| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Microbiology Department (Drs. Gallina, Troupioti, Sensalari, and Ms. Libanori), Division of General Thoracic Surgery (Dr. Rocco), Azienda Ospedaliera Eugenio Morelli, Sondalo, Italy.
Correspondence to: Massimo Gallina, MD, Laboratorio di Analisi Chimico-Cliniche e Microbiologia, Azienda Ospedaliera Eugenio Morelli, Via Zubiani 33, 23039 Sondalo (SO), Italy; e-mail: laborat{at}novanet.it
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Patients: Three hundred ten patients suspected of pulmonary mycobacterial infection or receiving antituberculous chemotherapy, accrued between 1996 and 1997.
Setting: Tertiary-care facility located in Northern Italy.
Design: We retrospectively compared the AMTDT results with the results of cultures. AMTDT results were also compared with those of acid-fast bacilli (AFB) staining of the same specimens. The study included 360 respiratory specimens from 310 patients collected between 1996 and 1997. In 1996, we used the initial version of AMTDT (50 µL of sediment); in 1997, we used the new version of AMTDT (450 µL of sediment).
Results: Compared with cultures, AMTDT and AFB staining had sensitivities of 87.2% and 68.4%, and specificities of 70.0% and 89.7%, respectively. When AMTDT and AFB staining were both positive, the sensitivity and specificity were 89.3% and 96.9%, respectively. When AMTDT and AFB staining were in disagreement, the sensitivity and specificity of AMTDT were 81.8% and 18.1%, respectively.
Conclusion: We conclude that when AMTDT is used to predict culture outcome, the results should be evaluated in conjunction with AFB staining results before making decisions about TB management.
Key Words: acid-fast bacilli staining • amplified mycobacterium tuberculosis direct test • tuberculosis
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Previously, the only rapid test for predicting the outcome of cultures has been the detection of acid-fast bacilli (AFB) using microscopy. More recently, however, the in vitro nucleic acid direct amplification tests (DATs), such as the Gen-Probe Amplified Mycobacterium Tuberculosis Direct Test (AMTDT; Gen-Probe; San Diego, CA) have been heralded as being more efficient. These tests initially aroused great enthusiasm, due to their higher sensitivity and specificity2 3 4 5 6 ; but the expectation that DATs would supplant AFB microscopy, accurately predict culture results, and provide an immediate definitive diagnosis was premature, and has been replaced with a more realistic view of the limitations and practical value of molecular diagnostics for TB.7
While culture results remain the "gold standard" for the presence of mycobacteria in a specimen during diagnosis and follow up,1 8 9 10 11 the clinical applications of AMTDTs must still be discussed and defined.7 8 9 12 With a view to optimizing DATs and sample number without compromising diagnostic accuracy, in the present article we evaluate the simultaneous use of AMTDT and AFB microscopy as routine laboratory procedures to predict culture results.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
50 cfu/slant) culture-positive specimens was 15.8%.
Decontamination Procedures
Respiratory specimens were digested and decontaminated by the
N-acetyl-L-cysteine sodium hydroxide method and neutralized with 67 mM
phosphate buffer (pH, 6.8). After centrifugation (4,000g for
15 min), the precipitate was resuspended in 2 mL of sterile phosphate
buffer. One aliquot of precipitate was used for solid medium culture
and a second for AFB staining; a third aliquot was stored for up to 1
week at - 20°C until the AMTDT could be performed.
Cultures
A 0.2-mL precipitate for each slant was used to inoculate two
International Union of TB Medium (IUTM) slants (prepared in our
laboratory13
) that were incubated at 37°C. The
precipitate was spread homogeneously over the entire surface of the
tubed medium using a serologic pipette. Slants were inspected for
growth weekly for 8 weeks. All isolates were confirmed to be MT complex
organisms using the Bactec
p-nitro-
-acetylamino-§-hydroxypropiophenone test (Becton
Dickenson Microbiology Systems; Sparks, MD) or the Accuprobe DNA
hybridization assay (Gen-Probe). For each positive culture, the number
of mycobacterial colonies and growth time were recorded, according by
standard methods.14
Cultures were categorized as high-
grade culture-positive (> 50 cfu), low-grade culture-positive ( 50
cfu), or as culture-negative. Negative cultures were recorded as such
at the end of the eighth week, but were checked weekly until the end of
12 weeks.
AFB Microscopy
One hundred microliters of precipitate were smeared on a slide
to cover an area of approximately 1 x 2 cm. Smears were stained with
auramine (TB Fluorescent Stain Kit M; Difco Laboratories; Detroit,
MI) and observed at 200x magnification. Semiquantitative
reporting of smear results were graded following the methods of Kent
and Kubica.14
AMTDT
AMTDT has been used routinely in this laboratory since November
1995, when requested by our Division of Phtisiology, following the
protocols as supplied by the manufacturer. In 1996, we used the
original version (50 µL of processed sediment: AMTDT 1). In 1997, we
used the new protocol (450 µL of sediment: AMTDT 2).15
All runs of AMTDT 1 included positive and negative controls for both
amplification and hybridization; for AMTDT 2, only positive and
negative controls of amplification were included. Results were read
with a Berthold luminometer (LUMAT LB 9501; Laboratorium Prof. Berthold
GmbH & Co.KG; Bad Wildbad, Germany). As specified by the
Gen-Probe protocol for AMTDT 1, samples with > 500,000 relative light
units (RLU) were considered positive; samples with < 30,000 RLU were
recorded as negative, and samples with values between 30,000 and
500,000 RLU were recorded as undetermined. For AMTDT 2, however,
samples with > 30,000 RLU were recorded as positive, whereas samples
with values of < 30,000 RLU were considered negative.
For each specimen, results from an AMTDT were compared with those of AFB staining and the number of colony-forming units of MT complex obtained on the IUTM slants.
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A total of 117 cultures were recovered from 360 specimens; among them,
60 produced > 50 colonies on the slant, and 57 produced < 50
colonies. The fraction of specimens in the three categories was not
significantly different between the two study years
(
2 test, p = 0.57).
The results of IUTM cultures, AFB microscopy, and AMTDT from both study
years are summarized in Tables 1
2
3
. In each category in Tables 1
,2
, for which there were two or more
samples in both years, a 2 test was performed.
No significant differences were observed for any category for the two
study years; therefore, data from both years were pooled.
|
|
|
2 test, p < 0.001).
Of 60 high-grade culture-positive samples (> 50 cfu), 7 AFB smears
and 3 AMTDTs yielded false-negatives. Of 57 low-grade culture-positive
samples ( 50 cfu), 19 were AFB smear negative, 11 were AFB
smear ± (doubtful), and 12 were AMTDT negative. Overall, for the
243 culture-negative samples, AFB smears yielded 10 false-positives
(4.1%), whereas AMTDT yielded 73 false-positives (30.0%).
AMTDT was positive in 102 of 117 total positive cultures (sensitivity, 87.2%), in 57 of 60 high-grade positive cultures (sensitivity, 95.0%), and in 45 of 57 low-grade positive cultures (sensitivity, 78.9%). AMTDT was negative for 170 of 243 negative cultures (specificity, 70.0%).
AFB smear was positive in 80 of 117 total positive cultures (sensitivity, 68.4%), in 53 of 60 high-grade positive cultures (sensitivity, 88.3%), and in 27 of 57 low-grade positive cultures (sensitivity, 47.4%). AFB smears were negative for 218 of 243 negative cultures (specificity, 89.7%).
Not considering results in which AMTDT and AFB smear were contradictory or AFB smears were ± (doubtful), AMTDT and AFB smears were both positive for 75 of 84 total positive cultures (sensitivity, 89.3%), in 51 of 52 high-grade positive cultures (sensitivity, 98.1%), and in 24 of 32 low-grade positive cultures (sensitivity, 75.0%). The specificity was 96.8% (155 of 160). Both these tests were negative for 9 of 117 total positive cultures (7.7% false-negatives), and in only 1 of 60 high-grade positive cultures (1.7% false-negatives). For 5 of 243 culture-negative specimens, AMTDT and AFB smear were both positive (2.1% false-positives).
Of 116 specimens for which AMTDT and AFB smear results were contradictory, or AFB smear results were ± (doubtful), AMTDT was positive in 27 of 33 total positive cultures (sensitivity, 81.8%), and in 6 of 8 high-grade positive cultures (sensitivity, 75.1%). AMTDT was negative in 15 of 83 negative cultures (specificity, 18.1%).
Table 4 summarizes various predictive values of AMTDT with different AFB smear results. Rates of positivity on culture range from 92.4% for ++ to ++++ AFB-positive smears to 10.6% for AFB-negative smears. The maximum positive predictive value (PPV) was obtained when AFB smears were positive ++ or higher. A PPV of 88.9 was obtained for AFB-positive + smears. Elevated negative predictive values (NPVs) were obtained for AFB-± (doubtful) smears and AFB-negative specimens (90.9% and 94.5%, respectively).
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In contrast to Gamboa et al,15
we did not find differences
in sensitivity between the two protocols of AMTDT provided by Gen-Probe
(AMTDT 1, 85% [34 of 40]; AMTDT 2, 88.3% [68 of 77];
2 test, p = 0.61). Nonetheless, an advantage
of AMTDT 2 is that it never yielded undetermined results, while 19.6%
(34 of 175) of samples in 1996 were of this type.
Our study shows that when AMTDT is applied to a patient population like
the one described, it has questionable specificity and limited
sensitivity in predicting culture results. Previous studies have shown
that the sensitivity of AMTDT depends on the number of bacilli in the
specimen16
3
7
; likewise, we found in our study that AMTDT
was more likely to detect MT in specimens with > 50 cfu/slant. In
fact, the sensitivity of AMTDT increases from 78.9% in specimens with
50 cfu/slant to 95% in specimens with > 50 cfu/slant. The
sensitivity of AMTDT is also influenced by the number of specimens
containing inhibitors. Although inhibitors are relatively rare, they
are also unpredictable.12
In this study, only two
specimens were AMTDT negative when the respective AFB-positive smear
was +++ and the culture had > 50 cfu/slant, presumably because the
specimens contained inhibitors. However, in clinical specimens, a low
degree of positivity, sample inhomogeneity, inhibitors, and/or effects
of sample processing may decrease the sensitivity of the
test.1
17
For these reasons, there is a need to improve
sample collection and processing. In this setting, a test for
inhibitors needs to be performed to better assist in determining the
value of a negative AMTDT result.
Our results confirm that AMTDT can detect inviable organisms, as noted previously.7 16 In the present patient population, there were many patients undergoing antimycobacterial treatment; consequently, the low specificity of the test indicates that AMTDT is not particularly suitable for the rapid assessment of treatment efficacy.16
In this mixed patient population, and without considering NTM, AFB microscopy showed a higher specificity (89.7%) in predicting culture results than AMTDT, but a lower sensitivity.
When AMTDT and AFB smear results are concordant, the high specificity of AFB microscopy decreases the overall fraction of false-positives. In particular, when AFB smears are positive ++ to ++++, and AMTDT is positive, PPV is also high (95.2%). In our opinion, the clinician could consider this result a sufficient guarantee of correct sampling and can expect, in most cases, to obtain a positive culture result. She/he could confidently use this result to make management decisions, and, where required, the culture could most probably be used for drug sensitivity tests. In addition, this result should persuade the clinician to limit the number of successive samples to be analyzed, as suggested by several authors,18 19 to as few as two, thereby avoiding an inefficient use of laboratory resources.
The high specificity obtained when AMTDT results were in agreement with the AFB smear has also been noted by previous authors.1 10 However, we found that when a positive AMTDT result is associated with an AFB-positive + smear, the possibility of a negative culture rises to about 10%. Samples with such a combination of results may have low numbers of MT and, consequently, the possibility of a negative culture (or a positive one only after prolonged incubation) is high, and it would be wise for the clinician to send further samples to the laboratory.
A negative AFB smear associated with a negative AMTDT has a high NPV (94.5%) in predicting the culture result. This combination of results indicates that the possibility of a positive culture is very low, and in any case, a longer growth time is expected.
When the results obtained from AMTDT and AFB smear are in disagreement, the sensitivity of AMTDT is still acceptable (81.8%), but its specificity (18.1%) means the predictive value is almost nil. In this case, the clinician has no choice but to send better quality specimens for reevaluation. However, if the AMTDT is negative in the presence of an AFB-positive ++++ smear, the presence of inhibitors or NTM can be suspected.
If the AFB smear is negative and AMTDT is positive, the probability of a negative culture increases. In this case, the clinician should verify if an antimycobacterial therapy is or has been administrated to the patient. In fact, AMTDT can remain positive many months after the initiation or completion of therapy.8 For this reason, one author12 has proposed that AMTDT only be used for untreated patients.
In conclusion, when AMTDT is requested to predict the outcome of a culture under routine laboratory conditions, we would recommend that this test always be used in association with the semiquantitative AFB smear to be sure that the specimen gives useful information for the management of TB.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
This work was performed in the Hospital Azienda Ospedaliera Eugenio Morelli, Sondalo, Italy.
Received for publication June 23, 1999. Accepted for publication March 20, 2000.
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
