Positive NAAT test results for Neisseria gonorrhoeae do not require routine confirmatory testing

1
Positive NAAT test results for Neisseria
gonorrhoeae do not require routine confirmatory
testing
Matthew R. Golden MD, MPH Center for AIDS STD,
University of WA Public Health - Seattle King
County
2
Objectives

• To demonstrate that NAAT for gonorrhea do not
  require universal confirmatory testing when used
  in low prevalence environments.
• Issues I will not argue
• Should low prevalence populations be screened
• Selective screening currently do not exist
• PCR can be used in very low prevalence
• populations without confirmatory testing

3
Problems with an argument based on PPV

• Routine confirmatory testing is impractical
• Need for routine testing reflects a hyperbolic
  sense of risk, and excessive attention to
  relative rather than absolute risk
• Perceived need is based on notion that true test
  performance in a very low prevalence population
  cant be known. Its cynical.
• Maybe some tests are OK?

4
Approaches to confirmatory testing
5
Positive predictive value is primarily a function
of prevalence specificity
Positive Predictive Value ()
1
Prevalence
6
The absolute risk number of false positives
varies little with prevalence
false positives per 10,000 98 95 91
Absolute risk false positive 0.98/100 0.95/100 0.
91/100
Prevalence 1 4 8
PPV 49 80 89
Assumes 95 sensitive test and 99 specific
7
Are NAAT tests good enough?

• How good do they need to be?
• PPV gt90 in a very low prevalence
• population (i.e. 0.5)

8
Size, Specificity and Prevalence of Larges NAAT
Studies
Specificity estimates are imprecise. Recent
studies have avoided discrepant analysis,
probably erring on the side of underestimating
specificity
9
Number of specimens needed to define PPV gt90 in
a 0.5 prevalent population
True Specificity ()
10
Number of specimens needed to define PPV gt90 in
0.5 prevalent population
tested confirmatory testing design
Specificity ()
11
PPV Aptima for N. gonorrhoeae in a low prevalence
population
4 Labs in WA state test 59,664 specimens 280
(0.5) GC
265 Tested using alternative set rRNA primers
258 GC PPV 97.4 (95 CI 95.1-98.8)

• 0/194 negative specimens tested positive for N.
  gonorrhoeae

12
Limitations to Confirmatory Testing

• Using a NAAT for confirmatory testing assumes
  that the confirmatory NAAT is specific. If an
  organism other than N. gonorrhoeae has nucleic
  acid amplification sequences amplified by both
  tests, then a confirmatory NAAT results will not
  be valid.
• Routine confirmatory testing doesnt solve this
  problem unless all the patients come back for
  culture.
• This is not practical and culture is somewhat
  less sensitive than NAATs, so interpreting
  results will still be difficult.
• There is still a need for judgment in ordering
  and interpreting tests

13
Conclusions

• Confirmatory testing is not indicated for all
  specimens that test positive for N. gonorrhoeae
  using NAATs in low prevalence populations
• Confirmatory testing is not practical
• Perceived need based on low threshold for
  absolute risk of false positive
• Assumes test performance cannot be defined
• Need for confirmation depends on the test
• Aptima performs well and that confirmatory
  testing is not needed.
• PCR does not perform adequately and confirmatory
  testing is
• needed.
• Data are inadequate to assess the need for
  confirmatory
• testing with SDA Probetec some data low
  positive SDA may
• be more likely to be false positive