https://www.sciencedirect.com/science/article/abs/pii/S1568163726002965
This is behind a paywall, but one of the authors kindly allowed me access to do a chatGPT(5.6paid) analysis:
Overall assessment
This is an unusually useful conceptual review, but it does not resolve whether ageing is damage-driven or programmed. Its principal contribution is to expose, in a structured adversarial format, what each paradigm can and cannot presently explain.
The strongest conclusion is not that either side wins. It is that the dichotomy is partly false:
Ageing can be initiated by accumulating damage, amplified by regulated responses to that damage, and paced by evolved resource-allocation and repair systems—without requiring evolution to have selected ageing itself as an adaptation.
The programmed-ageing case presented by Peter Lidsky is intellectually provocative, especially the pathogen-control hypothesis, but remains much more speculative than the damage framework. Mikolaj Ogrodnik’s damage model is better grounded empirically, although it is so broad that it risks becoming difficult to falsify.
1. What kind of paper is this?
Lidsky and Ogrodnik, Aging’s Competing Paradigms: Another Decade of Investigation into Damage-Driven versus Programmed Theories, is an accepted 2026 review in Ageing Research Reviews.
It is not a systematic review, meta-analysis or primary experimental paper. It is a staged debate:
- Ogrodnik defends damage-driven ageing.
- Lidsky defends programmed, adaptively evolved ageing.
- They answer 14 questions covering cross-species longevity, cellular senescence, regeneration, ageing clocks, ecological influences, non-linear ageing, semelparity, long-lived proteins and falsifiability.
- Each author then directly criticises the opposing framework.
That format makes the paper valuable for clarifying assumptions, but it also means that evidence selection is partly advocative rather than systematic.
2. Summary
A. The damage-driven model
Ogrodnik argues that ageing ultimately results from progressive loss of function caused by molecular and structural damage.
The relationship between damage and ageing is not assumed to be simple. It is modified by at least five factors:
-
Rate of damage production
This includes reactive chemistry, metabolic by-products, replication and translation errors, environmental exposure and other insults. -
Repair or removal
DNA repair, proteostasis, autophagy, mitophagy, immune clearance and tissue turnover determine whether damage persists. -
Damage resistance
Some proteins, membranes and other structures tolerate chemical modification better than others. -
Redundancy
Organs and tissues can continue functioning despite the loss or impairment of some components. -
Replacement and regeneration
Organisms such as Hydra and planarians can replace damaged cells or structures sufficiently rapidly to avoid conventional senescence.
Ageing therefore need not track the amount of damage linearly. A tissue may function normally until redundancy is exhausted, after which a modest additional burden produces rapid deterioration.
Many apparently programmed features of ageing are interpreted as adaptive responses to damage. Examples include:
- cellular senescence;
- inflammation;
- myeloid skewing of haematopoiesis;
- mobilisation and depletion of hair-follicle stem cells;
- changes in epigenetic clocks;
- shifts in metabolism and tissue composition.
The analogy is starvation: starvation produces a highly ordered sequence of metabolic responses, but that does not mean the organism contains a programme whose selected function is to kill it. The responses are principally attempts to survive an underlying deficit.
The damage framework also explains the ageing of structures with little or no active gene regulation, such as:
- the lens;
- extracellular matrix;
- cartilage;
- long-lived nuclear-pore proteins;
- other extremely long-lived proteins and RNAs.
Ogrodnik’s proposed tests include simultaneously improving several damage-management systems, transferring resilience mechanisms from long-lived species to short-lived ones, and selectively preventing particular forms of damage in lifespan-limiting tissues.
B. The programmed-ageing model
Lidsky argues that ageing is a genetically regulated process that evolved to limit lifespan.
His main observations are:
- ageing is reasonably reproducible within a species;
- lifespan varies enormously between species;
- ageing can be accelerated, slowed or sometimes reversed;
- biological age is reset during embryogenesis;
- negligible senescence exists;
- eusocial castes with closely related genomes can have radically different lifespans;
- semelparous species execute coordinated post-reproductive death;
- ageing-related biochemical transitions appear organised and non-linear;
- some animals deploy rejuvenation or diapause only under particular environmental conditions.
He argues that these findings are difficult to reconcile with ageing as an unavoidable, uncoordinated accumulation of damage.
Three adaptive-programme theories
The paper distinguishes three families of programmed-ageing theory:
-
Overpopulation control
Ageing prevents populations from exhausting their resources. -
Acceleration of evolution
Shorter lifespans increase generational turnover and adaptability. -
Pathogen-control hypothesis — PCH
Ageing evolved to prevent old individuals carrying chronic infections from transmitting them to genetically related individuals.
The authors regard the first two as having serious weaknesses. Lidsky concentrates on PCH.
C. The pathogen-control hypothesis
PCH proposes that ageing is an altruistic, kin-selected defence against infection.
Its essential scenario is:
- a chronic pathogen reduces or eliminates the host’s reproductive capacity;
- the infected host nevertheless survives and can transmit infection;
- local population structure causes the host to interact disproportionately with relatives;
- continued host survival therefore reduces inclusive fitness;
- mechanisms that remove older, infection-carrying individuals can consequently be selected.
Under this model, the ageing immune system is not merely failing. It undergoes an evolved functional shift from protecting the individual to protecting relatives or the population from pathogens.
PCH also invokes “security harnesses” to explain why simple non-ageing mutants do not arise. The idea is that pathogens would benefit by inhibiting host ageing and extending their infectious period, so ageing would evolve redundant, pathogen-resistant control mechanisms.
The proposed predictions include:
- lifespan should be related to the kin structure of pathogen-transmission networks;
- pathogens infecting relatives should be more closely related in short-lived than long-lived species;
- dispersing species such as birds and bats should experience weaker selection for programmed ageing;
- eusocial caste longevity should reflect differences in pathogen transmission and kin protection;
- naked mole-rats should be unusually sensitive to infectious disease, using rapid infection-induced death rather than ageing to protect the colony;
- ageing should involve specific, detrimental, regulated changes in immune function.
D. Cellular senescence
The disagreement here illustrates a central definitional problem.
Ogrodnik argues that:
- genotoxic damage efficiently induces senescence;
- senescence is often a protective response to tissue damage;
- impaired immune clearance allows senescent cells to accumulate;
- senescent cells contribute to pathology;
- current senolytic interventions improve health or median lifespan more reliably than maximum lifespan;
- this suggests senescence is an amplifier or secondary consequence, rather than the deepest cause of ageing.
Lidsky responds that senescence is clearly a cellular programme, comparable to apoptosis. He connects it to antiviral immunity: infected non-immune cells can enter senescence and generate a local inflammatory state. If these cells are not cleared, chronic inflammation and ageing could follow.
The two are partly using “programme” differently. Ogrodnik denies that organismal ageing is an evolved death programme; he does not deny that cells execute regulated responses. Lidsky sometimes moves from “senescence is a regulated cellular programme” to “organismal ageing is programmed,” but the former does not establish the latter.
E. Declining maintenance
Ogrodnik proposes a damage–repair-failure–damage loop:
[
\text{damage} \rightarrow \text{impaired repair} \rightarrow \text{more damage}
]
Repair enzymes, immune cells, metabolic substrates and regulatory systems can themselves become damaged. This could progressively weaken maintenance and create non-linear deterioration.
Lidsky counters that earlier error-catastrophe models, particularly Orgel’s translational-error theory, were not supported. He instead proposes:
[
\text{ageing programme} \rightarrow \text{repair suppression} \rightarrow \text{damage}
]
Examples offered include age-related mTORC1 activity and DREAM-complex repression of DNA-repair genes. Partial reprogramming is also presented as evidence that age-associated maintenance failure is regulatory and reversible rather than irreparable.
F. Ageing clocks and rejuvenation
Ogrodnik interprets epigenetic clocks mainly as measures of damage responses or altered tissue state, rather than direct measurements of accumulated molecular lesions. Acute stress, injury, inflammation, senescence and DNA-repair deficiencies can therefore alter clock readings.
Lidsky regards their orderly progression as compatible with an ageing programme, but acknowledges that reversible clock movement does not decisively distinguish the theories. The same methylation marker might be:
- persistently activated during ageing; and
- transiently activated during an immune response.
Thus, an apparent reversal of biological age may represent removal of a reversible state rather than actual reversal of all age-related damage.
G. Non-linear ageing
The damage model explains non-linearity through:
- thresholds;
- loss of redundancy;
- positive feedback;
- compensatory responses;
- organ-specific tipping points.
The programmed model interprets distinct age transitions as scheduled changes in life-history function—for example, reproductive, parental, grandparental or social-role transitions followed by terminal frailty.
The paper cites multi-omic evidence for discrete transitions, but neither interpretation is uniquely entailed by such transitions.
H. Semelparity
Lidsky argues that programmed post-reproductive death in salmon, octopuses and some marsupials proves that an evolved death programme is biologically possible.
Ogrodnik agrees that semelparous death can be programmed but stresses that it differs from ordinary ageing:
- it is rapid;
- it has a relatively identifiable hormonal trigger;
- it involves coordinated resource consumption and tissue breakdown;
- it is directly linked to reproduction;
- it is mechanistically much clearer than slow, heterogeneous ageing.
Semelparity therefore establishes possibility, not generality.
3. What is novel?
1. The adversarial format
The clearest novelty is presenting both paradigms in one paper, with advocates answering the same questions and directly responding to one another.
This exposes disagreements that ordinary reviews frequently conceal, particularly differences in:
- definitions;
- standards of evidence;
- views of falsifiability;
- distinctions between primary causes and secondary adaptations.
2. Integration of recent findings into an old debate
The paper updates the damage-versus-programme debate using:
- epigenetic clocks;
- partial reprogramming;
- ageing trajectories with discrete transitions;
- reversible caste-associated longevity;
- negligible senescence and rejuvenation;
- modern senescence biology;
- recent immunological findings;
- cross-species comparative biology.
These observations individually are not new, but bringing them together as tests of competing paradigms is useful.
3. Elevation of PCH into an explicit ageing framework
The most substantively novel element is the extended presentation of the pathogen-control hypothesis as a theory of adaptive ageing.
In particular, PCH attempts to overcome two classic objections:
- Group-selection problem: replaced by kin selection and inclusive fitness.
- Absence of immortal mutants: explained through redundant “security harnesses” that prevent pathogens from disabling host ageing.
The claim that ageing forms part of an immune strategy which increasingly protects kin rather than the ageing individual is distinctive.
4. Specific ecological predictions
The proposed comparison between:
- relatedness of interacting hosts;
- genetic relatedness of their pathogens; and
- species lifespan
is a relatively concrete and interesting comparative prediction.
The prediction that naked mole-rats should exhibit unusual infection sensitivity is also memorable and experimentally approachable, although it is not necessarily unique to PCH.
5. A stronger formulation of the damage model
Ogrodnik usefully expands “damage accumulation” beyond oxidative damage. The proposed explanatory variables—production, resistance, repair, redundancy and replacement—form a more sophisticated framework than the old free-radical theory.
This is conceptual clarification rather than an entirely new theory.
4. Critique
A. The paper constructs too sharp a dichotomy
The main weakness is that “damage” and “programme” operate at different explanatory levels.
A complete account of ageing may include:
flowchart TD
A["Evolutionary trade-offs"] --> B["Maintenance and repair settings"]
B --> C["Molecular and organelle damage"]
C --> D["Regulated stress responses"]
D --> E["Inflammation, senescence and remodelling"]
E --> C
E --> F["Functional decline"]
Evolution can programme repair investment, metabolic rate, reproductive allocation and stress responses. These settings can then permit damage to accumulate. Damage subsequently activates programmes such as senescence, inflammation, fibrosis and altered differentiation.
That is neither purely programmed ageing nor purely passive damage accumulation.
The decisive question is not whether programmes participate in ageing—they clearly do—but whether natural selection favoured those programmes because they shorten lifespan, rather than because they confer early-life benefits or constitute imperfect responses to damage.
B. “Programmed process” is conflated with “adaptively programmed ageing”
This is particularly evident in the discussion of cellular senescence.
Senescence is genetically regulated, but that does not show that its evolutionary function is to age or kill the organism. Fever, clotting, fibrosis and apoptosis are also programmed responses. Their late-life pathology does not mean that their selected purpose is organismal ageing.
For programmed adaptive ageing to be established, the authors would need evidence that:
- the mechanism reliably limits lifespan;
- its lifespan-limiting effect is not merely a pleiotropic side effect;
- disabling it improves late-life survival without destroying a more immediate protective function;
- the responsible alleles spread because shortening lifespan increased inclusive fitness.
The paper supplies little evidence at this level.
C. PCH currently lacks foundational comparative evidence
Ogrodnik’s criticism is persuasive here. PCH predicts relations among:
- chronic sterilising infection;
- kin-structured transmission;
- population viscosity;
- immune architecture;
- evolved lifespan.
But the paper does not present a comparative dataset demonstrating these relationships.
Before PCH can be treated as a leading theory, it would need phylogenetically controlled evidence showing that its ecological variables predict lifespan after adjustment for:
- body size;
- developmental time;
- metabolic rate;
- predation;
- flight;
- reproductive strategy;
- social organisation;
- environmental temperature;
- cancer resistance;
- regenerative capacity.
Without this, bats, birds, eusocial animals and naked mole-rats can be explained retrospectively through PCH, but such explanations are vulnerable to storytelling.
D. The “security harness” risks being an ad hoc rescue
PCH explains the lack of non-ageing mutants by proposing that pathogens exert selection for redundant mechanisms preventing them from disabling host ageing.
This is possible, but it is presently an auxiliary hypothesis introduced to explain missing evidence. It needs independent predictions—for example:
- multiple ageing-effectors with pathogen-resistant redundancy;
- evidence that pathogens actively target these effectors;
- signatures of host–pathogen arms races in the relevant genes;
- experimental pathogen evolution towards suppression of host-ageing mechanisms.
Without such evidence, the security-harness idea reduces rather than increases falsifiability.
E. The programme model does not yet identify the programme
Lidsky refers to a master regulator and to detrimental immune functions, DREAM and mTORC1, but no defined organismal ageing circuit is demonstrated.
A genuine programme ordinarily has:
- inputs;
- a regulatory architecture;
- an ordered sequence of effectors;
- a selectable output;
- mutations that disrupt or alter the programme predictably.
Semelparity meets much of this description. Ordinary mammalian ageing currently does not.
The absence of a single non-ageing mutation is not fatal—development itself is distributed across many genes—but it weakens claims of a discrete death programme.
F. Age resetting does not show that adult ageing is programmed
Embryonic rejuvenation establishes that some age-associated states can be reset during reproduction. It does not demonstrate that somatic ageing is deliberately imposed.
The germline and early embryo employ:
- strong cellular selection;
- mitochondrial bottlenecks and quality control;
- dilution by rapid cell division;
- proteostatic clearance;
- extensive epigenetic reprogramming;
- elimination of defective embryos.
These mechanisms are compatible with damage management. Resetting a consequence of damage is not evidence that the original damage was programmed.
Furthermore, epigenetic-clock resetting is not equivalent to the removal of all:
- nuclear mutations;
- mtDNA mutations;
- protein crosslinks;
- extracellular-matrix damage;
- chromosomal abnormalities.
G. Rejuvenation and ageing plasticity do not discriminate cleanly
The ability to slow or reverse some ageing phenotypes demonstrates plasticity, but both frameworks predict plasticity.
A damage model can accommodate rejuvenation through:
- damage dilution;
- selective cell replacement;
- proteolysis and autophagy;
- mitochondrial turnover;
- reprogramming of maladaptive stress responses;
- regeneration from relatively protected stem cells.
A programme model can explain the same observations as switching between life-history states.
The experiment must therefore measure both regulatory state and physical damage. Clock reversal alone is insufficient.
H. The damage framework is too inclusive
Ogrodnik’s framework encompasses virtually every detrimental molecular alteration plus the organism’s response to it. That makes it empirically plausible but potentially unfalsifiable.
If an intervention fails, the explanation may be:
- the wrong damage was targeted;
- insufficient damage types were targeted;
- the critical tissue was missed;
- damage had already crossed a threshold;
- redundancy was exhausted;
- the repair intervention had pleiotropic effects.
Those may all be true, but together they protect the theory from decisive tests.
The paper would have been stronger if it defined a measurable total damage burden or specified which lesions are expected to be lifespan-limiting in particular organisms.
I. Evidence of damage causality remains uneven
That molecular damage accumulates with age is not in doubt. The harder issue is whether it is:
- primary;
- rate-limiting;
- sufficient;
- necessary;
- or partly downstream of altered regulation.
Some types of damage are strongly causal in particular pathologies. But failure of antioxidant interventions and several simplistic repair theories shows that “more damage equals faster ageing” is not generally adequate.
The damage advocate recognises this, but the broadened framework consequently becomes less predictive.
J. The critique of damage-based therapies is unfair
Lidsky argues that damage theories have failed to generate effective anti-ageing interventions. This is too strong.
Several interventions plausibly act partly through damage prevention or clearance, including:
- UV protection;
- control of hypertension and hyperglycaemia;
- enhancement of autophagy;
- removal of senescent or otherwise dysfunctional cells;
- prevention of infectious and inflammatory injury;
- protection from smoking and environmental toxins.
These do not abolish ageing, but a theory need not yield complete rejuvenation to have explanatory or therapeutic value.
Conversely, IL-11 inhibition, VEGF manipulation or partial reprogramming do not prove programmed ageing. They may improve repair, perfusion or adaptation to damage.
K. Semelparity is not a strong bridge to ordinary ageing
Semelparity proves that programmed death can evolve. It does not establish that gradual ageing has the same evolutionary origin.
The relevant comparison is not:
Can evolution ever programme death?
It is:
Does the slow, stochastic, multi-organ, individually heterogeneous deterioration observed in iteroparous mammals possess the causal and evolutionary signatures of a selected death programme?
The paper does not demonstrate that.
L. The falsifiability asymmetry is overstated
PCH generates testable correlations, but several are not exclusive to PCH.
For example:
- bats and birds may be long-lived because flight reduces extrinsic mortality;
- eusocial queens may live longer because of reproduction-associated endocrine states and protected environments;
- naked mole-rat infection sensitivity may reflect specialised immune evolution unrelated to adaptive ageing.
Confirmation of these predictions would strengthen PCH but would not necessarily falsify damage or conventional evolutionary theories.
A strong test would require a prediction whose direction differs between the competing models.
M. The review is not systematic
The article does not describe:
- search methods;
- inclusion criteria;
- evidence grading;
- systematic treatment of contrary findings;
- formal comparison of competing models.
It should therefore be read as a sophisticated position paper, not an exhaustive or neutral evaluation of the literature.
5. What experiments would genuinely discriminate the theories?
The paper’s proposals could be strengthened by several sharper tests.
1. Experimental evolution with controlled pathogen transmission
Use short-lived organisms with manipulated population structures:
- high kin interaction versus random mixing;
- chronic sterilising pathogen versus no pathogen;
- matched mortality and resource conditions.
PCH predicts evolution of shorter intrinsic lifespan specifically where persistent infection is transmitted predominantly among kin.
2. Remove the proposed ageing programme without broadly improving repair
Identify a specific late-life immune or neuroendocrine circuit and disable it after reproduction.
A programme model predicts:
- substantial extension of maximum lifespan;
- coordinated postponement of multiple ageing phenotypes;
- little initial reduction in accumulated molecular damage;
- a fitness cost expressed through increased pathogen transmission to kin.
That last element is particularly important for demonstrating adaptive function.
3. Reduce multiple forms of damage without resetting regulatory state
Simultaneously improve:
- proteostasis;
- mitochondrial quality control;
- DNA repair;
- removal of damaged cells;
- extracellular-matrix maintenance.
If this profoundly extends maximum lifespan while the putative ageing regulatory programme remains intact, the damage model gains support.
4. Track damage and regulation during genuine rejuvenation
In rejuvenating animals or partial-reprogramming models, measure:
- somatic mutations;
- mtDNA heteroplasmy and deletions;
- protein aggregates and crosslinks;
- lipofuscin;
- extracellular-matrix modifications;
- epigenetic age;
- transcriptional state;
- mitochondrial function.
If function and clocks reverse while physical damage remains, regulatory causation becomes more plausible. If damage is removed in parallel, rejuvenation supports active repair rather than a pure programme-switch interpretation.
5. Phylogenetically controlled PCH analysis
Construct a multi-species dataset of:
- maximum lifespan;
- kin structure;
- dispersal;
- chronic pathogen load;
- pathogen effects on fertility;
- host–pathogen genetic relatedness;
- immune response and infection mortality.
This is a necessary foundation before PCH can be regarded as a general theory.
6. Relevance to a mitochondrial/acetyl-CoA model of ageing
The paper does not meaningfully engage with the possibility that mitochondrial dysfunction changes nuclear regulation through metabolic signalling.
A pathway such as:
[
\text{mitochondrial dysfunction}
\rightarrow \text{reduced citrate export}
\rightarrow \text{reduced nuclear acetyl-CoA}
\rightarrow \text{altered histone acetylation}
\rightarrow \text{transcription/splicing changes}
]
cuts across the paper’s dichotomy.
It could be interpreted as:
- damage-driven: mitochondrial or mtDNA damage reduces metabolic competence;
- programmatic: cells actively reconfigure chromatin and transcription in response to metabolic state;
- hybrid: initially adaptive regulatory responses become self-reinforcing and pathogenic.
This illustrates why the most productive distinction may be:
- initiating lesions;
- sensing and signalling mechanisms;
- adaptive responses;
- maladaptive feedback loops;
- irreversible structural consequences.
Calling the whole sequence either “damage” or “programme” obscures those causal layers.
Bottom line
The paper is novel mainly as a well-constructed debate and as a prominent presentation of the pathogen-control hypothesis. It successfully shows that modern damage theories must explain plasticity, rejuvenation, species differences and coordinated regulatory changes. It also shows that programmed-ageing theories must move beyond suggestive patterns to identifiable mechanisms and comparative evolutionary evidence.
At present:
- The damage framework has substantially stronger empirical foundations.
- PCH is interesting and testable but remains speculative.
- Neither framework alone adequately explains ageing.
- A hybrid causal model—damage interacting with evolved, regulated, sometimes maladaptive responses—is better supported than either extreme.