CLINICAL RESEARCH STUDY

Admission Hypoglycemia and Increased Mortality in

Patients Hospitalized with Pneumonia
John-Michael Gamble, BScPharm, MSc,
a
* Dean T. Eurich, PhD,
a
* Thomas J. Marrie, MD,
b
Sumit R. Majumdar, MD, MPH
a,b
a
Department of Public Health Sciences, School of Public Health, University of Alberta, Edmonton, Alberta, Canada;
b
Department of
Medicine, Faculty of Medicine and Dentistry, University of Alberta Hospital, Edmonton, Alberta, Canada.
ABSTRACT
BACKGROUND: The relationship between spontaneous admission hypoglycemia and mortality in patients
hospitalized with community-acquired pneumonia is unclear.
METHODS: From 2000 to 2002, clinical data were prospectively collected on all patients with community-
acquired pneumonia who were admitted to all 6 hospitals in Edmonton, Alberta, Canada. Patients with
admission glucose greater than 6.1 mmol/L (n 1996) were excluded; the remaining patients were
categorized as having admission hypoglycemia (4.0 mmol/L [n 54]) or normoglycemia (4.0 to 6.1
mmol/L [n 902]). Multivariable Cox proportional hazards models were used to examine the relationship
between hypoglycemia and all-cause mortality in-hospital, at 30 days, and at 1 year.
RESULTS: The mean age was 65 (standard deviation20) years, 48% were female, 8% had diabetes, and 56%
had severe pneumonia. Overall, admission hypoglycemia was present in 2% (54/2990) of the entire cohort and
6% of those with glucose of 6.1 mmol/L or less. Total deaths were 89 (9%) in-hospital, 96 (10%) at 30 days,
and 247 (26%) at 1 year. In-hospital mortality was higher among patients with admission hypoglycemia (11
[20%] deaths) compared with those with normoglycemia (78 [9%]; adjusted hazards ratio [aHR] 2.96; 95%
condence interval [CI], 1.39-6.31; P.005). An increased risk of mortality was observed at 30 days (11 [20%]
vs 85 [10%]; aHR 2.89; 95% CI, 1.32-6.29) and remained elevated at 1 year (19 [35%] vs 228 [25%]; aHR1.80;
95% CI, 1.02-3.17). These results were not inuenced by treatment for diabetes (P.4 for interaction).
CONCLUSION: In a population-based sample of patients with community-acquired pneumonia, spontaneous
admission hypoglycemia was independently associated with increased mortality during hospitalization that
persisted to 1 year. Patients with hypoglycemia are an easily identied group that may warrant more
intensive inpatient and postdischarge follow-up.
2010 Elsevier Inc. All rights reserved. The American Journal of Medicine (2010) 123,
556.e11-556.e16
KEYWORDS: Cohort study; Community-acquired pneumonia; Mortality hypoglycemia
Community-acquired pneumonia is a major cause of mor-
bidity and mortality.
1
Despite advances in treatment, com-
munity-acquired pneumonia-related mortality ranges from
5% to 10% in the short term, and 20% to 65% of patients
hospitalized with pneumonia die within 5 years.
2
Several
independent prognostic factors have been associated with
mortality in patients with community-acquired pneumonia
(eg, age, comorbidities, functional status, laboratory nd-
Funding: An establishment grant fromAlberta Heritage Foundation for Med-
ical Research; grants-in-aid from Capital Health; and unrestricted grants from
Abbott Canada, Pzer Canada, and Janssen-Ortho Canada (all to TJM); an oper-
ating grant from the Canadian Institutes of Health Research (200809MOP-
191604). DTE and SRM receive salary support awards from the Alberta Heritage
Foundation for Medical Research, and DTE also receives salary support from
Canadian Institutes of Health Research. JMGholds a Canadian Institutes of Health
Research doctoral award and a full-time health research studentship through the
Alberta Heritage Foundation for Medical Research.
Conict of Interest: None of the authors have any conicts of interest
associated with the work presented in this manuscript.
Authorship: All authors had access to the data and played a role in
writing this manuscript.
*JMG and DTE contributed equally to this work.
Reprint requests should be addressed to Dean T. Eurich, PhD, 2-040
Health Research Innovation Facility, University of Alberta, Edmonton,
Alberta, Canada, T6G2E1.
E-mail address: deurich@ualberta.ca
0002-9343/$ -see front matter 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.amjmed.2009.11.021
ings),
3,4
including, more recently, blood glucose levels at
time of hospital admission.
5,6
Admission hyperglycemia in
patients hospitalized for community-acquired pneumonia is
associated with an increased risk of short-term severe ad-
verse events, including death;
5,6
however, the association
between low blood glucose levels
(hypoglycemia) and adverse events
remains unclear.
Spontaneous hypoglycemia has
numerous causes, including but
not limited to severe systemic
illness (eg, sepsis, liver failure, ad-
renal insufciency, shock); ad-
vanced malnutrition; various med-
ications (eg, oral hypoglycemic
agents and insulin); malignancy;
and frailty.
7
Although the inci-
dence of hypoglycemia is rela-
tively low at 2% to 6%,
8,9
it is
associated with increased mortal-
ity in a variety of inpatient
8-14
and
outpatient populations.
15
Patients with severe pneumonia
are often elderly and may be par-
ticularly susceptible to adverse outcomes from spontaneous
hypoglycemia as a result of depleted metabolic reserves
because of advanced age, multiple comorbidities, poor ox-
ygenation, release of inammatory mediators,
16
and polyp-
harmacy. In the only study conducted to date in patients
with community-acquired pneumonia, hypoglycemia on
hospital admission was independently associated with in-
creased 30-day mortality and acute complications.
14
Nota-
bly, however, the independent effect of admission glucose
levels on longer-term outcomes in patients with community-
acquired pneumonia is less certain. We hypothesized that
routinely measured and easy to capture hypoglycemia in
patients with community-acquired pneumonia could serve
as a marker for poor outcomes in both the acute and longer
term. Therefore, we estimated the rates and potential asso-
ciations between hypoglycemia at the time of admission in
patients hospitalized with community-acquired pneumonia
and all-cause mortality in-hospital, at 30 days, and at 1 year
after admission.
MATERIALS AND METHODS
Population and Setting
Data were collected on a prospective cohort of adults
(17 years) with community-acquired pneumonia who
were admitted to 6 hospitals in Edmonton (population 1
million), Alberta, Canada, between 2000 and 2002, and
managed according to a standardized clinical pathway. A
detailed description of the cohort and data-collection
methods has been reported.
17,18
Only patients with tuber-
culosis or cystic brosis, and those who were immuno-
compromised or pregnant were excluded. This study was
approved by the Health Research Ethics Board at the
University of Alberta.
Data collected included: patient demographics (ie, age,
gender, comorbidities); clinical characteristics (ie, smoking
status, vital statistics, laboratory values, presence of ad-
vanced directives); prescription
and nonprescription drug use
within the week before admission;
functional status; and pneumonia
severity. Functional status was
based on patient or proxy inter-
view and classied into 3 mutu-
ally exclusive groups on the basis
of the week before admission: in-
dependent in ambulation, consis-
tently requiring any type of walk-
ing aid or a wheelchair, and
bedridden. Pneumonia severity at
the time of presentation was deter-
mined using the well-validated
measure of illness severity, the
Pneumonia Severity Index (based
on 3 demographic characteristics,
5 concurrent illnesses, 5 physical
ndings, and 7 laboratory tests).
19
On discharge, subjects were followed up to 5 years
through linkage to the provincial administrative data-
sets.
2
The Alberta Provincial Ministry of Health main-
tains and updates demographic information, vital statis-
tics, and health services data on all registered subjects.
Trained health record administrators are responsible for
diagnostic coding, and coding accuracy is routinely val-
idated both provincially and through a central Canadian
agency.
2
Study Sample
The source population consisted of all subjects with a
random venous blood glucose measurement at hospital
admission that was conducted as part of the clinical
pathway (n 2990). Because previous studies have sug-
gested that casual blood glucose levels greater than 6.1
mmol/L are associated with adverse outcomes in patients
with community-acquired pneumonia,
5,6
subjects pre-
senting with casual blood glucose levels greater than 6.1
mmol/L were excluded (n 1996). In addition, of the
remaining 994 subjects, we excluded another 38 patients
(4%) who could not be successfully linked to our admin-
istrative databases. We then proceeded to classify the
remaining 956 subjects, as others have,
5,20
according to
their admission glucose levels: those presenting with
hypoglycemia dened as a random blood glucose level
less than 4.0 mmol/L
21,22
and those with random blood
glucose levels within the normal range of 4.0 to 6.1
mmol/L. Those with normoglycemia served as the refer-
ence group for all analyses.
CLINICAL SIGNIFICANCE
In a prospective cohort of patients
hospitalized with pneumonia, sponta-
neous admission hypoglycemia was
uncommon (2%).
However, patients with glucose less
than 4 mmol/L at admission have a sub-
stantially increased risk of in-hospital,
30-day, and 1-year mortality.
More intensive follow-up might be war-
ranted in this population.
556.e12 The American Journal of Medicine, Vol 123, No 6, June 2010
Outcomes
Our primary outcome, dened a priori, was in-hospital mor-
tality. We secondarily assessed 30-day and 1-year mortality.
Statistical Analysis
First, we explored the relationship between admission blood
glucose and mortality within all subjects with a random
blood glucose measurement at hospital admission (n2990)
using locally weighted scatter-plot smoothing (LOWESS)
curves. Then, within our cohort of patients with admission
blood glucose of 6.1 mmol/L or less, we used Cox proportional
hazards models to assess the relationship between hypoglyce-
mia at admission and time to in-hospital, 30-day, and 1-year
mortality. All patients were followed until time of death, ter-
mination of health insurance coverage (eg, departure from
province), or March 31, 2006, whichever occurred earliest.
Crude and adjusted hazard ratios and 95% condence intervals
were reported for each outcome. Multivariable regression
models were dened a priori and included numerous potential
confounding variables, such as age, gender, history of comor-
bidities (eg, diabetes, neoplasm, cardiovascular disease,
chronic obstructive pulmonary disease, renal disease), total
number of prescription medications, and clinical character-
istics (Pneumonia Severity Index, premorbid functional sta-
tus, smoking status, nursing home residence, presence of
advanced directives, direct admission to the intensive care
unit, and immunization history). In addition, because pa-
tients with diabetes may be at increased risk of hypoglyce-
mia at the time of illness and because hypoglycemia may
affect them differently, rst we adjusted for diabetes in our
base models and then we tested for the presence of an
interaction term (hypoglycemia diabetes). Diabetes was
dened as a documented history requiring treatment with oral
hypoglycemic agents or insulin; we did not routinely collect
A1c. All analyses were conducted using Stata/IC 10.1
(StataCorp LP, College Station, Tex).
Sensitivity Analysis
To further evaluate the independent effects of hypoglycemia
on adverse outcomes in patients with community-acquired
pneumonia, we undertook a series of additional analyses.
First, we adjusted for the presence or absence of systemic
inammatory response syndrome using established criteria:
2 or more instances of a temperature more than 38C or less
than 36C, heart rate more than 90 beats/min, respiratory
rate more than 20 breaths/min or PaCO
2
less than 32 mm Hg,
or white blood cell count more than 12 10
9
/L, less than
410
9
/L, or more than 10% immature (band) forms.
23
Second, because patients who are directly admitted to the
intensive care unit may confound the relationship between
hypoglycemia and mortality, we conducted an analysis
whereby these patients were excluded. Finally, because sub-
jects with diabetes are at a higher risk of hypoglycemia and
diabetes itself is associated with poorer outcomes in patients
with community-acquired pneumonia,
5
we excluded pa-
tients with diabetes before admission in our analysis.
RESULTS
Of the 956 subjects who formed the study cohort, the mean
age was 65 years (standard deviation20), 48% were fe-
male, 8% had diabetes on admission, 56% had severe
(Pneumonia Severity Index class IV/V) pneumonia, and the
mean follow-up was 2.78 years (standard deviation 1.83)
(Table 1).
Overall, spontaneous admission hypoglycemia was un-
common, occurring in 2% (54/2990) of the entire popula-
tion with pneumonia. In our study sample (ie, those with
admission glucose 6.1 mmol/L; n 956), there were 54
subjects (6%) with hypoglycemia on admission. Patient
characteristics were similar between groups except that pa-
tients with hypoglycemia had more severe pneumonia
(Pneumonia Severity Index class IV/V) (72% vs 55%), were
more likely to be directly admitted to the intensive care unit
(30% vs 8%), and took more medications (28% vs 14%)
than patients with normoglycemia (Table 1). Of note, pa-
tients with diabetes were far more likely to present with
Table 1 Baseline Cohort Characteristics
Characteristics
Admission Glucose (mmol/L)
4.0 4.0 to 6.1 P Value
n (%) 54 6% 902 94%
Age, y, mean (SD) 62.9 19.1 64.9 20.0 .47
Gender, female, n (%) 24 44% 438 49% .56
Comorbidities
Cardiovascular disease 25 46% 320 35% .11
Neoplasm 5 9% 142 16% .20
COPD 18 33% 273 30% .63
Renal disease 10 19% 135 15% .48
Prior diabetes mellitus 25 46% 47 5% .001
Pneumonia Severity Index,
n (%)
.04
Class I or II 7 13% 226 25%
Class III 8 15% 182 20%
Class IV 22 41% 322 36%
Class V 17 31% 172 19%
Premorbid functional status,
n (%)
.99
Walking with/without
assistance
49 91% 821 91%
Wheelchair/prosthesis 4 7% 63 7%
Bedridden 1 2% 18 2%
Smoking status, n (%) .14
Nonsmoker 29 54% 361 40%
Ex-smoker 12 22% 259 29%
Current smoker 13 24% 282 31%
Directly admitted to ICU 16 30% 72 8% .001
Previous pneumococcal
vaccination, n (%)
11 20% 191 21% .89
Presence of living will or
advanced directive
6 11% 72 8% .42
Nursing home resident 8 15% 139 15% .91
5 medications 15 28% 128 14% .01
SD standard deviation; COPD chronic obstructive pulmonary dis-
ease; ICU intensive care unit.
556.e13 Gamble et al Hypoglycemia and Mortality
admission hypoglycemia than patients without diabetes
(46% vs 5%, P.001).
Mortality
A J-shaped relationship was observed between admission
blood glucose and mortality within our entire population of
patients with pneumonia (n2990) (Figure 1). In our sam-
ple of patients without hyperglycemia (n956), 89 subjects
(9%) with blood glucose levels of 6.1 mmol/L or less died
in-hospital. Figure 2 displays our unadjusted analyses. Hy-
poglycemia was associated with a signicantly higher risk
of in-hospital mortality (11 [20%] vs 78 [9%] of those with
normoglycemia; hazard ratio 2.07; 95% condence interval,
1.10-3.89; P.03). The unadjusted increased risk of mor-
tality in the hypoglycemia group persisted up to 30 days (11
[20%] deaths vs 85 [10%] deaths; hazard ratio [HR] 2.36;
95% condence interval [CI], 1.26-4.42; P.01) and at 1
year after admission (19 [35%] deaths vs 228 [25%] deaths;
HR 1.59; 95% CI, 1.00-2.54; P.05), although the risk
seemed to diminish over time. The most common causes of
death within 1 year of hospital admission for community-
acquired pneumonia were related to the initial pneumonia
itself (41%) or attributed to cardiopulmonary events (26%;
one half cardiac, one half pulmonary), cancer (18%), or
other causes (15%).
After adjustment for age, gender, comorbidities, total
number of medications, clinical characteristics, pneumonia
severity, and intensive care unit admission, we continued to
observe an increased risk of mortality associated with hy-
poglycemia compared with normoglycemia on admission
for in-hospital mortality (adjusted HR 2.96; 95% CI, 1.39-
6.31; P.005), 30-day mortality (adjusted HR 2.89; 95%
CI, 1.32-6.29; P.008), and at 1 year (adjusted HR 1.80;
95% CI, 1.02-3.17; P.042). Furthermore, no statistically
signicant interaction between diabetes status and hypogly-
cemia (P for interaction .4) existed for any of our 3
outcomes, suggesting our results are consistent for patients
with or without diabetes and reect an association with
spontaneous hypoglycemia.
Sensitivity Analyses
The risk of death for all outcomes was nearly identical to
our primary analysis after further adjusting for the presence
of systemic inammatory response syndrome: in-hospital
mortality (adjusted HR 3.13; 95% CI, 1.46-6.71; P.003);
Figure 1 J-shaped curve of casual admission glucose and
mortality.
Figure 2 Mortality rates for patients with spontaneous hypoglycemia and normoglycemia at
admission.
556.e14 The American Journal of Medicine, Vol 123, No 6, June 2010
30-day mortality (adjusted HR 3.01; 95% CI, 1.37-6.60;
P.006); and 1-year mortality (adjusted HR 1.82, 95% CI,
1.03-3.21, P.04). Analyses that excluded patients who
were directly admitted to the intensive care unit (n88)
increased the risk associated with hypoglycemia compared
with normoglycemia for in-hospital mortality (adjusted HR
4.61; 95% CI, 1.88-11.29; P.001), 30-day mortality (ad-
justed HR 4.73, 95% CI, 1.94-11.52, P.001), and 1-year
mortality (adjusted HR 2.41; 95% CI, 1.22-4.78; P.012).
Similarly, consistent results also were found when we ex-
cluded all subjects with diabetes (n 72) from analyses:
in-hospital mortality (adjusted HR 2.96; 95% CI, 1.25-
7.01), 30-day mortality (adjusted HR 2.99; 95% CI, 1.19-
7.49), and 1-year mortality (adjusted HR 2.27; 95% CI,
1.13-4.54).
DISCUSSION
In our population-based cohort of adults who were admitted
to the hospital for community-acquired pneumonia, we
found an increased risk of both acute and long-term mor-
tality associated with spontaneous hypoglycemia at presen-
tation. Although hypoglycemia was relatively uncommon
(2% of all patients with pneumonia, 6% of those without
overt hyperglycemia), given the magnitude of the increased
risk in mortality that we observed, it is an important and
previously not well-recognized prognostic factor that merits
further consideration.
Despite a 10-fold greater risk of hypoglycemia in those
with diabetes compared with the nondiabetic population, in
our study diabetes did not affect the relationship between
hypoglycemia and mortality and was not associated with
any effect modication. Previous studies also have reported
that hypoglycemia adversely affects outcomes irrespective
of diabetes.
8,10,14
The fact that our ndings were unaltered
when we excluded diabetic patients further supports our
assertion that most hypoglycemic episodes were spontane-
ous and not the result of iatrogenesis.
Our results are consistent with the only previous study in
patients with community-acquired pneumonia, whereby ad-
mission hypoglycemia was associated with a 2-fold in-
creased risk of death at 30 days, irrespective of diabetes
status, that is nearly identical to our ndings.
14
Notably,
however, our data suggest mortality remains elevated for up
to 1 year. Collectively, the magnitude of risk (2-fold)
seems to be similar among studies and across several dif-
ferent patient populations.
8-10,12-15,24
Our ndings also have implications for previous studies
evaluating the effects of elevated blood glucose and mor-
tality. As others have noted,
15,20
we also observed a well-
dened J-shaped relationship between casual blood glucose
at admission and mortality. If this J-shaped relationship is
usually present, it means the adverse effects of hyperglyce-
mia have been systematically underestimated because most
studies have grouped all patients with a blood glucose level
less than 6.1 mmol/L into 1 reference category for purposes
of comparison.
5,25
In fact, the Pneumonia Severity Index
itself accords 10 points to severe hyperglycemia (14
mmol/L); if the reference group were 4 to 6.1 mmol/L, it
is likely that more than 10 points would have accrued to
severe hyperglycemia, and perhaps even lesser degrees of
hyperglycemia (8 to 10 mmol/L or 10 to 14 mmol/L)
would have been accorded points.
6
Alternately, as a rough
calculation using the methods of the Pneumonia Severity
Index developers,
26
admission hypoglycemia would receive
a score of approximately 20 points in our study.
STUDY LIMITATIONS
First, we may have failed to balance groups with respect to
all important prognostic factors. Nevertheless, we adjusted
for many clinical characteristics, such as functional status
and disease severity, that have not been available to oth-
ers.
12,13
Second, we limited our analyses to the rst labo-
ratory glucose measurement on admission, and we do not
know whether these values were associated with symptoms
or treated. Third, the cut-points used in our analyses to
categorize subjects as normoglycemic or hypoglycemic may
have affected our results, although we would note these are
the same cut-points for hypoglycemia in North American
guidelines.
21,22
Fourth, our study did not focus on the po-
tential mechanisms leading to increased mortality (ie, neu-
roglycopenia) or potential causes for hypoglycemia (ie, in-
terleukin-mediated hypoglycemia).
27
Finally, clinical data
were only available at the time of presentation to the hos-
pital, and changes in therapies and comorbidities during the
rst year of follow-up were not captured.
CONCLUSIONS
Although hypoglycemia in patients with community-ac-
quired pneumonia is relatively uncommon, it is associated
with a heretofore underappreciated and substantial independent
increased risk of in-hospital, 30-day, and 1-year mortality. We
would not necessarily advocate modifying existing pneu-
monia severity scores because admission hypoglycemia is
not often present. Instead, we believe it is better considered
an easy-to-measure red ag to alert clinicians to the fact
that all else equal, the patient with hypoglycemia is at high
risk of adverse events and deserving of more intensive
follow-up and surveillance during hospitalization and after
discharge.
ACKNOWLEDGMENTS
The authors thank all of the community-acquired pneumo-
nia pathway research nurses for their dedication and hard
work.
References
1. Statistics Canada. Ten leading causes of deaths in Canada, 2004 and
2005. Available at: www.statcan.gc.ca. Accessed June 30, 2009.
2. Johnstone J, Eurich DT, Majumdar SR, et al. Long-term morbidity and
mortality after hospitalization with community-acquired pneumonia: a
556.e15 Gamble et al Hypoglycemia and Mortality
population-based cohort study. Medicine (Baltimore). 2008;87:329-
334.
3. Mortensen EM, Kapoor WN, Chang CC, Fine MJ. Assessment of
mortality after long-term follow-up of patients with community-ac-
quired pneumonia. Clin Infect Dis. 2003;37:1617-1624.
4. Fine MJ, Smith MA, Carson CA, et al. Prognosis and outcomes of patients
with community-acquired pneumonia. A meta-analysis. JAMA. 1996;275:
134-141.
5. Kornum JB, Thomsen RW, Riis A, et al. Type 2 diabetes and pneu-
monia outcomes: a population-based cohort study. Diabetes Care.
2007;30:2251-2257.
6. McAlister FA, Majumdar SR, Blitz S, et al. The relation between hyper-
glycemia and outcomes in 2,471 patients admitted to the hospital with
community-acquired pneumonia. Diabetes Care. 2005;28:810-815.
7. Service FJ. Hypoglycemic disorders. N Engl J Med. 1995;332:1144-
1152.
8. Krinsley JS, Grover A. Severe hypoglycemia in critically ill patients:
risk factors and outcomes. Crit Care Med. 2007;35:2262-2267.
9. Kagansky N, Levy S, Rimon E, et al. Hypoglycemia as a predictor of
mortality in hospitalized elderly patients. Arch Intern Med. 2003;163:
1825-1829.
10. Kosiborod M, Inzucchi SE, Goyal A, et al. Relationship between
spontaneous and iatrogenic hypoglycemia and mortality in patients
hospitalized with acute myocardial infarction. JAMA. 2009;301:1556-
1564.
11. Svensson AM, McGuire DK, Abrahamsson P, Dellborg M. Associa-
tion between hyper- and hypoglycaemia and 2 year all-cause mortality
risk in diabetic patients with acute coronary events. Eur Heart J.
2005;26:1255-1261.
12. Mannucci E, Monami M, Mannucci M, et al. Incidence and prognostic
signicance of hypoglycemia in hospitalized non-diabetic elderly pa-
tients. Aging Clin Exp Res. 2006;18:446-451.
13. Shilo S, Berezovsky S, Friedlander Y, Sonnenblick M. Hypoglycemia
in hospitalized nondiabetic older patients. J Am Geriatr Soc. 1998;46:
978-982.
14. Singanayagam A, Chalmers JD, Hill AT. Admission hypoglycaemia is
associated with adverse outcome in community-acquired pneumonia.
Eur Respir J. 2009;34:932-939.
15. Wei M, Gibbons LW, Mitchell TL, et al. Low fasting plasma glucose
level as a predictor of cardiovascular disease and all-cause mortality.
Circulation. 2000;101:2047-2052.
16. Mizock BA. Alterations in carbohydrate metabolism during stress: a
review of the literature. Am J Med. 1995;98:75-84.
17. Marrie TJ, Wu L. Factors inuencing in-hospital mortality in commu-
nity-acquired pneumonia: a prospective study of patients not initially
admitted to the ICU. Chest. 2005;127:1260-1270.
18. Basi SK, Marrie TJ, Huang JQ, Majumdar SR. Patients admitted to
hospital with suspected pneumonia and normal chest radiographs:
epidemiology, microbiology, and outcomes. Am J Med. 2004;117:305-
311.
19. Aujesky D, Fine MJ. The pneumonia severity index: a decade after
the initial derivation and validation. Clin Infect Dis. 2008;47:S133-
S139.
20. Kosiborod M, Inzucchi SE, Krumholz HM, et al. Glucometrics in
patients hospitalized with acute myocardial infarction: dening the
optimal outcomes-based measure of risk. Circulation. 2008;117:1018-
1127.
21. Canadian Diabetes Association Clinical Practice Guidelines Expert
Committee. Canadian diabetes association 2008 clinical practice
guidelines for the prevention and management of diabetes in Canada.
Can J Diabetes. 2008;32:S1-S201.
22. American Diabetes Association. Standards of medical care in diabe-
tes2009. Diabetes Care. 2009;32:S13-S61.
23. Members of the ACCP/SCCM Consensus Conference Committee.
American College of Chest Physicians/Society of Critical Care Med-
icine Consensus Conference: denitions for sepsis and organ failure
and guidelines for the use of innovative therapies in sepsis. Crit Care
Med. 1992;20:864-874.
24. Vriesendorp TM, DeVries JH, Van Santen S, et al. Evaluation of
short-term consequences of hypoglycemia in an intensive care unit.
Crit Care Med. 2006;34:2714-2718.
25. Baker EH, Janaway CH, Philips BJ, et al. Hyperglycaemia is associ-
ated with poor outcomes in patients admitted to hospital with acute
exacerbations of chronic obstructive pulmonary disease. Thorax. 2006;
61:284-289.
26. Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify
low-risk patients with community-acquired pneumonia. N Engl J Med.
1997;336:243-250.
27. del Rey A, Monge-Arditi G, Besedovsky HO. Central and peripheral
mechanisms contribute to the hypoglycemia induced by interleukin-1.
Ann N Y Acad Sci. 1998;840:153-161.
556.e16 The American Journal of Medicine, Vol 123, No 6, June 2010