Indian J Med Res 123, June 2006, pp 807-814

Association of conversion & cure with initial smear grading

among new smear positive pulmonary tuberculosis patients
treated with Category I regimen
P.G. Gopi, V. Chandrasekaran, R. Subramani, T. Santha, A. Thomas, N. Selvakumar & P.R. Narayanan

Tuberculosis Research Centre (ICMR), Chennai, India

Received April 27, 2005

Background & objective: Early diagnosis of tuberculosis (TB) is important for initiating treatment
to gain cure. The present investigation was undertaken to study the association of conversion
and cure with initial smear grading among pulmonary tuberculosis (TB) patients registered in
a directly observed treatment – short course (DOTS) programme in Tiruvallur district, south
India.

Methods: All new smear positive cases registered from May, 1999 to December, 2002 were analysed
for conversion and cure related to initial smear grading.

Results: Of the 1463 patients, 1206(82.4%) were converted at the end of the intensive phase and
1109 (75.8%) were declared ‘cured’ after the completion of treatment. The cure rate decreased
as the initial smear grading increased and the decrease in trend was statistically significant
(P=0.01). Similarly, a significant decrease in conversion rate was also observed with increase in
initial smear grading (P<0.001). In multivariate analysis, lower cure rate was significantly
associated with patient’s age (AOR=1.5, 95% CI=1.1-2.1), alcoholism (AOR=1.7, 95% CI 1.2-
2.4) and conversion at the end of intensive phase (AOR=3.5, 95% CI= 2.6-4.8).

Interpretation & conclusion: Cure and conversion rates were linearly associated with initial
smear grading. High default and death rates were responsible for low cure and conversion. The
proportion of patients who required extension of treatment and those who had an unfavourable
treatment outcome were significantly higher among patients with a 3+ initial smear grading.
This reiterates the need to pay more attention in motivating these patients to return to regular
treatment and sustained commitment in the control of tuberculosis. There is a need to extend
the treatment for one more month in the intensive phase of treatment.

Key words Cure - default - DOTS - smear grading - tuberculosis

The Revised National Tuberculosis Control the most cost-effective strategy and developed
Programme (RNTCP) based on the World Health based on scientific evidence, is to cure at least 85
Organization’s DOTS (directly observed treatment- per cent new smear-positive cases and to detect
short course) strategy was introduced in India in at least 70 per cent of new smear-positive cases
1993. The goal of DOTS, which has emerged as in the population. Patients diagnosed with
807
808 INDIAN J MED RES, JUNE 2006

tuberculosis are treated under DOTS in accordance of a case as per RNTCP guidelines. The community
with the RNTCP guidelines 1 . Diagnosis of survey was carried out to assess the epidemiological
pulmonary tuberculosis is based on three sputum impact of DOTS implemented in this area. In
examination by smear microscopy in accordance community survey, case detection is early compared
with the guidelines 2,3. Early diagnosis of TB and to passive case detection at health facility. So, cases
initiating treatment under DOTS would not only detected in community would have reported
enable the patients to get cured but also reduce voluntarily at health facilities for diagnosis, and
the transmission of infection and disease to others. cases detected by these methods can be merged.
The anti- tuberculosis regimens used for category
In RNTCP, the sputum smears are graded and I, II, and III patients were 2H 3 R 3 Z 3 E 3 /4H 3 R 3 ,
reported based on the bacillary load. The present 2H 3 R 3 Z 3 E 3 S 3 /1H 3 R 3 Z 3 E 3 /5H 3 R 3 E 3 and 2H 3 R 3 Z 3 /
study was carried out on new smear-positive 4H 3R 3, respectively. (H= isoniazid; R= rifampicin;
pulmonary tuberculosis cases treated with category Z= pyrazinamide; E= ethambutol; S= streptomycin.
I regimen under RNTCP in Tiruvallur district, south Numbers before the letters indicate the duration
India, to find the association of conversion and cure of the treatment phase in months and numbers in
related to smear grading at the start of treatment, subscript indicate the number of times the drug is
and to examine other factors like sex, age, habit given each week). Intensive phase (IP) treatment
on smoking and alcoholism, patient delay, for category I and II patients was extended by
conversation rate, smear grade on admission etc. another month if the sputum smear remained positive
influencing the patient’s treatment outcome ‘cure’ at the end of IP. The treatment was under
at the end of treatment. observation for every dose in the IP and at least
the first of the three doses of every week during
Material & Methods continuation phase (CP). All category I patients
registered from May, 1999 to December, 2002
Study population, diagnostic algorithm and
formed the study population.
treatment regimens: The DOTS strategy was
implemented in a rural population of 580,000 in Procedure for sputum microscopy: When patients
Tiruvallur district, south India in May, 1999 and is with chest symptoms suggestive of TB reported
being monitored by Tuberculosis Research Centre at the health facility, three specimens were collected
(TRC), Chennai. The area covered by 209 villages by spot-morning-spot technique as per the
and 9 town blocks has 17 peripheral health procedures described elsewhere 4 . Direct smears
institutions (PHI), of which seven have smear were made immediately after sputum collection and
microscopy facilities. read for acid fast bacilli (AFB) using Ziehl-Neelsen
(ZN) technique 5. In cases of discordant results,
The investigation was based on two types of the highest smear grading was taken as per RNTCP
patients – health facility and community survey. guidelines. The smear readings were graded as
Patients reporting at one of the PHIs with symptoms negative (no AFB in 100 oil immersion fields), scanty
suggestive of TB were diagnosed based on three (1 to 9 AFB in 100 oil immersion fields), 1+ (10 to
sputum examinations. All patients diagnosed were 99 AFB in 100 oil immersion fields), 2+ (1 to 10
given directly observed treatment as per RNTCP bacilli per oil immersion field in at least 50 fields)
guidelines1. Patients detected by sputum examination and 3+ (more than 10 AFB per oil immersion field
based on symptom elicitation and chest X-ray in in at least 20 fields).
an ongoing epidemiological disease survey
(community survey) in the same area were also Utmost care was taken to obtain three good
referred for treatment if they satisfied the definition quality specimens. The quality of the specimens
 GOPI et al: INITIAL SMEAR GRADING VERSUS CONVERSION & TREATMENT OUTCOMES 809

was also recorded in the relevant records. Quality (iii) Cure— A patient is declared ‘cured’ if his/
assurances were followed to ensure the quality of her sputum smear is positive initially, completed
sputum microscopy. This was done by a senior treatment and had negative sputum smears, on at
tuberculosis laboratory supervisor (STLS) who least two occasions, one of which was at the end
checked all positive slides and a 10-20 per cent of treatment.
random sample of negative slides in an unblinded
fashion. In addition, another 10 per cent random (iv) Default— A patient is said to have defaulted
sample of positive as well as negative slides was if he/she discontinued taking anti-TB treatment for
rechecked in a blinded fashion at TRC6. The quality two months or more consecutively.
assurance method adopted during the cohort period
ensured the quality of sputum AFB microscopy. (v) Failure— A smear-positive case who is smear-
positive at 5 months or more after starting the
Culture of Mycobacterium tuberculosis: Two treatment or who was initially smear-negative but
samples of sputum specimen were collected from became smear-positive during treatment is a
each patient immediately after diagnosis or within ‘failure’ case.
a week of starting treatment and transported these
specimens in cetylpyridinium chloride (CPC) solution Data collection and statistical analysis: The data
to the central laboratory for culture examination on the smear grading at the start of the treatment
and susceptibility test to assess the drug sensitivity and at the end of the intensive phase of the
profile. The delay between the collection and treatment, and the treatment outcomes of the
processing of the specimen varied from 7 to 10 patients were collected from the Tuberculosis
days. The isolated cultures were confirmed for M. Register maintained in the Tuberculosis Unit (TU)
tuberculosis by a Niacin test, a 68 0 C catalase of the study area. These patients were visited by
test and growth on para-nitrobenzoic acid 7. All the trained health workers at the health facility or their
cultures were processed and confirmed residence to collect the socio-demographic profile,
independently. The culture results were graded as duration of symptoms, care seeking behaviour and
follows: colonies = 1 to 19 colonies, 1+ = 20 or treatment details using a structured questionnaire.
more but less than 100 colonies, 2+ = innumerable Also, data on personal habits like smoking and
colonies and 3+ = confluent growth. alcoholism were collected. Informed consent was
obtained from all patients included in the study.
Definitions: Standard international definitions 8 were
The data after scrutiny were computerized and
followed to classify TB patients according to
edited incorporating all corrections. Statistical
category and outcome as follows:
analysis and test of significance such as Chi-square
(i) Case— A case of category I patient is the and trend Chi-square were undertaken using Epi-
one whose smear is positive on at least two sputum Info version 6.04d (Centers for Disease Control,
specimens and had no previous history of treatment Atlanta, GA). P<0.05 was considered to be
of more than a month or one smear positive and statistically significant for interpretation of the
chest X-ray abnormal suggestive of tuberculosis. results. Crude odds ratio (OR) and 95 per cent
confidence interval (C.I.) were used for the
(ii) Conversion— A smear-positive patient is said interpretation of factors analysed using univariate
to be ‘converted’ if his/her sputum collected and method. Logistic regression analysis was performed
examined at the end of the intensive phase (or at using SPSS/PC+, Version 4.0 (SPSS Inc, Chicago,
the end of intensive extension phase) found to be IL, 1990) to find out the independent risk factors
smear negative. and adjusted odds ratio (AOR). The criterion
810 INDIAN J MED RES, JUNE 2006

for inclusion of variables in the multivariate model Of the 1463 patients put on treatment, sputum
was set at P≤0.1. was not examined for 99 (41% with 3+ smear
grading) patients due to default or death by the
Results end of two months of treatment. The remaining
1364 patients were subjected to smear examination
Of the 1463 patients, 1206 (82.4%) became
and 354 (26.0%) patients remained smear positive
negative on smear microscopy at two months of
at 2 months requiring extension for an additional
the intensive phase of treatment or at three months
month. Among those with positive smear at 2
in case of extension of treatment for another month
months, 182 (41.1%) of the 443 patients had a high
(Table I). The conversion rate decreased as the
(3+) smear grading compared to 172 (18.7%) of
smear grading increased and the decrease in trend
921 patients with a lower (scanty or 1+ or 2+)
was statistically significant (Trend Chi-square =
smear grading and the difference was statistically
33.1; P < 0.001). Significantly more patients (13.2%;
significant (Chi-square = 77.0; P < 0.001). This
64 of 484) remained smear positive at 3 months
showed that more patients with a higher smear
among those with a higher smear grading than that
among patients with a lower smear grading (5.3%; grading required extension of treatment compared
52 of 979) (Chi-square: 26.7 ; P < 0.001). to that with a lower grading.

Of the 1463 patients, 1109 (75.8%) patients

Table I. Smear and conversion of new smear positive were declared cured (Table II). It is observed that
patients registered in a DOTS programme from May, 1999 the cure rate decreased as the grading of the smear
to December, 2002 in Tiruvallur district, south India result increased and the decrease in trend was
Initial grading No. of patients No. converted (%)
statistically significant (Trend Chi-square =
6.1; P = 0. 01).
Scanty 43 40(93.0)
1+ 562 486(86.5) The distribution of cured patients for various
factors is shown in Table III. Univariate analysis
2+ 374 325(86.9)
demonstrated that the proportion of patients cured
3+ 484 355(73.3)
was significantly lower among older patients
Total 1463* 1206(82.4) (>45 yr), males, smokers, alcoholics, illiterates, those
* 141 patients from whom sputum was not collected due to who had cough for 4 wk, those with 3+ smear
default or death by the end of intensive phase were also grading and those who did not convert at the end
included in the denominator for estimation of conversion as
per the RNTCP policy. of IP of treatment. In multivariate analysis, older

Table II. Treatment outcome among new smear-positive patients registered for treatment in a DOTS programme from May,
1999 to December, 2002 in Tiruvallur district, south India

Smear Treatment outcome

grading
Cured (%) Defaulted Expired Failed Others Total
Scanty 38 (88.4) 5 - - - 43
1+ 432 (76.8) 82 18 28 2 562
2+ 292 (78.1) 44 13 21 4 374
3+ 347 (71.7) 81 27 25 4 484
Total 1109 (75.8) 212 58 74 10 1463
 GOPI et al: INITIAL SMEAR GRADING VERSUS CONVERSION & TREATMENT OUTCOMES 811

Table III. Risk factors for cure among new smear positive patients registered for treatment in a DOTS programme from
May, 1999 to December, 2002 in Tiruvallur, south India

Risk factor No. Cure (%) OR (95%C.I.) P value AOR (95%C.I.)

Age >45 yr 699 490 (70.1)

<45 764 619 (81.0) 1.8 (1.4-2.3) <0.001 1.5 *(1.1-2.1)

Sex Male 1125 817 (72.6)

Female 338 292 (86.4) 2.4 (1.7-3.4) <0.001 1.4 (0.9-2.3)

Education Illiterate 599 488 (74.8)

Literate 766 611 (79.8) 1.3 (1.0-1.7) <0.05 1.2 (0.8-1.6)

Smoking No 708 599 (84.6) 2.4 (1.8-3.1) <0.001 1.1 (0.8-1.7)

Yes 659 461(70.0)

Alcoholism No 849 716 (84.3) 2.7 (2.1-3.6) <0.001 1.7*(1.2-2.4)

Yes 519 345 (66.5)

Cough <4 wk 383 317 (82.8) 1.6 (1.1-2.1) <0.01 1.3 (0.9-1.9)

>4 wk 983 742 (75.5)

Patient delay <4 wk 901 708 (78.6)

>4 wk 439 334 (76.1) 1.2 (0.9-1.5) 0.3 -

Body weight > 40 kg 648 499 (77.0)

< 40 kg 632 466 (73.7) 1.2 (0.9-1.6) 0.2 -

Diagnosis Com. survey 296 217 (73.3)

Health facility 1167 892 (76.4) 1.2 (0.9-1.6) 0.3 -

Smear grade Low 979 762 (77.8) 1.4 (1.1-1.8) <0.01 1.1 (0.8-1.5)
High 484 347 (71.7)

Conversion Yes 1010 885 (87.6) 4.1 (3.1-5.5) <0.001 3.5*(2.6-4.8)

No 354 224 (63.3)

* Statistically significant

patients, alcoholics and those who did not convert Among the 1387 patients, 158 were culture
were more likely to have a less cure rate. negative and 3 were found to be contaminated on
susceptibility test. Of the remaining 1226 patients
Culture results were available for 1387 of 1463 for whom drug sensitivity results were available,
patients. The association between smear and culture 1094 (89.2%) had sensitive organism to izoniazid
grading is shown in Table IV. There was a linear (H) and rifampicin (R), 111(9.1%) had resistant
association between the extent of smear grading organism to H alone, 5 (0.4%) had resistant
and positive cultures (Trend Chi-square= 83.6; organism to R alone and 16 (1.3%) had multi-drug
P<0.001). resistant (MDR-TB) organisms to H and R.
812 INDIAN J MED RES, JUNE 2006

Table IV. Smear and culture gradings of patients registered for treatment in a DOTS programme from May, 1999 to December,
2002 in Tiruvallur, south India

Culture
Smear Total
Negative Colonies 1+ 2+ 3+ NC*

Scanty 23 7 6 3 0 4 43
1+ 89 117 137 154 42 23 562
2+ 27 51 72 145 56 23 374
3+ 19 37 86 208 108 26 484
Total 158 212 301 510 206 76 1463
* NC- sputum not collected

However, no difference in the proportions between alcoholic and being diagnosed by community survey.
patients with low and high grades was found (data It is observed that there was an annual decline of
not shown). 12 per cent in the default rate over a four year
period (data not shown) and it is an indication that
Discussion the performance of the programme has improved.
The death and the failure rates were 4.0 and
Our study shows that conversion and cure rates 5.1 per cent respectively.
were associated with the smear grading of the
patients at the time of start of treatment. Both the More patients with a higher smear grading had
rates decreased as the grading increased and the extension of one month’s treatment in the intensive
trend was statistically significant. Smear positivity phase compared to those with a lower smear
depends on the extent of lesion or the presence of grading. In a similar study 12 conducted by the New
cavitations and the smear grading is associated with Delhi Tuberculosis Centre to assess the importance
the infectiousness of the case 9. Smear positive of initial smear grading as a predictor of treatment
cases are infectious cases transmitting the infection outcome revealed that a larger proportion of
and disease to others especially those who are close previously untreated sputum positive patients with
contacts 10 . In our study, the conversion and the 3+ grading required extension of intensive phase.
cure rates of the patients were 82 and 76 per cent In addition to the above findings, we have observed
respectively which were lower than the national that there was a linear association between smear
average 1 . The defaulter rate was as high as 14.5 grading and positive cultures. The proportion of
per cent compared to expected level of 5 per cent culture positives among patients with scanty positives
or less in RNTCP. A study 11 from a cohort of all was 41 per cent only. The culture examination was
patients from the same area has shown a default done on using different samples collected from the
rate as high as 19 per cent of which nearly three- same patients. The delay between collection and
fourths did so by the end of the intensive phase, a processing of these specimens for culture varied
period during which the symptoms usually declines. from 7 to 10 days. This might have contributed to
In that study, a higher default rate was reported the observed difference in the proportion of culture
to be associated with irregular drug intake, being positives. The association between the smear
male, having a history of previous treatment, being microscopy by fluorescent technique and cultures
 GOPI et al: INITIAL SMEAR GRADING VERSUS CONVERSION & TREATMENT OUTCOMES 813

based on the same specimen showed that great in RNTCP is to be viewed further in lieu of the
majority (94%) of the smear positives were culture above observations and necessitates more attention
positives (data not shown). on patients with higher smear grading to motivate,
counsel and ask them to come for treatment with
A study 11 on an earlier cohort of patients from sustained commitment in the control of tuberculosis.
the same area has reported that high rate of default For these patients, there is a need to extend the
and deaths were responsible for low cure rates treatment for one more month in the IP of treatment
and found various risk factors for unfavourable
treatment outcomes like default, failure and death Acknowledgment
of tuberculosis patients treated under the DOTS
programme and recommended various measures The authors thank Shriyut S. Radhakrishnan (Senior
to improve the programme performance. Among Treatment Supervisor) and E. Prabhakaran (Senior Tuberculosis
patients with a higher smear grading, the proportion Laboratory Supervisor) for maintaining patient Tuberculosis
of patients whose smear had not converted, was register, Shriyut Abdul Kudhoos, A. Narasimhan, R. Sasidharan,
not only higher at two months of treatment but S. Arjunan and L. Krishnamacharya of the Field Epidemiology
also at 3 months. We have further analyzed the Unit for patient interviews. Drs C. Kolappan, K. Sadacharam
data to identify other factors associated with cure and P. Paul Kumaran for carrying out the community survey
rate of patients. Another important finding was and the laboratory assistance provided by the staff of
that patients reported at health facilities with Bacteriology department. The authors are grateful for the
symptoms suggestive of tuberculosis and diagnosed assistance and co-operation of the Joint Director of Health,
were with high grades of smear compared to those Deputy Director (Thoracic Medicine), Deputy Director Health
identified through community survey (data not Services of Tiruvallur district, Tamil Nadu state and all the
shown). This supports the findings of an earlier medical and paramedical staff including treatment observers
study 13 comparing the results of patients detected who participated in this work. The assistance rendered by
at health facility with those by community survey. the staff of the Epidemiology Unit in checking data and arranging
Patients identified at the health facilities are more for computerization is highly appreciated. The staff of EDP
likely to be infectious due to the reason that they division is gratefully acknowledged for data entry and data
voluntarily report late after the onset of symptoms management. Lastly, the authors are grateful to all the patients
suggestive of tuberculosis. The need for early for their cooperation.
diagnosis after onset of symptoms should be
This work was funded in part by a grant from the United
popularized in the community. Since the inception
States Agency for International Development provided through
of DOTS strategy in the area, health education
the World Health Organization, SEARO, New Delhi.
programme through information, education and
communication (IEC) has been ongoing to motivate
References
symptomatic patients to report earlier for diagnosis.
The awareness about tuberculosis and the need 1. Khatri GR, Frieden TR. The status and prospectus of
for early diagnosis and regular treatment is now tuberculosis control in India. Int J Tuberc Lung Dis
on the increase in the area. 2000; 4 : 193-200.

In conclusion, our findings showed that among 2. World Health Organization, Treatment of Tuberculosis:
patients with severe disease in terms of higher Guidelines for National Programmes, 2 nd ed. WHO/
smear grade, the conversion and cure rates were TB/97.220, Geneva, 1997.
lower compared to other patients and more had
unfavourable treatment outcomes in terms of higher 3. Enarson DA, Rieder HL, Arnadottir T, Trebucq A.
default and death. The smear grading that exists Management of tuberculosis. A guide for low income
814 INDIAN J MED RES, JUNE 2006

countries. Fifth edition. Paris: International Union 8. World Health Organization. Revised international
against Tuberculosis and Lung Disease, 2000. definitions in tuberculosis control. Int J Tuberc Lung
Dis 2001; 5 : 213-5.
4. Gopi PG, Subramani R, Selvakumar N, Santha T,
Eusuff SI, Narayanan PR. Smear examination of two 9. Toman K. Toman’s tuberculosis. Case detection,
specimens for diagnosis of pulmonary tuberculosis in treatment and monitoring. Questions and answers.
Tiruvallur district, south India. Int J Tuberc Lung Dis Geneva: World Health Organization, 2004.
2004; 8 : 824-8.
10. British Thoracic Association. A study of standardized
5. Revised National Control Programme Central contact procedure in tuberculosis. Tubercle 1978;
Tuberculosis Division. Manuals for Laboratory 59 : 245-59
Technicians; New Delhi, India: Directorate General of
Health Services, Ministry of Heath and Family Welfare, 11. Santha T, Garg R, Frieden TR, Chandrasekaran V,
1999, Available at www.tbcindia.org/Labmanual.pdf, Subramani R, Gopi PG, et al. Risk factors associated
accessed on. with default, failure and death among tuberculosis
patients treated in a DOTS programme in Tiruvallur
6. Selvakumar N, Prabhakaran E, Fathima R, Chandu NA, District, South India. Int J Tuberc Lung Dis 2002;
Srinivasan S, Santha T, et al. Blinded rechecking 6 : 780-8.
of sputum smears for acid- fast bacilli to ensure
quality and usefulness of re-staining of smears to 12. Rajpal S, Dhingra VK, Aggarwal JK. Sputum grading
assess false-positive errors. Int J Tuberc Lung as predictor of treatment outcome in pulmonary
Dis 2003; 7 : 1077-82. tuberculosis. Indian J Tuberc 2002, 49 : 139-41.

7. Selvakumar N, Fathima R, Garg R, Rajasekaran S, 13. Santha T, Garg R, Frieden TR, Subramani R, Gopi PG,
Nalini S, Thyagarajan K, et al. Evaluation of phenol Chandrasekaran V, et al. Are community surveys to detect
ammoniaum sulphate sedimentation smear microscopy tuberculosis in high prevalence areas useful? Results
method for diagnosis of pulmonary tuberculosis. J Clin of a comparative study from Tiruvallur district, south
Microbiol 2002; 40 : 3017-20. India. Int J Tuberc Lung Dis 2003; 7 : 1-8.

Reprint requests: Dr P.R. Narayanan, Director, Tuberculosis Research Centre, Mayor V.R. Ramanathan Road
Chetput, Chennai 600031, India
e-mail: [email protected]