"High-protein diets above 1.6 g/kg body weight damage kidneys in healthy people."
This claim is false. Multiple large-scale research reviews consistently find that high-protein diets do not damage kidneys in healthy adults — and one large study found they may actually lower kidney disease risk.
What Was Claimed?
The idea that eating a lot of protein — specifically more than 1.6 grams per kilogram of body weight per day — harms your kidneys is widely repeated in fitness and wellness circles. For a 75 kg (165 lb) person, that threshold works out to about 120 grams of protein per day, a level commonly reached by people following high-protein diets or athletes. If this claim were true, a large portion of people pursuing common dietary advice would be quietly damaging their kidneys.
What Did We Find?
The research evidence against this claim is strong and consistent. A 2018 meta-analysis published in the Journal of Nutrition — the most direct test of the claim — pooled results from 28 randomized controlled trials involving 1,358 healthy adults. These were controlled experiments where researchers assigned participants to high- or normal-protein diets and measured kidney function before and after. The result: high-protein intake did not adversely affect the rate at which kidneys filter blood. The groups showed no meaningful difference in how kidney function changed over time.
A separate and independent 2018 systematic review in Advances in Nutrition reached the same conclusion, examining both controlled trials and observational studies in people consuming protein above the standard US dietary recommendation. The researchers found that higher protein intake was consistent with normal kidney function in healthy individuals.
Perhaps most striking is a 2024 meta-analysis that followed 148,051 people across six cohort studies and tracked who developed chronic kidney disease. Rather than finding harm, the study found that higher protein intake was associated with lower risk of kidney disease — for total protein, plant protein, and animal protein alike.
One counterargument worth addressing: high protein does cause the kidneys to filter at a slightly higher rate temporarily, a phenomenon called glomerular hyperfiltration. Some have speculated this could cause damage over time. But this is a normal physiological response, like a heart beating faster during exercise — not a sign of damage. Controlled trials have not found this mechanism to produce actual harm in healthy adults.
What Should You Keep In Mind?
This proof applies specifically to people with healthy kidneys. If you have existing kidney disease, the picture is different: clinical guidelines do recommend limiting protein for that population, and that recommendation is well-established. The claim conflates two distinct groups.
The 1.6 g/kg figure in the claim is not a kidney safety limit set by doctors or kidney disease specialists. It comes from sports nutrition research as an approximate ceiling for muscle-building benefits — a completely different question. No nephrology guideline treats it as a threshold beyond which kidneys are at risk.
The studies here, while large and well-designed, are mostly short to medium term. Some researchers have raised theoretical concerns about very long-term high protein intake, and those questions aren't fully settled. What is clear is that no controlled trial has demonstrated actual kidney damage in healthy adults.
How Was This Verified?
This verdict is based on three independent systematic reviews and meta-analyses, all verified by retrieving the source documents directly and confirming the quoted conclusions. The full reasoning behind each source, the adversarial checks performed, and the logic connecting the evidence to the verdict are documented in the structured proof report and the full verification audit. To inspect or reproduce the verification process yourself, see re-run the proof yourself.
What could challenge this verdict?
1. Is there evidence that high-protein diets cause rapid kidney function decline in healthy adults?
A search for "high protein diet rapid GFR decline healthy adults" identified Jhee et al. (2019, Nephrology Dialysis Transplantation), a Korean community-based prospective cohort of 9,226 healthy adults. The study reported that the highest protein intake quartile had 1.32× higher odds of rapid eGFR decline versus the lowest quartile. This is observational data, not a controlled experiment. Observational associations cannot establish causation — high animal protein consumption co-occurs with high sodium, red meat, and other dietary factors that independently affect kidney function. Crucially, the RCT meta-analysis (B1) — the highest level of evidence — found no adverse GFR change when protein was experimentally controlled. The observational association does not override the controlled trial evidence. Does not break the proof.
2. Does glomerular hyperfiltration from high protein intake cause long-term kidney damage in healthy people?
A search for "glomerular hyperfiltration high protein long-term damage healthy kidneys" identified Ko et al. (2020, JASN) and Kalantar-Zadeh et al. (2020, Nephrology Dialysis Transplantation). Ko et al. acknowledge the hyperfiltration mechanism and note "It is possible that long-term high protein intake may lead to de novo CKD" — but they also note that "long-term trials have not observed an increase in proteinuria" in those without kidney disease. Kalantar-Zadeh et al. argue for protein caution in vulnerable groups but explicitly state "persons with healthy intact kidneys may not be affected by this harmful impact." Adaptive hyperfiltration (transiently elevated GFR) is a normal physiological response to protein loading, not the same as kidney damage. No RCT has demonstrated histological damage or irreversible GFR decline in healthy adults from protein intake. Does not break the proof.
3. Is 1.6 g/kg body weight a clinically established kidney safety limit?
Searches for "1.6 g/kg protein kidney safety limit" and National Kidney Foundation/KDIGO guidelines found no support for this specific threshold as a nephrology safety cutoff. The 1.6 g/kg figure originates in Morton et al. (2018, British Journal of Sports Medicine) as an approximate upper bound for muscle protein synthesis optimization, not a kidney-safety boundary. Clinical nephrology guidelines address protein restriction for people with CKD; they set no upper-limit threshold for healthy adults. Does not break the proof.
Sources
detailed evidence
Evidence Summary
| ID | Fact | Verified |
|---|---|---|
| B1 | Devries et al. (2018) Journal of Nutrition — meta-analysis of 28 RCTs, 1358 healthy adults: HP intakes do not adversely influence GFR | Yes |
| B2 | Van Elswyk et al. (2018) Advances in Nutrition — systematic review of RCTs and observational studies: higher protein consistent with normal kidney function | Yes |
| B3 | Cheng et al. (2024) Frontiers in Nutrition — meta-analysis of 6 cohort studies, 148,051 participants: higher protein associated with lower CKD risk | Yes |
| A1 | Verified source count (sources rejecting the claim) | Computed: 3 verified sources ≥ threshold of 3 |
Proof Logic
The claim asserts that protein intake above 1.6 g/kg/day causes kidney damage in otherwise healthy individuals. To disprove this, the proof requires ≥3 independently sourced systematic reviews or meta-analyses whose verified quotes confirm the claim is false.
B1 — Devries et al. (2018), Journal of Nutrition. This is the most direct evidence: a pre-registered systematic review and meta-analysis of 28 randomized controlled trials with 1,358 participants, explicitly restricted to adults without kidney disease. The RCTs compared high-protein intake (≥1.5 g/kg body weight, or ≥20% of energy, or ≥100 g/day) against normal- or lower-protein intake. The result: "Our analysis indicates that HP intakes do not adversely influence kidney function on GFR in healthy adults." While post-intervention GFR was slightly higher in the HP group (consistent with adaptive hyperfiltration), the change in GFR did not differ between groups — indicating no progressive kidney function loss.
B2 — Van Elswyk et al. (2018), Advances in Nutrition. An independent systematic review covering both RCTs and observational studies in healthy individuals consuming protein above the US RDA (0.8 g/kg). Conclusion: "These data further indicate that, at least in the short term, higher protein intake within the range of recommended intakes for protein is consistent with normal kidney function in healthy individuals." This review is independent of Devries et al. — different author team, different journal, different search protocol — providing a second line of evidence.
B3 — Cheng et al. (2024), Frontiers in Nutrition. The most recent meta-analysis: 6 prospective cohort studies totaling 148,051 participants with 8,746 CKD cases. Rather than finding harm, higher total, plant, and animal protein intake were each associated with reduced CKD incidence (total protein RR = 0.82, 95% CI 0.71–0.94). Conclusion: "The data showed a lower CKD risk significantly associated higher-level dietary total, plant or animal protein (especially for fish and seafood) intake." This further undermines the claim.
A1 — Source count. All three citations were verified live on their source pages (full quote match). Confirmed count: 3 ≥ threshold of 3. Claim holds (in the disproof direction), yielding verdict: DISPROVED.
Conclusion
Verdict: DISPROVED
Three independently verified systematic reviews and meta-analyses (B1, B2, B3) — totaling randomized controlled trial data from 1,358 healthy participants and prospective cohort data from 148,051 participants — consistently find that high-protein diets do not damage kidneys in healthy individuals. The most direct evidence (B1) comes from 28 RCTs specifically measuring GFR changes in adults without kidney disease; GFR changes did not differ between high-protein and normal-protein groups.
The claim confuses two distinct populations: people with pre-existing kidney disease (for whom protein restriction is clinically indicated) and people with healthy kidneys (for whom high protein is not shown to cause damage). The 1.6 g/kg threshold in the claim is not a nephrology safety limit; it is a sports-nutrition figure. No clinical guideline identifies it as a kidney-damage threshold for healthy adults.
All three citations were fully verified on their source pages. No adversarial check broke the proof.
audit trail
All 3 citations verified.
Original audit log
Source: proof.py JSON summary
B1 — Devries et al. (2018) - Status: verified - Method: full_quote - Fetch mode: live - Coverage: N/A (full quote match)
B2 — Van Elswyk et al. (2018) - Status: verified - Method: full_quote - Fetch mode: live - Coverage: N/A (full quote match)
B3 — Cheng et al. (2024) - Status: verified - Method: full_quote - Fetch mode: live - Coverage: N/A (full quote match)
Source: proof.py JSON summary
| Field | Value |
|---|---|
| Subject | High-protein diet above 1.6 g/kg body weight per day |
| Property | Whether it causes measurable kidney damage (GFR decline, proteinuria, or CKD development) in adults without pre-existing kidney disease |
| Operator | >= |
| Threshold | 3 (independent verified sources that reject the claim) |
| Proof direction | disprove |
| Operator note | The claim is DISPROVED if ≥3 independent systematic reviews or meta-analyses confirm that high-protein intake does NOT damage kidneys in healthy adults. "Damage" is interpreted as measurable adverse change in kidney function markers (GFR decline, increased proteinuria, or elevated creatinine). The 1.6 g/kg threshold cited in the claim is not a clinically established kidney-safety boundary; it is commonly cited in sports nutrition literature (Morton et al. 2018) as an approximate upper limit for muscle protein synthesis optimization. The meta-analyses used here study protein intakes ≥1.5 g/kg or "above the US RDA (0.8 g/kg)" — both categories encompass the >1.6 g/kg range in the claim. Sources in empirical_facts reject the claim; sources supporting the claim are in adversarial_checks. |
Natural language claim: High-protein diets above 1.6 g/kg body weight damage kidneys in healthy people.
Formal interpretation: The claim is evaluated as: does high-protein intake (≥1.5–1.6 g/kg/day) cause measurable kidney damage — defined as adverse change in kidney function markers (GFR decline, increased proteinuria, or elevated creatinine) — in adults without pre-existing kidney disease?
Proof direction: Disproof. The sources in this proof reject the claim. Sources arguing in favor of the claim are documented in Counter-Evidence Search.
Operator note: The claim is DISPROVED if ≥3 independent systematic reviews or meta-analyses confirm that high-protein intake does NOT damage kidneys in healthy adults. The 1.6 g/kg threshold is not a clinically established kidney-safety boundary — it appears in sports nutrition literature (Morton et al. 2018) as an approximate upper limit for muscle protein synthesis optimization. The meta-analyses here study protein intakes ≥1.5 g/kg or "above the US RDA (0.8 g/kg)," both of which encompass the >1.6 g/kg range in the claim.
Source: proof.py JSON summary
| Fact ID | Domain | Type | Tier | Note |
|---|---|---|---|---|
| B1 | nih.gov | government | 5 | Government domain (.gov) — PubMed/NCBI abstract page |
| B2 | nih.gov | government | 5 | Government domain (.gov) — PubMed Central full-text |
| B3 | frontiersin.org | academic | 4 | Known academic/scholarly publisher |
All sources have tier ≥ 4. No low-credibility sources were used.
Source: proof.py inline output (execution trace)
[✓] devries_2018: Full quote verified for devries_2018 (source: tier 5/government)
[✓] van_elswyk_2018: Full quote verified for van_elswyk_2018 (source: tier 5/government)
[✓] cheng_2024: Full quote verified for cheng_2024 (source: tier 4/academic)
Confirmed sources: 3 / 3
verified source count vs threshold (disproof: sources rejecting the claim): 3 >= 3 = True
Source: proof.py JSON summary
| Description | Sources Consulted | Sources Verified | Agreement |
|---|---|---|---|
| Multiple independent systematic reviews consulted | 3 | 3 | Yes |
Source statuses: - devries_2018: verified - van_elswyk_2018: verified - cheng_2024: verified
Independence note: B1 (Devries 2018) and B2 (Van Elswyk 2018) are independent systematic reviews published simultaneously in different journals (Journal of Nutrition and Advances in Nutrition). B3 (Cheng 2024) is a more recent independent meta-analysis from a different author group. All three draw on overlapping but not identical underlying study pools. Independence is at the publication and author-team level; as is typical for meta-analyses in nutrition, they may draw on some of the same primary studies.
Source: proof.py JSON summary
Source: proof.py JSON summary
Check 1: Is there evidence that high-protein diets cause rapid kidney function decline in healthy adults?
- Verification performed: Searched "high protein diet rapid GFR decline healthy adults" and "high protein kidney function decline cohort study"; found Jhee et al. (2019, Nephrology Dialysis Transplantation), a Korean community-based prospective cohort of 9,226 adults without kidney disease, reporting that the highest protein intake quartile had 1.32× higher odds of rapid eGFR decline vs the lowest quartile.
- Finding: Jhee et al. (2019) is an observational association study, not an RCT. Observational findings cannot establish causation due to confounding (high animal protein often co-occurs with high sodium, red meat, etc.). The RCT meta-analysis by Devries et al. (2018) — the highest level of evidence — found no adverse GFR change when protein intake was experimentally manipulated in controlled trials. The observational association does not override the controlled trial evidence.
- Breaks proof: No
Check 2: Does glomerular hyperfiltration from high protein intake cause long-term kidney damage in healthy people?
- Verification performed: Searched "glomerular hyperfiltration high protein long-term damage healthy kidneys"; reviewed Ko et al. (2020, JASN, PMC7460905) and Kalantar-Zadeh et al. (2020, Nephrology Dialysis Transplantation, doi:10.1093/ndt/gfz216). Ko et al. state in the abstract that "evidence suggests that worsening renal function may occur in individuals with and perhaps without impaired kidney function" and note "It is possible that long-term high protein intake may lead to de novo CKD." Kalantar-Zadeh et al. argue protein restriction is warranted for vulnerable populations.
- Finding: Both papers acknowledge the hyperfiltration mechanism but qualify their concerns for healthy people. Ko et al. (2020) also note that "long-term trials have not observed an increase in proteinuria" in those without kidney disease. Kalantar-Zadeh et al. (2020) explicitly state "persons with healthy intact kidneys may not be affected by this harmful impact." The concerns are speculative for healthy individuals; no RCT has demonstrated actual kidney damage (not just adaptive hyperfiltration) in healthy adults.
- Breaks proof: No
Check 3: Is 1.6 g/kg body weight a clinically established kidney safety limit?
- Verification performed: Searched "1.6 g/kg protein kidney safety limit" and "1.6 g/kg protein kidney damage threshold"; searched National Kidney Foundation guidelines and nephrology clinical guidelines.
- Finding: The 1.6 g/kg figure originates in sports nutrition literature as the upper bound for muscle protein synthesis optimization (Morton et al. 2018, Br J Sports Med), not as a nephrology safety threshold. Clinical nephrology guidelines (National Kidney Foundation, KDIGO) address protein restriction for people WITH CKD; they do not identify 1.6 g/kg as a risk cutoff for healthy adults. There is no established clinical threshold beyond which healthy kidneys are damaged by protein intake.
- Breaks proof: No
- Rule 1 — No hand-typed extracted values: N/A — qualitative proof; no numeric values extracted from quotes. Auto-passed by validator.
- Rule 2 — Citations verified by fetching: All three citations verified live.
verify_all_citations()imported and called. ✓ - Rule 3 — Anchored to system time:
date.today()used forgenerated_at. No time-dependent claim logic required. ✓ - Rule 4 — Explicit claim interpretation:
CLAIM_FORMALdict present withoperator_notespecifying disproof direction, threshold rationale, and interpretation of "damage." ✓ - Rule 5 — Adversarial checks searched for counter-evidence: Three adversarial checks performed: (a) observational evidence for GFR decline (Jhee 2019), (b) hyperfiltration-to-damage speculation (Ko 2020, Kalantar-Zadeh 2020), (c) origin of 1.6 g/kg threshold. ✓
- Rule 6 — Cross-checks used independently sourced inputs: Three independent systematic review teams (Devries, Van Elswyk, Cheng) from different journals and institutions. ✓
- Rule 7 — Constants/formulas from computations.py:
compare()imported and used for claim evaluation. No hard-coded formulas. ✓ - validate_proof.py result: PASS — 15/15 checks, 0 issues, 0 warnings.
Source: proof.py JSON summary
For qualitative proofs, extraction records reflect citation verification status per source rather than numeric value extraction.
| Fact ID | Value (status) | Value in Quote | Quote Snippet (first 80 chars) |
|---|---|---|---|
| B1 | verified | True | "Our analysis indicates that HP intakes do not adversely influence kidney functio" |
| B2 | verified | True | "These data further indicate that, at least in the short term, higher protein int" |
| B3 | verified | True | "The data showed a lower CKD risk significantly associated higher-level dietary t" |
Source: proof.py JSON summary. No numeric extraction was performed — this is a qualitative consensus proof.
Cite this proof
Proof Engine. (2026). Claim Verification: “High-protein diets above 1.6 g/kg body weight damage kidneys in healthy people.” — Disproved. https://doi.org/10.5281/zenodo.19489789
Proof Engine. "Claim Verification: “High-protein diets above 1.6 g/kg body weight damage kidneys in healthy people.” — Disproved." 2026. https://doi.org/10.5281/zenodo.19489789.
@misc{proofengine_high_protein_diets_above_1_6_g_kg_body_weight_damage_kidneys_in_healthy_people,
title = {Claim Verification: “High-protein diets above 1.6 g/kg body weight damage kidneys in healthy people.” — Disproved},
author = {{Proof Engine}},
year = {2026},
url = {https://proofengine.info/proofs/high-protein-diets-above-1-6-g-kg-body-weight-damage-kidneys-in-healthy-people/},
note = {Verdict: DISPROVED. Generated by proof-engine v1.3.1},
doi = {10.5281/zenodo.19489789},
}
TY - DATA TI - Claim Verification: “High-protein diets above 1.6 g/kg body weight damage kidneys in healthy people.” — Disproved AU - Proof Engine PY - 2026 UR - https://proofengine.info/proofs/high-protein-diets-above-1-6-g-kg-body-weight-damage-kidneys-in-healthy-people/ N1 - Verdict: DISPROVED. Generated by proof-engine v1.3.1 DO - 10.5281/zenodo.19489789 ER -
View proof source
This is the exact proof.py that was deposited to Zenodo and runs when you re-execute via Binder. Every fact in the verdict above traces to code below.
"""
Proof: High-protein diets above 1.6 g/kg body weight damage kidneys in healthy people.
Generated: 2026-03-31
"""
import json
import os
import sys
PROOF_ENGINE_ROOT = os.environ.get("PROOF_ENGINE_ROOT")
if not PROOF_ENGINE_ROOT:
_d = os.path.dirname(os.path.abspath(__file__))
while _d != os.path.dirname(_d):
if os.path.isdir(os.path.join(_d, "proof-engine", "skills", "proof-engine", "scripts")):
PROOF_ENGINE_ROOT = os.path.join(_d, "proof-engine", "skills", "proof-engine")
break
_d = os.path.dirname(_d)
if not PROOF_ENGINE_ROOT:
raise RuntimeError("PROOF_ENGINE_ROOT not set and skill dir not found via walk-up from proof.py")
sys.path.insert(0, PROOF_ENGINE_ROOT)
from datetime import date
from scripts.verify_citations import verify_all_citations, build_citation_detail
from scripts.computations import compare
# 1. CLAIM INTERPRETATION (Rule 4)
CLAIM_NATURAL = "High-protein diets above 1.6 g/kg body weight damage kidneys in healthy people."
CLAIM_FORMAL = {
"subject": "High-protein diet above 1.6 g/kg body weight per day",
"property": (
"whether it causes measurable kidney damage (GFR decline, proteinuria, or CKD "
"development) in adults without pre-existing kidney disease"
),
"operator": ">=",
"operator_note": (
"The claim is DISPROVED if >=3 independent systematic reviews or meta-analyses "
"confirm that high-protein intake does NOT damage kidneys in healthy adults. "
"'Damage' is interpreted as measurable adverse change in kidney function markers "
"(GFR decline, increased proteinuria, or elevated creatinine). "
"The 1.6 g/kg threshold cited in the claim is not a clinically established "
"kidney-safety boundary; it is commonly cited in sports nutrition literature "
"(Morton et al. 2018) as an approximate upper limit for muscle protein synthesis "
"optimization. The meta-analyses used here study protein intakes >=1.5 g/kg or "
"'above the US RDA (0.8 g/kg)' -- both categories encompass the >1.6 g/kg range "
"in the claim. This is a disproof (proof_direction='disprove'): the sources in "
"empirical_facts reject the claim; sources supporting the claim are in "
"adversarial_checks."
),
"threshold": 3,
"proof_direction": "disprove",
}
# 2. FACT REGISTRY
FACT_REGISTRY = {
"B1": {
"key": "devries_2018",
"label": (
"Devries et al. (2018) Journal of Nutrition — meta-analysis of 28 RCTs, "
"1358 healthy adults: HP intakes do not adversely influence GFR"
),
},
"B2": {
"key": "van_elswyk_2018",
"label": (
"Van Elswyk et al. (2018) Advances in Nutrition — systematic review of RCTs "
"and observational studies: higher protein consistent with normal kidney function"
),
},
"B3": {
"key": "cheng_2024",
"label": (
"Cheng et al. (2024) Frontiers in Nutrition — meta-analysis of 6 cohort "
"studies, 148,051 participants: higher protein associated with lower CKD risk"
),
},
"A1": {"label": "Verified source count (sources rejecting the claim)", "method": None, "result": None},
}
# 3. EMPIRICAL FACTS — sources that REJECT the claim (confirm it is false)
empirical_facts = {
"devries_2018": {
"quote": (
"Our analysis indicates that HP intakes do not adversely influence "
"kidney function on GFR in healthy adults."
),
"url": "https://pubmed.ncbi.nlm.nih.gov/30383278/",
"source_name": (
"Devries MC et al. (2018) Changes in Kidney Function Do Not Differ "
"between Healthy Adults Consuming Higher- Compared with Lower- or "
"Normal-Protein Diets: A Systematic Review and Meta-Analysis. "
"Journal of Nutrition, 148(11):1760-1775. DOI:10.1093/jn/nxy197"
),
},
"van_elswyk_2018": {
"quote": (
"These data further indicate that, at least in the short term, higher "
"protein intake within the range of recommended intakes for protein is "
"consistent with normal kidney function in healthy individuals."
),
"url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC6054213/",
"source_name": (
"Van Elswyk ME et al. (2018) A Systematic Review of Renal Health in "
"Healthy Individuals Associated with Protein Intake above the US "
"Recommended Daily Allowance in Randomized Controlled Trials and "
"Observational Studies. Advances in Nutrition, 9(4):404-418. "
"DOI:10.1093/advances/nmy026"
),
},
"cheng_2024": {
"quote": (
"The data showed a lower CKD risk significantly associated higher-level "
"dietary total, plant or animal protein (especially for fish and seafood) "
"intake."
),
"url": "https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2024.1408424/full",
"source_name": (
"Cheng Y et al. (2024) Association between dietary protein intake and risk "
"of chronic kidney disease: a systematic review and meta-analysis. "
"Frontiers in Nutrition, 11:1408424. DOI:10.3389/fnut.2024.1408424"
),
},
}
# 4. CITATION VERIFICATION (Rule 2)
citation_results = verify_all_citations(empirical_facts, wayback_fallback=True)
# 5. COUNT SOURCES WITH VERIFIED CITATIONS
COUNTABLE_STATUSES = ("verified", "partial")
n_confirmed = sum(
1 for key in empirical_facts
if citation_results[key]["status"] in COUNTABLE_STATUSES
)
print(f" Confirmed sources: {n_confirmed} / {len(empirical_facts)}")
# 6. CLAIM EVALUATION — MUST use compare(), never hardcode claim_holds
claim_holds = compare(
n_confirmed,
CLAIM_FORMAL["operator"],
CLAIM_FORMAL["threshold"],
label="verified source count vs threshold (disproof: sources rejecting the claim)",
)
# 7. ADVERSARIAL CHECKS (Rule 5)
# These are sources that SUPPORT the claim (i.e., argue high protein DOES harm kidneys).
# Searched for counter-evidence before writing this proof.
adversarial_checks = [
{
"question": (
"Is there evidence that high-protein diets cause rapid kidney function "
"decline in healthy adults?"
),
"verification_performed": (
"Searched 'high protein diet rapid GFR decline healthy adults' and "
"'high protein kidney function decline cohort study'; found Jhee et al. "
"(2019, Nephrology Dialysis Transplantation), a Korean community-based "
"prospective cohort of 9,226 adults without kidney disease, reporting that "
"the highest protein intake quartile had 1.32x higher odds of rapid eGFR "
"decline vs the lowest quartile."
),
"finding": (
"Jhee et al. (2019) is an observational association study, not an RCT. "
"Observational findings cannot establish causation due to confounding "
"(high animal protein often co-occurs with high sodium, red meat, etc.). "
"The RCT meta-analysis by Devries et al. (2018) — the highest level of "
"evidence — found no adverse GFR change when protein intake was "
"experimentally manipulated in controlled trials. The observational "
"association does not override the controlled trial evidence."
),
"breaks_proof": False,
},
{
"question": (
"Does glomerular hyperfiltration from high protein intake cause long-term "
"kidney damage in healthy people?"
),
"verification_performed": (
"Searched 'glomerular hyperfiltration high protein long-term damage healthy "
"kidneys'; reviewed Ko et al. (2020, JASN, PMC7460905) and Kalantar-Zadeh "
"et al. (2020, Nephrology Dialysis Transplantation, doi:10.1093/ndt/gfz216). "
"Ko et al. state in the abstract that 'evidence suggests that worsening renal "
"function may occur in individuals with and perhaps without impaired kidney "
"function' and note 'It is possible that long-term high protein intake may "
"lead to de novo CKD.' Kalantar-Zadeh et al. argue protein restriction is "
"warranted for vulnerable populations."
),
"finding": (
"Both papers acknowledge the hyperfiltration mechanism but qualify their "
"concerns for healthy people. Ko et al. (2020) also note that 'long-term "
"trials have not observed an increase in proteinuria' in those without kidney "
"disease. Kalantar-Zadeh et al. (2020) explicitly state 'persons with healthy "
"intact kidneys may not be affected by this harmful impact.' The concerns are "
"speculative for healthy individuals; no RCT has demonstrated actual kidney "
"damage (not just adaptive hyperfiltration) in healthy adults."
),
"breaks_proof": False,
},
{
"question": (
"Is 1.6 g/kg body weight a clinically established kidney safety limit?"
),
"verification_performed": (
"Searched '1.6 g/kg protein kidney safety limit' and '1.6 g/kg protein "
"kidney damage threshold'; searched National Kidney Foundation guidelines "
"and nephrology clinical guidelines."
),
"finding": (
"The 1.6 g/kg figure originates in sports nutrition literature as the upper "
"bound for muscle protein synthesis optimization (Morton et al. 2018, Br J "
"Sports Med), not as a nephrology safety threshold. Clinical nephrology "
"guidelines (National Kidney Foundation, KDIGO) address protein restriction "
"for people WITH CKD; they do not identify 1.6 g/kg as a risk cutoff for "
"healthy adults. There is no established clinical threshold beyond which "
"healthy kidneys are damaged by protein intake."
),
"breaks_proof": False,
},
]
# 8. VERDICT AND STRUCTURED OUTPUT
if __name__ == "__main__":
any_unverified = any(
cr["status"] != "verified" for cr in citation_results.values()
)
is_disproof = CLAIM_FORMAL.get("proof_direction") == "disprove"
any_breaks = any(ac.get("breaks_proof") for ac in adversarial_checks)
if any_breaks:
verdict = "UNDETERMINED"
elif claim_holds and not any_unverified:
verdict = "DISPROVED" if is_disproof else "PROVED"
elif claim_holds and any_unverified:
verdict = (
"DISPROVED (with unverified citations)"
if is_disproof
else "PROVED (with unverified citations)"
)
elif not claim_holds:
verdict = "UNDETERMINED"
else:
verdict = "UNDETERMINED"
FACT_REGISTRY["A1"]["method"] = f"count(verified citations) = {n_confirmed}"
FACT_REGISTRY["A1"]["result"] = str(n_confirmed)
citation_detail = build_citation_detail(FACT_REGISTRY, citation_results, empirical_facts)
extractions = {}
for fid, info in FACT_REGISTRY.items():
if not fid.startswith("B"):
continue
ef_key = info["key"]
cr = citation_results.get(ef_key, {})
extractions[fid] = {
"value": cr.get("status", "unknown"),
"value_in_quote": cr.get("status") in COUNTABLE_STATUSES,
"quote_snippet": empirical_facts[ef_key]["quote"][:80],
}
summary = {
"fact_registry": {
fid: {k: v for k, v in info.items()}
for fid, info in FACT_REGISTRY.items()
},
"claim_formal": CLAIM_FORMAL,
"claim_natural": CLAIM_NATURAL,
"citations": citation_detail,
"extractions": extractions,
"cross_checks": [
{
"description": "Multiple independent systematic reviews consulted",
"n_sources_consulted": len(empirical_facts),
"n_sources_verified": n_confirmed,
"sources": {k: citation_results[k]["status"] for k in empirical_facts},
"independence_note": (
"B1 (Devries 2018) and B2 (Van Elswyk 2018) are independent "
"systematic reviews published simultaneously in different journals "
"(Journal of Nutrition and Advances in Nutrition). B3 (Cheng 2024) "
"is a more recent independent meta-analysis from a different "
"author group. All three draw on overlapping but not identical "
"underlying study pools."
),
}
],
"adversarial_checks": adversarial_checks,
"verdict": verdict,
"key_results": {
"n_confirmed": n_confirmed,
"threshold": CLAIM_FORMAL["threshold"],
"operator": CLAIM_FORMAL["operator"],
"claim_holds": claim_holds,
"proof_direction": "disprove",
},
"generator": {
"name": "proof-engine",
"version": open(os.path.join(PROOF_ENGINE_ROOT, "VERSION")).read().strip(),
"repo": "https://github.com/yaniv-golan/proof-engine",
"generated_at": date.today().isoformat(),
},
}
print("\n=== PROOF SUMMARY (JSON) ===")
print(json.dumps(summary, indent=2, default=str))
Re-execute this proof
The verdict above is cached from when this proof was minted. To re-run the exact
proof.py shown in "View proof source" and see the verdict recomputed live,
launch it in your browser — no install required.
Re-execute the exact bytes deposited at Zenodo.
Re-execute in Binder runs in your browser · ~60s · no installFirst run takes longer while Binder builds the container image; subsequent runs are cached.
machine-readable formats
Downloads & raw data
found this useful? ★ star on github