Carbidopa/Levodopa

Limits on what this shows

• Key negative/null findings: González-Rodríguez et al. 2021 (Nature) and the broader complex-I-deficiency literature in Parkinson's disease attribute mitochondrial complex I impairment to the disease process itself, genetic and idiopathic, not to levodopa treatment; this deficiency is present even in untreated, de novo patients. No human peripheral-tissue (platelet, lymphocyte, muscle) study of levodopa's own effect on mitochondrial function was located after multiple search attempts.
• At-risk subgroups: None established in humans.

Regulatory status

Not addressed in the current US label.

Population evidence

None at Tier A in humans. Tier B mechanistic evidence only: Miyazaki et al. 2007 (FEBS Lett), Asanuma et al. 2010 (Ann Neurol), and Takeshima et al. 2011 (Neurol Res) each use levodopa/dopamine-quinone as the specific toxic agent in cell or animal models, describing dopamine-quinone-induced dopaminergic neurotoxicity and testing candidate protective compounds against it.

Mechanism

Named. Dopamine-quinone formation from levodopa-derived dopamine generates reactive oxygen species and can covalently modify proteins, a route mechanistically adjacent to, but not equivalent to, direct mitochondrial electron-transport-chain inhibition.

Counter-evidence held on record

The disease-pathophysiology literature functions as the counter-evidence constraining this entry to yellow: it demonstrates mitochondrial complex I deficiency exists in Parkinson's disease independent of levodopa exposure, the confounding-by-indication caution that applies here. This is cell and animal mechanistic evidence only, using levodopa's own oxidation chemistry, and must not be read as "levodopa damages your mitochondria" in humans. The well-established mitochondrial finding in Parkinson's disease is a feature of the disease itself, not a demonstrated consequence of taking this medication.

Limits on what this shows

• Key negative/null findings: These are small pilot or crossover studies; a larger controlled comparison cleanly separating drug-attributable oxidative stress from disease-intrinsic oxidative stress, present even in untreated de novo Parkinson's disease per the broader pathophysiology literature, is not established at scale.
• At-risk subgroups: None specifically defined in the available small studies.

Regulatory status

Not addressed in the current US label.

Population evidence

Müller et al. 2016 (J Neural Transm), pilot study: plasma oxidative-stress markers and repulsive guidance molecule A rose after levodopa administration in Parkinson's patients. Muhlack et al. 2017 (Neurol Res), crossover: 100 mg oral levodopa raised homocysteine and cysteinyl-glycine over 3 hours, an effect not seen with transdermal rotigotine or placebo. Martignoni et al. 1999 (Free Radic Biol Med), peripheral oxidative-stress markers examined with a specific focus on the role of L-DOPA. Buhmann et al. 2004 (Neurobiol Dis), plasma/CSF oxidative markers elevated in Parkinson's disease and influenced by antiparkinsonian medication.

Mechanism

Named. Dopamine and levodopa auto-oxidize to reactive quinones and reactive oxygen species; separately, COMT methylation consumes SAMe. Both are plausible, independently documented routes to the measured biomarker changes.

Counter-evidence held on record

Parkinson's disease itself carries an intrinsic oxidative burden independent of treatment (dopamine metabolism, nigral iron handling), a competing explanation for any global oxidative-marker difference versus healthy controls; the rotigotine-versus-oral-levodopa crossover design in Muhlack et al. helps isolate an incremental drug effect from baseline disease state. This reflects levodopa's own redox chemistry, not a nutrient the patient is failing to obtain, and is not a reason to add high-dose antioxidant supplements independently; large antioxidant trials in Parkinson's disease have not shown benefit from doing so (see Fails verification).

Limits on what this shows

• Key negative/null findings: No evidence found of levodopa lowering serum iron, ferritin, or transferrin saturation; this is a one-directional bioavailability interaction, not a nutritional-depletion signal.
• At-risk subgroups: Patients concurrently taking iron supplements or iron-fortified multivitamins for anemia, prenatal vitamins, or as part of a GI blood-loss workup.

Regulatory status

Tier A. Current US label: iron salts "can form chelates with levodopa and carbidopa and consequently reduce the bioavailability of carbidopa and levodopa."

Population evidence

Label-level pharmacokinetic statement only. A dedicated search for levodopa's effect on serum iron, ferritin, or iron-deficiency-anemia status in Parkinson's patients, tried with multiple phrasings, returned no indexed primary studies; the existing iron-and-Parkinson's-disease literature concentrates heavily on brain (nigral) iron accumulation as a feature of disease pathology, a different question from peripheral iron status and outside this drug-specific scope.

Mechanism

Named. Direct chelation of levodopa/carbidopa by iron cations in the gastrointestinal tract, reducing absorbed drug; not a mechanism by which the drug consumes or wastes the body's iron.

Counter-evidence held on record

None needed; the direction of the interaction is already established by the label itself, and no contrary evidence changes it. Do not read this as levodopa causing iron deficiency, and do not increase iron intake in response to it. The correct action is timing separation, generally at least 2 hours between iron-containing products and carbidopa-levodopa dosing, on a prescriber's guidance, to protect the drug's effectiveness, not to protect iron status.

Limits on what this shows

• Key negative/null findings: No human data were located in either direction. The Weiner 1978 finding is a single, small, 47-year-old animal study that a dedicated search this session found no later replication or contradiction of, in guinea pigs or any other species, brain or peripheral copper.
• At-risk subgroups: None established in humans.

Regulatory status

Not addressed in the current US label; no copper interaction, monitoring recommendation, or caution appears in the labeling sections reviewed.

Population evidence

Weiner, Nausieda, and Klawans 1978 (Neurology), animal study in guinea pigs: chronic administration of levodopa-carbidopa, bromocriptine, or lergotrile altered brain trace-metal concentrations across multiple brain regions, consistently elevating manganese while decreasing copper. The levodopa-carbidopa combination itself, not levodopa alone, was one of the tested regimens, so this passes the cohort test for this drug. Rehmani et al. 2017 (Toxicol In Vitro), in vitro mechanistic study: L-DOPA and dopamine undergo copper-mediated oxidative DNA damage, with copper acting as the redox catalyst for L-DOPA's own pro-oxidant chemistry, a direct chemical interaction between the drug's active moiety and copper. No Tier A human study measuring serum, CSF, or tissue copper or ceruloplasmin specifically in carbidopa-levodopa-treated patients was located.

Mechanism

Named at Tier B only, and two distinct routes. First, an observed but mechanistically unexplained brain trace-metal redistribution with chronic levodopa-carbidopa dosing in the 1978 animal study. Second, copper's independent, well-characterized role as a redox-active transition metal that catalyzes L-DOPA/dopamine auto-oxidation to reactive quinones and reactive oxygen species, the same auto-oxidation chemistry underlying this dossier's oxidative-stress and mitochondrial-burden entries (see Overlap map).

Counter-evidence held on record

No source located directly contradicts either finding, but neither has been independently replicated. The Weiner 1978 result is old, small-scale, guinea-pig-only, and offers no mechanism for the copper decrease specifically; a near-half-century gap with no follow-up study is itself a reason for caution about how much weight this carries. This is animal and in vitro mechanistic evidence only, not a demonstrated human copper deficiency or excess from this drug. It is not a reason to supplement or restrict dietary copper, and no clinician-ordered copper or ceruloplasmin monitoring is currently indicated by this evidence base alone.

Limits on what this shows

• Key negative/null findings: No source located shows carbidopa-levodopa itself reducing gut microbial diversity or damaging the microbiome as an ecosystem; the evidence base concerns a specific subset of bacteria consuming the drug, not drug-induced dysbiosis.
• At-risk subgroups: Patients with higher fecal tyrosine-decarboxylase-positive bacterial abundance show more inter-individual variability in dose requirement and motor response; this remains an active pharmacomicrobiomics research area, including bacterial-enzyme inhibitor co-administration strategies not yet approved.

Regulatory status

Not addressed in the current US label.

Population evidence

Maini Rekdal et al. 2019 (Science), isolated human gut bacteria capable of decarboxylating levodopa to dopamine in the gut lumen via bacterial tyrosine decarboxylase, identified a downstream Eggerthella lenta pathway that further degrades that dopamine, and demonstrated a small-molecule inhibitor of the bacterial enzyme in a mouse model. Hu et al. 2024 (CNS Neurosci Ther) and Hor et al. 2025 (Parkinsonism Relat Disord), fecal tyrosine-decarboxylase activity or abundance associated with inter-patient variability in levodopa motor response.

Mechanism

Named and direction-checked: this is the gut microbiome affecting the drug's bioavailability, the reverse of a drug harming the microbiome. Carbidopa, a peripheral decarboxylase inhibitor, does not reach or act on the bacterial enzyme, so it does not protect against this specific pathway.

Counter-evidence held on record

This is a mechanistic and translational research area rather than a validated clinical intervention; no approved product currently targets the bacterial pathway. This does not mean levodopa harms gut health, and probiotics or prebiotics are not a validated fix for it. It means an individual's own gut bacteria can partly explain why the same oral dose produces a different clinical response between patients, a question for a treating neurologist, not a supplement aisle.

Limits on what this shows

• Key negative/null findings: No published case series quantifying incidence was located; this reads as a genuine rare, idiosyncratic postmarketing signal rather than a common, dose-related effect.
• At-risk subgroups: None identified in the sources reviewed.

Regulatory status

Tier A. Current FDA label(s) list hemolytic and non-hemolytic anemia, thrombocytopenia, and leukopenia among hematologic adverse reactions; no incidence rate is given.

Population evidence

Label-level postmarketing/adverse-reaction listing only. A dedicated PubMed search for carbidopa-levodopa case reports of hemolytic anemia, tried with multiple phrasings, returned no indexed results.

Mechanism

Not established in the sources located. Plausible general routes for drug-induced hemolysis include immune-mediated mechanisms or oxidative red-cell membrane damage; the latter is conceptually adjacent to the oxidative-stress entry above, but no source specific to carbidopa-levodopa ties the two together directly, so this link is noted as speculative rather than established.

Counter-evidence held on record

No source contradicts the label listing; the limitation is absence of quantification, not a competing finding. This is a clinical hematologic outcome, not a nutrient deficiency. Per the general rule that anemia is not the same as iron or B12 deficiency, any anemia in a patient on this drug warrants standard hematologic workup, reticulocyte count, direct Coombs test, hemolysis markers, rather than empiric iron or B12 supplementation.

Limits on what this shows

• Why no grade: This category was re-searched in full for this repair pass, with a dedicated tool budget and multiple additional phrasings combining levodopa/carbidopa-levodopa with ATP production, mitochondrial complex I activity, and oxidative phosphorylation, in general and peripheral-tissue-specific (platelet, lymphocyte, muscle) forms. One promising, recently indexed title surfaced on a PubMed search-listing page early in the session, "Molecular analysis of mitochondrial complex I in the levodopa short duration response in Parkinson's disease" (Bispo AG et al., Parkinsonism Relat Disord, 2026), but its abstract could not be retrieved and its cohort (human or animal, tissue studied, and whether ATP production itself, rather than complex I genetics or expression, was measured) could not be confirmed: the PubMed abstract page began returning a Google reCAPTCHA challenge, and repeated OpenAlex title-search, Crossref, and Europe PMC lookups for it were blocked (429 or 403) before a PMID or DOI could be captured. It is disclosed here as a named, unconfirmed lead for a future revision, not cited as supporting evidence, since the mandatory cohort test could not be completed. This session's network proxy status log confirms outright 403 policy-level connection denials, not merely rate-limiting, for api.crossref.org, www.ebi.ac.uk (Europe PMC), dailymed.nlm.nih.gov, eutils.ncbi.nlm.nih.gov, and api.fda.gov; the Semantic Scholar Graph API and OpenAlex's search/filter endpoints (as distinct from its per-PMID single-record endpoint, which did work, see the Copper entry above) returned repeated 429 responses throughout; and PubMed's own search-listing pages, reachable early in the session, began serving a reCAPTCHA challenge to non-browser fetches partway through. No drug-specific consumer claim that carbidopa-levodopa itself depletes cellular ATP was located, distinct from the general claim that Parkinson's disease itself is a mitochondrial energy disease, addressed separately under Fails verification (creatine entry), which concerns the disease, not the drug. Assessment: still most likely a genuine literature gap for the drug specifically at the tiers that could be verified this session, now with one specific, named lead (the Bispo 2026 title) that a future revision should check first once these channels are reachable, rather than re-running the same blocked queries.