Predictive epidemiology

Document Sample
Predictive epidemiology Powered By Docstoc
					          Control




        Interventions




Agent   Individual      Community
 Types of Approaches to the
Epidemiology of Tuberculosis
  o Etiologic epidemiology
  o Descriptive epidemiology
  o Predictive epidemiology
                     A Model for the Epidemiology of Tuberculosis



 Risk                   Risk                    Risk                         Risk
factors                factors                 factors                      factors

                                                           Infectious
                                                          tuberculosis
                                 Subclinical
          Exposure                infection                                              Death

                                                         Non-infectious
                                                          tuberculosis



                                                                   Rieder HL. Infection 1995;23:1-4
     Exposure to
    M. tuberculosis

o Etiologic epidemiology
o Descriptive epidemiology
o Predictive epidemiology
    Risk factors for exposure to
          M. tuberculosis

o    Incidence of infectious tuberculosis

o    Duration of infectiousness

o    Number of contacts with susceptibles
     per unit of time of infectiousness
                                     Number of Children per Private Household by
                                       Census Year, Switzerland, 1920 - 1990
                                2
No. of children per household




                                1




                                0
                                    1920           1940             1960             1980
                                                          Year of Census
                                      Swiss Federal Office of Statistics. Statistisches Jahrbuch 1994
                                            Change in the Median Age of Tuberculosis Patients,
                                             Finland, Among Males and Females, 1954 - 1995
                                  70
Age-adjusted median age (years)




                                  60



                                  50                         Males



                                  40                                  Females



                                  30


                                                      60             70           80           90
                                                                 Year of notification
                                       Härö AS. Tuberculosis and Respiratory Diseases Yearbook 1998;24:101-3
In just 30 years the median age of tuberculosis patients in
Finland has moved from the generation of parents to that
of grandparents.

People preferentially socialize with peers of the same age,
thus interactions are intra-generational. In the case of
parents, the important interactions are both inter- and intra-
generational. In Europe, it is largely the parents (not the
grandparents) who take care of the youngest. In other
societies this might differ.

In any case, the age in which tuberculosis predominantly
occurs, will have important repercussions on the risk of
exposure for different population segments.
     Infection with
    M. tuberculosis

o Etiologic epidemiology
o Descriptive epidemiology
o Predictive epidemiology
    Risk of Infection Given Exposure:
       Largely Exogenous Factors

     Particles    x     Exposure time
     Volume

Particles:       Production of infectious droplet nuclei
Volume:          Volume of air and ventilation
Exposure time:   Time of inhaling air with droplet nuclei
                                   Settling of Water Droplets in Saturated Air

                    10000
Time (sec) of falling
  2 m (log scale)




                        1000


                         100


                          10


                           1


                               0.001           0.01                0.1                    1
                                         Diameter of droplet (mm) (log scale)

                                                       Wells WF. Am J Hyg 1934;20:611-8
                               Evaporation Time of Water in Unsaturated Air at 18o C
                         700
Evaporation time (sec)




                         600
                         500
                         400
                         300
                         200
                         100
                           0


                                  0.05     0.1        0.2          0.5        1         2
                                                 Diameter of droplet (mm)
                                                            Wells WF. Am J Hyg 1934;20:611-8
Falling and Evaporation Times of Droplets of Varying Diameters
                                                   Droplet diameter (mm)
                             0.00           0.05           0.10            0.15               0.20
Falling time (seconds)




                         1




                                    To evaporation
                         2


                                                                                  To ground
                                                                                    (2m)
                         3

                                                                  Wells WF. Am J Hyg 1934;20:611-8
 Production of Infectious
Droplets and Droplet Nuclei

  o Aerosolization
  o Site and form of disease
                             Number of Droplets produced by Different
                                 Aerosol Producing Maneuvers
                     50000


                     40000
Number of droplets




                     30000


                     20000


                     10000


                         0
                                 Coughing         Talking         Singing

                             Loudon RG, et al. Am Rev Respir Dis 1968;98:297-300
                              Cumulative Percentage of Droplets Expelled During
                                Aerosol Producing Maneuvres, by Droplet Size
                        100
Cumulative percentage




                         80


                         60              Coughing


                         40

                                         Singing
                         20

                                                    Talking
                          0

                              1      3     5 7 10                 100               1000
                                             Droplet diameter (micrometers)
                                          Loudon RG, et al. Am Rev Respir Dis 1968;98:297-300
                   Proportion of Contacts <15 Years with Tuberculous
                  Infection, in Relation to Cough Frequency of Source
                          50
Per cent with infection




                          40


                          30


                          20


                          10


                           0
                                      <12.0        12.0-47.9      48.0+
                                              Mean cough frequency

                               Loudon RG, et al. Am Rev Respir Dis 1969;109-111
                         Tuberculous Infection Among Close Contacts (Children <15 yr)
                                    by Bacteriologic Status of Index Case
                    70       s+/c+


                    60
                                     s-/c+
                    50
Per cent infected




                    40                       s-/c-

                                                         s+/c+
                    30
                                                                                    s+/c+
                    20
                                                                         s-/c-
                                                                                            s-/c+   s-/c-
                    10                                           s-/c+

                     0
                              Bedfordshire                   Rotterdam                Saskatchewan
                              1948 - 1952                   1967 - 1969                1966 - 1971
                                                            Shaw JB, et al. Am Rev Respir Dis 1954;69:724-32
                                                     van Geuns HA, et al. Bull Int Union Tuberc 1975;50:107-21
                                                      Grzybowski S, et al. Bull Int Union Tuberc 1975;50:90-106
                                     Sensitivity of Direct Sputum Smear Examination
                                 in Identifying Pulmonary Tuberculosis and Transmitters
                               1.0                                                     Fraction due to
Fraction of cases / infected



                                                                                       smear-neg cases
                                                  Smear-neg
                               0.8                Culture-pos

                               0.6
                                                                                       Fraction due to
                                                                                       smear-pos cases
                               0.4
                                                  Smear-pos
                                                  Culture-pos
                               0.2

                               0.0
                                                   Cases of               Infected
                                                  pulmonary               contacts
                                                 tuberculosis              < 15 yr

                                                                               Calculated from data from:
                                     Grzybowski S, et al. Bull Int Union Tuberc Lung Dis 1975;50:90-106
                Smear-Microscopy Results in Culture-Confirmed
             Pulmonary Tuberculosis, by HIV Status, Lusaka, Zambia
                              100
                                      >20
(No. of bacilli / field FM)




                                                            >20
                               80
  Per cent with result




                                     11-20

                               60
                                      1-10                 11-20

                               40
                                                            1-10
                               20     neg
                                                            neg
                                0
                                    HIV-pos              HIV-neg

                                        Elliott AM, et al. BMJ 1990;301:412-5
                    Tuberculous Infection in Young High-Risk Contacts
                        by Isoniazid Susceptibility of Index' Strain
                    40
Per cent infected




                    30


                    20


                    10


                     0
                                    INH-resistant       INH-susceptible

                         Snider DE, Jr, et al. Am Rev Respir Dis 1985;132:125-32
 Clearance of Air

o Dispersion of bacilli
o Survival of bacilli
                      Killing of BCG by Sunlight and in Open Sky Without
                            Direct Exposure to Direct Rays of the Sun
                      7
                     10

                      6
                     10
Number of colonies




                      5
                     10
   (log scale)




                      4
                     10
                      3
                     10
                                                                                 Skyshine
                      2
                     10                     Sunlight
                      1
                     10

                           05 15 30    60                                              240
                                            Exposure time (minutes)
                                      Edwards LB, et al. Bull World Hlth Organ 1952;5:333-6
Extent of Contact

o Proximity
o Length of contact
                         Tuberculous Infection Among Children by Type of
                          Contact and Bacteriologic Status of Index Case,
                         British Columbia and Saskatchewan, 1966 - 1971
                    40        Close
                    35
                    30
Per cent infected




                    25
                    20
                    15                 Casual                Close
                    10
                                                                      Casual
                     5
                     0
                                 Smear +                       Smear -
                              Grzybowski S, et al. Bull Int Union Tuberc 1975;50:90-106
                    Prevalence of Tuberculous Infection by Proximity of Contact
                        to a Smear-Positive Case, Jinan, China, 1993-1996

                    40
Per cent infected




                    30


                    20


                    10


                     0
                             Household   Household    Household    Community
                              intimate    regular      sporadic     member

                                           Extent of contact
                                          Lutong L, et al. Int J Tuberc Lung Dis 2000;4:275-7
                                      Prevalence of Tuberculous Infection Among Inflight
                                     Personal in Contact with a Tuberculous Staff Member
                               100
Per cent tuberculin positive




                                80


                                60


                                40


                                20


                                 0
                                      0      10       20        30        40         50        60
                                              Total flight time with index case (hours)
                                                            Driver CR, et al. JAMA 1994;272:1031-5
Risk of Infection Among Contacts as a Function of the Proximity of Contact
Potential Transmitters of M. tuberculosis

Persons who cough
Persons with sputum positive for acid-fast bacilli
Persons not on chemotherapy
Persons just started on chemotherapy
Persons with a poor response to chemotherapy
Surgery

          30 meters


                      TB
                      Amb

                            District Hospital
            Advising the District Medical Officer

The number of infectious particles (tubercle bacilli) expelled into
the air in this environment must be considered far above the average
encountered in other areas of the city. While all patients are on
treatment, some have commenced treatment only recently.

The volume of air into which these bacilli are expelled to is, however,
infinite given a distance from the surgical ward of 30 meters (outdoors
proximity matters).

The survival of bacilli outdoors in East Africa is likely to be limited
because of frequent sunshine.

The length of exposure of surgical patients is by definition limited.

The District Medical Officer might be advised that the risk assessment
allows construction of the ambulatory as planned.
     Infection with
    M. tuberculosis

o Etiologic epidemiology
o Descriptive epidemiology
o Predictive epidemiology
  Descriptive epidemiology of
     tuberculous infection

o Methodological issues in determining
   prevalence of tuberculous infection

o Prevalence of tuberculous infection
  Milestones in the History of Tuberculin

1890         Koch Old Tuberculin
1907         von Pirquet scratch test
1909         Moro patch test
1909         Mendel intradermal test
1910         Mantoux intradermal test
1924         Hastings no-lesion, reactor cattle
1926         Crawford heterologous antigens
1934         Seibert PPD-S
1950ies      WHO surveys
1958-65      US Navy recruit study
1965-95      Korea surveys
   Standardization of the
    Tuberculin Skin Test

o Standardization of tuberculin
o Standardization of administration
o Standardization of reading
Cumulative Percentage of Children Reacting to Increasing Doses
of Tuberculin, Among Healthy Children and Tuberculosis Patients
                                                                       5 TU PPD-S
                    100
                                     Tuberculosis patients
                     80
Per cent reacting




                     60
                                                                                                 Healthy children
                     40


                     20


                      0
                               -9        -8        -7        -6        -5        -4        -3        -2        -1        0
                          10        10        10        10        10        10        10        10        10        10
                                                        Dose of tuberculin (mg)
                                              Furcolow ML, et al. Public Health Rep 1941;56:1082-1100
                         Comparison of Tuberculin PPD-S and PPD RT23 in
                          Eskimo Children and United States Navy Recruits
                    20          Eskimo children                2 TU PPD RT23
                    15

                    10
                                                      5 TU PPD-S
                     5
Per cent reacting




                     0

                    20
                                      2 TU PPD RT23         US Navy recruits
                    15

                    10

                     5       5 TU PPD-S

                     0
                         0        5        10         15      20       25        30
                                                Induration (mm)
                              Comstock GW, et al. Bull World Health Organ 1964;31:161-170
This slide is not available
This slide is not available
This slide is not available
                                Tuberculin Skin Test Reaction Sizes Among
                               Aboriginals with a BCG Scar, Alberta, Canada

                    0.20
Fraction reacting




                    0.15


                    0.10


                    0.05


                    0.00
                           0             10                20               30
                                                 Induration (mm)

                                   Enarson DA. Int J Tuberc Lung Dis 1998;2(suppl 1):S16-S22
                Distribution of Reaction Sizes to 5 TU Tuberculin PPD-S
                      in 5,544 Tuberculosis Patients, United States
                    20
Per cent reacting




                    15


                    10


                     5


                     0
                         0           5         10        15        20        25         30
                                                    Induration (mm)
                             Edwards LB, et al. Am Rev Respir Dis 1969;99(4, part 2 of 2):1-132
                             Tuberculin Skin Test Reaction Size Distribution
                           in Two Tuberculin Surveys, Korea 1965 and 1995

                    0.06                                                          1965
Fraction reacting
                    0.04


                    0.02


                    0.00
                           0      5      10         15        20        25        30


                    0.06
                                                                                  1995
Fraction reacting




                    0.04



                    0.02



                    0.00
                           0      5      10         15        20        25        30
                                              Induration (mm)
                                              Korean Institute of Tuberculosis 1966:1-181
                                         Korean National Tuberculosis Institute 1996:1-180
                            Frequency Distributions of Tuberculin Skin Test
                           Reaction Sizes in Surveys in Tanzania and Djibouti
                    0.10
                                Tanzania
                    0.08

                    0.06

                    0.04

                    0.02
Fraction reacting




                    0.00
                            0              5   10         15      20       25      30      35
                    0.10

                    0.08        Djibouti

                    0.06

                    0.04

                    0.02

                    0.00
                            0              5   10         15      20       25      30       35
                                                         Induration (mm)
                                                     Styblo K. TSRU Progress Report 1998;1:31-66
                                                    Data courtesy Djibouti: Trébucq A, IUATLD, 1995
                                     Frequency Distribution of Tuberculin Skin Test Results
                                         by Age, in Different Areas of the United States


                    15       50 +
                                                                                  10
                                           Minnesota                                                       North Carolina

                                  40-49                                                           40 +
Per cent reacting




                                                              Per cent reacting
                    10


                                                                                   5              20-39
                                   30-39
                     5


                                   18-29                                                           10-19
                                                                                           0-9

                     0                                                             0
                         0   10             20           30                            0     10              20             30

                                                       Induration (mm)

                                                                        Edwards LB, et al. Acta Tuberc Scand 1959;47(suppl):77-97
Sensitivity to M. intracellulare in the United States




      Edwards LB, et al. Am Rev Respir Dis 1969;99:(No 4, part 2):13
                    Recovery of M. avium complex from Savannah River
                  Floodplain (USA) Soil Samples, as a Function of Soil pH
                400
CFU / mg soil




                300


                200


                100


                  0
                          4               5                 6                  7
                                                pH
                                Brooks RW, et al. Am Rev Respir Dis 1984;130:630-3
                        Example of a Composite Distribution Resulting from
                         Specific and Non-Specific Reactions to Tuberculin
                  100
                                                                 Tuberculous infection
                                                                  yes    no total
                   80
                                                           pos      a     b    a+b
                                                      Test neg      c     d    c+d
                                                           total   a+c b+d       N
Number reacting




                   60
                                         Test criterion


                   40

                                                   b
                   20

                                           c
                   0
                        0     5     10           15        20      25      30       35
                                            Induration (mm)
                                               Rieder HL. Tuber Lung Dis 1995;76:114-21
Example for the Dependence of the Predictive Value
       of a Positive Test on the Prevalence
     Example 1:                               Example 1:
      90% Sensitivity                          90% Sensitivity
      90% Specificity                          90% Specificity
      40% Prevalence                           10% Prevalence


        Condition                             Condition
Test present absent      total         Test present absent total
 pos 360        60        420           pos 90        90 180
 neg    40     540        580           neg 10       810 820
 tot   400     600       1000           tot 100      900 1000


Predictive value of a positive test:   Predictive value of a positive test:
   360 / 420 = 86%                        90 / 180 = 50%
                         Probability of Tuberculous Infection Among US Navy Recruits
                             by Tuberculin Skin Test Size and History of Contact
                               100           History of contact
Probability of Infection (%)




                                80


                                60


                                40

                                                           No history of contact
                                20


                                 0
                                     0   5         10             15          20          25
                                                  Induration (mm)

                                                        Rust P, et al. Am J Epidemiol 1975;101:311-22
                    Correlation Between Prevalence of Tuberculin Reacting
                        Cattle and Frequency of Cattle with no Lesion
                    60
Per cent reacting




                    40




                    20




                    0
                         0       10              20              30             40
                                      Per cent with no lesion

                                      Hastings EG, et al. Am Rev Tuberc 1930;22:218-25
With progress of the tuberculosis eradication
program among cattle in the United States, the
prevalence of tuberculous infection among cattle
decreased.

As a result, the predictive value of a positive
tuberculin skin test decreased: an increasing
proportion of reactors had no tuberculous infection
at necropsy.
                      Distribution of Tuberculin Skin Test Reactions 2 1/2 Months to
                    4 to 5 Years After BCG Vaccination in Madras and in Puerto Rico
                    30                                       At 2 1/2 months
                             Chingleput
                               At 4 years
                    20


                    10
Per cent reacting




                     0
                         0         5        10          15           20        25   30
                    20
                                                 Induration (mm)to 5 years
                                                             At 4
                             Puerto Rico

                    10




                     0
                         0          5       10          15           20        25   30
                                                 Induration (mm)
           Tuberculosis Prevention Trial, Madras. Indian J Med Res 1980;72(suppl):1-74
                               Comstock GW, et al. Am J Public Health 1974;64:283-91
                                    Mixture and Component Distributions                  Sensitivity and Specificity at Different Cut-Offs

                      0.30                                                         1.0
                                                                                                                             Specificity
                      0.25                                                         0.8
Fraction reacting




                      0.20
                                                                                   0.6
                      0.15
                                                                                   0.4
                      0.10

                      0.05                                                         0.2                                       Sensitivity

                      0.00                                                         0.0
                             0         5      10      15      20      25      30         0      5       10      15      20        25       30
                                                Induration (mm)
                                 Probability to Belong to the Left Distribution          Probability to Belong to the Right Distribution

                       1.0                                                         1.0

                       0.8                                                         0.8
        Probability




                       0.6                                                         0.6

                       0.4                                                         0.4

                       0.2                                                         0.2

                       0.0                                                         0.0
                             0         5      10      15      20      25      30         0      5       10      15      20        25       30
                                                   Abscissa                                                  Abscissa
Observed Distribution of Tuberculin Skin Test Reaction Sizes,
Mixture Distribution, and Estimated Prevalence of Tuberculous
                 Infection, Tumkur, South India
           0.08                                                       Males
                                                                      25-29 yrs

           0.06
Fraction




           0.04



           0.02



           0.00
                  0     5         10         15        20        25         30
                                       Induration (mm)
                      Data: Narain R, et al. Bull World Health Organ 1963;29:641-64
                      Mixture analysis: Neuenschwander BE, unpublished data, 2000
Summary Tuberculin Test

 o Standardized tuberculin
 o Standardized technique
 o Unpredictable specificity
 o Weakness of cut-off points
 o Mixture analysis promising
        Infection with
       M. tuberculosis

Prevalence of tuberculous infection
in different population segments
                              Age-Specific Prevalence of Tuberculous Infection
                                     in Healthy Children, Paris, 1910
                    100


                     80
Per cent reacting




                     60


                     40


                     20


                      0
                          0                 5                 10                 15
                                                Age (years)
                                                 Mantoux C. Presse Méd 1910;2:10-13
                                            Prevalence of Tuberculous Infection
                                           by Age and Sex, Denmark, 1950 - 1952
                                                                                             Males
                              80                                                             Females
Prevalence of infection (%)




                              60


                              40


                              20


                               0
                                   0   5    10 15        25       35       45       55
                                                               Age (years)

                                           Groth-Petersen E, et al. Bull World Health Organ 1959;5:5-49
                                       Prevalence of Tuberculous Infection by Age
                                              and Sex, India, 1961 - 1962
Prevalence of infection (%)



                                                                              Males
                              60

                                                                              Females
                              40



                              20



                               0
                                   0   5   10 15   25     35     45    55
                                                        Age (years)
National Tuberculosis Institute Bangalore. Bull World Health Organ 1974;51:473-88
                                Male-to-Female Ratio in Prevalence of Tuberculous,
                             Infection by Age, India 1961-1962 and Denmark, 1950-52
                       1.8
Male-to-female ratio




                       1.6

                       1.4
                                                                                        India
                       1.2
                                                                                        Denmark
                       1.0

                       0.8


                                0   5   10 15       25       35      45       55
                                                          Age (years)
                       National Tuberculosis Institute Bangalore. Bull World Health Organ 1974;51:473-88
                                           Groth-Petersen E, et al. Bull World Health Organ 1959;21:5-49
                             Prevalence of Tuberculous Infection by Age
                                    and Residence, Korea, 1965
                                                                                 Urban
                    80
Per cent infected




                    60
                                                                                 Rural

                    40


                    20


                     0
                         0       5       10       15         20         25         30
                                              Age (years)
                                                 Korean Institute of Tuberculosis, 1996
                             Prevalence of Tuberculous Infection by Age
                                 and Residence, in Zanzibar, 1961
                                                                                      Urban

                    40
Per cent infected




                                                                                      Rural



                    20




                     0
                         0      5       10      15       20        25   30       35         40
                                                     Age (years)
                             Roelsgaard E, et al. Bull World Health Organ 1964;30:459-518
                                   Urban-to-Rural Ratio of Age-Specific
                                 Tuberculous Infection, Zanzibar and Korea
                       2.5
                                                    Zanzibar
Urban-to-rural ratio




                       2.0



                       1.5

                                                                     Korea
                       1.0

                             0             10              20              30              40
                                                       Age (years)

                                                          Korean Institute of Tuberculosis, 1966
                                 Roelsgaard E, et al. Bull World Health Organ 1964;30:459-518
Prevalence of Tuberculous Infection by Age and Socioeconomic
    Status in Heidelberg, Trondheim, and Hué, 1910 - 1920

                               60                              Trondheim
 Prevalence of infection (%)



                                                                                           o   r
                                                                          or            Po
                                                                     Po

                                                                                                   Hué
                               40
                                                          or                            Rich
                                                     Po



                                                                      Rich
                               20       Heidelberg
                                                      ch
                                                     Ri




                               0
                                    0                           10                 20                    30
                                                                     Age (years)

                                          D'Arcy Hart P. Br Med Res Council Spec Ser 1932;No. 164:5-132
           Frequency of Large Tuberculin Skin Test Reactions Among
             High School Students, Washington County, USA, 1963
                             Education               Housing characteristics
                             of parents                Persons per room
                     25
Reactors per 1,000




                     20

                     15

                     10

                      5

                      0
                          12+ yr     < 12 yr             <1           1+
                               Kuemmerer JM, et al. Am Rev Respir Dis 1967;96:885-92
                    Frequency of Tuberculin Reaction >15mm, Among Males,
                        by Age, Selected African Countries, 1955 - 1960
                    1.0

                                                                                 Nigeria
                    0.8
Fraction reacting




                                                                                 Sierra Leone
                    0.6                                                          Ghana


                    0.4


                    0.2
                                                                                 Kenya

                    0.0
                          0   5      10      15      20              30               40
                                             Age group (years)

                                  Roelsgaard E, et al. Bull World Health Organ 1964;30:459-518
                                    Age-Specific Prevalence of Tuberculous Infection
                                       in Cross-Sectional Surveys in Switzerland
                              100                                                          1920
Prevalence of infection (%)


                                                                                           1953/54
                               80                                                          1967

                               60                                                          1978

                               40

                                                                                           1990
                               20
                                                                                           (modeled)

                                0
                                    0     10         20         30         40         50          60
                                                           Age (years)

                                           Zuberbühler JG. Thesis, University of Zürich, 1981:1-121
                                                     Hofer S. Thesis, University of Zürich 1982;1-45
                                                    Haefliger E. Prax Klin Pneumol 1982;36:335-64
                                                Haefliger E. Therapiewoche Schweiz 1989;5:855-66
                                           Rieder HL, et al. Schweiz Med Rundschau 1990;79:675-9
                              Prevalence of Tuberculous Infection by Age and Sex,
                                  Korea, 1965-95 - Analysis by Mixture Models
                    100       Females
Per cent infected
                     80

                     60                                  65
                                                    19
                     40                                            85
                                                    75        19
                                               19                            95
                     20                                                 19

                      0
                          0           5                  10             15        20    25          30

                    100       Males
Per cent infected




                     80

                     60                                  65
                                                    19
                                                                   85
                     40                                       19
                                                    75                       95
                                               19                       19
                     20

                      0
                          0           5                  10             15        20    25          30
                                                              Age (years)
                                          Neuenschwander BE, et al. Int J Tuberc Lung Dis 2000;4:719-29
                              Correlation Between Tuberculous Herds (1937 - 1939) and
                             Human Tuberculin Skin Test Reactors (1950 - 1952), Denmark

                             80
Per cent of human reactors




                             70
                             60
                             50
                             40
                             30
                                              Urban                                          Rural
                             20
                             10
                              0
                                  0      20    40     60   80      100   0      20   40    60        80   100
                                      Per cent tuberculous herds             Per cent tuberculous herds
                                                           Magnus K. Bull World Health Organ 1966;35:483-508
     Interval Between Ingestion of M. tuberculosis and Tuberculin
     Skin Test Conversion, 97 Children, Lübeck, Germany, 1930
                      25


                      20
Per cent converting




                      15


                      10


                       5


                       0
                           1   2   3   4     5    6   7    8    9   10   11   12   13
                                             Week first time reacting

                                           Böcker M. Arbeiten a d Reichsges Amt 1935;69:306-24
     Infection with
    M. tuberculosis

o Etiologic epidemiology
o Descriptive epidemiology
o Predictive epidemiology
Epidemiologic uses of tuberculin
      skin test surveys

   o    Predict course of the
        tuberculosis epidemic

   o    Estimate burden of
        tuberculosis
Algebraic Derivation of the Average Annual Risk of Infection from Prevalence


Notation:
b         Calendar year cohort was born
a         Age of cohort at calendar time survey was conducted
Pb+a         Prevalence of infection in cohort at time of survey
R            Average annual risk of infection between calendar time b and calendar
             time b+a

 At birth:            1 – Pb            =        1
 At age = 1:          1 – Pb+1          =        1 (1 – R)1
 At age = 2:          1 – Pb+2          =        1 (1 – R)1 (1 – R)2
 …
 At age = a:          1 – Pb+a          =        1 (1 – R)1 (1 – R)2 … (1 – R)a
                      1 – Pb+a          =        (1 – R)a
                      (1 – Pb+a)1/a     =        (1 – R)
                      R                 =        1 – (1 – Pb+a)1/a
                                 Difference between Calculated Annual Risk
                                  and Actual Annual Incidence of Infection
Risk or incidence of infection



                                  Inci
                                      den
                                            ce 1


                                 Calculated risk


                                              e   2
                                     Incidenc




                                 b                        b+(a/2)       b+a
                                                      Calendar time
                                Annual Risk of Tuberculous Infection in
                              Low Income Countries of Five WHO Regions
                        4
                                                         Afghanistan
Risk of infection (%)




                        3                                                              Africa
                                                                                       Americas
                                            Thailand
                                                            Lesotho           Korea    South-East Asia
                        2   Gambia
                                           Pakistan                                    Eastern Mediterranea
                                                                         Philippines

                                                China           Botswana               Western Pacific
                        1                                              Tanzania

                                               Samoa
                               Argentina        Malaysia       Algeria
                                  Libya                      Bahrain
                                                Kuwait
                        0
                             70                75                 80              85
                                                           Year
                                           Cauthen GM, et al. WHO Document 1988;WHO/TB/88.154:1-34
                                            Secular Trend in Annual Risk of Infection,
                                                 Selected European Countries
                                     Engl
                                            and                  Serbia
                             10                   a nd
                                                         Wale
                                                             s     Slo              Poland
                                                                         ven
Per cent risk (log scale)




                                                                               ia       No
                                    Slope reference:
                                    % decline / year                                      rw
                                                                                            ay
                              1                             0%
                                                            5%                                                Fr
                                                                                                                an
                                                                                                                   ce
                                                           10%                          Ne
                                                                                           t   he
                             0.1                                                                 rla
                                                           15%                                         nd
                                                                                                          s


                            0.01

                                   1900              1920           1940                       1960                  1980
                                                                  Calendar year
                      Waaler H, et al. Bull Int Union Tuberc 1975;50:5-61                        Sutherland I, et al. Tubercle 1983;64:241-253
                      Sutherland I, et al. Bull Int Union Tuberc 1971;45:75-114            Styblo K, et al. Bull Int Union Tuberc 1969;42:5-104
                      Lotte A, et al. Int J Epidemiol 1973;2:265-82                      Vynnycky E, et al. Int J Tuber Lung Dis 1997;1:389-96
                                        Age-specific Prevalence of Tuberculous Infection
                                                       in Western Europe
Per cent of total population   20



                                                                                        Not infected
                               15
                                                                                         Infected


                               10




                                5




                                0
                                    0     10    20    30     40    50     60    70    80     90     100
                                                              Age (years)
                                            Data courtesy: ten Dam HG, World Health Organization, 1990
                                      Annual Risk of Tuberculous Infection
                                      WHO Eastern Mediterranean Region
                         5
                                                                                       Algeria
                                                                                       Afghanistan
                         2                                                             Bahrain
Risk of infection (%)




                                                                                       Kuwait
                                                                                       Libya
     (log scale)




                         1                                                             Pakistan
                              Slope reference:                                         Syria
                              % decline / year
                        0.5
                                                     1%

                                                     5%
                        0.2
                                                    10%

                        0.1


                                    50                60            70           80
                                                            Year
                                                 Cauthen GM. WHO Document 1988;WHO/TB/88.154:1-34
                               Annual Risk of Tuberculous Infection
                               WHO African Region (Except Algeria)

                         5



                         2
Risk of infection (%)
     (log scale)




                         1
                              Slope reference:
                              % decline / year
                        0.5
                                                      1%

                                                      5%       Botswana        Gambia
                        0.2                                    Burundi         Lesotho
                                                      10%
                                                               Cameroon        Nigeria
                                                               Ethiopia        Tanzania
                        0.1


                              50                 60            70         80
                                                        Year
                                Cauthen GM. WHO Document 1988;WHO/TB/88.154:1-34
                                        Age-specific Prevalence of Tuberculous Infection
                                                in Tropical and Southern Africa
Per cent of total population   20



                                                                                        Not infected
                               15
                                                                                           Infected


                               10




                                5




                                0
                                    0     10    20    30     40    50    60    70     80       90     100
                                                              Age (years)
                                            Data courtesy: ten Dam HG, World Health Organization, 1990
                                   Annual Risk of Tuberculous Infection
                                     WHO South-East Asia Region

                         5



                         2
Risk of infection (%)
     (log scale)




                         1
                              Slope reference:
                              % decline / year
                        0.5
                                                 1%

                                                 5%           India
                        0.2                                   Indonesia
                                                 10%
                                                              Thailand

                        0.1


                                 50               60          70          80
                                                       Year
                                  Cauthen GM. WHO Document 1988;WHO/TB/88.154:1-34
                                      Age-specific Prevalence of Tuberculous Infection
                                    in South-East Asia and Oceania (Excl. AUS and NZ)
                               20
Per cent of total population


                                                                                     Not infected
                               15
                                                                                     Infected


                               10




                                5




                                0
                                    0   10   20    30     40    50    60     70    80    90     100
                                                           Age (years)
                                         Data courtesy: ten Dam HG, World Health Organization, 1990
   The Epidemiologic Situation of Tuberculosis and Economic Realities

                                 Annual risk of infection (%)        Health
   Country or area                                                  resource
                                Current level   Annual decline     availability

 I Industrialized                0.01 – 0.1          > 10           Excellent

II Middle-income Latin            0.5 – 1.5         5 – 10            Good
   America, West Asia, North
   Africa
III Middle-income East and        1.0 – 2.5          <5               Good
    Southeast Asia
IV Sub-Saharan Africa, Indian     1.0 – 2.5         0–3               Poor
   subcontinent

                                                    Kochi A. Tubercle 1991;72:1-6
           Which situation do you prefer?


Situation A

Initial capital           100 US$

Interest                  6%



Situation B

Initial capital           200 US$

Interest                  4%
                                           Compound Interest
Accumulated capital   1000

                          800
   (linear scale)

                          600                     Situation B

                          400                                            Situation A

                          200

                           0
                                0   10               20                  30            40
                                         Years accumulating interest
    Accumulated capital




                          800
        (log scale)




                                    Situation B
                          400
                                                                Situation A
                          200


                          100
                                0   10               20                  30            40
                                         Years accumulating interest
        Estimate Disease Burden from Risk of Infection?


It has been stated:
1% risk of infection                 <=>   50 incident sputum smear-
                                           positive cases / 100,000


Underlying assumptions:
o 10 transmissions / case / year
o 2 years of infectiousness / case


Thus:
100 cases / 100,000                       2,000 transmissions / 100,000
 50 cases / 100,000                       1% risk of infection
 Incidence, Point Prevalence, and Period Prevalence




      Year 1        Year 2          Year 3      Year 4

In year 3:
      Incidence:                         12 cases
      Point prevalence, March 15:        24 cases
      Period prevalence year 3:          288 person-months
 Incidence, Point Prevalence, and Period Prevalence




      Year 1        Year 2          Year 3      Year 4

In year 3:
      Incidence:                         12 cases
      Point prevalence, March 15:        6 cases
      Period prevalence year 3:          72 person-months
     Tuberculosis
o Etiologic epidemiology
o Descriptive epidemiology
o Predictive epidemiology
        Tuberculosis Case Rates Among Navy Recruits During
       Follow-up, According to Size of Initial Tuberculin Reaction
                    400
Cases per 100,000




                    300


                    200


                    100


                      0
                                0-5            6 - 11         >= 12
                                          Induration (mm)
                          Edwards LB, et al. Am Rev Respir Dis 1973;108:1334-9
  As elaborated previously, the probability of tuberculous infection
increases with increasing size of the tuberculin skin test reaction: the
specificity of the test is larger, the larger the reaction size.

  Therefore, the majority of recruits with very small reactions has no
tuberculous infection, and a large proportion of those with
intermediate reaction sizes reacts because of some non-specific
cross-reaction.

  As only a person with a tuberculous infection can develop
tuberculosis, the risk is highest in the group with the largest proportion
of infected persons among all reactors.

 This example illustrates that in the following it will be essential to be
certain that risk factors for developing tuberculosis are strictly limited
to persons who have a tuberculous infection.
                 Lifetime Risk of Pulmonary Tuberculosis, Adjusted
                   for Age, by Calendar Time, England and Wales
           20
                       All pulmonary

           15
Per cent




           10



            5          Smear-positive pulmonary


            0
                1900        1920        1940          1960          1980
                                            Year

                              Vynnycky E, et al. Am J Epidemiol 2000;152:247-63
                                                   Incidence of Tuberculosis Among Initial Reactors,
Average annual case rate (per 100,000)
                                                    by Age when Tuberculosis was First Diagnosed
                                         400



                                         300



                                         200



                                         100



                                           0
                                               0        5     10        15       20        25   30    35      40
                                                                             Age (years)
                                                                   Comstock GW, et al. Am J Epidemiol 1974;99:131-8
                                   Percentage of Eskimos with 5mm or more Induration
                                     to 5 TU PPD-S, by Sex and Age, Alaska, 1957
                             100
Per cent positive reactors




                             80


                             60
                                                          Female
                             40                           Male


                             20


                              0
                                   0     10         20       30       40       50       60
                                                         Age group (years)
                                              Comstock GW, et al. Am Rev Respir Dis 1967;95:935-43
                                 Male-to-Female Ratio of Tuberculous Infection
                                            Among Eskimos, 1957
                       1.2

                       1.1
Male-to-female ratio




                        1

                       0.9


                       0.8


                       0.7




                             0      10      20       30       40       50       60
                                                 Age group (years)
                                      Comstock GW, et al. Am Rev Respir Dis 1967;95:935-43
                                 Tuberculosis Cases Rates in Per Cent
                                 Among Eskimos, Alaska, 1958 - 1964
                        10


                        8
Case rate in per cent




                        6


                        4                          Female
                                                   Male
                        2


                        0
                             0   10         20       30        40       50       60
                                                 Age group (years)
                                      Comstock GW, et al. Am Rev Respir Dis 1967;95:935-43
                             Male-to-Female Ratios of Tuberculosis, Adjusted for
                             Tuberculous Infection Among Eskimos, 1958-1964
                       1.7
Male-to-female ratio



                       1.4

                       1.2

                        1


                       0.8

                       0.7



                             0     10       20       30       40       50       60
                                                 Age group (years)
                                     Comstock GW, et al. Am Rev Respir Dis 1967;95:935-43
     Incidence of Tuberculosis During Follow-up in the
Placebo Groups of Preventive Chemotherapy and BCG Trials
                           12
Cases per 1,000 reactors




                           10

                            8

                            6

                            4

                            2       Preventive                                BCG trial
                                    therapy trial
                            0
                                0                   5           10               15              20
                                                        Year of observation
                                                           Ferebee SH. Adv Tuberc Res 1969;17:28-106
                                                             D'Arcy Hart P, et al. Br Med J 1977;2:293-5
                               Incidence of Tuberculosis Among Tuberculin-Positive,
                                  HIV Infected Persons in Three Selected Studies
                              18
Incidence per 100 person-yr


                              16
                              14
                              12
                              10
                               8
                               6
                               4
                               2
                               0
                                      New York City      Spain            Italy
                                             Selwyn PA, et al. N Engl J Med 1989;320:545-50
                                                         Guelar A, et al. AIDS 1993;7:1345-9
                                                   Antonucci G, et al. JAMA 1995;274:143-8
                                  Risk of Tuberculosis Among HIV Infected, Tuberculin-
                                     Positive Persons, by CD4+ Lymphocyte Count
                             30
Tuberculosis incidence per
 100 person-yr (log scale)




                             10

                              5


                              2

                              1



                                          < 0.20            0.20 - 0.35                 > 0.35

                                               CD4+ lymphocyte count x 109 / L

                                                   Antonucci G, et al. J Am Med Assoc 1995;274:143-8
                                 Correlation Between Extent of HIV-Induced Immuno-
                                Suppression and Clinical Manifestation of Tuberculosis
                              500                    Pulmonary tuberculosis
Median CD4 cell count / mm3



                              400
                                                 Lymphatic, serous tuberculosis


                              300                             Tuberculous meningitis


                              200                                          Disseminated tuberculosis



                              100



                                0
                                                Duration of HIV infection

                                                De Cock KM, et al. J Am Med Assoc 1992;268:1581-7
                              Cumulative Percentage Tuberculosis-Free by Category
                                 of Silicosis Among Gold Miners, South Africa
                        100

                                                                                          Category 0
                         90
Per cent disease-free




                                                                                          Category 1
                         80                                                               Category 2


                         70


                         60
                                                                                          Category 3

                         50


                               0     1      2       3        4       5       6        7
                                                Year of observation
                                                Cowie RL. Am J Respir Crit Care Med 1994;150:1460-2
                   Relative Odds for Tuberculosis Among Male
                 Smokers, Aged 30 Years and Older, Great Britain

                                                                     95% upper
             6                                                       confidence interval


             5
Odds ratio




             4

             3

             2                                                       95% lower
                                                                     confidence interval

             1

                 0        10          20           30           40
                         Number of cigarettes smoked per day

                                    Edwards JH. Br J Prev Soc Med 1957;11:10-11
                     Age-Specific Relative Risks for Incident Tuberculosis
                           Among Diabetics, Korea, 1988 - 1990
                                                                All bacteriologically
                30
                                                                confirmed cases
                                                                Smear-positive cases
Relative risk




                10
 (log scale)




                3



                1


                        30        40        50            60         70
                                            Age (years)
                                         Kim SJ, et al. Tuber Lung Dis 1995;529-33
                                         Incidence of Infectious Tuberculosis by (Age-Adjusted) Body
                                           Mass Index, 15+ Years Old, by Sex, Norway,1963 - 1975
Cases per 1 million observation years



                                        200



                                        150

                                                             M
                                                                 al
                                                                   e
                                        100     Fe
                                                   m   ale

                                         50



                                          0
                                                 21          23             25        27        29        31

                                                                      Body mass index (kg/m2 )
                                                                            Tverdal A. Eur J Respir Dis 1986;69:355-62
   Predicted and Observed Mean Number of Tubercles Following
Inhalation of a Single Bacillus per Tubercle in Experimental Animals
                                  20
  Predicted number of tubercles



                                                                   15 hamsters

                                  15


                                                                 10 hamsters
                                  10           12 rats

                                                          15 hamsters
                                   5

                                           22 rats
                                   0
                                       0   5             10                15             20
                                           Observed number of tubercles
                                                              Ratcliffe HL. Am J Hyg 1952;55:36-48
Percentage of Secondary Cases Among Tuberculin-Positive
 Contacts, by Type of Source Case, Canada, 1966 - 1971
                                7   sm+/cul+
Per cent cases among reactors




                                6

                                5
                                          sm-/cul+
                                4                         sm+/cul+

                                3

                                2
                                               sm-/cul-         sm-/cul+
                                1
                                                                     sm-/cul-
                                0
                                       Intimate              Casual
                                        contact              contact
                                         Grzybowski S, et al. Bull Int Union Tuberc 1975;50:90-106
                                                   Rieder HL, et al. Epidemiol Rev 1989;11:79-98
                                      Example of Risk Differences in Individuals
                                      Following Infection with M. tuberculosis
                               1000                                                ??
Cases per 1,000 person-years




                                100
         (log scale)




                                 10




                                  1
                                       Long-standing    Recent        Super-     Underlying
                                         infection     infection     imposed HIV infection
                                                                   HIV infection
Correlation Between Tuberculous Herds (1937 - 1939) and Human Tuberculin
   Skin Test Reactors and Tuberculosis Morbidity (1950 - 1952), Denmark

                               80                                                                       8




                                                                         Morbidity per 1,000 reactors
  Per cent of human reactors




                               60                                                                       6


                               40                                                                       4


                               20                                                                       2


                                0                                                                       0
                                    0      20    40   60    80   100                                        0    10 20 30 40 50 60 70 80
                                        Per cent tuberculous herds                                              Per cent tuberculous herds
                                                                       Magnus K. Bull World Health Organ 1966;35:483-508
                               Estimated Morbidity Rates Following Infection
                                with M. bovis or M. tuberculosis, Denmark
                                      urban
Cases per 1,000 reactors



                           6
                                              rural
                           5

                           4

                           3

                           2

                           1                              urban
                                                                  rural
                           0
                                     Infection with      Infection with
                                    M. tuberculosis         M. bovis
                                 Magnus K. Bull World Health Organ 1966;35:483-508
                         Selected Risk Factors for Tuberculosis Given
                           that Tuberculous Infection has Occurred
Referent: Infection >7 yr past
          Infection <1 yr past
                HIV infection
              Fibrotic lesions
                     Silicosis
  Carcinoma of head or neck
                  Hemophilia
Immunosuppressive treatment
                Hemodialysis
                 Underweight
                     Diabetes
             Smoking, heavy
                Gastrectomy
           Jejunoileal bypass
               Infecting dose


                                 1   2         5      10     20        50   100
                                         Relative risk / odds (log scale)
           Public Health Importance of a Risk Factor

 There are two elements that make a risk factor to be important for public
health: the prevalence of the factor in the population and the strength of
the factor

  It is a bit of a judgment call, but we consider recent infection to be the
most important on this list. Everywhere in the world, whenever a new case
of infectious tuberculosis emerges, the diagnosis and commencement of
treatment are usually only made after that case has already infected one or
more other persons.

 Although jejunoileal bypass surgery is a very strong risk factor, its
application is limited to very few people with extreme obesity in very affluent
countries.

 HIV infection is certainly the most important risk factor in many countries
of sub-Saharan Africa, but it is not (yet) the case in India and China which
account together for about one third of the world population.
     Tuberculosis

o Etiologic epidemiology
o Descriptive epidemiology
o Predictive epidemiology
Improved Notification of Tuberculosis in Two London Hospitals
                          100
                                                        Chest medicine
                                                        Other medical specialty
Per cent cases notified




                          90
                                                        Surgical specialty
                                                        Other / unknown
                          80


                          70


                          60


                          50
                                1985-89         1992-93

                                          Brown JS, et al. Br Med J 1995;310:974
                                        Notification Rates of Sputum Smear-Positive
                                        Tuberculosis in Three Sub-Saharan Countries
                            80
Notifications per 100,000



                                                              Senegal
                            60

                                            Tanzania
                            40


                            20            Malawi



                            0
                                 1982    1984   1986   1988     1990    1992   1994   1996   1998   2000
                                                         Year of notification
                                 Reports of National Tuberculosis Programs, 1982-1999
                                    KNCV. Mission Report No.6 Tanzania, March 2000
          KNCV / Norwegian Health Association. Mission Report No. 32 Malawi, Jan 2000
                                     IUATLD Mission Report No. 28 Senegal, Aug 2000
                            Yearly Percentage Changes in Notifications
                            of Tuberculosis cases, Senegal, 1991 - 1999
      + 50
Relative change (%)




      + 20


                      0                                                            Smear-positive
                                                                                   All forms
                                                                                   Smear-negative
          - 20

          - 30                                                                     Extrapulmonary

                          1992   1993   1994    1995    1996   1997    1998    1999
                                           Year of notification

                                               IUATLD Progress Report Senegal No. 28, August 2000
                     Change in Notifications of Sputum Smear Positive
                  Tuberculosis over Previous Year in Malawi and Tanzania


                  60                                                  Malawi
Per cent change




                                                                      Tanzania
                  40


                  20


                  0




                   1984   1986   1988   1990   1992    1994    1996     1998     2000
                                        Year of notification

                                         Annual reports of National Tuberculosis Programs
                                                     Malawi and Senegal, 1999 and 2000
                                         Reported Tuberculosis Cases, Western Europe, 1974 - 1990
Notifications per 100,000 (log scale)


                                        30




                                        20




                                        15




                                                   1975             1980            1985            1990
                                                                  Year of notification
                                                   Raviglione MC, et al. Bull World Health Organ 1993;71:297-306
                                         Reported Tuberculosis Cases, Western Europe, 1974 - 1990
Notifications per 100,000 (log scale)

                                        80



                                        40



                                        20



                                        10



                                         5


                                                 75               80                85                90
                                                                   Year of notification
                                                  Raviglione MC, et al. Bull World Health Organ 1993;71:297-306
                           Changes in Tuberculosis Notifications in
                          Western Europe From One to the Next Year
                  + 40

                  + 20
Per cent change




                    0

                  - 20

                  - 40

                  - 60

                         1976   1978   1980     1982   1984    1986   1988   1990   1992
                                              Year of notification
                          Raviglione MC, et al. Bull World Health Organ 1993;71:297-306
                                    Tuberculosis Notification Rates Among Males,
                                  Cross-Sectional Observations, Finland 1954 -1994

                            500                    1954

                                                   1964
Notifications per 100,000




                            100                    1974
                            50
        (log scale)




                                                   1984
                            10
                             5
                                                  1994


                             1
                            0.5

                                  0          20               40           60              80
                                                          Age (years)
                                             Härö AS. Tuberc Respir Dis Yearbook 1998;24:1-151
                                      Tuberculosis Notification Rates Among Males,
                                          by Birth Cohort, Finland 1954 -1994

                            500                                               1902               1892
                                                                                                         1954
                                                                1932          1922        1912
Notifications per 100,000




                            100                           194
                                                                2
                             50               1952
        (log scale)




                                                                                                         1994


                             10                     2
                                              196
                              5



                              1        1972
                            0.5

                                  0             20                     40            60                 80
                                                                Age (years)
                                                    Härö AS. Tuberc Respir Dis Yearbook 1998;24:1-151
                                      Tuberculosis Notification Rates Among Females,
                                            by Birth Cohort, Finland 1954 -1994

                            500
                                                                      19
                                                                           22
                                                                                      1912           1892
Notifications per 100,000




                                                                                             1902
                                                           194
                            100                                  2
                                                                                193                          1954
                             50                     1952                              2
        (log scale)




                                                                                                             1994
                             10
                                                2
                              5           196

                                                1972
                              1
                            0.5

                                  0                  20                     40                  60          80
                                                                     Age (years)
                                                     Härö AS. Tuberc Respir Dis Yearbook 1998;24:1-151
                    Average Annual Incidence of Tuberculosis Among Women,
                     Cross-Sectionaly and by Birth Cohort, Upstate New York
                    100
                                                                                    1940-44
                                                     19




                                            19
                    30                                    20              19
Cases per 100,000




                                             30
                                                                19             00
                                     1940                            10
   (log scale)




                    10
                                                                                    1970-74
                      3
                              1950
                      1


                    0.3


                          0          20               40                   60         80
                                                  Age (years)
                                             Hinman AR, et al. Am J Epidemiol 1976;103:486-97
                                        Age-Specific Notification Rates of Sputum
                                       Smear-Positive Tuberculosis, Senegal, 1999
                                                  2.9        2.9
                                                                        2.5
Notifications per 100,000



                            200

                                                                              2,2
                            150                                                      3.0
                                                                                           Males

                            100             1.7


                             50                                                            Females
                                      0.9
                             0
                                  0         15    25    35         45    55    65
                                                         Age group (years)
                                                        IUATLD Progress Report No. 28, August 2000
                                  Notifications of Pulmonary Tuberculosis In Denmark
                                    (excluding Copenhagen), by Sex, 1921 - 1957
                            200          Females
Notifications per 100,000




                            100
        (log scale)




                                      Males


                             50


                             25




                                  1920           1930           1940           1950           1960
                                                                Year
                                         Groth-Petersen E, et al. Bull World Health Organ 1959;21:5-49
                            Male-to-Female Pulmonary Tuberculosis Case Rate Ratios
                                in Denmark (excluding Copenhagen), 1921 - 1957
                            1.3
Male-to-female rate ratio
       (log scale)




                             1



                            0.8

                            0.7




                                  1920       1930           1940           1950          1960

                                                            Year
                                         Groth-Petersen E, et al. Bull World Health Organ 1959;21:5-49
  Possible Epidemiologic Explanations for the
Change in the Tuberculosis Sex Ratio in Denmark
A change in risk of infection and an increase in the age of patients:

In the 1920ies, perhaps only a quarter escaped infection by age 20 years,
and sex-specific differences in the prevalence of infection at that age might
have been small. Therefore, the higher incidence rates among young
women (because of their larger disease risk given infection).

With the decrease in infection risk, fewer young adults are infected, and
sex-differences in infection may become more prominent. Males become
the higher infection prevalence group and start offsetting the inverse
disease risk ratio.

With declining risk and prevalence of infection, the age of tuberculosis
patients increases, further increasing the female-to-male disease ratio
(because men’s risk increases with age relative to females).
                             Incidence of Pulmonary Tuberculosis Among Married
                                  Men, by Residence, Denmark, 1960 - 1968
                                                                                  Capital
                        60
Incidence per 100,000




                        40
                                                                                   Rural
                                                                                  districts
                        20



                         0
                              20          40              60               80
                                                  Age (years)
                                        Christensen O. Scand J Respir Dis 1978(suppl):21-7
                                   Incidence Rates of New Cases of Tuberculosis by
                                     Socioeconomic Level, Upstate New York, 1973
                             120
Cases per 100,000 per year




                             100

                              80

                              60

                              40

                              20

                               0
                                      0     2       4       6       8      10      12
                                                  Socioeconomic level
                                          Hinman AR, et al. Am J Epidemiol 1976;103:486-97
                                      Standardized Tuberculosis Notification Rates by
                                      Deprivation Score, London Burroughs, UK, 1996
Standardized notification rate


                                 80
  (per 100,000 population)




                                 60


                                 40


                                 20


                                  0

                                       -20       0           20           40            60           80
                                                 Jarman underpriviledged area score
                                                     Bradsley M, et al. Health of Londoners Project, 1998
                                                               McKee M, et al. Lancet 2000;356:665-70
                                 Incidence of pulmonary tuberculosis Among Males
                              by Marital Status in Copenhagen, Denmark, 1960 - 1968


                        200
Incidence per 100,000




                                                                                   Divorced
                        150
                                                                                   Single
                        100                                                        Widowed

                                                                                   Married
                         50


                          0
                                 20          40              60               80
                                                     Age (years)
                                           Christensen O. Scand J Respir Dis 1978(suppl):21-7
                                Reported Tuberculosis Cases in the United States, 1953 - 1999
                              80000
Number of cases (log scale)




                              40000




                              20000




                                      1950    1960     1970        1980         1990         2000
                                                      Year of notification
                                                            Centers for Disease Control and Prevention
                                                        Reported Tuberculosis in the United States, 1999
    Tuberculosis Notification Rates in New York City, 1920 - 1999

                            200
Notifications per 100,000




                            100
        (log scale)




                             50




                             20



                                  1920     1940            1960           1980           2000
                                                   Year of notification
                                         New York City Department of Health, New York, 2000:25
                                        Reported Tuberculosis Cases Among Blacks
                                          and Whites, United States, 1980 - 1992
                           16000
Number of notified cases




                           13000
                                                            White
     (log scale)




                           10000



                                                         Black
                            8000


                                   80        82     84           86        88        90         92
                                                          Year of report
                                                    US Centers for Disease Control and Prevention
                                                   Tuberculosis Statistics in the United States, 1992
                         Tuberculosis Cases Rates Among Children Aged less
                         than Five Years, by Race, United States, 1981 - 1992
                    16
                                                              Black
Cases per 100,000




                     8



                     4

                                                             White

                     2


                             1982      1984       1986       1988       1990       1992
                                               Year of notification
                                                   Centers for Disease Control and Prevention
                                    Tuberculosis Statistics in the United States, 1991 and 1992
     Global Migration
o Foreign labor
o Asylum seekers, refugees
o Internally displaced persons
                               Foreign Born-to-Canada Born Risk Ratios
                                for Tuberculosis in Canada, 1970 - 1972
                         16    Philippines
                                                  China
                               India
                          8
Risk ratio (log scale)




                                                  Finland
                          4    Portugal
                                                  Yugoslavia
                               Greece
                          2                       Austria
                               Hungary

                          1    Canada
                                                  United Kingdom and Germany
                               United States
                                                  Benelux
                         0.5
                                   Enarson D, et al. Am Rev Respir Dis 1979;119:11-8
                               Age-Adjusted Rate Ratios for Tuberculosis among
                               Foreign-Born Persons, United States, 1986 - 1993
                                                    Other Asian countries
                         16    Sub-Saharan Africa
                                                    India
                               Mainland China
Risk ratio (log scale)




                          8                         Latin America

                                                    Middle East
                          4


                          2                         Former socialist economies

                          1                         Established market economies


                         0.5
                                      McKenna MT, et al. N Engl J Med 1995;332;1071-6
                                          Reported Tuberculosis Cases in Sweden,
                                             by Country of Origin, 1984 - 1998
                             600
Reported cases (log scale)


                                                            Born in Sweden



                             300



                                             Born abroad

                             150




                                   1982    1984   1986   1988   1990   1992   1994   1996   1998   2000
                                                           Year of notification
            Data courtesy: Romanus V. Swedish Institute for Infectious Disease Control, 1997
                                      Swedish Institute for Infectious Disease Control, 2000
               Proportion of Tuberculosis Cases Reported as
                 Foreign-Born, WHO Europe Region, 1997
      Slovakia
      Moldova
      Armenia
      Hungary
       Estonia
         Latvia
     Lithuania
Czech Republic
        Ireland
       Finland
       Croatia
          Malta
        France
       Norway
   Switzerland
      Sweden
     Denmark
          Israel


                   0       20        40         60           80          100
                                 Per cent foreign-born
                                                     EuroTB, September 1999
Notification Rates of Culture-Confirmed Tuberculosis Cases,
        by Country of Birth, Switzerland, 1995 - 1998
                                   Foreign-born
                             (without asylum seekers)
                    30
Cases per 100,000




                    20


                    10
                                           Swiss-born

                     0
                         0        20            40              60            80
                                            Age group (years)
                                                     Swiss Federal Office of Public Health
                                                           Bull SFOPH 2000;(No. 1):8-11
                                     Interval Between Entry or Re-Entry to Britain and
                                    Notification of Tuberculosis Among Asians in London
                           40
                           30                                                               No visit to Asia
                           20                                                              (time from entry)
                           10
                            0
Number of notified cases




                                1   2   3   4   5   6   7    8   9 10 11 12 13 14 15 16 17 18 19 20 21 22
                           40
                           30                                                                  Visit to Asia
                           20                                                               (time from entry)
                           10
                            0
                                1   2   3   4   5   6    7   8   9 10 11 12 13 14 15 16 17 18 19 20 21 22

                           40
                           30                                                                 Visit to Asia
                           20                           Years after entry / re -entry     (time from re-entry)
                           10
                            0
                                1   2   3   4   5   6    7   8   9 10 11 12 13 14 15 16 17 18 19 20 21 22


                                                                  McCarthy OR. Br J Dis Chest 1984;78:248-53
  High risk of tuberculosis following immigration

 Persons who never returned to their country of origin had
the highest risk immediately following immigration. This
suggests that most of disease in these persons was the
result of progression from recent infection acquired in the
country of origin.

 The distribution among those who did return shows an even
distribution of case emergence over time when analyzed
relative to the first entry. When arranged relative to re-entry,
the distribution resembles that among those who never
returned.

  This analysis strongly suggests the importance of recently
acquired infection and points to the differences in infection
risk in low and high incidence countries.
 Impact of HIV on the Epidemiology of Tuberculosis

Direct:   o   Reactivation of tuberculous infection acquired before
              HIV infection

          o   Progression of tuberculous infection acquired after
              HIV infection

Indirect: o   Transmission to the population not infected with HIV

                              Sutherland I. Br Comm Dis Rep 1990:10
    Nosocomial Tuberculosis Outbreak in a
          Hospital in Verona, Italy

         HIV pos
         patients                                    Personnel




0   20      40      60   80     100   120   140    160   180     200
           Days after occurrence of index case
                              di Perri G, et al. Lancet 1989;2:1502-4
              Factors Associated with Extrapulmonary Tuberculosis
                at AIDS Diagnosis, Selected European Countries
                  Italy
              Denmark
           Switzerland
       United Kingdom
             Germany
                France
                 Spain
              Portugal

 Homo- /bisexual men
    Injecting drug user
Hemophilia / transfusion
  Heterosexual contact

                  Male
                Female

                           0.5        1          2           4          8          16
                                       Adjusted odds ratio (log scale)

                                 European Centre for the Epidemiological Monitoring of AIDS
                                                             Quart Rep 1995;(No 46):37-48
                          Notifications of Tuberculosis, by Type of Disease,
                                        Tanzania, 1984 - 1999
                  25000                                                          Smear-positive

                  20000
Number of cases




                                                                                 Smear-negative
                  15000

                  10000                                                          Extrapulmonary


                   5000
                                                                                 Relapses
                      0
                            1985           1990                  1995          2000
                                          Year of notification

                                            Tanzania NTLP / IUATLD. Progress Report 1996;No. 36
                                                       Tanzania NTLP / KNCV. Report 2000;No. 7
                         Change in HIV-Seroprevalence among Tuberculosis
                            Patients in Tanzania, 1991-93 and 1994-98

                                60
Per cent infected (log scale)




                                50                                                    Female
                                                    Urban

                                40                                                    Male
                                                    Rural


                                30




                                20
                                     1991-93   1994-98               1991-93   1994-98

                                           Range N, et al. Int J Tuberc Lung Dis 2001;5:405-12
        Notification Rates of Sputum Smear-Positive Tuberculosis,
                by Age, Tanzania Mainland, 1984 and 1995
                            200                    1995
Notifications per 100,000




                            150


                            100
                                                            1984


                             50


                              0
                                  0   15      25    35    45     55    65
                                                     Age group (years)
                                           Tanzania NTLP / IUATLD. Progress Report 1996;No. 36
                                 Impact of HIV Infection on Tuberculosis Notifications
                                        in Chiang Rai, Thailand, 1985 - 1994
                           500
No. of cases (log scale)




                                                                                    All cases

                           400



                           300

                                                                                    HIV-neg cases


                           200


                                  85                     90                        95
                                                   Year of notification
                                                               Yanai H, et al. AIDS 1996;10:527-31
Tuberculosis among AIDS Patients, by Race / Ethnicity,
         Florida, United States, 1981 - 1986
   Per cent with tuberculosis   30



                                20



                                10



                                0
                                     White, non- Hispanic Black      Haitian
                                      Hispanic           American
                                       Rieder HL, et al. Arch Intern Med 1989;149:1268-73
Determinants for the Frequency of HIV-Associated Tuberculosis in a Community



              Total population

                                                     Prevalence and incidence of
                                                     infection with M. tuberculosis
                    Infected with
                                                     Prevalence and incidence
                    M. tuberculosis
                                                     of HIV infection

                                                     Overlap of the two respective
                                                     population segments



                                 Infected with HIV
             Proportion of Tuberculosis Cases Attributable to Primary Infection, Exogenous
             Reinfection, or Endogenous Reactivation, Among Males, by Age, Netherlands
                         100                                                                     Primary


Per cent of all cases
                          80

                                      15 to 19 years old
                          60


                          40


                          20
                                                                                                 Endogenous
                           0                                                                     Exogenous
                               1950         1955           1960          1965          1970

                         100
                                      60 to 64 years old
                                                                                                 Endogenous
 Per cent of all cases




                          80


                          60


                          40


                          20
                                                                                                 Exogenous
                                                                                                 Primary
                           0
                               1950         1955           1960          1965          1970
                                                            Year
                                                                   Sutherland I, et al. Tubercle 1982;63:255-68
    Estimated Number of Incident Tuberculosis
    Cases in the World, by WHO Region, 1997

                                               Southeast Asia
                               Western Pacific
                       Africa

        Eastern Mediterranean
       Europe
      Americas


0          1               2               3              4
            Number of cases (in million)
                 Dye C, et al. J Am Med Assoc 1999;282:677-86
     Tuberculosis

o Etiologic epidemiology
o Descriptive epidemiology
o Predictive epidemiology
                                      Projected Tuberculosis Incidence among Males
                                             in Finland, Years 2000 and 2010
                                400                                          Observed in
Cases per 100,000 (log scale)



                                                                             birth cohorts
                                                                                             Projected from
                                100                                                          birth cohorts

                                 30

                                 10


                                  3             2000

                                  1                         2010


                                       20     30       40          50   60          70         80
                                                        Age (years)
                                                        Härö AS. Tuberc Respir Dis Yearbook 1988;18:66
                      WHO Forecasts for Tuberculosis Incidence, 1990 - 2005
                3.0
                                                                            Africa
Million cases




                2.0




                1.0

                                                                            Industrialized
                                                                                countries
                0.0
                             1990       1995           2000          2005
                                      Dolin PJ, et al. WHO Document 1993;WHO/TB/93.173
 Tuberculosis Death

o Etiologic epidemiology
o Descriptive epidemiology
o Predictive epidemiology
Risk Factors for Death from
       Tuberculosis
o   Site and form of tuberculosis

o   Patient’s and doctor’s delay
                          Cumulative Case Fatality from Untreated Sputum
                             Smear-Positive Pulmonary Tuberculosis
                100
                                                    United Kingdom
                                                                                     Sweden
                 80
Per cent dead




                                                           Denmark
                 60


                 40


                 20


                  0
                      0      2    4      6      8     10      12     14   16    18      20
                                        Years after first admission
                                                         Thompson BC. Br Med J 1943;2:721
                                             Berg G. Acta Tuberc Scand 1939;(suppl):1-207
                                      Buhl K, et al. Bull World Health Organ 1967;37:907-25
                 Fate of Untreated Pulmonary Tuberculosis in Sanatorium
                 Patients, Long-Term Follow-Up, Barmelweid, Switzerland
                100

                                                        "Open" tuberculosis
                 80
Per cent dead




                 60


                 40
                                                      "Closed" tuberculosis
                 20


                  0
                      0   2   4     6      8     10     12      14     16     18
                                  Years after first admission
                                         Krebs W. Beitr Klin Tbk 1930;74:345-79
                            Tuberculosis Cases by Form and Site and
                          Expected Case Fatality from Untreated Disease
                    100
                                         Other
                                     Osteoarticular            Other
                    80
Per cent of cases




                                      Lymphatic

                    60                Pulmonary,
                                      smear-neg
                                                            Pulmonary,
                    40                                      smear-pos
                                      Pulmonary,
                    20                smear-pos


                     0
                                        Cases                 Fatal
                                                              cases
                                Data on cases: Rieder HL. Tubercle 1985;66:179-186
                                                    Data on fatality: Various sources
                                         Mortality-to-Morbidity Ratio of Pulmonary
                                     Tuberculosis, Denmark, 1935 and 1940 Combined
                                                                                               Female
Mortality-to-morbidity ratio



                               0.8
                                                                                               Male
                               0.6


                               0.4


                               0.2


                               0.0
                                     0   5   15 20 25       35     45      55      65
                                                        Age group (years)
                                              Groth-Petersen E, et al. Bull Wld Health Org 1959;21:5-49
                               Male-to-Female Case Fatality from Pulmonary
                             Tuberculosis, Denmark, 1935 and 1940 Combined
                       1.4
                       1.3
Male-to-female ratio




                       1.2
    (log scale)




                       1.1

                        1

                       0.9


                       0.8            Small
                                     numbers


                             0   5      15 20 25          35      45       55       65
                                                   Age group (years)
                                 Groth-Petersen E, et al. Bull World Health Organ 1959;21:5-39
 Age-Specific Frequency of Death from Pulmonary Tuberculosis
by Amount of Drugs Taken, North Arcot District, India, 1986 - 1988
                                                                    < 50%
                60
                                                                   50 - 79%
Per cent dead




                40



                20                                                 80 - 100%



                 0
                     15   25      45               65
                               Age group (years)

                                   Datta M, et al. Tuber Lung Dis 1993;74:180-6
                                       Proportion of Tuberculosis Cases Diagnosed at
                                         Death, by Age, United States, 1985 - 1988
                              20
Per cent diagnosed at death




                              15



                              10



                              5



                              0
                                   0       15   25   35    45    55     65    75    85       100
                                                          Age (years)
                                                             Rieder HL, et al. Chest 1991;100:678-81
                 Mycobacterium tuberculosis Bacteremia in
                Febrile Patients in Dar es Salaam, Tanzania
           25
                                                  HIV pos
           20


           15
Per cent




                                                   survived

           10


           5                                        died
                           HIV neg
           0
                      Archibald LK, et al. Clin Infect Dis 1998;26:290-6
  Tuberculosis Death

o Etiologic epidemiology
o Descriptive epidemiology
o Predictive epidemiology
                              Tuberculosis Mortality in Three European Cities,
                                Modeled From Available Data, 1750 - 1950
                     1000   London   Stockholm
Deaths per 100,000




                                                 Hamburg
                      800


                      600


                      400


                      200


                        0
                             1750        1800         1850          1900          1950
                                                      Year
                                       Grigg ERN. Am Rev Tuberc Pulm Dis 1958;78:151-72
                           Tuberculosis Mortality Rates in Germany, 1892 - 1940
                     250
Deaths per 100,000




                     200

                     150

                     100

                      50

                       0
                             1890       1900        1910          1920      1930         1940
                                                           Year
                       Redeker F. In: Handbuch der Tuberkulose (Hein J, et al, eds) 1958;1:473
Mortality from Respiratory Tuberculosis Among Males
 Cross-sectionally and by Birth Cohort, Switzerland

                                                                              18
                                                                                   46




                                                                                        18
                      300                                                    185




                                                                                        36
 Deaths per 100,000




                                                                                   6
                                                                   186
                                                                         6
                                                      187                                    1901
                                                               6
                      200                                                                    1921
                                                                   188
                                                     189                 6                   1941
                                           191             6
                      100                        6
                                                                                             1961
                                          1926       1906
                                                                                             1981
                        0
                            0   10   20     30        40            50        60        70    80
                                                 Age (years)
                                     Rieder HL, et al. Soz Präventivmed 1998;43:162-6
                     Mortality from Respiratory Tuberculosis Among Females,
                       Cross-Sectionally and by Birth Cohort, Switzerland


                     300
Deaths per 100,000




                                                                                    183
                                                                             1846         6
                     200                                       185                            1901
                                                                       6
                                                                            1866
                                                                                              1921
                                          19
                                            06
                                                   18
                                                                                              1941
                                                        96                 1876
                     100                                     188
                                                                   6
                                                                                              1961
                                         1926      1916
                                                                                              1981
                       0
                           0   10   20      30          40     50           60      70         80
                                                 Age (years)
                                         Rieder HL, et al. Soz Präventivmed 1998;43:162-6
      Estimated Number of Tuberculosis Deaths
         in the World, by WHO Region, 1997

                                             Southeast Asia
                       Western Pacific

                                  Africa
         Eastern Mediterranean
      Europe
      Americas


0.0       0.2          0.4         0.6          0.8         1.0
                Number of deaths (in million)
                   Dye C, et al. J Am Med Assoc 1999;282:677-86
  Tuberculosis Death

o Etiologic epidemiology
o Descriptive epidemiology
o Predictive epidemiology
           Number of Tuberculosis Cases and Deaths Preventable
            by Implementation of the DOTS Strategy, 1997 - 2020
                       20   Southeast
                              Asia                          Tuberculosis cases
                                    Sub-Saharan             Deaths from tuberculosis
or deaths (millions)
Preventable cases




                       15              Africa



                       10                    Western
                                             Pacific
                                                         Eastern
                        5                              Mediterranean
                                                                             Americas
                                                                    Europe
                        0

                                                  Dye C, et al. Lancet 1998;352:1886-91
“…The art of epidemiologic reasoning is to
draw sensible conclusions from imperfect
data…”

     Comstock GW. Am J Epidemiol 1990;131:206

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:16
posted:10/7/2012
language:Unknown
pages:178