Key Words: sepsis; premature infants; pentoxifylline; tumor necrosis factor-alpha; interleukin-1; interleukin-6; cytokines; septic shock; mortality rate
One of the most common causes of death in prematurely delivered, very low birth weight infants requiring intensive therapy is sepsis of bacterial origin. Endotoxin, the lipopolysaccharide component of the cell wall of Gram-negative bacteria, induces the release of a cascade of cytokines that causes tissue damage, leading to multiple organ failure . Between 65% and 85% of these immature patients with Gram-negative sepsis complicated by septic shock die despite wide-spectrum antibiotics and intensive supporting care [2,3]. It is known that plasma tumor necrosis factor (TNF), interleukin-1 beta (IL-1), and interleukin-6 (IL-6) levels are elevated in infants with sepsis and that the magnitude of elevation correlates with the mortality rate [4,5]. Several animal studies have shown that the treatment of sepsis with pentoxifylline, a methylxanthine derivative, causes an inhibitory effect on production of TNF, improves the survival rate, and attenuates circulatory shock [6,7]. In adult patients with septic shock, pentoxifylline was well tolerated and caused a decrease in plasma TNF concentration .
However, the elimination of TNF is not necessarily the prudent course in the treatment of septic shock as numerous studies examining monoclonal antibodies against this cytokine have demonstrated . Recently, it was found that phosphatidic acid is an important molecule in intracellular signal transduction pathways, which is used by several of the inflammatory mediators such as endotoxins or cytokines. Rice et al.  suggested that inhibitors of phosphatidic acid formation may have significant clinical potential in the treatment of sepsis and septic shock. One of the inhibitors of phosphatidic acid formation is pentoxifylline and a unique metabolite of this drug in humans, lisofylline.
In a pilot study , we showed that pentoxifylline inhibits production of TNF-alpha and may have therapeutic value in the treatment of premature infants with sepsis. These preliminary observations prompted this multicenter study to assess the variability of plasma TNF, IL-1, and IL-6 concentrations in the course of sepsis and to evaluate the influence of pentoxifylline infusion on the magnitude of synthesis of these cytokines. The prospective, randomized, double-blind study was also designed to determine the effects of intravenous administration of pentoxifylline on the clinical course of sepsis in prematurely delivered infants. In contrast with the previous study, we extended the duration of the clinical observation and administration of pentoxifylline from 3 to 6 days. Some septic patients received pentoxifylline on 3 successive days; after a visible improvement in their clinical status, a transient aggravation was found within the next 1 or 2 days of management of the sepsis.
PATIENTS AND METHODS
This study was approved by the Ethical Committee of the Jagiellonian University Medical College and the Polish Mother-Child Hospital. Informed written consent was obtained from parents of all infants.
Patients. All premature infants (gestational age, <36 wks) with suspected sepsis who were admitted to the neonatal intensive care unit of the neonatal department of the Medical College Jagiellonian University in Cracow and the neonatal intensive care unit of the Polish Mother-Child Hospital in Lodz, between January 1, 1995, and July 30, 1996, were candidates for this study. Only neonates with "late onset sepsis" (diagnosed after the first week of life) were enrolled in the study. Infants with major congenital malformations, intraventricular hemorrhage (Grade III or IV), or symptoms of a congenital infection were excluded.
The entry criteria of sepsis were physical and laboratory signs of infection or rapid deterioration of respiratory and cardiovascular function. Physical symptoms of infection were defined by the presence of at least three of the following: feeding intolerance, abdominal distension, temperature instability, disordered peripheral circulation (defined as paleness or peripheral cyanosis and mottled skin with a delayed capillary refill >3 sec), lethargy or irritability, hepatosplenomegaly. The following laboratory measurements were determined in all infants with suspicion of sepsis: leukocyte count, leukocyte differentiation, and concentration of C-reactive protein. Leukopenia <5,000/mm3, leukocytosis >20,000/mm3, and a C-reactive protein >50 mg/L were considered indicative of an inflammatory phenomena. Respiratory dysfunction was established by the presence of tachypnea (>70 breaths/min) or episodes of apnea and a >15% increase in the fraction of FiO2. Tachycardia (heart rate [HR], >190 beats/min) or bradycardia (HR, <90 beats/min) and disordered peripheral circulation were regarded as symptoms of circulatory dysfunction.
A positive blood culture was required for confirmation of sepsis. Samples for blood culture were obtained before the start of antibiotic therapy. Blood specimens for culture were never obtained from the catheters. Two positive blood culture specimens from separate sites were required for confirmation of septicemia with Staphylococcus epidermidis, and only a significant growth of bacteria was considered necessary for a diagnosis of sepsis. Criteria for septic shock included sepsis accompanied by a mean arterial pressure below the fifth percentile for age after appropriate volume resuscitation or requirement for vasopressors to maintain a mean arterial pressure above the fifth percentile for age.
Study Protocol. Computer-generated random numbers were used to number and code identical bottles of pentoxifylline and placebo, which were released by study number. The code was held by the representative of Pharmaceutical Inc.-KRKA-Slovenia. All patients with the presumptive (before the results of blood culture were obtained) diagnosis of sepsis were, thus, randomly allocated to receive pentoxifylline (pentoxifylline group) or 0.9% saline as placebo (placebo group). Therefore, the medical care givers and the laboratory workers were blind as to which infant received pentoxifylline or the placebo.
The pentoxifylline (Pentilin, KRKA, Slovenia), given simultaneously with first-line antibiotics, was administered intravenously 5 mg/kg/hr for 6 hrs. However, when peripheral circulation was disordered, the administration of antibiotics was postponed until approximately 30 mins after the first infusion of pentoxifylline or placebo was started. This postponement was based on the suggestion by Hinshaw  and on our own clinical observations (unpublished data) that disturbed perfusion might diminish penetration of antibiotics to the site of infection. The identical infusions of pentoxifylline were repeated on 6 successive days of treatment. The infused volumes of drug or saline were comparable. It is understood that apart from pentoxifylline or placebo, dopamine and dobutamine were introduced immediately in both groups when the symptoms of disordered peripheral circulation occurred.
The routine medical management of sepsis in both groups was comparable. Augmentin and amikacin were used as first-line antibiotics. After the results of blood cultures were obtained, antibiotics were adjusted according to the sensitivity of the isolated bacteria. When the clinical situation deteriorated during treatment with these antibiotics, therapy was changed to imipenem. Essential parts of the management of septic shock were as follows: obligatory mechanical ventilation (patient sedated with midazolam and opiate), improvement of blood perfusion by volume resuscitation (crystalloid and colloid solutions) and/or vasopressors adhibition (protocols of dopamine and dobutamine or epinephrine administration were identical for both groups), compensation of acid-base balance disturbances, and treatment with antibiotics. When disseminated intravascular coagulation developed, infants were additionally treated with infusions of fresh frozen plasma or/and concentrate of fibrinogen, administration of ethamsylate and vitamin K (newborns are vitamin K-depleted), packed red blood cells transfusions, and/or antithrombin III supplementation (after evaluation of the plasma antithrombin III level). Immunoglobulins, steroids, and granulocyte transfusions were not given. The criteria either for artificial ventilation or for weaning from the ventilator were comparable. The only difference between groups in the medical management of sepsis and septic shock was the infusion of pentoxifylline (or placebo).
Evaluation of Infant's Clinical Condition in the Course of Sepsis. Systemic arterial pressure (radial or umbilical artery catheterization), heart and respiratory rates, as well as skin temperature were recorded continuously throughout the 6 days. A 5-min period of mean arterial blood pressure below the value of the tenth percentile for age was considered to be necessary for a diagnosis of hypotension. Acid-base balance was determined by arterial blood sampling at 4- to 6-hr intervals. Metabolic acidosis was defined as a pH of <7.23 and a plasma bicarbonate value of <17 mEq/L in the arterial blood sample. Apart from white blood cells, also platelet count and plasma protein levels were determined daily. Daily urinary volume was measured during the first 6 days of sepsis. Renal insufficiency was defined as oliguria or anuria when the urine volume measures were <20 mL/kg/day. Neonates with abdominal distension had a radiographic examination at the outset. Necrotizing enterocolitis was diagnosed according to Bell et al. . There was no radiologic evidence of necrotizing enterocolitis at the time of enrolling infants in the study. Criteria for disseminated intravascular coagulation included generalized hemorrhagic diathesis with bleeding from venipuncture sites, elevated fibrin split products (>10 [micro sign]g/mL), hypofibrinogenemia (<150 mg/dL), thrombocytopenia (<100 x 103/mm 3), and prolongation of screening tests (partial thromboplastin time, >80 sec; prothrombin time, >17 sec; thrombin clotting time, >35 sec). A hepatic failure was diagnosed when increased serum concentrations of alanine and aspartate aminotransferases (ALT, AST; >100 units/L) and significantly elevated prothrombin time (>17 sec) and partial thromboplastin time (>80 sec) were found.
TNF, IL-1, and IL-6 Determination. For determination of cytokines, six plasma samples were collected from newborns: on the first day of treatment, immediately before and after pentoxifylline or placebo infusion, and similarly on the third and sixth days of therapy. Blood specimens were obtained by venipuncture or from an arterial catheter and immediately transported on ice to the laboratory. Plasma was separated from the blood within 30 mins. Aliquots were stored at -70[degree sign]C (-94[degree sign]F) until assayed. The volume of plasma required for each cytokine analysis was 50 [micro sign]L. TNF was determined by the immunoenzymetric test (TNF-alpha EASIA; Medgenix, Fleurus, Belgium). The minimum detectable concentration was estimated to be 3 pg/mL and was defined as the TNF concentration corresponding to the mean optical density of 20 replicates of the zero standard + 2 SD. IL-1 beta and IL-6 were determined by Endogen Interleukin-ELISA test. The minimum detectable concentrations were estimated to be 1 pg/mL (for IL-1 beta and IL-6). Blood samples were also obtained from ten healthy newborn infants, who were weight- and gestational age-matched on the 2nd or 3rd wk of life; these blood samples were used for comparative cytokine determination.
Primary End Point. Evaluation of plasma cytokine levels was undertaken to determine clinical status during the 6-day course of sepsis.
Secondary End Point. The clinical condition of infants at discharge was determined to estimate the possible effects of pentoxifylline for future patients.
Statistics. All the results of plasma cytokine measurements were distributed in a parametric fashion. Data are expressed as mean +/- SD. Additionally, the values of the median +/- 95% confidence interval are presented on the graphs in Figure 1, Figure 2 and Figure 3. Two-factor analysis of variance for repeated measures was performed to compare the differences in plasma cytokines concentration. Post hoc comparisons of differences at specific days were then made with the Tukey Honestly Significant Difference test . The degree of association between plasma concentrations of the two different cytokines, measured in one blood sample, was quantified by using the Pearson's product moment linear correlation coefficient. Statistical significance of differences in categorical values were analyzed by the chi-square test or the Fisher's exact test. The unpaired Student's t-test was used to examine the differences between body weight, gestational age, and Apgar score in the pentoxifylline and placebo groups. Also, the differences between plasma cytokine levels evaluated in patients with sepsis caused by Gram-negative or Gram-positive bacteria were compared by means of the unpaired t-test. All statistical analyses were performed using a statistical package (Statistica for Windows, StatSoft, Tulsa, OK). A p value of <or=to .05 was considered to be significant.
One hundred patients were initially recruited into the study. However, in 10 of 50 infants in the pentoxifylline group and in 12 of 50 patients in the placebo group, sepsis was not confirmed by blood cultures and these 22 infants were excluded from further analysis. Therefore, the pentoxifylline group and the placebo group consisted of 40 and 38 subjects, respectively.
Alterations in Plasma Cytokine Concentrations. Plasma TNF, IL-1, and IL-6 concentrations determined before the first infusion of pentoxifylline or placebo (initial data) were within the range of 15 to 2114 pg/mL for TNF, 0.0 to 223.9 pg/mL for IL-1, and 0.0 to 1488.0 pg/mL for IL-6. The mean values of plasma TNF, IL-1, and IL-6, evaluated before the first pentoxifylline or placebo infusion, were not significantly different between both groups (Table 1). In the group of ten healthy, birth weight- and gestational age-matched infants, plasma TNF and IL-1 concentrations were not detectable. IL-6 plasma levels in the control group were detectable in five of ten healthy newborns (23 +/- 11 pg/mL).
The alterations in plasma TNF, IL-1, and IL-6 levels in both groups during the 6-day course of the study are shown in Figure 1, Figure 2 and Figure 3. There was a steady decline in plasma TNF concentrations observed in the pentoxifylline group during the study. However, the only statistically significant difference was found between the values obtained before the pentoxifylline infusion on the first day and on the 6th day of the study (347.2 +/- 452.3 pg/mL vs. 68.1 +/- 127.5 pg/mL; p = .009). On the contrary, in the placebo group, a slight decrease in plasma TNF levels was found within the first 3 days of therapy, but then the elevation of this cytokine concentration in blood occurred. However, these changes were not statistically significant.
Plasma IL-6 concentrations were continuously decreasing in the course of the study, and the mean differences in plasma IL-6 levels observed between initial data and values obtained after the last infusion were statistically significant in both groups (pentoxifylline group, 413.1 +/- 705.3 pg/mL, vs. 24.4 +/- 58.6 pg/mL; p = .009; placebo group, 517.1 +/- 588.2 pg/mL, vs. 117.8 +/- 417.6 pg/mL; p = .046). As shown in Figure 2, changes in plasma IL-6 concentrations found in the placebo group, although not statistically significant, were more distinct compared with respective data obtained in the pentoxifylline group. On the 6th day of study, mean plasma IL-6 concentrations measured in the pentoxifylline group were statistically significantly lower in comparison with respective data obtained in the placebo group (pentoxifylline group, 13.6 +/- 16.6 pg/mL, vs. placebo group, 197.5 +/- 280.2 pg/mL; p = .04).
The fluctuations of plasma IL-1 concentrations in both groups were not statistically significant. The direction of changes was similar in the pentoxifylline and the placebo groups.
There were several statistically significant correlations between plasma cytokine concentrations on the first day of study, found before pentoxifylline or placebo administration in both groups (pentoxifylline group: TNF/IL-1, r2 = .91, p = .005; TNF/IL-6, r2 = .91, p = .004; IL-1/IL-6, r2 = .82, p = .01; placebo group: TNF/IL-1, r2 = .85, p = .01; TNF/IL-6, r2 = .56, p = .04; IL-1/IL-6, r2 = .78, p = .02). Then, on the third and sixth days of therapy with pentoxifylline, plasma TNF concentrations did not correlate either with plasma IL-1 or with IL-6 levels in all further blood samples. On the contrary, in the placebo group, the following significant correlations between plasma cytokine concentrations were found on the 3rd day of the study: TNF/IL-6, r2 = .75, p = .02; TNF/IL-1, r2 = .68, p = .04; and on the 6th day of the study: TNF/IL-1, r2 = .83, p = .01; TNF/IL-6, r2 = .85, p = .01; IL-1/IL-6, r2 = .76, p = .02.
The values of plasma TNF and IL-6 levels, evaluated in neonates in both groups on the 1st day of study, were significantly higher in cases of sepsis caused by Gram-negative bacteria (for TNF: Gram-negative, 535.19 +/- 690.5 pg/mL, vs. Gram-positive, 93.7 +/- 122.4 pg/mL, p = .000002; for IL-6: Gram-negative, 627.3 +/- 801.2 pg/mL, vs. Gram-positive, 136.5 +/- 290.5 pg/mL, p = .00004). However, data obtained on the 3rd and 6th days of therapy did not show statistically significant differences in infants of both groups. In contrast, plasma IL-1 concentrations measured in the course of sepsis were not statistically dependent on causative organisms isolated from blood culture in both groups. Plasma TNF concentrations determined on the 1st day in infants with sepsis complicated by shock were significantly higher compared with the results obtained in cases of sepsis when symptoms of shock were not observed (812.3 +/- 659.7 pg/mL vs. 129.5 +/- 139.1 pg/mL; p = .003). However, there were no similar differences found with respect to IL-1 or IL-6.
Clinical Condition of Infants at Randomization. The onset of sepsis in all infants occurred between the 8th and 23rd days of life. There were no significant differences at randomization between groups with regard to the following parameters: birth weight, gestational age, Apgar score, and gender. A slightly increased number of infants with septic shock in the placebo group (8 vs. 6) influenced an insignificantly higher incidence of hypotension and number of episodes of metabolic acidosis in this group of patients (Table 1). On the other hand, disordered peripheral circulation was found at randomization in 12 infants in the pentoxifylline group and in 10 newborns in the placebo group. The clinical symptoms of necrotizing enterocolitis were observed in two patients in the pentoxifylline group and in one infant in the placebo group. There were no symptoms of disseminated intravascular coagulation observed at entry in either group. The incidence of anuria or oliguria was established based on the measurement of diuresis within 12 to 24 hrs. There is an explanation as to why this symptom was not analyzed at randomization.
Clinical Condition of Infants in the Course of Sepsis. To evaluate a possible beneficial influence of pentoxifylline on the clinical course of sepsis of different severity, patients in each group were divided into two categories: A, infants, who developed septic shock at randomization, and B, the others. Because of the small number of patients in category A, only the mortality rate was compared between the pentoxifylline and the placebo groups.
Category A. One patient (27 wks of gestational age; birth weight, 830 g) of six infants in the pentoxifylline group and four patients (range of gestational ages, 29-32 wks; range of birth weights, 1120-1670 g) of eight newborns in the placebo group, who had symptoms of shock at time of enrollment into the study died. However, the difference was not significant. In the pentoxifylline group, the infant with septic shock died on the first day of therapy, whereas in the placebo group, death occurred on the 3rd (3 cases) and 4th days of treatment. In all five cases, the main symptoms leading to death were a severe metabolic acidosis, generalized bleeding (venipuncture sites, alimentary tract, and pulmonary hemorrhages), and at least two organ failures (renal and hepatic dysfunction).
Category B. Comparison of the incidence of metabolic acidosis, oliguria or anuria, hepatic failure, necrotizing enterocolitis, hypotension, and disseminated intravascular coagulation between the pentoxifylline and the placebo groups is presented in Table 2.
There were four neonates in the placebo group who developed disseminated intravascular coagulation on the 2nd to 5th days of therapy. In two of them, severe symptoms of organ failure with shock occurred and they died on 4th or 6th days of therapy.
In 15 of the 40 infants in the pentoxifylline group and in 14 of the 38 neonates in the placebo group, infection was caused by Gram-negative bacteria. Sepsis complicated by shock was exclusively caused by Gram-negative bacteria.
Finally, in the pentoxifylline group, only 1 of 40 treated infants died of sepsis, whereas in the placebo group, 6 of 38 patients did not survive; diagnosis was confirmed by the findings of diffused foci of inflammation and hemorrhagic diathesis during the autopsy. The difference in mortality was statistically significant (p = .046). All infants in both groups who survived sepsis (39/40 infants in the pentoxifylline group and 32/38 patients in the placebo group) were discharged. The clinical course of their further stay in the hospital until discharge (between primary and secondary end points) was comparable according to the duration of hospitalization and the number of subsequent bacterial or fungal infections in both groups.
Our previous studies implicated pentoxifylline as an important inhibitor of TNF synthesis in neonates with sepsis . We also suggested the beneficial effect of pentoxifylline on the course of severe systemic infections in premature infants. Recently, Bienvenu et al.  demonstrated that modulation of cytokine release induced by pentoxifylline is not to be restricted solely to TNF. They found strongly diminished production of IL-2, interferon-gamma, TNF, and IL-10 in septic patients' blood at different concentrations of pentoxifylline. However, IL-6 biosynthesis was unaffected by pentoxifylline. This is not in keeping with the results obtained by Mandi et al. . These authors showed that the administration of pentoxifylline to septic patients resulted in the normalization of TNF synthesis and in a moderate decrease in IL-6 production. Our present data clearly demonstrate that pentoxifylline significantly diminished plasma TNF and IL-6 levels but has no influence on plasma IL-1 concentrations. Moreover, the effect of pentoxifylline on the modulation of cytokine release was strong enough to disturb the correlations between evaluated cytokines, which normally existed in infants with sepsis in the placebo group. As shown in Figure 1, the most visible (although not significant statistically) decrease in the plasma TNF concentration occurred after the first infusion of pentoxifylline. However, the plasma TNF level still remained elevated, and similar mean values of plasma cytokine concentrations were found on the 3rd and 6th days of therapy. It is known that the half-life of TNF in plasma is approximately 7 to 9 mins . Thus, we conclude that synthesis of this cytokine was continued during the treatment but was significantly suppressed. The possible explanation for persistent production of TNF may be the administration of antibiotics, which leads within the next 3 to 4 days to the destruction of the cell wall of bacteria and endotoxin release . This explanation can be partially supported by the sudden increase in plasma TNF levels on the 6th day of therapy, which was found in infants treated with the placebo. A highly elevated plasma TNF concentration was associated with an increased incidence of shock and death. This is in keeping with data obtained previously [11,19]. The modulatory effect of pentoxifylline on plasma IL-6 concentrations was less expressed than in the case of TNF. However, as it is shown in Figure 2, the fluctuations in plasma IL-6 levels in the pentoxifylline group were diminished in comparison with the respective values found in infants treated with placebo. According to our results, IL-1 seems to be the least influenced by pentoxifylline administration. One could suspect that synthesis of that cytokine was not significantly dependent on plasma TNF or IL-6 concentrations.
The difference in the overall mortality rate between both groups was statistically significant (p = .046). It confirms our previous results and strongly suggests the beneficial effect of pentoxifylline on the course of life-threatening systemic infections in premature infants. A significantly lower incidence of metabolic acidosis, renal or hepatic failure, and the occurrence of clinical symptoms of necrotizing enterocolitis observed in the pentoxifylline group might be a consequence of improved organ blood flow in response to pentoxifylline. The one possible explanation for the mechanism of vasomotor response to pentoxifylline consists of an increased intracellular cyclic adenosine monophosphate and release of vasodilator prostanoids . According to Wang et al. , pentoxifylline maintains vascular endothelial cell function during sepsis. Also, the attenuation of the severity of the clinical course of sepsis after pentoxifylline administration was likely associated with an influence of pentoxifylline, either on plasma cytokine concentrations or a decrease in endothelial adhesions of leukocytes. Recent studies demonstrate that continuous administration of pentoxifylline reduces circulating soluble adhesion molecules in critically ill patients and inhibits the expression of ICAM-1 on stimulated mononuclear cells [22-24]. Moreover, Steeb et al.  found that pentoxifylline prevented small-intestine vasoconstriction and preserved microvascular blood flow in animals with sepsis.
The higher incidence of disseminated intravascular coagulation in the placebo group, followed by death resulting from shock in two infants, confirmed our previous observations. It is known that TNF and bacterial lipopolysaccharide endotoxins increase blood viscosity and activate the endothelial-associated coagulation . According to Boldt et al. , the continuous intravenous administration of pentoxifylline for 5 days beneficially influenced the thrombomodulin/protein C/protein S system in the septic patients. This could partially explain why, in pentoxifylline-treated premature neonates with sepsis, the frequency of hemorrhagic diathesis was lower in comparison with the placebo group.
Our study showed a significantly reduced incidence of episodes of hypotension in pentoxifylline-treated infants. We assume that pentoxifylline prevents the albumin accumulation in several organs and stabilizes the oncotic pressure of the blood. This hypothesis may be supported by the findings of Sato et al. , who demonstrated that pentoxifylline influenced albumin clearance across the endothelium of the pulmonary artery in animals.
In summary, we conclude that pentoxifylline significantly affects the synthesis of TNF and IL-6 in infants with sepsis. Moreover, its influence on endothelial cells and a broad spectrum of rheologic properties indicates that pentoxifylline is a promising immunotherapeutic agent. Dosage and schedule of treatment, as proposed in the study, seem to be safe and have therapeutic value in the treatment of sepsis in premature infants.
1. Michie HR, Manogue KR, Spriggs DR, et al: Detection of circulating tumor necrosis factor after endotoxin administration. N Engl J Med 1988; 318: 1481-1486
2. Baley JE: Neonatal sepsis: the potential for immunotherapy. Clin Perinatol 1988; 4:755-766
3. Hervas JA, Alomar A, Salva F, et al: Neonatal sepsis and meningitis in Mallorca, Spain. Clin Invest Dis 1993; 16:719-723
4. de Bont ES, Martens A, van Raan J, et al: Tumor necrosis factor-alpha, interleukin-beta 1, and interleukin-6 plasma levels in neonatal sepsis. Pediatr Res 1993; 33:380-383
5. Sullivan JS, Kilpatrick L, Costarino AT, et al: Correlation of plasma cytokine elevations with mortality rate in children with sepsis. J Pediatr 1992; 120:510-515
6. Masouye P, Gunning K, Pittet JF, et al: Xanthine derivative HWA 138 attenuates hemodynamic and oxygen uptake dysfunction secondary to severe endotoxin shock in sheep. Circ Shock 1992; 36:113-119
7. Lundblad R, Ekstrom P, Giercksky KE: Pentoxifylline improves survival and reduces tumor necrosis factor, interleukin-6, and endothelin-1 in fulminant intraabdominal sepsis in rats. Shock 1995; 3:210-215
8. Zeni F, Pain P, Vindimian N, et al: Effects of pentoxifylline on circulating cytokine concentrations and hemodynamics in patients with septic shock: results from a double-blind, randomized, placebo-controlled study. Crit Care Med 1996; 24:207-214
9. Aderka D: The potential biological and clinical significance of the soluble tumor necrosis factor receptors. Cytokine Growth Factor Rev 1996; 7:231-240
10. Rice GC, Brown PA, Nelson RJ, et al: Protection from endotoxic shock in mice by pharmacologic inhibition of phospatidic acid. Proc Natl Acad Sci U S A 1994; 91:3857-3861
11. Lauterbach R, Zembala M: Pentoxifylline reduces plasma tumour necrosis factor-alpha concentration in premature infants with sepsis. Eur J Pediatr 1996; 155:404-409
12. Hinshaw LB: Sepsis/septic shock: participation of the microcirculation and abbreviated review. Crit Care Med 1996; 24:1072-1078
13. Bell M, Ternberg J, Feigin R: Neonatal necrotizing enterocolitis: therapeutic decision based upon clinical staging. Ann Surg 1978; 187:1-7
14. Tukey J: The philosophy of multiple comparisons. Stat Sci 1991; 6:100-116
15. Bienvenu J, Doche C, Gutowski MC, et al: Production of proinflammatory cytokines and cytokines involved in the TH1/TH2 balance is modulated by pentoxifylline. J Cardiovasc Pharmacol 1995; 25(Suppl 2):80-84
16. Mandi Y, Farkas G, Ocsovszky I: Effects of pentoxifylline and pentaglobin on TNF and IL-6 production in septic patients. Acta Microbiol Hung 1995; 42:301-308
17. Beutler B, Cerami A: Cachectin: more than a tumor necrosis factor. N Engl J Med 1987; 316:379-385
18. Mock CN, Jurkovich GJ, Dries D, et al: The clinical significance of endotoxin released by antibiotics: what is the evidence? J Endotox Res 1996; 3:253-260
19. Endo S, Inada K, Inoue Y, et al: Two types of septic shock classified by the plasma levels of cytokines and endotoxin. Circ Shock 1992; 38:264-274
20. Krysztopik RJ, Bentley FR, Wilson MA, et al: Vasomotor response to pentoxifylline mediates improved renal blood flow to bacteremia. J Surg Res 1996; 63:17-22
21. Wang P, Wood TJ, Ba ZF, et al: Pentoxifylline maintains vascular endothelial cell function during hyperdynamic and hypodynamic sepsis. Surgery 1996; 120:367-373
22. Boldt J, Muller M, Heesen M, et al: Influence of different volume therapies and pentoxifylline infusion on circulating soluble adhesion molecules in critically ill patients. Crit Care Med 1996; 24:385-391
23. Boldt J, Muller M, Heesen M, et al: The effects of pentoxifylline on circulating ad hesion molecules in critically ill patients with acute renal failure treated by continuous veno-venous hemofiltration. Intensive Care Med 1996; 22:305-311
24. Mandi Y, Farkas G, Ocsovszky I, et al: Inhibition of tumor necrosis factor production and ICAM-1 expression by pentoxifylline: beneficial effects in sepsis syndrome. Res Exp Med (Berl) 1995; 195:297-307
25. Steeb GD, Wilson MA, Garrison RN: Pentoxifylline preserves small-intestine microvascular blood flow during bacteriemia. Surgery 1992; 112:756-764
26. Davidson LW, Mollitt DL: The effect of endotoxin on red blood cell deformability and whole blood viscosity. Curr Surg 1990; 47:341-345
27. Boldt J, Muller M, Heyn S, et al: Influence of long-term continuous intravenous administration of pentoxifylline on endothelial-related coagulation in critically ill patients. Crit Care Med 1996; 24:940-946
28. Sato K, Stelzner TJ, O'Brien RF, et al: Pentoxifylline lessens the endotoxin-induced increase in albumin clearance across pulmonary artery endothelial monolayers with and without neutrophils. Am J Respir Cell Mol Biol 1991; 4:219-227