We summarized the baseline demographics using descriptive statistics, including means, medians, proportions, standard deviations (SD), interquartile ranges (IQR), and 95% confidence intervals (95% CI).
We used both unadjusted and standardized phase angle (SPA) for analysis, where SPA = (phase angle − age/sex/body mass index specific mean phase angle value)/standard deviation of the age/sex/body mass index specific standard deviation of the healthy population.31 We used the Kaplan Meir method for survival analysis, and log rank tests to compare between groups.
We conducted multivariate Cox Proportional Hazards regression analysis with backward selection incorporating variables with P-value <0.10 in univariate survival analysis. These variables included the PaP score, PPI, serum albumin, fat free mass, unadjusted phase angle, hand grip strength, maximal inspiratory pressure and standardized phase angle. Age and sex were also included because hand grip strength and maximal inspiratory pressure are dependent on these variables. Palliative Performance status and Karnofsky Performance Status were not included in the multivariate model because they were already part of PaP and PPI, and correlated strongly with these prognostic scores.
We also determined the association between phase angle (both unadjusted and standardized) with various prognostic variables using the Spearman Correlation test.
The sample size justification was based on having at least 10 events (i.e. deaths) for each prognostic variable in the multivariable Cox Proportional Hazards regression model. We anticipated observing at least 120 deaths in 200 patients, thereby providing enough information to include up to 12 prognostic variables in the model. In total, we recruited 222 patients to ensure at least 200 patients completed all 3 functional measures.
The Statistical Analysis System (SAS version 9.2, SAS Institute, Cary, North Carolina) were used for statistical analysis. A P-value of <0.05 was considered significant.
We used both unadjusted and standardized phase angle (SPA) for analysis, where SPA = (phase angle − age/sex/body mass index specific mean phase angle value)/standard deviation of the age/sex/body mass index specific standard deviation of the healthy population.31 We used the Kaplan Meir method for survival analysis, and log rank tests to compare between groups.
We conducted multivariate Cox Proportional Hazards regression analysis with backward selection incorporating variables with P-value <0.10 in univariate survival analysis. These variables included the PaP score, PPI, serum albumin, fat free mass, unadjusted phase angle, hand grip strength, maximal inspiratory pressure and standardized phase angle. Age and sex were also included because hand grip strength and maximal inspiratory pressure are dependent on these variables. Palliative Performance status and Karnofsky Performance Status were not included in the multivariate model because they were already part of PaP and PPI, and correlated strongly with these prognostic scores.
We also determined the association between phase angle (both unadjusted and standardized) with various prognostic variables using the Spearman Correlation test.
The sample size justification was based on having at least 10 events (i.e. deaths) for each prognostic variable in the multivariable Cox Proportional Hazards regression model. We anticipated observing at least 120 deaths in 200 patients, thereby providing enough information to include up to 12 prognostic variables in the model. In total, we recruited 222 patients to ensure at least 200 patients completed all 3 functional measures.
The Statistical Analysis System (SAS version 9.2, SAS Institute, Cary, North Carolina) were used for statistical analysis. A P-value of <0.05 was considered significant.
