BALF=bronchoalveolar lavage 8uid; CMV=cytomegalovirus; Kcc=Kendall correlation coefficient; LDH=lactate dehydrogenase;

Transcription

1 Amount of Pnumocystis carinii and Dgr of Acut Lung Inflammation in HIVassociatd P carinii Pnumonia* J_rgn Vstbo, M.D.; Thyg L. Nilsn, M.D.; Jtt Jung, M.D.; and Jns D. Lundgrn, M.D. Corrlations btwn smiquantitativ amounts of PnumocyBtiB carini; (PC), th dgr of inftammation, and th svrity of pnumonia wr analyzd in 58 patints with PC pnumonia (PCP). Matrial from both transbronchial biopsis (TBBs; n=39) and bronchoalvolar lavag 8uid (BALF; n = 57) was xamind. In th TBB th amount of PC corrlatd strongly with ovrall inftammation in th intrstitium (Kndall corrlation cofficint [Kl=.59; p<o.oooi), typ 2 pnumocyt prolifration, and dma formation. Th amount of PC in th TBB also corrlatd with intrstitial accumulation of nutrophils (K=.54; p=o.oooi), lymphocyts, and macrophags. In BALF th amount of PC corrlatd with dma formation and typ 2 pnumocyt prolifration in th TBB but not with th prcntag of nutrophils, lymphocyts, or macrophags in vral charactristics of Pnumocystis carinii pnumonia (PCP) in patints with AIDS hav bn S proposd as prognostic indicators, partly supportd by studis corrlating spcially th prcntag of nutrophils in bronchoalvolar lavag fluid (BALF) with th svrity and prognosis of PCE 1-3 Th amount of Pnumocystis carinii (PC) could also ba prognostic indicator, although our prsnt knowldg dos not support this." No study has attmptd to corrlat th following thr aspcts of PCP associatd with th human immunodficincy virus (HIV): (1) amounts of PC; (2) inflammatory changs in th lung; and (3) th svrity and prognosis of PCE In this study, w xamind th rlationship btwn th amount of PC with markrs of inflammation in transbronchial biopsis (TBBs) and BALF of HIVassociatd PCE In ordr to study th clinical rlvanc of ths findings, th amount of PC and th dgr of inhammation wr corrlatd with indicators of th svrity of th pnumonia, t, P 2 and th srum lactat dhydrognas (LDH) lvl at th tim of diagnosis, and with short- and long-trm survival of th patints. lbtints MATERIALS AND METHODS Thr wr 189 fibroptic bronchoscopic procdurs prformd on 153 HIV-positiv patints with pulmonary symptoms in th *From th Dpartmnts of Infctious Disass (Drs. Vstbo, Nilsn, and Lundgrn) and of Pathology (Dr. Jung), Hvidovr Hospital, Univrsity of Copnhagn, Dnmark. Manuscript rcivd Sptmbr 16; rvision accptd Dcmbr 23. BALF. Th amount of PC in th BALF and th prcntag of nutrophils in th BALF corrlatd significantly with POI and th srum lactat dhydrognas (LDH) lvl. Nithr short-trm nor long-trm survival was affctd by th amount of PC, in8ammatory markrs in th TBB, inoammatory clls in BALF, POI' or th srum LDH lvls. In conclusion, th amount of PC is associatd with th xtnt of th acut inftammatory raction in th lung in PCP associatd with human immunod6cincy virus (HIV). (Chst 1993; 14:19-13) BALF=bronchoalvolar lavag 8uid; CMV=cytomgalovirus; Kcc=Kndall corrlation cofficint; LDH=lactat dhydrognas; MGG =May-Griinwald-Gimsa stain; PC =Pnumo Cf/BIiB carini; pcp=pnulllocflblu carini pnumonia; TBB = transbronchial biopsy priod from Sptmbr 1989 to Jun Thirty-six procdurs wr follow-up xaminations du to ithr prsistnt symptoms or just to assss th ffct of th tratmnt. Sixty-nin patints had pcp during th priod, and of thos, w rport th findings in a group of 58 conscutiv patints who had a primary pisod of PCE Clinical data, including pcp prophylaxis prior to bronchoscopy, wr collctd prospctivly. Chst x-ray films and laboratory valus, including CD4 count, srum LDR lvl, and artrial blood gas lvls, wr gathrd lss than 24 h bfor bronchoscopy. Outcom was rcordd from th mdical fils. Bronchoscopy Bronchoscopy was don undr local ansthsia as prviously dscribd." Th BAL and TBB wr prformd in th right middl lob in th cas ofdiffus infiltrats on th chst x-ray film (othrwis from th sit of localizd infiltration). Th BAL was prformd with instillation of up to 24 ml of warmd salin solution in 4 to 6 aliquots. Th rcovrd BALF was poold; 6 to 8 prcnt of th instilld volum was aspiratd. Two or thr TBBs wr takn in anothr subsgmnt in th sam lob. BALF; CU Diffrntial Counts, and TBBs Immdiatly aftr bronchoscopy, a standardizd volum of th obtaind BALF (3.5 ml) was cntrifugd for 5 min at 1,5 g, and smars wr prpard from th dsposit by aliquoting 1 drop pr slid. Th following staining mthods wr usd: May-Griinwald Gimsa stain (MGG); Papanicolaou stain; Grocott mthnaminsilvr stain; and an immunoproxidas stain using monoclonal antibodis against PC (DAKO-Pnumocystis, M 778; DAKO-DK). Th biopsis wr fixd in 4 prcnt buffrd Formalin, mbddd in paraffin, and staind with hmatoxylin-osin, van Gison- Hansn stain, priodic acid-schiff (PAS), Grocott mthnamin-silvr stain, and th immunoproxidas stain. Biopsis wr considrd satisfactory for valuation if thy containd mor than 1 wll-prsrvd alvolar lumina. Evaluation of th BALF and TBBs was don prospctivly and without knowldg of th patint's clinical status. Also, th obsrvr was blindd to th information on th BALF whn studying th TBB and vic vrsa. CHEST I 14 I 1 I JUL'f,

2 Tabl 1-Smiquantitativ Grading ofoccurrnc ofpc in BALF and TBBs Grad Occurrnc o PC in BALF PCinTBB No foamy matrial or cysts No PC Scattrd foamy matrial and/or fw scattrd cysts Fw scattrd PC organisms in singl alvolar lumn Foamy matrial and svral cysts Fw PC organisms in svral lumina Numrous pnumocysts all ovr whol smar in svral stains Numrous PC organisms filling almost all lumina Th prsnc of PC was diagnosd if a honycombd foamy matrial (i, trophozoits) could b dmonstratd by ithr MGG or Papanicolaou stain and confirmd on th immunoproxidas staind slid or if cysts wr found in th silvr-staind smar (or both). In both BALF and TBB, trophozoits contributd mostly in th ovrall scor of PC. Thy wr gradd in ithr MGG staining or using immunoproxidas staining. In only two cass wr only cysts dtctd. Th amount of PC was gradd smiquantitativly on a scal of to 3 as dscribd in Tabl 1. In a prvious sris of patints with P C th intraobsrvr variation for grading th amount of PC and indics of inhammation was xamind. For th amount of PC in both th TBB and BALF, a K valu of.81 was found. In th BALF Cytospin prcipitats, a diffrntial count of th inflammatory clls was prformd in a rprsntativ ara of th MGG-staind slid containing a monolayr of clls. In th TBB th inflammatory raction in th alvolar walls was stimatd smiquantitativly and gradd on a scal of to 3 by rcording th following paramtrs: dma in th alvolar walls (, not prsnt; 1, slight widning of th alvolar spta with a loos txtur of th tissu; incrasing to 3, svr incras of th thicknss of th alvolar wall du to intrstitial 8uid accumulation); typ 2 pnumocyt prolifration (, not prsnt; 1, scattrd cuboidal clls protruding into th alvolar spac; incrasing to 3, numrous cuboidalclls protrudinginto thalvolar spac in a hobnail fashion); and th numbr and typ ofinflammatory clls (, no in8ammatory clls prsnt); 1, scattrd inflammatory clls in fw alvolar walls; incrasing to 3, numrous in8ammatory clls in all alvolar walls prsnt). Th prsnc of fibrin and th numbr and typ of inflammatory clls in th alvolar lumina wr valuatd in th sam way. For inflammation, pnumocyt prolifration, and dma in th TBB, w hav prviously found K varying from.52 to.67. For cll typs involvd in inflammation, K varid from.59 to 1.. Statistical Mthoda A nonparamtric masur of corrlation, i, th Kndall rankordr corrlation cofficinr'" (Kcc) has bn usd; this tst maks no assumptions about linar associations and distributions. Th logrank tsts was usd for analysis of survival. A Kcc =.54, p <.1 B Kcc =.52, p =.1 C Kcc =.3, p =.2 c:: 3 : CD CD.s s 2 1 III 2 II o 2 c:: E. I) I) I) E 1 III 1 E ;;: E :J.!: II a. 1 2 J J PC in TBB PC in TBB PC in TBB Kcc =.54, p =.1 E Kcc =.46, P =.1 F Kcc =.53, p =.1 :;3 II :;3 :;3.s 62 2 II 2., In I) I).!? Ilil III III s: 1 II III..J:. :J O...J J PC in TBB PC in TBB PC in TBB FIGURE 1. Association btwn smiquantitativ valuation of PC in TBBs and smiquantitativ scor of inflammation in TBB (A), pnumocyt prolifration in TBB (B), dma in TBB (C), nutrophils in intrstitium oftbb (D), macrophags in intrstitium oftbb (E), and lymphocyts in intrstitium oftbb (F). Valus for Kcc and p ar shown. 11 Amountof P carinii and Dgrof Acut Inflammation in Lungs (Vstbo t 81)

3 RESULTS Th total group consistd of 58 conscutiv patints with PCP for whom mdian valus and rangs for background data ar shown in th following tabulation: Ag, yr 38 (22-76) Po 2, mm Hg 8.4 ( ) Pco 2, mm Hg 4.3 ( ) Srum LDH 567 (259-1,293) CD4 count (mm") 14 (-2) No. of daths 5 (9%) No diffrncs in ths charactristics wr found whn comparing all patints with th 57 patints with complt information on th BALF and th 39 patints with complt information on th BALF and TBB. All patints had low CD4 counts, i, 2/mm 3 or lss. Within th rang to 2/mm 3, no association btwn th CD4 count and th amount of PC was sn, ithr in th BALF or in th TBB. Only two patints had rcivd PCP prophylaxis, i, pntamidin inhalation; in ths patints th amounts of PC in th BALF and TBB wr similar to thos of th othrs. Pathologic Corrlations Figur 1 shows that smiquantitativ valuation of PC in th TBB corrlatd with th smiquantitativ scor of in8ammation in th TBB (Fig la), th typ 2 pnumocyt prolifration in th TBB (Fig IB), dma in th TBB (Fig lc), nutrophils in th intrstitium of th TBB (Fig ID), lymphocyts in th intrstitium of thtbb (Fig IE), and macrophags in th intrstitium of th TBB (fig IF). Valus for Kcc and p ar shown. Th amounts of PC in th TBB wr significantly associatd with a smiquantitativ scor of macrophags in th alvolar spac in th TBB (Kcc=.27; p=.3). No significant numbrs of nutrophils or lymphocyts wr dtctd in this ara of th TBB. Th association btwn smiquantitativ valuation of PC in th BALF and th prcntag of nutrophils in th BALF was insignificant (Kcc = -.1; p =.46), as was th association btwn PC in th BALF and lymphocyts (Kcc= -.8; p=.25) and macrophags (Kcc=.5; p=.46) in th BALF. Associations btwn PC in th BALF and inhammatory markrs in th TBB wr also xamind. Th PC in th BALF corrlatd significantly with typ 2 pnumocyt prolifration in th TBB (Kcc =.35; p =.6) and dma in th TBB (Kcc=.26; p=.3). Clinical Findings at TIm ofdiagnosis Primarily th P 2 but also th srum LDH lvl at th tim of bronchoscopy wr studid as markrs of th svrity of PCE Information on th P 2 and th srum LDH lvl at tim of bronchoscopy was availabl for 53 and 43 patints, rspctivly; missing data wr unrlatd to any of th faturs undr study Th Tabl 2-Associations Btwn PC in Both BALE' and TBBs, Inflammatory Markrs in TBB, Nutrophils in Both BALE'and TBB, and Markrs ofsvrity ofpcp P 2 Srum LDU Data Kcc p Kcc P BALF PC Nutrophils TBB PC Alvolar spac Macrophags Intrstitium Ovrall inflammation Edma Pnumocyt prolifration Nutrophils Macrophags amounts of PC in th BALF, nutrophils in th BALF, and dma in th TBB corrlatd significantly with Po 2, whras only PC in th BALF corrlatd significantly with th srum LDH (Tabl 2). Inflammatory markrs othr than dma in th TBB wr consistntly associatd with P 2 and LDH, but corrlations wr only marginally significant. Bactria wr found in th BALF of fiv patints. Nithr inflammatory markrs in th TBB (including nutrophils) nor th man P 2 or srum LDH in ths patints diffrd from patints without bactria in th BALF. Cytomgalovirus (CMV) was culturd from th BALF of 12 patints; a positiv cultur was not rlatd to cll diffrntial counts in th BALF, TBB findings, Po 2, or srum LDH. Survival Th amount of PC and th inflammatory changs dtctd in th BALF and TBB at th tim of diagnosis wr without prognostic information for survival from th PCP pisod. Fifty-thr patints survivd th initial PCE Thy hav bn followd for 77 to 73 days (mdian, 346 days); 16 patints hav did during this follow-up (12-month survival rat, 65 prcnt). Nithrsignificant associations nor any trnd could b found whn looking at associations btwn long-trm mortality and th following variabls: th amount of PC in th BALF or TBB; in8ammatory markrs in th TBB; cll typs in th TBB; cll diffrntial counts in th BALF; P 2 at th tim of bronchoscopy; or srum LD H at th tim of bronchoscopy. DISCUSSION To our knowldg, this study is th first to corrlat th amount of PC with th apparanc of in8ammation in th lung of patints with AIDS who had PCE Dscriptions of th histologic findings in PCP usually mphasiz th infiltration of lymphocyts and macro- CHEST I 14 I 1 / J U L

4 phags into th pulmonary Intrstitium.l-v'" In this study th amount of PC in th TBB corrlatd with th dgr of inflammation in th intrstitium, including th accumulation of nutrophils and lymphocyts. Th amount of PC in th TBB corrlatd with smiquantitativ masurs of alvolar macrophags sn in th TBB but not with th prcntag of inhammatory clls in th BALF. Unfortunatly, th total numbr of inhammatory clls pr millilitr of BALF was not availabl; thus, w wr not abl to analyz for quantitativ masurs of th inflammatory raction in th BALF. Also, quantitating th amount of PC in th BALF appars to b mor unrliabl than quantitating th amount of PC in th TBB, sinc PC adhrs to th alvolarwall.ii Othr invstigators hav dscribd ways of quantitating th amount of PC in th BALF, ithr rlativ to th numbr of inflammatory clls" or pr millilitr ofbalf, but nithr of ths mthods sms suprior to th rlativly simpl but standardizd tchniqu usd in this rport. A varity of possibl causs for th incras in th numbr of nutrophils in PC-infctd pulmonary intrstitium xists, g, tobacco smok and contmporary infction with bactria, C M or hrps simplx virus. In our patints, no data wr availabl on xposur to tobacco smok. Th prsnc of bactria was associatd with highr prcntags of nutrophils in th BALF but not in th TBB. Cytomgalovirus was not associatd with th accumulation of nutrophils in our study Othrs hav implmntd an incrasd prcntag of nutrophils in th BALF as a prdictor of lowr Po! at th tim of diagnosis," W found a similar association whn corrlating th prcntag of nutrophils in th BALF and initial Po!. Thr was a statistically wakr invrs corrlation btwn th amount of nutrophils in th intrstitium and Po! (p=.7), which may b xplaind by th smallr numbr of TBBs than BALs prformd (39 vs 51). Also, th amount of PC corrlatd with Po! (BALF, p =.2; TBB, p =.9). Ths findings ar somwhat in contrast to thos of a prvious study by Limpr t al," who found no association btwn th numbr of cysts pr millilitr of BALF and th initial POI in patints with AIDS who hav P C In th TBB, only th dgr of intrstitial dma in our patints corrlatd significantly with Po!. Th associations btwn th amount of PC, th accumulation of nutrophils, and dma may suggst that PC and a rsultant acut inhammation lad to dma formation in th pulmonary intrstitium and consquntly to impaird oxygn transportation; howvr, it is notworthy that this oftn svr inhammatory rspons in HIV-associatd PCP is not sufficint or ffctiv in claring th causal microorganism. Srum LDH was includd in this study as a markr of th svrity of th pnumonia basd on prvious 112 studis l3-15 and appard to hav prognostic significanc. Srum LDH valus wr markdly highr in 3 of our patints with a fatal cours of th pnumonia than in 4 surviving patints (mans, 1,95 and 63 U/L, rspctivly); th small numbr of daths prcluds statistical analysis. In our study th amount of PC in th BALF corrlatd with srum LD H lvls. Also, th prcntag of nutrophils in th BALF and th dgr of dma formation in th pulmonary intrstitium wr associatd with srum LDH. Ths associations hav not prviously bn dscribd but could b du to backhow of LDH from pulmonary clls ovr th alvoloartrial mmbran, as suggstd by Smith t ali W did not find th prcntag of nutrophils in th BALF nor th initial Po! to prdict survival from th PCP pisod, whras svral prvious studis hav found a high prcntag of nutrophils and a low POI to b associatd with poor survival W did not find any association with short-trm survival whn studying changs in th TBB, as othrs hav faild to do,'? howvr, only 5 (9 prcnt) of our patints did in th cours of th PCE W also studid whthr changs in th TBB and BALF could prdict changs in long-trm survival, but our data faild to show such an association. Ths ngativ rsults contrast with prvious findings, 17 whr patints with HIV-rlatd PCP who had larg amounts of PC and a svr dgr of dma in th initial TBB had significantly lowr long-trm survival (i, obsrvation tim, 12 months). Th tim span btwn th prsnt study and th studis by Brnnr t al'? and othrs l 3.13 may xplain th discrpancis, sinc antiviral thrapy tratmnt of opportunistic infctions, and prsumably arlir diagnosis of PCP hav improvd in th last 5 yars. I 9-!1 Th data on morphology prsntd in this rport could hav bn inhuncd by th pathologist studying both th amount of PC and th dgr of inhammation in th sam spcimn; howvr, svral points argu against th data bing confoundd. First, th amount of PC and th changs in host clls wr valuatd using diffrnt stains of th matrial. Scondly, th study was don prospctivly; and th hypothss tstdin this matrialwraddrssdby thclinicians. Thirdly; th finding that th dgr of dma and typ 2 pnumocyt prolifration in th TBB corrlatd with th amount of PC in th BALF strongly argus for a ral association btwn ths paramtrs in th lung of patints with AIDS who hav PCE Futur invstigations should focus on th idntification of mdiators and clls of importanc in th inflammatory componnt of th dvlopmnt of PCE REFERENCES 1 MurrayJF, Mills J. Pulmonary infctious complications of human Amount of P carinll and Dgrof AcutInflammation in Lungs(Y8stbo t 81)

5 immunodficincy virus infction: part 2. Am Rv Rspir Dis 199; 141: Smith RL, l-sadr WM, Lwis ML. Corrlation of bronchoalvolar lavag cll populations with clinical svrity of Pnumacystis carinii pnumonia. Chst 1988; 93:6-4 3 Mason GR, Hashimoto CH, Dickman PS, Foutty LF, Cobb CJ. Prognostic implications of bronchoalvolar lavag nutrophilia in patints with Pnumocystls carinii pnumonia and AIDS. Am Rv Rspir Dis 1989; 139: Limpr AD, OfFord Smith TF, Martin WJ. Pnumocystls carinii pnumonia: diffrncs in lung parasit numbr and inflammation in patints with and without AIDS. Am Rv Rspir Dis 1989; 14: Orholm M, Lundgrn JD, Nilsn TL, Ivrsn J. Indication for fibroptic bronchoscopy in HIV-infctd patints suspctd for Pnumocystis carinii pnumonia. Dan Md Bull 199; 37: Sigl S, Castllan NJ Jr. Nonparamtric statistics for th bhavioral scincs. 2nd d. Nw York: McGraw-Hill Book Co, SPSS. Statistical packag for th social scincs. Nw York: McGraw-Hill Book Co, Pto R, Pik MC, Armitag Brslow NE, Cox DR, Howard S t al Dsign and analysis of randomizd clinical trials rquiring prolongd obsrvation of ach patint: 2. analysis and xampls. Br J Cancr 1977; 35: Wbr WR, Askin FB, Dhnr LE Lung biopsy in Pnumocystis carin" pnumonia: a histopathologic studyof typical and atypical faturs. Am J Clio Patholl977; 67: Gal AA, Koss MN, Strigl S, Angritt E Pnumocystis carinii infction in th acquird immun dficincy syndrom. Smin Diag Patholl989; 6: Pottratz ST, Paulsrud J, Smith JS, Martin WJ. Pnumocystis carin" attachmnt to culturd lung clls by Pnumocystis gp 12, a fibronctin binding protin. J Clin Invst 1991; 88: Baughman Strohofr S, Colanglo G, Fram PT. Smiquantitativ tchniqu for stimating Pnumocystis carinii burdn in th lung. J Clin Microbioll99; 28: Kals C Murrn JR, Torrs RA, Crocco JA. Early prdictors of in-hospital mortality for Pnumocyms carin«pnumonia in th acquird immunodficincy syndrom. Arch Intrn Md 1987; 147: Bnson CA, Spar J, Hins D, Pottag JC Jr, Ksslr HA, Trnholm GM. Combind APACHE II scor and srum lactat dhydrognas as prdictors of in-hospital mortality causd by first pisod Pnumocystis carinii pnumonia in patints with acquird immunodficincy syndrom. Am Rv Rspir Dis 1991; 144: Effrn LS, Nadarajah D, Palat DS. Survival following mchanical vntilation for Pnumocystis carini; pnumonia in patints with th acquird immunodficincy syndrom: a diffrnt prspctiv. Am J Md 1989; 87: Spich R, Wbr R, Kronaur CM, Opravil M, Luthy R, Russi EW Prognostic scor for Pnumocystis carin" pnumonia. Rspiration 199; 57: Brnnr M, Ognibn F Lack EE, Simmons JT, Suffrdini AF, Lan HC, t al. Prognostic factors and lif xpctancy of patints with acquird immunodficincy syndrom and Pnumocystis carinii pnumonia. Am Rv Rspir Dis 1987; 136: Garay SM, Grn J. Prognostic indicators in th initial prsntation of Pnumocystis carin" pnumonia. Chst 1989; 95: Pitt J, Mor Flisman J. Pattrns of survival with AIDS in th Unitd Stats. Halth Srv Rs1991; 26: Smith E, Orholm M. Trnds and pattrns of opportunistic disass in Danish AIDS patints Scand J Infct Dis 199; 22: Fridmann Y, Franklin C, Frls S, Wil MH. Long-trm survival of patints with AIDS, Pnumocystis carinii pnumonia and rspiratory failur. JAMA 1991; 266:89-92 CHEST I 14 I 1 I J U 1993 L 113