A drug label is the regulator-approved account of what a medicine is, who it is for, and how it should be used, and its central promise is the approved indication: the specific disease, population, and conditions of use for which a regulator has concluded that the benefits outweigh the risks. The detail that makes the whole system click into place is this. That indication is not a summary of everything the molecule might do. It is a compressed description of the trials that were actually run. Once you see the label as evidence made legible rather than biology described in full, the difference between on-label and off-label stops looking bureaucratic and starts telling you something useful at the bedside.
Key points#
- The approved indication can be no broader than the studies that supported it, so the wording tracks the trial population, dose, and setting.
- On-label means prescribing within that approved use; off-label means outside it, and off-label prescribing is legal and often reasonable.
- A clinician may prescribe off-label, but a manufacturer generally may not promote off-label; the label marks the boundary of what has been proven to the regulator.
- A narrow indication is not a weak drug. It usually signals a focused development program with room to expand as more trials report.
Start with the distinction that matters at the bedside#
The clearest way into the topic is to work backward from how prescribers use the word. Once an indication exists, it draws a line. Prescribing a medicine for the approved condition, age group, and dose is on-label. Prescribing it for a different disease, a different population, or a different regimen is off-label.
Off-label use is legal in most systems and is frequently the right thing to do. A physician may have sound reasons, sometimes strong published evidence, for using an approved drug outside its labeled use, and in oncology and pediatrics this is routine. The point is not that off-label use is forbidden. It is that a regulator has not formally reviewed and endorsed that particular use, so the specific weighing of benefit against harm that a label represents has not been carried out for it. Good data may exist. They simply have not passed through the approval gate.
The asymmetry falls on the manufacturer, not the prescriber. A company generally cannot promote a drug for uses the label does not cover, because the label defines the outer edge of what has been demonstrated to the regulator's standard. This is why sponsors run additional trials to widen an indication rather than leaning on real-world habit. A new population means a new claim, which means new evidence.
The indication is a compressed description of the trials#
Here is the part that surprises people who have not sat inside drug development. A regulator does not approve a molecule in the abstract. It approves a claim, and the claim can be no wider than the studies that support it.
If the pivotal trials enrolled adults with one condition, at one dose, on one background therapy, the indication will describe adults with that condition, at that dose, in that setting. Studying only adults usually means no pediatric claim. Excluding patients with meaningful kidney impairment usually leaves the label silent on them or attaches a dosing caveat. Even the chosen primary endpoint shapes the words a regulator is willing to grant. An indication, read carefully, is a short paragraph that encodes a trial population and a study design.
That is why the wording of an indication is negotiated so closely at submission. Two teams can read the same dataset and press for a broader or a narrower sentence, and the regulator's task is to make sure the sentence does not promise more than the data can carry. The indication is often settled late, after the efficacy and safety picture is fully assembled, because it has to reflect what was shown rather than what was hoped.
Where the label lives and what it must contain#
In the United States the label is formally the Prescribing Information, sometimes called the Full Prescribing Information. It forms part of the New Drug Application or Biologics License Application, and once approved it becomes the legally binding statement of the conditions under which the medicine may be marketed. The European counterpart is the Summary of Product Characteristics. These documents follow a fixed structure on purpose, so that a clinician opening any modern label knows where to find dosing, contraindications, warnings, adverse reactions, and the data behind each.
The opening Indications and Usage section is the one clinicians read first and the one that is contested most. Under the governing US regulation it must state the approved indication concisely and may include only uses supported by substantial evidence of safety and effectiveness. That phrase, substantial evidence, is a legal standard, and the entire clinical development program exists to meet it.
Why the evidence has to be trustworthy first#
None of this holds up unless the underlying data are credible, which is the job of Good Clinical Practice, the international ethical and scientific standard for designing, running, and reporting trials. The framework was refreshed in 2025. The ICH E6(R3) guideline reached its final step on 6 January 2025, with the European Medicines Agency applying it from 23 July 2025 and the US FDA publishing it on 9 September 2025. The revision leans into quality-by-design, risk-based oversight, and modern data governance, reflecting trials that increasingly run across multiple sites and rely on connected technology. It sits alongside ICH E8 on general trial considerations and ICH E9 on statistical principles, including how a study defines the precise question it means to answer. The chain is straightforward: careful design and conduct yield trustworthy evidence, and only trustworthy evidence earns a place on the label.
How to read a label as a clinician#
Reading a label with all of this in mind turns a dense document into a practical tool.
- The Indications and Usage section tells you what was proven, not the full range of what the drug might plausibly do.
- The trial population described there tells you how far you are extrapolating whenever your patient differs from it.
- Warnings and contraindications are earned from observed data, not written as theoretical caution.
- A narrow indication often reflects a focused program with room to grow. Some pathways, such as orphan-drug designation for rare diseases, exist precisely to make well-defined narrow indications viable.
The label, in the end, is the visible output of an otherwise invisible process: a protocol, a population, a set of endpoints, a statistical plan, and a standard of evidence, all compressed into a few paragraphs a prescriber can act on. Once the indication reads as a description of the evidence rather than of the molecule, the whole system becomes legible.
Sources and further reading
Questions and answers
Is off-label prescribing allowed?
Yes, in most systems a physician may prescribe an approved drug outside its labeled indication when clinical judgment supports it, and in fields such as oncology and pediatrics this is common. What differs is that the regulator has not formally reviewed that specific use, and manufacturers generally cannot promote it.
Why do two similar drugs carry different indications?
Because the indication follows the trials, not the biology. Two medicines with comparable mechanisms can end up with different labels if their sponsors studied different populations, doses, endpoints, or settings.
Does a narrow indication mean a drug is less effective?
Not necessarily. A narrow indication usually reflects the boundaries of the evidence submitted rather than a limit on the drug itself, and indications often widen as further trials report.