The alarmin interleukin-1α causes preterm birth through the NLRP3 inflammasome.

Sterile intra-amniotic inflammation is a clinical condition frequently observed in women with preterm labor and birth, the leading cause of neonatal morbidity and mortality worldwide. Growing evidence suggests that alarmins found in amniotic fluid, such as interleukin (IL)-1α, are central initiators of sterile intra-amniotic inflammation. However, the causal link between elevated intra-amniotic concentrations of IL-1α and preterm birth has yet to be established. Herein, using an animal model of ultrasound-guided intra-amniotic injection of IL-1α, we show that elevated concentrations of IL-1α cause preterm birth and neonatal mortality. Additionally, using immunoblotting techniques and a specific immunoassay, we report that the intra-amniotic administration of IL-1α induces activation of the NOD-like receptor family, pyrin domain containing 3 (NLRP3) inflammasome in the fetal membranes, but not in the decidua, as evidenced by a concomitant increase in the protein levels of NLRP3, active caspase-1, and IL-1ß. Lastly, using Nlrp3-/- mice, we demonstrate that the deficiency of this inflammasome sensor molecule reduces the rates of preterm birth and neonatal mortality caused by the intra-amniotic injection of IL-1α. Collectively, these results demonstrate a causal link between elevated IL-1α concentrations in the amniotic cavity and preterm birth as well as adverse neonatal outcomes, a pathological process that is mediated by the NLRP3 inflammasome. These findings shed light on the mechanisms underlying sterile intra-amniotic inflammation and provide further evidence that this clinical condition can potentially be treated by targeting the NLRP3 inflammasome.

Regional myocardial blood flow, edema formation, and magnetic relaxation times during acute myocardial ischemia in the canine.

This study was designed to measure early changes in myocardial perfusion after acute coronary occlusion, and to examine the relationships among blood flow, myocardial edema, and magnetic relaxation times of ex vivo myocardial tissue. In ten dogs, the left anterior descending coronary artery was occluded for 4 hours prior to sacrifice of the animals. Regional myocardial blood flow was measured using radiolabeled microspheres (15 micron), which were injected into the left atrium 5 minutes prior to sacrifice. Multiple subendocardial tissue samples from the left ventricular free wall were obtained for measurement of magnetic relaxation times, percent water content and tissue radioactivity. Mild, moderate, and severe ischemia were defined as reductions in myocardial blood flow to 30% to 50%, 15% to 30%, and less than or equal to 15% of control, respectively. Myocardial water content was increased with mild ischemia (79.6 +/- 0.7%), moderate ischemia (79.9 +/- 0.4%), and severe ischemia (80.3 +/- 0.6%), all P less than .005 vs. control. T1 relaxation times rose with mild (544 +/- 10 msec, P less than .005 vs. control), moderate (543 +/- 11 msec, P less than .005 vs. control), and severe ischemia (574 +/- 10 msec, P less than .001 vs. control). T2 relaxation times behaved in a similar manner, being prolonged in the mildly, moderately, and severely ischemic groups (38.3 +/- 0.3, 38.1 +/- 0.3 and 38.2 +/- 0.3 msec, respectively; all P less than .001 vs. control).(ABSTRACT TRUNCATED AT 250 WORDS)

Effects of beta-adrenergic blockade on the natural progression of myocardial infarct size and compensatory hypertrophy.

Using contrast-enhanced computed tomography, the effects of beta-adrenergic blockade were assessed on experimentally produced myocardial infarcts in dogs evaluated serially over the course of approximately 1 month. Infarct size, initial perfusion defect (jeopardized segment) and noninfarcted muscle mass were studied in two groups of conditioned mongrel dogs. Group 1 (n = 11) served as the control group and Group 2 (n = 10) was pretreated with propranolol (2 mg/kg). Each animal in the propranolol-treated group was given identical amounts of the agent twice daily for 7 days after coronary occlusion. Both groups developed increases in the noninfarcted muscle mass of the left ventricle (compensatory hypertrophy). The mean increase averaged 19.8% over 30 days when the two groups were included together. Infarct size was smaller in the propranolol-treated group, and averaged 28% less (p less than 0.05) than that of the control group 30 days after initial myocardial infarction. Thus, pharmacologic interventions were shown by computed tomography to alter the size of an acute experimental myocardial infarct, particularly when examined over the time course of infarct healing. Moreover, compensatory hypertrophy occurred in both the control and propranolol-treated groups.

Effect of prolonged renal dysfunction on intravascular and extravascular pulmonary fluid volumes during left atrial hypertension.

To examine the development of pulmonary edema during experimental renal dysfunction, left atrial pressure was altered in 14 mongrel dogs divided into two groups. Group 1 was composed of seven control animals, and Group 2 was composed of seven animals with surgically induced renal failure (1 week of bilateral ureteral ligation). Data were obtained at two levels of matched transmural pulmonary vascular pressure (defined as mean left atrial pressure less serum protein osmotic pressure). In the animals with renal dysfunction, extravascular lung water (EVLW) (thermal-green dye technique) was higher at moderately (-1 to -2 mm Hg) and severely elevated (11 to 12 mm Hg) vascular driving pressures (11.5 +/- 1.2 cc/kg vs 10.6 +/- 0.8 cc/kg and 14.8 +/- 1.3 cc/kg vs 13.0 +/- 1.9 cc/kg, respectively, both P less than 0.05 vs control). Because protein osmotic pressure was lower in the renal failure group (15.0 +/- 1.8 mm Hg vs 18.4 +/- 1.4 mm Hg, P less than 0.05), greater accumulations of extravascular lung water occurred at lower levels of left atrial pressure (14.2 +/- 1.4 mm Hg vs 17.1 +/- 1.2 mm Hg, P less than 0.05; 26.8 +/- 2.6 mm Hg vs 29.5 +/- 2.3 mm Hg, P less than 0.01). In addition, when the ratio of EVLW/PBV (pulmonary blood volume) was examined in both groups at each stage of the experiment, the ratio was greater in the Group 2 animals at each elevated pressure, suggesting increased permeability with renal dysfunction. In conclusion, pulmonary edema formation occurs at lower left atrial pressures in the setting of sustained renal dysfunction, this phenomenon can be partially explained by lower protein osmotic pressure though altered pulmonary microvascular permeability may contribute to edema formation.

In vivo validation of the thermal-green dye technique for measuring extravascular lung water.

Assessment of extravascular lung water (EVLW) is imprecise in vivo, yet of both clinical and investigative relevance in patients with cardiac disease. Recently, a double-indicator method using thermal-green dye has been proposed as a nondestructive technique for in vivo quantification of EVLW. In our 5-yr study, indocyanine green dye was used as the intravascular indicator and heat as the diffusible indicator in 44 control dogs, 74 dogs administered intravenous oleic acid, 63 dogs in whom left atrial pressure was altered with a left atrial balloon, and 31 dogs with low-output cardiac failure (electrical shock and complete heart block). In these animals, in vivo measures of EVLW correlated closely with standard gravimetric techniques (r = 0.87, p less than .001), although the indicator dilution technique tended to underestimate actual lung water at higher volumes. In an additional 26 dogs, fluid (lactated Ringer's solution) was administered directly into a distal pulmonary airway, producing alveolar rather than interstitial edema. In these animals, as the infused volume was increased, the thermal technique underestimated consistently the actual amount of infused fluid. Nonetheless, we conclude that in most clinical and experimental situations where moderate changes in lung water are anticipated, this technique can provide reasonable estimates of extravascular fluid accumulation.

Extravascular lung water: effects of intravenous ionic and non-ionic (lopamidol) contrast media during ischemia.

Intravenous injections of ionic contrast media increase extravascular lung water in patients with elevated left atrial pressure, particularly in the presence of myocardial ischemia. The authors compared bolus injections of sodium methylglucamine diatrizoate and iopamidol on extravascular lung water at several levels of left atrial pressure in dogs. Methylglucamine increased lung water by a maximum of approximately 25-30% above baseline levels at low (less than 3 mm Hg), moderate (approximately equal to 15 mm Hg), and elevated left atrial pressures (greater than or equal to 25 mm Hg). At matched pressures, the peak change in lung water in the dogs given iopamidol was +4%, +7%, and +6%, respectively. In dogs with myocardial ischemia, the differences were even more pronounced (+45%, +60%, and +70%, respectively, for ionic media, and +7%, +12%, and +21% for iopamidol). The authors caution against using ionic contrast media in patients with left ventricular dysfunction, particularly associated with ischemia. In such cases, non-ionic media appear safer.

Tissue distribution and magnetic resonance spin lattice relaxation effects of gadolinium-DTPA.

Gadolinium-DTPA complex (Gd-DTPA) is a potential clinical magnetic resonance (MR) contrast agent that enhances images primarily by decreasing spin-lattice relaxation time (T1) in tissues in which it localizes. This study was designed to determine the immediate tissue distribution of intravenously administered Gd-DTPA in selected organs of interest as a function of administered dose and tissue Gd-DTPA concentration. An intravenous bolus of Gd-DTPA with a tracer quantity of Gd-153 was administered to three groups of rabbits at the following doses: 0.01 mM/kg (n = 6); 0.05 mM/kg (n = 6); 0.10 mM/kg (n = 6). A control group received sham injections. Five minutes after Gd-DTPA was administered, all animals were killed; samples of serum, lung, heart, kidney, liver, and spleen were analyzed in a 0.25 T MR spectrometer to measure T1, and then in a gamma well counter to determine tissue concentration of Gd-DTPA. Tissue distribution (per cent dose/tissue weight in g) at five minutes after injection was proportionally constant over the range of doses given. Tissue concentration varied linearly with injected dose (r greater than 0.98 for all tissues). Relaxation rate (1/T1) varied linearly with injected dose and with tissue Gd-DTPA concentration (r greater than 0.97 for all tissues). The order of tissue relaxation rate response to a given dose was: kidney greater than serum greater than lung greater than heart greater than liver greater than spleen. We conclude that because of its extracellular distribution and linear relaxation rate versus concentration relationship, Gd-DTPA enhancement in MR images may be a good marker of relative organ perfusion.

Magnetic resonance relaxation times of percutaneously obtained normal and abnormal body fluids.

Seventy-three fluid samples obtained via percutaneous aspiration and drainage were analyzed by proton magnetic resonance spectroscopy. The fluids included abscess contents, bile, ascitic fluid, cyst and pseudocyst fluid, urine, hematomas, pleural fluid, lymphoceles, seromas, cerebrospinal fluid, pancreatic ductal fluid, and necrotic tumor. They were grouped by their clinical etiology and analyzed with respect to their inherent magnetic relaxation properties. In addition, some of the samples were tested for the following chemical parameters which were correlated with T1 and T2 values: total protein content (n = 36), osmolality (n = 24), specific gravity (n = 11), and amylase levels (n = 23). A large overlap was found in the T1 (spin-lattice) and T2 (spin-spin) relaxation times of the fluids; however, the mean T1 values of abscesses and hematomas were significantly lower than those of other fluids. Similar variability was seen in T2 values, though hematomas and abscesses again could be distinguished by shorter relaxation times. The spin-lattice (1/T1) and spin-spin relaxation rates (1/T2) showed a moderate correlation with total protein content, osmolality, and specific gravity. It is concluded that there is some predictability to MR analysis of body fluids, though the overlap in magnetic relaxation times limits specificity.

Hemodynamic effects of rapid and slow infusions of manganese chloride and gadolinium-DTPA in dogs.

The acute hemodynamic effects of two paramagnetic contrast materials, manganese chloride and gadolinium-DTPA, were examined in dogs using ultrasonic dimension gauge crystals. Slow infusions (more than 15 minutes) of MnCl2 (0.1 mm/kg) or Gd-DTPA (0.1 mm/kg) via an infusion pump had no significant hemodynamic effects. When given in just over 1 minute, Gd-DTPA produced elevated left ventricular (LV) end diastolic pressure and minor dilation of the ventricle and slowed diastolic filling. MnCl2, given rapidly, reduced systemic vascular resistance, resulting in hypotension. It also reduced LV volume and had less marked diastolic effects (probably secondary to the amount of hypotension created). With both agents, these side effects waned after 5-10 minutes. It is concluded that both Gd-DTPA and MnCl2 can be given safely in 0.1-mm/kg doses when administered as a slow, continuous infusion. The acute hemodynamic effects of Gd-DTPA are consistent although minor. While the acute effects of Gd-DTPA are small, neither agent is sufficiently innocuous to be given as a rapid injection in clinically unstable patients. Slow, intravenous infusion of Gd-DTPA or MnCl2 is likely to be tolerated well by even severely ill individuals.

Nuclear magnetic resonance analysis of acute myocardial infarction in dogs: the effects of transient coronary ischemia of varying duration and reperfusion on spin lattice relaxation times.

The purposes of this study were to evaluate the effects of acute myocardial ischemia with reperfusion on T1 (spin-lattice) nuclear magnetic resonance (NMR) relaxation times in a canine model and correlate these changes with bulk myocardial water content (%H2O). In 15 dogs the left anterior descending coronary artery was occluded for either 40 minutes (n = 5), 1 hour (n = 5), or 2 hours (n = 5). In 15 additional dogs, matched occlusion periods were followed by 3 hours of reperfusion. T1 of tissue from normal and ischemic myocardium was measured in vitro with a 2.5 kg NMR spectrometer. In the reperfusion animals, the 2-hour group showed significant increases in %H2O and T1 when the ischemic segment of myocardium was evaluated (both p less than 0.01). All but one animal in the 1-hour (reperfusion) group showed increases in both %H2O or T1 in the ischemic segment of myocardium when compared to control segments. The mean values from the ischemic myocardium in the 1-hour group were significantly higher than the values from the matched control segment for %H2O (p less than 0.05) and T1 (p less than 0.05). In the group undergoing 40 minutes of ischemia followed by 3 hours of reperfusion, neither %H2O nor T1 changed significantly. In the nonreperfused animals, neither T1 nor %H2O content increased significantly after 40 minutes. Significant increases were seen in the 1-hour (p less than 0.05) and 2-hour groups (without reperfusion) (p less than 0.01). In addition, the 2-hour occlusion followed by reperfusion animals had significantly greater T1 relaxation times and %H2O than control ischemic animals.(ABSTRACT TRUNCATED AT 250 WORDS)

Chest radiographs in congestive heart failure: response to therapy in acute and chronic heart disease.

Thirty-four men with left ventricular mechanical dysfunction were admitted to an intensive care unit with either an acute myocardial infarction (Group 1, n = 18) or worsening of clinical respiratory signs and symptoms in the setting of a chronic congestive cardiomyopathy (Group 2, n = 16). On admission, all individuals had pulmonary venous hypertension classified as at least Grade 3 by standard radiographic criteria. In each subject, mean pulmonary capillary wedge pressure (mm Hg), extravascular lung water (EVLW) (ml/kg), and chest radiographs were serially evaluated. In the patients in whom pharmacologic therapy successfully returned left ventricular filling pressures to near normal levels (less than or equal to 15 mm Hg), the chest radiograph returned to its baseline level (defined by the discharge radiograph) later in the patients with chronic heart failure (5.1 +/- 1.0 days) than in the patients with acute myocardial infarctions (2.1 +/- 1.2 days, p less than 0.01). Radiographic changes in extravascular water (interstitial and alveolar edema) mirrored changes in EVLW, although EVLW was initially greater in Group 2 (16.3 +/- 1.8 ml/kg) than in Group 1 (10.7 +/- 1.3 ml/kg, p less than 0.01). In the patients in whom filling pressures either worsened or changed less than 3 mm Hg, EVLW and chest radiographs did not markedly change. It is concluded that changes in radiographic pulmonary edema mirror changes in indicator-dilution measurements of EVLW. Radiographic phase lag represents a slow decline in EVLW after therapy for heart failure, which is prolonged in patients with chronic failure and greater EVLW.

Furosemide during sustained left atrial hypertension in functionally anephric dogs: intravascular and extravascular pulmonary fluid volumes. V.

The response of intravascular (PBV) and extravascular (EVLW) pulmonary fluid volume was examined using double-indicator techniques (thermal-green dye) in 11 open-chest anesthetized dogs during the production of sustained left atrial (LA) hypertension by a LA balloon over a period of 195 min. In 6 of these animals data were also acquired after the intravenous administration of furosemide (1 mg/kg). The renal effects of the diuretic were blocked by tying off the ureters and the vascular supply of both kidneys. Left atrial pressure (N = 11) was abruptly increased from 2.2 +/- 2.1 mm Hg to 30.2 +/- 4.0 mm Hg (P less than 0.01) and maintained at that level for 120 min. Data were obtained prior to pressure elevation, immediately upon pressure elevation, and then every 60 min for a total of 120 min. At that point EVLW had increased (8.1 +/- 0.8 cc/kg at control to 21.7 +/- 2.0 cc/kg at 120 min, P less than 0.001), as had PBV (6.2 +/- 2.1 cc/kg to 9.1 +/- 3.1 cc/kg P less than 0.01). After furosemide injection (N = 6), LA pressure declined (mean peak reduction of approximately 6 mm Hg at 60-75 min, P less than 0.01), aortic and pulmonary arterial pressure both declined (P less than 0.01). However, EVLW remained unchanged, though PBV decreased significantly (peak decrease at 75 min after furosemide administration of 2.0 +/- 0.4 cc/kg, P less than 0.01). In the untreated dogs, EVLW continued to climb (P less than 0.05 vs treated dogs at 75 min postfurosemide).(ABSTRACT TRUNCATED AT 250 WORDS)

Computed tomographic analysis of the effects of hyperosmolar mannitol and methylprednisolone on myocardial infarct size.

Using contrast-enhanced computed tomography, the effects of hyperosmolar mannitol and methylprednisolone on experimentally produced myocardial infarcts were evaluated serially over the course of approximately 1 month. Infarct size, initial perfusion defect (jeopardized segment) and noninfarcted muscle mass were determined in three groups of conditioned mongrel dogs. Group 1 (n = 11) served as the control group, groups 2 and 3 were pretreated with mannitol (375 mg/kg, n = 10) or methylprednisolone (7.5 mg/kg, n = 11). Each animal in the treatment groups was treated with identical doses of the originally administered agent twice daily for 7 days after coronary occlusion. Each group developed increases in the noninfarcted muscle mass of the left ventricle (compensatory hypertrophy). The mean increase averaged more than 20% over 30 days when all groups were included together. Infarct size was smaller in both of the treatment groups. However, at 4 days after infarction, mannitol-treated dogs had a mean infarct size that was 68 +/- 8% (+/- standard error of the mean) of the size of control infarcts (p less than 0.01) and methylprednisolone-treated dogs had a mean infarct size that was 77 +/- 6% of the size of control infarcts (p less than 0.01) (referenced to the initial perfusion defect). At 30 days, these differences were less substantial (though still significant), 88 +/- 4% and 85 +/- 5%, respectively. Pharmacologic interventions can be shown to alter the size of an acute myocardial infarction, particularly when examined over the time course of infarct healing.(ABSTRACT TRUNCATED AT 250 WORDS)

Analysis of phase-angle histograms from equilibrium radionuclide studies: correlation with semiquantitative grading of wall motion.

Quantitative wall motion assessment from gated radionuclide left ventriculograms using phase analysis was studied in 14 subjects (6 normal volunteers and 8 patients with previous acute myocardial infarction). The standard deviation and skewness of the phase-angle histograms were determined from both global and segmental left ventricular (LV) regions of interest (septal, apical and posterolateral). Studies were performed at rest, after administration of atropine and after combined administration of phenylephrine and atropine. Both the standard deviation and skewness showed significant correlations with semiquantitative wall motion scoring. From the global analyses, the highest correlations were found after atropine administration (r = 0.86, p less than 0.001 for standard deviation and r = 0.72, p less than 0.001 for skewness). Nevertheless, deterioration in global wall motion scores correlated poorly with directional changes in standard deviation (r = 0.06, difference not significant) or skewness (r = 0.33, p less than 0.05). No significant correlation between skewness or change in skewness and wall motion scores were found with the segmental analyses. The maximal correlation between segmental standard deviation and segmental wall motion grading was again noted after atropine administration (r = 0.68, p less than 0.001), but deterioration in grading did not correlate with similar deterioration of the standard deviation (r = -0.05, difference not significant). Based on 90% confidence limits for normal standard deviation and skewness, an abnormal standard deviation (greater than 14.5) identified 13 of 28 wall motion disorders (sensitivity 46%), whereas an abnormal skewness (greater than 1.4) identified 1 of 28 wall motion disorders (sensitivity 4%).(ABSTRACT TRUNCATED AT 250 WORDS)

Magnetic resonance tissue analysis of acute renal vascular occlusion in the rabbit: enhancement with gadolinium-DTPA complex.

In order to assess the sensitivity of nuclear magnetic resonance (NMR) in detecting acute renal vascular insufficiency, in vitro NMR spectroscopy (at 0.25 T) was performed on rabbit renal cortices following 45 min of unilateral renal artery (RAO) or renal vein occlusion (RVO). Data were obtained both with and without paramagnetic enhancement with gadolinium-DTPA (Gd-DTPA). In the absence of contrast material, RVO was distinguished by markedly elevated spin-lattice (T1) and spin-spin (T2) relaxation times when compared to the contralateral control kidney [mean increase of 29% in T1 (p less than 0.001) and 19% in T2 (p less than .001)]. RAO produced no change in T1 (p = N.S.) and a small change in T2 (mean increase of 11%, p less than .01). Five min following injection of 0.05 mM/kg of Gd-DTPA, relaxation times of control kidneys were markedly shortened [mean decrease 75% in T1 (p less than .001) and 12% in T2 (p less than 0.01)]. With Gd-DTPA, kidneys with RVO continued to have elevated T1 and T2 relaxation time, and kidneys with RAO maintained their essentially normal pre-contrast relaxation time values. We conclude that non-contrast NMR tissue analysis clearly differentiated normal from congested (RVO) kidneys, but not from acutely ischemic (RAO) kidneys. Paramagnetic enhancement with Gd-DTPA allows the differentiation of normally perfused from acutely ischemic or congested kidneys.

Usefulness of the chest x-ray for predicting abnormal left ventricular function after acute myocardial infarction.

We evaluated 229 patients discharged after a definite acute myocardial infarction. Pulmonary venous congestion determined from chest x-ray films during the hospitalization and at discharge and the cardiothoracic ratio at discharge were compared to the left ventricular ejection fraction measured at discharge by a gated radionuclide technique. During hospitalization, pulmonary venous congestion was found on at least one x-ray frame in 94 patients (41%). At discharge 134 patients (59%) had abnormal ejection fraction (less than 0.51) and 35 had pulmonary venous congestion (15%). The sensitivity of the x-ray for detecting an abnormal ejection fraction was 20% when pulmonary venous congestion was observed on the discharge x-ray film (specificity 92% and predictive value 77%), 52% if pulmonary venous congestion was present on any x-ray film during the hospitalization (specificity 74% and predictive value 73%), and 47% if the cardiothoracic ratio was abnormal (greater than or equal to 0.50) on the discharge x-ray film (specificity and predictive value 66%). We conclude that an abnormal x-ray film at discharge or during the hospitalization will identify approximately one-half of the abnormal ejection fractions at the time of hospital discharge. Therefore, to reliably assess left ventricular function, either for prognostic or therapeutic purposes in the individual patient, a more direct measure of left ventricular function such as radionuclide angiography must be obtained.

Nuclear magnetic resonance tissue analysis of acute renal artery and renal vein occlusion: the effect of a paramagnetic contrast agent.

In order to assess the potential sensitivity of nuclear magnetic resonance for the assessment and detection of renal vascular insufficiency, nuclear magnetic resonance spectroscopy was performed on rabbit renal cortices following 45 minutes of unilateral renal artery or vein occlusion with and without the injection of a paramagnetic contrast material (manganese chloride). Renal vein occlusions produced a marked increase in T1 and T2 relaxation times when compared to the contralateral control kidney (mean increase of 18 per cent in T1 and 27 per cent in T2). Renal artery occlusions produced no change in T1 and only a slight increase in T2 (mean increase of 5 per cent). After the intravenous injection of 0.03 mM/kg. of manganese chloride, T1 of the control kidneys was reduced by almost 50 per cent. Relaxation parameters from the kidneys undergoing arterial occlusion did not change after manganese injection when compared to the occluded kidneys without contrast. In the kidneys undergoing venous occlusion, a moderate decrease in T1 resulted from manganese chloride. We conclude that in the absence of a paramagnetic contrast material, nuclear magnetic resonance discriminates well between normal and (venous occlusion) congested kidneys, but not between normal and acutely ischemic (arterial occlusion) kidneys. In the presence of a renal paramagnetic contrast material, NMR aids in the differentiation of acute unilateral arterial insufficiency, as relaxation times from the normal kidney decline while the ischemic kidney does not change.

Extravascular lung water in patients with mitral stenosis: relationship to pulmonary capillary wedge pressure and Kerley B lines.

To evaluate the relationship between extravascular lung water, pulmonary capillary wedge pressure, and chest radiographic findings, extravascular lung water (EVLW) was assessed using double indicator-dilution techniques in 34 adult patients with mitral stenosis. Seven patients were studied 6 to 12 months after successful mitral valve replacement. In the 27 preoperative patients, septal lines were found to be indicative of elevated EVLW only in the presence of intravascular congestion. An excellent correlation between EVLW and pulmonary capillary wedge pressure was observed in these patients (r = .81, p less than .001), whereas only a rough correlation between scored radiographic findings and EVLW was observed. Postoperatively, septal lines were invariably associated with normal filling pressures and EVLW. Thus in both preoperative and postoperative mitral stenosis patients, interstitial Kerley B lines are insensitive markers of elevated extravascular lung water in the absence of pulmonary vascular engorgement. This emphasizes the importance of interpreting radiographic findings of extravascular fluid in conjunction with evaluation of the vascular bed in patients with chronic postcapillary hypertension.