Printed in UK all rights reserved - PDF document

Printed in UK - all rights reserved Copyright ERS Journals Ltd 1997European Respiratory JournalISSN 0903 - 1936 with tense cirrhotic ascites , G. Laffi Respiratory mechanics in patients with tense cirrhotic ascites. R. Duranti, G. Laffi, G.Misuri, D. Riccardi, M. Gorini, M. Foschi, I. Iandelli, R. Mazzanti, M. Mancini, G.Scano, P. Gentilini.
ERS Journals Ltd 1997.ABSTRACT: Lung volumes are decreased by tense ascites and increase after largevolume paracentesis (LVP). The overall effect of ascites and LVP on the respira-tory function is poorly understood.We studied eight cirrhotic patients with tense ascites before and after LVP.Inspiratory muscle force (maximal transdiaphragmatic pressure (P), and thelowest pleural pressure (P *Section of Pneumology, Istituto di MedicinaInterna ed Immunoallergologia and **Istitutodi Medicina Interna, Universitˆ di Firenze,Firenze, Italy. +Unitˆ di Terapia IntensivaPolmonare, Azienda Ospedaliera di Careggi,Firenze, Italy. ++Pro-Juventute Don GnocchiFoundation Pozzolatico, Firenze, Italy.Correspondence: R. DurantiIstituto di Medicina Interna ed Immuno-allergologiaUniversitˆ degli Studi di FirenzeItalyKeywords: Cirrhosisdynamic elastanceReceived: October 21 1996Accepted after revision March 19 1997Supported by grants from the Ministerodell'Universitˆ e della Ricerca Scientificae Tecnologica of Italy, and by the ItalianLiver Foundation, Florence, Italy. Although tense ascites has been reported to decreaselung volumes, and large volume paracentesis (LVP) toincrease them [1Ð3], the overall effect of ascites andLVP on the respiratory function is poorly understood.In patients with ascites, ABELMANNet al. [1] observeddecreased lung volumes, rapid shallow breathing and in-creased resting oxygen consumption; paracentesis cau-sed an increase in lung volumes, and a reduction bothof respiratory frequency and resting oxygen consump-tion. According to ABELMANNet al. [1], these alterationswere due to the increased intra-abdominal pressure thatwas transmitted to the chest, causing increase in pleuralpressure, elevation and relative fixation of the diaph-ragm, and increased stiffness of the chest wall: the move-ments of a more rigid thoracic cage required an increasedwork of breathing. If this hypothesis is correct, over- Eight patients (7 males and 1 female; mean age 57±8 )lism; 4) no parenchymal lung involvement or pleuraleffusion on routine chest radiographic image; and 5)absence of diagno

stic criteria for asthma, chronic bron-chitis or emphysema according to the American ThoracicSociety [6]. The anthropometric and clinical data of thepatients are presented in table 1. Two patients were non-smokers,and the sis had been diagnosed by history, clinical examination,been diagnosed by history, clinical examination,plained of dyspnoea.After admission to hospital, patients were given a dietcontaining 40 mEqáday-1of sodium, with a controlledfluid intake (1 Láday-1). In patients with hyponatraemia(serum sodium -1) water ingestion was res-tricted to 500 mLáday-1.The protocol of the study was approved by the Univer-sity's Ethics Committee and written informed consentwas obtained from each subject.ProtocolThe study was conducted on three consecutive days.On the first day, blood gas values, lung volumes, andinspiratory muscle strength were measured, while thepatients were seated in a comfortable high-backed arm-chair; they were then placed in a comfortable supine posi-tionand breathed quietly through the pneumotachograph.recorded during tidal breathing. On the second day, large day, pulmonary function testing was repeated using thesame methods as employed on the first day. On the sec-ond and third day, blood electrolytes and body weightwere recorded.MeasurementsRoutine spirometry was performed using a water-sealed spirometer (Pulmonart Godart; Sensormedics Corp.,Yorba Linda, CA, USA). Functional residual capacity(FRC) was measured by the helium dilution technique.The normal values for lung volumes were those of theEuropean Coal and Steel Community [9]. Arterial bloodsamples were taken while the subjects were breathingroom air, and blood gas values were analysed (ABL-3analyzer; Radiometer, Copenhagen, Denmark).For ventilation measurements, patients breathed througha Fleisch No. 3 pneumotachograph connected to a diffe-rential pressure transducer. Volume was obtained by elec-trical integration of the flow signal. From the spirogramthe following parameters were derived: inspiratory time(tI), expiratory time (tE), total time of the respiratorycycle (ttot) and tidal volume (VT). Respiratory frequen-cy (fR= 1/ttot×60) and minute ventilation (V'E= VT×fR) were also calculated. End-tidal carbon dioxidetension (PET,CO2) and arterial oxygen saturation (Sa,O2)were monitored continuously by an infra-red CO2meter(Datex Normocap, Helsinki, Finland) and an ear oxime-ter (Radiometer, Copenhagen, Denmark), respectively.Mou

th pressure during tidal breathing was measuredusing a pressure transducer (Statham P23ID; StathamLab. Inc., Hato Rey, Puerto Rico).Changes in thoracoabdominal dimensions were deter-mined in 6 of the 8 patients by linearized magnetome-ters. Pairs of magnetometer coils were attached on themidline to measure the anteroposterior diameters of thelower rib cage at the level of the nipples (fifth costalcartilage) and of the abdomen approximately 2 cm abovethe umbilicus. The magnetometer output voltages weredisplayed on an x-y Tektronics 5115 storage oscillo-scope (Tektronics Corp., Beaverton, OR, USA), and RESPIRATORYMECHANICSANDCIRRHOTICASCITES Table 1. Ð Anthropometric and clinical data of patients with cirrhotic ascitesPts Sex Age Weight Height Smoking Diagnosis Child-Pugh Amount of Weightclass fluid removed decreaseNo. yrs kg cm pack-yrs L kg1 M 71 76 168 40 HBV B 11.2 11.02 M 43 87 174 17 HBV, HCV C 5.1 5.53 M 59 92 166 10 CRYPT C 13.0 15.14 M 66 74 164 24 HCV C 6.8 6.75 M 55 85 170 0 HBV C 6.5 7.06 F 56 62 150 0 HCV C 4.1 6.07 M 53 106 183 36 HCV C 3.5 5.58 M 55 72 160 45 HCV C 7.2 9.0 Cambridge, MA, USA). Signal gains were adjusted sothat isovolume manoeuvres ("belly-in") produced a slopeof approximately -1 when the anteroposterior diameterof the abdomen was plotted against the anteroposteriordiameter of the lower rib cage [10].Oesophageal pressure (P) was measured with a con-ventional balloon-tipped catheter system [11] connectedto a differential pressure transducer (Validyne, Northridge,CA, USA), as described previously [12]. The balloonwas positioned in the mid-oesophagus and contained 0.4mL of air. Oesophageal pressure was used as an indexof pleural pressure (P). Gastric pressure (P) wassimultaneously measured with a similar balloon-tippedcatheter system connected to a second differential pres-sure transducer. This balloon was positioned in the stom-ach with the tip 65Ð70 cm from the nares, and contained2 mL of air. Transdiaphragmatic pressure (P) wasobtained by subtracting Pto Inspiratory muscle strength was assessed by measur-ing minimal (i.e. the greatest negative) inspiratory pleu-ral pressure (P) and maximal transdiaphragmaticpressure (P) at FRC during sniff manoeuvres [13].The patients were repeatedly encouraged to try as hardas possible, and they had a visual feedback of the pre-ssure generated. The manoeuvres were repeated untilthree measurements with less than 5% variability wererecorded. Th

e lowest Pand the highest Pval-ues obtained were used for analysis and compared withthose of age-matched normal subjects: the lower limitof normality was calculated as the mean value -SDand were also recorded during tidal breath-ing: P) and PP) swings were calculat-ed as the difference between the pressure measured atend-expiration and the peak value measured during in-spiration. These values were compared with those ofage-matched normal subjects: the upper limit of nor-mality was calculated as the mean value +SD1.65. Totallung resistance (RLwere measured during breathing at rest. Total lung re-sistance was obtained using the isovolume method ofFet al. [14]. Dynamic lung elastance was deter-mined by dividing the difference in Pbetween pointsof zero flow by VT. To evaluate end-expiratory alveo-lar pressure, we used the indirect method recently des-cribed by HALUSZKAet al. [15] and DALet al.[16], rather than the direct method of airway occlusion.This was because awake subjects react to airway oc- clusion in an unpredictable fashion, so that no reliablemeasurements of alveolar pressure can be obtained. Wethus looked for the presence of a time lag between thefall in Pat the onset of the inspiratory effort and theonset of inspiratory airflow, and measured the negativedeflection in Pthat preceded the start of inspiratoryflow (fig. 1). This negative deflection in Pwill be ref-erred to as intrinsic positive end-expiratory pressure(PEEP) for consistency with previous investigations [15,16].The change in Presulting from the contraction ofthe abdominal muscles during expiration was also asse-ssed. In agreement with NINANEet al. [17], the increasein during the expiratory phase of the breathing cyclewas taken as a reflection of the mechanical effect ofabdominal muscle contraction. The ratio of pleural pres-sure swing to tidal volume (PVT) was also calcu-lated, in order to assess the pressure necessary to producetidal volume.All signals were recorded continuously on a multi-channel chart recorder (TA4000; Gould, Valley View,OH, USA).Statistical analysisMean values and standard deviations of the mean werecalculated for all variables. Data obtained under controlconditions and after LVP were compared by Wilcoxontest for paired samples. Single and stepwise multipleregression analyses were performed to assess the rela-tionships between variables. The proportion of total vari-ance of the dependent variable accounted for by the

predictor variable(s) is reported as the square of corre-lation coefficient (r2), expressed as a percentage. Singleregression analysis was carried out by least square met-hod. All statistical analyses were carried out using theSPSS for Windows 6.0 package (SPSS Inc., Chicago,IL, USA).ResultsIn all patients, the decrease in body weight from thesecond to the third day was equal to or greater than theweight of fluid removed; moreover, there was a closerelationship between the amount of fluid removed andthe decrease in body weight (r=0.96; p)electrolytes did not significantly change after LVP. DURANTIETAL 1 L1 Lás-1+50 cmH2O-5+20+10 cmH2O01 s a)b) Fig. 1. Ð Recordings in a representative patient: a) before; and; b) after large volume paracentesis (LVP). Arrow indicates t was elevated (14.4±2.4 Láminto large tidal volume (0.82±0.23 L) and high respirato-ry frequency (18.5±4.3 breathsámin= 13.7±1.8; = 0.81±0.25 L;= 17.8±3.8 breathsámin RESPIRATORYMECHANICSANDCIRRHOTICASCITES Table 2. Ð Pulmonary function data, arterial blood gasParameter Before LVP After LVP p-valueVC % pred 90±10 98±17 FRC % pred 86±19 98±14 TLC % pred 94±10 97±11 RV % pred 107±21 103±24 % pred 87±12 100±17 /VC % 74±7 79±5 IC L 2.9±0.63 2.7±0.53 ERV L 0.58±0.30 0.98±0.45 mmHg 11.96±1.03 12.12±0.64 mmHg 4.37±0.43 4.52±0.36 pH 7.44±0.02 7.44±0.02 ás 5.9±1.6 5.0±1.5 11.4±2.6 10.0±2.0 O 4.3±3.5 1.1±1.3 : arterial par- 100a)90807060504030


F Ppl,min cmH2O

M 110












21181512963c)


F Ppl,sw cmH2O 101214161820222426d)








F di,sw cmH2O











Before LVPAfter LVP Before LVPAfter LVP Fig. 2. Ð a) Minimal pleural pressure (: male; : female; LVP (from 3.2 to 2.1 cmHversusPThe magnetometry data show that during quiet breath- El,dynwas high in all subjects and PEEPiwas presentin 7 out of 8 patients. After LVP, RLdid not changesignificantly, whereas both El,dynand PEEPidecreasedsignificantly (table 2 and fig. 1).Interrelated measurementsChanges in Ppl,EE, Ppl,swand PEEPiwere closely re-lated to the amount of fluid removed (fig. 5aÐc). More-over, changes in PEEPiwith LVP were significantlyrelated to changes in Ppl,EE(fig. 6). Changes in Ppl,sw/VTinduced by LVP were closely related to changes in El,dynand PEEPi(r=0.83; ptively). Stepwise multiple regression analysis (table 3)showed that changes in El,dyn(F=7.32; p=0.042) and DURANTIETAL 1.00.40.2000.10.20.30.4Rib cage displacement L
a) Pga cmH2O b)1.21.00.

80.60.40.2Abdominal displacement L 891011121314 Fig. 4. - a) Plots of rib cage displacement 9630 -3 -6 -9-12Ppl cmH2O 0369121518212427121518212427 ga cmH2O Fig. 3. Ð Plot of gastric pressure ( position, pleural and transdiaphragmatic pressure swings RESPIRATORYMECHANICSANDCIRRHOTICASCITES Table 3. Ð Stepwise multiple regression for changesVariable Coefficient model r1.087 80.5 0.0083.21 90.5 0.042 1412108642016b)Ppl,sw cmH2O







121086420 -2a)Ppl,EE cmH2O







10864202c)PEEPi cmH2O 4681012 Fig. 5. Ð Relationship of volume (
24681012 Fig. 6. Ð Relationship of changes in PEEPichanges in P hibited mild hypoxaemia (Nos. 4 and 8), but on aver-age Pa,O2was normal and unchanged after LVP.Inspiratory muscle strengthThe patients studied showed normal or decreased inspi-ratory muscle strength, which did not change with LVP.These findings are consistent with those of HOURANIetal. [24], who measured respiratory muscle strength in116 patients with severe liver disease of varying aetio-logy and found a decreased maximal inspiratory pres-sure (MIP) in 56% of them. The reasons as to whysure (MIP) in 56% of them. The reasons as to whyneum dialysate from 0 to 3 L induced progressive decre-ase of FRC and increase both of MIP and Pdi, measuredin sitting position. Two factors have been thought to beinvolved in determining the increase in diaphragmaticstrength in patients with raised intra-abdominal pressure[25, 26]: 1) the elevation of the diaphragm in the thoraxby the high intra-abdominal pressure could lengthendiaphragmatic fibres, increasing the area of appositionof the diaphragm to the rib cage and improving thelength-tension relationship of the diaphragm; and 2) thepresence of liquid in the abdomen could decrease theabdominal compliance, thus providing a more effectivefulcrum for the diaphragmatic action.Based on the above considerations, one could expectthat diaphragmatic strength is increased in ascitic pati-ents and decreases with LVP. However, in the patientsstudied, LVP induced inconsistent changes. At variancewith PREZANTet al. [25], SIAFAKASet al. [27] found thatMIP, measured in sitting position, was low during CAPDand increased after drainage of the fluid. There are seve-ral reasons that could explain these discrepancies. First-ly, the positive effect of diaphragmatic elevation maybe counterbalanced by an enlargement of the lower ribcage, leading to a greater radius of curvature of th

e The lack of effects of LVP on inspiratory musclestrength suggests that other mechanisms linked to liverdisease are probably involved in determining inspira-tory muscle weakness. In this context, the metabolicalterations characteristic of liver cirrhosis (such as hypo-proteinaemia and electrolyte abnormalities) might playa role, as confirmed by the observation of muscle wast-ing in patients with liver diseases [24], and, in particu-lar, in cirrhotic patients [30]. Finally, one has also toconsider that Pdi,maxis an effort-dependent manoeuvre,so that variations in effort pre- and postparacentesisso that variations in effort pre- and postparacentesisliquid subtraction induced an increase of abdominal DURANTIETAL by a high abdominal contribution to tidal volume gene-ration confirms that abdominal compliance is probablyvery high in these patients. Finally, the lack of changesboth in the contribution of the abdomen to the genera-tion of tidal breathing and in the VPwith LVPargues against change in abdominal compliance withLVP, and supports the hypothesis of HANSONet al. [4].Lung mechanicsAnother important finding of this study is the ele-vated load on the inspiratory muscles due both to an in-creased pulmonary dynamic elastic load (E) and thepresence of a threshold load (PEEPi). In order to over-come the increased lung mechanical load, ascitic pati-ents have to increase the Pat each breath (table 3).The interrelationships between changes in PVTchanges in Eand PEEPiindicate that after para-centesis the load on the inspiratory muscles decreasedand that a lower Pwas necessary to generate the sameVT. The following two factors are likely to contributeto the increased Ein cirrhotic patients. Firstly, clo-sure of alveolar units; this mechanism is thought to beinvolved in causing a decrease in lung compliance bothin obese patients [31, 32], and in normal subjects dur-ing chest wall strapping [33, 34]. The closure of air-ways in cirrhotic patients may be a consequence of thehigh intra-abdominal pressure and the consequent in-creased pleural pressure (see below). Secondly, increasein lung elastic recoil due to interstitial pulmonary oede-ma, which is thought to be present in patients with livercirrhosis [35].An interesting finding was the presence of a PEEPi.Although the reasons for the presence of a PEEPicirrhotic patients with ascites are complex, one has toconsider many possibilities. Firstly, an apparent PEEPimay be d

ue to a discrepancy between pleural and oeso-phageal pressure in the supine position, in which thepresent measurements were carried out. In fact, in thisposition oesophageal pressure may be higher than pleu-ral pressure [36, 37]. However, this cannot explain thepresence of a time lag between the beginning of the ne-gative deflection in pleural pressure and the beginningof inspiratory flow. Secondly, contraction of abdominalmuscles during expiration may result in an apparentPEEPor contribute to it [17]. This mechanism did notappear to contribute substantially to the production ofPEEPin the present patients, as they did not show anincrease in gastric pressure on expiration (fig. 1).Thirdly, high intra-abdominal pressure causes pleuralpressure to be high at end-expiration (fig. 4), and thisin turn can determine an early closure of the airways.As a consequence, the alveoli of the dependent lungregions do not empty during expiration and a positivepressure remains in them at end-expiration. This hypo-thesis is supported by the close interrelationship bet-ween changes in PEEPi, the amount of fluid removedand (figs. 5a, 5c and 6): the larger the amount offluid removed, the greater the reduction in PandPEEP, and the lower the P, the lower the PEEPi.Thus, we think that increased abdominal pressure play-ed a major role in determining the presence of PEEPi.However, other mechanisms, such as mechanical com- Finally, the observation that LVP had small effectson respiratory mechanics when less than 5Ð6 L of asci-tes was removed (fig. 5aÐc), suggests that a substantialreduction of respiratory load can be achieved only iflarge volumes of ascites are subtracted. This observa-tion may have importance in the clinical managementof ascitic patients.In conclusion, in supine cirrhotic patients, tense asci-tes determines an overload for inspiratory muscles dueboth to a high lung elastic load and the presence ofpositive end-expiratory alveolar pressure. Large volumeparacentesis unloads the inspiratory muscles and dec-reases their activation, thus significantly improving therespiratory function. Decreased inspiratory muscle strength,possibly determined by the metabolic alterations char-acteristic of liver cirrhosis, may be present in somepatients in the sitting position and is not modified bylarge volume paracentesis.References1.Abelmann WH, Frank NR, Gaensler EA, Cugell DW.2.Berkowitz KA, Butensky MS, Smith RL. Pulmonary3.Angueira

CE, Kadakia SC. Effects of large-volume4.Hanson CA, Ritter AB, Duran W, Lavietes MH. Ascites:5.Conn HO. Hepatic encephalopathy. : Schiff L, SchiffER, eds. Diseases of the Liver. Philadelphia, J.B. Lippin-6.American Thoracic Society. Chronic bronchitis, asthma7.Pugh RNH, Murray-Lyon IM, Daewson JL, PietroniMC, Williams R. Transection of the oesophagus for bleed-8.Arroyo V, Gines P, Planas R, PlanŽs J, RobŽs J. Paracen-9.European Coal and Steel Community. Standardization10.Konno K, Mead J. Measurements of the separate vol-11.Agostoni E, Rahn H. Abdominal and thoracic pressures12.Duranti R, Misuri G, Gorini M, Goti P, Gigliotti F,Scano G. Mechanical loading and control of breathing13.Miller JM, Moxham J, Green M. The maximal sniff in RESPIRATORYMECHANICSANDCIRRHOTICASCITES 14.Frank NR, Mead J, Ferris BG Jr. The mechanical be-15.Haluszka J, Chartrand DA, Grassino AE, Milic-Emili J.16.Dal Vecchio L, Polese G, Poggi R, Rossi A. "Intrinsic"17.Ninane V, Yernault JC, De Troyer A. Intrinsic PEEP18.Gins P, Tit—o L, Arroyo V, Planas R, . Randomized19.Berry MJ, McMurray RG, Katz VL. Pulmonary andventilatory responses to pregnancy, immersion and exer-20.Heinemann HO, Emirgil C, Mijnssen JP. Hyperventilation21.Stanley NN, Salisbury BG, McHenry LR Jr, CherniackNS. Effect of liver failure on the response of ventila-and in the goat. 22.Augusti AGN, Roca J, Bosch J, Rodriguez-Roisin R.23.Rodriguez-Roisin R, Agusti A, Roca J. The hepatopul-24.Hourani JM, Bellamy PE, Tashkin DP, Batra P, SimmonsMS. Pulmonary dysfunction in advanced liver disease:25.Prezant DJ, Aldrich TK, Karpel JP, Lynn RI. Adaptations 26.Gilroy RJ, Mangura BT, Lavietes MH. Rib cage and27.Siafakas NM, Argyrakopoulos T, Andreopoulos K,Tsoukalas G, Tzanakis N, Bouros N. Respiratory mus-28.Contreras G, Gutierrez M, Beroiza T, . Ventilatory29.Zocchi L, Garzaniti N, Newman S, Macklem PT. Effect30.Conn HO, Atterbury CE. Cirrhosis. : Schiff L, SchiffER, eds. Diseases of the Liver. Philadelphia, J.B. Lippin-31.Sharp JT, Henry JP, Sweany SK, Meadows WR, PetrasRJ. The total work of breathing in normal and obese32.Rochester DF, Enson Y. Current concepts in the patho-33.McIlroy MB, Butler J, Finley T. The effect of chest34.Caro CG, Butler J, Dubois AB. Some effects of restric-35.Ruff F, Hughes JMB, Stanley N, . Regional lung36.Mead J, Gaensler EA. Esophageal and pleural pres-37.Knowles JH, Hong SK, Rahn H. Possible errors using DURANTI