Table 3.
Multivariable linear regression model for an association between 24-h NT-proBNP levels and hand grip strength at 6 and 12 months after sepsis*
| Effect | Estimate | SE | p value |
|---|---|---|---|
| 6-month hand grip strength | |||
| Intercept | 57.18 | 3.86 | < 0.0001 |
| NT-proBNP (lowest tertile as reference) | 0.0015† | ||
| Middle tertile | − 1.63 | 2.02 | 0.4222 |
| Highest tertile | − 9.30 | 2.60 | 0.0006 |
| Sex (female) | − 16.18 | 1.88 | < 0.0001 |
| Race (non-Caucasian) | 2.80 | 3.41 | 0.4138 |
| Charlson comorbidity index | − 0.09 | 0.52 | 0.8588 |
| Age | − 0.32 | 0.09 | 0.0003 |
| CCI‡ | − 3.81 | 2.29 | 0.0990 |
| APACHE | − 0.03 | 0.12 | 0.8372 |
| 12-month hand grip strength | |||
| Intercept | 56.87 | 5.21 | < 0.0001 |
| NT-proBNP (lowest tertile as reference) | 0.0427† | ||
| Middle tertile | − 1.20 | 2.23 | 0.5916 |
| Highest tertile | − 8.71 | 3.61 | 0.0177 |
| Sex (female) | − 15.61 | 2.21 | < 0.0001 |
| Race (non-Caucasian) | 3.65 | 2.83 | 0.2010 |
| Charlson comorbidity index | − 0.05 | 0.62 | 0.9380 |
| Age | − 0.31 | 0.12 | 0.0146 |
| CCI‡ | − 4.00 | 2.11 | 0.0617 |
| APACHE | − 0.07 | 0.11 | 0.5386 |
Inverse probability weighting based on concurrent Zubrod scores was used to account for missing follow-up data, as well as absence due to death; NT-proBNP levels: lowest tertile (29–711 ng/L), middle tertile (712–2378 ng/L), and highest tertile (2379–70,000 ng/L)
SE standard error, NT-proBNP N-terminal pro-brain natriuretic peptide, CCI chronic critical illness, APACHE Acute Physiology and Chronic Health Evaluation
†Overall p value
‡Rapid recovery (RAP) patients as a reference group