Developing New Medicines

The discovery of the antibiotic penicillin shows how an observation can reveal a potential medicine. In 1928, Alexander Fleming noticed that bacteria did not grow near a Penicillium mould contaminating a Petri dish. The mould released penicillin, which killed surrounding bacteria. This suggested that the chemical could be useful against bacterial infections, but discovering its effect did not by itself make it ready for use as a medicine.

Today, scientists at universities and pharmaceutical companies search for new medicines. Most new drugs are synthesised—made through chemical reactions—by chemists, although a chemical from a plant or microorganism may still provide the starting point. Researchers develop promising candidates for further testing. One continuing aim is to find antibiotics that work against bacteria resistant to existing treatments.

What testing needs to establish

Testing addresses three connected questions:

• Toxicity: does the drug cause harmful effects?
• Efficacy: does it produce the intended benefit?
• Dose: what amount produces an effective treatment without unacceptable harmful effects?

An effective drug is not necessarily a safe one. Researchers therefore seek an optimum dose, balancing benefit against harm and identifying the lowest dose that is effective.

Preclinical testing: before human trials

Preclinical testing takes place before a potential medicine is tested in humans. Laboratory tests use cells and tissues to investigate useful effects and toxicity.

Testing on live animals then allows researchers to examine effects in a whole organism rather than only in isolated cells or tissues. These tests investigate efficacy, toxicity and suitable doses. Evidence from preclinical testing helps researchers decide whether the candidate is suitable to progress to human trials.

Clinical testing: volunteers and patients

Clinical trials test potential medicines in humans. Trials generally begin with healthy volunteers, using very low doses to reduce the risk of harmful effects. Researchers monitor the volunteers, and doses can be increased cautiously if the initial results indicate that this is safe.

Further trials involve patients who have the condition the medicine is intended to treat. Researchers investigate whether it works against that condition, monitor side effects and compare doses to establish an optimum dose. Safety testing therefore continues: it is not something checked once and then forgotten.

Making a fair comparison

In some patient trials, one group receives the active drug and another receives a placebo, a treatment containing no active drug. Comparing the groups helps researchers judge whether the active drug provides a benefit beyond changes seen without it.

For example, an improvement in patients receiving a new painkiller is not enough on its own to demonstrate efficacy. Researchers need to compare their improvement with that of the placebo group.

In a double-blind trial, neither the patients nor the doctors assessing them know which treatment each patient receives during the trial. This reduces the influence of expectations on patients' reports and doctors' assessments, helping to reduce bias.

Checking the evidence

Research findings are peer-reviewed by other scientists before publication in scientific journals. This provides a check on how the research and its results are reported. Developing a medicine therefore involves more than finding a useful chemical: it requires evidence gathered and scrutinised across several stages.

1. Preclinical: cells and tissues, then live animals; investigate efficacy, toxicity and doses.
2. Clinical: healthy volunteers first, generally starting with very low doses; monitor harmful effects and increase doses cautiously.
3. Patient trials: test effectiveness against the condition, monitor side effects and establish the optimum dose.

Essential distinctions

• Toxicity: harmful effects. Efficacy: intended benefit. Dose: amount given.
• Placebo comparison: helps identify the benefit of the active drug.
• Double-blind: neither patient nor assessing doctor knows the treatment allocation; reduces bias.
• Peer review: other scientists evaluate the research before publication.