Received: 28 January 2002Accepted: 30 July 2002Published online: 6 September 2002© Springer-Verlag 2002

Abstract Objective: To investigatethe value of measuring total plasmaprotein and hemoglobin concentra-tions for the diagnosis of pulmonaryedema secondary to scorpion enveno-mation. Design and setting: Retro-spective study over a 4-year period inthe medical intensive care unit of auniversity hospital. Patients: 67 pa-tients older than 3 years admitted inthe intensive care unit for scorpionenvenomation and stratified into twogroups according to the presence ofpulmonary edema assessed by a med-ical committee that took into accountclinical, radiological, and blood gasdata at admission and after treatment.Total plasma protein and hemoglobinconcentrations were analyzed sepa-rately. Results: At admission all pa-tients with and without pulmonaryedema exhibited polypnea and tachy-cardia. The mean plasma protein andhemoglobin concentrations werehigher in patients with pulmonary

edema (74±6 and 14.2±2.0 g/dl, re-spectively) than in those without pulmonary edema (64±6 and12.3±1.4 g/dl). After 24 h plasmaprotein and hemoglobin concentra-tions decreased in the pulmonaryedema group (–11 and –1.9 g/dl) de-spite a negative fluid balance(–500 ml). A plasma protein concen-tration of 70 g/l or more predicted thepresence of pulmonary edema with asensitivity of 80% a specificity of96%, a positive predictive value of97%, and negative predictive valueof 77%. Conclusions: In scorpion-en-venomed patients with cardiorespira-tory manifestations high plasma pro-tein and hemoglobin concentrationssuggest the presence of pulmonaryedema.

Keywords Scorpion envenomation ·Plasma protein concentration · Hemoglobin level · Pulmonary edema

Intensive Care Med (2002) 28:1600–1605DOI 10.1007/s00134-002-1480-6 O R I G I N A L

Mabrouk BahloulMounir BouazizHassen DammakChokri Ben HamidaHichem KsibiNoureddine RekikHedi ChellyJean Louis TeboulHatem Kallel

Value of the plasma protein and hemoglobinconcentration in the diagnosis of pulmonaryedema in scorpion sting patients

Introduction

Scorpion envenomation is common in tropical and sub-tropical regions. In Tunisia almost 40,000 stung patientsare recorded per year [1]. Approximately 1000 of themhave systemic manifestations requiring hospitalization,and about 100 patients eventually die [1]. Cardiorespira-tory manifestations, mainly cardiogenic shock and pul-monary edema, are the leading causes of death afterscorpion envenomation [1]. Acute pulmonary edema hasbeen attributed to acute left ventricular failure resultingfrom massive release of catecholamines or myocardial

damage induced by the venom [2, 3]. The diagnosis ofpulmonary edema is suggested by the presence of ta-chypnea, cough, inspiratory retraction of intercostalspaces, and the presence of lung crackles on auscultationof one or both lungs. The diagnosis is confirmed by thepresence of arterial hypoxemia and alveolar infiltratesover one or both lungs on chest radiography. Neverthe-less, breathing disturbances have been also observed inpatients without pulmonary edema or myocardial dam-age after scorpion envenomation [4]. In addition, chestradiography may fail to diagnose pulmonary edema inthe intensive care unit [5]. Therefore, in some cases a

M. Bahloul (✉) · M. BouazizH. Dammak · C. Ben Hamida · H. KsibiN. Rekik · H. Chelly · H. KallelService de Réanimation médicale, CHU Habib Bourguiba, Route el Ain Km 1, 3029 Sfax Tunisiae-mail: bahloulmab@yahoo.frTel.: +216-9-8698267Fax: +216-7-4243427

J.L. TeboulService de réanimation médicale, CHU Bicètre, 78 rue du Général Leclerc,94275 le Kremlin-Bicetre cedex, Paris France

complementary tool may be necessary for diagnosticconfirmation. Previous studies indicate that acute cardio-genic pulmonary edema is associated with an increase intotal plasma protein and hemoglobin concentrations be-cause of the transfer of hypo-oncotic fluid from the intra-vascular compartment into the lung interstitium [6, 7, 8].

The purpose of our study was to investigate whetherplasma protein and hemoglobin blood concentrations andtheir early changes could be helpful for the diagnosis ofacute cardiogenic pulmonary edema in scorpion-enven-omed patients with cardiorespiratory symptoms.

Materials and methods

We retrospectively included all 85 patients admitted for scorpionenvenomation over a period of 4 years (1993–1996) in the 22-bedintensive care unit (ICU) of our university hospital. The diagnosisof scorpion envenomation was based on a history of scorpionsting. Three patients were excluded because they were aged under3 years. Fifteen were excluded because they had a doubtful diag-nosis of pulmonary edema. In all these 15 patients lung crackleswere present on auscultation. However, in 12 of them chest radi-ography was normal. In the three remaining patients some lung in-filtrates were present on chest radiography but they were not typi-cal of interstitial or alveolar edema and did not disappear at thetime of the discharge of the patients. Thus 67 patients were includ-ed in the analysis (35 men, and 32 women; mean age 19±16 years,range 3–72).

The following information was collected retrospectively: age,sex, vital signs (heart rate, respiratory rate, systolic and diastolicblood pressure), findings of lung auscultation, Simplified AcutePhysiology Score II calculated during the first 24 h of admission,use of mechanical ventilation, use of inotropic drugs, presence ofshock (systolic blood pressure less than 90 mmHg or decreasegreater than 40 mmHg), fluid intake volume and urinary output.The chest radiographs performed at admission and throughout thehospitalization period were reviewed. The following biochemicalvariables measured at admission were also collected: arterial pH,arterial oxygen tension (PaO2), arterial carbon dioxide tension(PaCO2), and blood urea nitrogen. The values of plasma proteinconcentration (Buret method) and hemoglobin level were mea-sured separately. In one patient the measurement of left ventricularejection fraction using echocardiography was performed at admis-sion, while in five other patients the measurement of pulmonaryartery occlusion pressure could be obtained from a pulmonary ar-tery catheter inserted just after the admission in the intensive careunit.

A medical committee of five ICU physicians examined retro-spectively all the available data in order to classify patients ac-cording to the presence of pulmonary edema at admission. The di-agnosis of pulmonary edema was based on the presence of clinicaland radiological features of cardiogenic pulmonary edema and onthe presence of arterial hypoxemia. The medical committee con-sidered particularly the presence of signs of respiratory distress(tachypnea, inspiratory retraction of intercostal spaces) and thepresence of lung crackles on auscultation of one or both lungs. Inaddition, the medical committee looked for signs of interstitialand/or alveolar pulmonary edema on the chest radiographs. Mani-festations of interstitial pulmonary edema [9] included the loss ofthe normal sharp definition of pulmonary vascular markings, hazi-ness, loss of demarcation of hilar shadows, thickening of interlob-ular septa, and peribronchial cuffing. Manifestations of alveolarpulmonary edema included unilateral or bilateral confluent acinarshadows creating irregular patchy increases in parenchymal densi-

ty of the lower two-thirds of the lung [9]. In patients receiving me-chanical ventilation, arterial hypoxemia was defined as aPaO2/FIO2 ratio less than 300. The response to treatment of theclinical, radiological, and gas exchange abnormalities was alsotaken into account for the diagnostic classification. Patients agedunder 3 years and those with uncertain diagnosis of pulmonaryedema because of discordance between clinical signs, chest radio-graphs, and blood gas measurements were excluded from the anal-ysis. For all the cases, the final consensus obtained by the medicalcommittee resulted in the following classification: patients withpulmonary edema, patients without pulmonary edema and patientswith exclusion criteria.

The medical committee was not aware of the value of plasmaprotein concentration and hemoglobin level measured at any timeof the hospitalization period in the patients who were included aswell as in the patients who were excluded. In patients included inthe pulmonary edema group, a second set of measurements ofplasma protein concentration and hemoglobin level obtained after24 h (between 24 and 48 h) following admission was analyzed. In19 patients of this group the net fluid balance at the 24th h couldbe calculated.

Statistics

The characteristics of patients in both groups with and withoutpulmonary edema were compared using the Mann-Whitney non-parametric test for continuous variables and the χ2 test for categor-ical variables. The Wilcoxon’s test was used to compare two con-tinuous variables within one group. A p value less than 0.05 wasconsidered statistically significant.

The value of plasma protein concentration and of hemoglobinlevel to predict the presence of pulmonary edema was analyzedusing Receiver Operating Characteristic (ROC) curves. The areaunder the curve, which was estimated by the method of Hanleyand McNeill [10], provides a measure of overall diagnostic accu-racy of the test. The optimal plasma protein concentration and he-moglobin level for the calculation of positive and negative predic-tive values was obtained from the ROC analysis.

Results

The group finally classified as pulmonary edema groupincluded 39 patients; 12 (31%) had radiological signs ofinterstitial pulmonary edema and 27 (69%) radiologicalsigns of alveolar pulmonary edema. The two groups ofpatients with and without pulmonary edema had similardemographic characteristics, but there were differencesin respiratory and hemodynamic parameters (Table 1).Almost patients of both groups exhibited polypnea andtachycardia at the time of ICU admission. There was nodifference in systolic blood pressure or diastolic bloodpressure at the admission between the two groups. Therewas the same proportion of patients with vomiting orsweating in the two groups (62% vs. 60% and 75% vs.70%, respectively). The plasma sodium concentrationwas normal and not different between the two groups(140±4 vs. 141±5 mmol/l). Eleven patients in the pulmo-nary edema group (one in the control group) developedcardiogenic shock during the ICU stay. At admission themean values of pH, PaO2, and PaCO2 were similar be-tween the two groups, since 59% of patients with pulmo-

1601

1602

nary edema were already ventilated before the ICU ad-mission. Total plasma protein and hemoglobin concen-trations were significantly higher in the pulmonary ede-ma group (74±6 and 14.2±2.0 g/dl, respectively) than inthe control group (64±4 and 12.3±1.4 g/dl; Table 1,Fig. 1).

Of the patients with pulmonary edema 80% had a to-tal plasma protein concentration above 70 g/l vs. 3.5% inthe control group. A value of total plasma protein con-centration above 70 g/l predicted the presence of pulmo-nary edema with a sensitivity of 80%, a specificity of96%, a positive predictive value of 97%, and a negativepredictive value of 77%. Figure 2 shows the receiver op-erating characteristic (ROC) curve for various plasmaprotein concentrations in the diagnosis of pulmonaryedema in the whole study population. The area under theROC curve was 0.92. A value of hemoglobin level above15 g/dl was correlated with the diagnosis of pulmonary

edema with a sensitivity of 35% and a specificity of93%. The area under the ROC curve was 0.79. Twenty-four hours after admission all the patients without pul-monary edema except one improved and did not requirefurther investigations. Table 2 shows the changes undertreatment of clinical and biological parameters in thepulmonary edema group at the 24th h. Although the netfluid balance (measured only in 19 patients (Fig. 3) wasnegative (–500 ml), significant decreases in the totalplasma protein (–11 g/l) and hemoglobin concentration(–1.9 g/dl) were observed (p<0.001 for both). The hemo-dynamic nature of the pulmonary edema was confirmedby a low value (less than 40%) of left ventricular ejec-tion fraction in the patient who underwent echocardiog-raphy and by a pulmonary artery occluded pressure value

Table 1 Demographic andclinical parameters at admis-sion of all study population(DBP diastolic blood pressure,Hb hemoglobin concentration,HR heart rate, MV mechanicalventilation, PaCO2 arterial car-bon dioxide tension, PaO2 arte-rial oxygen tension, PE pulmo-nary edema, RR respiratoryrate, SAPS Simplified AcutePhysiology Score, SBP systolicblood pressure

Overall PE group Control group p(n=67) (n=39) (n=28)

Age (years) 19±16 16±11 22±20 nsSex: M/F 35/32 22/17 13/15 nsSAPS II 18±10 20±12 14±6 0.02RR (breaths/min) 28±8 30±9 25±5 0.01HR (beats/min) 120±24 126±27 112±17 0.01SBP (mmHg) 114±24 115±30 110±10 nsDBP (mmHg) 66±21 68±23 63±18 nsShock 12 11 1 0.01pH 7.33±0.1 7.31±0.09 7.34±0.08 nsPaCO2 (mmHg) 35±8 36±7 33±8 nsPaO2 (mmHg) 113±66 110±75 117±51 nsMV 23 23 0 –Urea (mmol/l) 6±2 6±2 6±3 nsUse of inotropes 34 33 1 0.0001Plasma protein level (g/l) 70±7 74±6 64±4 <0.0001Hb (g/dl) 13.3±2 14.2±2 12.3±1.4 <0.0001Hospital stay (h) – 74±49 31±17 <0.0001Died/survivors – 3/39 0/28

Fig. 1 Individual values of plasma protein concentration at admis-sion in the two groups of patients with (PE+) and without pulmo-nary edema (PE–)

Fig. 2 Receiver-operating-characteristic (ROC) curve for abilityof plasma protein concentration to detect pulmonary edema in thewhole population

signs of pulmonary edema are usually present when theextravascular lung water increases by at least 35% [12,13].

Because of the lack of sensitivity and specificity ofusual criteria of pulmonary edema, some adjunctive so-phisticated tools have been proposed [14, 15, 16, 17].For instance, high levels of natriuretic peptides (atrial,brain) have been found to have good sensitivity andspecificity to detect left ventricular systolic dysfunctionand pulmonary edema [15, 16]. In our study we enrolledonly patients for whom the presence or the absence ofpulmonary edema was retrospectively ascertained by themedical committee on the basis of clinical, radiological,and gas exchange evaluation at admission and after treat-ment. In six patients classified in the pulmonary edemagroup we obtained additional information (echocardiog-raphy or pulmonary artery catheter) suggesting the car-diogenic nature of the pulmonary edema. In the 33 re-maining patients of this group the hemodynamic natureof pulmonary edema, although currently well established[2], was not absolutely ascertained. However, the cardiacorigin was likely on the basis of radiological characteris-tics and their response to treatment.

We found that increases in hemoglobin and plasmaprotein concentration were strongly correlated with thepresence of pulmonary edema, and that a level of plasmaprotein concentration higher than70 g/l predicted thepresence of pulmonary edema with high predictive val-ues. These results are in accordance with findings of pre-vious studies performed in patients with cardiogenic pul-monary edema related to usual causes of left ventriculardysfunction [6, 7, 8, 18].

Cardiogenic pulmonary edema formation results froman increase in pulmonary capillary hydrostatic pressurerelated to an increase in left ventricular filling pressure[6, 7, 8]. Fluid filtration across the pulmonary capillariessometimes reaches 50% of the normal plasma volume[6]. In cardiogenic pulmonary edema the protein concen-tration in this fluid is low and approximately one-half ofthat in the plasma [6, 7, 8]. Accordingly, filtration of alarge amount of hypo-oncotic fluid into the lung wouldresult in both a reduction in plasma volume and an in-crease in hemoglobin and plasma protein concentrations.In an animal study Weiser and Grande [19] measured si-multaneous increases in extravascular lung water and inhematocrit and plasma protein concentrations after in-duction of pulmonary edema by balloon obstruction ofthe left atrium. In patients with cardiogenic pulmonaryedema Figueras and Weil [8] reported lower values ofplasma and total blood volumes measured using isotopicmethods and higher values of colloid osmotic pressure,total protein concentration, and hematocrit than in pa-tients without cardiogenic pulmonary edema. Similarfindings have been reported in other series of patientswith acute cardiogenic pulmonary edema [6, 7, 18]. Inour study the resolution of pulmonary edema under treat-

1603

above 18 mmHg in the five patients who underwent pul-monary artery catheterization.

Discussion

This study of patients admitted after scorpion stingshows that total plasma protein and hemoglobin concen-trations are higher in patients suffering from pulmonaryedema than in those without pulmonary edema. Pulmo-nary edema is a life-threatening complication of scorpionsting from various species. It results from acute left ven-tricular dysfunction and myocardial damage [2, 3]. It canbe diagnosed from the analysis of clinical signs, gas ex-change, and chest radiographic abnormalities. However,breathing disturbances secondary to the venom effect onthe parasympathetic nervous system [4] can be observedin patients without cardiogenic pulmonary edema. Insome cases the diagnosis of cardiogenic pulmonary ede-ma can be difficult even using clinical judgment and ra-diological evaluation, in particular in patients with previ-ous pulmonary disease [11]. In addition, portable chestradiography may fail accurately to make the diagnosis ofpulmonary edema [5], in particular to detect moderateincrease in extravascular lung water, since radiological

Table 2 Evolution under treatment in the pulmonary edema group(DBP diastolic blood pressure, Hb hemoglobin concentration, HRheart rhythm, PaCO2 carbon dioxide tension, PaO2 arterial oxy-gen tension, SBP systolic blood pressure)

Before After ptreatment treatment

HR (beats/min) 126±27 124±19 0.37SBP (mmHg) 116±30 100±14 0.08DBP (mmHg) 68±23 60±9 0.42PaO2 (mmHg) 110±75 128±56 0.09PaCO2 (mmHg) 36±7 34±5 0.23pH 7.31±0.09 7.38±0.05 0.002Plasma protein concentration (g/l) 74±6 63±6 <0.001Hb (g/dl) 14.2±1.8 12.3±1.6 <0.001

Fig. 3 Changes in plasma protein concentration and fluid balanceover the first 24 h of treatment

ment was associated with large decreases in hemoglobinand protein plasma concentrations, despite a negativefluid balance. Similar findings have been reported in pre-vious studies in patients who exhibited successful resolu-tion of their pulmonary edema under diuretic and mor-phine therapy [6, 7, 8].

In the present study almost all patients with pulmona-ry edema received dobutamine and mechanical ventila-tion. At our institution we administer dobutamine assoon as we have enough evidence suggesting the pres-ence of pulmonary edema, since it has been demonstrat-ed that scorpion envenomation can result in pulmonaryedema secondary to acute left ventricular failure [2], andsince dobutamine is efficient in improving cardiac func-tion in this condition [20]. Mechanical ventilation anddobutamine induce a decrease in the pulmonary capillarypressure and in the central venous pressure. The decreasein the pulmonary capillary pressure prevents the pulmo-nary edema formation while the decrease in central ve-nous pressure should increase the lung lymph flow andhence the removal of fluid from the interstitium into thesystemic veins [21]. This should lead to expansion of theplasma volume and hemodillution with decrease in he-moglobin and plasma protein concentrations. This hasbeen confirmed in previous clinical studies in whichboth increase in plasma and/or blood volume and de-crease in hematocrit and plasma protein level were ob-served despite negative fluid balance, after reversal ofclinical signs of pulmonary edema [6, 7, 18]. A study of76 patients treated with diuretic therapy and morphinesulfate for cardiogenic pulmonary edema found that inthe 65 patients who improved hematocrit and plasmaprotein concentrations decreased to levels comparable tothose of a control group, while they remained elevated inthe 11 patients who did not improved [8]. This confirmsthat the resolution of pulmonary edema is associatedwith reexpansion of plasma volume related to fluid redis-tribution from the interstitial to the intravascular com-partment even when the fluid balance is negative.

The retrospective nature our study could represent amethodological limitation, at least because some datawere not available in all patients (e.g., as fluid balance).In this regard it is possible that in a patient with prior hy-povolemia due to severe vomiting or sweating pulmona-ry edema has occurred only after saline infusion in thepresence of cardiac dysfunction. In such a case normalplasma protein concentration and hemoglobin levelcould coexist with the presence of pulmonary edema. Itis possible that this phenomenon explains some of ourfalse-negative cases. On the other hand, it could be pos-tulated that in our patients the hemoconcentration wasrelated to external losses of body fluid (vomiting and/orsweating) rather than to internal transfer of plasma intothe lung interstitium (false positive). Although it is pos-sible by nature, we think that it was unlikely in our studyfor at least two reasons. First, it would be surprising if

hemoconcentration related to external losses fluid oc-curred only in the group for which the diagnosis of pul-monary edema was established. Furthermore, we foundthe same proportion of patients with a history of vomit-ing or a history of sweating in the two groups. Second,the plasma sodium concentration was normal and notdifferent between the two groups. The combination ofhemoconcentration with normal sodium concentration(as observed in the pulmonary edema group) should indi-cate that the losses of water and of sodium are equilibrat-ed and is more consistent a transfer of plasma into an in-terstitial tissue rather than with external hypotonic fluidlosses related to vomiting and sweating.

The retrospective diagnostic classification, even ifquestionable by nature, may have not represented a tooimportant limitation, because the medical committeetook advantage of the knowledge of the response totreatment of the main clinical, radiological, and bloodgases abnormalities for making its consensual diagnosticdecision. This type of methodology was previously usedin studies that assessed the value of a new method in thediagnosis of left heart failure in patients admitted foracute dyspnea [15, 17]. We excluded patients for whomthe diagnosis of pulmonary edema was doubtful in orderclearly to distinguish two diagnostic conditions (pres-ence or absence of pulmonary edema) and then reallytest the value of plasma protein concentration to detectthe presence of pulmonary edema. However, the plasmaprotein concentration, once it was validated as a relevantdiagnostic parameter, should be more useful for the diag-nosis of pulmonary edema in the cases for which theusual criteria are doubtful rather than in those for whichthe diagnostic is easy to make by using the conventionalcriteria.

Conclusion

In scorpion-envenomed patients with cardiorespiratorymanifestations high total plasma protein concentrationsand hemoglobin levels suggest the presence of cardio-genic pulmonary edema. These simple biochemical mea-surements can represent an adjunctive tool when usualcriteria fail to confirm or to infirm the diagnosis of pul-monary edema in such patients.

1604

1605

References

1. Abroug F, Elatrous S, Nouira S, Haguigua H, Touzi N, Bouchoucha S(1999) Serotherapy in scorpion en-venomation: a randomised control trial.Lancet 354:906–909

2. Abroug F, Ayari M, Nouira S, GamraH, Boujdaria R, Elatrous S, Ben FarhatM, Bouchoucha S (1995) Assessmentof left ventricular function in severescorpion envenomation: combined he-modynamic and echo-Doppler study.Intensive Care Med 2:629–635

3. Kumar EB, Soomro RS, al Hamdani A,el Shimy N (1992) Scorpion venomcardiomyopathy. Am Heart J123:725–729

4. Sofer S, Gueron M (1988) Respiratoryfailure in children following enveno-mation by the scorpion Leiurus quinq-uestriatus: hemodynamic and neurolog-ical aspects. Toxicon 26:931–939

5. Sivak ED, Richmond BJ, O’DonavanPB, Borkowski GP (1983) Value of ex-travascular lung water measurement vsportable chest x-ray in the managementof pulmonary edema. Crit Care Med11:498–501

6. Henning RJ, Weil MH (1978) Effect ofafterload reduction on plasma volumeduring acute heart failure. Am J Car-diol 42:823–827

7. Figueras J, Weil MH (1978) Blood vol-ume prior to and following treatmentof acute cardiogenic pulmonary edema.Circulation 57:349–355

8. Figueras J, Weil MH (1977) Increase inplasma oncotic pressure during acutecardiogenic pulmonary edema. Circula-tion 55:195–199

9. Fraser RG, Paré JAP (1970) Diagnosisof diseases of the chest. Saunders, Phil-adelphia, pp 854–860

10. Hanley JA, McNeill BJ (1982) Themeaning and use of the area under a re-ceiver operator-characteristic (ROC)curve. Radiology 143:29–36

11. Render ML, Weinstein AS, BlausteinAS (1995) Left ventricular dysfunctionin deteriorating patients with chronicobstructive pulmonary disease. Chest107:162–168

12. Laggner A, Kleinberger G, Haller J, K Lenz, Sommer G, Druml W (1984)Bedside estimation of extravascularlung water in critically ill patients:comparison of the chest radiograph andthe thermal dye technique. IntensiveCare Med 10:309–313

13. Halperin BD, Feeley TW, Mihm FG,Chiles C, Guthaner DF, Blank NE(1985) Evaluation of the portable chestroentgenogram for quantitating extra-vascular lung water in critically illadults. Chest 88:649–653

14. Zema MJ, Masters AP, Margouleff D(1984) Dyspnea: the heart or the lungs?Differentiation at bedside by use of thesimple Valsalva maneuver. Chest85:59–64

15. Davis M, Espiner E, Richards G, Billings J, Town I, Neill A, Drennan C,Richards M, Turner J, Yandle T (1994)Plasma brain natriuretic peptide in as-sessment of acute dyspnoea. Lancet343:440–444

16. McDonagh TA, Robb SD, MurdochDR, Morton JJ, Ford I, Morrison CE,Tunstall-Pedoe H, McMurray JJV, Dargie HJ (1998) Biochemichal detec-tion of left-ventricular systolic dys-function. Lancet 351:9–13

17. Ailani RK, Ravakhah K, DiGiovine B,Jacobsen G, Tun T, Epstein D,West BC (1999) Dyspnea differentia-tion index. A new method for the rapidseparation of cardiac vs pulmonarydyspnea Chest 116:1100–1104

18. Sedziwy L, Thomas M, Shilliigford J(1968) Some observations on hemato-crit changes in patients with acutemyocardial infarction. Br Heart J30:344–349

19. Weiser PC, Grande F (1974) estimationof fluid shifts and protein permeabilityduring pulmonary edemagenesis. AmJ Physiol 226:1028–1036

20. Elatrous S, Nouira S, Besbes-OuanesL, Boussarsar M, Marghli S, Abroug F(1999) Dobutamine in severe scorpionenvenomation: effects on standard he-modynamics, right ventricular perfor-mance, and tissue oxygenation. Chest116:748–755

21. Allen SJ, Drake RE, Katz J, Gabel JC,Laine GA (1987) Elevation of superiorvena caval pressure increases extravas-cular lung water after endotoxemia.J Appl Physiol 62:1006–1009