Preditores de mortalidade intra-hospitalar após correção cirúrgica de fratura de fêmur proximal em pacientes idosos: um estudo retrospectivo
Carregando...
Arquivos
Data
2021-02-26
Autores
Souza, André Moreira Fogaça de
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Introdução: A população brasileira está envelhecendo rapidamente e, consequentemente, a incidência de fratura de quadril vem aumentando nos últimos anos. Faltam estudos nacionais que investiguem os determinantes da mortalidade intra-hospitalar em pacientes idosos submetidos a tratamento cirúrgico de fratura do fêmur (FF). Objetivos: Descrever e analisar a interação entre estado clínico e comorbidades na mortalidade pós-operatória intra-hospitalar (desfecho primário) e em outros desfechos intermediários, como se seguem: atraso para a realização da cirurgia maior que 48h (Tpré≥48h), ocorrência de complicações pré-operatórias infecciosas (C-PREi) ou cardiovasculares (C-PRÉc) e seus determinantes. Métodos: Estudo observacional de coorte retrospectivo de base populacional, utilizando prontuários médicos e radiografias de pelve e quadril de todas as cirurgias para correção de fratura de fêmur proximal (FF) realizadas em hospital universitário terciário no Brasil, de agosto de 2012 a dezembro de 2018. Foram incluídos tratamento cirúrgico de FF. Critérios de exclusão: pacientes submetidos à artroplastia eletiva de quadril, com fratura patológica por neoplasia, trauma de alta energia, causa contralateral ou não especificada para segunda intervenção cirúrgica no fêmur proximal e intervenção cirúrgica para complicações pós-operatórias tardias. Pacientes que tiveram internação hospitalar superior a 30 dias também foram excluídos da amostra final. Análise estatística: foi composta por três etapas para o desfecho primário e intermediário. Primeiramente, foram realizadas análises univariadas para se verificar a associação entre as variáveis clínicas e os desfechos primários e intermediários. Após esta etapa, a regressão logística multivariada foi realizada para encontrar os preditores independentes para os mesmos desfechos (padronizados por idade e Tpré). O teste de Hosmer-Lemeshow foi realizado para calibração do modelo multivariado. A análise da área sob a curva (AUC) ROC foi realizada para testar a sensibilidade e especificidade do modelo. Para as análises uni e multivariadas, foi realizada a Odds Ratio (OR) com 95% do intervalo de confiança e o valor de p <0,05 foi considerado significativo. Resultados: Foram identificados 1027 pacientes elegíveis para o estudo. Destes, 335 pacientes foram excluídos e 692 pacientes entraram na análise final dos dados. A mortalidade intra-hospitalar pós-operatória foi de 3,6%. As variáveis pré-operatórias significativas para aumento da mortalidade foram a presença de C-PRÉi (OR: 3,32 (IC95% 1,10 - 10,02)), doença pulmonar obstrutiva cronica (DPOC) (OR: 3,95 (IC95% 1,37 - 11,40)) e hipertensão arterial sistêmica (HAS) (OR: 4,1 (IC95% 1,18 - 14,24)). A área sob a curva (AUC) ROC baseada neste modelo logístico multivariado, padronizado por idade e Tpré, foi considerada excelente (AUC: 0,80 (IC95% 0,71 - 0,88)). Não foram encontrados preditores que determinem atraso maior de 48h para a realização da cirurgia. O desenvolvimento de C-PRÉi foi associado à idade (anos) (OR: 1,04 (IC95% 1,01-1,08)), Tpré (dias) (OR: 1,16 (IC95% 1,10-1, 22)) e DPOC (OR: 4,55 (IC95% 2,00 - 10,36)). Além disso, os preditores de C-PRÉc foram Tpré (dias) (OR: 1,21 (IC95% 1,12 - 1,30)), ASA ≥ III (OR: 3,78 (IC95% 1,35 - 10, 61)) e DPOC (OR: 4,60 (IC95% 1,42 - 14,88)). Apesar disto, a avaliação da AUC de ambos os modelos demonstra que eles foram pobres para prever resultados intermediários ((AUC C-PRÉi: 0,63 (IC95% 0,52 – 0,74); AUC C-PRÉC: 0,60 (IC95% 0,48 – 0,73)). Conclusões: Nossos achados evidenciaram as características individuais e os eventos pré-operatórios, que podem impactar na mortalidade hospitalar após correção cirúrgica de FF em idosos. Este resultado pode ser predito pela presença de HAS, DPOC e a ocorrência de C-PRE-i. Além disso, os preditores de CPRÉ-i foram idade Tpré e DPOC. Nenhum determinante clínico-demográfico esteve associado ao atraso maior de 48 horas para a realização da cirurgia.
Background: The Brazilian population is aging fast, and consequently, the incidence of hip fracture has been increasing in the last years. There is a lack of national studies investigating the determinants of in-hospital mortality in elderly patients undergoing surgical treatment of femoral fractures (FF). Objectives: To describe and analyses the interaction between clinical status, comorbidities, and primary and secondary outcomes. Primary outcome: in-hospital post-operatory mortality. Intermediate outcomes: delay of surgery longer than 48h (Tpre ≥ 48h), occurrence of preoperative infectious complications (C-PREi) or preoperative cardiovascular complications (C-PREc), and their determinants. Also, to assess the impact of intermediate outcomes over primary outcome. Methods: Population-based cohort retrospective study, using medical records and pelvis and hip radiographs of all consecutive surgeries for correction of fracture in the proximal femur in a tertiary teaching Hospital in Brazil, from august 2012 to December 2018. Participants: 1027 consecutive patients aged 60 years or older who underwent a surgical treatment of proximal femoral fracture. We excluded 335 patients: those who undergoes elective hip replacement, with pathologic fracture due to malignancy, high energy trauma, contralateral or unspecified cause to second surgical intervention in the proximal femur, and surgical intervention for later postoperative complications. Patients who had hospital length of stay longer than 30 days were also excluded from the final sample. Statistical analyses: The statistical analyses were composed by three steps for primary and intermediate outcome. First, the univariate analyses were carried out to check the association between clinical variables and the primary and intermediate outcomes. Second, multivariate logistical regression was performed to find the independent predictors for the same outcomes (standardized by age and Tpre). The Hosmer-Lemeshow test was performed for calibration of the multivariate model. Finally, ROC curve analysis was performed to test the sensitivity and specificity of the model. For uni and multivariate analyses, Odds Ratio (OR) with 95% of the confidence interval was performed and p-value < 0,05 was considered significant. Results: Overall postoperative in-hospital mortality was 3.6%. Significant preoperative variables for increased mortality included the presence of C-PREi (OR: 3,32 (IC95% 1,10 – 10,02)), chronic obstructive pulmonary disease (COPD) (OR: 3,95 (IC95% 1,37 – 11,40)), and systemic arterial hypertension (SAH) (OR: 4,1 (IC95% 1,18 – 14,24)). The area under curve (AUC) of the ROC curve based on this multivariate logistical model, standardized by age and Tpré, was considered excellent (AUC: 0,80 (IC95% 0,71 – 0,88)). There are no predictors to delaying surgery beyond the 48h. Development of C-PRÉi was associated to age (years) (OR: 1,04 (IC95% 1,01-1,08)), Tpré(days) (OR: 1,16(IC95% 1,10 – 1,22)) and COPD (OR: 4,55 (IC95% 2,00 – 10,36)). Moreover, the predictors of C-PRÉc were Tpré(days) (OR: 1,21(IC95% 1,12 – 1,30)), ASA ≥ III (OR: 3,78 (IC95% 1,35 – 10,61)) and DPOC (OR: 4,60 (IC95% 1,42 – 14,88)). Despite this, the evaluation of the AUC of both models demonstrate that they were poor to predict intermediate outcomes ((AUCC-PRÉi: 0,63 (IC95% 0,52 – 0,74); AUCC-PRÉC: 0,60 (IC95% 0,48 – 0,73)). Conclusions: Our finding showed the individual characteristics and preoperative events that can impact on in-hospital mortality after surgical correction of FF in older people. This outcome can be accurately predicted by the presence of SAH, COPD and the occurrence of C-PREi. In addition, the predictors of C-PREi were Tpre age and COPD. No clinical-demographic determinants were found associated with a delay longer than 48 hours for surgery
Background: The Brazilian population is aging fast, and consequently, the incidence of hip fracture has been increasing in the last years. There is a lack of national studies investigating the determinants of in-hospital mortality in elderly patients undergoing surgical treatment of femoral fractures (FF). Objectives: To describe and analyses the interaction between clinical status, comorbidities, and primary and secondary outcomes. Primary outcome: in-hospital post-operatory mortality. Intermediate outcomes: delay of surgery longer than 48h (Tpre ≥ 48h), occurrence of preoperative infectious complications (C-PREi) or preoperative cardiovascular complications (C-PREc), and their determinants. Also, to assess the impact of intermediate outcomes over primary outcome. Methods: Population-based cohort retrospective study, using medical records and pelvis and hip radiographs of all consecutive surgeries for correction of fracture in the proximal femur in a tertiary teaching Hospital in Brazil, from august 2012 to December 2018. Participants: 1027 consecutive patients aged 60 years or older who underwent a surgical treatment of proximal femoral fracture. We excluded 335 patients: those who undergoes elective hip replacement, with pathologic fracture due to malignancy, high energy trauma, contralateral or unspecified cause to second surgical intervention in the proximal femur, and surgical intervention for later postoperative complications. Patients who had hospital length of stay longer than 30 days were also excluded from the final sample. Statistical analyses: The statistical analyses were composed by three steps for primary and intermediate outcome. First, the univariate analyses were carried out to check the association between clinical variables and the primary and intermediate outcomes. Second, multivariate logistical regression was performed to find the independent predictors for the same outcomes (standardized by age and Tpre). The Hosmer-Lemeshow test was performed for calibration of the multivariate model. Finally, ROC curve analysis was performed to test the sensitivity and specificity of the model. For uni and multivariate analyses, Odds Ratio (OR) with 95% of the confidence interval was performed and p-value < 0,05 was considered significant. Results: Overall postoperative in-hospital mortality was 3.6%. Significant preoperative variables for increased mortality included the presence of C-PREi (OR: 3,32 (IC95% 1,10 – 10,02)), chronic obstructive pulmonary disease (COPD) (OR: 3,95 (IC95% 1,37 – 11,40)), and systemic arterial hypertension (SAH) (OR: 4,1 (IC95% 1,18 – 14,24)). The area under curve (AUC) of the ROC curve based on this multivariate logistical model, standardized by age and Tpré, was considered excellent (AUC: 0,80 (IC95% 0,71 – 0,88)). There are no predictors to delaying surgery beyond the 48h. Development of C-PRÉi was associated to age (years) (OR: 1,04 (IC95% 1,01-1,08)), Tpré(days) (OR: 1,16(IC95% 1,10 – 1,22)) and COPD (OR: 4,55 (IC95% 2,00 – 10,36)). Moreover, the predictors of C-PRÉc were Tpré(days) (OR: 1,21(IC95% 1,12 – 1,30)), ASA ≥ III (OR: 3,78 (IC95% 1,35 – 10,61)) and DPOC (OR: 4,60 (IC95% 1,42 – 14,88)). Despite this, the evaluation of the AUC of both models demonstrate that they were poor to predict intermediate outcomes ((AUCC-PRÉi: 0,63 (IC95% 0,52 – 0,74); AUCC-PRÉC: 0,60 (IC95% 0,48 – 0,73)). Conclusions: Our finding showed the individual characteristics and preoperative events that can impact on in-hospital mortality after surgical correction of FF in older people. This outcome can be accurately predicted by the presence of SAH, COPD and the occurrence of C-PREi. In addition, the predictors of C-PREi were Tpre age and COPD. No clinical-demographic determinants were found associated with a delay longer than 48 hours for surgery
Descrição
Palavras-chave
Fratura de quadril, Mortalidade hospitalar, Cuidado perioperatorio, Complicações, Eventos adversos, Hip fractures, Hospital mortality, Preoperative care, Adverse outcomes