Improving MGIT PZA Susceptibility Testing –A Multi-laboratory Evaluation

of Alternative Inoculum Dilutions

Glenn P. Morlock

Frances Tyrrell

James Posey

National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention

Division of Tuberculosis Elimination

10th National Conference on Laboratory Aspects of Tuberculosis

April 18, 2017

Major Factor Affecting PZA DST

Inoculum size Large inoculum reduces PZA activity

• Much more pronounced than other drugs

• Increase from 106 to 107 cells/ml

o PZA became inactive

• 107 –108 cells/ml

o pH increase of 1 & 1.5 units, respectively

• 106 cells / ml (commonly used inoculum)

o pH increase of 0.3 units

Explanation for pH increase

• Ammonia produced by deamination of PZA

PZA False Resistance

Increasingly common Concurrent with transition from BACTEC 460 to MGIT

• Co-incidence?

Bactec 460 Minimal (no supplement) media

Radiometric detection

• 14C-labeled CO2

Mycobacteria Growth Indicator Tube (MGIT) Enriched (supplemented) media

• Fluorometric detection

o Oxygen depletion from bacterial respiration

PZA Susceptibility Testing in MGIT:Problem Summarized

Very temperamental Inoculum too large

• Media is neutralized

o False resistance

Inoculum density too low

• Test will terminate prior to a result

Inoculum density just right

• Accurate result

Study Hypotheses

1. Over-inoculation causes: False resistance

Poor reproducibility

2. Lower inoculum density will: Mitigate false resistance rate

Improve reproducibility

3. Inoculum effect not universal Some strains highly responsive

Others recalcitrant

Study Design

3 inoculum schemes Standard protocol (“BD”)

2 reduced inoculum protocols (“A” and “C”)

10 laboratories 9 extramural

CDC LB/ART

20 isolates 15 unique strains

5 duplicates

Standard Inoculation Protocol“BD” Method

Seed Tube

Inoculum

370C

1 to 10dilution

nodilution

370C

370C

Control TubeNo PZA

µg/ml

PZA

100

Growth = Resistant

No Growth = Susceptible

Day1 or 2

If Day3, 4, or 5

1:5 Dilution

Experimental Protocol “A”

Seed Tube

Inoculum

370C

1 to 25dilution

1 to 2.5dilution

370C

370C

Control TubeNo PZA

µg/ml

PZA

100

Growth = Resistant

No Growth = Susceptible

Day1 or 2

Experimental Protocol “C”

Seed Tube

Inoculum

370C

1 to 50dilution

1 to 5dilution

370C

370C

Control TubeNo PZA

µg/ml

PZA

100

Growth = Resistant

No Growth = Susceptible

Day1 or 2

Selection Criteria for Study Strains

Phenotypic Minimum Inhibitory Concentration (MIC) results from previous

studies

Genotypic pncA sequence

• Sanger

• Whole Genome

Data Collected

Quantity Per participant

• 3 inoculum densities x 20 isolates = 60 tests

o In duplicate = 120 tests

In Total

• 1200 tests

Type Categorical

• Resistant

• Susceptible

Study Strains

Characteristics of 20 M. tuberculosis strains.

Isolate No. (duplicate)

Susceptibility to PZA (100 µg/ml) pncA mutationOther Drug Resistance

(multiple tests) Nucleotide Amino

Consistently Predominately no. acid

1 Susceptible — C 195 → T Ser65Ser CIP

4 Resistant — A(-11) → G NA RIF, CIP

7 Susceptible — None NA none

8 Susceptible — None NA none

11 (23) Susceptible — None NA none

12 (24) — Susceptible None NA none

13 (25) — Susceptible None NA none

14 Resistant — T 37 → C Phe13Leu INH

16 (26) — Susceptible None NA FQ

17 — Resistant A 139 → G Thr47Ala INH, RIF, AMK, EMB

18 Susceptible — C 509 → T Ala170Val none

19 Inconclusive Inconclusive C 244 → T His82Tyr none

20 Susceptible — None NA none

21 (27) Inconclusive Inconclusive G 82 → A Ala28Thr INH, RIF, EMB

22 Susceptible — A 110 → T Glu37Val none

“Predominately” PZA Susceptible with Wild Type PncA

“The Likely Suspects”

Note• Appreciably effected by inoculum dilution

Susceptible or Resistant by MethodIsolate “BD” “A” “C”

S R S R S R12 2 17 13 7 16 424 3 17 11 9 15 513 14 6 19 1 18 125 14 6 20 0 18 016 10 10 16 3 19 126 14 6 17 3 19 1

“Predominately”PZA Resistant with pncA Mutation

“The Trouble Maker”

MIC near critical concentration

Thr 47 Ala pncA mutation

• known to give discrepant PZA results

• Also INH and RIF resistant

• Consideration

• Strains borderline resistant at critical concentration

• Likely to be impacted by inoculum density

• Isolate 17 supports this contention

Isolate

Susceptible or Resistant by Method

BD A C S R S R S R

17 2 18 10 10 8 9

Algorithmically Assigned Predicted Results

Note Inconclusive Isolates Excluded

True

Susceptible

82%

18%

True

Resistant

Predicted Results vs Study Results I

True

Susceptible

True

Resistant

82%

18%

Predicted

61.1%17.1%

21.1%

0.7% 0.7%

False

Resistant

False

Susceptible

No

result

“BD” Method

Predicted Results vs Study Results II

82%

18%

True

Susceptible

True

Resistant

Predicted

False

Resistant

False

Susceptible

No

result

“A” Method

76.4%

14.2%

5.7%

3.7% 3.3%

Predicted Results vs Study Results III

82%

18%

True

Susceptible

True

Resistant

Predicted

False

Resistant

False

Susceptible

No

result

“C” Method

79.2%

14.5%

2.8%

3.1%

5.6%

Study Results Comparison I

61.1%17.1%

21.1%

0.7% 0.7%

True

Susceptible

True

Resistant

False

Resistant

False

Susceptible

No

result

“BD” Method “A” Method

76.4%

14.2%

5.7%

3.7% 3.3%

Study Results Comparison II

True

Susceptible

True

Resistant

False

Resistant

False

Susceptible

No

result

“BD” Method “C” Method

79.2%

14.5%

2.8%

3.1%

5.6%

61.1%17.1%

21.1%

0.7% 0.7%

Study Results Comparison III

True

Susceptible

True

Resistant

False

Resistant

False

Susceptible

No

result

“C” Method

79.2%

14.5%

2.8%

3.1%

5.6%

“A” Method

76.4%

14.2%

5.7%

3.7% 3.3%

Data AnalysisPrecision and Accuracy

Accuracy Comparison

0

20

40

60

80

100

78.3

90.5

94.1

21.7

9.5 5.9

Accurate Inaccurate

“BD” Method “A” Method “C” Method

Precision Comparison

0

20

40

60

80

100

72.2

15.2 12.6

1.3

87.6

6.9 5.5 5.2

93.5

2.9 3.68.5

BD method A method C method

Precise &

Accurate

Precise &

Inaccurate

Not

Precise < 2 Results

Conclusions I

Inoculum size critical Too high

• Can produce false-resistant result

Just Right

• Accurate result

Too little

• Accurate Result

o Delayed

• No result

o Test terminated

Ideal Inoculum a difficult goal Dilution not universal solution

• Can be detrimental

o Very delayed or no result

MGIT PZA Test = PZA mono-resistantWhat do you do?

Consider False Resistance

No

Be Suspicious

PZAMono-resistance

is rare

PZA mono-resistant

M. bovis?Accurate

PZA ResultYes

M. Bovis is naturallyresistant to PZA

MGIT PZA Test = PZA mono-resistantConsiderations

Consider False Resistance

Possibility ofOver Inoculation

ActionPathways

Within Protocol?

No Validationrequired

MGIT PZA Test = PZA mono-resistantConsiderations

Yes

Within Protocol?

• Repeat Test• Use “younger” seed tube1

• Day 1 vs. day 2• Day 3 vs. day 4 or 5

• Record hours past positive• Avoid clumps

• Allow time to settle• Aspirate from top

1. Association of Public Health Laboratories, Issues in MTBC Drug Susceptibility Testing: PZA. February, 2016.

MGIT PZA Test = PZA mono-resistantConsiderations

Consider False Resistance

Possibility ofOver Inoculation

ActionPathways

Within Protocol?

No Validationrequired

Outside Protocol?

Validationrequired

MGIT PZA Test = PZA mono-resistantConsiderations

Yes

Within Protocol?

Outside Protocol?

Yes

• Repeat Test• Use “younger” seed tube1

• Day 1 vs. day 2• Day 3 vs. day 4 or 5

• Record hours past positive• Avoid clumps

• Allow time to settle• Aspirate from top

• Repeat Test• Use reduced inoculum• Use smaller volume• Use standardized inoculum

• Use Modified Test• Reduced inoculum

1. Association of Public Health Laboratories, Issues in MTBC Drug Susceptibility Testing: PZA. February, 2016.

Acknowledgements

Co-investigators

Frances Tyrrell

James Posey

Enablement

Angela Starks

Tracy Dalton

David Temporado

Participating Laboratories

Whole Genome Sequencing

Paige Chopra

Lauren Cowen

Intangible Support

Entire Laboratory Branch

The Devil is in the Seed Tubesand When to Use Them

BD Protocol• “Day 0” = day seed tube positive

• Use on “day 1 or 2”

• Theoretically

o Positive @ 23:00

o May be used from hours 1 to 49 past positive

o Positive @ 01:00

o May be used from hours 23 to 71 past positive

o Using 18 hour / 1 generation

o 0 to 4 generations difference

BD Protocol May Be Too Imprecise?

Too large a time window• Considering temperamentality of test system

o Differing growth rate between isolates

Is there a better measurement?• Smaller time window

o May be constrained by work processes

• Growth Units value

o Not linear but may still be better

• Calculation using both factors

o May compensate for differences in growth rates

Algorithm for AssigningPredicted Categorical Result:

Susceptible, Resistant or Inconclusive?

Yes

No

Yes

Yes

InconclusiveN = 3

ResistantN = 2

YesNo

PZA MICPredominately

≤ 100 µg/ml?

PZA MICConsistently> 100 µg/ml

?

PZA MICPredominately

> 100 µg/ml?

ResistantN = 1

SusceptibleN = 3

No

No

No

Yes

pncAMutation

?

SusceptibleN = 10

Non-Synonymousor regulatory

?

SusceptibleN = 1