bs_bs_banner
Journal of Evaluation in Clinical Practice ISSN 1365-2753
Hospital-acquired pressure ulcers and risk of hospital mortality in intensive care patients on mechanical ventilation Francisco Manzano MD PhD,1 Ana M. Pérez-Pérez RN,2 Susana Martínez-Ruiz RN,2 Cristina Garrido-Colmenero MD,4 Delphine Roldan RN,2 María del Mar Jiménez-Quintana MD,3 Emilio Sánchez-Cantalejo PhD5 and Manuel Colmenero MD PhD6 1
Principal Investigator 2Registered Nurse 3Medical Doctor, Intensive Care Unit, Hospital Universitario Virgen de las Nieves, Granada, Spain Medical Doctor, Department of Dermatology, Hospital Universitario Virgen de las Nieves, Granada, Spain 5 Professor, Department of Statistics, Andalusian School of Public Health, Granada, Spain 6 Medical Doctor, Intensive Care Unit, Hospital Universitario San Cecilio, Granada, Spain 4
Keywords cohort studies, critically ill, intensive care unit, mechanical ventilation, mortality, outcome studies, pressure ulcers, risk factor Correspondence Dr Francisco Manzano Unidad de Cuidados Intensivos Hospital Universitario Virgen de las Nieves C/Avenida de las Fuerzas Armadas, 2 Granada, 18014 Spain E-mail: [email protected] Dr Manuel Colmenero Unidad de Cuidados Intensivos Hospital Universitario San Cecilio C/Dr. Oloriz, 16 Granada, 18012 Spain E-mail: manuel.colmenero.sspa@ juntadeandalucia.es
Abstract Rationale, aims and objectives Pressure ulcers (PUs) are a common and serious complication in critically ill patients. The aim of this study was to evaluate the relationship between the development of a PU and hospital mortality in patients requiring mechanical ventilation (MV) in an intensive care unit (ICU). Methods A prospective cohort study was performed over two years in patients requiring MV for ≥ 24 hours in a medical-surgical ICU. Primary outcome measure was hospital mortality and main independent variable was the development of a PU grade ≥ II. Hazard ratios (HRs) were calculated using a Cox model with time-dependent covariates. Results Out of 563 patients in the study, 110 (19.5%) developed a PU. Overall hospital mortality was 48.7%. In the adjusted multivariate model, PU onset was a significant independent predictor of mortality (adjusted HR, 1.28; 95% confidence interval, 1.003– 1.65; P = 0.047). The model also included the Acute Physiology and Chronic Health Evaluation II score, total Sequential Organ Failure Assessment on day 3, hepatic cirrhosis and medical admission. Conclusion Within the limitations of a single-centre approach, PU development appears to be associated with an increase in mortality among patients requiring MV for 24 hours or longer.
Accepted for publication: 27 March 2014 doi:10.1111/jep.12137
Introduction Hospital-acquired pressure ulcers (PUs) continue to pose a major health problem for hospitalized adults [1,2]. PUs reduce the quality of life of patients and are associated with higher morbidity and increased health care costs [3–6], and PU onset is an established health care quality indicator [7]. Critically ill patients are at high risk for developing hospital-acquired PUs [8]. The increased risk of circulatory impairment from immobility, hemodynamic and respiratory instability, altered sensory perception, and organ 362
failure all contribute to this increased risk [8]. Although recent evidence suggests an apparent decline in PUs [9], incidence rates continue to be elevated, ranging from 3.3% to 29% depending on the type of intensive care unit (ICU) [10–15]. One reason for the failure to achieve the eradication of PUs in the ICU may be that the implementation of evidence-based preventive guidelines has not been a priority issue for many health care centres and professionals [16,17]. Thus, it was observed that registered nurses tended to pay little attention to PU prevention in patients at risk [17]. Health care professionals in the ICU generally focus on hemodynamic and
Journal of Evaluation in Clinical Practice 20 (2014) 362–368 © 2014 John Wiley & Sons, Ltd.
F. Manzano et al.
respiratory functions because of their immediate effect on the mortality of critical patients, whereas strict skin care may be less rigorous because it is not related to a situation that is considered life-threatening [16,18]. This apparently inadequate awareness of the importance of PU prevention may be in part attributable to the lack of empirical evidence on the potentially severe clinical repercussions of PU onset or on its possible association with increased ICU mortality risk. Thus, although numerous studies have shown that hospital-acquired PUs frequently increase the morbidity of ICU patients [3,5,8], it remains controversial whether they also increase the mortality risk. Comparisons of mortality rates between patients with and without PUs in different wards, hospital settings and nursing homes have yielded contradictory data, with reports that PUs do [19–27] and do not [25,28] increase mortality. In general, there is considered to be greater empirical evidence of an association between the presence of PUs and increased mortality in nursing home residents [19–21] than in acute hospitalized patients [22,23,26]. Few studies have investigated the impact of PUs on the mortality of ICU patients, and these have also reported conflicting results [24,29,30]. In critically ill patients, it has not been established whether an increase in PU-related mortality is attributable to complications of the PU itself or to the severity of their underlying disease or other co-morbidities [29]. This question was addressed in the present study by prospectively gathering a large body of data on the presence of PUs and clinical outcomes in ICU patients at high risk of PU, represented by those requiring mechanical ventilation (MV) for ≥ 24 hours. The objective was to analyse the association between the onset of an ICU-acquired PU and hospital mortality in ventilated adult ICU patients.
Methods Study design A prospective, single-centre, observational study was undertaken in patients under MV from 1 April 2006 through 31 March 2008 in the 26-bed medical-surgical ICU of the Virgen de las Nieves University Hospital, Granada (Spain). All consecutive adult patients (>18 years old) mechanically ventilated for ≥ 24 hours with either endotracheal intubation or non-invasive ventilation and who were PU-free on ICU admission were enrolled in the study. Patients were followed up to death in the hospital or to hospital discharge. The study was approved by the Ethics Committee of the Virgen de las Nieves University Hospital, which waived any need for written consent, given that this epidemiologic observational study did not require any intervention that was not routine medical practice in the unit.
Data collection Data were collected prospectively using pre-printed case report forms. Data were entered centrally using the Statistical Package for Social Sciences (SPSS) version 15.0 for Windows (SPSS Inc., Chicago, IL, USA). Then, all data were re-entered by two encoders, and in cases of inconsistency, were reviewed by a third researcher. An inter-examiner consistency of > 98% per variable and 99% per patient was observed for the entire data entry process. The main outcome variable was hospital mortality. The primary independent variable was the presence during ICU stay of a PU
© 2014 John Wiley & Sons, Ltd.
Pressure ulcers, ventilation and mortality
grade ≥ II according to the European Pressure Ulcer Advisory Panel [31]. PUs caused by medical devices were not considered as ulcers. Patients were examined for PUs before 1000 h every day. If the nurse attending the patient suspected the onset of a PU, the patient was then examined independently by two nurseresearchers, who assessed its presence and severity. In cases of discrepant evaluations, the examination was repeated by a third nurse-researcher to establish the final result. Before the study, all research nurses participated in a training programme that included the definition of a PU and the criteria for establishing its grade. Clinical data were also prospectively gathered in all participants on age, gender, body weight, body mass index (BMI), diabetes mellitus, hospital stay (days) before ICU admission, Norton scale [32], type of admission (medical or surgical), history of cancer, chronic renal failure, hepatic cirrhosis and peripheral vascular disease, the presence of severe anaemia (defined as haemoglobin < 8 g dL−1), and the diagnosis at MV onset (see Table 1). Clinical and laboratory data were collected for the Acute Physiology and Chronic Health Evaluation (APACHE) II [33]. Individual and total Sequential Organ Failure Assessment (SOFA) scores were collected at admission and every 24 hours thereafter [34]. Data were gathered at ICU discharge on ICU mortality, total time on MV and length of ICU stay. Post-ICU length of hospital stay was also recorded.
Statistical analyses Descriptive statistics were computed for all study variables. The Shapiro test was used, and stratified distribution plots were examined to verify the normality of continuous variable distribution. Nonparametric tests of comparison were used for non-normally distributed variables. Difference testing between groups was performed using the two-tailed t-test, Mann–Whitney U-test, chisquare test and Fisher’s exact test as appropriate. Mortality rates (all causes) were compared between PU and non-PU patients using the Kaplan–Meier estimate and the log–rank test. The starting time point was admission to the ICU, and the end point was either death or ICU discharge or PU onset. P < 0.05 was considered significant in all tests. The hazard ratio (HR) and 95% confidence intervals (CIs) for hospital mortality associated with PU acquisition were estimated using a Cox proportional hazard model with time-dependent covariates. This model included as covariates all fixed variables and the occurrence of PU as a time-dependent exposure. Multiple records (with consecutive start and end times) were created for each subject to account for every change in exposure to the timedependent variable over the study period. The time scale in the model was the time since the first appearance of a PU. Fixed variables considered in the Cox regression analysis included age, female gender, BMI, APACHE II score, diagnosis at MV onset, total SOFA score on days 1 and 3, type of admission (medical or surgical), diabetes mellitus, hepatic cirrhosis, hypoproteinaemia, severe anaemia, history of cancer and chronic renal failure. All variables were introduced in the model after exclusion of the presence of collinearity. A backward selection method was then applied, retaining those variables with P values < 0.05 in the final model. The proportional hazard assumption was tested for all variables included in the model by graphical methods. All statistical tests were two-sided. Data were analysed using SPSS version 363
Pressure ulcers, ventilation and mortality
Age, years , Mean ± SD >65 years, n (%) Female, n (%) BMI, kg m−2 Diabetes mellitus, n (%) Hepatic cirrhosis, n (%) APACHE II score at admission Total SOFA on day 1, points Total SOFA on day 3, points Type of admission Medical Surgical Diagnosis at MV onset, n (%) Neurological disease/trauma Heart disease Respiratory failure Heart surgery Gastrointestinal disease Septic Shock Anaemia (Hb < 8 g dL−1) on day 1 Peripheral vascular disease History of cancer, n (%) Chronic renal failure, n (%) Length of ICU stay, days MV duration, days Length of hospital stay, days
F. Manzano et al.
Non-survivors (n = 274)
Survivors (n = 289)
P value
65 ± 3 167 (59) 98 (35.8) 27.9 ± 4.7 91 (54) 28 (10.2) 26.2 ± 8.2 9.04 ± 3.5 9 ± 4.1
61 ± 9 116 (41) 100 (34.6) 28.1 ± 5.8 77 (46) 13 (4.5) 21.3 ± 6.9 7.4 ± 3.01 6.6 ± 3.3
0.09
Journal of Evaluation in Clinical Practice ISSN 1365-2753
Hospital-acquired pressure ulcers and risk of hospital mortality in intensive care patients on mechanical ventilation Francisco Manzano MD PhD,1 Ana M. Pérez-Pérez RN,2 Susana Martínez-Ruiz RN,2 Cristina Garrido-Colmenero MD,4 Delphine Roldan RN,2 María del Mar Jiménez-Quintana MD,3 Emilio Sánchez-Cantalejo PhD5 and Manuel Colmenero MD PhD6 1
Principal Investigator 2Registered Nurse 3Medical Doctor, Intensive Care Unit, Hospital Universitario Virgen de las Nieves, Granada, Spain Medical Doctor, Department of Dermatology, Hospital Universitario Virgen de las Nieves, Granada, Spain 5 Professor, Department of Statistics, Andalusian School of Public Health, Granada, Spain 6 Medical Doctor, Intensive Care Unit, Hospital Universitario San Cecilio, Granada, Spain 4
Keywords cohort studies, critically ill, intensive care unit, mechanical ventilation, mortality, outcome studies, pressure ulcers, risk factor Correspondence Dr Francisco Manzano Unidad de Cuidados Intensivos Hospital Universitario Virgen de las Nieves C/Avenida de las Fuerzas Armadas, 2 Granada, 18014 Spain E-mail: [email protected] Dr Manuel Colmenero Unidad de Cuidados Intensivos Hospital Universitario San Cecilio C/Dr. Oloriz, 16 Granada, 18012 Spain E-mail: manuel.colmenero.sspa@ juntadeandalucia.es
Abstract Rationale, aims and objectives Pressure ulcers (PUs) are a common and serious complication in critically ill patients. The aim of this study was to evaluate the relationship between the development of a PU and hospital mortality in patients requiring mechanical ventilation (MV) in an intensive care unit (ICU). Methods A prospective cohort study was performed over two years in patients requiring MV for ≥ 24 hours in a medical-surgical ICU. Primary outcome measure was hospital mortality and main independent variable was the development of a PU grade ≥ II. Hazard ratios (HRs) were calculated using a Cox model with time-dependent covariates. Results Out of 563 patients in the study, 110 (19.5%) developed a PU. Overall hospital mortality was 48.7%. In the adjusted multivariate model, PU onset was a significant independent predictor of mortality (adjusted HR, 1.28; 95% confidence interval, 1.003– 1.65; P = 0.047). The model also included the Acute Physiology and Chronic Health Evaluation II score, total Sequential Organ Failure Assessment on day 3, hepatic cirrhosis and medical admission. Conclusion Within the limitations of a single-centre approach, PU development appears to be associated with an increase in mortality among patients requiring MV for 24 hours or longer.
Accepted for publication: 27 March 2014 doi:10.1111/jep.12137
Introduction Hospital-acquired pressure ulcers (PUs) continue to pose a major health problem for hospitalized adults [1,2]. PUs reduce the quality of life of patients and are associated with higher morbidity and increased health care costs [3–6], and PU onset is an established health care quality indicator [7]. Critically ill patients are at high risk for developing hospital-acquired PUs [8]. The increased risk of circulatory impairment from immobility, hemodynamic and respiratory instability, altered sensory perception, and organ 362
failure all contribute to this increased risk [8]. Although recent evidence suggests an apparent decline in PUs [9], incidence rates continue to be elevated, ranging from 3.3% to 29% depending on the type of intensive care unit (ICU) [10–15]. One reason for the failure to achieve the eradication of PUs in the ICU may be that the implementation of evidence-based preventive guidelines has not been a priority issue for many health care centres and professionals [16,17]. Thus, it was observed that registered nurses tended to pay little attention to PU prevention in patients at risk [17]. Health care professionals in the ICU generally focus on hemodynamic and
Journal of Evaluation in Clinical Practice 20 (2014) 362–368 © 2014 John Wiley & Sons, Ltd.
F. Manzano et al.
respiratory functions because of their immediate effect on the mortality of critical patients, whereas strict skin care may be less rigorous because it is not related to a situation that is considered life-threatening [16,18]. This apparently inadequate awareness of the importance of PU prevention may be in part attributable to the lack of empirical evidence on the potentially severe clinical repercussions of PU onset or on its possible association with increased ICU mortality risk. Thus, although numerous studies have shown that hospital-acquired PUs frequently increase the morbidity of ICU patients [3,5,8], it remains controversial whether they also increase the mortality risk. Comparisons of mortality rates between patients with and without PUs in different wards, hospital settings and nursing homes have yielded contradictory data, with reports that PUs do [19–27] and do not [25,28] increase mortality. In general, there is considered to be greater empirical evidence of an association between the presence of PUs and increased mortality in nursing home residents [19–21] than in acute hospitalized patients [22,23,26]. Few studies have investigated the impact of PUs on the mortality of ICU patients, and these have also reported conflicting results [24,29,30]. In critically ill patients, it has not been established whether an increase in PU-related mortality is attributable to complications of the PU itself or to the severity of their underlying disease or other co-morbidities [29]. This question was addressed in the present study by prospectively gathering a large body of data on the presence of PUs and clinical outcomes in ICU patients at high risk of PU, represented by those requiring mechanical ventilation (MV) for ≥ 24 hours. The objective was to analyse the association between the onset of an ICU-acquired PU and hospital mortality in ventilated adult ICU patients.
Methods Study design A prospective, single-centre, observational study was undertaken in patients under MV from 1 April 2006 through 31 March 2008 in the 26-bed medical-surgical ICU of the Virgen de las Nieves University Hospital, Granada (Spain). All consecutive adult patients (>18 years old) mechanically ventilated for ≥ 24 hours with either endotracheal intubation or non-invasive ventilation and who were PU-free on ICU admission were enrolled in the study. Patients were followed up to death in the hospital or to hospital discharge. The study was approved by the Ethics Committee of the Virgen de las Nieves University Hospital, which waived any need for written consent, given that this epidemiologic observational study did not require any intervention that was not routine medical practice in the unit.
Data collection Data were collected prospectively using pre-printed case report forms. Data were entered centrally using the Statistical Package for Social Sciences (SPSS) version 15.0 for Windows (SPSS Inc., Chicago, IL, USA). Then, all data were re-entered by two encoders, and in cases of inconsistency, were reviewed by a third researcher. An inter-examiner consistency of > 98% per variable and 99% per patient was observed for the entire data entry process. The main outcome variable was hospital mortality. The primary independent variable was the presence during ICU stay of a PU
© 2014 John Wiley & Sons, Ltd.
Pressure ulcers, ventilation and mortality
grade ≥ II according to the European Pressure Ulcer Advisory Panel [31]. PUs caused by medical devices were not considered as ulcers. Patients were examined for PUs before 1000 h every day. If the nurse attending the patient suspected the onset of a PU, the patient was then examined independently by two nurseresearchers, who assessed its presence and severity. In cases of discrepant evaluations, the examination was repeated by a third nurse-researcher to establish the final result. Before the study, all research nurses participated in a training programme that included the definition of a PU and the criteria for establishing its grade. Clinical data were also prospectively gathered in all participants on age, gender, body weight, body mass index (BMI), diabetes mellitus, hospital stay (days) before ICU admission, Norton scale [32], type of admission (medical or surgical), history of cancer, chronic renal failure, hepatic cirrhosis and peripheral vascular disease, the presence of severe anaemia (defined as haemoglobin < 8 g dL−1), and the diagnosis at MV onset (see Table 1). Clinical and laboratory data were collected for the Acute Physiology and Chronic Health Evaluation (APACHE) II [33]. Individual and total Sequential Organ Failure Assessment (SOFA) scores were collected at admission and every 24 hours thereafter [34]. Data were gathered at ICU discharge on ICU mortality, total time on MV and length of ICU stay. Post-ICU length of hospital stay was also recorded.
Statistical analyses Descriptive statistics were computed for all study variables. The Shapiro test was used, and stratified distribution plots were examined to verify the normality of continuous variable distribution. Nonparametric tests of comparison were used for non-normally distributed variables. Difference testing between groups was performed using the two-tailed t-test, Mann–Whitney U-test, chisquare test and Fisher’s exact test as appropriate. Mortality rates (all causes) were compared between PU and non-PU patients using the Kaplan–Meier estimate and the log–rank test. The starting time point was admission to the ICU, and the end point was either death or ICU discharge or PU onset. P < 0.05 was considered significant in all tests. The hazard ratio (HR) and 95% confidence intervals (CIs) for hospital mortality associated with PU acquisition were estimated using a Cox proportional hazard model with time-dependent covariates. This model included as covariates all fixed variables and the occurrence of PU as a time-dependent exposure. Multiple records (with consecutive start and end times) were created for each subject to account for every change in exposure to the timedependent variable over the study period. The time scale in the model was the time since the first appearance of a PU. Fixed variables considered in the Cox regression analysis included age, female gender, BMI, APACHE II score, diagnosis at MV onset, total SOFA score on days 1 and 3, type of admission (medical or surgical), diabetes mellitus, hepatic cirrhosis, hypoproteinaemia, severe anaemia, history of cancer and chronic renal failure. All variables were introduced in the model after exclusion of the presence of collinearity. A backward selection method was then applied, retaining those variables with P values < 0.05 in the final model. The proportional hazard assumption was tested for all variables included in the model by graphical methods. All statistical tests were two-sided. Data were analysed using SPSS version 363
Pressure ulcers, ventilation and mortality
Age, years , Mean ± SD >65 years, n (%) Female, n (%) BMI, kg m−2 Diabetes mellitus, n (%) Hepatic cirrhosis, n (%) APACHE II score at admission Total SOFA on day 1, points Total SOFA on day 3, points Type of admission Medical Surgical Diagnosis at MV onset, n (%) Neurological disease/trauma Heart disease Respiratory failure Heart surgery Gastrointestinal disease Septic Shock Anaemia (Hb < 8 g dL−1) on day 1 Peripheral vascular disease History of cancer, n (%) Chronic renal failure, n (%) Length of ICU stay, days MV duration, days Length of hospital stay, days
F. Manzano et al.
Non-survivors (n = 274)
Survivors (n = 289)
P value
65 ± 3 167 (59) 98 (35.8) 27.9 ± 4.7 91 (54) 28 (10.2) 26.2 ± 8.2 9.04 ± 3.5 9 ± 4.1
61 ± 9 116 (41) 100 (34.6) 28.1 ± 5.8 77 (46) 13 (4.5) 21.3 ± 6.9 7.4 ± 3.01 6.6 ± 3.3
0.09
