Page 5 of 19
 ACCEPTED MANUSCRIPT

comorbidities, chronic obstructive pulmonary disease (COPD) has been previously shown to be a
particularly important comorbidity increasing risk of mortality. The association between number of
comorbidities and mortality has already been reported for a broader version of this dataset, inclusive of
hematologic malignancies.14 In a similar study with a different database, COPD had increased risk of
mortality in addition to renal disease, diabetes mellitus, congestive heart failure, and pulmonary
embolism.4 The Multinational Association for Supportive Care in Cancer risk index10 includes burden of
illness (which would encompass number and severity of comorbidities) and separately includes

PT
comorbid COPD in its calculation of risk for complications of FN. COPD (as well as chronic cardiovascular
disease) was found to be an important risk-predicting variable in the CISNE, developed on a stable,
ambulatory, solid tumor FN population from 1996-2004.15 Given the consistent reporting of number of

RI
comorbidities and specifically comorbid COPD as important risk predictors for mortality in FN, the
evidence provided by this study that lung cancer patients have considerably higher rates of these risk
factors than patients with other solid tumors helps explain the higher mortality seen in the lung cancer

SC
population.

The finding of pneumonia as an important variable associated with higher risk for mortality has been
supported by previous studies which found that pneumonia increased independent risk for mortality

U
and serious complications of FN patients both in presentation to emergency departments16 and during
their hospitalization.13 While this study reported a mortality rate of 19.1%, mortality rates from FN
AN
pneumonia have been reported as high as 37.1%.4 The higher frequency of pneumonia in the lung
cancer patients compared to the other solid tumors could be mechanistically explained with multiple
reasons. Comorbid lung disease, as expected, was associated with higher incidence of pneumonia, and
M

this comorbidity was much more frequent in the lung cancer population. However, the higher rate of
comorbid lung disease did not fully explain the higher incidence of pneumonia in the lung cancer
patients, as their incidence of pneumonia was higher even when correcting for the presence of
D

comorbid COPD; the higher incidence of pneumonia may partially be an inherent risk of the lung cancer
itself. Anatomic complications of solid tumors have been understood to predispose to particular
TE

infection types: in this case, an endobronchial tumor predisposes to post-obstructive pneumonia.17 In

lung cancer patients, anatomical complications of the solid tumor as well as increased rates of comorbid
lung disease may independently increase the risk for pneumonia, an infectious complication of FN which
EP

has been established to have high risk of mortality.

Interestingly, the mortality of pneumonia changed considerably with the presence or absence of both
lung cancer and COPD. Although overall, and in the absence of comorbid lung disease, lung cancer
C

patients had higher mortality from pneumonia than other solid tumors, this difference disappeared for
the subgroup of lung cancer and other solid tumor patients with comorbid lung disease. This suggests
AC

that, in terms of mortality from pneumonia, comorbid lung disease is a more important factor than lung
cancer. The high rates of comorbid lung disease found in the lung cancer population may increase not
only the incidence of pneumonia but also the mortality from pneumonia.

Limitations: This analysis is based on an administrative database and information about clinical
conditions was based on the presence (or absence) of the billing codes. This may have underestimated
the frequency of clinical events which may have occurred but were not included in the patient’s billing.
The database lacked any information about cancer stage, chemotherapy regimens and doses,
performance status, tobacco smoking, socioeconomic status, or laboratory values such as severity and
duration of neutropenia, all of which would likely add important prognostic information.7,8,17

Page 6 of 19
 ACCEPTED MANUSCRIPT

In summary, lung cancer patients presenting with febrile neutropenia have a higher mortality rate in
comparison to patients with other solid tumors. The lung cancer patients are older, have more
comorbidities, have a higher incidence of comorbid lung disease, and are more likely to have
pneumonia; all of these have been identified as independent risk factors for mortality both in the overall
solid tumor population and also specifically in patients with lung cancer. In order to reduce the mortality
of chemotherapy in lung cancer patients, careful pretreatment assessment and optimal supportive care
during therapy are critical. Greater awareness of risk factors associated with poor prognosis during

PT
hospitalization for febrile neutropenia may help to identify fragile patients at the beginning of
hospitalization or even earlier, helping to guide more personalized cancer treatment and supportive
care, potentially improving clinical outcomes.

RI
U SC
AN
M
D
TE
C EP
AC

Page 7 of 19
 ACCEPTED MANUSCRIPT

Acknowledgements: Thanks to the International Association for the Study of Lung Cancer (IASLC), which
approved this abstract as poster in the 2015 World Conference on Lung Cancer. The ANC Study Group is
supported by a research grant from Amgen to Fred Hutchinson Cancer Research Center (Dr. Lyman, PI).
The funding agency was not involved in study design or the collection, analysis, interpretation or
reporting of data. Additional research support provided by the Duke Stead Resident Research Grant.

PT
RI
U SC
AN
M
D
TE
C EP
AC

Page 8 of 19
 ACCEPTED MANUSCRIPT

Table 1: Patients’ Characteristics of Patients (Overall and Stratified by Lung Cancer vs. Other Solid
Tumors).

Category All Solid Tumors Lung Cancer Other Solid Tumors

n (%) n (%) n (%)
All 61,086 11,111 49,975

PT
Age
18-<40 7,227 (11.8) 180 (1.6) 7,047 (14.1)
40-<50 8,391 (13.7) 829 (7.5) 7,562 (15.1)

RI
50-<60 15,186 (24.9) 2,526 (22.7) 12,660 (25.3)
60-<70 17,080 (28.0) 4,086 (36.8) 12,994 (26.0)
70-<80 10,419 (17.1) 2,855 (25.7) 7,564 (15.1)

SC
80 or more 2,783 (4.6) 635 (5.7) 2,148 (4.3)
Gender
Female 34,955 (57.2) 5,361 (48.2) 29,594 (59.2)
Male 26,131 (42.8) 5,750 (51.8) 20,381 (40.8)

U
Race
AN
White 42,203 (69.1) 8,141 (73.3) 34,062 (68.2)
Black 9,545 (15.6) 1,802 (16.2) 7,743 (15.5)
Hispanic 3,266 (5.3) 261 (2.3) 3,005 (6.0)
Asian 1,759 (2.9) 213 (1.9) 1,546 (3.1)
M

Other/Unknown 4,313 (7.1) 694 (6.2) 3,619 (7.2)

Year of Hospitalization
2004 4,537 (7.4) 739 (6.7) 3,798 (7.6)
D

2005 4,788 (7.8) 858 (7.7) 3,930 (7.9)

TE

2006 5,368 (8.8) 1,024 (9.2) 4,344 (8.7)

2007 5,771 (9.4) 1,032 (9.3) 4,739 (9.5)
2008 6,823 (11.2) 1,253 (11.3) 5,570 (11.1)
2009 7,541 (12.3) 1,397 (12.6) 6,144 (12.3)
EP

2010 8,309 (13.6) 1,566 (14.1) 6,743 (13.5)

2011 8,685 (14.2) 1,540 (13.9) 7,145 (14.3)
2012 9,264 (15.2) 1,702 (15.3) 7,562 (15.1)
C
AC

Page 9 of 19
 ACCEPTED MANUSCRIPT

Table 2: Incidence of Comorbidities, Infection, and Stay at Intensive Care Unit.

Category All Solid Tumors Lung Cancer Other Solid Tumors

n (%) n (%) n (%)
All 61,086 11,111 49,975

PT
Comorbidities
Diabetes Mellitus 9,050 (14.8) 1,930 (17.4) 7,120 (14.2)
Heart Disease 31,978 (52.3) 7,427 (66.8) 24,551 (49.1)

RI
Hepatic Disease 3,729 (6.1) 419 (3.8) 3,310 (6.6)
Lung Disease 17,758 (29.1) 5,784 (52.1) 11,974 (24.0)
Renal Disease 12,677 (20.8) 2,475 (22.3) 10,202 (20.4)

SC
Cerebrovascular Disease 1,363 (2.2) 354 (3.2) 1,009 (2.0)
Peripheral Vascular Disease 1,513 (2.5) 628 (5.7) 885 (1.8)
Venous Thromboembolism 5,017 (8.2) 921 (8.3) 4,096 (8.2)

U
Number of Comorbidities
0 17,858 (29.2) 1,674 (15.1) 16,184 (32.4)
AN
1 18,172 (29.7) 3,069 (27.6) 15,103 (30.2)
2 14,250 (23.3) 3,365 (30.3) 10,885 (21.8)
3 7,499 (12.3) 2,054 (18.5) 5,445 (10.9)
M

4 2,705 (4.4) 789 (7.1) 1,916 (3.8)

5+ 602 (1.0) 160 (1.4) 442 (0.9)
Infection
D

Infection (any) 32,951 (53.9) 6,628 (59.7) 26,323 (52.7)

Sepsis 8,217 (13.5) 1,744 (15.7) 6,473 (13.0)
TE

Bacterial Infection without Sepsis 6,561 (10.7) 916 (8.2) 5,645 (11.3)
Pneumonia 8,100 (13.3) 2,937 (26.4) 5,163 (10.3)
Urinary Tract Infection 7,130 (11.7) 1,068 (9.6) 6,062 (12.1)
EP

Fungal Infection 8,085 (13.2) 1,570 (14.1) 6,515 (13.0)

Stay at Intensive Care Unit 9,887 (16.2) 2,323 (20.9) 7,564 (15.1)
C
AC

Page 10 of 19
 ACCEPTED MANUSCRIPT

Table 3: Multivariable Logistic Regression Model for In-Hospital Mortality in Patients with Lung Cancer
(n=11,111, died n=1,244)

Odds Wald's 95% CI Robust 95% CI *

Ratio and p-value and p-value
Age ≥60 Years 1.23 1.05-1.44 0.0118 1.04-1.44 0.0126
Comorbidities

PT
No. of Comorbidities ≥2 1.62 1.31-2.00 <.0001 1.30-2.02 <.0001
Lung Disease 2.02 1.68-2.43 <.0001 1.68-2.43 <.0001
Hepatic Disease 1.47 1.11-1.96 0.0078 1.07-2.03 0.0188

RI
Pulmonary Embolism 1.40 1.04-1.88 0.0286 1.01-1.93 0.0441
Renal Disease 1.60 1.37-1.87 <.0001 1.35-1.88 <.0001
Cerebrovascular Disease 1.65 1.20-2.27 0.0019 1.19-2.29 0.0026

SC
Infection
Any Infection 1.28 1.03-1.58 0.0239 1.04-1.56 0.0181
Sepsis 4.41 3.74-5.19 <.0001 3.66-5.32 <.0001

U
Pneumonia 1.52 1.30-1.78 <.0001 1.31-1.76 <.0001
ICU Stay 3.13 2.70-3.63 <.0001 2.61-3.76 <.0001
AN
In-Hospital Chemotherapy 1.62 1.32-1.98 <.0001 1.32-1.99 <.0001
C-statistics 0.86
CI=confidence interval; ICU=intensive care unit
M

*Empirical standard errors accounting for clustering of the observations from the same hospital
Results in the table are adjusted for gender, ethnicity and year of discharge.
D
TE
C EP
AC

Page 11 of 19
 ACCEPTED MANUSCRIPT

12

10 11.2

PT
9.3 9.1
8 8.7
Died (%*)

7.8 7.8
7.4 7.2 7.1 7.1

RI
6 6.9
5.5 5.4
5.1
4
4.1

SC
3.4 3.2 3.2
2

U
AN
M

*Percent and 95% confidence limit presented

D

Figure 1: In-Hospital Mortality by Cancer Type.

TE
C EP
AC

Page 12 of 19
 ACCEPTED MANUSCRIPT

45
Lung Cancer
40
Other Solid Tumors
35
35.635.1
30
Died (%*)

PT
25
25.524.9
20

RI
19.5
15 17.4

10 12.2
11.2

SC
7.9
5 2.0
6.1 0.8 4.7
3.1
0
All 0 1 2 3 4 5+

U
Patients
Number of Comorbidities
*Percent and 95% confidence limit presented
AN
Figure 2: Mortality of the Two Populations by Increasing Numbers of Comorbidities.
M
D
TE
C EP
AC

Page 13 of 19
 ACCEPTED MANUSCRIPT

35
32.4
30.2 30.3 Lung Cancer
30
27.6
Other Solid Tumors

PT
25
Patients (%)

21.8
20 18.5

RI
15.1
15
10.9

SC
10
7.1

5 3.8
1.4 0.9

U
0
0 1 2 3 4 5+
AN
Number of Comorbidities
M

Figure 3: Number of Comorbidities in Lung Cancer Patients vs. Other Cancer Types
D
TE
C EP
AC

Page 14 of 19
 ACCEPTED MANUSCRIPT

70 66.8

60 Lung Cancer
52.1
Patients (%*)

49.1 Other Solid Tumors

50

PT
40

30
24.0 22.3

RI
20.4
20 17.4
14.2

10 6.6 8.3 8.2

3.8 5.7

SC
3.2 2.0 1.8
0

U
AN
*Percent and 95% confidence limit presented
M

Figure 4: Frequencies of Specific Comorbidity Types.

D
TE
C EP
AC

Page 15 of 19
 ACCEPTED MANUSCRIPT

Clinical Events

70
Lung Cancer
60 Other Solid Tumors
59.7

PT
50
Patient (%*)

52.7
40

RI
30

20 26.4

SC
20.9
10 15.7
13.0 15.1
10.3
0

U
Infection Sepsis Pneumonia Stay at ICU
*Percent and 95% confidence limit presented
AN
Figure 5a: Frequencies of Selected Risk Factors.
M
D

In-Hospital Mortality
TE

50
Lung Cancer
Other Solid Tumors
40
39.6
EP
Died (%*)

30 33.1
28.8
20 23.9
21.2
C

16.2 17.8
10
AC

9.6

0
Infection Sepsis Pneumonia Stay at ICU
*Percent and 95% confidence limit presented

Figure 5b: Mortality Rates of Selected Risk Factors.

Page 16 of 19
 ACCEPTED MANUSCRIPT

Frequency of Pneumonia

40
Lung Cancer
Pneumonia (%*)

36.0
Other Solid Tumors
30

26.4

PT
20 22.8

16.1
10
10.3

RI
6.4
0
All Patients Patients Patients

SC
with without
*Percent and 95% confidence limit presented Lung Disease Lung Disease

Figure 6a: Frequency of Pneumonia by Tumor Type and Presence of Comorbid Lung Disease.

U
AN
Mortality in Patients with Pneumonia

30 Lung Cancer
M

Other Solid Tumors 28.9

26.1
Died (%*)

20 21.2
D

17.8

10
TE

9.2
5.5
0
All Patients Patients Patients
EP

with with without

Pneumonia Lung Disease Lung Disease
*Percent and 95% confidence limit presented

Figure 6b: Mortality of Patients with Pneumonia by Tumor Type and Presence of Comorbid Lung
C

Disease.
AC

Page 17 of 19
 ACCEPTED MANUSCRIPT

References

1. Lyman GH, Michels SL, Reynolds MW, Barron R, Tomic KS, Yu J. Risk of mortality in patients with
cancer who experience febrile neutropenia. Cancer. 2010;116(23):5555-5563.
2. Smith TJ, Khatcheressian J, Lyman GH, et al. 2006 update of recommendations for the use of
white blood cell growth factors: an evidence-based clinical practice guideline. Journal of clinical
oncology : official journal of the American Society of Clinical Oncology. 2006;24(19):3187-3205.

PT
3. Caggiano V, Weiss RV, Rickert TS, Linde-Zwirble WT. Incidence, cost, and mortality of
neutropenia hospitalization associated with chemotherapy. Cancer. 2005;103(9):1916-1924.
4. Chindaprasirt J, Wanitpongpun C, Limpawattana P, et al. Mortality, length of stay, and cost
associated with hospitalized adult cancer patients with febrile neutropenia. Asian Pacific journal

RI
of cancer prevention : APJCP. 2013;14(2):1115-1119.
5. Kuderer NM, Dale DC, Crawford J, Cosler LE, Lyman GH. Mortality, morbidity, and cost
associated with febrile neutropenia in adult cancer patients. Cancer. 2006;106(10):2258-2266.

SC
6. Freifeld AG, Bow EJ, Sepkowitz KA, et al. Clinical practice guideline for the use of antimicrobial
agents in neutropenic patients with cancer: 2010 update by the infectious diseases society of
america. Clinical infectious diseases : an official publication of the Infectious Diseases Society of
America. 2011;52(4):e56-93.

U
7. Crawford J, Dale DC, Lyman GH. Chemotherapy-induced neutropenia: risks, consequences, and
new directions for its management. Cancer. 2004;100(2):228-237.
AN
8. Wallington M, Saxon EB, Bomb M, et al. 30-day mortality after systemic anticancer treatment
for breast and lung cancer in England: a population-based, observational study. The Lancet
Oncology. 2016;17(9):1203-1216.
M

9. Lyman GH, Kuderer NM. Epidemiology of febrile neutropenia. Supportive cancer therapy.
2003;1(1):23-35.
10. Klastersky J, Paesmans M. The Multinational Association for Supportive Care in Cancer (MASCC)
risk index score: 10 years of use for identifying low-risk febrile neutropenic cancer patients.
D

Support Care Cancer. 2013;21(5):1487-1495.

11. Vizient. About UHC. 2016; https://www.uhc.edu/about. Accessed 3/23/2016, 2016.
TE

12. Weycker D, Barron R, Kartashov A, Legg J, Lyman GH. Incidence, treatment, and consequences
of chemotherapy-induced febrile neutropenia in the inpatient and outpatient settings. Journal
of oncology pharmacy practice : official publication of the International Society of Oncology
Pharmacy Practitioners. 2014;20(3):190-198.
EP

13. Dulisse B, Li X, Gayle JA, et al. A retrospective study of the clinical and economic burden during
hospitalizations among cancer patients with febrile neutropenia. Journal of medical economics.
2013;16(6):720-735.
C

14. Culakova E, Poniewierski M, Crawford J, Dale D, Lyman G. Impact of Comorbidities on Length of

Stay and Mortality in Hospitalized Patients with Cancer and Febrile Neutropenia. ASH
AC

Conference Abstracts. 2014.

15. Carmona-Bayonas A, Jimenez-Fonseca P, Virizuela Echaburu J, et al. Prediction of serious
complications in patients with seemingly stable febrile neutropenia: validation of the Clinical
Index of Stable Febrile Neutropenia in a prospective cohort of patients from the FINITE study.
Journal of clinical oncology : official journal of the American Society of Clinical Oncology.
2015;33(5):465-471.
16. Lynn JJ, Chen KF, Weng YM, Chiu TF. Risk factors associated with complications in patients with
chemotherapy-induced febrile neutropenia in emergency department. Hematological oncology.
2013;31(4):189-196.

Page 18 of 19
 ACCEPTED MANUSCRIPT

17. Baden LR, Bensinger W, Angarone M, et al. Prevention and treatment of cancer-related
infections. J Natl Compr Canc Netw. 2012;10(11):1412-1445.

PT
RI
U SC
AN
M
D
TE
C EP
AC

Page 19 of 19