Trends and Seasonal Patterns in Hospitalization Rates of Clostridium Difficile in the US Elderly Jyotsna S. Jagai, Sara M. Parisi, Meena P. Doshi, Elena N. Naumova Tufts University School of Medicine APHA Conference, Abstract # November 6, 2007
2 Study Introduction Clostridium difficile associated diarrhea Increasing incidence in the US Both in incidence and severity Major hospital acquired gastrointestinal infection Increased rates associated with hospital and long-term care facilities Risk factors Antibiotic use Advanced age
3 Antibiotic Resistance (AR) Worldwide increases in resistant strains of bacteria. Penicillin makes up the majority of resistance, but others are growing. AR rates vary considerably based on study location and bacteria studied. Many social factors affect resistance rates Prescription patterns Patient compliance Hospital environment Major concern that vancomycin resistance will increase.
4 Study Objective Explore differences in patterns between rates of C. difficile and percentage of C. difficile with antibiotic resistance By gender By race Geographic distribution Annual trend Seasonality
5 Description of Data Centers for Medicare & Medicaid Services (CMS) database Part A Hospitalization Claims, 1991 – 2004 ~Total U.S. population ,865,046 ~U.S. elderly (≥65yo) ,688,844 (12.75%) Medicare beneficiaries~ 95% Number of CMS records197,315,577 Number of CMS records/year14,093,969 Clostridium difficile cases 1,054,125 (62% Female) Clostridium difficile annual rate21.71 per 10,000 elderly (ICD9CM – ) AR C. difficile cases 19,654 (1.86% of C.difficile) (ICD9CM – V09.(0-9) codes)
6 Diagnosis Position of C.difficile
7 Estimation of Rates is center of the data range, Population data available from 1990 & 2000 census Calculated a population estimate to represent the middle of the range “Growth over 7.5 years”=(2000pop-1990pop)/10)*7.5) ~ 1997pop = 1990pop +”Growth over 7.5 years” This population estimate was used to calculate annual rates of Clostridium difficile, per 10,000 elderly.
8 Description of AR cases Represented by V-codes Supplementary ICD9CM codes V09 – Infection with Drug-Resistant Micro- Organisms The AR V-codes introduced in October Coded in the 3 rd -10 th diagnosis slots Supplement to the primary diagnoses V09.0-V09.9 all represent AR, but within different drug classes
9 Description of AR cases V09 Group Description Infection with microorganisms resistant to: V09.0Penicillin (Methicillin-resistant staphylococcus aureus (MRSA)) V09.1Cephalosporins and other B-lactam antibiotics V09.2Macrolides V09.3Tetracyclines V09.4Aminoglycosides V09.5Quinolones and fluoroquinolones V09.6Sulfonamides V09.7Other specified antimycobacterial agents Excludes: Amikacin (V09.4), Kanamycin (V09.4), Streptomycin [SM] (V09.4) V09.8Other specified drugs Includes: Vancomycin (glycopeptide) intermediate s. aureus VISA/GISA, Vancomycin (glycopeptide) resistant enterococcus VRE, Vancomycin (glycopeptide) resistant s. aureus V09.9Microorganisms, unspecified
10 Distribution of AR V-codes Penicillin Vancomycin
11 Rates by Age and Race Categories White All African American Hispanic North American Native Asian Hispanic African American Asian North American Native Counts for 85+ age category White5,158 African American 398 Hispanic56 Asian41 N. American Native 4
12 Distribution C. difficile Average annual rate of C. difficile hospitalizations for U.S. elderly, ages 65 and over,
13 Distribution C. difficile cases with AR Average annual rate of C. difficile hospitalizations for U.S. elderly, ages 65 and over, 1991 – 2004 superimposed with percentage of cases with AR (1195) (1136) (411) (50)
14 Trend Analysis
15 Seasonality
16 Annual Harmonic Regression Gamma ( ) Seasonal Peak (max) Seasonal Nadir (min) Intensity Peak timing
17 Seasonality of C. difficile Average peak for C. difficile was seen in the 36 th week (early Sept.) Peak timing ranged between the 31 st and the 39 th week.
18 Seasonality of AR Cases Amplitude levels off ~1998. Peak occurs between the 16 th and 38 th week.
19 Summary Females demonstrated rates of Clostridium difficile which were ~1.2 times higher than males however, males demonstrated a higher percentage of antibiotic resistant cases (~1.2 times). Rates of C. difficile are ~5 times higher in the white and African American populations compared to the Asian and Hispanic populations. However, the percentage of cases with antibiotic resistance were lower in white and African American. The highest rates of C. difficile were seen in the central Midwest and through Ohio and Pennsylvania. The percentage of AR cases demonstrated small, but significant correlation with C. difficile hospitalization rates.
20 Summary There is an increasing trend in hospitalization rates for C. difficile and percentage of AR over the 14 year study period. There is a stabilization in percentage of cases with AR in 1997 at ~ 2%. C. difficile demonstrated a seasonal peak in the 36 th week (early September) however, percentage of AR cases has a less defined seasonal pattern. This descriptive study demonstrates that there is a need to understand the relationship between Clostridium difficile and cases of antibiotic resistance.
21 Thank you! Support by the Gastroenteritis and Extreme Weather Events in Elderly (GEWEL) Project funded by NIEHS (NIEHS ES ). Guidance and comments from team members of the Tufts Initiative for the Forecasting and Modeling of Infectious Diseases (InForMID)* Conference sponsorship by the APHA Environmental Health Section and Friedman School of Nutrition Science and Policy at Tufts University. *
22 Percent US Elderly, Ages 65 and Over, in Nursing Homes per County
23 Hospital Locations Superimposed on the Number of Hospital Beds per 100 Residents Ages 65 and Older