Authors: Juanmanuel Gomez, MD, J Robert Cantey, MD & Kit N. Simpson, Dr.Ph
Title: Epidemiology of Prosthetic Joint Infections (PJI) in South Carolina
Addresse: 100 Doughty St, PO BOX 250752, Charleston, SC 29425.
Purpose: In the 1990's, utilization of Total Hip Arthroplasty (THA) grew from 81 to 92/105 and Total Knee Arthroplasty (TKA) from 60 to 104/105. Most reports on risk factors for PJI come from academic centers with programs focused on joint replacements. Little is known in regards to PJI risk factors in a community-wide setting. The purpose of this study was to assess risk factors for PJI among patients discharged from SC hospitals from 1998-2002.
Methods: The Office of Research and Statistics of SC maintains a statewide hospital discharge database, we analyzed years 1998-2002. We created two datasets using ICD-9-CM code 996.66 (PJI) and codes 8151 (THA) / 8154 (TKA). We excluded duplicate records, age < 25 years or residing in a non SC county. We analyzed demographics, 19 comorbid conditions, hospital charge, length of stay (LOS), Charlson comorbidity score, county of residency and probability of having arthoplasty in a small hospital (< 100 beds). We used multivariable logistic regression to explore variables associated with PJI. Statistical analysis done with SAS 9.0.
Results: There were 30479 records in the "Arthroplasty" dataset vs. 1417 in the PJI dataset. "Arthroplasty" mean age was 69 vs. 67 for PJI. Percent blacks was 18 vs. 22, female sex 67 vs. 61, respectively. The mean Charlson comorbidity score was 0.56 vs. 0.64. Mean LOS for PJI was 11.12 days with a mean hospital charge of $32790. There were194 PJI discharges in 1998 in contrast to 366 in 2002 (88% increase). TKA-THA discharges increased 34% in the same interval. By multivariable logistic regression we found male OR 1.21, 95% CI (1.09-1.36), black 1.22 (1.12-1.48), chronic renal insufficiency 2.54 (1.25-4.42), cirrhosis 3.95 (2.16-6.02) and malnutrition 6.23 (4.51-9.38) positively associated with PJI, even after controlling for county of residency. After controlling for all other variables only county "X" was significantly associated with PJI. When probability of having arthroplasty surgery in a small hospital was entered into the regression model, it was found to have a strong association with PJI OR 2.16 (1.49-3.13).
Discussion: Our study includes the 47 hospitals and their multiple surgical teams performing THA/TKA in SC. The diversity of this population gives significant strength to our findings. On multiple logistic regression models, black race had 22% (p = 0.0036) higher chance of developing a PJI. Also, male gender, chronic renal insuficiancy, cirrhosis and malnutrition were positively associated. Residence in County X was associated with a 39% (p = 0.0003) increased risk of developing PJI, but this finding seems to be better explained by the probability of having surgery in a small hospital.
Significance: Community wide analysis like ours will show novel risk factors associated with PJI. Black race must be considered a significant risk factor in future analysis. Geographic variation in the risk of PJI is an intriguing finding, probably related to the health care quality and experience of the surgical team
