TB Genotyping in California 200405

1
TB Genotyping in California 2004-05
CTCA Conference May 11, 2006

• Martin Cilnis, MS, MPH
• Tuberculosis Control Branch
• California Department of Health Services

2
Objectives

• Status of the Universal Genotyping Project in
  California 2004-05
• Epidemiology of genotype clusters
• Describe distribution and size of clusters
• Common clusters
• Characteristics of clustering

Genotype shared by at least 2 cases
3
Del Norte
Genotyping Initiative Participation by County
Siskiyou
Modoc
Humboldt
Trinity
Shasta
Lassen
Tehama
Mendocino
Plumas
Glenn
Butte
Sierra
Colusa
Lake
Nevada
Yuba
Sutter
Placer
Yolo
Sonoma
Amador
Napa
El Dorado
Sacramento
Solano
Marin
Alpine
Berkeley
Calaveras
Contra Costa
San Joaquin
San Francisco
Tuolumne
Alameda
37 (61)
San Mateo
Mono
Stanislaus
Santa Clara
Mariposa
Merced
Santa Cruz
11 (18)
Madera
San Benito
Fresno
13 (21)
Monterey
Inyo
Tulare
Kings
San Luis Obispo
Kern
Santa Barbara
Ventura
San Bernardino
Los Angeles
Pasadena
Orange
Long Beach
Riverside
San Diego
Imperial
4
Genotyping participation in CA 2004-05
4597 culture-positive TB cases reported in 2004-05
Universal 54
Selective 13
Not participating 33
1701 (37 of culturepositive) isolates submitted
for genotyping in 2004-05
Universal 62
Selective 14
Not participating 5
5
Percentage of Culture-Positive Isolates Submitted
by Year
6
CA Genotyping Results2004-05
7
Description of Genotype Clusters in California
2004-05

• Frequency and distribution of clusters
• Common cluster types

8
Frequency of Cluster SizeN190 clusters
9
Distribution of Clusters N190 clusters
10
Top 15 Cluster Names,CA 2004-05
11
Top 15 Cluster Names,CA 2004-05
12
Summary Description of Genotype Clusters

• The majority of clusters (54) contain 2 cases
• Most clusters (78) are found in more than one
  jurisdiction
• Local TB programs are unable to track clusters
  outside of their jurisdiction because each
  receives their own genotyping data
• Solution Currently, 12 jurisdictions agreed to
  share genotyping information

13
Summary Description of Genotype Clusters

• The largest clusters (CA_021 and CA_006) are
  widespread throughout the state/world
• RFLP analysis can split clusters
• RFLP done for 12 CA_006 and 14 CA_021 isolates
• 13 (50) unique
• 13 (50) in 5 clusters

14
Analysis of Clusters,CA 2004-05

• Is clustering associated with place of birth?
• U.S.-born and foreign-born?

Analysis includes data from universal
jurisdictions only and excludes clusters CA_021
and CA_006
15
Place of Birth and Clustering
Significant pX2 lt 0.05
16
Analysis of Clusters,CA 2004-05

• Is clustering associated with any TB case risk
  factors?
• Alcohol abuse
• Homelessness
• Injection and non-injection drug use
• Correctional facility resident/employee
• Health care worker
• HIV/AIDS
• Migratory agricultural worker
• Long-term care facility resident
• Multiple risk factrors

17
Case Risk and Clustering
Significant pX2 lt 0.05
18
Summary Place of Birth, Case Risk and Clustering

• U.S.-born cases are more likely to have clustered
  genotypes than FB cases
• Among the FB cases, there is no significance in
  the time since arrival into the U.S. and
  clustering
• Higher proportions of clustering in alcohol
  abuse, homelessness, IDU, and NIDU groups
• Cases with multiple risk factors are more likely
  to cluster than those with only one or none

19
- Multivariate Analysis

• Predictors include
• U.S.-born
• Age lt 45 years
• White or African-American race/ethnicity
• Case risk factors
• Homelessness
• Alcohol abuse
• Non-injection drug use
• Multiple risk factors

20
- Multivariate Analysis

• Having multiple risk factors is a significant
  predictor of clustering regardless whether the
  case is U.S.- or foreign-born
• In U.S. cases, clustering is associated with
  agelt45, African-American race/ethnicity, alcohol
  abuse, and having multiple case risk factors

21
Limitations

• Non-universal genotyping participation selection
  bias resulting in under- or over-estimate of
  clustering
• Incomplete matching to RVCT selection bias
  resulting in under- or over-estimate of
  clustering
• Limited discriminatory power of PCR genotyping
  methods for Beijing and Manila strains results
  in over-estimation of clustering

22
Next Steps

• Increase participation in universal genotyping
• Sharing of genotyping data throughout the state
• Use of genotyping and epidemiologic data in TB
  control

23
Questions about genotyping?

• Contact
• Martin Cilnis
• Epidemiologist, TB Control Branch
• mcilnis_at_dhs.ca.gov
• [phone]