Raynaud Phenomenon: Primary versus Secondary

Look at both hands and answer four questions. Do they both do the same thing? Did it start before you were 40? Have your fingers ever gone puffy or swollen? Is there a sore on a fingertip that will not heal?

Symmetrical, started early, no puffiness and no sores is the harmless version, and that describes most people who have this. Any other pattern is worth a doctor's appointment, not a panic.

Takes 30 seconds. Nothing to buy, nothing to prepare.

Go today, do not wait: if a finger or toe stays white or blue, is painful, and will not return to normal colour after warming, or if there is an ulcer or sore on a fingertip that will not heal. Those are the two findings that never belong to the harmless version.

The Verdict

Fingers going white in the cold is usually harmless. Four questions tell you when it is not.

Think of the small blood vessels in your fingers as taps. In the common version the tap is oversensitive and slams shut in the cold, but the pipe behind it is perfectly healthy, so everything opens up again and no damage is done. In the less common version the pipes themselves are narrowed and scarred, and that is why a doctor looks at the tiny vessels at the base of your fingernail under magnification. It is the one place in the body where you can see the pipes without cutting anything open.

Here is what is really happening: the small vessels in your fingers overreact to cold and shut down, and in most people the plumbing behind them is completely normal, which is why roughly 87 in 100 people diagnosed with the straightforward version never develop anything else.
What most people get wrong: a positive autoimmune blood test on its own means very little here, because 68 percent of people with these symptoms at specialist clinics test positive, and it is the blood test and the nailfold check together that change the picture.
What to watch for: fingers that go puffy or swollen, only one hand affected, symptoms starting after age 40, or a sore on a fingertip that will not heal.

Best for

Anyone whose fingers change colour in the cold and wants to know whether it needs investigating, and anyone using vibrating tools at work.

Skip if

You already have a diagnosed connective tissue disease, or you have an ulcer on a finger. Both need your specialist rather than a self-check.

Want the full evidence? Keep scrolling

What Works

Dark cinematic study of hand and digital circulation

The honest framing first. The interventions with the strongest evidence here are not delivered by a physical therapist. The highest-value action in the whole condition is working out which version you have and referring accordingly.

Order both tests, never just one HIGH

The antibody blood test and nailfold capillaroscopy together carry a relative risk of 16.96 for going on to develop a connective tissue disease. Either one alone carries 7.63 and 5.53. The combination is the referral trigger, not either half.

Evidence: STRONG. Systematic review and meta-analysis of 7 studies, 4,051 patients (Ingegnoli 2017), supported by a consensus of 110 experts across 85 European centres (Avouac 2011).

Find and remove a cause, if there is one HIGH

Vibrating tool use at work raises the odds of this condition roughly seven-fold. Chemotherapy drugs, beta-blockers and a newer class of cancer drugs can all cause it outright. Both are found by asking, not testing.

Evidence: STRONG for the association. Meta-analysis of 52 studies (Nilsson 2017); systematic review identifying 12 drug classes (Khouri 2016).

Calcium channel blockers, if a doctor prescribes them MODERATE

The first-line drug, and weaker than its reputation. Expect roughly 1.72 fewer attacks per week, with high-quality evidence that it does little for how severe each attack is.

Evidence: MODERATE. Cochrane review, 7 trials, 296 participants (Ennis 2016).

See the rest of the treatment hierarchy (Tier 2 and 3)

Topical nitrate ointment MODERATE
The largest pooled effect of any drug here, and suited to as-needed use because attacks are unpredictable. Effect size 0.70 overall, larger in secondary disease at 0.95. Meta-analysis of 7 trials, 346 patients (Curtiss 2018).

PDE-5 inhibitors, for secondary disease MODERATE
Cut about half an attack per day and roughly 15 minutes of daily attack time. Six double-blind trials (Roustit 2013). Specialist-led.

Prostanoids such as iloprost MODERATE
Alongside calcium channel blockers, one of only two interventions with clear favourable evidence in secondary disease (Huisstede 2011). Hospital-delivered, and covered by the 2024 national guideline for systemic sclerosis (Denton 2024).

Botulinum toxin injection EMERGING
For refractory digital ischaemia only, in specialist hands. Pooled pain reduction across 13 studies, with injection-site pain and hand weakness the commonest complications (Zhou 2023).

CO2 hand baths LOW
One trial, 24 patients, single session, measuring vessel diameter rather than symptoms. Genuinely interesting, not yet care (Schulz 2026).

Exercise Prescription

There is no exercise programme proven to treat this condition. No trial establishes a dose, a frequency or a progression. Anyone offering you a proven exercise cure for this is overselling it, and saying so is more useful than inventing a protocol to fill the gap.

What follows is the daily self-management routine that clinicians have recommended for decades based on how the condition behaves. It is worth doing. It is labelled honestly as practice-based rather than trial-proven. NO TRIAL EVIDENCE

What to do	How	How often
Warm your core, not just your hands	A warm layer on your body. Your body pulls blood away from the hands to protect your core, so warm hands start at the torso	Any cold day
Gloves on before you go out	Put them on while your hands are still warm indoors. Once an attack starts, gloves only trap cold hands	Every cold-weather outing
Carry a hand warmer	One in each pocket, rechargeable or disposable	Cold months
Warm water if an attack starts	Warm, not hot, running over the hands while you move the fingers gently	As needed
Keep a trigger diary	Date, what you were doing, how long it lasted. Most people find two or three triggers they can plan around	Daily for 4 weeks
Stop smoking, if you smoke	Ask your doctor about cessation support	Ongoing

What Doesn't Work

Oral vasodilator drugs other than calcium channel blockers. No evidence of effect across 8 studies and 290 participants, and one of them, enalapril, was associated with a small increase in attacks (Stewart 2012).
Acupuncture. The same meta-analysis that produced a positive headline result found nothing in its own network analysis, and rates its evidence very low quality (Zhou 2023).
Biofeedback. Seven of the eight trials in the behaviour-change review were biofeedback and the newest was published in 2002. It persists because it is intuitive and harmless, not because it works (Daniels 2018).
Any exercise programme sold as treatment for this condition. There is no trial establishing one.

Red Flags

Refer, do not treat

Dark cinematic study of compromised digital circulation

An ulcer, pitting scar or tissue loss on a finger. This never belongs to the harmless version. Refer urgently.
A painful, persistently discoloured digit that does not reperfuse on warming. Critical loss of blood supply. Same-day medical referral.
Asymmetric or one-sided attacks. Points to a structural or clot-related cause rather than spasm. Vascular referral.
Onset after age 40, or onset in a young child. Both sit outside the primary pattern.
Puffy fingers alongside a positive antibody test. The very-early systemic sclerosis red flag, carrying odds of 15.45 for progression. Rheumatology referral.
Skin tightening, telangiectasia, or new difficulty swallowing or breathlessness. Prompt rheumatology assessment.
New symptoms after starting chemotherapy, a beta-blocker or a tyrosine kinase inhibitor. Back to the prescriber.

Refer to: Rheumatology for suspected connective tissue causes. Vascular surgery or emergency care for critical ischaemia, non-healing ulceration or one-sided attacks. GP for baseline antibody testing, nailfold capillaroscopy and medication review. Occupational health where vibration exposure is the likely cause, because there the exposure is the treatment target.

Return to Training

This is a clearance checklist rather than a healing timeline. Primary Raynaud phenomenon places no restriction on training load at all. There is no injured tissue and nothing to heal, so reducing volume would be an intervention with nothing behind it. The one modification is subtractive.

Red flags screened and absent: no ulceration, no pitting scars, no non-reperfusing digit
Attacks confirmed symmetrical and affecting both hands
Secondary features screened and absent: no puffy fingers, no skin tightening, no telangiectasia
Occupational vibration exposure identified and addressed, or confirmed absent
Medication review completed, or confirmed not applicable
All deliberate cold exposure stopped: ice baths, cold plunges, whole-body cryotherapy
Patient can state their own red flags and knows when to seek same-day help

Conviction

MODERATE

Endpoint-stratified rather than a single grade, because the evidence is genuinely uneven across this condition:

Recognition and referral pathway HIGH
Low transition rate for isolated primary disease HIGH
Occupational vibration as a cause HIGH
Drug effect sizes MODERATE
Any physical therapy intervention, exercise dose or behavioural programme NO EVIDENCE
CO2 hand baths LOW

What would change my mind: the physical therapy claim

A multicentre, assessor-blinded trial of at least 200 adults with primary Raynaud phenomenon, comparing a structured self-management programme against usual advice across a minimum of two winters, with attack frequency and Raynaud's Condition Score as primary outcomes and adherence measured rather than assumed. A sustained between-group difference would give physical therapy its first real evidence base here. It would also be the first such trial since 2002.

What would change my mind: the referral pathway

A prospective, unselected primary-care cohort of 1,000 or more people presenting with new cold-induced colour change, all receiving antibody testing and nailfold capillaroscopy at baseline and followed for 5 years, reporting predictive values at a primary-care base rate. Every prediction figure on this page comes from specialist referral clinics, and that one study would tell us whether the strongest combination survives outside the clinic that produced it.

The Full Picture — Anatomy, Diagnosis & Evidence

What's Actually Going On

Dark cinematic anatomical study of digital arteries and microcirculation

Raynaud phenomenon is an exaggerated clamping-down of the small arteries in the fingers and toes in response to cold or emotional stress. The three-colour sequence follows the blood: white as the vessels shut, blue as oxygen-poor blood pools, then red and painful as flow floods back.

In the primary version the vessels are structurally normal. The problem is purely functional and fully reversible, with no fixed narrowing and no tissue damage.

In the secondary version the same spasm sits on top of real structural damage to the tiny vessels. That is why looking at the nailfold matters: it is direct visual evidence of the pipes themselves. The damage progresses through three recognisable stages, from enlarged and giant capillaries with small haemorrhages early on, to capillary loss and disorganisation late (Cutolo 2004, 241 patients).

Secondary causes fall into four groups, and three of the four are found by asking rather than testing: connective tissue disease (above all systemic sclerosis), occupational vibration exposure, drugs, and structural or clot-related causes behind one-sided attacks.

How to Identify It

Dark cinematic study of hand examination and nailfold vasculature

Nailfold capillaroscopy Sn: data unavailable | Sp: data unavailable
Reported in the literature as a predictive value rather than an accuracy pair. An abnormal nailfold pattern was the best single baseline predictor of transition, at a positive predictive value of 47% (Spencer-Green 1998).

Antinuclear antibodies (ANA) Sn: data unavailable | Sp: data unavailable
Positive predictive value 30% (Spencer-Green 1998). In the largest cohort of Raynaud patients at specialist centres, 68% were antibody positive (Minier 2014, 469 patients), which is precisely why a positive result on its own changes so little.

Both together RR 16.96 (95% CI 6.61–43.55)
The only figure here that justifies a referral decision on its own, against 7.63 for antibodies alone and 5.53 for capillaroscopy alone (Ingegnoli 2017).

Puffy finger observation OR 15.45 with a positive ANA
Observed rather than asked about, and the highest-value physical sign in the condition. Adding capillaroscopy or a specific antibody raises the odds to 19.52 (Siqueira 2022, 217 patients).

Why those first two rows say "data unavailable", stated rather than hidden. Across 85 papers retrieved for this condition, no bedside physical examination test carries a published sensitivity and specificity pair. The diagnosis is clinical in every source. The numbers that do exist are predictive values and relative risks from transition cohorts, and they are shown here in their real form rather than converted into accuracy figures no study ever measured. A number developed in a different study design would be an invented number here.

The Debate

Is the first-line drug actually any good?

Rirash 2017, Cochrane, 38 trials, 982 participants

Calcium channel blockers cut attacks by 6.13 per week against 13.7 on placebo.

VS

Ennis 2016, Cochrane, 7 trials, 296 participants

"Minimally effective." 1.72 fewer attacks per week, with high-quality evidence of little effect on severity.

Population and outlier handling, not drug effect. Rirash pools primary, secondary and mixed patients with very high statistical heterogeneity, and removing a single trial more than halves its estimate to 2.93. Follow Ennis for primary disease and set expectations at one or two fewer attacks per week.

Does warming the hands actually open the vessels?

Universal clinical advice, every patient leaflet

Keep the hands warm. Warmth is the foundation of self-management.

VS

Schulz 2026, randomised trial, 24 patients

Warm water at 40–42°C produced no significant vessel widening in systemic sclerosis patients, while a CO2 bath at 35°C did.

Keep advising warmth. One trial, 24 patients, a single session, measuring vessel diameter rather than attacks. It does not overturn the advice. It does show that the assumed mechanism is unproven in exactly the patients with the worst disease, which is the most interesting open question in this condition.

Honest Limitations

The prediction figures come from specialist clinics

The research finding: antibodies plus abnormal capillaroscopy carry a relative risk of 16.96, and 68% of Raynaud patients at specialist centres are antibody positive.

The real-world gap: those groups are referred populations, enriched many times over for disease. Around 4.85% of the general population has primary Raynaud phenomenon and the overwhelming majority never see a rheumatologist.

Clinical adjustment: screen on the same features, but quote absolute numbers to the patient. Relative risk decides who gets referred. Absolute risk is what you say out loud.

The interventions physical therapy owns have no evidence base

The research finding: 8 trials of behaviour change interventions, all published between 1978 and 2002, verdict "no evidence to support or refute".

The real-world gap: the advice every patient receives has never been tested to modern standards, and of 5 reviews and 19 trials for secondary disease, all but two concerned drugs.

Clinical adjustment: give the advice, label it accurately, and resist manufacturing a rehabilitation programme to fill the space.

The drug trials are short and the condition is seasonal

The research finding: the 38 trials in the largest review averaged 7.4 weeks, and 33 of 38 were cross-over designs.

The real-world gap: this condition is defined by cold exposure. A 7-week trial cannot cleanly separate a treatment effect from the weather.

Clinical adjustment: judge treatment across a season, not a fortnight, and expect winter deterioration that is not treatment failure.

The Nuance

Dark cinematic study contrasting healthy and compromised digital circulation

The two headline numbers point in opposite directions and both are true. Isolated primary Raynaud phenomenon converts to a connective tissue disease at just 2.65 per 100 person-years, and only 12.6% of patients transitioned across 2,531 patient-years of follow-up. That is a reassuring condition. Yet the right combination of findings carries a relative risk near 17. Both hold because the underlying rate is low: even a strong relative risk produces a modest chance for any individual. The best single baseline predictor still had a positive predictive value of only 47%, meaning it was wrong more often than not.

One study looks like it contradicts the others and does not. Ingegnoli found abnormal capillaroscopy without antibodies carried a relative risk of 5.53. Siqueira found Raynaud plus an abnormal nailfold pattern, without any other clinical sign, gave odds of 0.03 for progression. Read carelessly that says capillaroscopy is protective. It does not. Siqueira's cohort was built so that every patient already had at least one systemic sclerosis feature, so the comparison runs against other high-risk combinations rather than against healthy people. The real finding is that once a clinical sign is present, capillaroscopy adds the least of any combination.

Surgery barely enters the picture. Digital sympathectomy and botulinum toxin appear in current guidance only for severe refractory ischaemia. No comparative success rate appeared in the evidence retrieved. The closest thing to a conservative success rate is the natural history itself: 87.4% of people diagnosed with primary Raynaud phenomenon developed no secondary disorder at all.

And one occupational detail that reverses the usual order. In vibration-exposed workers, nerve injury appears at roughly a third of the exposure time needed to produce the colour change. The numbness usually arrives before the white fingers, not after, so waiting for the classic sign means waiting past the point where the more disabling injury has already started.

Sources

Ingegnoli F, et al. (2017). Outcomes, rates and predictors of transition of isolated Raynaud's phenomenon. Swiss Medical Weekly. PMID 28975961. Meta-analysis, 7 studies, 4,051 patients. Transition 2.65/100 person-years; both tests abnormal RR 16.96.
Spencer-Green G (1998). Outcomes in primary Raynaud phenomenon. Archives of Internal Medicine. PMID 9521223. Meta-analysis, 639 patients, 2,531 patient-years. 12.6% transitioned; capillaroscopy PPV 47%, ANA PPV 30%.
Garner R, et al. (2015). Prevalence, risk factors and associations of primary Raynaud's phenomenon. BMJ Open. PMID 25776043. Meta-analysis, 33 studies, 33,733 participants. Prevalence 4.85%.
Minier T, et al. (2014). VEDOSS EUSTAR multicentre study: puffy fingers as a pivotal sign. Annals of the Rheumatic Diseases. PMID 23940211. 469 Raynaud patients; 68% ANA positive.
Siqueira VS, et al. (2022). Predictors of progression to systemic sclerosis. Rheumatology (Oxford). PMID 35020814. 217 patients, median 4-year follow-up. Puffy fingers plus ANA, OR 15.45.
Avouac J, et al. (2011). Preliminary criteria for the very early diagnosis of systemic sclerosis. Annals of the Rheumatic Diseases. PMID 21081523. Delphi consensus, 110 experts, 85 centres.
Cutolo M, et al. (2004). Nailfold videocapillaroscopic patterns and serum autoantibodies. Rheumatology (Oxford). PMID 15026581. 241 patients; three progressive patterns established.
Ennis H, et al. (2016). Calcium channel blockers for primary Raynaud's phenomenon. Cochrane Database of Systematic Reviews. PMID 26914257. 7 trials, 296 participants. Minimally effective.
Rirash F, et al. (2017). Calcium channel blockers for primary and secondary Raynaud's phenomenon. Cochrane Database of Systematic Reviews. PMID 29237099. 38 trials, 982 participants.
Curtiss P, et al. (2018). Topical nitrates in the treatment of Raynaud's phenomenon. Journal of the American Academy of Dermatology. PMID 29408338. 7 trials, 346 patients.
Roustit M, et al. (2013). Phosphodiesterase-5 inhibitors for secondary Raynaud's phenomenon. Annals of the Rheumatic Diseases. PMID 23426043. 6 double-blind trials.
Stewart M, Morling JR (2012). Oral vasodilators for primary Raynaud's phenomenon. Cochrane Database of Systematic Reviews. PMID 22786498. 8 studies, 290 participants. No evidence of effect.
Daniels J, et al. (2018). Behaviour change interventions for the management of Raynaud's phenomenon. BMJ Open. PMID 30552281. 8 trials, 495 participants, published 1978–2002.
Huisstede BM, et al. (2011). Effectiveness of interventions for secondary Raynaud's phenomenon. Archives of Physical Medicine and Rehabilitation. PMID 21704799.
Nilsson T, et al. (2017). Hand-arm vibration and the risk of vascular and neurological diseases. PLoS One. PMID 28704466. 52 studies. Raynaud OR 6.9.
Khouri C, et al. (2016). Drug-induced Raynaud's phenomenon. British Journal of Clinical Pharmacology. PMID 26949933. 12 drug classes identified.
Zhou Y, et al. (2023). Botulinum toxins for the treatment of Raynaud phenomenon. Journal of Clinical Rheumatology. PMID 37011178. 13 studies.
Zhou F, et al. (2023). Acupuncture in patients with Raynaud's syndrome. Acupuncture in Medicine. PMID 35608095. 6 trials, 272 participants. Network analysis null.
Denton CP, et al. (2024). The 2024 British Society for Rheumatology guideline for management of systemic sclerosis. Rheumatology (Oxford). PMID 39255973. Current national guidance.
Schulz N, et al. (2026). Carbon dioxide vs. warm-water hand baths in systemic sclerosis. Microvascular Research. PMID 41724373. Randomised trial, 24 patients plus 12 controls.

Go Deeper

Most health advice never tells you how strong the evidence actually is, or where it runs out. The Verdict does, one topic a week, free.

Join The Verdict

Fingers going white in the cold is usually harmless.

Raynaud Phenomenon