The hardest outcome of all: HIV and suicide

Suicide rates are higher in people with HIV. But how much higher?
Photo credit: panitanphoto/
Photo credit: panitanphoto/

Occasionally, an apparently orthodox, peer-reviewed scientific study crosses our desks at aidsmap that prompts the reaction “Really??” One such came along back in April, when a paper appeared in the medical journal General Psychiatry, which comes from the publishers of the British Medical Journal.

It was the first attempt to do a systematic review and meta-analysis of the rates of suicide, suicide attempts and suicidal ideation (thoughts) in people with HIV. The paper, by Matt PeIton and colleagues from Penn State University in the US, reviewed 40 studies from all continents that between them included 185,199 people with HIV.

But we hesitated to report it, and there are two reasons why. The background reason is that suicide is a sensitive subject, and may be more sensitive to those of us who have overcome physical challenges to our lives. We wanted to make absolutely sure that what we reported was accurate and not sensationalised. So we asked three statisticians to critique the paper and interviewed a psychiatrist and a clinical psychologist about their views on suicide in people with HIV, to add background.

The second reason was that what the paper reported seemed shocking, and we wanted to understand if its findings were true.

Suicide rates – globally and in HIV

The World Health Organization (WHO) reports that 700,000 people worldwide died by their own hand in 2019, the last year we have figures for, and that globally, 1.3% of deaths in 2019 were the result of suicide. In the general population, it is the third most common cause of death in young women aged 15-29, and the fourth most common among young men. Rates in males of all ages are twice as high than they are in women, as are other deaths of violence such as assaults and road accidents.

There is no relationship between suicide and income, and rates in males are highest in high-income Europe and low-income Africa; for women, they are highest in south-east Asia.

There is a tendency for high-latitude rather than tropical countries to have higher rates. The figures in most Islamic countries and in strongly Catholic central and northern South America are notably low, so maybe a unified religious culture is protective. Rates peak at the age of 25 in lower-income countries, 55 in higher-income ones. One piece of good news is that rates have fallen in all global regions but one in the last 20 years, and have almost halved in Europe and east Asia. The exception is North America, where they have slightly increased.

The annual incidence of suicide worldwide in 2019 was, WHO found, 0.009% or one suicide in just over every 11,000 people per year (it’s one in just over 8000 men).

The General Psychiatry paper found 12 studies, all from North America and Europe, that documented rates of what’s rather chillingly called “completed suicide” in people with HIV. It found a rate of 1.02% a year, or one in every 98 people. One hundred times as many as in the general population.

That’s obviously the figure that attracted attention. But it also found higher rates of lifetime suicide attempts, and lifetime suicidal ideation.

These are both much more common than completed suicides, both because they’re “have you ever?” questions, but also because completed suicide is much rarer than attempted suicide. They found a lifetime rate of 15.8% in people with HIV for attempts, versus a WHO global estimate of 3% in the general population. The annual incidence of attempts in people with HIV was 2.04%.

As for suicidal ideation, we might think this was extremely common, but it’s usually asked within one of several validated questionnaires that gauge mental health, and generally implies active contemplation and some sort of planning, such as saving pills or making wills, rather than just waking up with dark thoughts. The ideation rate in people with HIV was 22.8%; the WHO general-population estimate is 9%.

Are suicide rates really that high in people with HIV?

From the start, however, it was clear that there was something odd about the very high completed-suicide figure. The WHO estimates that only one in every 286 suicide attempts actually ends in death. And yet in the meta-analysis the ratio of the annual incidence of attempts (in one set of studies) to the incidence of actual suicides (in a different set) was exactly 2:1. There seemed to be no reason why people with HIV should be 143 times ‘better’ at turning attempt into reality.

Then there was the timing of the research papers. Of the 40 studies, ten collected their data before 2000, while another seven were cohort studies that collected data over a long period that included the years prior to 2000 when treatment was either unavailable or only partially effective. In the 12 papers reporting actual suicides, only one collected all its data post-2000. In the 16 papers recording lifetime suicide attempts, 12 collected all their data post-2000.



A branch of medicine focused on the diagnosis, treatment and prevention of mental, emotional and behavioral disorders. Psychiatrists are doctors who trained in general medicine before specialising in mental health. They can prescribe medication.

exclusion criteria

Defines who cannot take part in a research study. Eligibility criteria may include disease type and stage, other medical conditions, previous treatment history, age, and gender. For example, many trials exclude women who are pregnant, to avoid any possible danger to a baby, or people who are taking a drug that might interact with the treatment being studied.


A feeling of unease, such as worry or fear, which can be mild or severe. Anxiety disorders are conditions in which anxiety dominates a person’s life or is experienced in particular situations.


When the statistical data from all studies which relate to a particular research question and conform to a pre-determined selection criteria are pooled and analysed together.


The study of the way people think, behave and interact. Psychological therapies are based on talking and working with people to understand the causes of mental health problems and develop strategies to deal with them. Psychologists have specialist training but are not medical doctors.

The lifetime rate of suicide attempts fell by an absolute percentage of 0.17% per year of publication. Being on antiretroviral therapy pushed the rate of attempts down more, by 1.35%, meaning that if that rate continued, it would nearly halve in ten years.

These differences were not statistically significant, but what was significant was that for every 10% fewer people with an AIDS-defining CD4 count or illness in a study, the rate of completed suicide fell by one-third (0.34% in absolute terms). Having a life-threatening illness is one of the factors most strongly predictive of suicide in the wider population, and these figures seemed to bear this out.

Our experts noticed other odd things about the meta-analysis. Professor Andrew Phillips of University College London noticed that the rate of completed suicide was amplified by two papers, one a review of death certificates in France in 2000, and the other of mortuary records from San Francisco from 1995-97. They reported annual incidences of suicide in people with HIV of 10.7% and 19.3%, which is not credible; one in five of all people with HIV did not die each year, even in the mid-90s, let alone specifically from suicide.

But, Phillips noted, what they were reporting was not the rate of suicide in all people with HIV, but the proportion of deaths in people with HIV that were due to suicide. The much larger group of people who did not die that year and so who, by definition, were not suicide cases, must be included if you’re to calculate incidence. So these studies should have been excluded.

Phillips’ colleague Professor Caroline Sabin commented that establishing whether a case actually is a suicide is hard. Deaths that appear to be suicide could be accidents; drug overdoses, for instance, or people who habitually self-harm but accidentally went too far.

She also noticed that several of the outlying studies with very high rates both of completed and attempted suicide concerned very vulnerable groups of people. One study reporting a suicide attempt rate of up to 22% a year was a study from Puerto Rico which included large numbers of drug users. Another study of suicide attempts was of largely Black perinatally infected children in the US. A third, which reported high rates of completed suicides in the Netherlands, was conducted in the early AIDS years (1984-92), and its title tells its own story: Death from suicide and overdose among drug injectors after disclosure of first HIV test result.

The two US studies quoted above were the only two from that country that reported an annual attempted-suicide incidence. Three studies from Europe all found a much lower annual incidence of attempts. As a result, the review’s figure for the annual incidence of attempted suicide is 51 times higher in the US than it is in Europe.

If outliers like these are excluded, we get something nearer to the global figures, though they are still higher. For the complete-suicide incidence rate, it’s hard to calculate without taking account of study sizes, but if we exclude outlying results it comes out as in the region of 0.7% a year. This is nearly eight times higher than the global rate and 5.5 times higher than the global rate in men, but is now within the same order of magnitude. Similarly, for lifetime suicide attempts, the rate is five times the global rate, and for suicidal ideation 2.5 times.

These are much nearer the figures reported from a study by Dr Sara Croxford of Public Health England which was presented to the 2017 British HIV Association conference; it was reported by aidsmap here. This study found that the suicide rate in people with HIV was more than twice that of the general global population (0.021% a year, accounting for 2% of all deaths in people with HIV). The rate in men was 0.032%.

Croxford emphasised to aidsmap that the suicide rate was five times the global population rate in the first year after diagnosis (0.052%), and that this accounted for 40% of all deaths by suicide in the cohort.

Still, even if the true suicide rates in people with HIV are not as raised above average at they first seemed on reading the General Psychiatry paper, they are still higher than in the general population. I interviewed two experts in mental health in people living with HIV, clinical psychologist Professor Lorraine Sherr, also of University College London, and Dr Pepe Catalan, former consultant psychiatrist at the Chelsea and Westminster Hospital, about their experience of the issue.

Dealing with suicide

“It is complicated to analyse suicidal ideation, attempts and suicide rates”, says Sherr. “Completed suicides are a mixture of ‘successes’ and ‘failures’ – people who meant to die and did, and people who didn’t really want to kill themselves but did so accidentally while they were self-harming (which is in reality usually a technique to control unmanageable anxiety, not to die) or because they accidentally overdosed.

“Equally, those who ‘attempt’ suicide are a mixture of people who wanted to die but ‘failed’ to – there weren’t enough pills, they were discovered in time, the train missed them – and people who are doing what used to be called cries for help. And they’re very different people. The first group want to die, the second group just don’t know how to live.”

Catalan agrees: “It also complicates the study of suicide because the demographics of these two groups are so different. Self-harmers are more often younger and female, planned suicides are more often older and male. People who plan suicide are more often people, especially men, who’ve experienced some great reversal of fortune, some humiliation or shame. They can’t adjust to the change of status.

“HIV diagnosis can of course be experienced as such a change. Back in the days when HIV often meant loss of employment, that often seemed to be a precipitating factor. The stigma of suicide also affects its recording – coroners may record ‘accidental death’ under pressure from families.

“But I think more often the precipitating factors for HIV infection and the precipitating factors for suicide are the same. We already know gay men have a higher rate of suicide than other men, and are also more likely to get HIV.

“Even sociopolitical factors matter. The closure of the coalfields did not immediately lead to an increase in suicides, perhaps because the shared experience and community support offset the impact of the job loss. In contrast, when people become unemployed alone, the impact is greater, and suicides in general went up during the deindustrialisation of the 1980s, but have been slowly declining since. I’ll be interested to see what COVID has done – will it be experienced as a time of shared experience and support, or of anxiety and isolation?"

Can suicides be predicted, and therefore prevented?

“No,” says Catalan. “We’re hopeless at predicting who will actually commit suicide. I mean yes, there were lots of people at my clinics one could point to and see that they were the type of person who might become suicidal, but there were lots of those, and most don’t actually end up killing themselves. We couldn’t give everyone the kind of intensive, prophylactic support that might be needed.”

“But you very much listen to them and let them talk,” says Sherr. “One of the biggest myths about suicide is that those who talk about it don’t do it. In my experience that’s not true. In fact, if someone who normally finds it hard to talk is suddenly saying they have suicidal feelings, they should be taken very seriously.”

“Yes, hang on to your chair, shut up and listen,” agrees Catalan. “And don’t rush in with solutions, don’t minimise their feelings, don’t tell them they’ll feel better soon and above all, don’t tell them to pull themselves together.

“Depression is not necessarily a sign of suicidality. In fact, often in the days before a suicide, people’s mood lifts. They’re leaving it all behind. It’s when a person talks about plans you really take notice.

“One of the worrying signs and precipitating factors is the complete loss of regular sleep patterns,” He adds. “It’s a sign of a state of free-floating anxiety and loss of role, and of course if you were not suicidal before, you might well be after a few nights’ lost sleep.”

“The best kind of prevention is the support you would provide for anyone,” says Sherr. “Have they got good social service and housing support? Do they have a GP and have a rapport with them? Medicines can make a big difference, and I don’t mean antidepressants necessarily, I mean treatment for chronic illnesses or pain.”

(Sara Croxford’s paper shows that, back in the days when antiretroviral therapy was not given immediately on HIV diagnosis, suicide rates in people with HIV improved dramatically as soon as they started it – both because they felt physically better, and because they felt in control of their health.)

“But it’s important not just to meet external needs,” Sherr adds. “Suicide is ultimately a psychological issue, and is about losing hope. I try and link people into a future, and I know I’m getting somewhere if I sense that they see one for themselves.

“Suicidal people have often lost their moorings, there is no structure in their lives, and I try to introduce a bit of that, even it it’s just to say ‘Call me tomorrow! At ten!’”

"A failed suicide attempt can even be interpreted as a signal from a better future,” says Catalan. “One of my first psychiatric jobs was working with people who’d jumped under tube trains. More often than not, they survive – they fall into the well between the rails. I had an HIV-positive client that happened to. He regarded it as a miracle, a sign that he was destined to live, and he never did it again.”

Suicide and its effects

What about the idea that suicide is ‘catching’? That it’s a learned behaviour? This is often why people are nervous to discuss it.

“There is some truth in the ‘copycat suicide’ idea,” says Catalan. “Back in the AIDS days, we would have clusters of suicides. It somehow gives people social permission.

“I do worry about social media – but not because it’s directly encouraging suicidal behaviour or self-harm. Rather the opposite: everyone is presenting themselves as successful, beautiful, loved, and above all happy. And if you don’t feel those things, you can feel excluded from everything. I worry that young people feel ashamed if they’re not happy.”

What would you advise to the families and loved ones of people who kill themselves? Or who are suicidal?

“The same advice we’d apply to ourselves, though families will feel guilt, shame and regret more intensely,” says Sherr.

“If someone is suicidal, try to listen, try not to get angry or directive, don’t suggest solutions too fast but do suggest possibilities. If someone has killed themselves, try to assure the family that it wasn’t anything they did or failed to do that made it happen. In reality, without their support, the person might have done it ten years earlier. They loved them for ten years and that kept them alive.”

 Professionals can be as affected by the suicides of patients as anyone. When I ask Sherr about how she understands the ones who want to die and do, she says “All I understand about them is how they make me feel!”

Catalan replies: “The first suicide patient I ever saw was a woman who had taken an overdose. I talked to her, we achieved some rapport, I arranged an appointment for the following week. She killed herself that week. What else could I have done?”

For everyone, suicide prompts a mixture or reactions, of simultaneous alienation and empathy. On the one hand, if while there’s life there’s hope, the degree of despair that may make someone end their lives is hard to imagine, and that may make us angry with them for leaving us.

On the other hand, there’s fellow feeling; there are probably few people who haven’t, one dark night, thought to themselves: “What’s the point?”. Even if suicide is not contagious, it can feel so.

Suicide can harm friends and family almost as much as it does the person who kills themselves, often because of guilt. “If only I’d seen they were desperate, if only I’d talked to them, if only I’d said the magic thing that would have stopped them”, we think. When it’s a parent, child or lover who goes, the effect can be devastating, and permanent.

Speaking personally, I’m fortunate never to have been that close to a suicide; but there was a strange period in 2008-9 when three HIV activists I knew, all loved and valued campaigners, killed themselves.

I asked an old school friend of one what he thought had brought on his friend’s suicide. “He’s been trying to do it all his life,” he shrugged. In other words, probably the most disturbing thing about suicide is that it’s unfathomable. We’ll never know.

If you've been affected by some of the issues in this article, Samaritans can be contacted in the UK on 116 123, and in the US, the National Suicide Prevention Lifeline is 1-800-273-8255. Other international helplines can be found at or via this Wikipedia page.


Pelton M at al. Rates and risk factors for suicidal ideation, suicide attempts and suicide deaths in persons with HIV: a systematic review and meta-analysis. General Psychiatry, 34:e100247, April 2021 (open access).