i-Base Q&A on the START study results

START-logoWhat is the START study?

The START study is a large international HIV study.

It looks at the risks and benefit of early HIV treatment.

“Early” is defined as having a CD4 count above 500. Results are compared to a “later” group that waited until the CD4 count was 350.

People were randomised to one of these two groups, so no-one was able to choose which group they joined. This study design makes the results more reliable and strong.

START stands for Strategic Timining of AntiRetroviral Treatment. It is one of the largest ongoing HIV studies.

What is the news about START?

On 27 May 2015, early results showed that the early treatment group was doing better.

The details were given in a press release from the US National Institute of Health, who are a major funder of the study.

One of the surprises is that even at very high CD4 counts, treatment reduces the risk of HIV related illnesses.

What were the study outcomes?

START looked at the risk of serious medical events. These events were looked at separately depending on whether they were related to HIV.

  • ‘AIDS events’ included a list of illnesses that were more common before we had effective treatment. These are all proven to have a clear link to HIV.
  • ‘Non-AIDS events’ included serious illnesses that also are common in people who do not have HIV. They may be related to HIV, but the link is not so clear. These include heart disease, liver disease, kidney disease and some non-AIDS cancers.

Early treatment was expected to have fewer serious events. Because everyone had a high CD4 count, early treatment was expected to reduce on non-AIDS events the most.

Why are the results so early?

Large studies often have a small group of independent experts called a DSMB. This stands for Data and Safety Monitoring Board. Unlike everyone else working on the study, this group can follow the ongoing results. They are responsible for overall safety in the study,

On 13 May 2015, the DSMB saw that early treatment was now significantly better. The DSMB recommended that the study be changed. Instead of continuing with both groups, all participants can now change to early treatment.

This was also a surprise because START results were not expected until 2017. So, the research question has been answered at least 18 months early. This was after an average of 3 years follow-up rather the planned 4.5 years.

Does this mean that the START study will now close?

No. The study will continue to monitor and treat everyone until at least the end of 2016.

This follow-up will provide important information from both groups.

How definite are the results?

The results are very clear.

Approximately half as many serious events (53% less) occurred in the group using early treatment.

When planning the study the research team thought that this reduction would only be about one third less (36%).

What happens to people in START?

  • People already on treatment will continue with monitoring and treatment.
  • People in START who are not yet on treatment will now have the chance to start treatment.

In many countries, this would not be possible if they were not in the study.

Clinics will be contacting people in the study over the next few weeks to talk about the results.

What if people want to start early treatment who are not in the study?

The START results are likely to make it easier to start treatment at higher CD4 counts.

Whether you can do this will depend on where you are treated. It will depend on the guidelines in your country.

Although guidelines sometimes take time to respond to new results, this news is likely to change guidelines in most countries.

As with all treatment questions, this is something to talk to your doctor about.

Who was in the study?

The study enrolled 4685 people from 35 countries.

  • 33% were in Europe.
  • 5% in South America or Mexico.
  • 21% in Africa.
  • 11% in the US.
  • 8% in Asia.
  • 2% in Australia.

Age ranged from 18-81, with an average age of 36 (and half were aged between 29 and 41).

More than half were gay men and more than a quarter were women.

Half the participants joined the study within a year of their HIV diagnosis.

Everyone had a CD4 count above 500 and 20% had a CD4 count above 800.

The number of other medical complications in people joining the study is important, given that this is generally a young group with early HIV infection.

  • Almost one-third were current smokers.
  • Half had at least one important risk for heart disease.
  • Almost 1 in 5 either had high blood pressure or were being treated for this.
  • About 1 in 12 either had high blood lipids or were on lipid lowering drugs.
  • Just over 3% had diabetes, were using diabetes treatment or had high fasting glucose.
  • Just under 3% had hepatitis B or and just under 4% had hepatitis C.

Other important health issues included about 150 (3%) people with alcohol or substance use issues and 270 people (6%) had a psychiatric diagnosis (including depression, bipolar and other conditions).

How many serious complications occurred?

The decision to change the study was based on 127 serious events: 41 in the early treatment group and 86 in the deferred treatment group. This was much lower than expected.

In 2013, the researchers worked thought that 213 events would be needed to see a clear difference between the groups.

It is very good that the study question was answered with fewer events.

What were the number of events in each group?

Tables 1a and 1b below give the results based on the analysis in March 2015. They show the different results for the two groups: early vs late. They show the results for the three different ways that serious events were categorised.

It also reports the results in three different ways.

  1. The actual number of events. For example, by category 1 there were 41 events in the early treatment group compared to 86 events in the late treatment group. (Table 1a).
  2. The event rate in each group per 100 patient years. For category 1, the rates were 0.6 compared to 1.25 per 100 patient years of follow up (PYFU).  (Table 1b).
  3. The difference between the two rates is then calculated so that the overall reduced risk from earlier treatment can be given as a percentage. For category 1, the 0.47 when subtracted from 100 shows that early treatment reduced the risk by 53%. The range in brackets, relates to the researchers being 95% confident that the actual percentage is somewhere between 32% and 68%. It is a way of allowing for the fact that some results happen by chance.  (Table 1b).

Table 1a. Number of primary endpoints in each arm (15 May 2015)

Number of events
Early arm (A) Later arm (B)
Category 1:AIDS, serious non-AIDS, or death (primary). 41 86
Category 2:AIDS or AIDS death. 14 46
Category 3:Serious non-AIDS or non-AIDS death. 28 41

Table 1b. Relative rates of primary endpoints in each arm (15 May 2015)

Rate per 100 PY

Hazard Ratio

Arm A/B (95% CI)

Early arm (A) Late arm (B)
Category 1:AIDS, serious non-AIDS, or death (primary). 0.60 1.25 0.47 (0.32 to 0.68)
Category 2:AIDS or AIDS death. 0.20 0.66 0.30 (0.17 to 0.55)
Category 3:Serious non-AIDS or non-AIDS death. 0.41 0.59 0.67 (0.42 to 1.09)
NS **

* PY = patient years, ** NS = not statistically significant

What were the main illnesses reported?

There are so far only some limited details.

Overall, 127 complications only occurred in both groups. This included 41 in the early group and 86 in the late group.

What is important is these serious events occurred about twice as often in the group not on treatment. The result was similar when looking at all serious events whether or not they were generally related to HIV.

Many of these complications were serious but treatable. Non-fatal AIDS-related complications included TB and cancer. Non-fatal non-AIDS complications included cardiovascular (heart, stroke) and other cancers. Both the AIDS-related and non-AIDS related events occurred less often in the early treatment group.

Non-AIDS related complications were a major focus for the study. These were defined as heart disease complications, liver or kidney complications or non-AIDS cancers.

Non-AIDS complications occurred less than expected.

Were the results the same in different countries?

START includes sites in many different countries. There are 2530 participants in low/middle income countries and 2,155 people in high income countries.

The benefits of earlier treatment were seen in both settings.

Were the results different for some groups of people?

More detailed analyses will look at whether the results are results to other factors including age, race, sex, HIV history etc.

Following thousands of people for more than three years has created a huge database.

This will allow many other questions to be asked about different risks.

When will more detailed results be available?

Only the main summary results were included in the press statement. This is because the detailed results are still being analysed.

The next results are likely to be presented at the International AIDS Conference being held in Vancouver in July 2015.

This conference might also include early results from some of the sub-studies.

Why are the results so important?

Up until now, very large database studies had not been able to find additional benefits of starting treatment at high CD4 counts.

Because START is a randomised study, this make the findings unlikely to just be by chance. Because it is such a large study, the results are also very significant.

The new results will affect treatment guidelines in rich and poor countries alike. They are likely to influence global guidelines from the World Health Organization (WHO).

The results may even cause stock prices to rise for a while for the drug manufacturers.

More importantly, the results will make treatment easier for people to access across the world.

Does this mean that drug companies are behind the results?

No. START was led by the large team of independent researchers in the INSIGHT network.

Main funders were public health institutions.

Drug manufacturers supported the study by donating their medicines. This was a tremendous help, but the companies did not design and run the study and they had no influence on the results.

Even within the study, participants and their doctors decided which treatments to use.

Some people say early treatment was always going to be better?

In designing the study, the researchers thought that early treatment would have advantages.

But START is an essential study for providing good quality evidence rather than opinion.

This was always a central part of the study, irrespective of the final results.

This huge study provided excellent health care for many years, often in settings where treatment would not be available at higher CD4 counts.

In providing the highest quality evidence, START will have one of the biggest influences globally in changing guidelines to make treatment available for everyone.

It took a brave and determined research team to set up and run the START study. It was also dependent on thousands of HIV positive people supporting the research in order to benefit the wider community.

It is significant that the study questions have been answered earlier than expected, and involved fewer people having to experience serious complications.

Why was it important to get this evidence?

Getting good evidence is a key success in START. Until now, many people had hoped and guessed that early treatment might be better. The evidence for this though was not very good. Some studies found benefits and other found little difference.

Although we rely on our doctors for information, often there might not be good quality evidence even for advice in treatment guidelines. Your doctor should really explain when the lack of evidence makes it difficult to know which option is best.

When rating evidence, randomised studies score highest, database studies score mid-way and expert opinion is rated last.

START is a randomised study. Previously, database studies disagreed on whether early treatment was good, so the recommendations to start earlier was mainly based on opinion.

Who ran the START study?

Four international research centers coordinated the work of the 215 sites in START.

  • The Medical Research Council (MRC) Clinical Trials Unit at University College London.
  • The Copenhagen HIV Programme at the Rigshospitalet, University of Copenhagen in Denmark.
  • The Kirby Institute at the University of New South Wales in Sydney, Australia.
  • The Veterans Affairs Medical Center affiliated with George Washington University in Washington, DC.

The University of Minnesota served as the statistical and data management center and trial sponsor.

Who funded the START study?

The main funder was the US National Institute of Allergy and Infectious Diseases (NIAID).

Additional funding for the START trial was also provided by other US organisations:

  • National Cancer Institute.
  • National Heart, Lung and Blood Institute.
  • National Institute of Mental Health.
  • National Institute of Neurological Disorders and Stroke.
  • Eunice Kennedy Shriver National Institute of Child Health and Human Development.
  • NIH Clinical Center.
  • National Institute of Arthritis and Musculoskeletal and Skin Diseases.

Funding was also provided by the many national health agencies:

  • Agence Nationale de Recherches sur le SIDA et les Hépatites Virales (ANRS) in France.
  • Bundesministerium für Bildung und Forschung in Germany.
  • European AIDS Treatment Network.
  • Government organizations based in Australia, Denmark, and the United Kingdom.

Further information

i-Base community article about the START results

HTB technical report on the START results

NIAID statement: Starting Antiretroviral Treatment Early Improves Outcomes for HIV-Infected Individuals NIH-Funded Trial Results Likely Will Impact Global Treatment Guidelines (27 May 2015).