Evaluation of non-HIV-related , drug-sensitive cluster outbreaks of tuberculosis with PCR-based DNA fingerprinting

OB JEC TIVES: To char ac ter ize two clus ter out breaks of tu ber cu lo sis by DNA fin ger print ing and to evalu ate the pos si bil ity of their as so cia tion with hu man im mu no de fi ciency vi rus (HIV) in fec tion and mul ti ple drugresistant My co bac te rium tu ber cu lo sis. SET TING: Two clus ters of tu ber cu lo sis cases in ad join ing abo rigi nal Ca na dian re serves and a con trol popu la tion. METH ODS: All cases of tu ber cu lo sis di ag nosed in the two com mu ni ties and a number of con trol iso lates were fin ger printed us ing a novel sim pli fied DNAbased tech nique. Demo graphic data, pu ri fied pro tein de riva tive (PPD) skin test re sponse to 5TU PPD, type of dis ease and HIV se rol ogy were also evalu ated. Posi tive cul tures were as sessed for sus cep ti bil ity to an ti tu ber cu lous drugs. RE SULTS: Two dis tinct clus ters with five cases in one com mu nity and 21 cases in the sec ond were iden ti fied. Iso lates from both com mu ni ties dif fered in their DNA pat tern but were iden ti cal within com mu ni ties. Thir teen cases had bac te rio logi cally proven dis ease and grew sus cep ti ble or gan isms. Se ro logi cal analy sis for HIV in fec tion was car ried out in 15 of 18 adults, and all were nega tive. CON CLU SIONS: The study shows the po ten tial for the rapid trans mis sion of tu ber cu lo sis in fec tion in HIVnegative sub jects with sus cep ti ble or gan isms. It fur ther high lights the use ful ness of DNA fin ger print ing for mo lecu lar epi de mi ol ogy in evalu at ing par ticu lar out breaks.

T here is con cern re gard ing an in creased preva lence of tu - ber cu lo sis (TB) in both de vel oped and de vel op ing countries (1).Much of this in crease is thought to be as so ci ated with hu man im mu no de fi ciency vi rus (HIV) in fec tion (2).In the United States multidrug-resistant TB (MDRTB) has particu larly been a prob lem (3).Re cent re ports of stud ies us ing DNA-based tech nol ogy have shown rapid trans mis sion of infec tion and de vel op ment of MDRTB in the pres ence of HIV in fec tion (4).In ad di tion, population-based mo lecu lar epi demi ol ogy stud ies have shown greater trans mis sion of dis ease than pre vi ously thought (5,6).
Among abo rigi nal Ca na di ans there is a high preva lence of TB -ap proxi mately 10 times the rate among the gen eral popu la tion (90/100,000 ver sus nine/100,000 popu la tion, respec tively) (7).So far MDRTB and HIV in fec tion have not been ma jor prob lems in this group (8), al though re cent trends in di cate that this may be chang ing (9).We re cently in ves tigated a number of clus ter out breaks of TB in this popu la tion group (10); two were geo graphi cally close, oc cur ring approxi mately 5 km apart.
We were in ter ested in es tab lish ing the preva lence of HIV in fec tion in both clus ters, the drug sus cep ti bil ity pat terns, and fi nally whether there was any link be tween the two outbreaks by us ing a novel DNA-based tech nique, which has been re cently de scribed (11).

PATIENTS AND METHODS
All cases of TB di ag nosed in both com mu ni ties be tween Janu ary and De cem ber 1992 were iden ti fied.Demo graphic data, risk fac tors for TB, re sponse to pu ri fied pro tein de rivative of tu ber cu lin (5TU PPD), chest x-ray find ings, HIV serol ogy, bac te rio logi cal re sults and drug sus cep ti bil ity pat terns were col lected.Spu tum and pleu ral fluid were cultured us ing the Bac tec tech nique.Speci mens that were positive on cul ture for My co bac te rium tu ber cu lo sis were proc essed for strain iden ti fi ca tion.
A newly de scribed method us ing a ligation-mediated polymerase chain re ac tion (PCR) pro ce dure was used to amplify the flank ing se quences on both sides of the in ser tion se quence 6110 from restriction-digested chro mo so mal DNA.This method has re cently been de scribed in de tail (11).Briefly, this pro cess in volves in ac ti vat ing the my co bac te ria with phe nol and break ing open the my co bac te ria by vortexing in the pres ence of glass beads.The de bris are then pel leted and the DNA su per natant is pu ri fied by ad her ence to glass pow der.Af ter re stric tion en zyme di ges tion a linker is ligated to the cut ends.Prim ers ho molo gous to the in ser tion se quence 986 and the linker are then used in PCR and the prod ucts run on an aga rose gel.The re sults are char ac ter is tic pat terns of mul ti ple bands of dif fer ent lengths de pend ing on the dis tance from the in ser tion se quences to the ex ter nal restric tion sites.

RESULTS
Two dis tinct clus ters with five cases in one com mu nity and 21 cases in the sec ond were iden ti fied.Iso lates from both com mu ni ties dif fered in their DNA pat tern but were iden ti cal within com mu ni ties.Thir teen cases had bac te rio logi cally proven dis ease and grew fully sen si tive or gan isms.Se ro logical analy sis for HIV in fec tion was car ried out for 15 of 18 adults; all were nega tive.All sub jects were in ter viewed by one in ves ti ga tor, and the con tact evalua tion rec ords for each out break were re viewed in the Brit ish Co lum bia Cen tre for Dis ease Con trol, Tu ber cu lo sis Di vi sion.Clus ter A: Both in dex and source case in clus ter A was a 22-year-old male.De tails of other cases di ag nosed within the com mu nity are given in Ta ble 1.He pre sented with a selfreported cough of six weeks' du ra tion in May 1992, but a history of cough ing for at least six months was sub se quently obtained from fam ily mem bers.He de nied a his tory of in tra ve nous drug use.In this com mu nity there were 21 cases (12 male).All but four had symp toms; 13 had docu mented PPD con ver sion, hav ing been nega tive on PPD skin test ing at ini tial screen ing in the com mu nity or hav ing pre vi ously been known to be nega tive from prior tu ber cu lin rec ords.Followup PPD skin tests were car ried out when sub jects pre sented with symp toms or where follow-up PPD skin test ing at least three months af ter the ini tial screen ing was car ried out.Recently docu mented PPD con ver sion in di cated re cent in fection.Seven oth ers had a posi tive PPD to 5 TU at pres en ta tion.HIV se rol ogy was com pleted in 12 of 15 adults and was nega tive in all.Dis ease was con firmed by my co bac te ri ol ogy in 10 cases and all had drug-susceptible or gan isms.Clus ter B: The in dex case in clus ter B was an in fant aged 20 months who pre sented with re cur rent pneu mo nia.Char ac teris tics of the other cases di ag nosed in this com mu nity are shown in Ta ble 2.Even tu ally the in fant had gas tric wash ings ob tained for my co bac te ri ol ogy and these grew M tu ber cu losis.Re verse con tact trac ing iden ti fied the fa ther of the infant's ba bysit ter as the source case.He had bi lat eral api cal ab nor mali ties noted 18 months pre vi ously but at that time he re fused to at tend for fur ther evalua tion.He had a his tory of in tra ve nous drug use.Three fur ther cases, all as ymp to matic, were iden ti fied in con tact trac ing.Two had PPD con ver sions and one had a posi tive PPD at evalua tion.Three sub jects had sus cep ti ble, bac te rio logi cally proven dis ease.The re main ing sub jects had pri mary dis ease.The three adults in this clus ter were HIV-negative.

DNA fin ger print ing:
The iso lates from both com mu ni ties were found to dif fer, show ing two unique pat terns.Twelve con trol speci mens, drawn at ran dom from cases di ag nosed through out the same pe riod and evalu ated at the same time, were all dif fer ent.The ex cep tions were iso lates from two HIV-negative broth ers who had the same PCR pat tern, which was dif fer ent from those in the clus ters.The con trol specimens were ob tained to pro vide in ter nal va lid ity, fur ther blind ing the labo ra tory to the epi de mi ol ogi cal data.Gel electro pho re sis pat terns pro duced by a sam ple of con trol and clus ter case iso lates from clus ter A are shown in Fig ure 1.
DISCUSSION TB is ree merg ing as a ma jor health con cern in both de veloped and de vel op ing coun tries (1).Tra di tion ally it has been thought that the ma jor ity of newly di ag nosed cases rep resented re ac ti va tion of pre vi ously ac quired in fec tion (12), although ex oge nous re in fec tion with a new or gan ism has been de scribed us ing phage typ ing (13).New in sight into the trans mis sion of TB in fec tion has been gained by evalu at ing clus ters of HIV-associated cases in whom in fec tion with a resis tant or gan ism has been ac quired while on ther apy for drug-susceptible dis ease (4).In ad di tion, population-based stud ies us ing re stric tion frag ment length poly mor phism as a fin ger print ing tech nique have in di cated that be tween 30% and 40% ( 6) of all newly diag nosed cases of dis ease are re lated.This in di cates that, contrary to the tra di tional be lief that ap proxi mately 10% of new cases are due to re cent ac qui si tion of in fec tion and pro gression to dis ease, much more in fec tion trans mis sion and progres sion to dis ease oc curs than here to fore thought.
We re port two sepa rate clus ters of TB cases with a to tal of 27 cases of ac tive TB.Ear lier non-HIV-related stud ies have tended to con cen trate on meth odo logi cal is sues (14-16) and dis tin guish ing among strains, but they have lacked a large number of simi lar iso lates and pro vided no epi de mi ol ogi cal in for ma tion.
The speed at which in fec tion oc curred and cases of ac tive TB dis ease de vel oped can be in ferred from the chrono logi cal pres en ta tion of the ac tive cases, es pe cially in clus ter A, follow ing the di ag no sis.Our data high light the ra pid ity with which in fec tion can be trans mit ted and dis ease de velop even in non-HIV-infected sub jects.This ra pid ity is fur ther supported by the high pro por tion of sub jects who had a con version of their Man toux PPD skin test re sponses while be ing evalu ated, in di cat ing that in fec tion had re cently been acquired.The two dis tinct pat terns in the clus ters sup port two dif fer ent sources for the two out breaks.
Para doxi cally the source case in clus ter A re ported the shorter his tory of symp toms but gen er ated the larger number of cases.Al though the fre quent com mu nal events on a North Ameri can In dian re serve and the gen er ally poor stan dard of hous ing may have fa cili tated the rapid trans mis sion of in fection, the to tal number of cases iden ti fied is sig nifi cant.It high lights the im por tance of early di ag no sis of TB and prompt ini tia tion of ther apy (17).Be cause of the po ten tial for trans mis sion of in fec tion within the com mu nity, it is pos si ble that the ho mo ge ne ity of iso lates evalu ated from this clus ter out break could be due to a ten dency of iso lates on abo rigi nal re serves to be simi lar.This hy pothe sis is dis proved by the het ero ge ne ity of iso lates from a number of North ern Al berta com mu ni ties (un pub lished data).
A cor ol lary of these data is that in clus ter situa tions, as de -scribed, the risk of pro gres sion from in fec tion to dis ease in cer tain epi de mi ol ogi cal situa tions is that much greater and re in forces the need to con sider prompt che mo pro phy laxis (18).It high lights the im por tance of promptly evalu at ing con tacts of all ac tive cases of TB and promptly ini ti at ing chemo pro phy laxis.For one pa tient in clus ter A there was a twomonth de lay be tween hav ing a nor mal chest x-ray and PPD con ver sion, and be ing of fered che mo pro phy laxis and ini ti ating it.A short time af ter che mo pro phy laxis was started the sub ject pre sented with a pleu ral ef fu sion.The phy si cian who as sessed her did not con sider TB as a pos si bil ity and con tinued the pa tient on isonia zid while the pre sump tive di ag no sis of a parap neu monic ef fu sion was be ing evalu ated.Six weeks later the pa tient pre sented with a TB em pyema which grew a fully sen si tive or gan ism.Af ter de cor ti ca tion and a full course of anti-TB ther apy there was com plete reso lu tion of the abnor mali ties.
The clus ter ing of cases among abo rigi nal pa tients is likely mul ti fac tor ial and, al though our data do not ad dress the is sue of hous ing, it is likely that bet ter hous ing and com mu nity edu ca tion re gard ing TB would fa cili tate a re duc tion in the over all rates in these com mu ni ties.
Al though much of our at ten tion has fo cused on HIV-r elated TB, this re port high lights the rapid pro gres sion to dis ease that can oc cur in the ab sence of HIV in fec tion and with fully sus cep ti ble or gan isms.The avail abil ity of DNA fin ger print ing of re spon si ble or gan isms pro vides a use ful epi de mi ol ogi cal tool for evalu at ing such out breaks and casts new light upon the patho gene sis of TB.Can Respir J Vol 3 No 5 Sep tem ber/Oc to ber 1996

FitzGer ald et al
Fig ure 1) My co bac te rial DNA pat terns af ter ligase-mediated po lymerase chain re ac tion.Lane 1 is an un re lated con trol, lanes 2 to 12 are from pa tients iden ti fied as epi de mi ol ogi cally re lated to clus ter 1 and lane 13 is a mo lecu lar weight (MW) marker.Lanes 2 to 10 have iden ti cal pat terns em pre inte gé né tique de l'ADN et éval uer l'éven tu al ité de leur as so cia tion à une in fec tion par le vi rus de l'im mu nodéfi cience hu maine (VIH) et à My co bac te rium tu ber cu lo sis mul ti résis tant aux médica ments.CON TEXTE : Deux grap pes de cas de tu ber cu lose dans des réserves ad ja cen tes d'a bo rigènes ca na di ens et une popu la tion témoin.MÉTH ODES : Tous les cas de tu ber cu lose di ag nos tiqués dans les deux com munau tés et un nom bre d'i so lats témoins ont été évalués par em pre inte gé né tique re po sant sur une nou velle tech nique simpli fiée d'iden ti fi ca tion de l'ADN.Les données démo graphiques, la réac tion au test cu tané à la tu ber cu line à 5 U.T. de PPD, le type de mala die et une sé rolo gie pour le VIH ont aussi été évalués.Les cultures posi tives ont été ex aminées rela tive ment à leur sus cep ti bil ité aux agents an ti tu ber cu leux.Deux grap pes dis tinc tes de 5 cas dans une communauté et de 21 cas dans la sec onde ont été iden ti fiées.Les isolats prove nant des deux com munau tés diffé raient dans la struc ture de leur ADN mais étaient iden tiques à l'inté rieur de chacune des com munau tés.Une tu ber cu lose ac tive a été démontrée à la cul ture dans 13 cas chez lesquels les or gan ismes se sont révé lés sen si bles aux an ti tu ber cu leux.Une ana lyse sé rolo gique vi sant à dépis ter une in fec tion par le VIH a été pra tiquée chez 15 adul tes sur 18. Tous les su jets testés étaient sé roné ga tifs.CON CLU SIONS : L'étude démontre qu'une trans mis sion rapide de l'in fec tion tu ber cu le use avec des or gan ismes sen si bles, chez des su jets sé roné ga tifs pour le VIH est pos si ble.En outre, elle démontre égale ment l'u til ité de l'em pre inte gé né tique de l'ADN en épidémi olo gie molé cu laire pour éval uer les flambées par ticu lières.
318 Can Respir J Vol 3 No 5 Sep tem ber/Oc to ber 1996 FitzGer ald et al cu lose par RÉSUL TATS :

fin ger print ing of tu ber cu lo sis
Can Respir J Vol 3 No 5 Sep tem ber/Oc to ber 1996 319 DNA

TABLE 1 Characteristics of individuals in cluster A who were in contact with the index case (number 1)
*Two read ings in di cate pu ri fied pro tein de riva tive (PPD) con ver sion dur ing the con tact evalua tion.CXR Chest x-ray; F Fe male; HIV Hu man im muno de fi ciency vi rus; M Male; ND Not done

TABLE 2 Characteristics of individuals in cluster B who were in contact with the source case (number 1)
*Two read ings in di cate pu ri fied pro tein de riva tive (PPD) con ver sion dur ing the con tact evalua tion.CXR Chest x-ray; F Fe male; HIV Hu man im muno de fi ciency vi rus; M Male; ND Not done