Arthritis & Rheumatism, Volume 60,
October 2009 Abstract Supplement

The 2009 ACR/ARHP Annual Scientific Meeting
Philadelphia October 16-21, 2009.

Giant Cell Arteritis: A Negative Association with Cancer?

Molloy,  Eamonn S., Hoffman,  Gary S., Langford,  Carol A.


To ascertain the odds of cancer associated with giant cell arteritis (GCA) as compared with non-GCA controls.


Data was obtained from the 2004 US Nationwide Inpatient Sample, a national hospital discharge database. Diagnoses were identified by ICD9 coding. Records coded for GCA (n=4,566) were matched with 18,264 controls (1:4); matching variables were age, sex, race and admission status (elective/non-elective). Conditional logistic regression was used to calculate odds ratios (OR) for GCA compared to controls, for cancer outcomes (all cancers, hematologic cancer, non-hematologic cancer, metastatic cancer) and non-malignant neoplasms. Demographic and clinical covariates were considered for inclusion in multivariate models, including age, sex, race, income, insurance payer, length of stay, AHRQ comorbidity measures.


Based on both upon univariate (table 1) and multivariate analyses (table 2), GCA was associated with significantly reduced odds for all cancers compared to controls but.the odds of non-malignant neoplasms did not differ between GCA and controls.

Table 1. Results of univariate analyses for GCA versus controls.

OutcomeOdds Ratio95% Confidence IntervalP Value
All cancer0.470.41–0.54<0.0001
Hematologic cancer0.540.39–0.750.0002
Non-hematologic cancer0.500.42–0.59<0.0001
Metastatic cancer0.390.30–0.50<0.0001
Non-malignant neoplasms0.970.79–1.210.81

Table 2. Results of multivariate analyses for GCA versus controls.

OutcomeOdds Ratio95% Confidence IntervalP Value
All cancer0.480.41–0.56<0.0001
Hematologic cancer0.490.35–0.68<0.0001
Non-hematologic cancer0.530.45–0.64<0.0001
Metastatic cancer0.380.29–0.50<0.0001
Non-malignant neoplasms0.860.69–1.060.15


GCA was associated with reduced odds of all measured cancer outcomes compared with non-GCA controls. There was however, no difference in the odds of non-malignant neoplasms between GCA and non-GCA controls, suggesting that the observed reduced odds of cancer associated with GCA was not related to under coding of GCA where that was not the primary diagnosis. Potential, but not evaluable, explanations for these observations include [1] detection and management of pre-cancerous lesions may have occurred during initial evaluation for GCA, [2] more frequent physician visits and thus adherence to malignancy screening guidelines may have occurred in the GCA group [3] an effect of medications for GCA such as prednisone, aspirin and non-steroidal anti-inflammatory drugs could have influenced malignancy outcomes and [4] patients with a predisposition to GCA may be inherently less vulnerable to cancer. Further study will be required to explore these hypotheses.

To cite this abstract, please use the following information:
Molloy, Eamonn S., Hoffman, Gary S., Langford, Carol A.; Giant Cell Arteritis: A Negative Association with Cancer? [abstract]. Arthritis Rheum 2009;60 Suppl 10 :177
DOI: 10.1002/art.25260

Abstract Supplement

Meeting Menu