Transcript Prevalence of Pulmonary Nontuberculous Mycobacterial Infections

Epidemiology of Pulmonary Nontuberculous
Mycobacterial Disease among U.S. Medicare
Beneficiaries, 1997-2007
Jennifer Adjemian, PhD
Epidemiology Unit, Laboratory of Clinical Infectious Diseases
National Institute of Allergy and Infectious Diseases (NIAID)
National Institutes of Health (NIH)
Nontuberculous Mycobacteria (NTM)
• Diverse group of small bacilli
– Over 140 species identified
– Widely distributed in the environment
• Species vary by geographic area
• Soil and water (natural and treated) sources
– Biofilms lining pipes for drinking water
NTM Exposure and Infection
• People infected from environmental exposures
– Hard to identify exact source of infection
• NTM species range greatly in human pathogenicity
– Most human infections in U.S. from:
•
•
•
•
•
M. avium complex (MAC) (water, soil, animals)
M. kansasii (tap water)
M. abcessus (more common in SE U.S.)
M. chelonae (keratitis cases from contact lenses/LASIK)
M. fortuitum (water, soil; salon whirlpool footbaths)
NTM Infection & Disease
• Ranges from asymptomatic infection to severe illness
• Infection can manifest in the:
– Lymphatic system
– Skin/soft tissue/bone
– Disseminated
– Lungs (80%)
• Treatment involves multiple antibiotics over
extended period of time
NTM Lung Disease
• Commonly presents as chronic pulmonary disease
– Symptoms:
•
•
•
•
•
•
Chronic or recurring cough (with or without blood)
Sputum production
Fatigue
Shortness of breath
Chest pain
Weight loss
• Often accompanied by coexisting lung diseases
– Bronchiectasis, COPD, and cystic fibrosis
PNTM Prevalence Estimates
• National prevalence estimates are limited
– Not a reportable disease
– Estimates vary by location, age and year
• 1.8/100,000 to 57/100,000
• Trends observed
– Increasing prevalence over time
– Older age groups (average age ~70 yrs)
– White females (≥ 60%), especially with low BMI
Need for National PNTM Prevalence Estimates
• National estimates needed:
– To determine true public health burden of NTM
– To identify most affected segments of US population
– To identify risk factors for disease
Study Objective
• To identify the prevalence and trends of NTM
lung disease in adults ≥ 65 yrs throughout US
Data Compilation
• Data from 5% sample of Medicare Part B enrollees
from 1997-2007
– Exclusion criteria:
• <65 yrs old, covered by an HMO, enrolled for <1
month, or resided outside US
• Data collected on:
– Demographics (sex, race/ethnicity, age, location)
– ICD-9 codes associated with claims
• Scanned for Pulmonary NTM (PNTM) – 031.0
Prevalence Estimates
• Annual prevalence (AP) for each year and period
prevalence (PP) for total time period (1997-2007)
– Stratified by sex, race/ethnicity, age group, state of
residence, geographic region
• Annual percent change (APC)
– Poisson regression models used to evaluate
significant changes in prevalence over time
Risk Factor Analysis
• Univariate logistic regression used to evaluate
associations between PNTM and demographic factors
Results: Population Summary
• 2.3 million individuals included from 5% sample
– Represents ~4% of all US adults ≥ 65 years old
– Demographics similar to general population
PNTM Cases
• 2,548 PNTM cases identified from 1997-2007
– 65% female, 90% white, and ranged 65-98 yrs old
– From all 50 states and DC
• 1/3 from the southeastern US
• > 50% of cases from 8 states: CA, FL, TX, NY, PA, IL, NC, GA
• 13 states averaged < 1 case/yr; ND only reported 1 case
U.S. Distribution of NTM Lung Disease Cases in Adults ≥ 65 years of age
Period Prevalence (PP) from 1997-2007
• PP was 112 (95% CI: 108-116) cases /100,000 persons
– Significantly higher among women than men
• 127 vs. 93 cases / 100,000 persons
– Varied by geographic location and race/ethnicity
PP of PNTM by Sex & Race/Ethnicity
(cases/100,000 persons)
350
300
250
200
150
100
50
0
Males
263
Females
201
135
89
89 92
79
40
White
Black
Asian /
Pacific
Islander
Hispanic
PP of PNTM by State
CA: 191
AZ: 179
LA: 178 MS: 165
FL: 176
HI: 396
Annual Prevalence (AP)
• Avg AP of 31/100,000 persons per yr
• Increased significantly by 8.2% per yr
60
50
40
PNTM 30
Cases/100,000
Persons 20
10
0
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
Male
Female
Total
AP by Region & Year
70
60
50
West
40
Southwest
Southeast
30
Northeast
20
Midwest
10
0
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
Univariate Logistic Regression Analysis
Variable
OR
95% CI
P-Value
Midwest
Northeast
Southeast
Southwest
West
Referent
1.2
1.7
1.7
1.9
-1.0-1.3
1.5-1.9
1.5-2.0
1.7-2.2
-0.01
<0.0001
<0.0001
<0.0001
Male
Female
Referent
1.4
-1.3-1.5
-<0.0001
White
Black
Asian/Pacific Islander
Hispanic
Referent
0.5
2.0
0.8
-0.4-0.6
1.6-2.5
0.6-1.1
-<0.0001
<0.0001
0.1
Region:
Sex:
Race/Ethnicity:
Discussion
• First study to provide nationally representative
estimates of PNTM disease in older adults throughout
the US and across various racial/ethnic groups
• Prevalence of PNTM is increasing significantly across all
regions of the US and among both men and women
– ICD-9 codes likely underestimate prevalence
Increasing Prevalence of PNTM Disease
• Recent shift from middle-aged male smokers to older
women without known risk factors
– Increasing segment of US population
• Improved diagnostics and testing, however increases
noted worldwide and throughout various points in time
• Greater concentration in drinking water
– Resistant to disinfectants and chlorination
Race/Ethnicity and PNTM Disease
• PNTM cases were predominantly white women, but
Asians/Pacific Islander men seem to be at greater risk
• Role of race/ethnicity complex and is likely driven by
behavioral, cultural, and biological/genetic factors
– In Japanese population, PNTM associated with certain
alleles of human leukocyte antigen markers
– Pacific Islanders more likely than Asian Americans to be
smokers, overweight, and have higher chronic disease and
mortality rates despite no differences in access to care
Geographic Differences in Prevalence
• Hawaii had highest prevalence
– ~ ½ Asian/Pacific Islander PNTM cases resided in HI, but
still had ↑est PP even when they were excluded,
suggesting environmental in Hawaii may also contribute
• Higher prevalence in SE and West is consistent with
prior reports of higher human exposure
– ↑ #s of NTM isolated from coastal swamps of SE
– NTM most often isolated from CF patients at study sites in
southern Louisiana and also California
Conclusions
• High risk groups and areas likely due to complex
interactions among environmental and host factors
• With prevalence rising throughout US, clinicians must
increase awareness of those at high risk for NTM
• More studies needed to better understand complex
relationships between host and environment
– Prospective epidemiologic case-control studies
– Spatial epidemiologic analyses evaluating environmental
factors associated with cases on a finer scale
Acknowledgements
• Support for this study was provided by the LCID,
NIAID, NIH, especially:
– Kenneth Olivier, MD
– Amy E. Seitz, MPH
– Steven M. Holland, MD
– Rebecca Prevots, PhD, MPH
Thank you!
Results: PP by Region
• Highest in western states (149/100,000 )
• > 1/3 of cases from Southeast (131/100,000)
• Lowest PP in Midwest (78/100,000 )
West
Midwest
Southwest
Southeast
Northeast
Results: COPD by Race/Ethnicity & Sex
100
87
76
80
60
40
40
86
83
33
43
33
% Female NTM Cases with
COPD
% Female non-NTM Cases
with COPD
20
0
White
100
88
Asian
88
Black
Hispanic
93
83
80
60
43
40
40
38
42
% Male non-NTM Cases
with COPD
20
0
White
Asian
Black
% Male NTM Cases with
COPD
Hispanic
Results: Bronchiectasis by Race/Ethnicity & Sex
60
50
40
30
20
10
0
53
50
30
% Female NTM Cases with
Bronchiectasis
24
1
2
1
White
Asian
Black
2
% Female non-NTM Cases
with Bronchiectasis
Hispanic
37
40
31
27
30
% Male NTM Cases with
Bronchiectasis
20
10
10
1
2
1
White
Asian
Black
0
1
Hispanic
% Male non-NTM Cases with
Bronchiectasis
Results: Medical Claims Analysis
• 16,508 records with pulmonary NTM-related claim
–
–
–
–
64% as primary claim
22% as secondary claim
11% as tertiary claim
4% as quaternary claim
• 75% had pulmonary NTM as primary claim at least once
• 6.5 pulmonary NTM claims submitted on avg per case
Results: AP by Sex & Age Group
60
Males
49
50
Females
43
39
40
33
30
30
27
29
25
21
20
17
10
0
65-69
70-74
75-79
80-84
≥ 85
Annual prevalence of PNTM cases among a sample of
U.S. Medicare Part B enrollees by sex from 1997-2007
60
50
40
PNTM
30
Cases/100,000
Persons
Male
20
Total
10
0
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
Female
Results: Age Distribution of NTM Cases
Average Annual % of Enrollees and Cases by Age Group
30
26
24 23
25
20
% of Total Population
27
% of NTM Cases
21
20
18
15
15
13
12
10
5
0
65-69
70-74
75-79
80-84
> 84
Results: Multivariate Logistic Regression
Variable
Region:
Southeast
Northeast
Midwest
Southwest
West
Age Class:
65-69 years
70-74 years
75-79 years
80-84 years
≥ 85 years
Sex:*
Male
Female
Race/Ethnicity:*
White
Black
Asian/Pacific Islander
Hispanic
North American Native
AOR
95% CI
P-Value
Referent
0.7
0.6
1
1.2
-0.6-0.8
0.6-0.7
0.9-1.2
1.0-1.3
-<0.0001
<0.0001
0.6
0.01
Referent
1.2
1.4
1.3
1
-1.1-1.4
1.2-1.5
1.1-1.5
0.9-1.2
-0.004
<0.0001
0.0002
0.6
Referent
NA
-NA
-NA
Referent
NA
NA
NA
NA
-NA
NA
NA
NA
-NA
NA
NA
NA
*Significant interaction detected; †Controlled for confounding by COPD and Bronchiectasis
Significant Interactions between Sex & Race/Ethnicity
Interaction Term
Level of Variables
Estimated
Race/Ethnicity by Sex: White vs. Black Females
AOR
95% CI
2.5
1.8-3.3
White vs. Hispanic Females
1.8
1.2-2.8
White Females vs. Males
1.4
1.3-1.7
Black Males vs. Females
1.8
1.2-2.7
Asian vs. White Males
2.5
1.7-3.3
Asian vs. Black Males
2
1.4-3.3
2.6
1.4-4.7
Asian vs. Hispanic Males
Estimating Disease Prevalence through U.S.
Medicare Data
• Centers for Medicare and Medicaid’s Carrier
Standard Analytic File (SAF)
– 5% sample of all medical claims submitted for individuals
enrolled in Medicare Part B (~44 million)
– Final action claims data from non-institutional providers
• i.e. physicians, independent clinical laboratories, ambulance
providers, and stand-alone ambulatory surgical centers
– Claims include primary ICD-9 code and up to 3 additional
codes for other conditions associated with the claim
• Representative of U.S. population ≥ 65 yrs
Mortality Analysis
• PNTM cases:
– 35% overall all-cause mortality rate
– 40% more likely to die than non-cases (OR=1.4 [1.3-1.6])
• Among PNTM cases, risk of death greater in:
– Men (OR=1.8 [1.6-2.2]) relative to women
– Blacks (OR=2.1 [1.4-3.1]) and Hispanics (OR=2.4 [1.2-4.7])
relative to whites
– Those with ≥ 3 comorbidities (OR=1.3 [1.0-1.6])
– Those also diagnosed with lung cancer (OR=1.7 [1.4-2.1],
asthma (OR=1.7 [1.1-2.7]) or pneumonia (OR=2.9 [2.3-3.6])
AP by Race/Ethnicity & Year
70
60
50
40
30
20
10
0
40
30
White
Males
20
Black
Males
White
Females
10
Black
Females
0
200
150
100
Asian/PI
Males
50
Asian/PI
Females
0
Diagnosis with NTM Lung Disease
• Clinical/Radiologic
– Pulmonary symptoms, nodular/cavitary opacities on chest
radiograph, or high-resolution computed tomography (HRCT)
scan showing multifocal bronchiectasis with multiple nodules
– Appropriate exclusion of other diagnoses. (i.e. TB)
• Microbiologic
– Pos. culture on ≥ 2 separate sputum samples
– Pos. culture on ≥ 1 bronchial wash or lavage
– Transbronchial or other lung biopsy with mycobacterial
histopathologic features
Annual Percent Change (APC)
•
Overall APC of 8.2% (95% CI: 7.2-9.2) per year
Sex
Age Group
Female --65-69
70-74
75-79
80-84
≥ 85
Male --65-69
70-74
75-79
80-84
≥ 85
APC
9.1
7.5
10.4
9.3
11.2
5.9
6.4
6.4
9.8
6.4
2.9
4.4
95% CI
6.5-11.7
5.3-9.8
6.8-14.2
6.9-11.7
7.7-14.8
3.3-8.5
3.7-9.2
2.7-10.2
5.9-13.9
2.9-9.9
-1.7-7.6
1.8-7.1
Comorbidity Analysis
Comorbidity
COPD*
Pneumonia
Acute Bronchitis
Bronchiectasis
Pulmonary Fibrosis
Mitral Valve Disorder
Diabetes Mellitus
Cancer of the Trachea,
Bronchus, & Lung
Pulmonary Hypertenstion
Asthma
Aspergillosis
Anomalies of the larynx,
trachea, and bronchus
HIV
% of
NTM
Cases
87
82
66
44
43
37
26
% of nonNTM Cases
Relative
Prevalence
40
33
32
1
7
21
29
2.2
2.5
2.1
38.1
6.2
1.7
0.9
17
5
3.8
7
3
2
2
1
0.1
2.7
3.4
29.3
0.4
0.05
8.1
0
0
0
* Prevalence of COPD likely overestimated due to issues associated with Medicare claims coding
Prevalence of selected comorbidities among
male and female PNTM cases
HIV
Diabetes Mellitus
Mitral Valve Disorder
Acute Bronchitis
Pulmonary Hypertenstion
Females
Pneumonia
Asthma
Males
Trachea, Bronchus, & Lung Cancer
Cystic Fibrosis
Pulmonary Fibrosis
Larynx, Trachea, & Bronchus Anomalies
Aspergillosis
Bronchiectasis
0
10
20
30
40
50
Comorbidities in PNTM Cases
• Significantly more likely to be diagnosed with other
and greater #s of comorbidities than non-cases
• Almost all PNTM cases were diagnosed with at least
one other lung-associated condition
• All prevalence estimates for PNTM and comorbidities
are limited to accuracy of ICD-9 code reporting
PP of PNTM by race/ethnicity and state
600
500
White
PNTM Cases
per 100,000
Persons
400
Black
300
200
Asian/Pacific
Islander
100
0
*The prevalence among blacks in Hawaii (2381/100,000 persons) likely skewed due to the presence of only one
PNTM case among a small sample of black individuals in this dataset residing in Hawaii (n=41).