The Translational Gap in Metabolic Disease

At 50 years of age, a female patient is living with rising blood sugar, hypertension on two drugs, aching knees, poor sleep, and liver enzymes quietly flagged but never pursued. Over the next few months, she may move between primary care, cardiology, hepatology, and sleep medicine — each visit focused on one organ, each note clinically appropriate, none quite capturing the metabolic story her body is telling.

That fragmentation is not an accident. It is the legacy of a century of specialization. Diabetes went to endocrinology. Obesity moved between primary care, obesity medicine, and bariatrics. Heart failure belonged to cardiology. Fatty liver disease was claimed by hepatology. Kidney disease sat in nephrology. Sleep apnea found a home in pulmonary medicine.

And specialization worked. It gave medicine sharper tools, cleaner trials, safer labels, deeper knowledge, and a structure for protecting patients. Specialists needed categories. Guidelines needed boundaries. Regulators needed clear endpoints. Clinical trials needed primary outcomes that could be measured, interpreted, and defended.

But the deeper that knowledge became, the more clearly it revealed the connections our systems were not built to hold.

Specialization worked — until biology exposed its limits

Clinical trials were built the same way the specialties were. Each chose a primary endpoint belonging to a single field: A1c for diabetes, ejection fraction for heart failure, biopsy for liver disease, creatinine for the kidney. The very design of evidence assumed that a drug would do one thing, in one organ, for one specialty's patients.

The trial was a lens pointed at a single organ system, deliberately excluding the rest.

But the body never agreed. A metabolic signal may begin in the gut, pancreatic islet, or adipose tissue, but it rarely stays there. It reaches the liver, skeletal muscle, kidney, blood vessel wall, and appetite centers of the brain, where it can influence hunger, satiety, reward, and behavior. An inflammatory cascade that begins in the liver can reshape vascular, renal, and systemic risk. A drug that changes how the kidney handles glucose can shift the trajectory of heart failure — even when the original trial was not built to see that outcome.

The body does not organize disease by specialty. It organizes disease through connected physiology.

For decades, this was true in principle. What has changed is that the clinical evidence has made it impossible to treat as background knowledge.

The evidence is crossing the boundaries

SELECT was a cardiovascular outcomes trial of semaglutide in people with overweight or obesity and established cardiovascular disease, without diabetes. The primary endpoint belonged to cardiology. The drug cut major adverse cardiovascular events by roughly 20 percent. But to get there, the trial had to enroll patients who would once have been seen mainly through an obesity or primary-care lens. The line between metabolic and cardiovascular disease had already blurred in the enrollment criteria. It dissolved further in the result.

FLOW was a kidney outcomes trial of semaglutide in people with type 2 diabetes and chronic kidney disease. Major kidney disease events fell by about 24 percent. A nephrology trial, paid for because of diabetes, in a drug first known for weight loss.

The same pattern repeats across fatty liver disease, sleep apnea, and broader cardiometabolic risk. The trials were designed with a specialty in mind. The results answered across organs.

The evidence is no longer subtle. Metabolic disease is not five unrelated diseases. It is a connected biology expressed in different tissues, noticed by different specialists, captured in different billing codes, and studied through endpoints that were often not designed to see the whole.

The new translational gap

This is the new translational gap in metabolic disease.

It is no longer only the familiar gap from bench to bedside. In metabolic disease — and likely across medicine — the harder gap is between compartmentalized proof and integrated biology: between the way we organize evidence and the way biology crosses organs, systems, and specialties.

The irony is that we now have more ways than ever to measure the body — genomics, biomarkers, imaging, continuous glucose data, digital signals, and real-world evidence — but far less agreement on how to assemble those measurements into one coherent picture of disease.

At first, the challenge was to find the biology. To identify the receptor, the pathway, the signal. That phase was hard, and it produced decades of work: the incretin system, amylin, glucagon, adipose-tissue hormones, inflammatory cytokines, renal glucose transporters, genetic variants that raised or lowered risk decades before symptoms appeared.

But now the biology is no longer scarce. Every conference, every journal, every pitch deck surfaces another intriguing mechanism, another elegant pathway, another signal that might — under the right conditions — become something important.

The hard part is no longer finding the biology. The hard part is deciding what it means — and more importantly, deciding what to bring together, and how.

For a century, medicine earned clarity by dividing: one organ, one specialty, one endpoint, one label. Now the evidence is pushing in the opposite direction. The question is no longer just "Which mechanism?" but "Which mechanisms belong together?" No longer just "Which endpoint?" but "Which endpoints, measured across which organs, tell a coherent story about this patient's trajectory?" No longer "Which trial design satisfies one endpoint?" but "Which evidence architecture can span cardiology, nephrology, hepatology, and metabolism without forcing each field to start from scratch?"

That is the shift the field has not yet fully absorbed.

SGLT2 inhibitors: the gap in miniature

SGLT2 inhibitors show this gap in miniature.

They were designed to lower glucose. The primary endpoint was A1c, for a diabetes indication, for a diabetes specialty. But as the class moved into cardiovascular-outcomes testing, signals appeared beyond the glucose story. EMPA-REG OUTCOME showed a striking reduction in heart failure hospitalization and cardiovascular death. CANVAS and DECLARE-TIMI 58 extended the pattern. Renal signals emerged as well.

The biology was already pointing across organs. Clinicians began to recognize that the story was larger than glucose. Yet cardiovascular and renal labels required prospective trials designed around those outcomes, not secondary signals from trials built around diabetes risk and glucose control.

The system demanded new trials, with cardiovascular and renal outcomes as primary endpoints, to prove what the earlier data had begun to suggest. So the outcome trials followed — DAPA-HF for heart failure, CREDENCE and DAPA-CKD for kidney disease — and they confirmed what the earlier signals had whispered.

The process had an almost absurd quality: not because regulators were wrong to require prospective evidence, but because the evidence architecture had to ask separately, in cardiology and nephrology, what the biology had already begun to reveal across organs.

That is the translational gap in miniature: biology had moved through the body, but proof still had to move one specialty, one endpoint, and one label at a time.

What history teaches

The gap is not new. It has shaped every major chapter in metabolic medicine, though it is rarely discussed in these terms.

The incretin story began long before GLP-1 drugs became household names. GLP-1 biology was elegant: a gut hormone that stimulated insulin secretion in a glucose-dependent way. But native GLP-1 disappeared from the bloodstream within minutes. The breakthrough was not simply a deeper understanding of the receptor. It was solving the practical problem that made the biology usable.

PCSK9 inhibitors moved differently. Human genetics gave the field unusually clean confidence that the target mattered: people with loss-of-function mutations had lower LDL cholesterol and fewer heart attacks. The target was not in doubt. The question was execution.

MASH is the cautionary tale. The biology was recognized for years, but drug development stalled because the disease was heterogeneous, progression was slow, and the endpoints were difficult. In that field, figuring out how to measure success became almost as important as the mechanism itself.

A common thread runs through these examples: discovery is necessary, but translation — deciding what to test, in whom, with what endpoint, and how to read the results — is where value is created or destroyed. And in metabolic disease, where the biology crosses boundaries by nature, translation increasingly means integration.

The quiet questions that shape strategy

In a field this crowded, the obvious things are already obvious. Diabetes is large. Obesity is large. Cardiometabolic disease is large. Pointing to market size no longer says much.

The serious question is whether the development strategy can weave evidence across the organ systems the biology actually engages — or whether it will be forced to chase separate labels, one specialty at a time.

Consider early intervention. In SURMOUNT-1, tirzepatide was associated with roughly 93 percent fewer new type 2 diabetes diagnoses during ongoing treatment. That number stops a room. But it is not the same as cure. Diagnoses rose after the drug was stopped. The way a company frames that evidence — whether it implies permanence or is honest about treatment dependence — will shape the credibility of the entire program.

Durability is another open question. Many metabolic benefits depend on continued treatment. The field still lacks rigorous evidence on lower-dose maintenance strategies.

Biomarkers remain a persistent gap. We still cannot reliably predict who will respond best, who will regain weight, who will benefit beyond weight loss, or who should receive a different therapy altogether. That shapes trial design, patient selection, and payer negotiations — and underscores how far we are from an integrated patient-level model of metabolic disease.

Endpoints are changing, too. Weight and A1c still matter, but cardiovascular outcomes, kidney outcomes, liver histology, sleep apnea metrics, physical function, and durability increasingly determine whether a therapy changes practice — or just adds another option to a crowded shelf.

And then there is positioning. In a crowded field, it is easy to build a story around a large market and a plausible mechanism. It is much harder to build a story that survives scrutiny — especially one that claims cross-organ benefit without the cross-organ evidence architecture to back it. That is where many programs become fragile.

What different audiences need

A founder needs to know whether their science can support a credible development plan before the wrong trial design turns a promising mechanism into a failed asset — and which organ systems their biology genuinely connects.

A large company needs to know which assets genuinely change the strategic picture: which ones force a competitor to react, create a new market segment, or unlock a combination strategy that was not possible before.

An investor needs to know whether the biology, the evidence, and the differentiation are strong enough to justify serious attention — or whether this is another well-told story that will collapse under due diligence.

A translational team needs to know which measurements are credible, feasible, and connected to a real decision — and which measurements, taken together, begin to tell a connected story rather than a collection of fragments.

Across these groups, the need is the same: a clear read on what matters, what remains uncertain, and what should happen next. Is the effect direct, or mostly weight-mediated? Is the endpoint clinically meaningful? Does the development path plan for integration across organs from the start? Is the company telling a genuinely differentiated story, or a familiar one dressed in cross-organ language it has not earned?

The discipline of knowing what not to claim

Overclaiming is seductive. An on-drug benefit is described as disease modification. A positive trial is treated as an approved indication. Cross-organ signals are presented as an integrated story before the evidence architecture exists to support it.

But sophisticated partners notice the gap. Regulators notice. Payers notice. Clinicians notice. Over time, credibility becomes part of the strategy. It separates the programs taken seriously from those that are merely well-marketed.

A credible metabolic-disease strategy is disciplined: on-drug benefit is not automatically durable disease modification. Trial evidence is not the same as a label. Cross-organ signals are not integrated evidence until the trials are designed to capture them together. Mechanistic plausibility is not enough.

Knowing what not to claim is how a strategy stays credible long enough to matter.

Where the work is heading

The next phase of metabolic medicine will still depend on discovery. But discovery alone will not be enough. The field will need teams that can do the harder, quieter work: understanding what the biology means, where it should be tested, how it should be measured, how the evidence can be assembled across the artificial boundaries we inherited, and how honestly it can be positioned.

That is the real work of translation. And in metabolic disease, the work of integration — across organs, across trials, across labels — is now central.