2002 metanalysis PcP

8/7/2019 2002 metanalysis PcP

1/8

Meta-analysis of diagnostic procedures for Pneumocystis cariniipneumonia in HIV-1-infected patients

M. Cruciani*, P. Marcati*, M. Malena*, O. Bosco*, G. Serpelloni*, C. Mengoli#

Meta-analysis of diagnostic procedures for Pneumocystis carinii pneumonia in HIV-1-infected patients. M. Cruciani, P. Marcati, M. Malena, O. Bosco, G. Serpelloni,C. Mengoli. #ERS Journals Ltd 2002.ABSTRACT: Sputum induction is a simple and noninvasive procedure for Pneumo-cystis carinii pneumonia (PCP) diagnosis in human immunodeficiency virus-1-positivepatients, although less sensitive than bronchoalveolar lavage (BAL). In order to obtainan overview of the diagnostic accuracy of sputum induction, a systematic review andmeta-analysis of studies reporting the comparative sensitivity and specificity of BAL(the "gold standard") and sputum induction was performed.

The odds ratio and related 95% confidence interval were calculated using summary

receiving operating characteristic curves as well as fixed-effect and random-effectmodels. Based on pooled data, the negative and positive predictive values werecalculated for a range of PCP prevalence using a Bayesian approach.

Seven prospective studies assessed the comparative accuracy of BAL and sputuminduction. On the whole, sputum induction demonstrated 55.5% sensitivity and 98.6%specificity. The sensitivity of sputum induction was significantly higher withimmunofluorescence than with cytochemical staining (67.1 versus 43.1%). In settingsof 2560% prevalence of PCP, the positive and negative predictive values ranged8696.7 and 66.289.8, respectively, with immunofluorescence, and 7994.4 and5383.5% with cytochemical staining.

In conclusion, in a setting of low prevalence of Pneumocystis carinii pneumonia,sputum induction, particularly with immunostaining, appears to be adequate for clinicaldecision-making.Eur Respir J 2002; 20: 982989.

*Center of Preventive Medicine, HIVOutpatient Clinic, Verona, and #Insti-tute of Microbiology, University ofPadua, Padua, Italy.

Correspondence: M. Cruciani, Centreof Preventive Medicine/HIV OutpatientClinic, ULSS 20, V. Germania, 20,37135 Verona, Italy.Fax: 39 0458622239E-mail: [email protected]

Keywords: Bronchoalveolar lavagediagnosismeta-analysisPneumocystis carinii pneumoniasputum induction

Received: February 18 2002Accepted after revision: May 2 2002

The use of effective prophylaxis for opportunisticinfections and, more recently, the availability of highlyactive antiretroviral therapy (HAART) have consid-erably altered the natural history of human immuno-deficiency virus (HIV)-1 infection [14]. The neweffective antiretroviral treatment provides substantialprotection against all major opportunistic infections.For instance, HAART has not only changed themorbidity and mortality of Pneumocystis cariniipneumonia (PCP) but also clinicians9 attitudestowards prophylaxis directed against this opportunis-tic infection [1, 410]. Despite these advances, PCP

has not lost its clinical relevance, particularly inpersons who are not aware of their HIV-1 infectionstatus or in those who fail to respond to antiretroviraltreatment.

In this dynamic context, reappraisal of strategiesfor the management of infectious complications ofHIV is the subject of intense investigation [1l0].Rapid and specific diagnosis of PCP is still neededin HIV-infected patients.

Laboratory techniques for the diagnosis of PCPusually rely on microscopic demonstration of P.carinii by means of conventional cytochemical proce-dures or immunocytochemical staining [1114]. Poly-merase chain reaction assays have been used to

diagnose PCP, but this technique is not yet suitablefor implementation in routine diagnostic laboratories[11]. Immunostaining seems to increase sensitivityand specificity, but the quality of the clinical mater-ial from the respiratory tract is of crucial importancefor diagnostic accuracy [1116]. Indeed, fibreopticbronchoscopy with bronchoalveolar lavage (BAL),with a sensitivity ofw90%, is recognised as the proce-dure of choice for the diagnosis of PCP [11, 12].

The sputum induction technique for PCP diag-nosis was originally described by BIGBY et al. [16].Since then, this technique has emerged as a simple

and noninvasive procedure, although rather widevariation in diagnostic accuracy (2590%) has beenreported [1117]. The wide range of accuracy reportedmay be related to factors in the population studied orfeatures related to study design, execution of the testand staining techniques [18]. For instance, data fromearly reports suggest that the sensitivity of sputuminduction techniques can be increased by use ofmonoclonal antibodies [15]. Moreover, althoughsensitivity and specificity remain constant when thetechnique is applied to populations with differentprevalences of disease, the related predictive values aresubject to wide variation [19].

Data from a systematic review and meta-analysis of

Eur Respir J 2002; 20: 982989DOI: 10.1183/09031936.02.01372002Printed in UK all rights reserved

Copyright #ERS Journals Ltd 2002European Respiratory Journal

ISSN 0903-1936


2/8


3/8

Table 1 summarises the characteristics of the sevenstudies included in the meta-analysis. Altogether, 322individuals (160 patients with PCP and 162 withother lung infections) were studied using the sputuminduction procedure in comparison to BAL. Sputumand BAL specimens from all of these patients wereprocessed using conventional staining methodologies,

which included P. carinii cytochemical stains for cystwall and/or intracystic bodies and trophozoites. Inthree of these studies, which included a total of 156patients, the sputum specimens were also processed by

immunofluorescence staining with monoclonal anti-bodies [41, 42, 44].

Study quality

The method of patient enrolment was consecutive

in all of the included studies. Four studies provided anadequate description of the independence of inter-pretation of test results [39, 4244]. Details of stainingtechniques and specimen collection procedures weresufficiently described in all but one [38] of the sevenpublications. Information on the study populationwas provided in all of the studies.

Results of the meta-analysis

Odds ratios and related 95% confidence intervals foreach study are shown in figure 1. The summary ROCcurves showing true-positive versus false-positive

rates from individual studies are shown in figure 2.The common diagnostic odds ratios and the corres-ponding sensitivities and specificities are shown intable 2. The fixed effect model was used for analysisof immunofluorescence data, as the homogeneityhypothesis was tenable (p=0.5).

The random-effect model was more appropriate for

Table 1. Characteristics of studies included in analysis

First author [ref.] Location (yr) Sputum staining Independenceof test

interpretation

Specimens(PCP by BAL)

n (n)

ROLSTON [38] USA (1988) Cytochemical No 58 (18)ZAMAN [39] USA (1988) Cytochemical Yes 35 (23)MILLER [40] UK (1991) Cytochemical No 40 (29)FORTUN [41] Spain (1992) Cytochemical, immunofluorescence No 17 (14)CHOUAID [42] France (1993) Cytochemical, immunofluorescence Yes 98 (26)SKT [43] Denmark (1995) Cytochemical Yes 33 (13)ROCHA [44] USA (1996) Cytochemical, immunofluorescence Yes 41 (35)

PCP: Pneumocystis carinii pneumonia; BAL: bronchoalveolar lavage.

Odds ratio

0.05 1

b)

ROCHA [44]

Overall

2500

FORTUN [41]

CHOUAID [42]

a)

ROLSTON [38]

Overall

MILLER [40]

ROCHA [44]

CHOUAID [42]

ZAMAN [39]

SKT [43]

Fig. 1. Forest plots showing odds ratios (&) and 95% confidenceintervals (horizontal bars) for studies investigating the diagnosticaccuracy of sputum induction with a) cytochemical and b)immunofluorescence staining (..........: overall odds ratio). Thescale on the x-axis is logarithmic.

False-positive rate

0.0 0.2

True-positiv

erate

0.2

0.0

1.0

1.0

0.4

0.6

0.8

0.4 0.6 0.8

Fig. 2. Summary receiver operating characteristic curves for twomethods of staining induced sputum specimens (: immuno-fluorescence; - - - - : cytochemical).

984 M. CRUCIANI ET AL.


4/8

data from cytochemical staining of sputum, as thehomogeneity hypothesis was rejected (p=0.001).

The summary ROC curves did not provide sig-nificant parameters for the regression line. Moreover,the regression coefficient b assumed a negative value,thus supporting the hypothesis that the odds ratiofor an association between cytochemical or immuno-fluorescent staining of induced sputum specimensand the standard reference is not dependent on thethreshold used.

Cytochemical versus immunofluorescent staining of

induced sputum specimens

The diagnostic odds ratio was considerably higherwhen specimens were examined by means of immuno-fluorescence compared to cytochemical staining. Thiswas also confirmed by comparison of sensitivity andspecificity. The overall sensitivity was significantlygreater with immunofluorescent staining than withcytochemical staining (67.1 versus 43.1%, p=0.001),whereas the specificity was comparable.

The a posteriori positive and negative predictivevalues are shown in table 3. The relationship betweenprevalence of PCP and positive and negative pre-dictive values is shown in figure 3. These overallvalues were calculated according to different preva-lences of PCP, and were based on pooled sensitivityand specificity data (67.1 and 96.5% for immuno-fluorescence and 43.1 and 96.2% for cytochemicalstaining, respectively).

Discussion

Owing to its high sensitivity and specificity, bron-choscopy with BAL is regarded as the gold standardfor PCP diagnosis in HIV-1-infected patients [1113].However, although safer than and at least as sensi-tive as fibreoptic bronchoscopy with transbronchialbiopsy, BAL with fibreoptic bronchoscopy is stillconsidered an invasive diagnostic method [12, 13, 45].As a consequence, the increased prevalence of PCPobserved concomitant with the explosion of theacquired immune deficiency syndrome (AIDS) epi-demic has brought about an increased need for less

invasive diagnostic methods [12, 46].Sputum induction has emerged as a sensitive andcost-effective diagnostic strategy for PCP in HIV-1-infected patients [11, 12, 16, 46]. This technique hasgained popularity because it is less expensive, saves

Table 2. Cumulative meta-analysis data

Staining Model Odds ratio Sensitivity Specificity

Immunofluorescence Fixed effects 56.9 (10.0324.6) 67.1 (58.169.9) 96.5 (87.899.3)Cytochemical Random effects 19.1 (3.0122.6) 43.1 (33.445.9) 96.2 (85.599.3)

95% confidence intervals are shown in parenthesis.

Table 3. Estimated positive predictive value (PPV) andnegative predictive value (NPV) of sputum inductionaccording to prevalence of Pneumocystis cariniipneumonia (PCP)

PCP prevalence%

PPV NPV

IF CS IF CS

1 16.4 10.2 99.7 99.45 50.5 37.2 98.2 97.010 68.3 55.6 96.4 93.815 77.4 66.6 94.3 90.520 82.9 73.8 92.2 87.125 86.6 79.0 89.8 83.530 89.3 82.9 87.3 79.835 91.3 85.9 84.5 75.840 92.8 88.3 81.5 71.745 94.1 90.2 78.2 67.450 95.1 91.9 74.6 62.855 95.9 93.2 70.6 58.060 96.7 94.4 66.2 53.0

IF: immunofluorescence; CS: cytochemical staining.

PCP prevalence %

0 20

NPV%

20

0

100

80

40

60

80

40 60

b)

PPV%

20

0

100

40

60

80

a)

Fig. 3. Relationship between prevalence of Pneumocystis cariniipneumonia (PCP) and a) positive predictive value (PPV) and b)negative predictive value (NPV) of sputum induction followed byimmunofluorescent () or cytochemical staining (- - - - -). ABayesian approach was followed.

985PNEUMOCYSTIS CARINII PNEUMONIA DIAGNOSIS IN HIV-1


5/8

time and is less invasive than bronchoscopy withBAL. In patients without HIV-1 infection, however,this procedure has not been studied extensively.Moreover, the reduced organism burdens observedin PCP associated with conditions other than AIDSseems to be a limiting factor for the widespread use ofsputum induction in patients without HIV-1 infection[47].

The reported sensitivities of the induced sputumtechnique often differ greatly from one report toanother [12, 13, 3844]. In this context, a meta-analysis can provide an overall summary of diag-nostic accuracy, as well as estimates of accuracyaccording to the characteristics of the study design,population examined and test used [2124]. Moreover,the decreased prevalence of PCP related to thewidespread use of prophylaxis, and more recently, tothe availability of HAART, provides further groundsfor a re-evaluation of P. carinii diagnosis.

Seven studies investigating the diagnostic accuracyof sputum induction in comparison with fibreopticbronchoscopy with BAL were included in the present

meta-analysis. In an observational evaluation of themethodological features of studies evaluating diag-nostic tests, selection of nonrepresentative patientsor use of different reference standards were foundto cause significant overestimation of the accuracy ofa diagnostic test [20]. Based on these observations,several studies with methodological shortcomings thatmay have affected the estimation of diagnosticaccuracy of sputum induction were excluded.

Quality assessment of the studies included in theanalysis detected methodological flaws in threereports, in which interpretation of test results wasnot blinded. All of the included studies were conse-cutive, representative of a relevant clinical population

of HIV-infected patients and applied the same refer-ence standard.Conversely, the experimental design of the studies

included in the present analysis may have introducedsome limits to the general applicability of the sum-mary estimates in routine clinical practice. Forinstance, the use of adequate techniques for induction,collection and processing of induced sputum speci-mens is of crucial importance to the diagnostic yieldof the procedure. Although these techniques do notrequire specialised skills or equipment, the availabilityof dedicated experienced personnel certainly increasesthe quality of induced specimens [40]. Moreover,interpretation of stained specimens for P. carinii can

be more problematic for induced sputum than BALspecimens. The use of monoclonal antibodies, how-ever, seems to overcome this problem even in labora-tories with little previous experience of identifying themicroorganism [15, 48].

Using meta-analysis, the present report providesan overall summary of diagnostic accuracy. Acrossthe whole population examined, sputum inductiondemonstrated 55.5% sensitivity and 98.6% speci-ficity. A significantly higher sensitivity from immuno-fluorescence staining compared to cytochemicalstaining of sputum specimens (67.1 versus 43.1%)was clearly identifiable. Conversely, the specificity ofthe two staining techniques was comparable.

The overall measurement of diagnostic accuracyapplied was the summary odds ratio computedusing the fixed-effect and/or random-effect model.Random-effect summaries yield higher estimatedvariances and, consequently, broader confidenceintervals than fixed-effect summaries when there isevidence of substantial heterogeneity among indivi-dual studies. If heterogeneity is found, the fixed-effect

model yields an artificially narrow confidence interval.In this case, the random-effect model, which incor-porates a moment estimator of the between-trialcomponents of variance, is more appropriate [49].Since with cytochemical staining of sputum the homo-geneity hypothesis was rejected, the odds ratio wascalculated using the random-effect model.

The estimate of the summary odds ratio for cyto-chemical staining of sputum induction was 19.1,corresponding to 43.1% sensitivity and 96.2% specifi-city. With the fixed-effect model, the summary diag-nostic odds ratio for immunofluorescence was 56.9,corresponding to 67.1% sensitivity and 96.5% specifi-city, and was reasonably consistent across the studies

(test for heterogeneity, p=0.5).With the summary ROC curve, the regression

coefficient did not differ significantly from zero, thussupporting the assumption that the odds ratio doesnot depend on threshold.

Whether sputum induction is appropriate for aparticular task depends ultimately on the predictivevalues in the intended setting, which, in turn, criticallydepend on the prevalence of the condition to bedetected. Based on the sensitivity and specificityobtained from meta-analysis, the interpretations ofthe sputum induction results for the different pretestprobabilities were examined using Bayes9 theorem togenerate post-test probabilities. For instance, in a

setting of 25% prevalence of PCP, a negative resultexcludes infection in 89.8% of sputum specimensprocessed with immunological staining, and in 83.5%of specimens processed with cytochemical staining.By contrast, in a setting of higher prevalence, anegative result gives an excessive chance of P. cariniibeing present (table 3 and fig. 3). A high prevalenceof PCP is still found in patients who failed to respondto HAART, are noncompliant with antiretroviraland/or prophylactic regimens, or are unaware of theirHIV status, or in countries with a large burden ofHIV-1 infection and no access to effective antiretro-viral treatment [6, 50]. In these circumstances, as inthe pre-HAART era, fibreoptic bronchoscopy with

BAL should be performed in patients with negativeinduced sputum results [51].Conversely, a positive result in a setting of 25%

PCP prevalence gave a 13 and 21% possibility of afalse-positive result with sputum specimens stainedwith monoclonal antibodies and conventional stainingtechniques, respectively. With a higher prevalence ofPCP, a positive result was almost invariably asso-ciated with the disease, independent of the stainingtechnique.

Another potential problem is that the induced tech-nique sputum may be less effective than bronchoscopywith BAL in the diagnosis of other lung diseases inHIV-1-infected patients. In two studies, bronchoscopy



6/8

with BAL performed concomitantly with or after theinduced sputum technique provided additional diag-noses of lung disease other than that caused by P.carinii [40, 52]. In two other comparative studies,however, sputum induction was also found to bereliable for the diagnosis of other infectious causes ofpneumonitis, including tuberculosis [38, 53].

The present meta-analysis supports the argument

that the sputum induction technique is a valuable toolfor the diagnosis of Pneumocystis carinii pneumoniain human immunodeficiency virus-infected patients.Since the specificity of induced sputum specimenresults is high, a positive result retains a high pre-dictive value even in a context of low Pneumocystiscarinii pneumonia prevalence. The interpretation ofa negative result may be more problematic, especiallyif obtained with cytochemical staining. Owing to itsgreater sensitivity, the immunofluorescent staining ofa sputum induction specimen is preferable to cyto-chemical staining. This test could be of particularvalue in settings in which the prevalence of Pneumo-cystis carinii pneumonia is low. In agreement with

the data of CHOUAID et al. [46], in such a situation,immunofluorescence staining of induced sputum maybe the most cost-effective diagnostic strategy. Further-more, a marked reduction in the need for broncho-scopy after the widespread use of protease inhibitorshas already been reported [54]. In conclusion, thepresent data provide compelling evidence of the valueof the sputum induction procedure in the diagnosisof Pneumocystis carinii pneumonia, and underscorethe importance of a reappraisal of strategies for thediagnosis of infectious complications in the era ofhighly active antiretroviral therapy.

Appendix

The following studies comparing induced sputumprocedures and bronchoscopy for the diagnosis ofPCP in HIV-1-infected patients were not includedin the analysis because they failed to meet the prede-fined inclusion criteria. Eleven studies were excludedbecause bronchoscopy with BAL was performed onlyin subsets of patients (usually in patients with negativeinduced sputum results) [6, 16, 46, 51, 5561]. Fifteenstudies (seven case-control [48, 6267] and eightretrospective and/or nonconsecutive [51, 6874]) wereexcluded on the basis of design and/or because theydid not permit direct comparison of accuracy for

every individual in the sample. Three studies used adifferent diagnostic standard reference [52, 75, 76].

References

1. Palella FJ Jr, Delaney KM, Moorman AC, et al.Declining morbidity and mortality among patientswith advanced human immunodeficiency virus infec-tion. N Engl J Med 1998; 338: 796800.

2. Jacobson MA, French M. Altered natural history ofAIDS-related opportunistic infections in the era ofpotent combination antiretroviral therapy. AIDS1998; 12: Suppl. A, S157S163.

3. Gebo KA, Diener-West M, Moore RD. Hospitaliza-tion rates in an urban cohort after the introduction ofhighly active antiretroviral therapy. J Acquir ImmuneDefic Syndr 2001; 27: 143152.

4. Segal R, Poznasky MC, Connors L, Sands K, BarlamT. Changing patterns of presentation of patients withHIV-related disease at a tertiary referral center and itsimplications for physician training. Int J STD AIDS

2001; 12: 453459.5. Detels R, Tarwater P, Phair JP, Margolick J, Riddler

SA, Munoz A. Effectiveness of potent antiretroviraltherapies on the incidence of opportunistic infectionsbefore and after AIDS diagnosis. AIDS2001; 15: 347355.

6. Benito N, Rano A, Moreno A, et al. Pulmonaryinfiltrates in HIV-infected patients in the highly activeantiretroviral therapy era in Spain. J Acquir ImmuneDefic Syndr 2001; 27: 3543.

7. Dworkin MS, Hanson DL, Navin TR. Survival ofpatients with AIDS, after the diagnosis of Pneumo-cystis cariniipneumonia, in the United States. J InfectDis 2001; 183: 14091412.

8. Arozullah AM, Yarnold PR, Weinstein RA, et al. A

new preadmission staging system for predictinginpatient mortality from HIV-associated Pneumocystiscarinii pneumonia in the early highly active anti-retroviral therapy (HAART) era. Am J Respir CritCare Med 2000; 161: 10811086.

9. Powderly WG. Prophylaxis for opportunistic infec-tions in an era of effective antiretroviral therapy. ClinInfect Dis 2000; 31: 597601.

10. Ledergerber B, Mocroft A, Reiss P, et al. Disconti-nuation of secondary prophylaxis against Pneumocys-tis carinii pneumonia in patients with HIV infectionwho have a response to antiretroviral therapy. EightEuropean Study Groups. N Engl J Med 2001; 344:168174.

11. Kroe DM, Kirsch CM, Jensen WA. Diagnosticstrategies for Pneumocystis carinii pneumonia. SeminRespir Infect 1997; 12: 7078.

12. Walzer PD. Pneumocystis carinii. In: Mandell GL,Bennett JE, Dolin R, eds. Mandell, Douglas, andBennett9s Principles and Practice of Infectious Dis-eases. 5th Edn. Churchill Livingstone, Philadelphia,PA, 2000; pp. 27812795.

13. Broaddus C, Dake MD, Stulbarg MS, et al. Broncho-alveolar lavage and transbronchial biopsy for thediagnosis of pulmonary infections in the acquiredimmunodeficiency syndrome. Ann Intern Med 1985;102: 747752.

14. Linder J, Radio SJ. Immunohistochemistry of Pneu-mocystis carinii. Semin Diagn Pathol1989; 6: 238244.

15. Kovacs JA, Ng VL, Masur H, et al. Diagnosis ofPneumocystis carinii pneumonia: improved detectionin sputum with use of monoclonal antibody. N EnglJ Med 1988; 318: 589593.

16. Bigby TD, Margolskee D, Curtis JL, et al. Theusefulness of induced sputum in the diagnosis ofPneumocystis carinii pneumonia in patients with theacquired immunodeficiency syndrome. Am Rev RespirDis 1986; 133: 515518.

17. Bedrossian CWM, Mason MR, Gupta PK. Rapidcytologic diagnosis of Pneumocystis: a comparisonof effective techniques. Semin Diagn Pathol 1989; 6:245261.

18. Deeks JJ. Systematic reviews of evaluations of



7/8

diagnostic and screening tests. BMJ 2001; 323: 157162.

19. Loy CT, Irwig LM, Katelaris PH, Talley NJ. Docommercial serological kits for Helicobacter pyloriinfection differ in accuracy? A meta-analysis. AmJ Gastroenterol 1996; 91: 11381144.

20. Lijmer JG, Mol BW, Heisterkamp S, Bonsel GJ, PrinsMH, van der Meulen JH. Empirical evidence of

design-related bias in studies of diagnostic tests.JAMA 1999; 282: 10611066.

21. Irwig L, Tosteson AN, Gatsonis C, Lau J, ChalmersTC, Mosteller F. Guidelines for meta-analyses eva-luating diagnostic tests. Ann Intern Med 1994; 120:667676.

22. Irwig L, Macaskill P, Glasziore P, Fahey M. Meta-analytic methods for diagnostic test accuracy. J ClinEpidemiol1995; 48: 119130.

23. Littenberg B, Moses LE. Estimating diagnosticaccuracy from multiple conflicting reports: a newmeta-analytic method. Med Decis Making 1993; 13:313321.

24. Vamvakas EC. Meta-analyses of studies of thediagnostic accuracy of laboratory tests. A review of

the concept and methods. Arch Pathol Lab Med 1998;122: 675686.

25. Moses LE, Shapiro D, Littenberg B. Combiningindependent studies of a diagnostic test into asummary ROC curve: data analytic approaches andsome additional considerations. Stat Med 1993; 12:12931316.

26. Gart JJ. Point and interval estimation of the commonodds ratio in the combination of 262 tables with fixedmarginals. Biometrika 1970; 57: 471475.

27. Gart JJ. The comparison of proportions: a review ofsignificance tests, confidence intervals and adjustmentsfor stratification. Rev Int Stat Inst 1971; 39: 1637.

28. Mantel N, Haenszel W. Statistical aspects of theanalysis of data from retrospective studies of disease.

J Natl Cancer Inst 1959; 22: 719748.29. Mantel N. Chi square tests with one degree of

freedom: extension of the Mantel-Haenszel procedure.J Am Stat Assoc 1963; 58: 690700.

30. DerSimonian R, Laird N. Meta-analysis in clinicaltrials. Control Clin Trials 1986; 7: 177188.

31. Cochran WG. The comparison of percentages inmatched samples. Biometrika 1950; 37: 256266.

32. Fleiss JL. Statistical Method for Rates and Propor-tion. 2nd Edn. New York, NY, John Wiley & Sons,1981; pp. 138143.

33. Motulsky H. Intuitive Biostatistics. Oxford, OxfordUniversity Press, 1995.

34. www.weasyma.com/biblio.html35. Cucherat M, Boissel JP, Leizorovicz A, Haugh MC.

EasyMA: a program for the meta-analysis of clinicaltrials. Comput Methods Programs Biomed 1997; 53:187190.

36. Cruciani M, Rampazzo R, Malena M, Lazzarini L,Todeschini G, Messori A. Prophylaxis with fluoro-quinolones for bacterial infections in neutropenicpatients: a meta-analysis. Clin Infect Dis 1996; 23:795805.

37. www.minitab.com38. Rolston KVI, Rodriguez S, McRory L, Botero GU,

Morice R, Mansell PWA. Diagnostic value of inducedsputum in patients with the acquired immuno-deficiency syndrome. Am J Med 1988; 85: 269.

39. Zaman MK, Wootens OJ, Suprahmanya B, Ankobiah W,

Finch PJP, Kamholz SL. Rapid noninvasive diagnosisofPneumocystis cariniifrom induced liquefied sputum.Ann Intern Med 1988; 109: 710.

40. Miller RF, Kocjan G, Buckland J, Holton J, Malin A,Semple SJG. Sputum induction for the diagnosis ofpulmonary disease in HIV positive patients. J Infect1991; 23: 515.

41. Fortun J, Navas E, Marti-Belda P, et al. Pneumocystis

cariniipneumonia in HIV-infected patients: diagnosticyield of induced sputum and immunofluorescent stainwith monoclonal antibodies. Eur Respir J 1992; 5:665669.

42. Chouaid C, Housset B, Poirot JL, Roux P, Lebeau B.Cost effectiveness of the induced sputum techniquefor the diagnosis of Pneumocystis carinii pneumonia(PCP) in HIV-infected patients. Eur Respir J 1993; 6:248252.

43. Skt J, Lerche AG, Kolmos HJ, Nielsen JO,Reinhardt Mathiensen L, Lundgren JD. Pneumocystiscariniiin bronchoalveolar lavage and induced sputum:detection with a nested polymerase chain reaction.Scand J Infect Dis 1995; 27: 363367.

44. Rocha P, Awe R, Guy ES, et al. A rapid inexpensive

method for processing induced sputum for thedetection of Pneumocystis carinii. Am J Clin Pathol1996; 105: 5257.

45. Bigby TD. Diagnosis of Pneumocystis carinii pneu-monia. How invasive. Chest 1994; 105: 650652.

46. Chouaid C, Housset B, Lebeau B. Cost-analysis offour diagnostic strategies for Pneumocystis cariniipneumonia in HIV-infected subjects. Eur Respir J1995; 8: 15541558.

47. Limper AH, Offord KP, Smith TF, Martin WJ II.Pneumocystis carinii pneumonia: differences in lungparasite number and inflammation in patients withand without AIDS. Am Rev Respir Dis 1989; 140:12041209.

48. Elvin KM, Bjorkman A, Linder E, Heurlin N, Hjerpe

A. Pneumocystis carinii pneumonia: detection ofparasites in sputum and bronchoalveolar lavage fluidby monoclonal antibodies. BMJ 1988; 297: 381384.

49. Hardy RJ, Thompson SG. A likelihood approach tometa-analysis with random effects. Stat Med 1996;15: 619629.

50. Ruffini D, Madhi SA. The high burden of Pneumo-cystis carinii pneumonia in African HIV-1-infectedchildren hospitalized for severe pneumonia. AIDS2002; 16: 105112.

51. Huang L, Hecht FM, Stansell D, Montanti R, HadleyWK, Hopewell C. Suspected Pneumocystis cariniipneumonia with a negative induced sputum examina-tion. Is early bronchoscopy useful? Am J Respir CritCare Med 1995; 151: 18661871.

52. Pitchenik AE, Ganjei P, Torres A, Evans DA, RubinE, Baier H. Sputum examination for the diagnosisof Pneumocystis carinii pneumonia in the acquiredimmunodeficiency syndrome. Am Rev Respir Dis 1986;133: 226229.

53. Conde MB, Soares SLM, Mello FCQ, et al. Compar-ison of sputum induction with fiberoptic broncho-scopy in the diagnosis of tuberculosis. Am J RespirCrit Care Med 2000; 162: 22382240.

54. Murri R, Cingolani A, Limongelli P, Visconti E,Maiuro G. Reduction of bronchoscopy performancesshould be included in the advantages of highly activeantiretroviral therapy (HAART). J Acquir ImmuneDefic Syndr 2000; 25: 286287.



8/8

55. Ognibene FP, Gill VJ, Pizzo PA, et al. Inducedsputum to diagnose Pneumocystis carinii pneumoniain immunosuppressed pediatric patients. J Pediatr1989; 115: 430433.

56. Leigh TR, Parson P, Hume C, Husain OAN, Gazzard B,Collins JV. Sputum induction for diagnosis of Pneumo-cystis carinii pneumonia. Lancet 1989; ii: 205206.

57. Ng VL, Gartner I, Weymouth LA, Goodman CD,

Hopewell PC, Hadley WK. The use of mucolysedinduced sputum for the identification of pulmonarypathogens associated with human immunodeficiencyvirus infection. Arch Pathol Lab Med 1989; 113: 488493.

58. Midgley J, Parson PA, Shanson DC, Husain OA,Francis N. Monoclonal immunofluorescence com-pared with silver stain for investigating Pneumocystiscarinii pneumonia. J Clin Pathol 1991; 44: 7576.

59. Kirsch CM, Azzi R, Yenokida GG, Jensen WA.Analysis of induced sputum in the diagnosis ofPneumocystis carinii pneumonia. Am J Med Sci 1990; 299: 386391.

60. Kirsch CM, Jensen WA, Kagawa FT, Azzi RL.Analysis of induced sputum for the diagnosis of

recurrent Pneumocystis cariniipneumonia. Chest 1992;102: 11521154.

61. Chouaid C, Roux P, Lavard I, Poirot JL, Housset B.Use of the polymerase chain reaction technique oninduced sputum samples for the diagnosis of Pneumo-cystis carinii pneumonia in HIV-infected patients. AmJ Clin Pathol 1995; 104: 7275.

62. Lipschik GY, Gill VJ, Lundgren JD, et al. Improveddiagnosis of Pneumocystis carinii infection by poly-merase chain reaction on induced sputum and blood.Lancet 1992; 340: 203206.

63. Willocks L, Burns S, Cossar R, Brettle R. Diagnosisof Pneumocystis carinii pneumonia in a populationof HIV-positive drug users, with particular referenceto sputum induction and fluorescent antibody tech-

niques. J Infect 1993; 26: 257264.64. Wazir JF, Macrorie SG, Coleman DV. Evaluation

of the sensitivity, specificity, and predictive valueof monoclonal antibody 3F6 for the detection ofPneumocystis carinii pneumonia in bronchoalveolarlavage specimens and induced sputum. Cytophatology1994; 5: 8289.

65. Evans R, Joss AWL, Pennington TH, Ho-Yen DO.The use of a nested polymerase chain reaction fordetecting Pneumocystis carinii from lung and bloodin rat and human infection. J Med Microbiol 1995;42: 209213.

66. Ortona E, Margutti P, Tamburrini E, et al. Detection

of Pneumocystis carinii in respiratory specimens byPCR-solution hybridization enzyme-linked immuno-assay. J Clin Microbiol 1997; 35: 15891591.

67. Khan MA, Farrag N, Butcher P. Diagnosis ofPneumocystis carinii pneumonia: immunofluores-cence staining, simple PCR or nPCR. J Infect 1999;39: 7780.

68. Blumenfeld W, Kovacs JA. Use of monoclonal

antibody to detect Pneumocystis carinii in inducedsputum and bronchoalveolar lavage fluid by immuno-peroxidase staining. Arch Pathol Lab Med 1988; 112:12331236.

69. Wolfson JS, Waldron MA, Sierra LS. Blindedcomparison of a direct immunofluorescence mono-clonal antibody staining method and Giemsa stainingmethod for identification of Pneumocystis carinii ininduced sputum and bronchoalveolar lavage speci-mens of patients infected with human immuno-deficiency virus. J Clin Microbiol1990; 28: 21362138.

70. Cregan P, Yamamoto A, Lum A, VanderHeide T,MacDonald M, Pulliam L. Comparison of fourmethods for rapid detection of Pneumocystis cariniiin respiratory specimens. J Clin Microbiol 1990; 28:

24322436.71. Stratton N, Hryniewicki J, Aarnaes SL, Tan G, de la

Maza LM, Peterson EM. Comparison of monoclonalantibody and calcofluor white stains for the detectionof Pneumocystis carinii from respiratory specimens.J Clin Microbiol 1991; 29: 645647.

72. Ng VL, Geaghan SM, Leoung G, et al. Lack of effectof prophylactic aerosolized pentamidine on the detec-tion of Pneumocystis carinii in induced sputum orbronchoalveolar lavage specimens. Arch Pathol LabMed 1993; 117: 493496.

73. Leibovitz E, Pollack H, Moore T, et al. Comparisonof PCR and standard cytological staining for detectionof Pneumocystis carinii from respiratory specimensfrom patients with or at high risk for infection by

human immunodeficiency virus. J Clin Microbiol1995;33: 30043007.

74. Rafanan AL, Klevjer-Anderson P, Metersky ML.Pneumocystis carinii pneumonia diagnosed by noninduced sputum stained with a direct fluorescentantibody. Ann Clin Lab Sci 1998; 28: 99103.

75. Bustamante EA, Levy H. Sputum induction comparedwith bronchoalveolar lavage by Ballard catheter todiagnose Pneumocystis cariniipneumonia. Chest 1994;105: 816822.

76. Roux P, Lavard I, Poirot JL, et al. Usefulness of PCRfor detection of Pneumocystis carinii DNA. J ClinMicrobiol 1994; 32: 23242326.


Documents

2002 metanalysis PcP