Autoimmune Disease and Fertility: What Is Actually Proven

For most autoimmune conditions, no — not directly. The things that reliably get in the way are the medicines used to control the disease, the disease being active rather than quiet, and, in two specific cases, an effect on pregnancy itself. Those two are antiphospholipid syndrome and thyroid disease where the thyroid is genuinely underactive.

What does not have evidence behind it is the broader idea of "immune infertility" as a diagnosis you can test for and treat. That distinction is the whole point of this page, because the tests and drugs sold on the back of it are expensive, are widely offered in India, and are specifically advised against by the European guideline that reviewed them.

Antiphospholipid syndrome is an autoimmune clotting disorder, and it is associated with pregnancy loss. It is also the only autoimmune condition on this page with a treatment that improved live births in pooled trial data.

Cochrane reviewed 11 studies in 1,672 women with persistent antiphospholipid antibodies and recurrent pregnancy loss. Adding heparin to low-dose aspirin raised the live birth rate by roughly a quarter in relative terms and roughly halved pregnancy loss, on low-certainty evidence, with the effect driven largely by one big trial. Aspirin on its own, against placebo, was tested in a single 40-woman trial and the result is too imprecise to say anything.

The diagnosis is where mistakes happen. Under the 2023 ACR/EULAR classification criteria, one positive antibody test is not a diagnosis. There is an entry requirement of at least one positive antiphospholipid antibody test within three years of a matching clinical event, and then a weighted score across six clinical domains and two laboratory domains; a patient needs at least three points from the clinical side and three from the laboratory side. In the validation cohort these criteria reached a specificity of 99% against 86% for the older 2006 criteria, at the cost of sensitivity, 84% against 99%. In plain terms: they are deliberately hard to meet, so that people without the condition are not treated for it.

If thyroid function is abnormal, it is treated. That is uncontroversial. The contested question — and the one Indian patients are most often billed for — is whether levothyroxine helps when the thyroid is working normally and only the antibodies are positive.

Two large randomised trials say it does not. TABLET screened nearly 20,000 women across 49 UK hospitals and randomised 952 who were euthyroid, had thyroid peroxidase antibodies and had a history of miscarriage or infertility. Live birth after 34 weeks was 37.4% on levothyroxine and 37.9% on placebo — an absolute difference of 0.4 percentage points in the wrong direction. POSTAL randomised 600 euthyroid antibody-positive women having IVF in Beijing and found miscarriage, clinical pregnancy and live birth all effectively unchanged.

Cochrane's 2019 review draws the line that matters. In women with subclinical hypothyroidism — a raised TSH — thyroxine may improve live birth, on one small trial of 64 women. In women with normal thyroid function and antibodies only, live birth was essentially identical across two trials in 686 women. ESHRE's recurrent pregnancy loss guideline carries this as a strong recommendation: euthyroid women with thyroid antibodies and recurrent loss should not be treated with levothyroxine. If you have been prescribed it on an antibody result alone, that is the sentence to take back to the consultation — alongside your actual TSH and free T4 values.

Context for why this matters here. In a cross-sectional study across eight Indian cities, anti-TPO antibodies were present in 21.85% of 5,360 participants, hypothyroidism in 10.95% of participants and subclinical hypothyroidism in 8.02%. A positive antibody test is a common finding in the general population, not a rare clue.

For systemic lupus, rheumatoid arthritis and the other rheumatic diseases, the reproductive problem is usually practical rather than immunological. The 2020 American College of Rheumatology reproductive health guideline — 131 graded recommendations and 12 good practice statements — is built around two principles: conceive while the disease is quiet, and conceive while on medicines compatible with pregnancy.

What that means in a consultation is a medication review well before you start trying, not at the first positive test. Some disease-modifying drugs have to be stopped and replaced months ahead; some gonadotoxic treatments raise the question of fertility preservation before they are started. The guideline itself notes that many of its recommendations are conditional because the evidence specific to these populations is thin — which is a reason to plan with a rheumatologist and an obstetric team together rather than to assume either the best or the worst.

This is the most useful section on the page. Every item below is offered commercially in India, and every item below is one that a major guideline has looked at and declined to recommend.

• Natural killer cell testing. ESHRE's 2023 recurrent implantation failure recommendations state plainly that peripheral NK cell testing is not recommended and that uterine NK cell testing is not recommended. Uterine T lymphocyte assessment and blood cytokine levels are in the same category.
• HLA matching between partners. Assessing HLA-C compatibility is not recommended for implantation failure, and HLA determination is not recommended in clinical practice for recurrent loss — the RPL guideline keeps only one narrow prognostic exception in a specific Scandinavian subgroup, on low-quality evidence.
• Intralipid infusion. Not recommended by ESHRE. A 2021 meta-analysis of 5 trials in 843 women reported a live birth risk ratio of 1.83 (95% CI 1.42-2.35), which sounds conclusive, but it pooled saline-controlled trials with unblinded no-treatment trials, rated overall risk of bias moderate, and its own authors concluded that the evidence is limited and more research is needed before adopting it. One randomised trial of empiric intralipid in this setting has since been retracted — and its results were null even as published.
• Immunoglobulin and lymphocyte immunisation. Cochrane reviewed 20 trials of immunotherapy for recurrent miscarriage and found no significant live-birth benefit for paternal cell immunisation, third-party donor leucocytes, trophoblast membranes or intravenous immunoglobulin. ESHRE's RPL guideline retains only a conditional statement, on low-quality evidence, that repeated high-dose immunoglobulin given very early in pregnancy may help women with four or more unexplained losses. That is a narrow, hedged exception — not a general indication.
• G-CSF. Not recommended for recurrent implantation failure, whether given into the uterus or under the skin.
• Anti-ovarian antibodies. The 2024 premature ovarian insufficiency guideline states that screening for anti-ovarian autoantibodies should not be used to diagnose autoimmune POI (strong recommendation, very low-quality evidence).
• Routine thyroid antibody screening as a test for autoimmune POI. Also advised against, because positive TPO antibodies are common in the general population. The same guideline recommends 21-hydroxylase autoantibodies instead in POI of unknown cause, because a positive result there points at adrenal insufficiency and needs endocrine referral.
• That endometriosis is an autoimmune disease. It is not classified as one in any source used here, and treating it as one is how immune therapy gets offered for it.

One structural point worth knowing: ESHRE's RPL guideline rests on 62 evidence-based recommendations of which only 12 are supported by moderate-quality evidence — the remaining 50 rest on low or very low. Where a guideline says "do not offer", it is usually saying the benefit has not been shown, not that harm has been proven. That is still the right basis for declining to pay for it. The same logic applies to inflammatory marker testing.

Short list, and the trigger matters as much as the test.

1. Thyroid function — TSH and free T4 — before starting treatment. Antibodies are a secondary question and, on their own, do not change management.
2. Antiphospholipid antibodies after recurrent pregnancy loss: lupus anticoagulant, anticardiolipin and anti-beta-2 glycoprotein I, repeated on a second occasion. One positive result on one day is not a diagnosis.
3. For implantation failure specifically, ESHRE suggests antiphospholipid testing where there are additional risk factors for thrombophilia, and says it can be considered without them. It does not suggest the immune panel.
4. In premature ovarian insufficiency of unknown cause: 21-hydroxylase autoantibodies, with endocrine referral if positive.
5. If you have a diagnosed rheumatic disease: a pre-pregnancy medication review, and an honest assessment of whether the disease is currently quiet.

It is worth saying that being handed a long immune panel can feel like finally being taken seriously after a run of losses, and being told the panel is not indicated can feel like being dismissed. Those are different things. Declining an unproven test is not the same as having no explanation — a proportion of recurrent loss genuinely has no identified cause, and ESHRE's guideline says so rather than filling the gap. Recurrent pregnancy loss and recurrent implantation failure are investigated along different lines, and the thresholds for starting are not the same.