A new analysis reveals genetic evidence for shared biopsychosocial traits among eight kinds of chronic pain, including headache. Many of those traits had causal effects on pain.
Many studies have found associations between common chronic pain conditions and a number of complex traits. These include psychiatric traits like anxiety and depression, lifestyle factors including smoking and sedentary behavior, and socioeconomic characteristics such as income and education, for example. But whether these traits have causal effects on pain has remained unclear. A new study from researchers at the University of Queensland in Australia takes a genetic approach to add valuable insight on this question.
The investigators report results from a phenome-wide association analysis of people with different types of chronic pain, including back pain, neck/shoulder pain, hip pain, knee pain, abdominal pain, facial pain, headache, or widespread pain. This kind of analysis examines associations between genetic variants and a large number of phenotypes, or collection of traits.
The group identified nearly 6,000 significant genetic correlations between these common pain conditions and biopsychosocial traits. A total of 570 correlations were of a causal nature, with 410 traits, including somatic pathologies, psychiatric traits, socioeconomic factors, and medical comorbidities, contributing to an increased risk of chronic pain. These findings offer researchers the potential to develop new treatments that target causal factors and provide support for the biopsychosocial model of pain.
Roger Fillingim, a pain researcher at the University of Florida College of Dentistry, Gainesville, US, called the findings “heartening” and “very compelling.”
“These results show us the biopsychosocial model of pain in action. The biopsychosocial model tells us that the experience of pain is driven by complex and dynamic interactions among biological, psychological, and social influences. This study shows us that not only are there biological and psychosocial traits associated with multiple pain conditions, but at least some of those associations appear to be genetically mediated,” said Fillingim, who was not involved with the study.
The research appeared in the March 2023 issue of the Journal of Pain.
Take it to the (Bio)bank
Scott Farrell, first author of the study and a postdoctoral research fellow at the University of Queensland, said the overlap between different biological, social, and psychological traits across chronic pain conditions has long been of interest. When his university colleagues who work in genetics gained access to a large dataset from the UK Biobank – a biomedical database containing genetic and health information from half a million UK study participants – they asked Farrell and his supervisor, Michele Sterling, also a co-author on the new study, what would be the most impactful research questions to address with this trove of data.
“This all started when we were asked what could be done with these tools to answer a meaningful question,” said Farrell. “We wanted to see how the genetics might influence these different aspects of the chronic pain experience.”
Genome-wide association studies (GWAS) had previously linked genetic variations called single nucleotide polymorphisms (SNPs) to chronic pain and had also demonstrated a shared genetic basis between certain pain conditions and some biopsychosocial traits. But those studies did not address causality – whether those traits had a causal effect on pain.
To add this missing piece of the puzzle, the researchers performed their own GWAS analyses using UK Biobank data, which not only includes genetic data, but also data on psychological traits, lifestyle factors, and medical history. It also contains information about pain, including whether participants said if they had experienced chronic pain (defined as pain lasting more than three months) in one of eight body sites.
The researchers would perform a series of GWAS focusing on back pain (79,089 cases), knee pain (77,996 cases), neck or shoulder pain (72,216 cases), hip pain (41,677 cases), headache (39,283 cases), abdominal pain (21,285 cases), widespread pain (6,063 cases), or facial pain (4,037 cases). The study included 239,125 controls who did not report chronic pain. Through these GWAS, Farrell and colleagues, including geneticist and study co-author Gabriel Cuéllar-Partida and co-author Trung Thanh Ngo, would identify hundreds of genetic variants associated with these conditions.
Next, the researchers performed what’s known as cross-trait bivariate linkage disequilibrium score regression, a method allowing them to estimate genetic correlations between chronic pain and nearly 1,500 complex traits across the eight pain types. They found 5,942 significant correlations, many of which corroborated phenotypic associations from past clinical studies.
“This analysis gave the opportunity to see how the individual genetic variants associated with a trait like, say, chronic back pain, overlapped with the genetic variants associated with another trait, like having heart disease,” Farrell said. “It can help [us] see how different kinds of factors relate to one another, but it also provides a partial explanation for why some of these comorbidities may occur together. It supports the notion of there being a genetic signature for chronic pain that overlaps between these different chronic pain conditions.”
The research team also used additional statistical methods to estimate the probability that a correlation reflected a causal relationship.
“The intuition behind these methods is that if a trait X like pain is causal for a trait Y like depression, for example, then you would expect that all genetic variants that affect trait X also affect trait Y, but not the other way around,” said Farrell. “This allows us to assess whether a trait is likely causal of another.”
The researchers’ analysis showed that across the 5,942 genetic correlations between the common pain conditions and biopsychosocial traits, 488 traits likely had a causal effect on chronic pain, with 410 of them contributing to an increased risk of chronic pain and another 78 traits leading to a decreased risk of pain. Conversely, 82 traits were causally affected by pain.
As with the genetic correlation analysis, the causal analysis also indicates a shared genetic signature across the different pain types, since the direction of the causal relationships with the complex biopsychosocial traits was consistent across each of the chronic pain types.
For instance, the same traits, including psychological/psychiatric traits, musculoskeletal pathologies, socioeconomic factors, medical comorbidities, blood markers, medication use, and addiction traits, all increased the risk of pain for most of the kinds of chronic pain included in the study.
Farrell said some of the causal relationships weren’t so surprising.
“We saw that having rheumatoid arthritis appears to cause knee pain. That’s not a surprise,” he said. “But other traits we looked at did raise some questions. For example, socioeconomic factors like being a junior level office worker had a causal impact on neck/shoulder pain. Yet being a software engineer had a causal effect on not having neck/shoulder pain. We initially thought that maybe computer time could explain why the junior level office worker had that pain. But then why wouldn’t the software engineer? It’s possible these jobs are proxies for socioeconomic standards, and it gives us things to investigate in the future.”
Frances Williams, a researcher at King’s College London, UK, who studies migraine genetics but was not involved with the current investigation, said the study’s methods are “very interesting and useful.”
“We’ve done similar work looking at twins, and what this group has done has taken an enormous biological resource and expanded the number of phenotypes in their analysis,” she said. “And overall it shows us that we need to do better for our patients by looking holistically at people living with chronic pain and not treat symptoms in isolation.”
What about headache?
The study yielded some interesting traits with causal effects on headache, ranging from painful gums, hernia, and being adopted as a child to depression, renting a home, death of a close relative, and intervertebral disk disorders, among roughly two dozen such traits. Things went in the other direction, too, with headache causally contributing to several biopsychosocial traits, such as the use of analgesics, multisite chronic pain, and psychological traits like depression, among others.
But Farrell said it is hard to determine how the different traits may relate to specific headache conditions like migraine.
“Like many of these big GWAS, we have to work with the data that are available to us,” he explained. “The question for this dataset was simply whether you had pain in the head for more than three months. It could be migraine, tension-type headache, any other type of headache, or something else entirely. Given that, we were not able to drill down closely enough to figure out whether some of these traits were related to specific conditions.”
Williams, whose study did look at migraine specifically, said a separate analysis focusing on specific headache conditions might show some unique results.
“In our data, the migraine genetics looked separate from other pain conditions,” she said. “Granted, our datasets were relatively small compared to some of the enormous studies that are coming out now. But we’ve known for a while that migraine seems different from other pain conditions. I suspect we might see somewhat different genetic associations than what we see in more musculoskeletal-type phenotypes, but we still need to do that research.”
Farrell said the new study, like most good research, likely provides more questions than answers for chronic pain researchers.
“These results raise interesting questions about why certain traits are associated,” he said. “We can use those questions to design new studies and try to drill down in a more targeted way to investigate what’s really happening with specific relationships, to better explain why so many of these factors occur together. You could also use these results to design new clinical trials, perhaps trying to treat cardiovascular disease or better manage socioeconomic factors, to help improve pain outcomes.”
Fillingim hopes the study’s evidence of a “biological footprint” for chronic pain can help change the way clinicians and researchers think about pain conditions.
“This provides additional support to the idea that psychological factors are not just making people think they have more pain – they are actually driving biological systems or are being driven by biological systems that change the experience of pain,” he said. “It’s exactly what the biopsychosocial model would predict. We have many therapies, like exercise and cognitive behavioral therapy, that are effective for pain and low risk but not nearly as accessible as many drug treatments. That’s something that needs to be addressed.”
Kayt Sukel is a freelance writer based outside of Houston, Texas.
Farrell et al.
J Pain. 2023 Mar;24(3):369-86.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Sign Up For An MSC Newsletter
Share this article: