d’Arc Lyra Batista, J; de Ftima Pessoa Milito de...

drsquoArc Lyra Batista J de Ftima Pessoa Milito de Albuquerque M deAlencar Ximenes RA Rodrigues LC (2008) Smoking increases therisk of relapse after successful tuberculosis treatment Internationaljournal of epidemiology 37 (4) pp 841-51 ISSN 0300-5771 DOIhttpsdoiorg101093ijedyn113

Downloaded from httpresearchonlinelshtmacuk6174

DOI 101093ijedyn113

Usage Guidelines

Please refer to usage guidelines at httpresearchonlinelshtmacukpolicieshtml or alterna-tively contact researchonlinelshtmacuk

Available under license httpcreativecommonsorglicensesby-nc-nd25

Smoking increases the risk of relapse aftersuccessful tuberculosis treatmentJoanna drsquoArc Lyra Batista1 Maria de Fatima Pessoa Militao de Albuquerque12Ricardo Arraes de Alencar Ximenes13 and Laura Cunha Rodrigues4

Accepted 13 May 2008

Background Recent tobacco smoking has been identified as a risk factor fordeveloping tuberculosis and two studies which have investigatedits association with relapse of tuberculosis after completion oftreatment had conflicting results (and did not control for confound-ing) The objective of this study was to investigate risk factors fortuberculosis relapse with emphasis on smoking

Methods A cohort of newly diagnosed TB cases was followed up from theirdischarge after completion of treatment (in 2001ndash2003) untilOctober 2006 and relapses of tuberculosis ascertained during thatperiod A case of relapse was defined as a patient who starteda second treatment during the follow up

Results Smoking (OR 253 95 CI 123ndash521) and living in an area wherethe family health program was not implemented (OR 361 95 CI146ndash893) were found to be independently associated with relapseof tuberculosis

Conclusions Our results establish that smoking is associated with relapse oftuberculosis even after adjustment for the socioeconomic variablesSmoking cessation support should be incorporated in the strategiesto improve effectiveness of Tuberculosis Control Programs

Keywords Tuberculosis control successful treatment relapse risk factorssmoking

IntroductionTuberculosis is a significant health problem worldwidecontrol of tuberculosis relies heavily on identification

and successful treatment of cases Treatment scheduleslast for 6 months One of the problems of tuberculosiscontrol is the fact that cases that were successfullytreated can relapse needing a second treatment Iden-tification of factors that increase the risk of relapseof tuberculosis after cure or completion of treatment isof great interest The monitoring of the effectiveness oftuberculosis programs is based on the routine classifi-cation of the outcome of each tuberculosis treatmentinto (i) treatment failure (patient still have positivesmear 5 months after start of treatment or later)(ii) death (patient who dies for any reason duringthe course of treatment) (iii) default (patient inter-rupts treatment for two consecutive months or more)(iii) cure (patient who is sputum smear negative in two

Corresponding author Centro de Pesquisas Aggeu MagalhaesFIOCRUZ Av Moraes Rego sn Campus da UniversidadeFederal de Pernambuco Cidade Universitaria Recife PECEP 50670ndash420 Brazil E-mail militaocpqamfiocruzbr

1 Universidade Federal de Pernambuco Recife Brazil2 Centro de Pesquisas Aggeu magalhaesFIOCRUZ Recife

Brazil3 Universidade de Pernambuco Recife Brazil4 London School of Hygiene and Tropical Medicine London

UK

The online version of this article has been published under an open access model Users are entitled to use reproduce disseminate or display the open access

version of this article for non-commercial purposes provided that the original authorship is properly and fully attributed the Journal and Oxford University Press

are attributed as the original place of publication with the correct citation details given if an article is subsequently reproduced or disseminated not in its entirety

but only in part or as a derivative work this must be clearly indicated For commercial re-use please contact journalspermissionsoxfordjournalsorg

Published by Oxford University Press on behalf of the International Epidemiological Association

The Author 2008 all rights reserved Advance Access publication 13 June 2008

International Journal of Epidemiology 200837841ndash851

doi101093ijedyn113

841

consecutive tests at least one in the last months oftreatment) (iv) treatment completed (a patient whohas completed treatment has no symptoms but doesnot meet the criteria for cure because of the absenceof a sputum test in the last month of treatmentusually because a sputum test was not conducted orthe patient could not produce sputum) Patients clas-sified as cure or completed treatment are consideredtreatment successes1

In Recife Brazil where this study is conducted 16 ofthe people starting treatment against tuberculosis hadhad at least one previous treatment Of this one-thirdhad cured or completed treatment and therefore was arelapse2 This is typical of many developing countriesSo far factors known to increase the risk of relapse arerestricted to deficiencies in treatment (taking the drugsirregularly taking the wrong drugs taking the rightdrugs in the wrong dosage) and (probably as indicatorsof deficiencies in treatment) patients missing consulta-tions having adverse events to the drugs used in thetreatment of tuberculosis and health services problemslike irregular supply of drugs3 Most exciting and ofpotential relevant to relapse recently tobacco smok-ing has been identified as a risk factor for develop-ing tuberculosis The proposed biological mechanismfor this is a decreased resistance to Mycobacteriumtuberculosis due to functional and morphological changesto macrophages in the alveoli of smokers45 It istherefore plausible that smoking may increase therisk of relapse by increasing risk of persistence ofM tuberculosis infection after treatment and risk of anyresidual M tuberculosis infection leading to disease5

Two studies have investigated this possibility and hadconflicting results36 None controlled for confounding

This article reports on a study of risk factors forrelapse based on the analysis of cases of relapse in acohort of cases of tuberculosis that was followed upfrom start of tuberculosis treatment to expected datefor completion of treatment for a previous study andthen followed up for this analysis for 3ndash5 years aftercompletion of treatment

MethodsStudy population and designSite of the study was Recife a city with a populationof one and half million people in the northeast ofBrazil In 2002 the incidence of tuberculosis in Recifewas 104 per 100 000 inhabitants7 and treatmentconsisted of an initial phase of isoniazid rifampicinand pyrazinamide lasting for 2 months followed bya second phase with isoniazid and rifampicin onlyfor a period of 4 months8 This is different from theinternationally recommended first-line treatmentwhich uses four drugs in the initial phase 2 monthsof ethambutol associated to isoniazid rifampicin andpyrazinamide9 The objective of the analysis reportedhere was to identify risk factors for tuberculosis

relapse leading to re-treatment in those who com-pleted their tuberculosis treatment successfully andsurvived the follow-up period Methods of the mainstudy (following cases from diagnosis to completionof treatment) were described elsewhere but in shortcases of tuberculosis diagnosed from May 2001 toJuly 2003 who were residents in Recife and aged13 years or older were recruited into the studycompleted a questionnaire on potential risk factorsand had blood collected for anti-HIV serology Imme-diate outcome of treatment was ascertained from theregister of the health unit where each patient wastreated and from each patientrsquos medical record soonafter the expected date of end of treatment and again6 months later and classified into failure defaultdeath confirmed cure or completion of treatmentResults of the initial study of determinants of theoutcome of tuberculosis treatment were publishedelsewhere1011 In October 2006 additional informa-tion was sought to identify cases from the cohort whorelapsed (started re-treatment for tuberculosis aftersuccessful completion of treatment) information onpatients who died was sought for exclusion from thecohort and for sensitivity analysis

The study population in this analysis consisted ofpatients from the original cohort who were dischargedfrom the first treatment after treatment success1 Exclu-sion criteria were not having completed treatment fail-ure of treatmenthaving a positive smear at the endof treatment (as these would remain in treatmentwith a different scheme) having died and having hada previous treatment for tuberculosis (as these werealready relapses when recruited into the first studyof determinants of outcome of treatment)

We did not examine separately those with negativesputum at the end of treatment and those with nosputum (cures vs competed treatment) as numberswere not large enough for subgroup analysis we didnot expect the two groups to be different becausethe main reason for not having a sputum test late inthe course of treatment was that the test was notrequested (data not shown) A case of relapse wasdefined as a patient from the original cohort dis-charged because of cure or completion of treatmentwho was diagnosed with tuberculosis and starteda second course of treatment during the follow upperiod Patients from the original cohort who survivedto the end of the follow up period and did not starta new treatment were the comparison group

We identified relapses by searching the SurveillanceSystem for Infectious Diseases (SINANMS) SINANMS is an electronic system with notifications of infec-tious diseases managed by the Ministry of Health(MS) in Brazil The search was done using name anddate of birth In Brazil tuberculosis treatment iscarried out only by the public sector and to receivetuberculosis treatment a patient must be notifiedso the notification system should include all diag-nosed cases receiving treatment We identified deaths

842 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

for exclusion and sensitivity analysis by searchingthe routine mortality system (SIMMS) an electronicsystem with all routine mortality data with causeof death coded to ICD managed by the Ministry ofHealth in Brazil SIMMS was searched based onname and date of birth and the cause of death wasabstracted for all deaths linked to the cohort

Exposure information had been collected as part ofthe previous study (follow up from first diagnosis oftuberculosis to discharge) from a standard pre-codedquestionnaire in an interview conducted face to faceby a trained interviewer and information abstractedfrom the tuberculosis treatment record and included

(i) Clinical and epidemiological factors treatment delay(defined as delay between onset of symptomsand start of treatment (the cut-off point adoptedwas according to Santos et al10) Clinical form oftuberculosis (pulmonary or extra pulmonary)and HIV co-infection history of contact with othercases of tuberculosis knowing somebody with TBand living in the same house as a TB patient

(ii) Biological factors age and sex(iii) Social factors employment status of patient and of

the head of the household literacy income andschooling of the head of the household access towater supply ownership of goods (radio refrig-erator video washing machine microwaveoven computer TV set air conditioning and acar) number of goods (the number of goods wasgrouped as follows first possession of each ofthe different goods was counted as 1 indepen-dent of the number of each good possessed (forexample owning two cars counted as owningone good) then the number of goods possessedwas summed and grouped into four categories0ndash1 2ndash3 4ndash6 and 7 or more goods

(iv) Lifestyle Alcohol consumption grouped as doesnot drink drinks rarely drinks at least once aweek drinks every day and once starts drinkingfinds it difficult to stop They were aggregatedduring analysis in two groups light drinker(none drinking rarely weekly) and heavydrinker (drinking every day or finding it difficultto stop once start drinking) Smoking groupedinto whether the person was a smoker at thetime of the interview (smokers) had neversmoked (never smokers) or used to smoke butstopped and the date when stopped (ex smok-ers) For the analysis because of small numbersand because the hypothesis was that the effectof smoking disappears with time those whonever smoked were grouped with those that hadgiven up smoking one year or more before theinterview)

(v) Access to health services number of health unitsattended with the complaint eventually diag-nosed as tuberculosis before it was diagnosedand treatment started whether unit attendedwas in the same neighborhood and in the

district of residence of case and whether theFamily Health Program (FHP) was implementedin the Health District of residence of case at thetime of the start of the treatment The FHP is anew program in which a team of trained healthvisitors visit families in the community Eachteam is responsible for approximately 3200 resi-dents in a defined geographical area and islinked to a FHP team of clinicians and otherhealth professionals in a supporting health unitThe FHP was being implemented during thestudy period and some but not all of the studyareas were covered

Statistical analysisLogistic regression was used to identify risk factorsfor relapse Time since end of treatment and ascer-tainment of re-treatment in October 2006 varieda variable indicating the year of the end of the firsttreatment was included in all models Variables asso-ciated with the outcome with a P-value 4025 or withan odds ratio of 515 were included in a multiplelogistic regression analysis which also included theyear of entry in the cohort A step-up procedure wasused Crude and adjusted ORs and 95 CIs werecalculated Since smoking is associated with povertyand alcohol consumption to increase the confidencethat the association between re-treatment and smok-ing was not a result of confounding by socioeconomicstatus additional unplanned analysis were under-taken exploring the effect of each of a large numberof social variables collected as part of the first studyon the relationship between smoking and relapseemployment status of patient and of the head of thehousehold literacy income and schooling of the headof the household access to water supply ownershipof each of the following goodsmdashradio refrigeratorvideo washing machine microwave oven computertelevision set air conditioning car number of goodsowned and alcohol consumption To explore thepotential consequences of excluding deaths a sensi-tivity analysis was conducted modelling the conse-quences of treating as relapses (i) all deaths coded totuberculosis and (ii) all deaths Data were stored by asoftware program (EPI-INFO version 604 Centers forDisease Control Atlanta GA) and analysed usingSTATA 90 (Stata-Corp LP College Station TX) Thestudy was approved by the Ethical Committee of theCenter of Health Sciences of the Federal Universityof Pernambuco All patients signed an informedconsent form

ResultsA total of 1353 diagnosed TB patients had been iden-tified for potential selection from the previous studyfrom May 2001 to July 2003 Of these 311 patientswere not eligible for the first study because they

SMOKING AND TUBERCULOSIS RELAPSE 843

referred previous treatment for tuberculosis (and sowere already relapses) and 42 patients because theywere under 13 years of age Thus 1000 patients ofpotential interest for this analysis were followedduring the period of the tuberculosis treatmentFrom this group 246 patients were not eligible forthis analysis because they had an unsuccessful out-come (146 defaulters 25 died during the course oftreatment 16 were treatment failures) 28 wereexcluded because they transferred out to a differenttuberculosis treatment unit and 31 because they hadno information on outcome of treatment The studypopulation therefore consisted of 754 patients withsuccessful treatment outcomes (cured or completedtreatment) and they were followed until 31 October2006 During the follow-up period 43 patients wereexcluded because they died after completion oftreatment therefore the final group consisted of 711patients of which 37 relapsed and 674 had notrelapsed by the end of follow up (Figure 1)

The mean age of cases in the analysis was 387 inmales and 356 in females In this study the rate ofrelapse was 5 (37711 patients) Approximately half(487) of the cases of relapse occurred in the firstyear after discharge

Tables 1ndash4 show the results of univariable analysisof the association between relapse and risk factorsadjusted by the year of entry in the cohort Amongbiological and lifestyle factors (Table 1) only smok-ing was associated with relapse (OR 234 95 CI

117ndash468) None of the socioeconomic factors wasassociated with relapse (Table 2) Among the vari-ables related to health services only not living in anarea that receives visits by the FHP was associated

Figure 1 Study flow diagram

Table 1 Association between relapse of tuberculosis and biologic and lifestyle characteristics of individuals ORmdashcrudeand adjusted by the year of entry in the cohort Recife 2006

Biological factorsand lifestyle

Relapse Control

N N Crude OR (CI) P OR (CI) Adjusted P

Age (years) 0154 0401

13ndash29 10 413 232 9587 100 ndash 100 ndash

30ndash49 22 696 294 9304 173 (081ndash374) 0159 172 (080ndash372) 0164

550 5 327 148 9673 078 (026ndash234) 0662 079 (026ndash235) 0668

Total 37 ndash 674 ndash

Sex

Female 14 556 238 9444 100 ndash 100 ndash

Male 23 501 436 9499 089 (045ndash177) 0755 088 (044ndash176) 0732


Alcohol consumption

None or light drinker 32 532 569 9468 100 ndash 100 ndash

Heavy drinker 3 405 71 9595 075 (022ndash252) 0643 076 (022ndash254) 0653


Cigarette smoking

Never smoking or givenup smoking

23 416 530 9584 100 ndash 100 ndash

Ever smoking 14 921 138 9079 233 (117ndash466) 0016 234 (117ndash468) 0016



Table 2 Association between relapse of tuberculosis and socioeconomic characteristics of individuals ORmdashcrudeand adjusted by the year of entry in the cohort Recife 2006

Social factors

Relapse Control


Literacy

Yes 33 537 582 9463 100 ndash 100 ndash

No 4 417 92 9583 076 (026ndash221) 0624 075 (025ndash217) 0596


Employment

Yes 12 545 208 9455 100 ndash 100 ndash

No 25 511 464 9489 093 (046ndash189) 0850 093 (045ndash189) 0844


Number of individuals per household


55 13 444 280 9556 074 (037ndash148) 0396 073 (037ndash147) 0391


Income of head of household (MW)

52 5 360 134 9640 100 ndash 100 ndash

lt2 31 614 474 9386 175 (067ndash459) 0254 175 (067ndash461) 0253


Employment of head of household

Yes 19 521 346 9479 100 ndash 100 ndash

No 18 520 328 9480 099 (051ndash193) 0998 099 (051ndash192) 0986


Schooling of the head of household

9 or more years 3 278 105 9722 100 ndash 100 ndash

Up to 8 years 31 558 525 9442 206 (062ndash688) 0237 206 (061ndash689) 0239


Piped water

Yes 35 529 627 9471 100 ndash 100 ndash

No 2 444 43 9556 083 (019ndash358) 0806 085 (020ndash369) 0835


Ownership of washing machine

Yes 7 486 137 9514 100 ndash 100 ndash

No 29 517 532 9483 106 (045ndash248) 0881 106 (045ndash247) 0890


Ownership of microwave oven

Yes 3 395 73 9605 100 ndash 100 ndash

No 33 529 591 9471 136 (041ndash454) 0619 135 (040ndash453) 0627


Ownership of computer

Yes 2 303 64 9697 100 ndash 100 ndash

No 35 552 599 9448 187 (044ndash795) 0397 186 (043ndash794) 0399


(continued)


Table 2 Continued

Social factors

Relapse Control


Ownership of a car

Yes 8 588 128 9412 100 ndash 100 ndash

No 29 507 543 9493 085 (038ndash191) 0702 084 (037ndash189) 0685


Ownership of air conditioning

Yes 2 244 80 9756 100 ndash 100 ndash

No 35 558 592 9442 236 (056ndash100) 0243 232 (054ndash985) 0255


Number of goods possessed 0312 0592

7ndash9 goods 3 469 61 ndash 100 ndash 100

4ndash6 goods 8 377 204 0744 080 (020ndash310) 0750 080 (020ndash312) 0750




Table 3 Association between relapse of tuberculosis and health services variables ORmdashcrude and adjusted by theyear of entry in the cohort Recife 2006

Access to healthservices factors

Relapse Control


Number of health units

Only 1 5 373 129 9627 100 ndash 100 ndash

52 31 550 533 9450 150 (057ndash393) 0409 147 (056ndash388) 0428


HS in same district of residence

Yes 23 518 421 9482 100 ndash 100 ndash

No 14 524 253 9476 101 (051ndash200) 0971 102 (052ndash204) 0935


HS in neighborhood of residence

Yes 7 778 83 9222 100 ndash 100 ndash

No 30 483 591 9517 060 (025ndash141) 0244 060 (025ndash141) 0244


Residence in areas of FHP visits

Yes 6 232 253 9768 100 ndash 100 ndash

No 30 699 399 9301 317 (130ndash772) 0011 325 (133ndash794) 0010


Health District of treatment 0412 0615

HD IV 6 351 165 9649 100 ndash 100 ndash

HD I 0 0 59 1000 ndash ndash ndash ndash

HD II 7 455 147 9545 131 (043ndash398) 0635 135 (042ndash429) 0609

HD III 11 821 123 9179 246 (088ndash683) 0084 243 (087ndash682) 0091

HD V 11 681 149 9313 203 (073ndash562) 0173 208 (074ndash577) 0160

HD VI 2 606 31 9394 177 (034ndash919) 0495 186 (035ndash979) 0461



with relapse (OR 325 95 CI 133ndash794) (Table 3)The only clinical or epidemiological factor associatedwith relapse was having had a delay of 60 or moredays before start of the first treatment of tuberculosisand even this was borderline (OR 223 95 CI099ndash498) (Table 4)

Table 5 shows the results of the multivariableanalysis The following variable remained in the

final model smoking (OR 253 95 CI 123ndash521)and living in an area in which the FHP was notimplemented (OR 361 95 CI 146ndash893)

The association between relapse and smoking wasadjusted for all socioeconomic variables and alcoholconsumption to exclude the possibility of confoundingby these variables There was no important change inthe magnitude of this association when these

Table 4 Association between relapse of tuberculosis and clinical and epidemiological characteristics of individualsORmdashcrude and adjusted by the year of entry in the cohort Recife 2006

Clinical andepidemiological factors

Relapse Control


Knows someone with tuberculosis

No 23 564 385 9436 100 ndash 100 ndash

Yes 14 475 281 9525 083 (042ndash164) 0602 083 (042ndash165) 0606


Household contact with a case of tuberculosis

No 32 522 581 9478 100 ndash 100 ndash

Yes 05 510 93 9490 097 (037ndash257) 0961 098 (037ndash259) 0974


Clinical form of tuberculosis

Pulmonary 31 512 574 9488 100 ndash 100 ndash

Extra pulmonary 05 481 99 9519 093 (035ndash246) 0892 093 (035ndash245) 0889


Treatment delay

460 days 8 301 258 9699 100 ndash 100 ndash

460 days 28 657 398 9343 227 (101ndash505) 0045 223 (099ndash498) 0050


HIV co-infection 0962 0972

No 1 714 13 9286 100 ndash 100 ndash

Yes 12 538 211 9462 074 (009ndash613) 0780 076 (009ndash640) 0807

Not known 24 537 423 9463 073 (009ndash587) 0774 076 (009ndash615) 0804


Table 5 Multivariate analysis of the association between relapse of tuberculosis and characteristics of individualsORmdashcrude and adjusted for each other and by the year of entry in the cohort Recife 2006

OR and 95 CIadjusted by

year of entry inthe cohort P


year of entry inthe cohort andby each other P paf

Cigarette smoking

Never smoking or given up smoking 100 ndash 100 ndash

Ever smoking 234 (117ndash468) 0016 253 (123ndash521) 0011 0228


Yes 100 ndash 100 ndash

No 325 (133ndash794) 0010 361 (146ndash893) 0005 0585


variables were included in the model the change inthe value of the crude and adjusted OR was lessthan 10 when adjustment was made for incomethe CI was wider and close to 1 but this wasassociated with a small change in the value of theOR (from 233 to 203) To evaluate the potentialeffect of misclassification of relapse in patients whodied we repeated the analysis considering the sixdeaths with tuberculosis as the cause and consideringall 43 deaths as relapses The inclusion of these casesdid not affect substantially our results as only minorchanges of the ORs were observed The OR was 205(95 CI 106ndash396) when six deaths were treated asrelapses and 197 (95 CI 119ndash325) when all deathswere treated as relapses

DiscussionSmoking and living in an area where the FHP was notimplemented were found to be independently asso-ciated with relapse of tuberculosis

The incidence of relapse in this cohort of patients(52) was lower than that found by Thomas et alin a study carried out in South India (123)6 Thisdifference might be related to some extent to differ-ences between the two populations (for example indegree of under nutrition) but we believe that thetwo most likely explanations are active ascertainmentin the Indian study and better treatment adherence inBrazil We expand on these below First in the Indiancase study finding was active with field workersvisiting the study subjects at 6 12 and 18 monthsafter completion of treatment and will have identifiedcases that might have progressed to cure withoutdiagnosis and cases that would have been diagnosedmuch later whereas in the Brazilian study relapseswere ascertained as they presented to health serviceswith symptoms Second the quality of the tubercu-losis control program in Brazil is probably better thanin India The quality of the programme is essentialin assuring adherence to treatment and low adher-ence is a well-established determinant of relapse Thetuberculosis control programme has been improvingin Brazil with a marked decrease in the percentage ofpatients notified with relapse in the city of Recife aswell as in Brazil from 2001 to 20057

In our study the bulk (648) of the relapses occur-red during the first year of follow-up This finding issimilar to that of Nogueira et al and Oliveira andMoreira who found respectively percentages of 65and 619 year after completion of treatment1213

In this study as expected neither sex nor age norsocial factors were risk factors for relapse of tuber-culosis Similar results were found by Oliveira andMoreira and Ormerod and Prescott314

Tuberculosis has been characterized as a diseaseof poverty but this is an aspect of developing tuber-culosis not of relapse Measures of low socioeconomicstatus like low family income illiteracy and low social

class have been found to be associated with anincreased risk of developing tuberculosis15 Althoughsocioeconomic variables have never been shown tobe associated with an increase in tuberculosis relapseinadequate treatment has and it is possible thatpoverty may lead to inadequate treatment in somecircumstances Even in countries where the tubercu-losis control program is well structured and diag-nosis and treatment are provided with no costs to thepatient cases may not be able to pay the indirect costsof treatment eg for transport to the health units16

In a study carried out in Salvador Northeast of Brazilthe costs to families because of loss of income andexpenses due to the disease were very high on average33 of the family income17 In Recife where this studywas conducted low socioeconomic status measuredas illiteracy or low income of the head of householdincreased the risk of all three negative outcomes oftuberculosis treatmentmdashdropout treatment failureand death11 However the present study in the samepopulation found no association of relapse to povertyThis may be because of low power or may indicate therelative magnitude of biological mechanisms onrelapse

The present study found no association betweenalcohol consumption and relapse which differs fromthe study of Selassie et al which described an increaseof almost 4-fold in the risk of relapse in the groupof alcoholics18 However Selassie et al did not controlfor smoking and if alcohol and tobacco consumptionare associated in that setting the association measuredmay be reflecting an unmeasured effect of smokingIn relation to the role of alcohol consumption andsmoking in the risk of developing tuberculosis (ratherthan relapsing) Ruffino-Neto and Ruffino19 severaldecades ago investigated the synergism between thesetwo factors and found that smoking was associ-ated with tuberculosis only in the group of drinkers(alcohol) whereas Brown and Campbell on examin-ing confounding between the two variables concludedthat alcohol had a stronger association with tuber-culosis20 Due to these findings smoking was adjustedfor alcohol consumption in the present study but therewas no confounding effect between alcohol consump-tion and smoking on the risk of relapse (data notshown) In addition as smoking is more frequentamong the poor the association between smoking andrelapse was adjusted for each of the socioeconomicvariables collected previously and there was only a verysmall change in the value of the OR Thomas et al whoalso described an association between smoking andrelapse did not control for the socioeconomic factors6

Despite the fact that in the second half of the lastcentury several studies pointed to an associationbetween smoking and developing tuberculosis192122

there was a gap of decades until scientific interest inthis topic was renewed More recently studies haveinvestigated and found evidence of smoking as a riskfactor for developing tuberculosis independently


or interacting with alcohol consumption and evidenceof progression from latent tuberculosis infection todisease45 The association between smoking anddeath from tuberculosis remains controversial whilein the review of Chiang et al smoking was associatedwith increased TB mortality23 Bates et al describedno clear effect of smoking on the case fatality rate inthose with active TB24

Smoking was found to be significantly associatedwith relapse in the present study (OR 234 95 CI117ndash468) corroborating the finding of Thomas et alwho described a similar association (OR 31 95 CI16ndash60)6 Several studies highlight the associationbetween smoking and the development of tuber-culosis4523ndash25 and the mechanism proposed toexplain this association is the neutralisation of thetumour necrosis factor a (TNF-a) in the pulmonarymacrophages by substances of the tobacco leavingthe patient more susceptible to a progressive devel-opment of disease from latent tuberculosis infec-tion5 The reduction in the TNF-a in this groupmay be explained by the high level of iron in thebronchoalveolar macrophages of smokers26 Thismechanism may also be behind the development ofrelapse

Tobacco consumption has been increasing mainly indeveloping countries In Brazil one of the greatestproducers and exporters of tobacco the NationalProgram of Control of Smoking of the Ministry ofHealth has been acting through the National Instituteof Cancer (INCA) and there has been some importantachievements as a decrease of nearly 50 in theprevalence of smokers between 1989 and 200327

Nevertheless it is still important to incorporate in thestrategies of the Tuberculosis Control Program theevidence that smoking increases the risk of relapseThis knowledge may become one more tool in thecontrol of the disease

Concerning the factors related to health services therisk of relapse was significantly lower among patientsliving in areas where the FHP was implemented TheMinistry of Health has defined a policy of decentra-lization and has transferred to the local authoritiesthe administration of the health services This policywhich aims to facilitate the access to health servicesand includes periodical visits of trained and reason-ably well-paid health agents to all families has beenimplemented with good results28 The TuberculosisControl Plan emphasizes the performance of thehealth agents and FHP teams as a tool to increasethe therapeutic adherence and prevent patients frominterrupting treatment829 During the period from2000 and 2006 the proportion of health services thatimplemented supervised treatment whether or notusing the FHP increased from 7 to 8130 In Recifeas in Brazil as a whole there has been a markeddecrease in the percentage of patients notified withrelapse from 2001 to 20057 It was not possible toestablish the degree of adherence in this study as

there is no record of treatment adherence However arobust explanation for the lower rate of relapse inpatients who received visits from the agent from theFHP is that the visits lead to better adherence totreatment during the 6 months long therapeuticregimen A closer monitoring of tuberculosis patientseven after discharge would have increased detectionof relapsed patients and tend to decrease themagnitude of the association

Since Brazil does not follow the internationallyrecommended first-line treatment9 it is of course notpossible to evaluate if the type of treatment affectsthe frequency of relapse in this study although theproportion of cases relapsing is not out of line withother countries Concerning the association betweenFHP and smoking with relapse as all patients studiedwere treated with the same drug regimen this wouldnot distort the association between these factors andrelapse but of course to generalize these findings toother treatment regimens the study must be repeatedin the appropriate settings

A limitation of this study was that we did not usethe CAGE standard questionnaire to ascertain alco-holism and therefore can only speak of alcoholconsumption Another limitation is that it is notpossible to differentiate relapse due to reactivation ofthe disease from tuberculosis resulting form a newinfection as genotyping was not used to identify thestrains of M tuberculosis This is not unusual Coxet al31 reported that only one among the 16 studiesincluded in their systematic review differentiated truerelapse from reinfection with a different strain ofM tuberculosis

Ascertainment of outcomes relied on a routinesurveillance system and studies relying on passivedetection of relapse are more likely to under-reportHowever we expect that in Brazil the risk relapsewithout diagnoses is low because the NationalProgram for Tuberculosis Control (PNCT) has stan-dard procedures for investigation and diagnosis in thewhole country and the standardized treatment regi-men is only delivered after the case is notified toSINAN The potential risk of differential detection ofrelapse for different groups eg by SES is minimizedby the partnership of the PNCT with the CommunityOutreach Program (PACS) and FHP which widens theactions and coverage of the PNCT especially in areasof low SES2829

A total of 43 patients were excluded because theydied after completion of the successful treatment andthis could generate survival bias Tuberculosis waspointed as basic cause of death in 6 of the 43 (datafrom mortality information system of the Ministry ofHealthmdashSIMMS) However sensitivity analysis con-firmed that the results are not affected by includingdeaths coded to tuberculosis and other deathssuggesting that any survival bias must be very small

It is possible that other factors might be associatedwith relapse including short temporary interruptions


of treatment diabetes and malnutrition Informationon these were not collected They may have asignificant impact on risk of relapse but we have noreason to think that they would be associated withsmoking

The sample size- the small number of relapsesmdashmaybe a limitation of this study Some associations espe-cially those related to socioeconomic conditions andalcohol consumption may have been present and nothave been identified because of the power of thestudy However this does not question the validity ofthe associations found

Our findings are relevant for Tuberculosis ControlPrograms Since just to cure patients may not beenough to achieve the control of the disease (as curedpatients can relapse) it is important to incorporate inthe strategies of the Tuberculosis Control Programstrategies to reduce the risk of relapse including

smoke cessation support Finally the effect of homevisits of the FHP confirms that in successful treat-ments of tuberculosis close monitoring of adherenceduring the 6 months of the treatment may helpprevent relapses

AcknowledgementsThe authors wish to thank the Brazilian NationalResearch Council (Conselho Nacional deDesenvolvimento Cientıfico e TecnologicomdashCNPq)British Council and REDE-TB do Brasil for fundingthis study The authors were partially supported byCNPq (scholarship 3059472006-0 to MFPM 3009172006-6 to RAAX and 1343612005-9 to JDLB)

Conflict of interest None declared

References1 World Health Organization The International Union

Against Tuberculosis and Lung Disease The RoyalNetherlands Tuberculosis Association Revised interna-tional definitions in tuberculosis control Int J Tuberc LungDis 20015213ndash15

2 Campos HS Albuquerque MFM Campelo ARL et alO retratamento da tuberculose no municıpio do Recife1997 uma abordagem epidemiologica J Bras Pneumol200026235ndash40

3 Oliveira HB Moreira Filho DC Recidiva em tuberculosee seus fatores de risco Rev Panam Salud Pub 20007232ndash41

4 Maurya V Vijayan VK Shah A Smoking and tubercu-losis an association overlooked Int J Tuberc Lung Dis20026942ndash51

5 Davies PDO Yew WW Ganguly D et al Smoking andtuberculosis the epidemiological association and immuno-pathogenesis Trans R Soc Trop Med Hyg 2006100291ndash98

6 Thomas A Gopi PG Santha T et al Predictors of relapseamong pulmonary tuberculosis patients treated in aDOTS programme in South India Int J Tuberc Lung Dis20059556ndash61

7 Governo do Estado de Pernambuco Secretaria Estadualde Saude Programa de Controle da Tuberculose [home-page on the Internet] Indicadores e informacoes em

saude SINANMSSUS Brasılia DF Ministerio da Saude2002 [cited 2007 Feb 19] Available from httpwwwsaudegovbr

8 Ministerio da Saude Secretaria de Polıticas de SaudeDepartamento de Atencao Basica Manual tecnico parao controle da tuberculose Brasılia DF Ministerio daSaude 2002 Cadernos de Atencao Basica n8 6

9 Migliori GB Hopewell PC Blasi F et al Improving the TBcase management The International Standards forTuberculosis care Eur Respir J 200628687ndash90

10 Santos MAPS Albuquerque MFPM Ximenes RAA et alRisk factors for treatment delay in pulmonary tuber-culosis in Recife Brazil BMC Public Health 2005525

11 Albuquerque MFPM Ximenes RAA Silva NL et al Factorsassociated with treatment failure dropout and death in acohort of tuberculosis patients in Recife PernambucoState Brazil Cad Saude Publica 200723105ndash14

12 Nogueira PA Belluomini M Almeida MMB et al Algumascaracterısticas dos reingressantes ao Sistema Experi-mental de Vigilancia segundo tipo de alta anteriorResumos do IV Congresso Paulista de Saude Publica1993 July 10ndash14 Sao Paulo Brazil

13 Oliveira HB Moreira Filho DC Abandono de tratamentoe recidiva da tuberculose aspectos de episodios previosCampinas SP Brasil 1993ndash1994 Rev Saude Publica200034437ndash43

KEY MESSAGES

Smoking was independently associated with relapse of tuberculosis after successful tuberculosistreatment even when adjusted for socioeconomic factors and alcohol consumption

Our findings provide evidence for incorporating smoking cessation support to the strategies to improveeffectiveness of tuberculosis control programs

Living in an area where the FHP (which provides home visits for routine health supervision) was notimplemented increased the risk of relapse of tuberculosis suggesting close monitoring of adherenceduring the 6 months of the treatment may not only increase successful treatment but also decreasetuberculosis relapses


14 Ormerod LP Prescott RJ Inter-relations betweenrelapses drug regimens and compliance with treatmentin tuberculosis Respir Med 199185239ndash42

15 Menezes AMB Costa JD Goncalves H et al Incidencia efatores de risco para tuberculose em Pelotas uma cidadedo Sul do Brasil Rev Bras Epidemiol 1998150ndash60

16 Tackling poverty in tuberculosis control Editorial Lancet20053662063

17 Costa JG Santos AC Rodrigues LC et al Tuberculosis inSalvador Brazil costs to health system and familiesRev Saude Publica 200539122ndash28

18 Selassie AW Pozsik C Wilson D et al Why pulmonarytuberculosis recurs a population-based epidemiologicalstudy Ann Epidemiol 200515519ndash25

19 Ruffino-Netto A Caron-Ruffino M Interacao de fatoresrisco em tuberculose Rev Saude Publica 19793119ndash22

20 Brown KE Campbell AH Tobacco alcohol and tuber-culosis Br J Dis Chest 196155150ndash58

21 Lowe CR An association between smoking and respira-tory tuberculosis Br Med J 195621081ndash83

22 Doll R Peto R Wheatley K et al Mortality in relation tosmoking 40 yearsrsquo observations on male British doctorsBr Med J 1994309901ndash11

23 Chiang CY Slama K Enarson DA Associations betweentobacco and tuberculosis Int J Tuberc Lung Dis 200711258ndash62

24 Bates MN Khalakdina A Pai M et al Risk of tuber-culosis from exposure to tobacco smoke a systematicreview and meta-analysis Arch Intern Med 2007167335ndash42

25 Alcaide J Altet MN Plans P et al Cigarette smoking as arisk factor for tuberculosis in young adults a case-controlstudy Int J Tuberc Lung Dis 199677112ndash16

26 Boelaert JR Gomes MS Gordeuk VR Smoking iron andTB Lancet 20033621243ndash44

27 Temporao JG Saude Publica e controle do tabagismo noBrasil Cad Saude Publica 200521670ndash71

28 Cavalcante MGS Samico I Frias PG et al Analise deimplantacao das areas estrategicas da atencao basica nasequipes de Saude da Famılia em municıpio de umaRegiao Metropolitana do Nordeste Brasileiro Rev BrasSaude Matern Infant 20066437ndash45

29 Ruffino-Netto A Villa TCS Tuberculosis treatment DOTSimplementation in some regions of Brazil Background andregional features Ribeirao Preto SP Millennium InstiitutendashRede TB 2000

30 Santos J Resposta brasileira ao controle da tuberculoseRev Saude Publica 200741 (Suppl 1)89ndash94

31 Cox HS Morrow M Deutschmann PW Long termefficacy of DOTS regimens for tuberculosis systematicreview Br Med J 2008336484ndash47


Smoking increases the risk of relapse aftersuccessful tuberculosis treatmentJoanna drsquoArc Lyra Batista1 Maria de Fatima Pessoa Militao de Albuquerque12Ricardo Arraes de Alencar Ximenes13 and Laura Cunha Rodrigues4

Accepted 13 May 2008

Background Recent tobacco smoking has been identified as a risk factor fordeveloping tuberculosis and two studies which have investigatedits association with relapse of tuberculosis after completion oftreatment had conflicting results (and did not control for confound-ing) The objective of this study was to investigate risk factors fortuberculosis relapse with emphasis on smoking

Methods A cohort of newly diagnosed TB cases was followed up from theirdischarge after completion of treatment (in 2001ndash2003) untilOctober 2006 and relapses of tuberculosis ascertained during thatperiod A case of relapse was defined as a patient who starteda second treatment during the follow up

Results Smoking (OR 253 95 CI 123ndash521) and living in an area wherethe family health program was not implemented (OR 361 95 CI146ndash893) were found to be independently associated with relapseof tuberculosis

Conclusions Our results establish that smoking is associated with relapse oftuberculosis even after adjustment for the socioeconomic variablesSmoking cessation support should be incorporated in the strategiesto improve effectiveness of Tuberculosis Control Programs

Keywords Tuberculosis control successful treatment relapse risk factorssmoking

IntroductionTuberculosis is a significant health problem worldwidecontrol of tuberculosis relies heavily on identification

and successful treatment of cases Treatment scheduleslast for 6 months One of the problems of tuberculosiscontrol is the fact that cases that were successfullytreated can relapse needing a second treatment Iden-tification of factors that increase the risk of relapseof tuberculosis after cure or completion of treatment isof great interest The monitoring of the effectiveness oftuberculosis programs is based on the routine classifi-cation of the outcome of each tuberculosis treatmentinto (i) treatment failure (patient still have positivesmear 5 months after start of treatment or later)(ii) death (patient who dies for any reason duringthe course of treatment) (iii) default (patient inter-rupts treatment for two consecutive months or more)(iii) cure (patient who is sputum smear negative in two

Corresponding author Centro de Pesquisas Aggeu MagalhaesFIOCRUZ Av Moraes Rego sn Campus da UniversidadeFederal de Pernambuco Cidade Universitaria Recife PECEP 50670ndash420 Brazil E-mail militaocpqamfiocruzbr

1 Universidade Federal de Pernambuco Recife Brazil2 Centro de Pesquisas Aggeu magalhaesFIOCRUZ Recife

Brazil3 Universidade de Pernambuco Recife Brazil4 London School of Hygiene and Tropical Medicine London

UK

The online version of this article has been published under an open access model Users are entitled to use reproduce disseminate or display the open access

version of this article for non-commercial purposes provided that the original authorship is properly and fully attributed the Journal and Oxford University Press

are attributed as the original place of publication with the correct citation details given if an article is subsequently reproduced or disseminated not in its entirety

but only in part or as a derivative work this must be clearly indicated For commercial re-use please contact journalspermissionsoxfordjournalsorg

Published by Oxford University Press on behalf of the International Epidemiological Association

The Author 2008 all rights reserved Advance Access publication 13 June 2008

International Journal of Epidemiology 200837841ndash851

doi101093ijedyn113

841





























Relapse Control


Age (years) 0154 0401



550 5 327 148 9673 078 (026ndash234) 0662 079 (026ndash235) 0668


Sex


Male 23 501 436 9499 089 (045ndash177) 0755 088 (044ndash176) 0732


Alcohol consumption




Cigarette smoking


23 416 530 9584 100 ndash 100 ndash





Social factors

Relapse Control


Literacy

Yes 33 537 582 9463 100 ndash 100 ndash

No 4 417 92 9583 076 (026ndash221) 0624 075 (025ndash217) 0596


Employment

Yes 12 545 208 9455 100 ndash 100 ndash

No 25 511 464 9489 093 (046ndash189) 0850 093 (045ndash189) 0844




55 13 444 280 9556 074 (037ndash148) 0396 073 (037ndash147) 0391



52 5 360 134 9640 100 ndash 100 ndash

lt2 31 614 474 9386 175 (067ndash459) 0254 175 (067ndash461) 0253



Yes 19 521 346 9479 100 ndash 100 ndash

No 18 520 328 9480 099 (051ndash193) 0998 099 (051ndash192) 0986






Piped water

Yes 35 529 627 9471 100 ndash 100 ndash

No 2 444 43 9556 083 (019ndash358) 0806 085 (020ndash369) 0835



Yes 7 486 137 9514 100 ndash 100 ndash

No 29 517 532 9483 106 (045ndash248) 0881 106 (045ndash247) 0890



Yes 3 395 73 9605 100 ndash 100 ndash

No 33 529 591 9471 136 (041ndash454) 0619 135 (040ndash453) 0627



Yes 2 303 64 9697 100 ndash 100 ndash

No 35 552 599 9448 187 (044ndash795) 0397 186 (043ndash794) 0399


(continued)


Table 2 Continued

Social factors

Relapse Control


Ownership of a car

Yes 8 588 128 9412 100 ndash 100 ndash

No 29 507 543 9493 085 (038ndash191) 0702 084 (037ndash189) 0685



Yes 2 244 80 9756 100 ndash 100 ndash

No 35 558 592 9442 236 (056ndash100) 0243 232 (054ndash985) 0255










Relapse Control



Only 1 5 373 129 9627 100 ndash 100 ndash

52 31 550 533 9450 150 (057ndash393) 0409 147 (056ndash388) 0428



Yes 23 518 421 9482 100 ndash 100 ndash

No 14 524 253 9476 101 (051ndash200) 0971 102 (052ndash204) 0935



Yes 7 778 83 9222 100 ndash 100 ndash

No 30 483 591 9517 060 (025ndash141) 0244 060 (025ndash141) 0244



Yes 6 232 253 9768 100 ndash 100 ndash

No 30 699 399 9301 317 (130ndash772) 0011 325 (133ndash794) 0010



HD IV 6 351 165 9649 100 ndash 100 ndash


HD II 7 455 147 9545 131 (043ndash398) 0635 135 (042ndash429) 0609


HD V 11 681 149 9313 203 (073ndash562) 0173 208 (074ndash577) 0160

HD VI 2 606 31 9394 177 (034ndash919) 0495 186 (035ndash979) 0461









Relapse Control



No 23 564 385 9436 100 ndash 100 ndash

Yes 14 475 281 9525 083 (042ndash164) 0602 083 (042ndash165) 0606



No 32 522 581 9478 100 ndash 100 ndash

Yes 05 510 93 9490 097 (037ndash257) 0961 098 (037ndash259) 0974






Treatment delay

460 days 8 301 258 9699 100 ndash 100 ndash

460 days 28 657 398 9343 227 (101ndash505) 0045 223 (099ndash498) 0050



No 1 714 13 9286 100 ndash 100 ndash

Yes 12 538 211 9462 074 (009ndash613) 0780 076 (009ndash640) 0807








Cigarette smoking





No 325 (133ndash794) 0010 361 (146ndash893) 0005 0585















































KEY MESSAGES




















































Relapse Control


Age (years) 0154 0401



550 5 327 148 9673 078 (026ndash234) 0662 079 (026ndash235) 0668


Sex


Male 23 501 436 9499 089 (045ndash177) 0755 088 (044ndash176) 0732


Alcohol consumption




Cigarette smoking


23 416 530 9584 100 ndash 100 ndash





Social factors

Relapse Control


Literacy

Yes 33 537 582 9463 100 ndash 100 ndash

No 4 417 92 9583 076 (026ndash221) 0624 075 (025ndash217) 0596


Employment

Yes 12 545 208 9455 100 ndash 100 ndash

No 25 511 464 9489 093 (046ndash189) 0850 093 (045ndash189) 0844




55 13 444 280 9556 074 (037ndash148) 0396 073 (037ndash147) 0391



52 5 360 134 9640 100 ndash 100 ndash

lt2 31 614 474 9386 175 (067ndash459) 0254 175 (067ndash461) 0253



Yes 19 521 346 9479 100 ndash 100 ndash

No 18 520 328 9480 099 (051ndash193) 0998 099 (051ndash192) 0986






Piped water

Yes 35 529 627 9471 100 ndash 100 ndash

No 2 444 43 9556 083 (019ndash358) 0806 085 (020ndash369) 0835



Yes 7 486 137 9514 100 ndash 100 ndash

No 29 517 532 9483 106 (045ndash248) 0881 106 (045ndash247) 0890



Yes 3 395 73 9605 100 ndash 100 ndash

No 33 529 591 9471 136 (041ndash454) 0619 135 (040ndash453) 0627



Yes 2 303 64 9697 100 ndash 100 ndash

No 35 552 599 9448 187 (044ndash795) 0397 186 (043ndash794) 0399


(continued)


Table 2 Continued

Social factors

Relapse Control


Ownership of a car

Yes 8 588 128 9412 100 ndash 100 ndash

No 29 507 543 9493 085 (038ndash191) 0702 084 (037ndash189) 0685



Yes 2 244 80 9756 100 ndash 100 ndash

No 35 558 592 9442 236 (056ndash100) 0243 232 (054ndash985) 0255










Relapse Control



Only 1 5 373 129 9627 100 ndash 100 ndash

52 31 550 533 9450 150 (057ndash393) 0409 147 (056ndash388) 0428



Yes 23 518 421 9482 100 ndash 100 ndash

No 14 524 253 9476 101 (051ndash200) 0971 102 (052ndash204) 0935



Yes 7 778 83 9222 100 ndash 100 ndash

No 30 483 591 9517 060 (025ndash141) 0244 060 (025ndash141) 0244



Yes 6 232 253 9768 100 ndash 100 ndash

No 30 699 399 9301 317 (130ndash772) 0011 325 (133ndash794) 0010



HD IV 6 351 165 9649 100 ndash 100 ndash


HD II 7 455 147 9545 131 (043ndash398) 0635 135 (042ndash429) 0609


HD V 11 681 149 9313 203 (073ndash562) 0173 208 (074ndash577) 0160

HD VI 2 606 31 9394 177 (034ndash919) 0495 186 (035ndash979) 0461









Relapse Control



No 23 564 385 9436 100 ndash 100 ndash

Yes 14 475 281 9525 083 (042ndash164) 0602 083 (042ndash165) 0606



No 32 522 581 9478 100 ndash 100 ndash

Yes 05 510 93 9490 097 (037ndash257) 0961 098 (037ndash259) 0974






Treatment delay

460 days 8 301 258 9699 100 ndash 100 ndash

460 days 28 657 398 9343 227 (101ndash505) 0045 223 (099ndash498) 0050



No 1 714 13 9286 100 ndash 100 ndash

Yes 12 538 211 9462 074 (009ndash613) 0780 076 (009ndash640) 0807








Cigarette smoking





No 325 (133ndash794) 0010 361 (146ndash893) 0005 0585















































KEY MESSAGES

























Social factors

Relapse Control


Literacy

Yes 33 537 582 9463 100 ndash 100 ndash

No 4 417 92 9583 076 (026ndash221) 0624 075 (025ndash217) 0596


Employment

Yes 12 545 208 9455 100 ndash 100 ndash

No 25 511 464 9489 093 (046ndash189) 0850 093 (045ndash189) 0844




55 13 444 280 9556 074 (037ndash148) 0396 073 (037ndash147) 0391



52 5 360 134 9640 100 ndash 100 ndash

lt2 31 614 474 9386 175 (067ndash459) 0254 175 (067ndash461) 0253



Yes 19 521 346 9479 100 ndash 100 ndash

No 18 520 328 9480 099 (051ndash193) 0998 099 (051ndash192) 0986






Piped water

Yes 35 529 627 9471 100 ndash 100 ndash

No 2 444 43 9556 083 (019ndash358) 0806 085 (020ndash369) 0835



Yes 7 486 137 9514 100 ndash 100 ndash

No 29 517 532 9483 106 (045ndash248) 0881 106 (045ndash247) 0890



Yes 3 395 73 9605 100 ndash 100 ndash

No 33 529 591 9471 136 (041ndash454) 0619 135 (040ndash453) 0627



Yes 2 303 64 9697 100 ndash 100 ndash

No 35 552 599 9448 187 (044ndash795) 0397 186 (043ndash794) 0399


(continued)


Table 2 Continued

Social factors

Relapse Control


Ownership of a car

Yes 8 588 128 9412 100 ndash 100 ndash

No 29 507 543 9493 085 (038ndash191) 0702 084 (037ndash189) 0685



Yes 2 244 80 9756 100 ndash 100 ndash

No 35 558 592 9442 236 (056ndash100) 0243 232 (054ndash985) 0255










Relapse Control



Only 1 5 373 129 9627 100 ndash 100 ndash

52 31 550 533 9450 150 (057ndash393) 0409 147 (056ndash388) 0428



Yes 23 518 421 9482 100 ndash 100 ndash

No 14 524 253 9476 101 (051ndash200) 0971 102 (052ndash204) 0935



Yes 7 778 83 9222 100 ndash 100 ndash

No 30 483 591 9517 060 (025ndash141) 0244 060 (025ndash141) 0244



Yes 6 232 253 9768 100 ndash 100 ndash

No 30 699 399 9301 317 (130ndash772) 0011 325 (133ndash794) 0010



HD IV 6 351 165 9649 100 ndash 100 ndash


HD II 7 455 147 9545 131 (043ndash398) 0635 135 (042ndash429) 0609


HD V 11 681 149 9313 203 (073ndash562) 0173 208 (074ndash577) 0160

HD VI 2 606 31 9394 177 (034ndash919) 0495 186 (035ndash979) 0461









Relapse Control



No 23 564 385 9436 100 ndash 100 ndash

Yes 14 475 281 9525 083 (042ndash164) 0602 083 (042ndash165) 0606



No 32 522 581 9478 100 ndash 100 ndash

Yes 05 510 93 9490 097 (037ndash257) 0961 098 (037ndash259) 0974






Treatment delay

460 days 8 301 258 9699 100 ndash 100 ndash

460 days 28 657 398 9343 227 (101ndash505) 0045 223 (099ndash498) 0050



No 1 714 13 9286 100 ndash 100 ndash

Yes 12 538 211 9462 074 (009ndash613) 0780 076 (009ndash640) 0807








Cigarette smoking





No 325 (133ndash794) 0010 361 (146ndash893) 0005 0585















































KEY MESSAGES
























Table 2 Continued

Social factors

Relapse Control


Ownership of a car

Yes 8 588 128 9412 100 ndash 100 ndash

No 29 507 543 9493 085 (038ndash191) 0702 084 (037ndash189) 0685



Yes 2 244 80 9756 100 ndash 100 ndash

No 35 558 592 9442 236 (056ndash100) 0243 232 (054ndash985) 0255










Relapse Control



Only 1 5 373 129 9627 100 ndash 100 ndash

52 31 550 533 9450 150 (057ndash393) 0409 147 (056ndash388) 0428



Yes 23 518 421 9482 100 ndash 100 ndash

No 14 524 253 9476 101 (051ndash200) 0971 102 (052ndash204) 0935



Yes 7 778 83 9222 100 ndash 100 ndash

No 30 483 591 9517 060 (025ndash141) 0244 060 (025ndash141) 0244



Yes 6 232 253 9768 100 ndash 100 ndash

No 30 699 399 9301 317 (130ndash772) 0011 325 (133ndash794) 0010



HD IV 6 351 165 9649 100 ndash 100 ndash


HD II 7 455 147 9545 131 (043ndash398) 0635 135 (042ndash429) 0609


HD V 11 681 149 9313 203 (073ndash562) 0173 208 (074ndash577) 0160

HD VI 2 606 31 9394 177 (034ndash919) 0495 186 (035ndash979) 0461









Relapse Control



No 23 564 385 9436 100 ndash 100 ndash

Yes 14 475 281 9525 083 (042ndash164) 0602 083 (042ndash165) 0606



No 32 522 581 9478 100 ndash 100 ndash

Yes 05 510 93 9490 097 (037ndash257) 0961 098 (037ndash259) 0974






Treatment delay

460 days 8 301 258 9699 100 ndash 100 ndash

460 days 28 657 398 9343 227 (101ndash505) 0045 223 (099ndash498) 0050



No 1 714 13 9286 100 ndash 100 ndash

Yes 12 538 211 9462 074 (009ndash613) 0780 076 (009ndash640) 0807








Cigarette smoking





No 325 (133ndash794) 0010 361 (146ndash893) 0005 0585















































KEY MESSAGES






























Relapse Control



No 23 564 385 9436 100 ndash 100 ndash

Yes 14 475 281 9525 083 (042ndash164) 0602 083 (042ndash165) 0606



No 32 522 581 9478 100 ndash 100 ndash

Yes 05 510 93 9490 097 (037ndash257) 0961 098 (037ndash259) 0974






Treatment delay

460 days 8 301 258 9699 100 ndash 100 ndash

460 days 28 657 398 9343 227 (101ndash505) 0045 223 (099ndash498) 0050



No 1 714 13 9286 100 ndash 100 ndash

Yes 12 538 211 9462 074 (009ndash613) 0780 076 (009ndash640) 0807








Cigarette smoking





No 325 (133ndash794) 0010 361 (146ndash893) 0005 0585















































KEY MESSAGES





































































KEY MESSAGES
























d’Arc Lyra Batista, J; de Ftima Pessoa Milito de...

Documents

Transcript of d’Arc Lyra Batista, J; de Ftima Pessoa Milito de...