Respiratory Medicine (2009) 103, 595e600

ava i lab le at www.sc ienced i rec t . com

j ourna l homepage : www.e lsev ier . com/ loca te / rmed

Influence of parecoxib (cox-2 inhibitor) inexperimental pleurodesis

Lisete R. Teixeira*, Francisco S. Vargas, Milena M.P. Acencio,Leila Antonangelo, Natalia P. Novaes, Jose D. Costa Jr, Evaldo Marchi

Laboratory of Pleura (Pulmonary Division), Heart Institute (InCor) of the University of Sao Paulo Medical School,Sao Paulo, Brazil

Received 8 August 2008; accepted 20 October 2008Available online 29 November 2008

KEYWORDSPleura;Parecoxib;Talc;Silver nitrate;Pleurodesis

* Corresponding author. CopacabanaE-mail addresses: lisetepneumo@y

0954-6111/$ - see front matter ª 200doi:10.1016/j.rmed.2008.10.016

Summary

Background: The intrapleural instillation of a sclerosing agent produces an inflammatoryprocess frequently followed by pain. The treatment can include the use of analgesics oranti-inflammatory drugs. Previously, it was demonstrated (experimental studies) that corti-coids and nonsteroidal anti-inflammatory drugs (diclofenac) reduce the inflammation andfibrosis produced by talc but not by transforming growth factor-b or silver nitrate. The objec-tive of this study was to determine whether parecoxib (COX-2 inhibitor) affects pleurodesisinduced by talc or silver nitrate.Methods: 140 rabbits received intrapleural injection (2 mL) of 400 mg/kg of talc or 0.5% silvernitrate. A subgroup of 70 animals received additional daily intramuscular parecoxib (1 mg/kg).They were sacrificed at 4, 24, 48, 72 h or 7, 14, or 28 days after the procedure. The pleuralfluid was quantified; biochemical examinations (glucose, lactic dehydrogenase, and proteins)and immunologic dosages (interleukin-8, vascular endothelial growth factor, and transforminggrowth factor-b1) were analyzed in pleural fluid and blood. Finally, macro- and microscopicpleura and lung studies were performed.Results: Evaluation after 28 days demonstrated that parecoxib reduced pleural and pulmonaryinflammation but not pleural adhesions. The changes were observed precociously ðy72 hÞ andwere more evident after silver nitrate injection.Conclusion: Systemic parecoxib injection does not interfere with talc or silver nitrate pleurod-esis. These results suggest that use of COX-2 inhibitors can be considered and depending of theresults of other studies, recommended in human pleurodesis.ª 2008 Elsevier Ltd. All rights reserved.

415/73, Sao Paulo e SP, Brazil. Tel.: þ55 11 3069 5695/5034/5125; fax: þ55 11 3069 5643.ahoo.com.br, lisete.teixeira@usp.br (L.R. Teixeira).

8 Elsevier Ltd. All rights reserved.

596 L.R. Teixeira et al.

Introduction

Recurrent accumulation of liquid or air in the pleural spacerepresents one of the biggest challenges in the treatmentof pleural diseases. Chemical pleurodesis is frequentlyindicated and performed by injecting a sclerosant agentinto the pleural space.1 The first step is the development ofan inflammatory process preceding a fusion of the pleuralmembranes and the obliteration of the thoracic cavity.2

This procedure causes a painful process of variedintensity whose treatment includes the administration ofanalgesic and anti-inflammatory drugs that can interferewith a cascade of events that start with inflammation andculminate with collagen deposition and adhesion for-mation.3e6 The relevance of this observation is clear whenwe consider that some of the candidates for pleurodesismay chronically use corticosteroids or even nonhormonalanti-inflammatories. Among those, COX-2 inhibitors areindicated in the management of pain, acute or chronic.7

Although, today we have several potent analgesic drugs,many patients (as rheumatologic diseases) can be usinganti-inflammatories agents during the indication ofa pleurodesis.

Previously (experimental studies),4e6,8 we demonstratedthat corticosteroids, as well as sodium diclofenac, decreaseinflammation and fibrosis produced by intrapleural injec-tion of talc or doxycycline, not interfering with pleurodesisinduced by TGF-b1 or silver nitrate.

In this context, the aim of this study is to answera practical question. Can the use of COX-2 inhibitorsdecrease the efficacy of pleurodesis? If no interference isobserved, these drugs could represent a therapeutic tool inpain control.

Material and methods

This study was developed in the Laboratory of Pleura of theHeart Institute e Brazil (InCor), after the approval of ourinstitutional ethics committee.

A group of 140 New Zealand rabbits weighing 2.0e3.0 kgwas randomized into 4 groups and then subdivided(5 animals) according to the time of sacrifice (4, 24, 48, 72 hor 7, 14, or 28 days).

According to methodology previously described, theanimals were anesthetized with 35 mg/kg of hydrochlorideketamine and 5 mg/kg of hydrochloride Xylazine (intra-muscular) and prepared for aseptic surgery. A 19-gaugeneedle attached to a 5-mL syringe containing the sclerosantwas inserted through the skin and thoracic muscles, atapproximately 1 cm of the parasternal line, between thesixth and seventh intercostal space. The plunger of thesyringe was removed, and the needle was slowly advanceduntil it reached the pleural space, where the subatmo-spheric intrapleural pressure allowed the fluid to enterspontaneously into the pleural cavity.9e11

The animals received (right hemithorax) 400 mg/kg oftalc or 0.5% silver nitrate, both in a total volume of 2 mL.The injected talc is asbestos-free [Mg6(Si8O20) (OH)4] of‘‘mixed size’’ (Magnesita, BA, Brazil), routinely used forpleurodesis. It contains particles of varied size (mean:25.4 mm; range: 6.4e50.5 mm); only 10% of particles are

smaller than 6.66 mm. The noninjected hemithorax (left)was used as the control. In the group of rabbits receivingparecoxib, the medicine was injected intramuscularly(1 mg/kg), 24 h before the procedure, daily during the first7 days, and weekly until the end of the experiment (28days).

According to the defined times, the animals wereeuthanized by an IV pentobarbital (40 mg/kg) injectioninto the marginal ear vein. Immediately, blood sampleswere collected, and a 21-gauge needle was insertedthrough the diaphragm to aspirate the pleural fluid.11 Thecollected fluids were centrifuged and separated forbiochemical analysis by automatized methods (glucose,proteins, and lactic dehydrogenase) and immunologicdosage (ELISA method) for interleukin-8 (IL-8), trans-forming growth factor-b1 (TGF-b1), and vascular endothe-lial growth factor (VEGF) analysis. In sequence, the thoraxwas removed en bloc, and the lungs were expanded with60 mL of 10% formalin and submerged in this solution for atleast 48 h. After this period, each pleural cavity wasexposed by using the method previously described by ourgroup.9,10 Finally, the thorax was opened, and a researcherblinded to the treatment (LRT) evaluated the presence ofmacroscopic pleural adhesions and collected fragmentsfrom the pleura and pulmonary parenchyma for latermicroscopic analysis.

Pleural adherence grades were established by thefollowing scores: (0) normal pleural space; (1) 1e3 smalladherences; (2) over 3 adherences, where lung can easilybe separated from the thorax; (3) generalized adhesions(showing areas where the lung could only be separatedfrom the thoracic wall with difficulty), and (4) completeobliteration of the pleural space by adhesions.

The pleura and lung parenchyma fragments were pro-cessed, stained (hematoxylineeosin), and evaluated for thepresence of inflammation (inflammatory cells as neutro-phils, eosinophils, neovascularization and fibrin) andfibrosis (mononuclear and fibroblasts cells and fibrosis)(graded from 0 e normal to 4 e marked changes). Themicroscopic analysis was done by an experienced examiner(LA) blinded to the treatment.9,10

Statistical analysis

A Sigma Stat 3.5 (System software e CA, USA) program wasused, and the results are reported as mean and SD.Statistical significance was determined via a t test andManneWhitney rank-sum test. Differences were consideredsignificant at p< 0.05.

Results

Pleural fluid volume

In the animals receiving talc, the volume of pleural fluidwas similar, independent of COX-2 administration, exceptin the evaluation performed 72 h after the procedure, whena reduction was observed in the animals treated with theanti-inflammatory (15.4� 4.8 vs 6.4� 2.7 mL; p Z 0.006).After 7 days, the volume was minimal or absent in bothgroups (0.7� 1.6 vs 0.0� 0.0 mL) (Fig. 1).

Figure 1 Effect of COX-2 inhibitor on pleural fluid volume and biochemical examinations after intrapleural talc (TL) or silvernitrate (SN).

Influence of parecoxib in pleurodesis 597

In the silver nitrate group, the fluid increased progres-sively, reaching the highest value at 72 h. In any of thetimes considered, differences due to parecoxib treatmentwere observed (Fig. 1).

Biochemical analysis

Total protein (PT) values in pleural fluid suggest, at alltimes, the presence of an exudate (>3 g/dL). Lacticdehydrogenase levels (DHL) were high in both pleural fluidand blood (pleura/serum> 0.6), characterizing an exuda-tive process. No differences between groups were found(Fig. 1).

Cytokine evaluation

Rabbits receiving talc maintained stable levels of IL-8. Nodifferences between blood and pleural fluid were observed.Four hours after the intrapleural silver nitrate injection,IL-8 levels increased; after 24 h, a reduction was observed.Evaluations performed from 4 to 48 h showed valuessignificantly higher in the groups that did not receive theCOX-2 treatment (Fig. 2).

VEGF, in the pleural fluid, increased progressively in thefirst 72 h of the study, the levels were statistically lower inthe animals receiving talc and COX-2 (p Z 0.027). VEGFlevels observed in the pleural fluid, in all groups and at alltimes, were higher than those obtained from blood(Fig. 2).

TGF-b1 levels obtained from pleural liquid and bloodwere similar throughout the study. Significant reductionafter 48 h of the intrapleural injection were observed in thegroup treated with the anti-inflammatory, independent ofthe sclerosant used (Fig. 2).

Anatomopathological exams

Results obtained are shown in Fig. 3. The left hemithorax(control) in all animals received a 0 score.

Macroscopically, the effect of the treatment with theCOX-2 inhibitor was observed transitorily in the animalsthat received talc. A reduction in pleural adhesions wasnoted after 48 (p Z 0.032), 72 h (p Z 0.048) and 7 days(p Z 0.008). From the second week on and at the finalevaluation (28 days), no differences were observed. In thesilver nitrate group, no differences were observed.

Microscopically, the inflammation due to talc instilla-tion was discrete. In the first 48 h, no differences werecharacterized due to COX-2 administration. We observeda reduction in the pleura (72 h: p Z 0.05; 7 days:p Z 0.038; and 14 days: p Z 0.004) and in the pulmonaryparenchyma (72 h: p Z 0.032; 7 days: p Z 0.008; and 14days: p Z 0.032) related to the COX-2 treatment. After 28days, the inflammation was discrete in both, pleura andlung.

Similar results were observed with silver nitrate. Pleuralinflammation was discrete up to 72 h, with no differencesdue to the parecoxib treatment. From this time, pleuralinflammation increased; however, a significant reductionwas observed after 7 (p Z 0.004), 14 (p Z 0.002), and 28days (p Z 0.011) due to COX-2 administration. The lungparenchyma inflammation showed a similar pattern, withstatistical differences after 72 h (p Z 0.032), and 7(p Z 0.008), 14 (p Z 0.032), and 28 days (p Z 0.005).

Pleural fibrosis became evident after 7 days (2.2� 0.3)of talc instillation and after 72 h (2.5� 0.5) of silver nitrateinstillation. Treatment with COX-2 resulted in a reductionin fibrosis after 72 h (p< 0.001) and 7 days (p Z 0.032) ofthe silver nitrate injection.

Figure 2 Effect of COX-2 inhibitor on IL-8, VEGF, and TGF levels (pleura and blood) after intrapleural talc (TL) or silvernitrate (SN).

598 L.R. Teixeira et al.

Discussion

This study demonstrates that systemic administration ofparecoxib interferes with the inflammatory process thatdevelops in the pleural space after the injection of thesclerosing agent (talc or silver nitrate), but does not reducethe intensity of pleurodesis. It is important to note that thisfact was verified with the talc recommended to inducepleurodesis, which contained mixed particles (90%>6.6 mm) being effective and well tolerated and alsoproducing few local and systemic collateral effects.12

The sclerosing agent when introduced into the pleuralspace damages the mesothelial cells, triggering an inflam-matory process characterized by a neutrophilic exudate.2

The initial changes culminate in the development ofadhesions between the visceral and parietal pleura, oblit-erating the virtual pleural space. Several factors modulatethese reactions including the degree of injury, the ability ofmesothelial cells and fibroblasts to secrete collagen, andthe balance between metalloproteinases and plasminogenactivators. In addition, the relationship among inflamma-tion, cytokines, and fibrinolysis should be considered.2,8 Asthe duration, extension, and intensity of inflammation

established in the pleural space influence the final outcomeof pleurodesis, the use of anti-inflammatory agents caninhibit this process, consequently reducing the pleuraladhesions. These findings have been previously docu-mented in experimental studies,5,8 demonstrating that theadministration of anti-inflammatory drugs (corticosteroidsor diclofenac) reduces the effectiveness of pleurodesisinduced by talc. Conversely, another study6 showed thatthe systemic administration of corticosteroids does notinterfere with pleurodesis induced by the intrapleuralinjection of the transforming growth factor TGF-b1, as wellas the use of diclofenac in pleurodesis induced by silvernitrate.8

The extrapolation of these results to humans raises thequestion of whether previous or concomitant parecoxibtreatment compromises pleurodesis intensity, interferingwith the control of pneumothorax or pleural effusion.

The mechanisms involved in inflammation are complex.It is interesting to observe, in this study, that talc or silvernitrate instilled into the pleural space determined (acutephase), a similar pleural exudate (Light’s criteria13), whoseinflammatory characteristics were more evident in the first24 h. No changes in the volume of fluid accumulated were

Figure 3 Effect of COX-2 inhibitor on gross and microscopic analysis after intrapleural talc (TL) or silver nitrate (SN).

Influence of parecoxib in pleurodesis 599

observed due to the systemic administration of the anti-inflammatory drug.

It is known that intrapleural instillation of a sclerosantcauses significant effects not confined to the pleural cavity;systemic repercussion is observed according to the agentused.14,15 This study confirms this fact, showing that talcand silver nitrate increase cytokine serum levels, demon-strating the presence of a systemic inflammatory processwith predominance in the pleural space.

The importance of interleukin-8 in the development ofpleural inflammation has been established. The immediateincrease (4 h) and the posterior decrement (24 h) of pleuralfluid IL-8 should be observed. In this study, the highestlevels were observed acutely in the pleural effusion afterintrapleural silver nitrate. Noteworthy is the reductionafter parecoxib treatment of IL-8 pleural and blood levelsin animals receiving silver nitrate but not in those injectedwith talc.

The vascular endothelial growth factor levels increasedsubstantially in pleural fluid and discretely in blood. Therelationship between VEGF and the volume of pleural fluidshould be noted; the superposition of the curves suggeststhe value of this angiogenic cytokine in pleural fluid accu-mulation. Besides the participation in the initial inflam-matory process stimulating the activation of othermediators and facilitating inflammatory cell migration, it isimportant in promoting vascular permeability.6,16

Finally, transforming growth factor (TGF-b1) increasedduring all study period with similar levels in serum andpleural fluid. It is a multifunctional cytokine with significantfibrotic action modulating the deposition of connectivetissue and immune reactions.6,16 An experimental studyshowed its importance in the development of pleurodesis;when injected into the pleural space, it is not influenced byanti-inflammatories.6

The pleural injury observed after instillation of scle-rosant agents and the interaction with anti-inflammatorydrugs demonstrate the complexity of the mechanismsinvolved. The laboratory analyses express part of theprocess while the final findings are represented bythe cellular and architectural changes in the pleura and inthe lung, more strongly observed after silver nitrateinstillation. However, a clear reduction in cellular affluxoccurs in parecoxib-treated groups, indicating a minorinflammatory response in both pleura and pulmonaryparenchyma.

In this context, pleural fibrosis was higher in the silvernitrate groups. Although collagen deposition had beenreduced in the acute phase of the animals treated withsilver nitrate and parecoxib, this drug did not compromisethe presence or the degree of macroscopic adhesionspresent in all evaluations. Curiously, in the animals injectedwith talc, parecoxib did not change the microscopic fibrosisnor the volume accumulated in the pleural space, preco-ciously. The reduction of the macroscopic adhesionsobserved in the first days did not interfere in the develop-ment or in the final result of pleurodesis.

A central question refers to the methodology. Theanti-inflammatory drug was administered daily for 7 daysand subsequently once a week until the end of the study(28 days). This plan is based on clinical practice; the useof anti-inflammatories in the induction of pleurodesis hasone single purpose, to relieve pain. During the first days,both the pain and the inflammation are morepronounced, justifying the schedule used. Another pointto be discussed could be the dose used daily insteadtwice a day as in the clinical practice. Although we havebeen used parecoxib once a day, due to be less invasivefor the animal, the dose administrated was the maximumrecommended.

600 L.R. Teixeira et al.

Finally, the study has an evident clinical implication; theextrapolation of these results suggests considering the useof COX-2 inhibitor anti-inflammatories in patients withpneumothorax or recurrent pleural effusion undergoingpleurodesis. The significance of this observation should beemphasized because many patients, who are candidates forpleurodesis, use these anti-inflammatories routinely. Clin-ical studies should be encouraged to define the analgesicprescription and the degree of interference in the intensityof pleurodesis.

Conflict of interest statement

None of the authors have a conflict of interest to declare inrelation to this work.

Acknowledgment

We thank laboratory technicians Carlos S. R. Silva and RickyT. W. Wu and the pharmacist Gabriela G. Carnevale for thehelp during the surgical procedure and with animal care.Foundation to Support Research from the State of Sao Paulo(FAPESP) and the National Board of Scientific and Techno-logic Development (CNPq).

References

1. Lee YC, Baumann MH, Maskell NA, et al. Pleurodesis practicefor malignant pleural effusions in five English-speaking coun-tries: survey of pulmonologists. Chest 2003;124:2229e38.

2. Light RW. Pleural diseases. 4th ed. Philadelphia, PA: LippincottWilliams & Wilkins; 2001.

3. Teixeira LR, Vargas FS, Carmo AO, Cukier A, Silva LM, Light RW.Effectiveness of sodium hydroxide as a pleural sclerosing agentin rabbits: influence of concomitant intrapleural lidocaine.Lung 1996;174:325e32.

4. Xie C, Teixeira LR, Mc Govern JP, Light RW. Systemic cortico-steroid decreases the effectiveness of talc pleurodesis. Am JRespir Crit Care Med 1998;157:1441e4.

5. Teixeira LR, Wu W, Chang DS, Light RW. The effect of corti-costeroids on pleurodesis induced by doxycycline in rabbits.Chest 2002;121:216e9.

6. Lee YCG, Devin CJ, Teixeira LR, et al. Transforming growthfactor beta-2 induced pleurodesis is not inhibited by cortico-steroids. Thorax 2001;56:643e8.

7. Kessenich CR. Cyclo-oxygenase 2 inhibitors: an important newdrug classification. Pain Manag Nurs 2001;2:13e8.

8. Teixeira LR, Vargas FS, Acencio MMP, et al. Influence of anti-inflammatory drugs (methylprednisolone and diclofenacsodium) on experimental pleurodesis induced by silver nitrateor talc. Chest 2005;128:4041e5.

9. Light RW, Wang N-S, Sassoon CSH, Gruer SE, Vargas FS. Talcslurry is an effective pleural sclerosant in rabbits. Chest 1995;107:1702e6.

10. Teixeira LR, Vargas FS, Acencio MM, Bumlai RU,Antonangelo L, Marchi E. Experimental pleurodesis inducedby antibiotics (macrolides or quinolones). Clinics 2006;61:559e64.

11. Marchi E, Vargas FS, Acencio MMP, et al. Pleurodesis: a novelexperimental model. Respirology 2007;12:500e4.

12. Genofre EH, et al. Talc pleurodesis: evidence of systemicinflammatory response to small size talc particles. Respir Med2009;103(1):91e7.

13. Light RW, MacGregor MI, Luchsinger PC, Ball Jr WC. Pleuraleffusions: the diagnostic separation of transudates andexudates. Ann Intern Med 1972;77:507e13.

14. Genofre EH, Vargas FS, Marchi E. Inflammation and clinicalrepercussions of pleurodesis induced by intrapleural talcadministration. Clinics 2007;62:627e34.

15. Marchi E, Vargas FS, Teixeira LR, Acencio MMP, Antonangelo L,Light RW. Intrapleural low-dose silver nitrate elicits morepleural inflammation and less systemic inflammation than low-dose talc. Chest 2005;128:1798e804.

16. Grove CS, Lee YCG. Vascular endothelial growth factor: the keymediator in pleural effusion formation. Curr Opin Pulm Med2002;8:294e301.