Table 3:
Mortality risk in older adults prescribed a LPD, according to prescription and measured protein intake over the follow-up period.
| All-cause mortality | ||||
|---|---|---|---|---|
| Prescribed diet (n = 1738) | Number of events/person-years | Crude incidence rate per 100 person-years (IQR) | Unadjusted OR | Adjusted ORa |
| Standard diet (n = 1329b) | 368/3298 | 11.2 (10.1–12.4) | Ref | Ref |
| LPD ≤0.8 g/kg (n = 631b) | 191/1529 | 12.5 (10.8–14.4) | 1.10 (0.93–1.31) | 1.15 (0.86–1.55) |
| LPD ≤0.6 g/kg (n = 363b) | 99/808 | 12.3 (10.1–14.9) | 1.06 (0.85–1.31) | 1.01 (0.73–1.40) |
| Restricted analysis according to measured protein intake (n = 778)b | All-cause mortality | |||
| Standard diet, adherent (n = 280) | 47/593 | 7.9 (5.9–10.6) | Ref. | Ref. |
| LPD ≤0.8 g/kg, adherent (n = 229) | 68/669 | 12.4 (9.8–15.7) | 1.49 (1.03–2.16) | 0.81 (0.46–1.43) |
| LPD ≤0.8 g/kg, non-adherent (n = 280) | 77/697 | 11.0 (8.8–13.8) | 1.33 (0.92–1.92) | 0.97 (0.60–1.58) |
| Standard diet, spontaneously low protein intake (n = 169) | 41/361 | 11.4 (8.4–15.4) | Ref. | Ref. |
| LPD ≤0.8 g/kg, adherent (n = 180) | 56/464 | 12.1 (9.3–15.7) | 1.04 (0.71–1.53) | 1.23 (0.68–2.20) |
Ref: reference.
Dynamic inverse probability of treatment weights and stabilized censoring weights included baseline sex, country, Charlson comorbidity index, diabetes, ischaemic heart disease, peripheral arterial disease, heart failure, cancer, education, marital status, smoking, alcohol intake, time-updated age, eGFR, SGA, BMI, haemoglobin, albumin, urea, potassium, phosphate, systolic and diastolic BP and time (natural cubic spline with 3 knots). The outcome marginal structural model further included age at inclusion, sex, country, comorbidity (diabetes, ischaemic heart disease, stroke, heart failure, peripheral arterial disease), eGFR at inclusion, SGA, smoking, BMI, haemoglobin and albumin at inclusion and time-varying kidney replacement therapy.
Time-varying exposure resulting in a patient that may count in several categories.