© 2010 by the American College of Gastroenterology The American Journal of GASTROENTEROLOGY
nature publishing group ORIGINAL CONTRIBUTIONS 117
LIVER AND BILIARY TRACT
INTRODUCTION
Pyogenic liver abscess (PLA) is associated with signiicant mor-
bidity, mortality, and health-care costs (1,2). Population-based
studies in Canada and Denmark have reported incidence rates
of PLA ranging from 1.1 to 2.3 per 100,000 population (1,3,4).
A recent population-based study from Taiwan quoted a higher
incidence rate of 17.6 per 100,000 population (5). However,
population-based studies evaluating the incidence of a PLA in
the United States are lacking.
Reported case-fatality rates from PLA are substantial, rang-
ing from 3 to 30% (6–11). Only three population-based studies,
from Canada and Denmark, have published case-fatality out-
comes: 6 (4), 10 (1), and 19% (3). In addition, the risk factors
associated with mortality from PLA have included age, bac-
teremia, and comorbidities such as cirrhosis, renal failure, and
malignancy (5,6,9,12). However, the low incidence of a PLA has
limited the capacity to identify independent predictors of mor-
tality from single-center studies.
A Population-Based Study of Pyogenic Liver Abscesses
in the United States: Incidence, Mortality, and
Temporal Trends
Liisa Meddings, MD
1
, Robert P. Myers, MD, MSc
1
, James Hubbard, MSc
1
, Abdel Aziz Shaheen, MD , MPH
1
, Kevin B. Laupland, MD, MSc
1,2
,
Elijah Dixon, MD, MPH
2,3
, Carla Coin, MD, MSc
1
and Gilaad G. Kaplan, MD , MPH, FRCPC
1,2
OBJECTIVES: Few population-based studies have evaluated pyogenic liver abscess (PLA) in North America. We
assessed the incidence of PLA and evaluated predictors of mortality.
METHODS: We used the Nationwide Inpatient Sample to identify all patients with discharges for PLA (ICD-9
572.0) between 1994 and 2005. Multivariable logistic regression analysis was performed to
determine whether mortality was associated with patient and hospital characteristics including
comorbidities, interventions, and bacterial cultures. We determined the annual incidence for
PLA in the US population and assessed for temporal changes using generalized linear regression
models.
RESULTS: We identified 17,787 PLA discharges for an overall incidence of PLA of 3.6 (95% confidence
interval (CI): 3.5–3.7) per 100,000 population. From 1994 to 2005, the annual average percent
increase in incidence was 4.1% (95% CI: 3.4 – 4.8; P < 0.0001). In-hospital mortality was
5.6% (95% CI: 5.3 – 6.0). Mortality was associated with older age (65 – 84 vs. 18 – 34: odds
ratio (OR) = 2.28 (1.48 – 3.51)); Medicaid (OR = 1.74 (1.36–2.23)) and Medicare (OR = 1.48
(1.18 –1.85) vs. private insurance; and comorbidities such as cirrhosis (OR = 2.48 (1.85–3.31)),
chronic renal failure (OR = 1.99 (1.28–3.09)), and cancer (OR = 2.32 (1.97–2.73)). Patients who
underwent percutaneous liver aspiration (OR = 0.45 (0.39–0.52)) had lower mortality, whereas
surgical drainage (OR = 0.87 (0.68–1.10)) and endoscopic retrograde cholangiopancreatography
(OR = 0.73 (0.52–1.03)) were not associated with mortality. The most commonly recorded
bacterial infections were Streptococcus species (29.5%) and Escherichia coli (18.1%). Patients
with bacteremia or septicemia (OR = 3.88 (3.36–4.48)) had an increased risk of death.
CONCLUSIONS: The incidence of PLA is increasing and is associated with significant mortality that is attributable
to several modifiable risk factors.
Am J Gastroenterol 2010; 105:117–124; doi:10.1038/ajg.2009.614; published online 3 November 2009
1
Department of Medicine, University of Calgary , Calgary , Alberta, Canada;
2
Department of Community Health Sciences, University of Calgary , Calgary , Alberta,
Canada;
3
Department of Surgery, University of Calgary , Calgary , Alberta, Canada. Correspondence: Gilaad G. Kaplan, MD, MPH, FRCPC, Department of Medicine,
Teaching Research and Wellness Center, University of Calgary , 3280 Hospital Drive NW , 6D17, Calgary, Alberta, Canada T2N 4N1. E-mail: ggkaplan@ucalgary.ca
Received 23 February 2009; accepted 18 August 2009