GH replacement and diabetes: reading the registry evidence

The registry question is about risk during treatment

Adults receiving GH replacement have been followed in large registries to examine diabetes risk. The reports do not all reach the same comparison with the general population. They are useful evidence for monitoring, but they cannot isolate a causal treatment effect as confidently as a randomized comparison, and they do not directly test compounded sermorelin. ^{[1] [2]}

This is a more useful starting point than declaring the entire hormone pathway either metabolically protective or inevitably harmful. A patient’s baseline risk and the reason for replacement matter. So do the investigators’ definitions, comparison population and duration of observation. The study name alone does not explain what its risk number means.

KIMS followed more than five thousand adults

The KIMS analysis included 5,143 adults with GH deficiency and 20,106 patient-years of observation, averaging 3.9 years. There were 523 new diabetes cases. The reported incidence was 2.6 per 100 patient-years, and observed-to-expected ratios were elevated against the selected population references. ^[1]

That incidence rate is not a statement that 2.6% of every patient group will develop diabetes in exactly one year. Patient-years account for differing follow-up time. It is also not the risk for an individual considering a different drug. A useful interpretation keeps the study population and the time denominator attached to the number instead of carrying a percentage alone into a sales consultation.

People who developed diabetes started with more risk

In KIMS, participants who developed diabetes were older and had higher body mass index, waist circumference, triglycerides and blood pressure, with lower HDL cholesterol. ^[1] Those differences make it difficult to attribute every subsequent diagnosis to GH rather than the underlying risk profile and clinical circumstances.

This is the central limitation of comparing a treated registry with an external reference population. Even when a statistical adjustment is appropriate, the groups can differ in ways the data do not fully capture. Our inference is that the signal deserves a monitoring response, while the size of a direct treatment effect remains a separate question. Association should inform the discussion without being relabeled as proof of causation.

For the measurement question behind this distinction, read sermorelin, glucose and insulin evidence.

HypoCCS reported a different overall interpretation

The HypoCCS analysis assessed 2,922 US and 3,709 European GH-treated patients. Overall diabetes incidence was 9.7 per 1,000 patient-years. Incidence varied by region and body mass index; the authors concluded that their analysis showed no evidence of increased diabetes incidence, while noting greater obesity prevalence among those more prone to diabetes. ^[2]

The units already differ from the KIMS headline. Ten events per thousand patient-years corresponds to one per hundred patient-years, not ten per hundred. That is arithmetic, not a clinical reconciliation. The underlying cohorts and methods still differ. Converting the units helps prevent an exaggerated apparent disagreement before the reader begins to examine the actual comparisons.

The reference population changes the story

HypoCCS reported US incidence after age, sex and BMI adjustment of 10.6 per 1,000 patient-years, compared with 7.1 in its US reference survey; the confidence intervals overlapped. French and German comparisons were similar to their references, while the Swedish point estimate was higher with overlapping intervals. GH dose did not correlate with incidence in that analysis. ^[2]

Those details explain why quoting a single country’s point estimate can produce a different impression from the authors’ overall conclusion. Overlapping intervals do not prove the risks are identical. They do indicate uncertainty in the estimates. A review should preserve that uncertainty rather than choosing the most reassuring or most alarming regional number to support a predetermined conclusion.

Clinical guidance still calls for individualized care

The Endocrine Society’s adult GH deficiency guidance emphasizes confirming the diagnosis, individualizing treatment and weighing benefits and risks. It also notes that treatment decisions require clinical judgment rather than applying a single laboratory threshold indiscriminately. ^[3] Registry findings belong within that framework.

For a person with established diabetes or several metabolic risk factors, the useful question is who will review glucose results and coordinate care. A provider should explain the plan in ordinary terms, including what is included in the recurring fee and what requires another consultation. Buying a hormone panel without that responsibility being clear does not itself create a safer or more informative follow-up arrangement.

Neither registry supplies a sermorelin risk percentage

These were reports about GH replacement in people with diagnosed deficiency. ^{[1] [2]} They did not randomize otherwise healthy adults to a current sermorelin service, compare injection with an oral blend or test a longevity subscription. A provider should not present them as a direct numerical estimate of any of those arrangements.

The distinction cuts both ways. The registries cannot prove a sermorelin program metabolically harmless, and they cannot quantify that program’s risk simply because the medicines interact with related signaling. Ask for evidence relevant to the actual medicine and clinical purpose. Where that evidence is limited, the honest response is to make the uncertainty visible rather than importing a convenient percentage.

How to use this evidence in a purchasing decision

Our assessment is that a credible service gives more than a claim about body composition. It establishes the treatment purpose, reviews relevant baseline risk, describes follow-up and makes coordination with existing care practical. Those features do not guarantee an outcome. They make the agreement easier to evaluate and help the patient understand who is responsible for the next decision.

The research conclusion remains measured: large registries provide useful information about diabetes occurring during GH treatment, with different external comparisons and substantial patient-level risk differences. They support attention to metabolic monitoring. They should neither be reduced to a universal scare statistic nor used as a blanket reassurance for a different medicine sold under a broad hormone-optimization label.

For a separate assessment of purchasing terms, see our Strut Health review; the service review does not establish a treatment benefit.