Claire D. is forty-two years old. She manages an agency, works long hours, still responds to messages late in the evening, and has maintained for several years a professional routine that she herself describes as demanding but under control. She does not seek help because something sudden has happened. She seeks help because a function she once took for granted seems gradually to be slipping away: she is no longer recovering.

At first, the change seems ordinary. A weekend is no longer quite enough to recover from the week. Then one short night takes two days to recover from. A workout that used to make her feel better now leaves her sore for much longer. Yet she continues to work, travel, exercise and maintain the same overall level of activity.

That is precisely what makes the situation difficult to interpret.

Claire has not collapsed. She is functioning.

But functioning and recovering are not the same thing.

“I can still do everything. The problem is that everything costs me more.”

The investigations available do not identify a major acute inflammatory syndrome capable, on its own, of explaining her condition.

A recent CRP measurement is 1.6 mg/L. Other biological and clinical parameters have changed over time without, when considered individually, producing a sufficiently specific signal to explain the fatigue, slower recovery and progressive reduction in her tolerance of ordinary demands.

At this stage, the mistake would be to make the CRP say something it does not say.

CRP is neither a global measurement of “inflammation in the body,” nor an independent marker of allostatic load, nor proof that an organism is or is not “under stress.”

It provides information about a particular inflammatory marker, in a particular context, at a particular point in time.

A reassuring value can therefore genuinely be reassuring for the question it measures, without necessarily resolving the clinical question that brought the patient to seek help.

When results reveal no major abnormality, continuing surveillance is often an entirely coherent decision.

The reasoning becomes incomplete only when the absence of abnormality in each individual measurement gradually becomes equated with the absence of change in the system as a whole.

In Claire’s case, the successive results do not tell the story of a silent acute inflammatory process that a laboratory somehow failed to detect.

They raise a different question.

Over several years, certain parameters have changed. Her recovery has changed. Her sleep has become shorter and more fragmented. Her tolerance of exertion has changed. Certain cardiovascular or metabolic measurements have shifted as well, without any one of them being sufficient, in isolation, to establish a diagnosis.

The question is therefore no longer to search at all costs for the abnormal value.

It is to determine whether several changes, until now interpreted separately, may or may not describe a common trajectory.

A biological value is generally interpreted against a reference range, a diagnostic threshold or a particular clinical context.

That interpretation is essential.

But it does not exhaust the information available when a patient has several years of comparable measurements.

A current value can then be compared with its previous level, the rate at which it has changed and simultaneous changes in other dimensions: blood pressure, glucose metabolism, weight, sleep, activity, medications, symptoms and capacity for recovery.

This approach does not turn a normal value into a pathological one.

It adds another dimension to its interpretation: time.

This is where the concept of allostasis becomes useful, provided it is not turned into a diagnosis.

Homeostasis classically refers to the maintenance of certain variables within ranges compatible with physiological functioning. Allostasis emphasises that this stability continuously requires adjustments across autonomic, neuroendocrine, immune, cardiovascular and metabolic systems.

A demand appears. The system responds. When that demand disappears, the response must also be capable of being modulated.

The theoretical problem arises when these adjustments are recruited repeatedly, for prolonged periods, or without sufficient resolution.

Research on allostatic load has attempted to characterise this accumulation across multiple biological systems. This is precisely why allostatic load cannot be reduced to a single cortisol measurement, CRP value or blood pressure reading: research models use combinations of neuroendocrine, immune, metabolic and cardiovascular biomarkers.

The purpose of the model is therefore not to declare: “Claire has a high allostatic load.”

The data presented here would not support that conclusion.

The relevant question is whether a multisystem interpretation explains her trajectory more effectively than a succession of results considered independently.

Claire frequently sleeps only four to five hours and remains exposed to professional and digital demands until late in the evening.

Here again, it would be excessive to conclude that insufficient sleep explains her entire clinical picture.

But it would be equally difficult to treat it as a minor detail.

Human data associate sleep disturbances with changes across several systems involved in stress responses and immune regulation. A meta-analysis involving more than 50,000 participants notably found an association between sleep disturbance and higher concentrations of CRP and IL-6.

Experimental findings are more nuanced, which matters: not every form of sleep restriction consistently produces the same inflammatory response. More recent meta-analytic evidence nevertheless suggests that repeated partial sleep restriction over several nights may increase certain inflammatory markers.

The reasonable conclusion is therefore not: “Claire is tired because she is inflamed.”

It is more demanding:

chronically insufficient sleep belongs within the system being analysed, not at the margins of the case.

What is progressively changing in Claire is not simply her level of fatigue.

It is the relationship between a demand and the time required to recover from it.

A demanding day now requires an entire evening to recover. A poor night affects the following day. Physical exertion leaves symptoms for longer than it once did. She continues to meet expectations, but the reserve that previously allowed her to absorb them is diminishing.

This information does not appear on any standard laboratory test.

It appears in the chronology.

It requires us to distinguish observed performance from the capacity for recovery that makes that performance possible.

“I’m still managing. But I have no reserve left.”

A few months later, unusual hip pain leads to a medical evaluation.

From that point onward, the bone-related finding cannot simply be absorbed into a general narrative about “stress” or “inflammation.”

Bone fragility or a stress injury requires its own diagnostic assessment: mechanical load, energy availability, hormonal status, vitamin D, medications, medical history, bone density and quality, endocrine, nutritional or other secondary causes must be considered according to the clinical context.

The literature clearly establishes that excess glucocorticoids can affect bone remodelling and bone strength. The evidence is particularly robust for prolonged pharmacological exposure and states of endogenous hypercortisolism.

But this literature does not justify the following shortcut: occupational stress → mildly elevated cortisol → bone loss → injury.

That would turn physiological plausibility into individual causality.

The new finding therefore does not confirm our hypothesis. It requires us to broaden the analysis and investigate what might explain the trajectory.

This is probably the most important distinction in the case.

It is established that Claire reports progressively less effective recovery and insufficient sleep. Changes in several parameters over time can be documented.

It is scientifically plausible that prolonged sleep disturbance and repeated demands may interact with neuroendocrine, autonomic, immune and metabolic systems.

What has not been demonstrated, on the basis of the data presented here, is that Claire has a chronic inflammatory syndrome, a specific dysregulation of the hypothalamic-pituitary-adrenal axis, or a measurable allostatic load responsible for all of her symptoms.

Separating evidence from hypothesis does not weaken the analysis.

It is precisely what prevents us from inventing a disease to explain a situation that still needs to be understood.

Thresholds are not the problem.

They are essential.

The problem arises when we ask them to answer a question they were not designed to answer.

CRP addresses a specific inflammatory question. Blood glucose addresses a metabolic question. Blood pressure describes a cardiovascular parameter. A cortisol measurement must itself be interpreted according to the time of day, sampling conditions, clinical indication and the question being investigated.

None of these results, on its own, constitutes a general measure of the level of physiological strain on an organism.

But when several systems change during the same period, medicine has access to an additional source of information: their temporal relationship.

It is this relationship that a purely cross-sectional reading can leave invisible.

From this point onward, the purpose of the analysis is no longer to accumulate additional investigations indiscriminately.

It begins by reconstructing what already exists.

When did her sleep change? Did her recovery deteriorate before or after the increase in professional demands? Do the metabolic changes precede the symptoms or follow them? Do certain medications, hormonal changes, infections, weight variations or changes in physical activity coincide with a shift in the trajectory?

The bone findings are treated as a specific medical question, rather than as automatic confirmation of a general theory.

Inflammatory markers are returned to their proper context. Endocrine measurements are interpreted according to their sampling conditions. Symptoms are examined in relation to sleep, exertion, treatments and events within the chronology.

This work does not necessarily produce a single explanation.

It allows overly simple explanations to be discarded and helps determine what genuinely needs to be investigated.

This is also where the earlier reasoning needs to be corrected.

It would be too simplistic to offer Claire three universal interventions — slow down, sleep better, eat better — and then interpret any subsequent improvement as proof that the presumed mechanism had been correct.

Improvement following an intervention is not, by itself, sufficient to demonstrate the mechanism that produced the symptoms.

Some findings may nevertheless justify action independently of proving that mechanism.

Chronically sleeping only four to five hours deserves reassessment. Recovery that has become unusually prolonged deserves to be characterised. A change in bone health requires appropriate medical investigation. An unfavourable metabolic or cardiovascular trajectory should be assessed according to the relevant clinical recommendations.

The objective is not to artificially “reset” the body.

It is to reduce modifiable constraints when they are identified, while continuing to investigate medical mechanisms that must not be overlooked.

Claire’s case does not demonstrate that hidden inflammation can be diagnosed behind normal results.

It shows something more useful.

A normal measurement and a reassuring trajectory are not synonymous.

A value can remain within its reference range while other dimensions of physiological functioning are changing. Conversely, changes in several parameters are not sufficient to prove that they all share the same cause.

Between these two errors — declaring that “everything is fine” because every result remains acceptable, or inventing a hidden pathology because several values are changing — lies a far more demanding clinical space.

That space is the reconstruction of the case.

What is changing?

In what order?

Since when?

Which systems are changing together?

And which relationships are demonstrated, merely plausible, or still entirely unknown?

The most informative finding in Claire’s case is ultimately not a CRP value.

It is the discordance between two observations: investigations that, considered separately, remain largely reassuring, and a capacity for recovery that is progressively deteriorating.

This discordance does not constitute a diagnosis.

It constitutes a reason not to stop the reasoning process.

The role of an in-depth clinical analysis is therefore neither to challenge the investigations nor to turn normality into disease.

It is to determine whether the available information has genuinely been examined as the elements of a single trajectory.

Sometimes, what is missing from a medical record is not another piece of data.

It is the relationship between the data that already exist.