3 replies to this topic

[Michael]

High protein intake is associated with low plasma NAD+ levels in a healthy human cohort
PLoS One. 2018 Aug 16;13(8):e0201968. doi: 10.1371/journal.pone.0201968. eCollection 2018.
Neda Seyedsadjadi, Jade Berg, Ayse A. Bilgin, Nady Braidy, Chris Salonikas, Ross Grant

...we used a cross-sectional study design of 100 apparently healthy middle-aged males (n = 48) and females, in which we assessed average dietary intakes of multiple components using a validated questionnaire. ... [P]lasma levels of NAD+ and Total NAD(H) were lower with increasing protein intake (F (2, 92) = 4.61, P = 0.012; F (2, 92) = 4.55, P = 0.013, respectively). The associated decrease in NAD+ and NAD(H) levels was even stronger with increasing plasma levels of the protein breakdown product urea (F (2, 93) = 25.11, P≤0.001; F (2, 93) = 21.10, P≤0.001).
 
[See the graph here. The mean protein intake was 79.29 ± 30.34 g/day; it's unclear from Table 1, but either the tertiles of protein intake were 70, 75, and 94 g protein/d, or the protein intake levels in the three tertiles of urea were thus -MR]
 
These associations were all independent of age, gender and energy intake. ...  but not after Bonferroni adjustment (P>0.002). ... However, no significant association was observed between protein intake or plasma urea, and plasma levels of NAD+ metabolites [[NADP⁺, cyclic ADP ribose (cADPR), nicotinamide (NAM), N-methylnicotinamide (MeNAM)]] ... We also observed that plasma levels of the inflammatory cytokine IL-6, and both Kyn, and Trp, but not the Kyn/Trp ratio were higher with increasing plasma urea levels (F (2, 94) = 3.30, P = 0.041; F (2, 95) = 7.41, P≤0.001; F (2, 96) = 4.23, P = 0.017, respectively). These associations were dependent on eGFR and energy intake, except for the urea and Trp association that was independent of all. In conclusion, we report for the first time, a novel association between protein intake, its metabolism, and plasma NAD+ levels with a possible link to inflammation.

PMID: 30114226 DOI: 10.1371/journal.pone.0201968

The fact that the association didn't survive Bonferroni adjustment may very well mean this is spurious. Repeating this study with this as the primary hypothesis, and in a larger cohort with a wider range of protein intakes, would resolve the issue. I'd also like to see it broken down by animal vs. plant protein.

[Oakman]

"However, no significant association was observed between protein intake or plasma urea, and plasma levels of NAD+metabolites."

 

"In conclusion, we report for the first time, a novel association between protein intake, its metabolism, and plasma NAD+ levels with a possible link to inflammation."

 

Is it just my lack of comprehension, or are these statements contradictory? And although I roughly get the idea of the Bonferroni adjustment, I definitely got lost in its details here. Sadly, I wasn't a good statistics pupil in college.

[Phoebus]

 Our results have shown for the first time that plasma levels of NAD+ and Total NAD(H) were lower with increasing protein intake (F (2, 92) = 4.61, P = 0.012; F (2, 92) = 4.55, P = 0.013, respectively). The associated decrease in NAD+ and NAD(H) levels was even stronger with increasing plasma levels of the protein breakdown product urea (F (2, 93) = 25.11, P≤0.001; F (2, 93) = 21.10, P≤0.001).

 

Edited by Michael, 07 September 2018 - 10:35 PM.

[Michael]

"However, no significant association was observed between protein intake or plasma urea, and plasma levels of NAD metabolites"
 
"In conclusion, we report for the first time, a novel association between protein intake, its metabolism, and plasma NAD+ levels with a possible link to inflammation."
 
Is it just my lack of comprehension, or are these statements contradictory?

No: they found an association of high protein/urea with levels of NAD+ (and NADH), but not with NAD+ metabolites (NADP⁺, cyclic ADP ribose (cADPR), nicotinamide (NAM), N-methylnicotinamide (MeNAM)). That's not a contradiction, tho' if the association were genuinely causal you'd certainly expect that a real effect on NAD(H) would also lead to downstream effects on (or be mediated by upstream effects on) the metabolites. Another reason for skepticism of the finding.
 

And although I roughly get the idea of the Bonferroni adjustment, I definitely got lost in its details here. Sadly, I wasn't a good statistics pupil in college.

Basically, the more possible associations you look at at once, the more likely it is that you'll find a nominally statistically significant effect by chance. Bonferroni adjustment is a way to statistically weed out likely false positives arising from this 'problem of multiple comparisons.' So here, the association collapsed when such an analysis was run, suggesting it may be mere noise in the data.