A thesis submitted to the Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, in fulfillment of the requirements for the degree of Doctor of Philosophy. Johannesburg, South Africa 2017. / Almost half of all cases of heart failure have a preserved ejection fraction. However, therapy targeting the mechanisms of this disorder has not improved outcomes. Left ventricular (LV) diastolic dysfunction is a characteristic feature of heart failure with a preserved ejection fraction. A more sound understanding of the mechanisms responsible for LV diastolic dysfunction produced by risk factors may lead to better approaches to preventing this syndrome. Although obesity is thought to be a major risk factor for LV diastolic dysfunction, this does not occur in all obese individuals. In the present thesis I have demonstrated in 737 randomly recruited participants from a community sample of African ancestry, that the relationship between insulin resistance (homeostasis model) and LV diastolic function, as assessed from trans-mitral velocity (E/A) and tissue Doppler imaging of the lateral and septal walls of the LV (e’ and E/e’), is markedly altered by the presence of a more concentrically remodelled LV (as indexed by LV relative wall thickness [RWT]). Importantly, insulin resistance was only associated with LV diastolic function or dysfunction in those with an RWT above a threshold value. In contrast no interactive effects on LV diastolic function between either blood pressure or age and RWT were noted. These data therefore suggest that obesity will only translate into LV diastolic dysfunction if it is associated with insulin resistance and a concentrically remodeled LV. Although hypertension is thought to play an important role in contributing to LV diastolic dysfunction, the pulsatile hemodynamic change primarily responsible for this effect is uncertain. In 524 randomly selected individuals from a community sample I have demonstrated that independent of confounders including left ventricular mass and RWT, aortic backward wave pressure effects (as determined using wave separation analysis), antedate the impact of aortic stiffness (indexed by aortic pulse wave velocity) or the factors determined by aortic stiffness (the time of backward wave return or forward wave pressures) on LV filling pressures (E/e’). These data therefore suggest that to adequately prevent LV diastolic dysfunction, targeting aortic backward wave pressures may be required. As conventional risk factors account for only a portion of the inter-individual variations in LV diastolic function, it is thought that the genetic factors may play a iv significant role. In 694 randomly recruited participants of African ancestry belonging to nuclear families, I demonstrated that independent of conventional risk factors, heritability accounts for approximately 50% of the variation in LV RWT, an important LV structural determinant of LV diastolic function. Moreover, in 442 randomly recruited individuals of African ancestry belonging to nuclear families, I also demonstrated that heritability accounts for approximately 50% of the variation in the index of LV filling pressures, E/e’, independent of LV mass or RWT remodeling and aortic function. These data provide strong evidence that genetic factors responsible for LV diastolic dysfunction and the structural determinants thereof should be sought. In conclusion, the results provided in the present thesis have advanced our knowledge of possible pathophysiological mechanisms that play a role in the development of LV diastolic dysfunction and hence possibly heart failure with a preserved ejection fraction. / MT2017

Issued also as thesis, Oslo.

Issued also as thesis, Oslo.

INTRODUCTION : L’insuffisance cardiaque congestive (ICC) induit remodelage pulmonaire et dysfonction ventriculaire droite (VD) qui contribuent de façon importante à la morbidité/mortalité. Malgré l’efficacité prouvée, l’antagonisme des récepteurs minéralocorticoïdes est sous-utilisé en ICC et ses mécanismes d’actions demeurent incompris. Nous avons évalué si l’Aldostérone contribue au remodelage pulmonaire et à la dysfonction VD en stimulant la prolifération des myofibroblastes (MYFs) pulmonaires. MÉTHODE ET RÉSULTATS : L’étude a été réalisée chez des rats avec infarctus du myocarde (IM) de taille modérée à grande permettant le développement de l’ICC. Deux semaines après l’IM, les rats ont été traités avec 100mg/kg/jour d’Aldactone ou non, pendant trois semaines et comparé à un groupe témoin (N=21;24;8). Comparativement au groupe témoin, les rats IM ont développé une ICC caractérisée par une réduction de la fraction de raccourcissement du VG (53±1%vs.16±2%, moyenne±ESM, P<0.0001), une hypertension pulmonaire (PSVD:27±1vs.40±3mmHg, P<0.01) et une hypertrophie VD (VD/(VG+Septum):24±1%vs.38±3%, P<0.05). L’Aldactone n’a eu aucun effet sur ces paramètres. Les rats IM ont développé un syndrome pulmonaire caractérisé par un abaissement de la courbe respiratoire pression-volume, un remodelage structurel pulmonaire avec doublement du poids poumon sec (P<0.01) et de la fibrose pulmonaire avec augmentation du taux de collagène dans les poumons (P<0.05). L’Aldactone n’a pas restauré la fonction pulmonaire. Enfin, les MYFs pulmonaires isolés n’ont pas proliféré avec l’exposition de 48h aux deux traitements d’Aldostérone (10-7M, 10-6M). CONCLUSION : L’Aldostérone ne contribue pas au remodelage pulmonaire et à la dysfonction VD associés à l’ICC. D’autres mécanismes d’actions sont responsables des effets bénéfiques de l’Aldactone. / BACKGROUND: Congestive heart failure (CHF) can induce pulmonary remodeling and RV dysfunction, which importantly contribute to morbidity and mortality. Despite proven efficacy, antagonism of mineralocorticoid receptors is underused in CHF and the mechanisms of its benefits still debated. We hypothesized that Aldosterone contributes to pulmonary remodeling and RV dysfunction by stimulating lung myofibroblasts (MYFs) proliferation. METHODS AND RESULTS: We studied rats with moderate to large myocardial infarcts (MI) to allow CHF development. Two weeks after MI, rats were treated with Aldactone 100mg/kg/day (N=21) or untreated (N=24) for three weeks and compared to a sham group (N=8). Five weeks after MI, infarct size was similar in the two MI groups, both by ultrasound and pathologic measures. Compared to sham, the MI-untreated group developed CHF with reduced LV fractional shortening (53±1%vs.16±2%; mean±SEM, P<0.0001), pulmonary hypertension (RVSP:27±1vs.40±3mmHg, P<0.01) and RV hypertrophy (RV/(LV+septum):24±1%vs.38±3%, P<0.05). Aldactone treatment had no effect on these parameters and did not improve LV or RV performance. CHF induced a restrictive respiratory syndrome characterized by a downward shift of the respiratory pressure-volume loop, important lung remodeling with nearly doubling of dry lung weight (P<0.01) and evidence of lung fibrosis demonstrated by histological lung collagen fractional area (P<0.05). The Aldactone therapy could not restore pulmonary function. Finally, isolated lung MYFs did not proliferate after 48hr exposure to aldosterone (10-7M and 10-6M). CONCLUSION: Aldosterone does not contribute to pulmonary remodeling and RV dysfunction associated with CHF. Other mechanisms of action must be responsible for the beneficial effects of Aldactone in CHF.

INTRODUCTION : L’insuffisance cardiaque congestive (ICC) induit remodelage pulmonaire et dysfonction ventriculaire droite (VD) qui contribuent de façon importante à la morbidité/mortalité. Malgré l’efficacité prouvée, l’antagonisme des récepteurs minéralocorticoïdes est sous-utilisé en ICC et ses mécanismes d’actions demeurent incompris. Nous avons évalué si l’Aldostérone contribue au remodelage pulmonaire et à la dysfonction VD en stimulant la prolifération des myofibroblastes (MYFs) pulmonaires. MÉTHODE ET RÉSULTATS : L’étude a été réalisée chez des rats avec infarctus du myocarde (IM) de taille modérée à grande permettant le développement de l’ICC. Deux semaines après l’IM, les rats ont été traités avec 100mg/kg/jour d’Aldactone ou non, pendant trois semaines et comparé à un groupe témoin (N=21;24;8). Comparativement au groupe témoin, les rats IM ont développé une ICC caractérisée par une réduction de la fraction de raccourcissement du VG (53±1%vs.16±2%, moyenne±ESM, P<0.0001), une hypertension pulmonaire (PSVD:27±1vs.40±3mmHg, P<0.01) et une hypertrophie VD (VD/(VG+Septum):24±1%vs.38±3%, P<0.05). L’Aldactone n’a eu aucun effet sur ces paramètres. Les rats IM ont développé un syndrome pulmonaire caractérisé par un abaissement de la courbe respiratoire pression-volume, un remodelage structurel pulmonaire avec doublement du poids poumon sec (P<0.01) et de la fibrose pulmonaire avec augmentation du taux de collagène dans les poumons (P<0.05). L’Aldactone n’a pas restauré la fonction pulmonaire. Enfin, les MYFs pulmonaires isolés n’ont pas proliféré avec l’exposition de 48h aux deux traitements d’Aldostérone (10-7M, 10-6M). CONCLUSION : L’Aldostérone ne contribue pas au remodelage pulmonaire et à la dysfonction VD associés à l’ICC. D’autres mécanismes d’actions sont responsables des effets bénéfiques de l’Aldactone. / BACKGROUND: Congestive heart failure (CHF) can induce pulmonary remodeling and RV dysfunction, which importantly contribute to morbidity and mortality. Despite proven efficacy, antagonism of mineralocorticoid receptors is underused in CHF and the mechanisms of its benefits still debated. We hypothesized that Aldosterone contributes to pulmonary remodeling and RV dysfunction by stimulating lung myofibroblasts (MYFs) proliferation. METHODS AND RESULTS: We studied rats with moderate to large myocardial infarcts (MI) to allow CHF development. Two weeks after MI, rats were treated with Aldactone 100mg/kg/day (N=21) or untreated (N=24) for three weeks and compared to a sham group (N=8). Five weeks after MI, infarct size was similar in the two MI groups, both by ultrasound and pathologic measures. Compared to sham, the MI-untreated group developed CHF with reduced LV fractional shortening (53±1%vs.16±2%; mean±SEM, P<0.0001), pulmonary hypertension (RVSP:27±1vs.40±3mmHg, P<0.01) and RV hypertrophy (RV/(LV+septum):24±1%vs.38±3%, P<0.05). Aldactone treatment had no effect on these parameters and did not improve LV or RV performance. CHF induced a restrictive respiratory syndrome characterized by a downward shift of the respiratory pressure-volume loop, important lung remodeling with nearly doubling of dry lung weight (P<0.01) and evidence of lung fibrosis demonstrated by histological lung collagen fractional area (P<0.05). The Aldactone therapy could not restore pulmonary function. Finally, isolated lung MYFs did not proliferate after 48hr exposure to aldosterone (10-7M and 10-6M). CONCLUSION: Aldosterone does not contribute to pulmonary remodeling and RV dysfunction associated with CHF. Other mechanisms of action must be responsible for the beneficial effects of Aldactone in CHF.

Introdução: A fluidoterapia é o tratamento de primeira linha para pacientes em choque hemorrágico ou choque séptico para restauração do volume circulante e da perfusão tecidual, mas diversas questões relacionadas a este tópico permanecem em debate, particularmente em relação às metas de ressuscitação representadas por variáveis fisiológicas a serem atingidas. A variação de pressão de pulso (VPP) já foi proposta como índice confiável para predição de fluido-responsividade em pacientes sob ventilação mecânica, mas requer avaliação complementar em variadas condições fisiopatológicas. Objetivo: O propósito do presente estudo foi comparar, em um modelo experimental de choque hemorrágico agudo com endotoxemia, uma estratégia de ressuscitação volêmica aguda guiada por VPP e pressão arterial média (PAM) a outra baseada em metas de ressuscitação convencionalmente empregadas envolvendo pressão venosa central (PVC), PAM e saturação venosa mista de oxigênio (SvO2). O modelo experimental foi desenvolvido para esta finalidade e cada variável empregada como meta foi adicionalmente avaliada quanto à capacidade de predição de fluido-responsividade. Métodos: Cinquenta e um porcos foram anestesiados, mecanicamente ventilados e, após preparo, aleatoriamente divididos em seis grupos: controle (Sham, n=8); infusão intravenosa de endotoxina em doses decrescentes (LPS, n=8); choque hemorrágico obtido por meio da retirada de 50% da volemia estimada em 20 minutos (Hemo, n=8); choque hemorrágico com endotoxemia conforme protocolos dos grupos LPS e Hemo (Hemo+LPS, n=9); choque hemorrágico com endotoxemia e, após 60 minutos, ressuscitação com cristalóides para atingir metas: PVC 12-15 mmHg, PAM &#8805; 65 mmHg e SvO2 &#8805; 65% (Conv, n=9); choque hemorrágico com endotoxemia e, após 60 minutos, ressuscitação com cristalóides para atingir as metas VPP &#8804; 13% e PAM &#8805; 65 mmHg (dPP, n=9). Tratamentos foram realizados por três horas. Além da avaliação hemodinâmica incluindo termodiluição e ecocardiografia transesofágica, foram realizadas gasometria arterial com mensuração de eletrólitos e lactato, gasometria venosa mista e tonometria intestinal. Ventilação regional foi avaliada por tomografia por impedância elétrica. Mensuração de citocinas séricas e exames histopatológicos pulmonares também foram efetuados. Resultados: Todos os animais dos quatro grupos que receberam a endotoxina desenvolveram hipertensão pulmonar e lesão pulmonar aguda ao longo do experimento. O grupo Hemo+LPS apresentou alta mortalidade (56%), com alterações hemodinâmicas mais acentuadas do que as observadas nos grupos Hemo e LPS. Os grupos Conv e dPP apresentaram o mesmo grau de comprometimento hemodinâmico observado inicialmente no grupo Hemo+LPS, mas houve rápida recuperação em reposta ao tratamento e todos sobreviveram. Entre os grupos tratados não houve diferenças significantes em relação ao volume de cristalóides administrado (volume total, P=0,066) ou ao débito urinário, mas a PVC no grupo Conv foi significantemente superior à dos grupos dPP (P=0,031) e Sham (P=0,048) ao final do protocolo. Entre as variáveis utilizadas como metas, áreas sob as curvas de características operacionais para predição de fluido-responsividade foram maiores para PVC (0,77; IC95%, 0,68-0,86) e VPP (0,74; IC95%, 0,65-0,83), sendo ambas estas variáveis selecionadas por regressão logística múltipla como variáveis independentes para predição de não-responsividade ao desafio volêmico (PVC: P=0,001, razão de chances, 1,7; IC95%, 1,25-2,32 e VPP: P=0,01, razão de chances, 0,91; IC95%, 0,84-0,98). O melhor valor de corte para VPP para maximização de sua função preditiva foi 15%, com sensibilidade 0,75 (IC95%, 0,63-0,85) e especificidade 0,64 (IC95% 0,49-0,77%). Resultados falso-positivos para VPP foram observados em condições de pressão arterial pulmonar média &#8805; 27 mmHg e gradiente transpulmonar &#8805; 14 mmHg, acompanhados de índice de resistência vascular pulmonar médio &gt; 3 unidades Wood. Resultados falso-negativos também foram constatados. Conclusões: O presente modelo experimental de choque hemorrágico agudo com endotoxemia produziu intenso comprometimento hemodinâmico, hipertensão pulmonar, lesão pulmonar aguda e, na ausência de tratamento, alta mortalidade. Nestas condições, a ressuscitação aguda com cristalóides guiada por VPP e PAM não produziu resultados inferiores à estratégia guiada por metas de ressuscitação convencionalmente estabelecidas, com base em PVC, PAM e SvO2. A principal diferença em desfecho entre as estratégias de ressuscitação foi indução de uma PVC significantemente maior no segundo grupo, ao final do protocolo. Apesar de seus desempenhos individuais terem sido considerados limitados em relação à predição de fluido-responsividade, PVC e VPP foram preditoras independentes de não-responsividade ao desafio volêmico, de modo que sua aplicação em conjunto deva ser investigada. VPP é proposta como uma variável adicional para auxiliar no monitoramento de pacientes, sendo o conhecimento de suas limitações indispensável. / Introduction: Fluid therapy is first-line treatment for patients in hemorrhagic or septic shock for the restoration of circulating volume and tissue perfusion, but several issues remain under debate, particularly regarding resuscitation goals represented by physiological variables to be achieved. Pulse pressure variation (PPV) has been proposed as a reliable index for the prediction of fluid responsiveness in mechanically ventilated patients, but further evaluation for its use in diverse conditions is required. Objective: To compare acute fluid resuscitation guided by PPV and mean arterial pressure (MAP) to another strategy consisting of conventionally-established goals, based on central venous pressure (CVP), MAP and mixed-venous oxygen saturation (SvO2), during experimental acute hemorrhagic shock with endotoxemia. An experimental model was developed to this end and each variable used as resuscitation goal was evaluated additionally for its ability to predict fluid-responsiveness. Methods: Fifty-one pigs were anesthetized, mechanically ventilated and, after preparation, randomized into six groups: control (Sham, n=8); intravenous infusion of endotoxin in decreasing doses (LPS, n=8); hemorrhagic shock of 50% the estimated blood volume in 20 minutes (Hemo, n=8); hemorrhagic shock with endotoxemia in accordance with protocols in groups LPS and Hemo (Hemo+LPS, n=9); hemorrhagic shock with endotoxemia followed by resuscitation with crystalloids, after 60 minutes, to achieve and maintain CVP 12-15 mmHg, MAP &#8805; 65 mmHg and SvO2 &#8805; 65% (Conv, n=9); hemorrhagic shock with endotoxemia followed by resuscitation with crystalloids, after 60 minutes, to achieve and maintain PPV &#8804; 13% and MAP &#8805; 65 mmHg (dPP, n=9). Treatments lasted for three hours. In addition to hemodynamic assessment including thermodilution and transesophageal echocardiography, arterial blood-gases with measurement of electrolytes and lactate, mixed-venous blood-gases and intestinal tonometry were performed. Regional ventilation was evaluated by electrical impedance tomography. Lung histopathology and measurement of serum cytokines were performed as well. Results: All animals from the four groups submitted to endotoxemia developed pulmonary hypertension and acute lung injury over the experimental period. Group Hemo+LPS presented with a high mortality rate (56%) and hemodynamic impairment which was more intense than that observed in groups Hemo or LPS. Groups Conv and dPP developed the same degree of hemodynamic compromise observed in group Hemo+LPS initially, but there was quick recovery in response to treatment and all pigs survived. Between treated groups there were no significant differences in amounts of crystalloids infused (total volume, P=0.066) or in urinary output, but CVP in group Conv was significantly higher than in groups dPP (P=0.031) and Sham (P=0.048) at the end of the study period. Among variables used as goals, areas under the receiver-operator characteristic curves regarding prediction of fluid-responsiveness were larger for CVP (0.77; 95%CI, 0.68-0.86) and PPV (0.74; 95%CI, 0.65-0.83), and both these variables were selected by multiple logistic regression as independent predictors of non-responsiveness to fluid challenge (CVP: P=0.001, odds ratio, 1.7; 95%CI, 1.25-2.32 and PPV: P=0.010, odds ratio, 0.91; 95%CI, 0.84-0.98). Best cutoff value to maximize the predictive function of PPV was 15%, with sensitivity 0.75 (95%CI, 0.63-0.85) and specificity 0.64 (95%CI 0.49-0.77). False positive results for PPV were observed at mean arterial pressure &#8805; 27 mmHg and transpulmonary gradient &#8805; 14 mmHg, with mean pulmonary vascular resistance index &gt; 3 Wood units. False negative results were also detected. Conclusions: This model of acute hemorrhagic shock with endotoxemia produced severe hemodynamic compromise, pulmonary hypertension, acute lung injury and, in the absence of treatment, a high mortality rate. In this setting, acute resuscitation with crystalloids guided by PPV and MAP was not inferior to the strategy guided by conventionally-established goals, based on CVP, MAP and SvO2. The main difference in outcome between resuscitation strategies was the induction of a significantly higher CVP in the second group, at the end of protocol. Although their individual performances were considered limited for the prediction of fluid-responsiveness, CVP and PPV were independent predictors of non-responsiveness to fluid challenge, so that their combined use should be investigated further. PPV is proposed as an additional variable to aid in patient monitoring, but awareness of its limitations is indispensable.

Introdução: A fluidoterapia é o tratamento de primeira linha para pacientes em choque hemorrágico ou choque séptico para restauração do volume circulante e da perfusão tecidual, mas diversas questões relacionadas a este tópico permanecem em debate, particularmente em relação às metas de ressuscitação representadas por variáveis fisiológicas a serem atingidas. A variação de pressão de pulso (VPP) já foi proposta como índice confiável para predição de fluido-responsividade em pacientes sob ventilação mecânica, mas requer avaliação complementar em variadas condições fisiopatológicas. Objetivo: O propósito do presente estudo foi comparar, em um modelo experimental de choque hemorrágico agudo com endotoxemia, uma estratégia de ressuscitação volêmica aguda guiada por VPP e pressão arterial média (PAM) a outra baseada em metas de ressuscitação convencionalmente empregadas envolvendo pressão venosa central (PVC), PAM e saturação venosa mista de oxigênio (SvO2). O modelo experimental foi desenvolvido para esta finalidade e cada variável empregada como meta foi adicionalmente avaliada quanto à capacidade de predição de fluido-responsividade. Métodos: Cinquenta e um porcos foram anestesiados, mecanicamente ventilados e, após preparo, aleatoriamente divididos em seis grupos: controle (Sham, n=8); infusão intravenosa de endotoxina em doses decrescentes (LPS, n=8); choque hemorrágico obtido por meio da retirada de 50% da volemia estimada em 20 minutos (Hemo, n=8); choque hemorrágico com endotoxemia conforme protocolos dos grupos LPS e Hemo (Hemo+LPS, n=9); choque hemorrágico com endotoxemia e, após 60 minutos, ressuscitação com cristalóides para atingir metas: PVC 12-15 mmHg, PAM &#8805; 65 mmHg e SvO2 &#8805; 65% (Conv, n=9); choque hemorrágico com endotoxemia e, após 60 minutos, ressuscitação com cristalóides para atingir as metas VPP &#8804; 13% e PAM &#8805; 65 mmHg (dPP, n=9). Tratamentos foram realizados por três horas. Além da avaliação hemodinâmica incluindo termodiluição e ecocardiografia transesofágica, foram realizadas gasometria arterial com mensuração de eletrólitos e lactato, gasometria venosa mista e tonometria intestinal. Ventilação regional foi avaliada por tomografia por impedância elétrica. Mensuração de citocinas séricas e exames histopatológicos pulmonares também foram efetuados. Resultados: Todos os animais dos quatro grupos que receberam a endotoxina desenvolveram hipertensão pulmonar e lesão pulmonar aguda ao longo do experimento. O grupo Hemo+LPS apresentou alta mortalidade (56%), com alterações hemodinâmicas mais acentuadas do que as observadas nos grupos Hemo e LPS. Os grupos Conv e dPP apresentaram o mesmo grau de comprometimento hemodinâmico observado inicialmente no grupo Hemo+LPS, mas houve rápida recuperação em reposta ao tratamento e todos sobreviveram. Entre os grupos tratados não houve diferenças significantes em relação ao volume de cristalóides administrado (volume total, P=0,066) ou ao débito urinário, mas a PVC no grupo Conv foi significantemente superior à dos grupos dPP (P=0,031) e Sham (P=0,048) ao final do protocolo. Entre as variáveis utilizadas como metas, áreas sob as curvas de características operacionais para predição de fluido-responsividade foram maiores para PVC (0,77; IC95%, 0,68-0,86) e VPP (0,74; IC95%, 0,65-0,83), sendo ambas estas variáveis selecionadas por regressão logística múltipla como variáveis independentes para predição de não-responsividade ao desafio volêmico (PVC: P=0,001, razão de chances, 1,7; IC95%, 1,25-2,32 e VPP: P=0,01, razão de chances, 0,91; IC95%, 0,84-0,98). O melhor valor de corte para VPP para maximização de sua função preditiva foi 15%, com sensibilidade 0,75 (IC95%, 0,63-0,85) e especificidade 0,64 (IC95% 0,49-0,77%). Resultados falso-positivos para VPP foram observados em condições de pressão arterial pulmonar média &#8805; 27 mmHg e gradiente transpulmonar &#8805; 14 mmHg, acompanhados de índice de resistência vascular pulmonar médio &gt; 3 unidades Wood. Resultados falso-negativos também foram constatados. Conclusões: O presente modelo experimental de choque hemorrágico agudo com endotoxemia produziu intenso comprometimento hemodinâmico, hipertensão pulmonar, lesão pulmonar aguda e, na ausência de tratamento, alta mortalidade. Nestas condições, a ressuscitação aguda com cristalóides guiada por VPP e PAM não produziu resultados inferiores à estratégia guiada por metas de ressuscitação convencionalmente estabelecidas, com base em PVC, PAM e SvO2. A principal diferença em desfecho entre as estratégias de ressuscitação foi indução de uma PVC significantemente maior no segundo grupo, ao final do protocolo. Apesar de seus desempenhos individuais terem sido considerados limitados em relação à predição de fluido-responsividade, PVC e VPP foram preditoras independentes de não-responsividade ao desafio volêmico, de modo que sua aplicação em conjunto deva ser investigada. VPP é proposta como uma variável adicional para auxiliar no monitoramento de pacientes, sendo o conhecimento de suas limitações indispensável. / Introduction: Fluid therapy is first-line treatment for patients in hemorrhagic or septic shock for the restoration of circulating volume and tissue perfusion, but several issues remain under debate, particularly regarding resuscitation goals represented by physiological variables to be achieved. Pulse pressure variation (PPV) has been proposed as a reliable index for the prediction of fluid responsiveness in mechanically ventilated patients, but further evaluation for its use in diverse conditions is required. Objective: To compare acute fluid resuscitation guided by PPV and mean arterial pressure (MAP) to another strategy consisting of conventionally-established goals, based on central venous pressure (CVP), MAP and mixed-venous oxygen saturation (SvO2), during experimental acute hemorrhagic shock with endotoxemia. An experimental model was developed to this end and each variable used as resuscitation goal was evaluated additionally for its ability to predict fluid-responsiveness. Methods: Fifty-one pigs were anesthetized, mechanically ventilated and, after preparation, randomized into six groups: control (Sham, n=8); intravenous infusion of endotoxin in decreasing doses (LPS, n=8); hemorrhagic shock of 50% the estimated blood volume in 20 minutes (Hemo, n=8); hemorrhagic shock with endotoxemia in accordance with protocols in groups LPS and Hemo (Hemo+LPS, n=9); hemorrhagic shock with endotoxemia followed by resuscitation with crystalloids, after 60 minutes, to achieve and maintain CVP 12-15 mmHg, MAP &#8805; 65 mmHg and SvO2 &#8805; 65% (Conv, n=9); hemorrhagic shock with endotoxemia followed by resuscitation with crystalloids, after 60 minutes, to achieve and maintain PPV &#8804; 13% and MAP &#8805; 65 mmHg (dPP, n=9). Treatments lasted for three hours. In addition to hemodynamic assessment including thermodilution and transesophageal echocardiography, arterial blood-gases with measurement of electrolytes and lactate, mixed-venous blood-gases and intestinal tonometry were performed. Regional ventilation was evaluated by electrical impedance tomography. Lung histopathology and measurement of serum cytokines were performed as well. Results: All animals from the four groups submitted to endotoxemia developed pulmonary hypertension and acute lung injury over the experimental period. Group Hemo+LPS presented with a high mortality rate (56%) and hemodynamic impairment which was more intense than that observed in groups Hemo or LPS. Groups Conv and dPP developed the same degree of hemodynamic compromise observed in group Hemo+LPS initially, but there was quick recovery in response to treatment and all pigs survived. Between treated groups there were no significant differences in amounts of crystalloids infused (total volume, P=0.066) or in urinary output, but CVP in group Conv was significantly higher than in groups dPP (P=0.031) and Sham (P=0.048) at the end of the study period. Among variables used as goals, areas under the receiver-operator characteristic curves regarding prediction of fluid-responsiveness were larger for CVP (0.77; 95%CI, 0.68-0.86) and PPV (0.74; 95%CI, 0.65-0.83), and both these variables were selected by multiple logistic regression as independent predictors of non-responsiveness to fluid challenge (CVP: P=0.001, odds ratio, 1.7; 95%CI, 1.25-2.32 and PPV: P=0.010, odds ratio, 0.91; 95%CI, 0.84-0.98). Best cutoff value to maximize the predictive function of PPV was 15%, with sensitivity 0.75 (95%CI, 0.63-0.85) and specificity 0.64 (95%CI 0.49-0.77). False positive results for PPV were observed at mean arterial pressure &#8805; 27 mmHg and transpulmonary gradient &#8805; 14 mmHg, with mean pulmonary vascular resistance index &gt; 3 Wood units. False negative results were also detected. Conclusions: This model of acute hemorrhagic shock with endotoxemia produced severe hemodynamic compromise, pulmonary hypertension, acute lung injury and, in the absence of treatment, a high mortality rate. In this setting, acute resuscitation with crystalloids guided by PPV and MAP was not inferior to the strategy guided by conventionally-established goals, based on CVP, MAP and SvO2. The main difference in outcome between resuscitation strategies was the induction of a significantly higher CVP in the second group, at the end of protocol. Although their individual performances were considered limited for the prediction of fluid-responsiveness, CVP and PPV were independent predictors of non-responsiveness to fluid challenge, so that their combined use should be investigated further. PPV is proposed as an additional variable to aid in patient monitoring, but awareness of its limitations is indispensable.