Read on Twitter
There is an interesting story hiding in the intersection of:
1. IDEAL
2. FHN Nocturnal
3. STARRT-AKI
#nephrology #dialysis
1. IDEAL
2. FHN Nocturnal
3. STARRT-AKI
#nephrology #dialysis
1/ Consider baseline kidney function.
IDEAL: mean eGFR = 10 mL/min
FHN Nocturnal: 52% of patients still in first year of dialysis, 52% with urine volume >500 mL/day
STARRRT-AKI: mean eGFR = 66-67 mL/min, median urine volume of 450-480 mL/day
IDEAL: mean eGFR = 10 mL/min
FHN Nocturnal: 52% of patients still in first year of dialysis, 52% with urine volume >500 mL/day
STARRRT-AKI: mean eGFR = 66-67 mL/min, median urine volume of 450-480 mL/day
2/ Mortality
IDEAL: HR = 1.04 with early vs. late start
FHN Nocturnal: sorry, 3 deaths (I'm uninterested in post hoc analyses with widespread treatment crossover)
STARRT-AKI: HR = 1.00 with accelerated vs. standard strategy
My conclusion: no effect.
IDEAL: HR = 1.04 with early vs. late start
FHN Nocturnal: sorry, 3 deaths (I'm uninterested in post hoc analyses with widespread treatment crossover)
STARRT-AKI: HR = 1.00 with accelerated vs. standard strategy
My conclusion: no effect.
3/ Infection
IDEAL: death/hospitalization HR = 0.87 (P = 0.20; death HR = 1.46, hospitalization HR = 0.81)
FHN Nocturnal: hospitalization HR = 2.0 (95% CI, 0.8-5.2)
STARRT-AKI: BSI rates of 2.5 vs. 0.4 per 1000 pt-mo (P = 0.07)
My conclusion: dialysis = potential for infection.
IDEAL: death/hospitalization HR = 0.87 (P = 0.20; death HR = 1.46, hospitalization HR = 0.81)
FHN Nocturnal: hospitalization HR = 2.0 (95% CI, 0.8-5.2)
STARRT-AKI: BSI rates of 2.5 vs. 0.4 per 1000 pt-mo (P = 0.07)
My conclusion: dialysis = potential for infection.
4/ Loss of kidney function
FHN Nocturnal: longer sessions lower risk of intradialytic hypotension (duh), but more rapid loss with 6 vs. 3 HD sessions/week
Note that in 6x arm, leading predictor of urine volume loss = change in nadir intradialytic SBP.
FHN Nocturnal: longer sessions lower risk of intradialytic hypotension (duh), but more rapid loss with 6 vs. 3 HD sessions/week
Note that in 6x arm, leading predictor of urine volume loss = change in nadir intradialytic SBP.
STARRT-AKI: hypotension rates of 66 vs. 39 per 1000 pt-mo (P = 0.001)
RR = 1.7 for KRT dependence at 90 days among survivors (95% CI, 1.2-2.4)
RR = 1.7 for KRT dependence at 90 days among survivors (95% CI, 1.2-2.4)
5/ Hypophosphatemia
FHN Nocturnal: 42% of patients in 6x arm required phosphorus in dialysate to avoid hypophosphatemia (vs. 0% in 3x arm)
STARRT-AKI: Phosphorus <0.5 mmol/L rates of 44 vs. 24 per 1000 pt-mo (P < 0.001)
FHN Nocturnal: 42% of patients in 6x arm required phosphorus in dialysate to avoid hypophosphatemia (vs. 0% in 3x arm)
STARRT-AKI: Phosphorus <0.5 mmol/L rates of 44 vs. 24 per 1000 pt-mo (P < 0.001)
6/ Hypokalemia
FHN Nocturnal: nothing significant
STARRT-AKI: nothing significant
FHN Nocturnal: nothing significant
STARRT-AKI: nothing significant
My feeling is that mortality risk is not the issue.
However, endogenous kidney function and dialysis are difficult to combine.
In particular, too much UF and hypotension are the fast tracks to loss of kidney function.
However, endogenous kidney function and dialysis are difficult to combine.
In particular, too much UF and hypotension are the fast tracks to loss of kidney function.
Furthermore, it's easy for dialysis to deplete those solutes that are not ordinarily included in dialysate (e.g., phosphorus).
Finally, dialysis creates an opportunity for infection.
Finally, dialysis creates an opportunity for infection.
For me, the line between how to use HD with vs. without kidney function has never been brighter. Makes me wonder even more about incremental HD (for ESKD).
