Depression and inflammation: when the immune system has a say in mood

Many people know this pattern: after an infection, mood stays at rock bottom even though the fever is long gone. The body is considered healthy, but the joy does not come back.

Perhaps you also landed here because of a lab result. A slightly raised CRP, a depression that feels stubborn despite treatment, and a search for “depression inflammatory markers” in the middle of the night.

I will show you what holds up about the link between inflammation and the mind, and just as clearly what does not. Nothing in this text is a reason to postpone, change or end psychotherapy or medication. The inflammation perspective is added on top. It replaces nothing.

When being ill feels like depression

Think back to your last proper flu. Not the fever, but everything around it. You did not want to see anyone. Food did not matter. What normally brings you joy felt meaningless, and your head felt as if it were wrapped in cotton wool.

Withdrawal, loss of appetite, joylessness, exhaustion, slowed thinking: these are also core symptoms of depression. Research calls this pattern sickness behaviour. It is not a defect but a protective programme. Imagine your immune system as a fire brigade that, when there is a fire, switches the whole house into rest mode. The messengers that carry this command to the brain are called cytokines.

The observation that founded the field

People treated with high-dose interferon alpha for melanoma or hepatitis C became depressed strikingly often. Interferon alpha is itself an immune messenger. So part of the body's own inflammatory response was given as a medicine, and mood tipped over.

Not an isolated finding: in a meta-analysis of 26 prospective studies on hepatitis C, the cumulative incidence of depression under interferon after 24 weeks was 0.25, roughly one person in four. People with high baseline IL-6, women and above all people with a previous depressive episode (odds ratio 3.96) were at particular risk. So inflammation does not affect everyone equally (Udina and colleagues, 2012) [Meta-analysis, observational studies].

The experiment in healthy volunteers

Interferon is an extreme situation. That is why researchers gave healthy people a small inflammatory stimulus.

And now you know why being ill and being depressed can sometimes feel so confusingly alike.

Inflammatory markers in depression: what CRP, IL-6 and TNF-alpha show and what they do not

Perhaps your lab report is lying next to you right now. A CRP value just above the reference range, a small asterisk next to it. And the question of whether that asterisk has anything to do with your mood.

You need three names for this. CRP, C-reactive protein, is made by the liver when inflammatory signals are circulating. It is the smoke detector in the hallway: it reports that something is smouldering somewhere, but not where. IL-6 and TNF-alpha are two of the cytokines that trigger this signal, the sparks, so to speak.

What the large meta-analyses say

A cumulative meta-analysis of 58 studies found higher IL-6 (d = 0.54) and higher CRP (d = 0.47) in depression, including in analyses restricted to people not taking antidepressants. TNF-alpha came in at d = 0.40, but is uncertain because of large differences between studies (Haapakoski and colleagues, 2015) [Meta-analysis, observational studies].

A quarter, not everyone: the most important number

In two Copenhagen population studies with 73,131 people, the likelihood of self-reported antidepressant use rose with each higher CRP category, up to an odds ratio of 2.70 at values above 10 mg/L. An association, not proof of causation (Wium-Andersen and colleagues, 2013) [Cohort, cross-sectional analysis].

What is inflammatory depression?

The term inflammatory depression is not a diagnosis of its own in the ICD classification system but a research term. It describes the idea that there might be a subgroup in which inflammation plays a greater role. A research group around Penninx calls a related picture immunometabolic depression.

Why silent inflammation is so closely tied to weight is explained in detail in Silent inflammation and weight.

And now you know why a single CRP value says so much less about depression than many people hope or fear.

How inflammation may affect the brain, and why the direction often remains open

Perhaps you are thinking now: the brain is well protected, isn't it? How is a messenger in the blood supposed to make my thoughts heavy?

The answer is not a single route but a network. That is exactly what neuroinflammation in depression means: not a pus-filled inflammation in the head, but altered communication between the immune system and the nervous system. From the perspective of Clinical Psychoneuroimmunology, KPNI for short, nerves, immune system, metabolism and hormones are not separate departments but one conversation.

Miller and Raison see behind this an ancient alarm programme against pathogens and predators, which today appears to contribute to the development of depression and to poorer response to antidepressants [Review].

Kynurenine: the turn away from serotonin

Imagine tryptophan as a train at a set of points. One track leads to serotonin, the other to kynurenine. Inflammatory messengers can switch the points via the enzyme IDO. On the kynurenine track, a second set of points follows: towards the rather protective kynurenic acid or towards quinolinic acid, which is considered potentially damaging to nerves.

A meta-analysis of 101 studies with 10,912 participants found an increased ratio of kynurenine to tryptophan in major depression. The team around Marx interprets this as a shift away from serotonin. The finding also appeared in schizophrenia and is not specific to depression [Meta-analysis, observational studies]. How much serotonin is made in the gut and how much of it reaches the brain is a story of its own, told in Depression and the gut-brain axis.

Neuroinflammation in depression: what can be seen in the brain

Microglia are the caretakers of the brain. They tidy up and maintain the connections between nerve cells. The hypothesis: under constant alarm, this maintenance work could suffer.

In 48 untreated people with depression, a higher CRP went along with weaker connectivity in the reward system and more pronounced joylessness, an association, not a cause (Felger and colleagues, 2016) [Cross-sectional, fMRI].

The blood-brain barrier: so far mainly mice

Imagine the blood-brain barrier as tightly grouted tiles. The grout consists of proteins such as claudin-5.

A second mouse model brings muscle into play: when a team around Agudelo activated a regulator in muscle that endurance training normally switches on, the muscle converted more kynurenine into kynurenic acid, which does not enter the brain, and the animals were protected against stress-induced depression-like behaviour. Purely animal data, not a training recommendation [In vivo, mouse].

Chicken or egg: why the direction in humans is often open

Does inflammation make you depressed, or does depression make you inflamed? If someone sleeps badly and barely moves for weeks, their inflammatory markers could also be the consequence. In humans, this can only be studied indirectly: with children followed over years, with genes, and with measurements during treatment.

In Copenhagen, too, a higher CRP predicted later hospital admissions with depression [Cohort]. At the same time, there are clear indications of a two-way relationship: in 15 longitudinal studies with 58,745 people, obesity raised the risk of later depression (odds ratio 1.55) and depression raised the risk of later obesity (odds ratio 1.58) (Luppino and colleagues, 2010) [Meta-analysis, observational studies].

A team around Kappelmann used genetic data. Genetic overlaps with depressive symptoms existed for CRP, similar to BMI. In Mendelian randomisation, however, genetically higher CRP showed no consistent associations for most individual symptoms, whereas a higher BMI was associated with joylessness, tiredness and changes in appetite [Mendelian randomisation].

One more piece of the puzzle: if inflammatory markers were merely a side effect of the episode, they would have to disappear with improvement. In 22 studies, medication did improve symptoms but did not lower TNF-alpha overall, while IL-1beta fell and IL-6 possibly did (Hannestad and colleagues, 2011) [Meta-analysis, before-after studies]. Antidepressants can therefore change inflammatory markers themselves, which makes it harder to interpret a value measured during treatment.

And now you know why nobody can seriously say whether, in your case, the inflammation came first or the depression.

The core: what anti-inflammatory treatments have shown in studies

If inflammation plays a part, then something against inflammation ought to lift mood too. That is exactly what was tested, with more contradictory results than it sounds online.

First of all: everything in this section is research. Neither the German National Care Guideline nor the Canadian CANMAT guideline of 2023 recommends anti-inflammatory drugs for the treatment of depression [Guideline]. Infliximab, minocycline and celecoxib are prescription-only.

Infliximab: a surprising pattern

Infliximab is an antibody against TNF-alpha, an immunosuppressant used in severe inflammatory diseases such as Crohn's disease. Because it dampens part of the body's defences, infections are among its known risks. It is not a treatment for depression.

Minocycline and celecoxib: the same pattern, then a setback

A team around Nettis gave 39 people with treatment-resistant depression and CRP of 1 mg/L or more the antibiotic minocycline or placebo for four weeks, in addition to their antidepressant. No difference overall. At CRP of 3 mg/L or more, however, the minocycline group showed the greatest improvement and a partial response in 83.3 percent, in very small subgroups [RCT].

For celecoxib, a painkiller from the COX-2 inhibitor group, early meta-analyses looked strongly positive. A team around Köhler found a benefit for anti-inflammatory treatments overall, in depression or depressive symptoms, across 14 studies with 6,262 participants, especially for celecoxib. But all studies had a high risk of bias, and the very high odds ratios for remission (7.89) and response (6.59) on celecoxib had very wide uncertainty ranges [Meta-analysis, RCTs].

Celecoxib and other NSAIDs come with known risks for gastrointestinal bleeding, the heart, circulation and the kidneys. The CANMAT guideline explicitly points out that SSRIs combined with NSAIDs such as ibuprofen are associated with an increased risk of gastrointestinal bleeding [Guideline].

What lies behind “52 percent fewer symptoms”

Online you can read that anti-inflammatory drugs reduce depressive symptoms by 52 percent and improve treatment success by 79 percent. The numbers come from a meta-analysis by Bai and colleagues with 30 RCTs and 1,610 participants, but they mean something different there: the response rate was 1.52 times and the remission rate 1.79 times that of the comparison groups. These are relative rates, not a percentage reduction in symptoms. In studies with women only there was no difference, and among the side effects, gastrointestinal events stood out [Meta-analysis, RCTs].

And something against inflammation as a precaution? In the ASPREE substudy, 19,114 healthy older people received low-dose aspirin or placebo for a median of 4.7 years. New depressive symptoms occurred equally often (hazard ratio 1.02). Aspirin did not prevent depression (Berk and colleagues, 2020) [RCT].

Omega-3 between meta-analysis, Cochrane and guideline

Omega-3 is where many people are most likely to take action themselves. That is why I deliberately set the data side by side. A meta-analysis of 26 RCTs with 2,160 participants found a small overall effect, more favourable for EPA-predominant than for DHA-predominant preparations (Liao and colleagues, 2019) [Meta-analysis, RCTs]. In a proof-of-concept analysis of a randomised trial with 155 people there was no difference overall, but with raised inflammatory markers there was more improvement on EPA than on placebo (Rapaport and colleagues, 2016) [RCT, secondary analysis].

As prevention, omega-3 gave no protection in the VITAL-DEP trial with 18,353 adults; the risk of depression was even slightly higher (hazard ratio 1.13). That is a different question from treatment and not directly comparable (Okereke and colleagues, 2021) [RCT].

The guidelines do not agree here. The German NVL strongly recommends not advising people with depression who have no proven deficiency to take dietary supplements, explicitly including omega-3 fatty acids. Its reasoning: uncertain benefit, costs, interactions and the risk that evidence-based treatments are pursued less consistently or rejected [Guideline]. CANMAT 2023 downgraded omega-3, previously second line for mild episodes, to third line because a Cochrane review found insufficient high-quality evidence [Guideline]. A WFSBP and CANMAT taskforce recommends omega-3 as an add-on to standard treatment, not as sole treatment [Guideline]. The ISNPR considers pure EPA or EPA-predominant preparations effective in major depression, but sees selection by inflammatory markers as a field of research [Guideline].

Over-the-counter preparations can also have side effects and interactions, and alongside an ongoing treatment they belong in a conversation with your doctor. More in Measuring the omega-3 index and ALA, EPA and DHA.

The subgroup idea, and why it remains open

From these puzzle pieces came the hypothesis of a subgroup with inflammatory depression: dampening inflammation might, if at all, make a difference only with raised inflammation, and without clearly raised inflammatory markers might even do worse than placebo.

Established Across all people with depression, anti-inflammatory agents show no reliable benefit. Plausible A signal with raised inflammation appears repeatedly. Open The stratified celecoxib trial was negative, and remissions did not become more frequent.

If psychotherapy and antidepressants have not been able to help enough, there are further, better-studied treatment options in medical hands. I describe one of them in Ketamine for treatment-resistant depression. What overall treatment of depression can look like from an integrative perspective is described in Treating depression holistically.

And now you know why “lowering inflammation” in depression is a research question today and not a therapy.

Where the inflammation may come from: causes worth looking at

Several sources can lie behind silent inflammation, often small ones at the same time, like dripping taps that together fill a bucket. One sentence first: none of these factors is a personal failing, and nobody could have reliably prevented depression with the right lifestyle. The list is a map, not a register of guilt.

• Visceral fat and metabolism. In 25 people with severe obesity, IL-6 in the portal vein, which drains abdominal fat, was about 50 percent higher than in an artery and was related to CRP (Fontana and colleagues, 2007) [Human study, invasive]. On insulin and inflammation: Insulin resistance and weight loss.
• Disturbed sleep. In 72 studies, sleep disturbances went along with slightly higher CRP and IL-6, whereas experimental sleep deprivation did not (Irwin and colleagues, 2016) [Meta-analysis, observational studies]. It is more about persistently disturbed sleep than about individual short nights. More in Sleep and depression and, with snoring and pauses in breathing, in Recognising sleep apnoea.
• Stress and strain. Acute laboratory stress raised IL-6 and TNF-alpha within minutes to hours, but not CRP (Marsland and colleagues, 2017) [Meta-analysis, experimental studies]. Slavich and Irwin describe how social threat and rejection could switch on inflammatory pathways [Review]. Adversity in childhood went along with small, consistent elevations of CRP, IL-6 and TNF-alpha in adulthood, a biological trace, not a fate (Baumeister and colleagues, 2016) [Meta-analysis, observational studies]. On the stress axis: Cortisol and the HPA axis.
• Lack of exercise. What is mainly established is the reverse: training can lower CRP, more on this below. On sitting: Sitting and lack of exercise.
• Smoking. In a cohort of 1,794 people with cardiovascular disease, quitting smoking was associated with a CRP 0.40 mg/L lower at follow-up after a median of 9.9 years (van 't Klooster and colleagues, 2020) [Cohort].
• Autoimmune diseases and severe infections. In a Danish registry of 3.56 million people, a hospital contact for an autoimmune disease raised the risk of a later mood disorder by 45 percent, and for an infection by 62 percent (Benros and colleagues, 2013) [Cohort]. More in Hashimoto's and the immune system and AIP in autoimmune diseases.
• Gum inflammation. In periodontitis, CRP was on average 1.56 mg/L higher than in controls and 0.50 mg/L lower after periodontal treatment (Paraskevas and colleagues, 2008) [Meta-analysis, observational studies]. Depression is associated with periodontitis (odds ratio 1.30), with very low to low certainty (Kloeckner and colleagues, 2026) [Meta-analysis, observational studies]. Both directions are conceivable, because depression also makes oral care harder.
• Gut-brain axis. The microbiome and the gut barrier are discussed as possible backgrounds of low-grade inflammation through which the immune system could connect gut and mood. The studies on this are in Depression and the gut-brain axis, for chronic inflammatory bowel disease in Crohn's disease and colitis from an integrative view, and on the gut barrier in Leaky gut and zonulin.

And now you know why the question “Where does my inflammation come from?” often gets you further than the question “How high is my CRP?”.

Lab values in depression: what hs-CRP can tell you and what it cannot

At some point the practical question comes up: which test, which values? The short answer: you can gather clues. Depression cannot be measured, and which treatment fits cannot be read from a blood value.

CRP and hs-CRP: where the 3 mg/L comes from

CRP and hs-CRP measure the same protein. The high-sensitivity method also measures very low concentrations accurately, and that is the range that matters in silent inflammation.

The thresholds of 1 and 3 mg/L used in the depression studies do not come from psychiatry but from a 2003 cardiology statement by the CDC and the American Heart Association on cardiovascular risk [Guideline]. According to summaries of that statement, two measurements taken about two weeks apart are averaged in metabolically stable people, and at values above 10 mg/L a source of infection or inflammation is looked for and the measurement is repeated.

I am deliberately not giving you a table of normal values: reference ranges depend on the laboratory, the method and the unit. In addition, acute infections, injuries, hospital stays, systemic inflammatory diseases and chronic infections can raise CRP considerably. What I observe clinically in addition are spikes after dental treatment or very intensive training; that is experience and not a study finding. A single value is therefore a snapshot, not a characteristic.

Which values can be useful in a work-up

The NVL calls for a careful history that includes physical illnesses, a review of medications that can be associated with depressive symptoms, and, if a physical cause is suspected, further diagnostics, weighed together and with a view to possible therapeutic consequences. Overdiagnosis is to be avoided [Guideline]. The NVL does not mention inflammatory markers.

What I observe clinically is that when inflammatory or metabolic involvement is suspected, it can be worth looking at hs-CRP, blood count, blood sugar and lipid values, thyroid and ferritin. Ferritin rises along with inflammation as an acute-phase protein and can mask iron deficiency; more on this in Iron and inflammation and Iron deficiency and the mind. Which values make sense for you is decided in a conversation with your doctor.

hs-CRP and response to antidepressants

In the CO-MED trial with 106 people, the estimated remission rate with allocation by CRP was 53.1 percent compared with an actual 41.5 percent, a model calculation (Jha and colleagues, 2017) [RCT, secondary analysis]. The counterweight: CANMAT 2023 names the association between CRP and poorer response to SSRIs, but explicitly does not recommend the measurement for routine selection of an antidepressant because the effects are small [Guideline].

And now you know why “crp level depression” has no number as its answer, but an interpretation.

Lifestyle levers that can lower inflammatory markers, and their honest limits

Perhaps you are now wondering: what can I contribute myself, naturally and in addition to my treatment? The answer needs a clean separation.

Established: these levers can lower inflammatory markers

In 17 studies lasting at least twelve weeks, a Mediterranean diet lowered hs-CRP by an average of 0.98 mg/L and also lowered IL-6, with large differences between studies (Schwingshackl and colleagues, 2014) [Meta-analysis, RCTs]. In 123 older people with chronic insomnia, cognitive behavioural therapy was associated after 16 months with a lower risk of CRP above 3 mg/L than a sleep seminar (odds ratio 0.26, wide uncertainty range), and it improved depressive symptoms more (Irwin and colleagues, 2014) [RCT].

Further reading: Exercise as medicine, CBT-I and sleep restriction, Pro-inflammatory foods and, because it is mentioned so often, Turmeric and inflammation.

Not established: that they improve depression by this route

That these levers improve depression by lowering inflammation has not been shown. Independently of that, the NVL strongly recommends motivating and supporting people without contraindications to take up physical activity, ideally in a group, and structured, supervised training, and just as strongly recommends encouraging a healthy diet [Guideline]. More in Lifestyle as therapy and Ketogenic diet and mental illness.

In the end, this is not about a lower CRP. It is about a morning feeling full of possibilities again. Joy is not a luxury. It carries a day.

And now you know why, in depression, I do not say “lower inflammation” but “look more closely, and do so in addition”.

Frequently asked questions about depression and inflammation

Where this article connects