Acta Paediatr 81:227-30. 1992
Modification of some markers of inflammation during treatment for acute respiratory exacerbation in cystic fibrosis EA Valletta', A Rigo?, L Bonazzi?, L Zanolla3 and G Mastella' Cystic Fibrosis Center'. /he lnslilure ?f Chemisrry and Clinical Microscopy qf rhe Universiry2 and rhe DeparImenr,bf Cardiology. Verona. rtaiy3
Valletta EA, Rigo A, Bonazzi L, Zanolla L, Mastella G. Modification of some markers of inflammation during treatment for acute respiratory exacerbation in cystic fibrosis. Acta Paediatr 1992;81:227-30. Stockholm. ISSN 0803-5253 A n objective approach for monitoring the treatment of acute pulmonary exacerbation in cystic fibrosis was evaluated. Eleven biochemical markers of inflammation (erythrocyte sedimentation rate, neutrophil count, C-reactive protein, a-I antitrypsin, haptoglobin, ceruloplasmin, fibronectin, a-l glycoprotein, a-2 macroglobulin,C3, granulocyte elastase and anti-Pseudomonas IgG) were measured in blood serum and plasma from 46 cystic fibrosis patients with chronic Pseudomonas aeruginosa colonization before and after treatment. The overall outcome in each patient was evaluated by means of a pondered sum of clinical, chest X-ray and lung function scores. Biochemical markers were related to the overall clinical improvement: haptoglobin, ceruloplasmin, fibronectin and a- I glycoprotein showed a good sensitivity (64-70'%), specificity (60-70%) and positive predictive value (86-89%). Granulocyte elastase showed a similar sensitivity(67%) and positive predictive value (850/) but a lower specificity (33%).The negative predictive value was generally poor (32-39%). Our data suggest that the combined measurement of some markers of inflammation and of conventional clinical parameters, may help in evaluating the efficacy of anti-infective treatment in cystic fibrosis. 0 Acute-phaseproteins, cystic fibrosis. injammation, pulmonary exacerbation treatment
C Mastella. Cjstic Fibrosis Center, Ospedale Cioile Maggiore, Piazzale Stqfani 2, 37126 Verona. Italy
Assessment of treatment efficacy during respiratory exacerbation in patients with cystic fibrosis (CF) is often subjective, especially when only clinical aspects are evaluated ( I , 2 ) . Functional, radiologic and microbiologic criteria may offer a more reliable picture of the situation although they are variably influenced by different therapeutic supports (antibiotics, physiotherapy, nutritional support, etc.) ( I -5). A more objective approach could be based on biological markers of infection, but definite agreement on this has not yet been reached (6-9). As a consequence of our preliminary experience ( I 0), we decided to investigate if the combination of conventional criteria and some biological markers of inflammation would help to estimate the efficacy of treatment of respiratory exacerbation in CF.
Methods Forty-six CF patients ( 3 2 male, 14 female; age range 333 years, mean 15.8 f6.9 years) were included in the study. All subjects had been colonized by Pseudomonas aeruginosa strains for at least six months. Patients were hospitalized presenting clinical and/or radiologic signs of pulmonary exacerbation according to the criteria suggested by Smith et al. (7). Patients were not included
if any of the following conditions were present: glucocorticoid therapy, parenteral antibiotic therapy in the previous week, and vital capacity approximately < 40% of the predicted value. Anti-P. aeruginosa treatment (according to in vitro susceptibilities) was administered iv three times a day for 10-14 days. The role of viral agents was not investigated. None of the patients received any anti-inflammatory drugs. Before and after treatment all patients were assessed for clinical status and pulmonary function if cooperating. Standard chest roentgenograms were obtained at the time of the enrolment and again at the end of the therapy if new infiltrates had been observed. Blood specimens were collected for laboratory investigation before and after treatment. Pat ient eualua t ion Patients were assessed clinically by means of a scoring system based on 1 1 variables: physical activity, sleep, body temperature, cough, sputum, chest auscultation, dyspnea, cyanosis, appetite, weight and capillary PCO2. Each variable was graded on a scale from 0 (absent or normal) to 3 (severe or pathological). In each patient an average clinical score (range 0-3) was obtained by dividing the sum of the scores by the number of the variables evaluated. The chest X-ray was blind evalu-
228
EA Vullettu r t ul.
ated by one of the authors (GM) according to the Chrispin-Norman score. Pulmonary function was measured with a computerized bell spirometer (Biomedin, Padova, Italy) by a technician not involved in the study. The forced expiratory volume in one second (FEVI) was expressed as a percentage of the predicted value. The overall outcome after treatment was evaluated in each patient by averaging the changes (post-treatment vs pre-treatment) observed in clinical score, X-ray score and pulmonary function. Briefly, a modification of these three variables was expressed as a percentage of the best relative result obtained among all subjects. To obtain an average final score, the percentage value for each criterion was graded (score I (0-33‘1/0) up to 3 (66IOO‘X)) and the sum of the scores was divided by the number of criteria (radiologic, clinical and functional) evaluated in each subject. According to this result (distributed from 1 to 3), the patients were then assigned to one of the following groups: group 1 (score 1 to 1.7), unchanged; group 2 (score 1.8 to 2.4), moderate improvement; group 3 (score 2.5 to 3), evident improvement. Lahoraiory investigaf ions ESR was evaluated by conventional methods. Neutrophi1 count was performed automatically. CRP, a-1 antitrypsin, haptoglobin, ceruloplasmin, a- I glycoprotein, a-2 macroglobulin and C3 fragment were measured in blood serum by rate nephelometry (Auto Immunochemical System, Beckman Analytical, Milano, Italy) ( 1 I). Fibronectin concentration ( 1 2 ) was assessed in plasma by the same method. Granulocyte elastase-a- 1 protease inhibitor complex was determined in plasma by an enzyme-linked immunoassay (ELISA) (Merck, FRG) ( 1 3). Serum anti-P. aeruginosa IgG were measured by an ELISA system. The P. aeruginosu antigens (0-1 7 0-groups, protein concentration 12.8 g/l, kindly provided by Dr N Hoiby, State Serum Institute, Copenhagen, Denmark) were diluted 1 :500 in phosphatebuffered saline (PBS), pH 7.4 and incubated overnight at 4°C in a 96-well microplate (Nunc Immunoplate I). After extensive washings a quenching solution of 5 mg/ ml bovine serum albumin in PBS (PBS-BSA), pH 7.4 was added to the wells. After 2 h of incubation at room temperature, the wells were washed and patients’ sera, diluted 1 : lo4 in PBS, were incubated in triplicate for 1 h at room temperature. A pool of sera with high anti-P. aeruginosa titre, diluted 1 :2000 to 1 : 128 000 was used as a standard curve on each plate. Blank wells with PBS were included. After washings, rabbit anti-human IgG (Miles Scientific, Milano, Italy), diluted I : 1000 in PBS/ BSA was incubated for 1 h at room temperature. Goat anti-rabbit immunoglobulin horse-radish peroxidase (HRP) conjugate (Bio-Rad Laboratories, Milano, Italy), diluted I :3000 was then seeded for 1 h at room temperature. A fresh 0-phenylenediamine/citratephosphate buffer 0.15 M, pH 5 solution was added to
ACTA PRDIATR 81 (1992)
react with the H R P bound to the wells. The reaction was stopped after 10 min with 1 N H2S04.Absorbance at 490 nm was measured immediately. Anti-P. aeruginosa IgG titre was calculated by the method of Brett et al. (14). The logarithm of the titre was used for statistical analysis. Statistical analysis The chi-square ‘goodness-of-fit’ test was used for each variable in order to test the hypothesis that the observations were normally distributed. The comparison between different groups of patients was performed using parametric methods (analysis of variance, Student’s t-test) for normally distributed variables and distribution-free methods (Kruskall-Wallis test, MannWhitney U-test) for non-normal variables. In order to assess sensitivity, specificity, positive and negative predictive values of selected variables as a “diagnostic” criterion, a cut-off value was selected, according to the normal distribution (setting a at 0.05 and fl at 0.10) for normally distributed data, and by the evaluation of the plot of sensitivity and specificity for non-normal variables ( 1 5).
Results According to the overall outcome, the 46 patients were divided in three groups: I0 subjects (mean age 15.6f 7 . 2 years) were assigned to group 1 (unchanged); 19 ( I 5.7 f6.1 years) to group 2 (moderate improvement); and 17 (16f6.6 years) to group 3 (evident improvement). There was no statistical difference in mean age between the three groups or between each and the whole population. Mean values of the biochemical markers, assessed before and after treatment, are shown in Table 1. Pre-treatment data in the whole population (46 patients) were analyzed to assess the characteristics of their distribution. A normal distribution ( p < 0.05) was observed for concentrations of haptoglobin, a-2 macroglobulin, ceruloplasmin, fibronectin, a-1 glycoprotein, C3 and anti-P. aeruginosa IgG. Mean values of each parameter in the three groups before treatment were statistically comparable. The pre-treatment mean value of each marker in the 46 patients was assumed to be the pre-treatment reference value and it was compared to the post-treatment mean level in the three groups separately (Table I). No significant modification was observed in group 1 patients after therapy. Haptoglobin, ceruloplasmin and a- 1 glycoprotein were significantly changed in subjects in group 2. Haptoglobin, ceruloplasmin, fibronectin, a- 1 glycoprotein and granulocyte elastase concentrations were significantly modified in patients in group 3. Assuming a moderate or evident improvement (groups 2 and 3, respectively) as a positive result, sensitivity, specificity, positive and negative predicting values of
Markers ?f influmniution in cystic fibrosis
ACTA PAEDIATR 81 (1992)
229
Tuhle 1. Biochemical markers before and after treatment. Values are mean+SD.
Before therapy” Biochemical marker
All patients
ESR (mm/h) Neutrophils ( x 103/mm3) CRP (mg/dl) a-1 antitrypsin (mg/dl) Haptoglobin (mg/dl) Ceruloplasmin (mg/dl) Fibronectin (mg/dl) r*- I glycoprotein (mg/dl) r*-2 macroglobulin (mgidl) C3 (mg/dl) Granulocyte elastase (pg/l) Anti-Pseudomonas IgG (log)
31 +22 4.9 2.4 0.95 k I .9 216+33 166+53 40+6 27+7 82 k 20 210k32 154+27 221 k 9 9 2.42 f 0.77
+
Group I
+
35 26 4.1k1.8 0.95 k 0.28 205 36 175+50 41 + 6 27k7 8 4 k 16 209 f33 162+24 242+ 100 0.79k0.65
+
After therapyh
Group 2
Group 3
Group I
Group 2
Group 3
24+ I 7 5.0k2.5 I .03 2.46 217+ 34.2 154+55 38+5 28+6 78k I9 212+34 148+27 227+ 101 2.3k0.66
35+22 5.3 2.7 1.12+1.62 224+ 28 174+50 40k6 26k5 85+21 208 29 156+27 202 89 2.3350.89
31 +28 4.4 I .8 0.36 k 0.65 199+50 146+42 38+5 26k9 68 2 23 208 40 138+27 200 46 2.5+0.9
22+7 4.2 2.4 0.28 k2.51 203+21 120f 42*** 35+3** 29+7 62+ l3*** 212k15 140+ 15 172+ 102 2.3k0.2
23+ 13 4.2 f2.7 0.31 1.4 207 27 118*33*** 36+5* 32 f7** 59+ 12*** 217+24 144+20 140 64*** 2+0.7
+
+
+
+ +
+
+
+
+ +
+
a All patients, n=46; group 1 (unchanged), n = 10; group 2 (moderate improvement), n = 19; group 3 (evident improvement), n= 17. Mean post-treatment values in the three groups were compared to those observed before treatment in the whole population (all patients): * p i 0 . 0 5 , ** p < 0.01, *** p < 0.001.
Table 2. Cut-off value, sensitivity, specificity, positive (PPV) and negative (NPV) predictive values in detecting the overall clinical improvement after treatment.
Biochemical marker Haptoglobin (mgidl) Ceruloplasmin (mg/dl) Fibronectin (mg/dl) a-I glycoprotein (mg/dl) Granulocyte elastase (pg/l)
cut-off value
Sensitivity
130.9 36.4 28.7 64.2 175.8
64 67 70 67 67
(X)
haptoglobin, ceruloplasmin, fibronectin, a- 1 glycoprotein and granulocyte elastase levels were calculated (Table 2). A good sensitivity (64-70%) and a good positive predictive value (86-89%) were observed for all the five biochemical markers. Specificity ranged from 60% to 70%, except for granulocyte elastase (specificity 33%). The negative predictive value was between 32% and 39‘%.
Discussion The criteria (clinical, radiologic, bacteriologic and functional) that are conventionally used to assess the efficacy ofan antimicrobial therapy in CF are often subjective or show great intrinsic variability ( I , 10, 16, 17). Moreover, they are strongly influenced by other therapeutic measures (hospitalization, physiotherapy, nutritional support) that are usually undertaken when an acute exacerbation occurs ( I , 5 , 18). Different scoring systems have been proposed to assess more reliably the patient’s clinical condition (7, 9). Some interesting data, related to the usefulness of serum CRP (6) and of indices of endobronchial infection and inflammation, were recently collected (7). To provide a more objective approach to the evaluation of the efficacy of anti-infective treatment in CF exacerbation, we extended our observation to a large
Specificity
PPV
NPV
(%)
(”/)
(”/I
70 70 70
88 88 89 86 85
35 37 39 33 32
60
33
number of plasma proteins (the so-called “acute phase proteins”) which play an important role in the inflammatory process (19, 20). We also evaluated the plasma concentrations of fibronectin-a multifunctional glycoprotein involved in opsonizing bacteria and immune complexes (2 I)-and granulocyte elastase, used recently to monitor pulmonary infection in CF (22). Our data suggest that five biochemical markers (haptoglobin, ceruloplasmin, fibronectin, a-I glycoprotein and granulocyte elastase) have good sensitivity and satisfactory positive predictive value in detecting clinical improvement after treatment for respiratory exacerbation in CF patients. As far as specificity is concerned, a larger variability was observed and the negative predictive value was generally poor. According to previous observations, ESR and peripheral neutrophil count were unreliable markers of the efficacy of anti-infective measures in CF exacerbation ( I , 9). In our experience, the anti-P. aeruginosa IgG did not show any significant change after treatment. Although levels of specific antibodies seem to be related to the clinical condition of the patient (23), the time elapsed was probably too short to allow any significant modification for the evaluation. “Normal” basal levels of the biochemical markers in CF were not evaluated in our study. Very few data are available on this subject and the variability of chronic pulmonary infection in CF must be taken into account. The concentration of many “acute phase proteins”
230
EA Vulleiiu 1’1 aI.
appears to be increased in CF patients with chronic P. aeruginosa infection, even if their modifications d o not
always correlate with the clinical condition (24). In conclusion, our study suggests that some biological markers of inflammation are significantly correlated with clinical improvement observed after treatment of respiratory exacerbation in CF with chronic P. aerugin o ~ acolonization. Nevertheless, none of the reactants, when considered alone, was reliable enough to be used as the “ideal” marker. We believe that the combined assessment of objective biological criteria and conventional clinical criteria may give a more reliable evaluation of the actual response to treatment. Further studies are needed. A~knolc,/edgen~eni.s.The authors are grateful to Drs Rossella Rolfini and Cristina Fiorentini for editing the manuscript.
References I . Smith AL. Antibiotic therapy in cystic fibrosis: evaluation of clinical trials. J Pediatr 1986:108:866 70 2. Wood RE. What is “pulmonary exacerbation” in cystic fibrosis? J Pediatr 1987:l I1:841-2 3. Van den Broek PJ. Antibiotic therapy in cystic fibrosis. Discussion. Chest 1988:94(suppI):155-6 4. Redding GJ, Restuccio R, Cotton EK. et al. Serial changes in pulmonary functions in children hospitalized with cystic fibrosis. Am Rev Respir Dis 1982;I26:3 1-6 5. Cerny FJ, Cropp GJA, Bye MR. Hospital therapy improves exercise tolerance and lung function in cystic fibrosis. Am J Dis Child 1984;138:261 -5 6. Glass S. Hayward C. Govan JRW. Serum C-reactive protein in assessment of pulmonary exacerbations and antimicrobial therapy in cystic fibrosis. J Pediatr 1988:l 13:76-9 7. Smith AL. Redding G, Doershuk C. et al. Sputum changes associated with therapy for endobronchial exacerbation in cystic fibrosis. J Pediatr 1988;112:547 54 8. Moreton RE, Kennedy CR. C reactive protein concentrations in cystic fibrosis. Arch Dis Child 1988;63:958-60 9. Bosso JA, Walker KB. Lack of correlation between objective indicators and clinical response scores during antimicrobial
ACTA PRDIATR R I (1992)
therapy for acute pulmonary exacerbations of cystic fibrosis. Clin Pharm 1988;7:897-901 10. Mastella G. Antibiotic therapy in cystic fibrosis. Discussion. Chest I988:94(suppl):154-5 I . Sternberg JC. A rate nephelometer for measuring specific proteins by immunoprecipitating reactions. Clin Chem 1977:23:1456-62 2. Gabrielli GB, Casaril M. Bonazzi L, Baracchino F, Bellisola G . Corrocher R. Plasma fibronectin in liver cirrhosis and its diagnostic value. Clin Chim Acta 1986:160:289-96 3. Neumann S. Gunzer G. Hennrich N, Lang H. “PMN-elastase assay”: enzyme-immunoassay for human polymorphonuclear elastase in complex with a-I proteinase inhibitor. J Clin Chem Clin Biochem 1984;22:693 7 14. Brett MM. Ghonein ATM, Littlewood JM. Serum antibodies to Pseudomonas aeruginosa in cystic fibrosis. Arch Dis Child 1986;61:1II4-20 15. Gerhardt W, Keller H. Evaluation of test data from clinical studies. Scand J Clin Lab Invest 1986;46(suppl I):l 74 16. Wientzen R, Prestidge C D , Kramer RI, McCracken G H , Nelson JD. Acute pulmonary exacerbations in cystic fibrosis. Am J Dis Child 1980:1341134-8 17. MacLusky 1, Levison H, Gold R, MacLaughin J. Inhaled antibiotics in cystic fibrosis: is there a therapeutic effect? J Pediatr 1986;108:861-5 18. Gold R. Mild to moderate chest exacerbations: d o antibiotics help? Pediatr Pulmonol 1987;(suppl 1):38-9 19. Whicher JT. Dieppe PA. Acute phase proteins. CIin lmmunol Allergy 1985;5:425-46 20. Stuart J, Whicher JT. Tests for detectingand monitoring theacute phase response. Arch Dis Child 1988;63:115 7 21. Mosher DF. Physiology of fibronectin. Annu Rev Med 1984:35:561-75 22. Ericsson Hollsing A, Lantz B. Bergstrom K, Malmborg AS, Strandvik B. Granulocyte elastase-a-lantiproteinase complex in cystic fibrosis: sensitive plasma assay for monitoring pulmonary infections. J Pediatr 1987;l I1:206-1 I 23. Doring G, Albus A. Hoiby N . Immunologic aspects of cystic fibrosis. Chest 1988;94(suppl):109-14 24. Hoiby N, Jacobson L, Jorgensen BA, Lykkegaard E. Weeke B. Pseudonionasaeruginosa infection in cystic fibrosis. Occurrence of precipitating antibodies against Pseudomonas aeruginosa in relation to the concentration ofsixteen serum proteins and theclinical and radiographical status of the lungs. Acta Paediatr Scand 1974;63:843-8 Received June 25, 1990. Accepted May 3, 1991
Modification of some markers of inflammation during treatment for acute respiratory exacerbation in cystic fibrosis EA Valletta', A Rigo?, L Bonazzi?, L Zanolla3 and G Mastella' Cystic Fibrosis Center'. /he lnslilure ?f Chemisrry and Clinical Microscopy qf rhe Universiry2 and rhe DeparImenr,bf Cardiology. Verona. rtaiy3
Valletta EA, Rigo A, Bonazzi L, Zanolla L, Mastella G. Modification of some markers of inflammation during treatment for acute respiratory exacerbation in cystic fibrosis. Acta Paediatr 1992;81:227-30. Stockholm. ISSN 0803-5253 A n objective approach for monitoring the treatment of acute pulmonary exacerbation in cystic fibrosis was evaluated. Eleven biochemical markers of inflammation (erythrocyte sedimentation rate, neutrophil count, C-reactive protein, a-I antitrypsin, haptoglobin, ceruloplasmin, fibronectin, a-l glycoprotein, a-2 macroglobulin,C3, granulocyte elastase and anti-Pseudomonas IgG) were measured in blood serum and plasma from 46 cystic fibrosis patients with chronic Pseudomonas aeruginosa colonization before and after treatment. The overall outcome in each patient was evaluated by means of a pondered sum of clinical, chest X-ray and lung function scores. Biochemical markers were related to the overall clinical improvement: haptoglobin, ceruloplasmin, fibronectin and a- I glycoprotein showed a good sensitivity (64-70'%), specificity (60-70%) and positive predictive value (86-89%). Granulocyte elastase showed a similar sensitivity(67%) and positive predictive value (850/) but a lower specificity (33%).The negative predictive value was generally poor (32-39%). Our data suggest that the combined measurement of some markers of inflammation and of conventional clinical parameters, may help in evaluating the efficacy of anti-infective treatment in cystic fibrosis. 0 Acute-phaseproteins, cystic fibrosis. injammation, pulmonary exacerbation treatment
C Mastella. Cjstic Fibrosis Center, Ospedale Cioile Maggiore, Piazzale Stqfani 2, 37126 Verona. Italy
Assessment of treatment efficacy during respiratory exacerbation in patients with cystic fibrosis (CF) is often subjective, especially when only clinical aspects are evaluated ( I , 2 ) . Functional, radiologic and microbiologic criteria may offer a more reliable picture of the situation although they are variably influenced by different therapeutic supports (antibiotics, physiotherapy, nutritional support, etc.) ( I -5). A more objective approach could be based on biological markers of infection, but definite agreement on this has not yet been reached (6-9). As a consequence of our preliminary experience ( I 0), we decided to investigate if the combination of conventional criteria and some biological markers of inflammation would help to estimate the efficacy of treatment of respiratory exacerbation in CF.
Methods Forty-six CF patients ( 3 2 male, 14 female; age range 333 years, mean 15.8 f6.9 years) were included in the study. All subjects had been colonized by Pseudomonas aeruginosa strains for at least six months. Patients were hospitalized presenting clinical and/or radiologic signs of pulmonary exacerbation according to the criteria suggested by Smith et al. (7). Patients were not included
if any of the following conditions were present: glucocorticoid therapy, parenteral antibiotic therapy in the previous week, and vital capacity approximately < 40% of the predicted value. Anti-P. aeruginosa treatment (according to in vitro susceptibilities) was administered iv three times a day for 10-14 days. The role of viral agents was not investigated. None of the patients received any anti-inflammatory drugs. Before and after treatment all patients were assessed for clinical status and pulmonary function if cooperating. Standard chest roentgenograms were obtained at the time of the enrolment and again at the end of the therapy if new infiltrates had been observed. Blood specimens were collected for laboratory investigation before and after treatment. Pat ient eualua t ion Patients were assessed clinically by means of a scoring system based on 1 1 variables: physical activity, sleep, body temperature, cough, sputum, chest auscultation, dyspnea, cyanosis, appetite, weight and capillary PCO2. Each variable was graded on a scale from 0 (absent or normal) to 3 (severe or pathological). In each patient an average clinical score (range 0-3) was obtained by dividing the sum of the scores by the number of the variables evaluated. The chest X-ray was blind evalu-
228
EA Vullettu r t ul.
ated by one of the authors (GM) according to the Chrispin-Norman score. Pulmonary function was measured with a computerized bell spirometer (Biomedin, Padova, Italy) by a technician not involved in the study. The forced expiratory volume in one second (FEVI) was expressed as a percentage of the predicted value. The overall outcome after treatment was evaluated in each patient by averaging the changes (post-treatment vs pre-treatment) observed in clinical score, X-ray score and pulmonary function. Briefly, a modification of these three variables was expressed as a percentage of the best relative result obtained among all subjects. To obtain an average final score, the percentage value for each criterion was graded (score I (0-33‘1/0) up to 3 (66IOO‘X)) and the sum of the scores was divided by the number of criteria (radiologic, clinical and functional) evaluated in each subject. According to this result (distributed from 1 to 3), the patients were then assigned to one of the following groups: group 1 (score 1 to 1.7), unchanged; group 2 (score 1.8 to 2.4), moderate improvement; group 3 (score 2.5 to 3), evident improvement. Lahoraiory investigaf ions ESR was evaluated by conventional methods. Neutrophi1 count was performed automatically. CRP, a-1 antitrypsin, haptoglobin, ceruloplasmin, a- I glycoprotein, a-2 macroglobulin and C3 fragment were measured in blood serum by rate nephelometry (Auto Immunochemical System, Beckman Analytical, Milano, Italy) ( 1 I). Fibronectin concentration ( 1 2 ) was assessed in plasma by the same method. Granulocyte elastase-a- 1 protease inhibitor complex was determined in plasma by an enzyme-linked immunoassay (ELISA) (Merck, FRG) ( 1 3). Serum anti-P. aeruginosa IgG were measured by an ELISA system. The P. aeruginosu antigens (0-1 7 0-groups, protein concentration 12.8 g/l, kindly provided by Dr N Hoiby, State Serum Institute, Copenhagen, Denmark) were diluted 1 :500 in phosphatebuffered saline (PBS), pH 7.4 and incubated overnight at 4°C in a 96-well microplate (Nunc Immunoplate I). After extensive washings a quenching solution of 5 mg/ ml bovine serum albumin in PBS (PBS-BSA), pH 7.4 was added to the wells. After 2 h of incubation at room temperature, the wells were washed and patients’ sera, diluted 1 : lo4 in PBS, were incubated in triplicate for 1 h at room temperature. A pool of sera with high anti-P. aeruginosa titre, diluted 1 :2000 to 1 : 128 000 was used as a standard curve on each plate. Blank wells with PBS were included. After washings, rabbit anti-human IgG (Miles Scientific, Milano, Italy), diluted I : 1000 in PBS/ BSA was incubated for 1 h at room temperature. Goat anti-rabbit immunoglobulin horse-radish peroxidase (HRP) conjugate (Bio-Rad Laboratories, Milano, Italy), diluted I :3000 was then seeded for 1 h at room temperature. A fresh 0-phenylenediamine/citratephosphate buffer 0.15 M, pH 5 solution was added to
ACTA PRDIATR 81 (1992)
react with the H R P bound to the wells. The reaction was stopped after 10 min with 1 N H2S04.Absorbance at 490 nm was measured immediately. Anti-P. aeruginosa IgG titre was calculated by the method of Brett et al. (14). The logarithm of the titre was used for statistical analysis. Statistical analysis The chi-square ‘goodness-of-fit’ test was used for each variable in order to test the hypothesis that the observations were normally distributed. The comparison between different groups of patients was performed using parametric methods (analysis of variance, Student’s t-test) for normally distributed variables and distribution-free methods (Kruskall-Wallis test, MannWhitney U-test) for non-normal variables. In order to assess sensitivity, specificity, positive and negative predictive values of selected variables as a “diagnostic” criterion, a cut-off value was selected, according to the normal distribution (setting a at 0.05 and fl at 0.10) for normally distributed data, and by the evaluation of the plot of sensitivity and specificity for non-normal variables ( 1 5).
Results According to the overall outcome, the 46 patients were divided in three groups: I0 subjects (mean age 15.6f 7 . 2 years) were assigned to group 1 (unchanged); 19 ( I 5.7 f6.1 years) to group 2 (moderate improvement); and 17 (16f6.6 years) to group 3 (evident improvement). There was no statistical difference in mean age between the three groups or between each and the whole population. Mean values of the biochemical markers, assessed before and after treatment, are shown in Table 1. Pre-treatment data in the whole population (46 patients) were analyzed to assess the characteristics of their distribution. A normal distribution ( p < 0.05) was observed for concentrations of haptoglobin, a-2 macroglobulin, ceruloplasmin, fibronectin, a-1 glycoprotein, C3 and anti-P. aeruginosa IgG. Mean values of each parameter in the three groups before treatment were statistically comparable. The pre-treatment mean value of each marker in the 46 patients was assumed to be the pre-treatment reference value and it was compared to the post-treatment mean level in the three groups separately (Table I). No significant modification was observed in group 1 patients after therapy. Haptoglobin, ceruloplasmin and a- 1 glycoprotein were significantly changed in subjects in group 2. Haptoglobin, ceruloplasmin, fibronectin, a- 1 glycoprotein and granulocyte elastase concentrations were significantly modified in patients in group 3. Assuming a moderate or evident improvement (groups 2 and 3, respectively) as a positive result, sensitivity, specificity, positive and negative predicting values of
Markers ?f influmniution in cystic fibrosis
ACTA PAEDIATR 81 (1992)
229
Tuhle 1. Biochemical markers before and after treatment. Values are mean+SD.
Before therapy” Biochemical marker
All patients
ESR (mm/h) Neutrophils ( x 103/mm3) CRP (mg/dl) a-1 antitrypsin (mg/dl) Haptoglobin (mg/dl) Ceruloplasmin (mg/dl) Fibronectin (mg/dl) r*- I glycoprotein (mg/dl) r*-2 macroglobulin (mgidl) C3 (mg/dl) Granulocyte elastase (pg/l) Anti-Pseudomonas IgG (log)
31 +22 4.9 2.4 0.95 k I .9 216+33 166+53 40+6 27+7 82 k 20 210k32 154+27 221 k 9 9 2.42 f 0.77
+
Group I
+
35 26 4.1k1.8 0.95 k 0.28 205 36 175+50 41 + 6 27k7 8 4 k 16 209 f33 162+24 242+ 100 0.79k0.65
+
After therapyh
Group 2
Group 3
Group I
Group 2
Group 3
24+ I 7 5.0k2.5 I .03 2.46 217+ 34.2 154+55 38+5 28+6 78k I9 212+34 148+27 227+ 101 2.3k0.66
35+22 5.3 2.7 1.12+1.62 224+ 28 174+50 40k6 26k5 85+21 208 29 156+27 202 89 2.3350.89
31 +28 4.4 I .8 0.36 k 0.65 199+50 146+42 38+5 26k9 68 2 23 208 40 138+27 200 46 2.5+0.9
22+7 4.2 2.4 0.28 k2.51 203+21 120f 42*** 35+3** 29+7 62+ l3*** 212k15 140+ 15 172+ 102 2.3k0.2
23+ 13 4.2 f2.7 0.31 1.4 207 27 118*33*** 36+5* 32 f7** 59+ 12*** 217+24 144+20 140 64*** 2+0.7
+
+
+
+ +
+
+
+
+ +
+
a All patients, n=46; group 1 (unchanged), n = 10; group 2 (moderate improvement), n = 19; group 3 (evident improvement), n= 17. Mean post-treatment values in the three groups were compared to those observed before treatment in the whole population (all patients): * p i 0 . 0 5 , ** p < 0.01, *** p < 0.001.
Table 2. Cut-off value, sensitivity, specificity, positive (PPV) and negative (NPV) predictive values in detecting the overall clinical improvement after treatment.
Biochemical marker Haptoglobin (mgidl) Ceruloplasmin (mg/dl) Fibronectin (mg/dl) a-I glycoprotein (mg/dl) Granulocyte elastase (pg/l)
cut-off value
Sensitivity
130.9 36.4 28.7 64.2 175.8
64 67 70 67 67
(X)
haptoglobin, ceruloplasmin, fibronectin, a- 1 glycoprotein and granulocyte elastase levels were calculated (Table 2). A good sensitivity (64-70%) and a good positive predictive value (86-89%) were observed for all the five biochemical markers. Specificity ranged from 60% to 70%, except for granulocyte elastase (specificity 33%). The negative predictive value was between 32% and 39‘%.
Discussion The criteria (clinical, radiologic, bacteriologic and functional) that are conventionally used to assess the efficacy ofan antimicrobial therapy in CF are often subjective or show great intrinsic variability ( I , 10, 16, 17). Moreover, they are strongly influenced by other therapeutic measures (hospitalization, physiotherapy, nutritional support) that are usually undertaken when an acute exacerbation occurs ( I , 5 , 18). Different scoring systems have been proposed to assess more reliably the patient’s clinical condition (7, 9). Some interesting data, related to the usefulness of serum CRP (6) and of indices of endobronchial infection and inflammation, were recently collected (7). To provide a more objective approach to the evaluation of the efficacy of anti-infective treatment in CF exacerbation, we extended our observation to a large
Specificity
PPV
NPV
(%)
(”/)
(”/I
70 70 70
88 88 89 86 85
35 37 39 33 32
60
33
number of plasma proteins (the so-called “acute phase proteins”) which play an important role in the inflammatory process (19, 20). We also evaluated the plasma concentrations of fibronectin-a multifunctional glycoprotein involved in opsonizing bacteria and immune complexes (2 I)-and granulocyte elastase, used recently to monitor pulmonary infection in CF (22). Our data suggest that five biochemical markers (haptoglobin, ceruloplasmin, fibronectin, a-I glycoprotein and granulocyte elastase) have good sensitivity and satisfactory positive predictive value in detecting clinical improvement after treatment for respiratory exacerbation in CF patients. As far as specificity is concerned, a larger variability was observed and the negative predictive value was generally poor. According to previous observations, ESR and peripheral neutrophil count were unreliable markers of the efficacy of anti-infective measures in CF exacerbation ( I , 9). In our experience, the anti-P. aeruginosa IgG did not show any significant change after treatment. Although levels of specific antibodies seem to be related to the clinical condition of the patient (23), the time elapsed was probably too short to allow any significant modification for the evaluation. “Normal” basal levels of the biochemical markers in CF were not evaluated in our study. Very few data are available on this subject and the variability of chronic pulmonary infection in CF must be taken into account. The concentration of many “acute phase proteins”
230
EA Vulleiiu 1’1 aI.
appears to be increased in CF patients with chronic P. aeruginosa infection, even if their modifications d o not
always correlate with the clinical condition (24). In conclusion, our study suggests that some biological markers of inflammation are significantly correlated with clinical improvement observed after treatment of respiratory exacerbation in CF with chronic P. aerugin o ~ acolonization. Nevertheless, none of the reactants, when considered alone, was reliable enough to be used as the “ideal” marker. We believe that the combined assessment of objective biological criteria and conventional clinical criteria may give a more reliable evaluation of the actual response to treatment. Further studies are needed. A~knolc,/edgen~eni.s.The authors are grateful to Drs Rossella Rolfini and Cristina Fiorentini for editing the manuscript.
References I . Smith AL. Antibiotic therapy in cystic fibrosis: evaluation of clinical trials. J Pediatr 1986:108:866 70 2. Wood RE. What is “pulmonary exacerbation” in cystic fibrosis? J Pediatr 1987:l I1:841-2 3. Van den Broek PJ. Antibiotic therapy in cystic fibrosis. Discussion. Chest 1988:94(suppI):155-6 4. Redding GJ, Restuccio R, Cotton EK. et al. Serial changes in pulmonary functions in children hospitalized with cystic fibrosis. Am Rev Respir Dis 1982;I26:3 1-6 5. Cerny FJ, Cropp GJA, Bye MR. Hospital therapy improves exercise tolerance and lung function in cystic fibrosis. Am J Dis Child 1984;138:261 -5 6. Glass S. Hayward C. Govan JRW. Serum C-reactive protein in assessment of pulmonary exacerbations and antimicrobial therapy in cystic fibrosis. J Pediatr 1988:l 13:76-9 7. Smith AL. Redding G, Doershuk C. et al. Sputum changes associated with therapy for endobronchial exacerbation in cystic fibrosis. J Pediatr 1988;112:547 54 8. Moreton RE, Kennedy CR. C reactive protein concentrations in cystic fibrosis. Arch Dis Child 1988;63:958-60 9. Bosso JA, Walker KB. Lack of correlation between objective indicators and clinical response scores during antimicrobial
ACTA PRDIATR R I (1992)
therapy for acute pulmonary exacerbations of cystic fibrosis. Clin Pharm 1988;7:897-901 10. Mastella G. Antibiotic therapy in cystic fibrosis. Discussion. Chest I988:94(suppl):154-5 I . Sternberg JC. A rate nephelometer for measuring specific proteins by immunoprecipitating reactions. Clin Chem 1977:23:1456-62 2. Gabrielli GB, Casaril M. Bonazzi L, Baracchino F, Bellisola G . Corrocher R. Plasma fibronectin in liver cirrhosis and its diagnostic value. Clin Chim Acta 1986:160:289-96 3. Neumann S. Gunzer G. Hennrich N, Lang H. “PMN-elastase assay”: enzyme-immunoassay for human polymorphonuclear elastase in complex with a-I proteinase inhibitor. J Clin Chem Clin Biochem 1984;22:693 7 14. Brett MM. Ghonein ATM, Littlewood JM. Serum antibodies to Pseudomonas aeruginosa in cystic fibrosis. Arch Dis Child 1986;61:1II4-20 15. Gerhardt W, Keller H. Evaluation of test data from clinical studies. Scand J Clin Lab Invest 1986;46(suppl I):l 74 16. Wientzen R, Prestidge C D , Kramer RI, McCracken G H , Nelson JD. Acute pulmonary exacerbations in cystic fibrosis. Am J Dis Child 1980:1341134-8 17. MacLusky 1, Levison H, Gold R, MacLaughin J. Inhaled antibiotics in cystic fibrosis: is there a therapeutic effect? J Pediatr 1986;108:861-5 18. Gold R. Mild to moderate chest exacerbations: d o antibiotics help? Pediatr Pulmonol 1987;(suppl 1):38-9 19. Whicher JT. Dieppe PA. Acute phase proteins. CIin lmmunol Allergy 1985;5:425-46 20. Stuart J, Whicher JT. Tests for detectingand monitoring theacute phase response. Arch Dis Child 1988;63:115 7 21. Mosher DF. Physiology of fibronectin. Annu Rev Med 1984:35:561-75 22. Ericsson Hollsing A, Lantz B. Bergstrom K, Malmborg AS, Strandvik B. Granulocyte elastase-a-lantiproteinase complex in cystic fibrosis: sensitive plasma assay for monitoring pulmonary infections. J Pediatr 1987;l I1:206-1 I 23. Doring G, Albus A. Hoiby N . Immunologic aspects of cystic fibrosis. Chest 1988;94(suppl):109-14 24. Hoiby N, Jacobson L, Jorgensen BA, Lykkegaard E. Weeke B. Pseudonionasaeruginosa infection in cystic fibrosis. Occurrence of precipitating antibodies against Pseudomonas aeruginosa in relation to the concentration ofsixteen serum proteins and theclinical and radiographical status of the lungs. Acta Paediatr Scand 1974;63:843-8 Received June 25, 1990. Accepted May 3, 1991
