Recurrent tuberculosis in Finland 1995-2013: a clinical and epidemiological cohort study

RESEARCH ARTICLE Open Access Recu en ube culosis in Finland 1995–2013: a clinical and epidemiological coho s udy Vi e Ko honen 1,2,3* , Hanna Soini 1 , Tuula Vasanka i 4,5 , Jukka Ollg en 1 , Pie e W. Smi 1,6 and Pe i Ruu u 1 Abs ac Backg ound: We in es iga ed he epidemiology and p e alence o po en ial isk ac o s o ube culosis (TB) ecu ence in a popula ion-based egis y coho o 8084 TB cases be ween 1995 and 2013. Me hods: An episode o ecu en TB was de ined as a case e- egis e ed in he Na ional In ec ious Disease Regis e a leas 360 days om he da e o he ini ial egis a ion. A eg ession model was used o es ima e isk ac o s o ecu ence in he na ional coho . To desc ibe he p esence o known isk ac o s o ecu ence, pa ien eco ds o he ecu en cases we e e iewed o TB diagnosis con i ma ion, po en ial ac o s a ec ing he isk o ecu ence, he ea men egimens gi en and he ou comes o he TB episodes p eceding he ecu ence. Resul s: TB egis y da a included 84 pa ien s, o whom mo e han 1 TB episode had been egis e ed. A e a ca e ul clinical e iew, 50 ecu en TB cases (0.6%) we e iden i ied. The o e all incidence o ecu ence was 113 cases pe 100,000 pe son-yea s o e a median ollow up o 6.1 yea s. Fo he i s 2 yea s, he incidence o ecu ence was o e 200/100000. In mul i a ia e analysis o he na ional coho , younge age emained an independen isk ac o a all ime poin s, and male gende and pulmona y TB a 18 yea s o ollow-up. Among he 50 ecu en cases, 35 pa ien s (70%) had ecei ed adequa e ea men o he i s episode; in 12 cases (24%) he ea ing physician and in wo cases (4%) he pa ien had discon inued ea men p ema u ely. In one case (2%) he ea men ou come could no be assessed. Conclusions: In Finland, he a e o ecu en TB was low despi e no sys ema ic di ec ly obse ed he apy. The i s 2 yea s a e a TB episode had he highes isk o ecu ence. Among he ecu en cases, he obse ed p ema u e discon inua ion o ea men in he i s episode in nea ly one ou h o he ecu en cases calls o imp o ed aining o he physicians. Keywo ds: Tube culosis ecu ence, Tube culosis epidemiology, Tube culosis ea men , Tube culosis Backg ound Tube culosis (TB) emains a majo global heal h p oblem wi h es ima ed 10.4 million new TB cases wo ldwide in 2015. In 2013, 0.3 million TB cases we e epo ed as ecu en [1]. A e success ul ea - men , ecu en TB is es ima ed o occu in 0–14% o all TB pa ien s wi hin 1–3 yea s [2]. Recu ence o TB ollowing ea men o an ini ial disease episode can occu due o endogenous e-ac i a ion wi h he same s ain o Mycobac e ium ube culosis ( elapse) o exogenous in ec ion wi h a new s ain ( e-in ec ion). In low-incidence coun ies, ecu ence a es ha e a ied be ween 0.4% and in a p ospec i e clinical ial up o 6% [3–5]. The p opo ion due o e-in ec ion has been epo ed o a y be ween 4 and 27% [3, 4]. In high-incidence coun ies he majo i y o ecu en cases, up o 77%, a e caused by e-in ec ion [6]. Finland is a low-TB-incidence (<10/100000) coun y since 2001, and in 2015 TB incidence was 5/100000 [7]. In 2015, 1% o TB cases had HIV in ec ion. Howe e , eme ging challenges o he TB con ol p og am include g adually inc easing esis ance o M. ube culosis, wi h 3% o all isola es mul i d ug esis an (MDR) in 2015 [7], concomi an ly wi h a apid inc ease in he p opo ion o TB cases occu ing in immig an s [8]. Finland did no implemen a comp ehensi e DOT (di ec ly obse ed he apy) s a egy in pa ien managemen un il 2013 [9]. * Co espondence: [email p o ec ed] 1 Depa men o Heal h Secu i y, Na ional Ins i u e o Heal h and Wel a e, Helsinki, Finland 2 Depa men o Pulmona y Diseases, Tampe e Uni e si y Hospi al, Tampe e, Finland Full lis o au ho in o ma ion is a ailable a he end o he a icle © The Au ho (s). 2017 Open Access This a icle is dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0 In e na ional License (h p://c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e Commons license, and indica e i changes we e made. The C ea i e Commons Public Domain Dedica ion wai e (h p://c ea i ecommons.o g/publicdomain/ze o/1.0/) applies o he da a made a ailable in his a icle, unless o he wise s a ed. Ko honen e al. BMC In ec ious Diseases (2017) 17:721 DOI 10.1186/s12879-017-2818-6 Ou p e ious s udy shows, ha in Finland mo e han 80% o ecu en cases du ing 1995–2013 we e elapses [10]. The aim o he p esen s udy was o in es iga e in a na ional, popula ion-based TB coho he occu ence o ecu en TB and a ew po en ial ac o s a ec ing he isk o ecu ence in Finland du ing he yea s 1995–2013. In ecu en cases, we desc ibe ea men egimens adminis- e ed and ea men ou come in he i s episode, and o he po en ial ac o s a ec ing he isk o ecu ence, in o de o s eng hen he TB ea men p og am in he changing epidemiologic en i onmen . Me hods Su eillance sys em and s udy popula ion The s udy popula ion consis ed o all TB cases in Finland epo ed om Janua y 1, 1995, o Decembe 31, 2013, o he Na ional In ec ious Disease Regis e (NIDR), main- ained a he Na ional Ins i u e o Heal h and Wel a e (THL). Clinical mic obiology labo a o ies manda o ily no i y new M. ube culosis isola ions o NIDR and submi isola es o he Mycobac e ial Re e ence Labo a o y (NRL) a THL o d ug suscep ibili y es ing (e hambu ol, isoniazid, py azinamide, i ampicin and s ep omycin). Physicians manda o ily no i y o NIDR labo a o y- con i med cases o TB: he labo a o y epo o a posi i e es esul o he clinician au oma ically includes a eminde o no i y he case. Since 2007, also clinically diagnosed TB cases, when a decision o gi e a ull cou se o TB ea men is made, a e no i ied. In o ma ion on HIV posi i i y is ob ained by linking da a wi hin NIDR. Da a on he coun y o o igin and he da e o dea h we e e ie ed om he na ional popula ion egis y. Da a om he di e en sou ces a e au oma ically linked as a case by a unique pe son iden i ying numbe . Case de ini ions and da a collec ion o he subg oup o ecu en cases An episode o ecu en TB was de ined as a case e- egis e ed in NIDR a leas 360 days om he da e o he ini ial egis a ion o a TB in ec ion episode. Fo he cases who had a ecu en episode in he egis e , da a on ana omical si e o disease (pulmona y/ex apulmona y), adiological, his ological and mic obiological esul s, HIV es esul s, subs ance abuse, he d ug egimen in he i s episode and ad e se e ec s we e ex ac ed om pa ien cha s. In pulmona y TB cases, also spu um smea and cul u e esul s a mon hs 0 and 2, and a he end o he ea men we e ob ained. Based on ca e ul e iew o hese da a, a numbe o cases we e excluded om u he analysis as ecu en cases, as hey did no mee he c i e ia o ecu en TB (Fig. 1). In cul u e nega i e episodes, he diagnoses we e based on clinical c i e ia [11], including adiological indings in combina ion wi h ei he his ological con i ma ion o posi i e nucleic acid ampli ica ion es esul s, excep o one case in which he diagnosis was only clinical and adiological. Managemen o he TB episode p eceding ecu ence The ea men egimen, ee o cha ge o he pa ien , adminis e ed in he i s TB episode was g ouped in o six ca ego ies, based on he na ional ecommenda ions: un il 2006 as desc ibed ea lie [12, 13], and om 2006 in he Na ional TB Con ol P og am [9]. The ea men egimens and he ou comes o he i s TB episodes o he ecu en cases, including cul u e nega i e and ex a- pulmona y cases, we e assessed by wo clinical TB expe s, and classi ied [11] as cu ed, comple ed ea - men , los o ollow-up and no e alua ed (no a al cases no ailu es). ‘Los o ollow-up’was u he di ided in o subg oups: physician’s decision o s op p ema u ely (including cases ha ecei ed ine ec i e ea men ), and de aul ed (in e up ion due o he pa ien ). G oup ‘no e alua ed’was u he di ided in o subg oups: s ill on ea men a 12 mon hs, and ea men ou come no assessed. In wo cases he de ails o ea men in he i s episode we e no a ailable, bu in one o hese cases he ea men ou come could be assessed. S a is ical analysis To calcula e he incidence o ecu en TB among all epo ed cases o he na ional coho , ollow-up ime in days was calcula ed o all cases in he na ional TB coho om 360 days a e hey we e no i ied un il an e en ( e-no i ica ion), dea h o censo ed (Decembe 31, 2013). The da e o dea h was acqui ed by linkage o popula ion egis e using pe son iden i ie . The eg es- sion model employed, uses pseudo-obse a ions o model censo ed da a [14] wi h S a a e sion 14.02 (S a aCo p LLC, 4905 Lakeway D i e, Collage S a ion, TX 77845 USA) o es ima e e ec s ( hei ela i e isks) o isk ac o s. The censo ing was assumed o be inde- penden , condi ional on he co a ia es [15]. Gende and ana omical si e o disease we e allowed o ha e ime dependen e ec s, and p edic i e ma gins we e calcu- la ed o es ima e cumula i e isk di e ences be ween gende s and ana omical si es o disease, espec i ely, adjus ing o o he ac o s in he model. Resul s we e quali a i ely checked using he ex ended ime dependen Cox model. All he explana o y a iables wi h uni a ia e p- alues <0.2 we e included in he mul i a ia e model, and included gende , ana omical si e o disease and age. Cause speci ic cumula i e isks o ime poin s 1 yea , 2 yea s and 18 yea s we e calcula ed, as he chosen ea ly ime poin s had mos ecu en e en s o explana o y a iables, and i has been p e iously epo ed ha he i s 1–2 yea s ha e he highes haza d o ecu ence [3, 16]. The maximum su eillance ime poin se a 18 yea s gi es he inal o e all di e ence es ima e o Ko honen e al. BMC In ec ious Diseases (2017) 17:721 Page 2 o 7 he cumula i e isk o ecu ence. The dis ibu ion o con inuous a iables be ween g oups was compa ed using Wilcoxon ank-sum es . E hics The e hics app o al o he s udy was gi en by he E hics Commi ee o Tampe e Uni e si y Hospi al, Tampe e, Finland. Resul s A o al o 8084 TB cases we e egis e ed in Finland du ing he s udy pe iod (Table 1): 43% we e emale, and 13.6% o o eign o igin (inc eased om 4.8% in 1995 o 32.2% in 2013). The median age was 70 yea s (in e qua ile ange [IQR], 56–79 yea s) o Finnish-bo n and 30 yea s (IQR 23–40 yea s, p< 0,001) o o eign-bo n cases. Cha ac e is ics and incidence o ecu en TB A e a ca e ul e iew o he 84 cases wi h mo e han one episode egis e ed, 50 TB cases (0.6% o all cases in he coho ) we e classi ied as ecu en (Fig. 1). The mean o e all incidence o ecu ence was 112.9 (95% con idence in e al [CI], 85.6–148.9); o he i s yea o ollow-up he o e all incidence was 236.4 (95%CI 140.0–399.2) and o he second yea o ollow-up 206.7 (95%CI 114.5– 373.2) pe 100,000 pe son-yea s. Ou o he 50 ecu en cases, wo had h ee disease episodes and 48 wo disease episodes. Fo y cases we e cul u e posi i e in all episodes. The median age was 51.5 yea s ( ange 6–95 yea s) a he egis a ion o he i s episode; ele en cases (22%) we e emale. In he i s episode 44 cases (88%) and in he sec- ond episode 39 cases (78%) we e classi ied as pulmona y TB. Two cases we e HIV posi i e; o 64% o cases HIV had no been es ed. Nine ecu en cases (18%) we e o o eign o igin. A his o y o subs ance abuse, mos ly alco- hol, was egis e ed in he pa ien eco ds in a leas 1 TB episode o 59% o males and none o emales; 49% o males had subs ance abuse eco ded in bo h episodes. Managemen o he TB episodes p eceding ecu ence Among he 48 ecu en cases (96%) wi h comple e pa ien eco ds, 36 cases (75%) ecei ed s anda d ea men in he i s episode (Table 2). Among hese, he du a ion o Fig. 1 S eps in iden i ying ecu en cases o ube culosis, Finland, 1995–2013. a Non- ube culous mycobac e ium egis e ed as M. ube culosis (N= 2), egis a ion o alse iden i y (N= 1) o inco ec egis a ion da e by he no i ie (N= 10). b Only one long con inuous TB episode (N= 5), o ea ing physician decided o gi e TB ea men wi hou speci ic e idence o TB and in subsequen expe assessmen by he s udy eam he e was ano he mo e likely cause o he disease (N=9) Table 1 Dis ibu ion o demog aphic and po en ial isk ac o s o ecu en ube culosis in a na ional TB coho o 8084 cases, Finland 1995–2013 Va iable TB ecu ence (n= 50) No TB ecu ence (n= 8034) All (n= 8084) Median age, yea s 51, 5 66 66 Gende emale n(%) 11 (22%) 3457 (43%) 3468 (42.9%) Fo eign o igin n(%) 9 (18%) 1066 (13.6%) a 1074 (13.6%) b Pulmona y si e o disease n(%) 44 (88%) c 5599 (69.2%) 5605 (69.3%) Cul u e posi i e n(%) 48 (96%) c 6631 (82.5%) 6680 (82.6%) P io o yea 2007 n(%) 45 (90%) c 5762 (71.7%) 5807 (71.8%) a O igin known o 7862 cases b O igin known o 7912 cases c Da a o he i s episode Ko honen e al. BMC In ec ious Diseases (2017) 17:721 Page 3 o 7 ea men was sho (<5.5 mon hs) in six cases, and he in- ensi e phase (<54 days) in h ee cases. Twel e cases (25%) ecei ed non-s anda d ea men due o d ug esis ance, ad e se e ec s, o he ea ing physician’s decision. O hese, six cases ecei ed TB ea men egimens ha we e assessed as p obably e ec i e, one o hem o a oo sho ime pe iod. T ea men egimens o six cases we e assessed as p obably ine ec i e. Six cases (12.5%) ecei ed di ec ly obse ed he apy (DOT). Among he 49 cases wi h ea - men ou come, i was success ul in 27 cases (55.1%). In 12 cases (24.5%) he ea ing physician had s opped he ea - men p ema u ely, and in wo cases (4.1%) in e up ion was due o he pa ien . Eigh cases (16.3%) we e s ill on ea - men a 12 mon hs (all inally comple ed ea men ). The e we e no ecu en cases wi h an MDR isola e in ei he episode. Fi e cases we e ini ially in ec ed wi h isoniazid- esis an isola es, bu in h ee o hese cases, he ea men egimen was no modi ied acco dingly. In wo cases, addi ional esis ance o s ep omycin and in one case esis ance o py azinamide de eloped du ing ea men . In one case, ini ially ully suscep ible isola e de eloped esis ance o py azinamide. Risk ac o s o ecu ence in he na ional coho The median ollow-up ime o cases in he coho o 8084 TB cases was 6.1 yea s (IQR 2.7–11.1 yea s). The ecu ence occu ed wi hin less han 2 yea s in 25 (50%), wo o less han 4 yea s in 8 (16%), and la e in 17 cases (34%) (Fig. 2a). No ecu ences occu ed in emales and o ex apulmona y cases a e he i s 2 yea s (Fig. 2b and c). In uni a ia e analysis o a iables a ailable o he na ional coho , he cumula i e isks o ecu ence be ween males and emales, and be ween pulmona y and ex apulmona y TB did no di e s a is ically signi ican ly a 1 and 2 yea s o ollow-up (Table 3). Howe e , a 18 yea s o ollow-up, he cumula i e isk o males was nea ly ou old compa ed o emales (Fig. 2b), and mo e han i e old o pulmona y TB com- pa ed o ex apulmona y TB (Fig. 2c). The isk o ecu - ence dec eased wi h e e y addi ional 10 yea s o age (Table 3). When only cases ha we e cul u e posi i e in all episodes we e included, he ecu ence a e was simi- la o ha seen in he whole ecu en cases coho (Fig. 2d). Whe he he i s episode occu ed p io o e sus a e 2007 did no ha e a signi ican associa ion wi h he isk o ecu ence (Table 3). In he mul i a ia e analysis (Table 3) younge age emained an independen isk ac o o cumula i e isk o ecu ence a all ime poin s, and male gende and pulmona y TB a 18 yea s o ollow-up. Discussion We in es iga ed he epidemiology and he p e alence o isk ac o s associa ed wi h ecu ence o TB in Finland in a comp ehensi e na ional, egis e -based coho o 8084 TB cases om 1995 o 2013, and ound in a ollow-up o up o 18 yea s ha 0.6% o cases du ing he s udy pe iod we e ecu en . The o e all incidence o ecu ence is 10–20 imes highe , and o he i s 2 yea s o ollow-up 20–40 imes highe han he inci- dence o TB in he gene al popula ion in Finland du ing he s udy pe iod [7]. Pa ien cha e iew o he ecu - en cases e ealed ha in nea ly one ou h o he ecu en cases, he physician had discon inued he ea men o he i s episode p ema u ely. In egis e -based s udies on ecu en TB, alida ion o he da a equi es conside able e o o ensu e da a Table 2 Dis ibu ion o ea men egimens in he i s episodes o 50 TB cases wi h a ecu ence T ea men g oup To al in g oup In ensi e phase sho In ensi e phase adequa e Du a ion o ea men sho Du a ion o ea men adequa e S anda d ea men A a 15 NA 15 1 14 S anda d ea men B b 4NA4 2 2 S anda d ea men wi h sho in ensi e phase C c 33 NA2 1 S anda d ea men D d 14 0 14 1 13 O he p obably e ec i e combina ion o an i-TB d ugs e 6NANA1 5 O he p obably ine ec i e combina ion o an i-TB d ugs 6NANANA NA No e alua ed 2 ––– – NA no applicable, Hisoniazid, RRi ampicin, Zpy azinamide, Ee hambu ol a HRZ in in ensi e phase, HR in con inua ion phase, adequa e du a ion o ea men ≥5.5 mon hs b HRE in in ensi e phase, HR in con inua ion phase, adequa e du a ion o ea men ≥8 mon hs c Sho in ensi e phase <54 days in s anda d ea men A o B d ≥4 an i-TB d ugs, including HRZ (adequa e du a ion o ea men ≥5.5 mon hs) o HRE (adequa e du a ion o ea men ≥8 mon hs) e Non-s anda d combina ions guided by d ug esis ance o due o ad e se e ec s, he adequacy o ea men du a ion assessed by he s udy g oup D ug esis ance igno ed o inapp op ia e dosing Re e ences [9,12,13] Ko honen e al. BMC In ec ious Diseases (2017) 17:721 Page 4 o 7 quali y, as demons a ed in ou s udy using a egis e wi h p o en high sensi i i y and speci ici y o TB [17], a e which au oma ed no i ica ion om he labo a o ies and manda o y eminde s om he labo a o y o he clinician on he need o no i y we e in oduced. We obse ed ha ou o he 84 cases ini ially iden i ied in he egis e da a as ecu en , only 50 cases we e uly ecu en . Majo easons o inaccu acies in ou egis e da a included inco ec no i ica ion da es, and no i ying clinical TB cases wi hou mic obiological con i ma ion, which he cha e iew e ealed o be inco ec . Wi hou alida ion o he da a o ecu en TB cases, he p opo ion o ecu en a b c d Fig. 2 Cumula i e isk o ecu ence o TB by ollow-up ime (and i s poin wise 95% con idence limi s). aO e all coho ; (b) By gende ; (c)By ana omical si e o disease; (d) Only cul u e posi i e TB Table 3 Uni a ia e and mul i a ia e analysis o isk ac o s o TB ecu ence in a na ional coho o TB cases in Finland, 1995–2013 Va iables Uni a ia e RR 95% CI pMul i a ia e RR 95% CI p Males a 1 yea a 1,43 0,48–4,26 0,52 1,93 0,40–9,31 0,41 Males a 2 yea s a 1,18 0,53–2,62 0,69 1,39 0,44–4,38 0,58 Males a 18 yea s a 3,92 1,95–7,91 < 0,001 5,88 2,21–15,66 < 0,001 Pulmona y TB a 1 yea b 1,72 0,48–6,19 0,40 4,13 0,58–29,67 0,16 Pulmona y TB a 2 yea s b 1,87 0,70–4,99 0,21 2,00 0,43–9,33 0,38 Pulmona y TB a 18 yea s b 5,54 2,17–14,14 < 0,001 15,15 4,98–46,08 < 0,001 Age + 10 yea s 0,87 0,77–0,98 0,03 0,83 0,70–0,99 0,04 Finnish o igin 0,64 0,31–1,33 0,23 ––– 1.episode be o e 2007 1,33 0,81–2,21 0,26 ––– a Re e ence emales. O e all p- alue o male gende in uni a ia e analysis < 0,0001 and in mul i a ia e analysis 0,0015 b Re e ence ex apulmona y TB. O e all p- alue o pulmona y TB < 0,0001 Ko honen e al. BMC In ec ious Diseases (2017) 17:721 Page 5 o 7 cases ou o he o al coho in ou s udy would ha e been almos double, s essing he need o igid pa ien cha e iew when assessing TB ecu ence. O e 80% o ecu en cases in Finland a e elapses o he p e ious in ec ion [8], as elsewhe e in low-incidence coun ies [3, 18, 19]. In egis e -based in es iga ions o ecu en TB in low incidence-coun ies, ca e ul e iew o he ac ual ea men in he i s TB pe iod o ecu - en cases has a ely been epo ed [20, 21]. We obse ed ha app oxima ely wo hi ds, including hose who ecei ed ea men las ing o e 12 mon hs, had ecei ed an adequa e ea men . Jus o e one hal among he ecu en cases had in he i s episode a success ul ou come acco ding o WHO c i e ia [11]. Impo an o aining policy was he inding ha clea ly mo e equen ly han in e up ion due o he pa ien , he ea ing physician had discon inued ea men p e- ma u ely, as desc ibed in ou p e ious epo [12]. Inad- equa e ea men s we e caused by he p esence o d ug esis ance wi hou app op ia e ea men egimen modi- ica ion, absence o app op ia e ex ension o du a ion o ea men when ea men was modi ied, ad e se e ec s, o he physician’s decision o s op wi hou he eason being documen ed in he pa ien eco d. Theo e allincidenceo ecu encein hena ional coho , wi h a median ollow-up pe iod o 6 yea s, and up o 18 yea s, was 113/100000, 10–20 imes highe han o he gene al popula ion in Finland, in line wi h long- e m ollow-up in low-incidence coun- ies (71–410/100000) [3, 16]. Fo he i s yea o ollow-up, s a ing a 12 mon hs om he egis a ion o he i s episode, he incidence o ecu ence was 236/100000, o he same magni ude as in ecen s udies om Aus alia [3] and Denma k [18], bu clea ly lowe han in a numbe o ea lie s udies om indus ialised coun ies [22]. Ou obse a ion in he na ional coho o male gende as a isk ac o o ecu ence is in line wi h some p e ious epo s om low-incidence coun ies [23–25], bu his inding has been inconsis en [20, 21]. Subs ance abuse da a is no collec ed in he NIDR o he na ional coho , bu we ound in nea ly 60% o he ecu en male cases a his o y o subs ance abuse in pa ien cha s, bu none in emales, which could con ibu e o he excess isk seen in males. An associa ion be ween ea men adhe ence and alcoholism has been epo ed in ecu en TB in he USA [20]. We ound in he na ional coho he isk o ecu - ence highe o pulmona y han o ex apulmona y TB, in line wi h ea lie s udies [16, 23, 26]. Mo e han 40% o ou ecu en cases (da a no shown) had bo h pulmona y and ex apulmona y in ec ion in he i s episode, which has been epo ed o be a isk ac o o ecu ences [25]. An unexpec ed inding in he na ional coho was ha he isk o ecu ence was associa ed wi h younge age, whe eas in wo ea lie s udies, age > 65 yea s [20] o age be ween 25 and 64 yea s [23] ha e been epo ed as isk ac o s o ecu ence. In ea lie s udies, ei he being an immig an [16, 20, 25] o being bo ne in he coun y [21] we e epo ed as isk ac o s, while o igin was no a isk ac o o ecu ence in ou na ional coho s udy. Limi a ions o ou coho s udy include he ac ha we may ail o iden i y some ecu ences, as ecu ences be o e 360 days om he da e o he ini ial episode (ea ly ecu ences) we e no analysed om he egis e da a. The s anda d cu -o ime ecommended by WHO, a which ea men ou come is eco ded, is 12 mon hs [11]. The e- o e, we chose his imepoin as a cu -o o ecu ence, in line wi h eg a la ge UK coho [16]. The p opo ion o ea ly ecu ences in e ospec i e s udies is small [3, 18]. In addi ion, in e ospec i e s udies, i may be di icul o dis inguish be ween ea men ailu es and ea ly ecu - ences as spu um samples a e no sys ema ically collec ed du ing ea men , and in ou s udy we also included cul u e nega i e and ex apulmona y cases. Almos one hi d o ou ecu en cases do no mee he WHO ea - men egimen desc ip ion and ou come c i e ia o a ecu en case [11]. Howe e , he ca e ul alida ion p ocess o ou egis e da a demons a es ha he same challenges a e likely o be p esen in o he egis e -based s udies, unless ca e ul alida ion has been pe o med. As a coun y wi h e y low incidence o HIV [7], he absence o sys ema ic es ing o all TB cases o HIV is unlikely o in oduce a bias in he isk analysis. The obse a ions on he sho comings o ea men among he i s episodes o he ecu en cases a e impo an o guiding aining and sys em de elopmen o he in eg a ed TB con ol p og am. In 2013–2015, ea men ou come in Finland was success ul (cu e o comple ed ea men ) in 75–78% o pulmona y TB cases [7], as in he Eu opean Region on a e age [27]. Conclusions In he absence o a comp ehensi e DOT s a egy, he a e o TB ecu ence was ound o be low in Finland. An impo an inding was ha in one ou h o he ecu en cases, he physician had discon inued he ea men p e- ma u ely, which implies ha aining o physicians needs o be imp o ed and, as TB becomes a e, ea men should possibly be p o ided in ewe cen e s. The i s 2 yea s a e a TB episode is a e y high- isk pe iod o ecu ence: his could be inco po a ed as an au oma ed high- isk signal in he de eloping in eg a ed elec onic pa ien managemen sys ems o educing he delays in implemen ing TB diagnos ics and ea men . Abb e ia ions CI: Con idence in e al; DOT: Di ec ly obse ed he apy; HIV: Human immunode iciency i us; IQR: In e qua ile ange; MDR: Mul i d ug esis an ; Ko honen e al. BMC In ec ious Diseases (2017) 17:721 Page 6 o 7 NIDR: Na ional in ec ious disease egis e ; NRL: Mycobac e ial e e ence labo a o y; TB: Tube culosis; THL: Na ional Ins i u e o Heal h and Wel a e Acknowledgemen s We hank Pi jo Tu iainen, Jan-E ik Lö lund and Teemu Mö önen a he Na ional Ins i u e o Heal h and Wel a e o echnical assis ance and P o esso Lau i Leh imäki a he School o Medicine, Uni e si y o Tampe e o c i ically e iewing his a icle. Funding The s udy was suppo ed by he Tampe e Tube culosis Founda ion, Väinö and Laina Ki i Founda ion and The Resea ch Founda ion o he Pulmona y Diseases. The unde s had no ole in s udy design, da a collec ion and analysis, decision o publish, o p epa a ion o he manusc ip . A ailabili y o da a and ma e ials All da a gene a ed o analysed du ing his s udy a e included in his published a icle. Au ho s’con ibu ions VK pa icipa ed in designing he s udy, collec ed he clinical da a, and was in ol ed in analysing he da a and w i ing he manusc ip . HS pa icipa ed in designing he s udy, supe ised he labo a o y wo k and was in ol ed in analysing he da a and w i ing he manusc ip . TV pa icipa ed in designing he s udy, and was in ol ed in analysing especially he clinical da a and w i ing he manusc ip . JO pa icipa ed in designing he s udy and pe o med he s a is ical analysis. PS pa icipa ed in designing he s udy and w i ing he manusc ip . PR pa icipa ed in designing he s udy, and was in ol ed in analysing he da a and had a majo ole in w i ing he manusc ip . All au ho s ead and app o ed he inal manusc ip . E hics app o al and consen o pa icipa e The e hics app o al o he s udy was gi en by he E hics Commi ee o Tampe e Uni e si y Hospi al, Tampe e, Finland. Consen o pa icipa e: No applicable. Consen o publica ion No applicable. Compe ing in e es s The au ho s decla e ha hey ha e no compe ing in e es s. Publishe ’sNo e Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in published maps and ins i u ional a ilia ions. Au ho de ails 1 Depa men o Heal h Secu i y, Na ional Ins i u e o Heal h and Wel a e, Helsinki, Finland. 2 Depa men o Pulmona y Diseases, Tampe e Uni e si y Hospi al, Tampe e, Finland. 3 School o Medicine, Uni e si y o Tampe e, 33014 Tampe e, Finland. 4 Finnish Lung Heal h Associa ion (Filha), Helsinki, Finland. 5 Facul y o Medicine, Uni e si y o Tu ku, Tu ku, Finland. 6 Depa men o in ec ious diseases, Public heal h labo a o y, GGD Ams e dam, Ams e dam, The Ne he lands. Recei ed: 13 Ap il 2017 Accep ed: 5 No embe 2017 Re e ences 1. Wo ld Heal h O ganisa ion. Global ube culosis epo 2015. Gene a: WHO; 2015. 2. Cox HS, Mo ow M, Deu schmann PW. Long e m e icacy o DOTS egimens o ube culosis: sys ema ic e iew. BMJ. 2008;336:484–7. 3. Doble CC, C aw o d ABH, Jel s PJ, Gilbe GL, Ma ks GB. Recu ence o ube culosis in a low-incidence se ing. Eu Respi J. 2009;33:160–7. 4. Jasme RM, Bozeman L, Schwa zman K, Ca e MD, Saukkonen JJ, Me chock B, e al. Recu en ube culosis in he Uni ed S a es and Canada: elapse o ein ec ion? Am J Respi C i Ca e Med. 2004;170:1360–6. 5. Kim L, Moonan PK, Yelk Wood u RS, Kamme e JS, Haddad MB. Epidemiology o ecu en ube culosis in he Uni ed S a es, 1993-2010. In J Tube c Lung Dis. 2013;17:357–60. 6. Ve e S, Wa en RM, Beye s N, Richa dson M, an de Spuy GD, Bo gdo MW, e al. Ra e o ein ec ion ube culosis a e success ul ea men is highe han a e o new ube culosis. Am J Respi C i Ca e Med. 2005;171:1430–5. 7. Jaakola S, Lyy ikäinen O, Rimhanen-Finne R, Salmenlinna S, Pi honen J, Sa olainen-Kop a C, Lii sola K, Jala a J, To opainen M, Nohynek H, Vi anen M, Lö lund J-E, Kuusi M, Salminen M. In ec ious diseases in Finland 2015. THL, 2016. h p://u n. i/URN:ISBN:978-952-302-710-7. Accessed 7 Aug 2017. 8. Raisanen PE, Soini H, Vasanka i T, Smi PW, Nuo i JP, Ollg en J, e al. Tube culosis in immig an s in Finland, 1995-2013. Epidemiol In ec . 2016; 144:425–33. 9. Minis y o social a ai s and heal h. Na ional ube culosis con ol p og amme 2006. Minis y o Social A ai s and Heal h. Helsinki; 2006. h p:// u n. i/URN:NBN: i- e201504225758. Accessed 7 Aug 2017. 10. Ko honen V, Smi P, Haanpe ä M, Casali N, Ruu u P, Vasanka i T, e al. Whole genome analysis o mycobac e ium ube culosis isola es om ecu en episodes o ube culosis, Finland, 1995-2013. Clin Mic obiol In ec . 2016;22:549–54. 11. Wo ld Heal h O ganisa ion. De ini ions and epo ing amewo k o ube culosis: Wo ld Heal h O ganiza ion; 2013. h p://apps.who.in /i is/ bi s eam/10665/79199/1/9789241505345_eng.pd . Accessed 7 Aug 2017 12. Vasanka i T, Kokki M, Holms om P, Liippo K, Sa na S, Ruu u P. Su eillance epo : g ea di e si y o ube culosis ea men in Finland. Eu o Su eill. 2007;12:17–21. 13. Lääkin öhalli us. Tube kuloosi ja sen lääkehoi o. Finnish Na ional Boa d o Heal h. Tube culosis and i s ea men . Kapseli 15. Lääkin öhalli uksen julkaisu 1985. [In Finnish]. Helsinki: Finnish Na ional Boa d o Heal h; 1985. 14. Ande sen PK, Pe me MP. Pseudo-obse a ions in su i al analysis. S a Me hods Med Res. 2010;19:71–99. 15. Pa ne ET, Ande sen PK. Reg ession analysis o censo ed da a using pseudo- obse a ions. S a a J. 2010;10:408–22. 16. C o s JP, And ews NJ, Ba ke RD, Delpech V, Abubaka I. Risk ac o s o ecu en ube culosis in England and Wales, 1998-2005. Tho ax. 2010;65:310–4. 17. Kokki M, Holms om P, Ruu u P. High sensi i i y o ube culosis in a na ional in eg a ed su eillance sys em in Finland. Eu o Su eill. 2005;10:90–3. 18. Bang D, Ande sen AB, Thomsen VO, Lillebaek T. Recu en ube culosis in Denma k: elapse s. e-in ec ion. In J Tube c Lung Dis. 2010;14:447–53. 19. In e an e JD, Haddad MB, Kim L, Gandhi NR. Exogenous Rein ec ion as a cause o la e ecu en ube culosis in he Uni ed S a es. Ann Am Tho ac Soc. 2015. doi:10.1513/AnnalsATS.201507-429OC. 20. Selassie AW, Pozsik C, Wilson D, Fe guson PL. Why pulmona y ube culosis ecu s: a popula ion-based epidemiological s udy. Ann Epidemiol. 2005;15:519–25. 21. Pascopella L, De ieme K, Wa JP, Flood JM. When ube culosis comes back: who de elops ecu en ube culosis in Cali o nia? PLoS One. 2011;6:e26541. 22. Panjabi R, Coms ock GW, Golub JE. Recu en ube culosis and i s isk ac o s: adequa ely ea ed pa ien s a e s ill a high isk. In J Tube c Lung Dis. 2007;11:828–37. 23. Kim L, Moonan PK, Heilig CM, Wood u RSY, Kamme e JS, Haddad MB. Fac o s associa ed wi h ecu en ube culosis mo e han 12 mon hs a e ea men comple ion. In J Tube c Lung Dis. 2016;20:49–56. 24. Pe i AC, Kal enbach LA, Ma u i F, Cummins J, Smi h TR, Wa ken in JV, e al. Ch onic lung disease and HIV in ec ion a e isk ac o s o ecu en ube culosis in a low-incidence se ing. In J Tube c Lung Dis. 2011;15:906–11. 25. Mille J-P, O cau A, de Olalla PG, Casals M, Rius C, Cayla JA. Tube culosis ecu ence and i s associa ed isk ac o s among success ully ea ed pa ien s. J Epidemiol Communi y Heal h. 2009;63:799–804. 26. El Sahly HM, W igh JA, Soini H, Bui TT, Williams-Bouye N, Escalan e P, e al. Recu en ube culosis in Hous on, Texas: a popula ion-based s udy. In J Tube c Lung Dis. 2004;8:333–40. 27. Eu opean Cen e o Disease P e en ion and Con ol/WHO Regional O ice o Eu ope. Tube culosis su eillance and moni o ing in Eu ope 2016. S ockholm: Eu opean Cen e o Diseasennn; 2016. h ps://ecdc.eu opa.eu/ si es/po al/ iles/media/en/publica ions/Publica ions/ecdc- ube culosis- su eillance-moni o ing-Eu ope-2016.pd . Accessed 7 Aug 2017 Ko honen e al. BMC In ec ious Diseases (2017) 17:721 Page 7 o 7