Clinical synthesis

Can chest pain be purely of chest wall origin?

In this chapter: the real frequency of each cause in primary care, what points to the heart, the lung, the aorta or the digestive tract, the numerical value of each clinical feature, the decision tree, and why pain reproduced on palpation never closes the file.

Yes, and it is even the commonest situation. But "chest wall" is a conclusion, not a way in. The patient who walks into the practice with a hand on the sternum does not have a diagnosis yet: he has chest pain, and a worry. The sequence that follows therefore places triage before description, because that is the order in which the clinician must think, and because that is the order in which the patient needs to be answered.

What lies behind chest pain in primary care?

The question has been addressed by a systematic review with meta-analysis: Haasenritter and his Marburg team retained 11 studies representing about 6,500 patients, of which 6 studies at low risk of bias totalling nearly 3,900 patients1. Heterogeneity between studies being too great for a single figure, the authors deliberately presented ranges rather than a pooled estimate. It is a methodological honesty that should be carried over: anyone quoting "30 % of chest wall pain" without an interval is quoting a figure the literature does not support.

A Swiss cohort run by Verdon with 59 general practitioners gives the same order of magnitude with a different methodology: of 24,620 patients examined, 672 (2.7 %) had chest pain. After twelve months of follow-up, the distribution came out as 49 % musculoskeletal causes, 16 % cardiovascular, 11 % psychogenic, 10 % respiratory, 8 % digestive. The three commonest diagnoses were chest wall syndrome (43 %), coronary artery disease (12 %) and anxiety (7 %). Unstable angina, infarction and pulmonary embolism together accounted for 1.8 % of cases2.

But the same study carries a second figure, which is almost always left out when the first is quoted: potentially serious conditions (cardiac, respiratory and neoplastic) accounted for 20 % of cases. And twenty-five patients died during follow-up, twelve of them from a cause directly related to their chest pain: seven cancers and five cardiac causes. One consultation in five for chest pain in general practice therefore concerns a disease that matters. That is not an argument for alarmism, it is an argument for method.

A third source, from the same Marburg team but published a year earlier, completes the framing with the denominator. Of all the patients consulting their general practitioner, 0.7 % do so for chest pain; 55.9 % of these patients are women, the mean age is 59 years, and 29.6 % consult for acute pain of less than 48 hours. The aetiological distribution there is identical: 46.6 % of chest wall origin, 11.1 % stable ischaemic heart disease, 9.5 % psychogenic causes and 3.6 % acute coronary syndromes3. Three independent cohorts, three countries, one and the same order of magnitude: that is what makes it legitimate to rely on them.

What distinguishes chest wall pain from cardiac, pulmonary or aortic pain?

The table below gathers what the sources cited establish about each broad family. It is read in a precise direction: one does not first look for what argues for the chest wall, one first looks for what would argue against it. A single item from the red column is enough to suspend chest wall reasoning, whatever the number of items in the purple column.

Two rows of this table deserve comment, because they contradict widespread shortcuts.

Site sorts nothing out. It is readily taught that left-sided lateral pain argues for the chest wall and that retrosternal pain argues for the heart. Bösner's German cohort, on 1,212 patients, shows the opposite: among the chest wall syndromes, the pain was retrosternal in 52.0 % of cases and on the left in 69.2 %4. Left retrosternal pain is therefore perfectly compatible with a chest wall origin, and chest wall pain is not marked out by its side.

The relationship with exertion does not sort things out in the expected direction either. Still in Bösner's cohort, 72.3 % of the patients with chest wall syndrome reported no temporal association of their pain with anything at all, neither exertion, nor meals, nor position. The absence of a triggering factor is therefore a frequent feature of chest wall pain, not an argument against it. Conversely, pain reproducibly triggered by exertion and settling with rest within a few minutes remains one of the few features whose likelihood ratio markedly increases the probability of coronary disease.

What is each clinical feature worth, in figures?

The review by Swap and Nagurney, published in the JAMA in 2005, remains the reference synthesis on the value of the history in chest pain. It establishes that four characteristics lower the probability of acute coronary syndrome with a likelihood ratio of 0.2 to 0.3: stabbing pain, pleuritic pain, positional pain, and pain reproducible on palpation6. Conversely, radiation to one or both shoulders or to the arms, or triggering by exertion, carry ratios of 2.3 to 4.7.

The authors' conclusion is reproduced here without softening, because it is the foundation of this whole chapter: "although some elements of the chest pain history are associated with an increased or decreased probability of acute coronary syndrome, none of them, alone or in combination, identifies a group of patients who can be sent home without further investigation".

The practical reading of this chart fits into one sentence. In a patient whose initial probability of coronary syndrome is 10 %, which is the order of magnitude observed in emergency departments according to Fanaroff9, pain reproduced on palpation brings that probability down to around 2 to 3 %. That is low. It is not zero. Out of a hundred patients, two or three still have a coronary syndrome.

Should one be more wary in women?

One should be equally wary, but not for the reason people think. The narrative of the "atypical female presentation" is firmly established, and it is largely refuted by recent data. Ferry and his team analysed the symptoms reported by 1,941 patients presenting to the emergency department with suspected coronary syndrome, using a high-sensitivity troponin assay and sex-specific thresholds. The result: chest pain was the presenting symptom in 91 % of men and 92 % of women. And typical symptoms were more frequent in the women who had an infarction than in the men (77 % against 59 %, p = 0.007)10.

The real issue documented by this work lies elsewhere: on moving to a high-sensitivity assay with sex-specific thresholds, 30 % of the women diagnosed (27 out of 90) were patients whom the conventional single-threshold assay had missed, against 5 % of the men. The risk is therefore not that women present differently, it is that they are measured against the wrong threshold. For the physiotherapist, the consequence is simple: do not play down typical chest pain in a woman on the grounds that it "does not look like a textbook heart attack", and do not turn it into an extra reason to worry either.

Which decision tree should be followed in practice?

The tree below formalises the sequence. It is written for the physiotherapist who sees a patient by direct access or on referral, and whose responsibility is not to make a cardiological diagnosis but to avoid treating what must not be treated. It follows the logic of the joint 2021 AHA/ACC guideline on the evaluation of chest pain13, transposed to the setting of private practice.

What do we really know about costochondritis?

In this chapter: what the literature establishes and what it leaves in the dark, the anatomy of the anterior chest wall, what the only available prospective series really measures, the patient profile, and a prognosis appreciably less brief than the one usually announced.

One has to begin with what the literature does not say, because that is unusual and because it is decisive for the way it is discussed with the patient. Costochondritis is a condition of unknown aetiology. No inflammation has been demonstrated in it, despite the suffix. No pathophysiological model is established. No controlled trial has evaluated its treatment, and the reference clinical review in American general practice puts it bluntly: "Clinical trials of treatment are lacking"8. And there is no published population prevalence.

This void does not prevent work. It only requires carefully distinguishing what is measured from what is assumed, and never offering the patient a mechanistic explanation invented to fill the silence of the data. Saying "we do not know exactly why, but we do know that it is the commonest situation and that it is not your heart" is more accurate (and, in practice, more reassuring) than improvising a story about cartilage inflammation.

Where exactly does the pain sit?

Costochondritis affects the costochondral junctions, that is the zone where the bone of the rib becomes cartilage, and the chondrosternal joints, where the cartilage meets the sternum. The descriptions converge on the 2^nd–5^th joints12. Involvement is frequently multiple, which already distinguishes it from Tietze syndrome, almost always single.

This location explains two clinical features that the patient reports spontaneously. First, the pain is pointed to with one finger: the patient puts a fingertip where it hurts, whereas coronary pain is shown with the flat hand or a closed fist. Second, it is modulated by deep breathing and by trunk movements, because the rib cage is an articulated system in which every breath moves the painful joints.

What does the only available prospective series measure?

Only one study has prospectively analysed costochondritis in consecutive patients: that of Disla and his team, published in the Archives of Internal Medicine in 1994. It included 122 patients presenting to the emergency department with non-traumatic chest pain, without fever or known cancer, and compared them with a group of chest pain without costochondritis5.

Its results are widely quoted, often distorted. Here is exactly what it establishes:

The distinction between the first two figures is crucial and almost always lost along the way. Thirty per cent of the patients had costochondritis; only fifteen per cent had a palpation that reproduced the pain that had brought them in. Chest wall tenderness therefore exists in many people, including when it does not explain their reason for consulting. That is exactly the trap of chest wall reasoning: finding a tender point is easy, proving that it is that point that is speaking is far harder.

The fourth figure is the one to keep in mind for a whole career. Six per cent of the patients labelled costochondritis had a myocardial infarction. The rate is four to five times lower than in the control group, which confirms the orienting value of the sign. It is not zero, which forbids making it a rule-out criterion.

The study offers two other useful lessons. The erythrocyte sedimentation rate sorts nothing out: 44 ± 31 mm/h in the costochondritis group against 41 ± 31 in the controls, so requesting an ESR to "confirm costochondritis" is pointless. And costochondritis is associated with fibromyalgia only in a minority of cases (3 patients out of 36, that is 8 %), even though widespread pain is clearly more frequent there than in the controls (42 % against 5 %). The shortcut "costochondritis = fibromyalgia starting" does not hold.

Who is affected?

The profile drawn by the available sources is coherent without being precise. Costochondritis affects women more often : 69 % in Disla's series, 17 women out of 25 in Freeston's rheumatology series26. The mean age in that last series was 50 years, with a range of 26 to 75 years. In children, adolescents and young adults, the American general practice review considers that the history and the examination are generally enough for the diagnosis, whereas beyond the age of 35, or where there is coronary risk, or in the face of any cardiopulmonary symptom, an electrocardiogram and possibly a chest X-ray are required8.

The triggering circumstances reported belong to clinical common sense more than to demonstration: unusual upper limb effort, carrying a load, prolonged coughing, a repeated sporting movement. The sports literature adds the disciplines that load the rib cage: rowing, golf, throwing, racquet sports12. A published paediatric case describes severe costochondritis arising a few months after a COVID-19 infection, resistant to anti-inflammatories and to corticosteroids, controlled with colchicine37 : an isolated case, not to be generalised, but one that recalls that post-infectious chest wall pain exists.

How long does it last?

It is the question every patient asks, and the one on which the usual account is most optimistic. "Self-limiting", "benign", "a few weeks": the three terms recur in almost every review. The only two studies that really followed these patients say something else.

The practical consequence is twofold. First, the message has to be adapted: speak of a course that is "favourable but often long, with fluctuations" rather than of rapid recovery, on pain of losing the patient's trust after six weeks. Second, one must plan a reassessment: pain that persists beyond a few weeks is not abnormal, but it deserves reopening the diagnosis rather than repeating the treatment.

The cost of not doing so has been measured. In Freeston's British series, the mean delay between the onset of symptoms and the diagnosis of costochondritis was 9.4 months, with extremes reaching 57 months. Among these 25 patients there were 39 admissions for chest pain before the rheumatological work-up against 6 afterwards, 169 minor investigations against 17, and 30 major investigations against none26. This is a retrospective study without a control group, and the authors themselves refrain from attributing the whole fall to their intervention. But the order of magnitude says something simple: naming chest wall pain correctly has a value, including in investigations avoided.

Costochondritis or Tietze syndrome: where does the boundary lie?

In this chapter: why the two terms are not interchangeable, the single criterion that separates them, what a swelling makes it mandatory to exclude, and why this distinction is a service to the patient rather than a subtlety of vocabulary.

The two words circulate as synonyms, including in professional documents. They are not, and the confusion has a direct clinical consequence: it causes serious diagnoses to be missed.

Tietze syndrome was described as a painful, non-suppurative swelling of the upper costosternal region. The reference critical review, by Aeschlimann and Kahn, notes that its aetiology and its pathology remained unknown, and suggests that it might belong to the seronegative rheumatic group more than was thought, while concluding that "the existence of this condition as an entity cannot be totally denied"11. It is a cautious formulation, written in 1990, and nothing since has replaced it.

The most important row of this table is the last. A swelling of the anterior chest wall is not a sign of costochondritis: it is a red flag. Three published case reports are enough to make the point.

A woman presenting with high right parasternal pain and a progressive swelling near the first costosternal junction was examined with ultrasound before the planned injection: the scan revealed a hypervascular mass, which turned out to be a Hodgkin lymphoma invading the chest wall29. A second case report describes a non-Hodgkin lymphoma of the medial end of the clavicle, initially diagnosed as Tietze syndrome30. A third describes a man of 60, diabetic, who had had for twenty-three years a recurrent chest swelling with discharge, treated all that time as sebaceous cysts: imaging and biopsy finally concluded on tuberculous costochondritis, cured after nine months of antituberculous treatment31.

Should anything else be considered when facing a chronic swelling?

Yes, and one entity deserves to be known to the physiotherapist because it passes through their hands before being diagnosed: chronic non-bacterial osteomyelitis of the sternocostoclavicular region, also called sternocostoclavicular hyperostosis, which belongs to the SAPHO syndrome spectrum. A meta-analysis of 40 studies and 2,030 patients paints its portrait: female predominance (67 %, 95 % CI 60-73), a mean diagnostic delay of five years (95 % CI 3-7), chest pain in 89 % and swelling in 79 %16.

Two details of this meta-analysis should raise the alarm. The first: the inflammatory markers are only inconsistently raised, autoantibodies and HLA-B27 are of normal prevalence, and the histology is not specific, in other words a normal blood work-up rules nothing out. The second: palmoplantar pustulosis is present in 53 % of patients. Looking at the palms and the soles of a patient who has complained for months of a swollen, painful sternoclavicular joint is a ten-second act that can save years.

How should a painful chest wall be examined?

In this chapter: the palpation technique and its traps, the two validated scores for chest wall syndrome and what they are really worth, the paradox of palpation in a picture, the place of imaging, and the thoracic spine examination that completes, without replacing, the examination of the chest wall.

How should one palpate?

The manoeuvre is simple, which is what makes it deceptive. One palpates in turn, from top to bottom and comparing right with left: the sternoclavicular joints, the manubrium and the angle of Louis, then each chondrosternal junction from the 2^nd to the 7^th, then the costochondral junctions further out to the side, the xiphoid process, and finally the lower costal margin. The pressure must be firm and progressive, with the pad of one or two fingers, never with the fist.

Three precautions determine the value of what is found.

Look for the patient's pain, not for a pain. Enough pressure on a costal cartilage is unpleasant in just about everybody. The only question that counts is: "is this exactly it, the pain that brings you here?". The difference between the two is not theoretical: in Disla's series, 30 % of patients had costochondritis and only 15 % a reproduction of the initial pain5.

Palpate before concluding, not in order to conclude. If the history has picked up a red flag, a positive palpation does not neutralise it. The reverse order (palpate, find, feel reassured, stop asking) is the commonest error mechanism with this complaint.

Look before touching. A swelling, a redness, an asymmetry of contour can be seen. They change everything, and they are invisible through clothing.

Is there a validated score?

Two, both built on chest wall syndrome in the broad sense, not on costochondritis in particular, which no tool isolates.

The first, from Bösner's German cohort (1,212 patients, 74 general practitioners), retains four determinants: the presence of localised muscle tension, of stabbing pain, of pain reproducible on palpation and the absence of cough. The area under the ROC curve is 0.78 (95 % CI 0.75-0.81)4.

The second, developed by Ronga on the Swiss cohort of 672 patients then validated on the German cohort, retains six variables: pain neither retrosternal nor oppressive, stabbing, well localised, the absence of coronary history, the absence of concern on the doctor's part, and pain reproducible on palpation, the last counting double, for 2 points out of 7. At a threshold of 6 points, specificity reaches 89 % but sensitivity falls to 45 %; on external validation, it drops to 22 % for a specificity of 93 %7.

This profile (high specificity, low sensitivity) says exactly what these scores are for and what they are not for. A high score confirms a chest wall origin reasonably well: that is what allows Ronga to estimate that 65 additional investigations could have been avoided in 43 correctly classified patients. A low score says nothing: it misses more than half of chest wall syndromes. These tools are confirmation tools, not rule-out tools.

What really happens when palpation reproduces the pain?

This is the central point of this article, and it gains from being seen rather than read.

Ronga's exact words deserve to be repeated one last time, because they are written by the authors of the tool that gives palpation its greatest weight: "the reproduction of chest pain by palpation, the most important characteristic for diagnosing chest wall syndrome, is not pathognomonic". Among the patients their rule classified as positive, three had stable angina.

Is imaging needed?

In the absence of a red flag, no. No imaging confirms costochondritis: there is nothing to see. The clearest demonstration comes from paediatrics: in Massin's Belgian series, chest X-ray, electrocardiogram and blood tests were performed in most of the children whose pain was presumed non-organic, and in no case did they diagnose an organic disease28.

Ultrasound, on the other hand, has two precise indications, both tied to a visible or palpable abnormality. The first: to investigate a swelling before any procedure; this is the scanning-before-injection that revealed a Hodgkin lymphoma in a woman referred for injection of costochondritis29. The second: dynamic ultrasound, which is today the reference tool for slipping rib syndrome18.

And where does the thoracic spine fit into all this?

It is part of the examination, without merging with it. The costovertebral and costotransverse joints link each rib to two vertebrae; pain felt at the front may have its source at the back, and Gregory's review explicitly places referred pain from the thoracic spine among the diagnoses to consider when facing chest wall pain12. In practice it is therefore justified to examine thoracic segmental mobility and to test the costovertebral joints at the painful level.

But the two subjects do not overlap, and that is what separates this article from that one. This page covers the anterior chest wall: sternum, cartilages, chondrosternal and costochondral joints, and the entities that sit there. Chest pain of spinal origin (posterior, segmental, with its own differential diagnosis and its own rehabilitation) is the subject of a dedicated article: thoracic spine pain (chest pain of spinal origin). A patient whose pain is at the front and whose anterior palpation is silent, while pressure on the costovertebral joints reproduces the pain, belongs to that other article.

Which other chest wall pains should be recognised?

In this chapter: seven entities that present as costochondritis and are not, each with the landmark that unmasks it, the examination manoeuvre that confirms it, and what it costs to miss it.

"Costochondritis" too often serves as a generic label for any chest wall pain. The cost of that imprecision can be measured: in Foley's series of athletes, the mean delay between the onset of symptoms and the correct diagnosis was 15.4 months, after an average of 2.3 specialist consultations per patient17. Naming the entity precisely means shortening that delay.

What can physiotherapy really do for chest wall pain?

In this chapter: the real level of evidence for costal and thoracic manual therapy (modest, and that has to be said), what the only randomised trial in the field shows at four weeks then at one year, the table of modalities by level of evidence, and why reasoned reassurance is probably the most useful act of the session.

A preliminary remark is called for, because it conditions the reading of everything that follows: there is no controlled trial on the treatment of costochondritis as such8. Everything available bears on musculoskeletal chest pain, a broader category that takes in the anterior chest wall, the thoracic spine and the costovertebral joints. Transposing these results to costochondritis is reasonable, but it is a transposition, and the patient has a right to know it.

What does the only randomised trial in the field show?

A Danish trial, run by Stochkendahl and her team, randomised 115 consecutive patients presenting with acute chest pain without coronary syndrome, recruited from an acute cardiology department and four practices. Two arms: four weeks of manual treatment including spinal manipulation, or self-management serving as a minimal intervention.

À 4 and 12 weeks, the trial shows statistically significant differences in favour of manual treatment: on self-perceived change in pain at four weeks, on numerical intensity at twelve weeks22.

À one year, the same team published the follow-up. Both groups improved, and no statistically significant difference remains between them. The authors write that they could identify no consistent trend in favour of either intervention23.

The two publications must be read together, resisting the temptation to quote only the first. What this trial establishes is that manual therapy probably speeds up the journey without changing its destination. That is a real benefit, relieving pain sooner has value, but it is not a lasting benefit, and presenting it as such would be dishonest.

What does the systematic review say?

The OPTIMa collaboration sifted the literature on non-invasive interventions in musculoskeletal chest pain. The most telling result is arithmetical: 6,988 articles screened, two studies retained after critical appraisal of the risk of bias24. The authors conclude that "quality evidence on the management of musculoskeletal chest pain is scarce".

What these two studies show:

• The manipulation of the thoracic spine, compared with placebo electrotherapy, produces a short-term pain reduction that is statistically significant but clinically unimportant.
• Acupuncture gives results equivalent to placebo.
• The multimodal programme (manual therapy, soft tissue techniques, exercise, hot and cold, advice), compared with a single education session produces a pain reduction, again statistically significant and clinically unimportant. But the patients receiving multimodal care were more numerous in reporting an important improvement in their chest pain.

This last nuance deserves a pause. The mean difference in intensity between the two groups is small; the proportion of patients who declare themselves clearly better does differ. This suggests that the benefit is not evenly distributed: some patients draw real value from active care, others do not, and the average crushes the difference. It is an argument for offering, reassessing, and stopping if it does not work, not for giving up from the start.

Which modality, for which level of evidence?

So what is actually done?

Three things, in decreasing order of probable value.

1. Name it and explain it. This is the most underestimated act of management, and the one for which the indirect rationale is the most solid. Aygun's series, covering 782 children and adolescents referred to paediatric cardiology, measures what nobody usually measures: 70.8 % of the parents and 90.2 % of the patients thought the pain came from the heart, whereas a cardiac cause was retained in only 8 patients, that is 1 %27. The gap between 90 % of fear and 1 % of risk is the real subject of the consultation.

Reassuring usefully does not consist in saying "it is nothing". It consists in saying what was looked for, what was not found, why one concludes what one concludes, and on what conditions the patient must come back. A form of words that works: "I have checked the features that would point to the heart, the lung and the aorta, and none of them is present. Your pain comes from the chest wall, and I can reproduce it. It is the commonest case, and the least serious. On the other hand, if you feel this, that or the other, you do not wait for me: you call 15." The explicit list of the conditions for coming back is part of the reassurance: it is what makes it credible.

2. Treat what responds. A costal or thoracic mobilisation, work on the soft tissues, mobilisation with breathing: if the patient feels better at the end of the session and the effect holds for a few days, the modality deserves to be continued. If three or four sessions have changed nothing, it no longer does. The modest evidence base for manual therapy justifies precisely that discipline: try, measure, stop.

3. Get the patient moving again. Chest wall pain spontaneously leads to shallow breathing, to an immobile trunk and to avoidance of the movements that wake the pain. That protection is useful for a few days; beyond that, it keeps stiffness and fear going. Gradually regaining chest expansion, rotations, arm elevation and deep inspiration is part of the treatment.

How should breathing be managed and activity resumed?

In this chapter: why breathing is both symptom and lever, what the only published series of respiratory physiotherapy in this setting is worth, the return to the sporting movement, and the special case of the rower.

Why does breathing matter more here than elsewhere?

Because it is caught in a loop. Chest wall pain is made worse by deep inspiration: that is even one of its distinctive features. The patient therefore spontaneously reduces the range of his breathing. This high, short breathing keeps costal stiffness going, increases the work of the accessory muscles, and in some patients sets up a hyperventilation that produces its own chest symptoms. Fear of a cardiac cause, widely documented, closes the loop.

There is an old series, often quoted and rarely read properly, that describes exactly this situation. Evans and Lum reported in the Lancet in 1977 the case of 50 patients referred to a cardiology clinic to confirm or rule out angina, and who were all habitual hyperventilators, thirteen of them also having genuine organic heart disease. Simple physiotherapy aimed at restoring a normal breathing pattern rendered 76 % of them symptom-free, over a follow-up of 11 to 68 months25.

This figure must be handled with care. It is an uncontrolled series from 1977: no randomisation, no comparator, no blinding, and nearly fifty years of methodological hindsight. It is not proof of effectiveness. It is a signal, and it points to two lessons that still hold: functional breathing disorder produces genuine chest pain, and it coexists at times with real heart disease, since a quarter of the patients in that series had one.

What should actually be worked on?

Four lines of work, to be dosed according to what the examination found.

Restore diaphragmatic breathing. Let the patient feel the difference between high thoracic breathing and abdominal-diaphragmatic breathing, first lying down, hand on the belly. The aim is not ventilatory performance: it is to give the patient back the sense that he can breathe fully without it hurting.

Regain chest expansion in stages. Progressively deeper breaths, staying below the pain threshold, then tolerating moderate discomfort. The landmark that reassures both patient and clinician is simple: pain that rises during the exercise and settles in the minutes that follow is acceptable; pain that persists for several hours afterwards means one has gone too fast.

Make the trunk mobile. Rotations, side bends, thoracic extension, arm elevations. These movements are exactly the ones the patient avoids because they wake the pain: they are therefore the ones to win back, progressively and without surprises.

Name the loop. Explaining to the patient that breathing shallowly because it hurts ends up hurting because one breathes shallowly is not a figure of speech: it is information that gives him back a grip on his problem.

How should sport be resumed?

The answer depends entirely on the entity settled on, which justifies in retrospect all the sorting work of the previous chapter.

What do published case reports teach us?

In this chapter: seven indexed case reports, chosen because each documents a different mode of error. No case in this chapter is invented or composite: each carries its PubMed identifier.

The case that justifies looking before injecting

A woman consults for high right parasternal pain with a swelling of gradual onset near the first costosternal junction. The picture is one of refractory costochondritis, and a local corticosteroid injection is considered: a procedure that experienced rheumatologists readily perform without guidance. The authors nonetheless choose to run the ultrasound probe over it before needling. The scan reveals a hypervascular mass, which would turn out to be a Hodgkin lymphoma invading the chest wall29.

What the case teaches: the swelling is the signal, and the local procedure is the moment of danger. An injection into a tumour mass achieves nothing and delays the diagnosis. The rule is simple and without exception: no local procedure on an uninvestigated swelling. An independent case report describes the same mechanism with a non-Hodgkin lymphoma of the medial end of the clavicle initially taken for Tietze syndrome30.

The case that forbids all-or-nothing reasoning

A man of 79, diabetic, who had had a partial gastrectomy, consults for epigastric pain that gradually worsens on exertion. The picture is that of unstable angina. Coronary angiography confirms a tight coronary stenosis. The angioplasty is successful. And the pain persists.

It is only after that procedure that a meticulous physical examination finds an exquisitely tender point on the xiphoid process. The diagnosis of xiphodynia is made, and xiphoidectomy immediately abolishes the pain34.

What the case teaches: two things, in both directions. On the one hand, chest wall pain can mimic a coronary syndrome to the point of leading to coronary angiography. On the other hand, and this matters more, this patient really did have a tight coronary stenosis. Finding the chest wall cause would not have removed the need to treat the coronary disease, and treating the coronary disease did not relieve the pain. The reasoning "it is one or the other" is the error mode this case demolishes.

The case that shows what a label can cost

A man of 60, diabetic, has had for twenty-three years a recurrent chest swelling with a greenish discharge. For all that time, the picture was treated as repeated sebaceous cysts. During one flare, imaging and biopsy finally establish the diagnosis: tuberculous costochondritis, a very rare form of extrapulmonary tuberculosis. Nine months of antituberculous treatment bring complete resolution of the symptoms31.

A Belgian case report documents the same territory with a different organism: a man of 43, immunocompetent and with no medical history, who had had for four months a painful swelling of the left chest wall. CT and MRI show an abscess of the ninth costal cartilage; the guided biopsy isolates Prevotella nigrescens. Antibiotic treatment is enough, with complete disappearance of the abnormalities at six months32.

What these cases teach: chest wall pain with swelling that progresses over months is not a costochondritis dragging on. It is a diagnosis not made. And the absence of a risk background offers no protection: the second patient had no medical history at all.

The case that recalls that infection can be afebrile

A Japanese man of 68, in good health, consults for pain and redness near the right sternoclavicular joint. A week of oral antibiotics changes nothing. The CT scan then shows an abscess containing air around the pectoralis major. Transferred to a tertiary centre, he undergoes emergency surgery: the intraoperative findings are necrotic tissue and destruction of the sternoclavicular joint. Blood cultures and samples isolate a Staphylococcus aureus sensitive to meticillin33.

What the case teaches: this patient had none of the classic risk factors. The review of 180 cases confirms that this is frequent: 23 % of sternoclavicular septic arthritis cases occur without a risk factor, and 35 % of patients are not febrile15. Focal sternoclavicular pain that does not settle and that comes with local signs must never be followed in physiotherapy without a medical opinion.

The case that shows what ultrasound is really for

A man of 70 consults for left chest pain present for several weeks. Examination finds focal tenderness of the left third costochondral joint, and ultrasound demonstrates a swelling of that joint, which makes the diagnosis Tietze syndrome rather than plain costochondritis. Three ultrasound-guided corticosteroid injections abolish the pain, the swelling and the tenderness35.

What the case teaches: ultrasound is not there to "confirm costochondritis", which it cannot do. It serves to see what is visible (a swelling, a mass, an effusion) and to guide what must be guided. The difference between this case and the lymphoma one lies in the content of the picture, not in the indication for the scan: in both cases, someone looked before acting.

The case that recalls that the lung is just behind

A woman of 64 has had for four months a refractory chest pain that came on while playing golf. The context is that of an overuse chest wall injury: the sports review places golf precisely among the activities that generate costochondritis, rib stress fracture and intercostal injury. No investigation is conclusive, no medical treatment relieves. Ultrasound performed over the painful areas finally shows focal pleural effusions; an ultrasound-guided injection of ropivacaine and triamcinolone at their level relieves the patient dramatically36.

What the case teaches: the parietal pleura lies immediately beneath the chest wall, and pain of mechanical appearance can come from it. When four months of chest wall treatment have achieved nothing, the right question is not "which other technique should be tried?" but "is the diagnosis the right one?".

The case that documents a post-infectious costochondritis

A child develops severe costochondritis a few months after a COVID-19 infection, with repeated emergency department attendances for intense chest pain. Neither non-steroidal anti-inflammatory drugs nor corticosteroids control it; colchicine does. The authors call for it to be borne in mind so as to avoid useless corticosteroid courses and repeated recourse to emergency departments37.

What the case teaches: this is a single case, and nothing allows a general rule to be drawn from it. What it usefully recalls is that chest wall pain can follow an infectious episode, and that a refractory costochondritis in a child deserves a specialist opinion rather than a therapeutic escalation.

How is this applied in practice?

In this chapter: the sequence of the first session, what to say and how to say it, the referral criteria, the follow-up indicators, and the five mistakes that recur most often.

What should be done at the first session?

Take the history before touching. The mode of onset, the character of the pain, what triggers it and what calms it, the relationship with exertion, with breathing and with trunk movements, the associated signs, and the cardiovascular, thromboembolic, infectious and neoplastic background. This step is not negotiable: it is the one that decides all the others.

Look. Chest uncovered, in decent light: asymmetry of contour, swelling, redness, skin rash, scar. Look at the palms and the soles as well if the pain is sternoclavicular and chronic.

Palpate methodically, from top to bottom and comparing sides, asking at each point the only question worth asking: "is this exactly the pain that brings you here?".

Complete with the targeted manoeuvres: hooking manoeuvre under the costal margin if the pain is low, isolated palpation of the xiphoid, examination of the thoracic spine and of the costovertebral joints at the same level.

Conclude explicitly, out loud, in front of the patient: what was looked for, what was not found, what is settled on, what is planned, and on what conditions to come back or to call.

How should it be said to the patient?

The way the message is worded matters, because worry is the dominant symptom. Three principles serve as method.

Say what was looked for, not only what was found. "It is nothing" reassures nobody. "I have checked the signs that would point to the heart, the lung and the aorta, and I find none of them" does reassure, because the patient understands that someone looked.

Give the frequency. "It is the commonest cause of chest pain among people who consult, about one case in three to one case in two" places the patient inside a population, not inside an exception.

Give the conditions for coming back. Reassurance without a safety net is fragile; reassurance that comes with a precise list of reasons to seek review is solid, because it shows that the door has not been closed. It is also what protects the patient, and the practitioner.

One point deserves to be anticipated: the duration. The available data show that about 55 % of patients are still in pain at six months as at one year. Announcing "it will go in two weeks" invites a contradiction that will destroy trust. Announcing "it settles favourably, but often in waves, and sometimes over several months" is true and holds up over time.

When to refer, and to whom?

What should be measured to know whether progress is being made?

• The intensity on a numerical scale, but above all its change week after week rather than its value at a single moment.
• The functional impact: which everyday or sporting movements have become possible again. That is what speaks to the patient, and it is what the OPTIMa review shows to be most discriminating: the proportion of patients reporting an improvement that is important.
• The breathing range tolerated without pain, a simple and telling measure: how far the patient can breathe in without seizing up.
• The level of worry: ask the patient explicitly whether he still fears for his heart. In children, that fear was present in 90.2 % of patients for 1 % of cardiac causes: it is an outcome indicator in its own right.

The five recurring mistakes

Frequently asked questions

Does pain reproduced by pressing on the ribs rule out a heart attack?

No. It makes the coronary cause less likely, without eliminating it. Pain reproducible on palpation carries a likelihood ratio of 0.2–0.3 for acute coronary syndrome6: it divides the probability by three to five, it does not abolish it. In practical terms, coronary artery disease is present in 3–6 % of the adults who consult for chest pain with chest wall tenderness8, and 6 % of the patients labelled costochondritis in an emergency department series had an infarction5.

What is the difference between costochondritis and Tietze syndrome?

The swelling. Tietze syndrome involves a visible or palpable swelling of the costosternal junction, most often single and high up: the second or third cartilage11. Costochondritis is pain without swelling, often over several junctions, typically from the second to the fifth12. This is not a nuance of vocabulary: the presence of a swelling makes imaging mandatory before any procedure, because it may reveal a tumour or an infection.

How long does costochondritis last?

Longer than the usual account suggests. In Disla's prospective series, 11 patients out of 21 reviewed at one year still had chest pain, even though only a third still had clear-cut costochondritis5. In Bösner's German cohort, 55.4 % of the patients with chest wall syndrome were still in pain at six months4. The course is favourable, but it is often long and fluctuating.

Is physiotherapy effective for costochondritis?

No controlled trial has evaluated a treatment for costochondritis as such8. For musculoskeletal chest pain in the broad sense, the OPTIMa review retained only two trials: thoracic manipulation does better than placebo in a way that is statistically significant but clinically unimportant, and a multimodal programme does not lower pain any more than a single education session, even though patients receiving multimodal care more often report an important improvement24. The only randomised trial in the field shows an advantage at 4 and 12 weeks that had disappeared at one year23. The evidence base is therefore modest, and the patient should be told so.

Which signs mean chest pain needs urgent medical assessment?

Constricting pain triggered by exertion and relieved by rest; radiation to both arms or to the jaw; profuse sweating, faintness or syncope; acute breathlessness; pain maximal from the outset with asymmetrical blood pressure or a neurological deficit; haemoptysis; fever or chest swelling. None of these signs is cancelled because palpation reproduces the pain.

Is rib pain in a child serious?

Rarely. In a series of 782 children and adolescents referred to paediatric cardiology for chest pain, a cardiac cause was retained in only 8 patients, that is 1 %, against 33 % musculoskeletal causes and 28.4 % psychogenic causes27. The same work reports that 70.8 % of the parents and 90.2 % of the young patients thought of a cardiac origin: the gap between fear and risk is the real subject of the consultation. Precordial catch, a brief needle-like pain at rest, is recognised from the history alone and calls for no investigation20.

Is an X-ray or a blood test needed?

In the absence of a red flag, no. No imaging confirms costochondritis. The erythrocyte sedimentation rate does not separate costochondritis from other chest pain: 44 ± 31 against 41 ± 31 mm/h in Disla's series5. And in children, X-ray, ECG and blood tests performed for pain presumed non-organic diagnosed no organic disease28. An electrocardiogram, on the other hand, is justified beyond the age of 35, where there is coronary risk, or in the face of any cardiopulmonary symptom8.

My pain is on the left: is that more worrying?

No, and site sorts nothing out. In Bösner's cohort, among the patients whose pain was genuinely of chest wall origin, it was retrosternal in 52.0 % of cases and on the left in 69.2 %4. Chest wall pain readily sits on the left and in the middle: that is an argument neither one way nor the other.

Do this article and the one on thoracic spine pain cover the same thing?

No, and the boundary is clear. This one covers the anterior chest wall: sternum, costal cartilages, chondrosternal and costochondral joints, with the entities that sit there, costochondritis, Tietze syndrome, slipping rib, xiphodynia, sternoclavicular disorders. The article thoracic spine pain (chest pain of spinal origin) covers pain that is posterior, of vertebral and costovertebral origin, with its own segmental assessment and its own rehabilitation. A patient whose anterior palpation is silent but in whom pressure on the costovertebral joints reproduces the pain belongs to the other page.

Two pages, two territories

This article and the one on thoracic spine pain answer each other without overlapping, and it is worth saying precisely where the line runs. Here: the anterior chest wall, sternum, costal cartilages, chondrosternal and costochondral joints, xiphoid process, sternoclavicular joints, lower costal margin. The entities covered are costochondritis, Tietze syndrome, slipping rib, xiphodynia, precordial catch, rib stress fracture and sternoclavicular disorders. Over there: chest pain of spinal origin, thoracic vertebral segments, costovertebral and costotransverse joints, with its segmental assessment, its occupational risk factors and its own rehabilitation: thoracic spine pain (chest pain of spinal origin).

The test that points the way is a single manoeuvre: if anterior palpation of a chondrocostal junction reproduces the patient's pain, it is this page; if it is silent and pressure on the costovertebral joints at the same level reproduces it, it is the other. And if both respond, which happens, both pages serve, in this order: triage first, treat afterwards. On the thoracolumbar junction, a third article completes the picture: Maigne syndrome. And for the neurovascular compressions of the cervicothoracic outlet, often confused with high chest wall pain: thoracic outlet syndrome.

References

Thirty-seven references, resolved one by one through the PubMed E-utilities API on 16 August 2026. For each of them, the abstract was read in full and every figure quoted in the article checked at source. The nine case reports in chapter 8 are genuine published and indexed cases: no case in this article is invented or composite.