Signal Evaluation: Fluoroquinolones and Tendon Injury and Tendon Rupture

Introduction

The association between fluoroquinolone antibiotics and tendon injury is one of the classic examples of a pharmacovigilance signal that developed from individual clinical observations into a well-characterised class safety concern.

It is also an unusually useful signal-evaluation case because the evidence contains several different components:

• spontaneous reports of tendon injury;
• a distinctive clinical phenotype;
• a relatively short temporal relationship in many cases;
• epidemiological evidence of increased risk;
• increased susceptibility among older patients and those receiving corticosteroids;
• evidence involving several fluoroquinolone substances;
• biological and experimental evidence supporting several possible mechanisms;
• and progressive regulatory action culminating in substantial restrictions on fluoroquinolone use in the European Union.

The signal is therefore useful for understanding how pharmacovigilance moves from:

"something unusual has been reported"

to:

"the drug is associated with a clinically meaningful adverse effect, particularly in defined circumstances, and prescribing should be modified accordingly."

It also illustrates an important limitation of mechanistic reasoning.

The association between fluoroquinolones and tendon injury is well established.

The precise molecular mechanism by which an individual patient develops tendon injury is not.

That distinction matters.

1. Define the Signal

The basic drug-event pair is:

fluoroquinolone exposure → tendon injury/tendinopathy → tendon rupture.

The signal should not, however, be treated as a single undifferentiated musculoskeletal event.

The clinical spectrum includes:

• tendon pain;
• tendon inflammation or tendinitis;
• tendinopathy;
• tendon degeneration;
• and tendon rupture.

The Achilles tendon is particularly important, although other tendons can also be affected.

This distinction matters because a signal evaluator needs to understand whether the available evidence concerns:

symptoms, pathological changes, or a clinically serious structural outcome.

A report of tendon pain and a surgically repaired Achilles tendon rupture are not equivalent events.

2. Why This Was a Good Pharmacovigilance Signal

Tendon rupture is clinically distinctive.

Unlike a nonspecific symptom such as fatigue, a tendon rupture can have:

• a clear anatomical location;
• a recognisable clinical presentation;
• an identifiable temporal relationship to drug exposure;
• and objective confirmation through imaging or surgery.

This makes the event relatively suitable for signal evaluation.

The Achilles tendon is also exposed to substantial mechanical loading.

Consequently, a medication that changes tendon structure or repair could plausibly produce a clinically recognisable pattern.

3. The Early Clinical Observations

Reports linking quinolone antibiotics with tendon disorders appeared in the medical literature before the signal became a major regulatory issue.

Early case reports described patients developing:

• Achilles tendon pain;
• Achilles tendinitis;
• and tendon rupture

during or after treatment with quinolone or fluoroquinolone antibiotics.

The reports were particularly notable because tendon injury is not an expected consequence of antibacterial activity.

That disconnect between:

the intended pharmacological action

and:

the observed adverse event

is often what makes an unusual adverse-event report worthy of attention.

4. Case Reports: Useful but Limited

Case reports can be particularly informative for rare adverse events.

They can reveal:

• unusual anatomical patterns;
• unexpected temporal relationships;
• risk factors;
• dechallenge;
• rechallenge;
• and clinical features that later become part of a formal case definition.

But they cannot establish incidence.

They also cannot reliably distinguish:

drug causation

from:

coincidental occurrence.

Tendon rupture has a substantial background incidence, particularly in active populations and older adults.

Therefore, epidemiological investigation was required.

5. The Temporal Relationship

One of the striking features of fluoroquinolone-associated tendon injury is that symptoms can develop relatively soon after treatment begins.

EMA's later review noted that tendon damage, particularly involving the Achilles tendon, can occur within approximately 48 hours of starting treatment, while injury may also appear months after treatment has stopped. 1

This is an important feature of the signal.

The temporal pattern is not simply:

long-term cumulative exposure → delayed toxicity.

Instead, the reported spectrum includes:

rapid onset during treatment

as well as:

delayed presentation after treatment.

That makes temporal reasoning more complicated than simply looking for a fixed latency period.

6. Why Delayed Events Matter

A delayed adverse event can be difficult to recognise.

If a patient develops Achilles tendon pain three months after receiving a short course of an antibiotic, the prescribing episode may no longer be prominent in the clinical history.

This creates a pharmacovigilance challenge.

The evaluator must therefore consider:

• exposure history extending beyond the immediate treatment period;
• the natural history of tendon injury;
• the timing of symptoms;
• and whether the observed latency is compatible with known evidence.

The EMA specifically recognised that tendon injury can occur months after discontinuation. 2

7. The Achilles Tendon

The Achilles tendon became particularly associated with the signal.

This is clinically useful because the signal was not simply:

fluoroquinolones → any musculoskeletal problem.

Instead, a relatively distinctive pattern emerged involving the Achilles tendon.

Other tendons can also be affected, including:

• shoulder tendons;
• hand tendons;
• and other tendons subjected to mechanical stress.

The Achilles tendon nevertheless remains the best-known clinical manifestation.

8. Epidemiological Evidence

One of the important steps in signal evaluation was determining whether fluoroquinolone exposure was associated with tendon rupture at a population level.

An early epidemiological study by van der Linden and colleagues examined tendon rupture in patients receiving fluoroquinolones.

The study provided evidence supporting an association between fluoroquinolone exposure and Achilles tendon rupture.

Later studies refined the relationship by examining:

• age;
• corticosteroid exposure;
• timing;
• individual fluoroquinolones;
• and duration of exposure.

The accumulating epidemiological evidence moved the signal substantially beyond spontaneous reports.

9. The Corticosteroid Interaction

One of the most important observations in this signal was the increased risk among patients receiving corticosteroids.

The association between fluoroquinolones and tendon injury appeared stronger when corticosteroids were used concurrently.

This is clinically important for two reasons.

First, it identifies a potentially higher-risk population.

Second, it provides a form of effect modification that is biologically and epidemiologically informative.

The signal therefore evolved from:

fluoroquinolone → tendon injury

to:

fluoroquinolone + patient susceptibility factors → greater tendon-injury risk.

EMA continues to identify systemic corticosteroid treatment as an important risk factor for tendon injury with fluoroquinolones. 3

10. Age as a Risk Factor

Older age is another consistently recognised risk factor.

EMA identifies older patients as being at increased risk of tendon damage associated with fluoroquinolones. 4

This makes biological and clinical sense because tendon structure and repair capacity change with age.

But again, biological plausibility should not be confused with proof of mechanism.

Age may represent several correlated processes:

• altered tendon structure;
• reduced tissue repair;
• changes in physical activity;
• comorbidity;
• concomitant medications;
• and greater exposure to corticosteroids.

Therefore, an observed age interaction should be interpreted within the broader clinical context.

11. Other Recognised Risk Factors

Regulatory assessments have also identified increased concern among patients with:

• renal impairment;
• solid-organ transplantation;
• and concomitant corticosteroid treatment.

EMA specifically highlights these populations in its current safety information. 5

These factors are important because they demonstrate that drug safety is not always adequately represented by an average risk across the entire treated population.

The same drug can produce substantially different risk in different patients.

12. A Simple Clinical Comparison

Consider two patients receiving the same fluoroquinolone.

Patient A

• 25 years old;
• no corticosteroid treatment;
• normal renal function;
• short course of treatment;
• no previous tendon disorder.

Patient B

• 75 years old;
• receiving systemic corticosteroids;
• renal impairment;
• previous tendon problems.

The drug exposure is similar.

The expected susceptibility is not.

This illustrates why pharmacovigilance evaluation should examine:

effect modification

rather than assuming that all exposed patients have the same risk.

13. A Case That Fits the Signal

Consider a 70-year-old patient receiving a fluoroquinolone for a bacterial infection.

The patient is also taking systemic corticosteroids.

Five days after starting the antibiotic, the patient develops acute Achilles tendon pain while walking.

Imaging shows tendon injury.

This case has several signal-supporting features:

• known suspect drug;
• characteristic anatomical site;
• plausible temporal relationship;
• recognised risk factor;
• and objective evidence of tendon injury.

It would therefore warrant careful evaluation as a suspected fluoroquinolone-associated tendon disorder.

14. A Case That Fits Less Well

Now consider a 35-year-old athlete who develops an Achilles tendon rupture during explosive exercise six months after receiving a fluoroquinolone.

There is no corticosteroid exposure.

There is no preceding tendon pain.

There is no imaging evidence suggesting an unusual tendon phenotype.

The temporal relationship is weak.

The patient also has a strong mechanical explanation.

This does not mean the drug contribution is impossible.

It means the case provides much less evidence for causality.

This distinction is critical.

A pharmacovigilance evaluator should not classify every tendon rupture occurring after fluoroquinolone exposure as drug-induced.

15. Dechallenge

Dechallenge is less straightforward for tendon injury than for reversible symptoms.

Stopping an antibiotic does not immediately restore tendon integrity.

However, discontinuation can prevent further exposure and may reduce the risk of progression or additional injury.

At the population level, the relationship between exposure and subsequent risk can also be studied.

The EMA's current information recognises that tendon injury may occur after treatment has stopped, meaning that the absence of immediate improvement after discontinuation does not invalidate the signal. 6

This is an important methodological point:

The absence of a classic immediate dechallenge response does not necessarily weaken a signal when the suspected event involves structural tissue injury.

16. Rechallenge

Intentional rechallenge would generally be inappropriate for a serious suspected tendon reaction.

Nevertheless, accidental or clinically necessary re-exposure can provide informative evidence in individual cases.

If a patient develops similar tendon symptoms repeatedly after exposure to the same or related fluoroquinolones, the observation may strengthen the causal hypothesis.

However, rechallenge evidence should be interpreted cautiously because:

• the event may recur naturally;
• the underlying susceptibility may persist;
• and deliberate rechallenge is ethically inappropriate for serious reactions.

17. The Biological Mechanism

Several mechanisms have been proposed for fluoroquinolone-associated tendon injury.

Experimental research has identified effects involving:

• tenocyte biology;
• extracellular-matrix metabolism;
• collagen;
• matrix metalloproteinases;
• oxidative stress;
• mitochondrial function;
• and cell proliferation.

A 2019 systematic review identified several proposed mechanisms, including altered collagen metabolism, effects on cell signalling and migration, ion-chelation effects and reactive oxygen species generation. The review also emphasised that there was no definitive structure-damage relationship and that the dose-effect relationship remained unclear. 7

This is exactly the distinction we should preserve in signal-evaluation writing.

18. What Is Established About the Mechanism?

Established

Fluoroquinolones are associated clinically with tendon injury and tendon rupture.

Strongly plausible

Fluoroquinolones may adversely affect tendon cells and extracellular-matrix homeostasis.

Supported experimentally

Experimental studies suggest effects on:

• collagen metabolism;
• matrix metalloproteinases;
• reactive oxygen species;
• cell proliferation;
• and mitochondrial function.

Not fully established

The precise sequence of molecular events responsible for tendon rupture in an individual patient.

The mechanism therefore supports the clinical association but does not independently prove causality.

19. Why Chelation Has Been Proposed

Fluoroquinolones can chelate divalent and trivalent metal ions.

Some proposed mechanisms have suggested that chelation may influence enzymes or cellular processes important to tendon maintenance.

This is biologically interesting.

But it would be an error to state:

"Fluoroquinolones cause tendon rupture because they chelate magnesium."

That is much stronger than the evidence supports.

The appropriate formulation is:

Chelation has been proposed as one contributor to the biological effects of fluoroquinolones on tendon tissue, but the precise causal mechanism remains incompletely established.

This distinction between:

mechanistic hypothesis

and:

established mechanism

is essential.

20. Collagen and Matrix Metalloproteinases

Another proposed mechanism involves altered extracellular-matrix metabolism.

Experimental studies have reported changes in collagen-related processes and matrix metalloproteinase activity.

Because tendon strength depends heavily on organised collagen architecture, disruption of extracellular-matrix homeostasis provides a plausible explanation for tissue weakening.

However, laboratory findings do not automatically establish that the same pathway is responsible for clinical tendon rupture.

The strongest conclusion is therefore:

experimental findings support biological plausibility.

They should not be presented as a complete causal pathway.

21. Oxidative Stress

Reactive oxygen species have also been proposed as contributors.

Experimental work has demonstrated oxidative stress and mitochondrial effects in tendon cells exposed to fluoroquinolones.

Again, the evidence is mechanistically interesting.

But the clinical relevance of each experimental pathway remains uncertain.

A useful hierarchy is:

clinical association

→ strongest evidence for the actual adverse effect.

epidemiological interaction

→ evidence about who is at increased risk.

experimental mechanism

→ evidence that the association is biologically plausible.

These are complementary forms of evidence, not interchangeable ones.

22. Epidemiological Quantification

A 2018 case-control study involving 1,118 Achilles tendon rupture cases confirmed an association between fluoroquinolone exposure and Achilles tendon rupture, with an odds ratio of approximately 2.20. 8

The magnitude is much more modest than some spontaneous-report disproportionality estimates.

This is not surprising.

Different study designs answer different questions.

A spontaneous-report database asks whether:

the event is reported disproportionately with the drug.

A case-control study asks whether:

exposure is more common among cases than controls.

Neither number should be interpreted as if it were the same measure.

23. Why Spontaneous-Report RORs Can Be Huge

Modern pharmacovigilance analyses of FAERS have reported very large disproportionality measures for fluoroquinolone-associated tendon disorders.

For example, a 2016–2021 FAERS analysis found positive signals for tendonitis and tendon rupture for ciprofloxacin, levofloxacin and moxifloxacin. The reported RORs were very large, including approximately 98.5 for ciprofloxacin and 76.4 for levofloxacin for the respective strongest tendon outcomes. 9

These numbers should not be interpreted as:

98-fold clinical risk.

Disproportionality measures quantify reporting patterns.

They are affected by:

• reporting behaviour;
• stimulated reporting;
• indication;
• notoriety;
• co-medications;
• differential exposure;
• and reporting completeness.

Therefore:

A very large disproportionality statistic is evidence that the reporting pattern deserves attention, not a direct estimate of patient-level relative risk.

24. A Useful Comparison of Evidence Types

The fluoroquinolone signal demonstrates the value of combining these evidence streams.

25. A Key Signal-Evaluation Principle: Do Not Mix Measures

Suppose a pharmacovigilance database produces:

ROR = 80

and an epidemiological study produces:

OR = 2.2.

It would be incorrect to conclude that the studies disagree simply because the numbers are different.

They are measuring different things.

The ROR describes disproportional reporting.

The epidemiological OR estimates the association between exposure and disease in a defined study population.

The evaluator should therefore ask:

What does each study design tell us?

rather than:

Which number is the correct one?

26. The Role of Background Incidence

Tendon rupture has a background incidence.

That incidence varies with:

• age;
• sex;
• physical activity;
• comorbidities;
• tendon location;
• and population.

Therefore, the clinical meaning of an increased relative risk depends on the baseline risk.

If the baseline risk is low, even a two-fold increase may produce a relatively small absolute excess.

Conversely, in a high-risk subgroup, the same relative increase may have greater clinical importance.

This is why identifying susceptible populations became an important part of the signal.

27. Corticosteroid Interaction

The fluoroquinolone-corticosteroid interaction is particularly useful for teaching effect modification.

Imagine:

Risk with fluoroquinolone alone = X

and:

Risk with corticosteroid alone = Y.

The combined exposure produces a risk greater than expected from either exposure alone.

This does not automatically prove a biological interaction.

But it raises the possibility that corticosteroid treatment modifies susceptibility.

Regulatory authorities have considered this clinically important enough to recommend avoiding concomitant systemic corticosteroid treatment with fluoroquinolones where possible. 10

28. Why This Interaction Matters

A drug safety signal becomes much more actionable when risk can be stratified.

Compare:

"Fluoroquinolones may cause tendon injury."

with:

"Fluoroquinolones may cause tendon injury, with greater risk in older patients, patients with renal impairment or solid-organ transplantation, and patients receiving systemic corticosteroids."

The second statement is much more useful.

It allows prescribers to modify decisions according to patient characteristics.

29. A More Difficult Question: Is It a Class Effect?

The early signal was associated with individual quinolone and fluoroquinolone medicines.

Over time, evidence accumulated across multiple fluoroquinolones.

Regulatory action eventually treated tendon injury as a class concern.

EMA's 2018 review involved multiple fluoroquinolones, including:

• ciprofloxacin;
• levofloxacin;
• lomefloxacin;
• moxifloxacin;
• norfloxacin;
• ofloxacin;
• pefloxacin;
• prulifloxacin;
• and rufloxacin.

The Agency concluded that the remaining fluoroquinolone medicines should have their use restricted and that prescribing information should describe serious tendon and other disabling adverse effects. 11

30. Why Class Extrapolation Is Not Automatic

A class effect should not be assumed merely because drugs share a therapeutic class.

The evaluator should examine:

• pharmacological similarity;
• target interaction;
• tissue exposure;
• clinical reports;
• epidemiological evidence;
• consistency across substances;
• and biological plausibility.

In this case, the convergence of clinical and regulatory evidence supported a class-level interpretation.

That does not mean every fluoroquinolone necessarily has identical potency for tendon toxicity.

31. Regulatory Evolution

The regulatory history is particularly instructive.

The signal progressed through several stages:

clinical reports

↓

recognition of tendon injury

↓

warnings concerning tendon disorders

↓

identification of high-risk populations

↓

broader class recognition

↓

review of serious disabling reactions

↓

substantial restrictions on use

This progression demonstrates that regulatory action can evolve as the scope and severity of a signal become clearer.

32. The 2018 EMA Review

In 2018, EMA conducted a broader EU-wide review of serious, disabling and potentially permanent adverse effects associated with quinolone and fluoroquinolone antibiotics.

The review considered adverse effects involving:

• tendons;
• muscles;
• joints;
• the nervous system;
• psychiatric effects;
• and sensory systems.

EMA concluded that some quinolone medicines should be suspended and that use of the remaining fluoroquinolones should be restricted. 12

This was an important change in regulatory perspective.

The issue was no longer simply:

"Does this antibiotic cause tendon rupture?"

It had become:

"Does the overall adverse-effect profile justify continued use in a particular clinical context?"

33. From Adverse Event to Benefit-Risk Problem

This distinction is fundamental.

Fluoroquinolones are effective antibiotics.

They can be particularly useful for selected infections and circumstances.

But they also have serious adverse effects.

Therefore, the regulatory question becomes:

When does the benefit of a fluoroquinolone outweigh its risks?

The answer depends heavily on the indication.

A potentially serious adverse effect may be acceptable when treating a severe infection for which alternatives are unsuitable.

The same risk may be unacceptable when the antibiotic is being used for a mild infection for which safer alternatives are readily available.

34. Risk-Minimisation Logic

The EMA recommendations therefore included restrictions on use.

The 2018 review recommended that fluoroquinolones generally should not be used:

• for infections that may resolve without antibacterial treatment;
• for certain non-severe infections;
• for prevention of some conditions;
• or in patients with previous serious fluoroquinolone-associated reactions.

It also recommended caution in patients at increased risk of tendon injury, particularly older patients, those with renal problems, transplant recipients and those receiving systemic corticosteroids. 13

This is a classic example of targeted risk minimisation.

35. The "Stop at First Sign" Recommendation

EMA recommends discontinuation at the first sign of tendon pain or inflammation. 14

This is an important example of translating pharmacovigilance evidence into a practical intervention.

The underlying reasoning is:

early symptom recognition

→

reduced continued exposure

→

potential reduction in progression or further injury.

The measure is therefore not merely descriptive.

It is designed to alter clinical behaviour.

36. Historical Product Information

A particularly useful historical signal-evaluation exercise is to compare product information across time.

For fluoroquinolones, the evaluator can examine:

Early product information

Was tendon injury mentioned?

Intermediate versions

Were tendonitis and tendon rupture explicitly identified?

Later versions

Were:

• older age;
• corticosteroid exposure;
• renal impairment;
• transplantation;
• and discontinuation advice

added or strengthened?

Post-2018 versions

How did the wording change after the broader EU review?

This provides a visible record of how safety evidence translated into regulatory language.

37. What Historical SmPC Comparison Can Teach

Historical product information can reveal whether the regulatory response evolved from:

event recognition

to:

risk characterisation

to:

risk minimisation.

This is particularly useful for training signal evaluators.

The evaluator should not simply quote the current SmPC.

The important question is:

What did regulators know at each point in time, and how did the wording change as the evidence developed?

That requires using genuine historical versions rather than reconstructing old wording from memory.

38. A Hypothetical Historical Signal Timeline

A simplified reconstruction might look like:

This is the essence of signal evolution.

39. A Clinical Case With Risk Modification

Consider an older patient receiving systemic corticosteroids who requires antibacterial treatment.

The clinician is considering a fluoroquinolone.

The signal evaluation tells us several things:

• the patient belongs to a higher-risk population;
• tendon injury is a recognised class effect;
• tendon symptoms may appear rapidly;
• tendon injury may also occur after treatment;
• and safer alternatives may be preferable when clinically appropriate.

The signal therefore changes prescribing behaviour without requiring the conclusion that fluoroquinolones are never appropriate.

40. A Case Where the Benefit May Be Different

Now consider a patient with a serious infection for which:

• first-line alternatives are unsuitable because of resistance;
• the infection is severe;
• and effective fluoroquinolone therapy is clinically important.

The risk-benefit calculation is different.

The presence of a known serious adverse effect does not automatically make the drug contraindicated.

Instead, the clinician must assess:

severity of infection

versus:

individual susceptibility to adverse effects.

This is the practical meaning of benefit-risk assessment.

41. Signal Detection Versus Signal Evaluation

The initial reports answered:

"Could fluoroquinolones cause tendon injury?"

The later evidence answered more sophisticated questions:

• How strong is the association?
• Which tendons are affected?
• How quickly can injury occur?
• Which patients are at greater risk?
• Does corticosteroid treatment modify risk?
• Is the effect associated with several fluoroquinolones?
• How serious can the outcome be?
• What prescribing restrictions are justified?

That is the progression from signal detection to signal evaluation.

42. What the Evidence Does Not Establish

It would be inappropriate to conclude that:

• every tendon rupture after fluoroquinolone exposure is caused by the drug;
• every fluoroquinolone has identical tendon toxicity;
• the exact molecular mechanism is fully known;
• tendon injury can never occur after a single short course;
• or stopping the drug guarantees prevention of rupture.

The evidence supports a much more nuanced conclusion.

43. What Is Established?

Established

• Fluoroquinolones are associated with tendon injury and tendon rupture.
• The Achilles tendon is particularly affected.
• Tendon injury can occur during treatment and may occur after treatment has ended.
• Older age increases risk.
• Systemic corticosteroid exposure increases risk.
• Renal impairment and solid-organ transplantation are recognised higher-risk contexts.
• The association applies across the fluoroquinolone class.
• Regulatory authorities have implemented substantial restrictions and warnings.
• The risk can be serious and, in some cases, prolonged or disabling.

Strongly supported

• Susceptibility varies between patients.
• The tendon injury signal is not explained simply by the antibacterial pharmacology.
• Several biological pathways may contribute to tendon damage.

Not fully established

• The exact molecular mechanism producing tendon rupture.
• The precise contribution of each proposed experimental pathway.
• Whether individual fluoroquinolones have clinically meaningful differences in tendon toxicity.
• The exact mechanism responsible for delayed tendon injury after treatment cessation.

44. What Would Strengthen the Mechanistic Evidence?

Mechanistic evidence would be strengthened by:

• reproducible human tissue findings;
• biomarkers predicting tendon injury;
• consistent dose-response relationships in relevant models;
• clear relationships between tissue exposure and tendon damage;
• and experimental findings that explain the clinical risk modifiers.

The absence of a fully established mechanism does not invalidate the epidemiological association.

This is an important general pharmacovigilance lesson.

45. The Role of Pharmacovigilance Databases

Spontaneous-report databases remain useful even after a signal has been established.

They can identify:

• rare presentations;
• unusual tendon sites;
• new patient populations;
• timing patterns;
• serious outcomes;
• and possible differences between products.

A 2016–2021 FAERS analysis continued to detect strong disproportionality signals for tendonitis and tendon rupture across ciprofloxacin, levofloxacin and moxifloxacin. 15

However, the database cannot determine absolute incidence because the denominator of exposed patients is not known.

It is therefore a surveillance tool, not a replacement for epidemiological studies.

46. Why the Signal Remains Relevant After Regulatory Action

Once a risk is recognised, pharmacovigilance does not stop.

The questions change.

Before recognition:

Is there a safety problem?

After recognition:

Is the risk changing?

Are risk-minimisation measures effective?

Are new populations being affected?

Are clinicians prescribing appropriately?

Are serious events still occurring despite warnings?

Are new fluoroquinolone products behaving consistently?

This is the transition from:

signal identification

to:

ongoing safety surveillance.

47. A Modern Signal-Evaluation Framework Applied to This Case

Step 1 — Define the drug-event pair

Fluoroquinolones and tendon injury/tendon rupture.

Step 2 — Define the phenotype

Particularly Achilles tendinitis, tendinopathy and rupture.

Step 3 — Establish the background rate

Tendon rupture occurs without fluoroquinolone exposure.

Step 4 — Review spontaneous reports

Look for repeated clinical patterns and serious outcomes.

Step 5 — Review epidemiology

Assess relative and absolute risk.

Step 6 — Examine susceptibility

Assess age, corticosteroids, renal impairment and transplantation.

Step 7 — Examine timing

Assess onset during treatment and after discontinuation.

Step 8 — Examine dechallenge

Interpret cautiously because tendon injury is structural.

Step 9 — Assess rechallenge

Use only accidental or clinically unavoidable evidence.

Step 10 — Assess mechanism

Distinguish established pharmacology from experimental hypotheses.

Step 11 — Assess class effect

Compare evidence across fluoroquinolones.

Step 12 — Review regulatory evolution

Determine how the warning and restrictions changed.

Step 13 — Assess risk minimisation

Determine whether prescribing restrictions and warnings are proportionate.

Step 14 — Perform benefit-risk assessment

Consider the clinical indication and availability of alternatives.

Step 15 — State uncertainty

Separate established conclusions from unresolved mechanistic questions.

48. Overall Signal Assessment

The totality of evidence supports a causal association between systemic fluoroquinolone exposure and tendon injury, including tendon rupture.

The evidence is supported by:

• repeated clinical reports;
• a characteristic anatomical phenotype;
• epidemiological associations;
• temporal relationships;
• increased susceptibility in older patients and patients receiving corticosteroids;
• evidence across multiple fluoroquinolone substances;
• biological plausibility;
• and regulatory review.

The signal is therefore substantially stronger than a simple spontaneous-report association.

At the same time, the precise mechanism remains incompletely established.

Experimental research supports several possible pathways involving tendon cells, extracellular matrix, collagen metabolism, oxidative stress and other cellular processes, but no single pathway adequately explains the entire clinical phenomenon. 16

The clinically important conclusion is therefore not dependent on identifying one definitive molecular mechanism.

49. Benefit-Risk Interpretation

The safety concern must be interpreted in context.

Fluoroquinolones can provide important antibacterial treatment.

However, their adverse-effect profile includes rare but potentially serious tendon injury and other disabling reactions.

Therefore, the benefit-risk balance depends strongly on:

• infection severity;
• availability of effective alternatives;
• patient susceptibility;
• previous fluoroquinolone reactions;
• and the clinical consequences of delaying effective treatment.

This explains why European regulatory action focused on restricting inappropriate use rather than declaring that all fluoroquinolone treatment should cease.

50. Final Conclusion

The fluoroquinolone tendon-injury signal illustrates how pharmacovigilance transforms a rare clinical observation into an actionable safety conclusion.

The signal began with unusual cases of tendon injury.

The repeated phenotype generated concern.

Epidemiological studies demonstrated an association.

Further studies identified important susceptibility factors, particularly older age and corticosteroid exposure.

Evidence accumulated across multiple fluoroquinolones.

Experimental studies provided plausible explanations involving tendon-cell and extracellular-matrix effects, although the precise mechanism remains incompletely established.

Regulators progressively strengthened warnings and ultimately restricted fluoroquinolone use in the European Union because the broader adverse-effect profile could no longer be considered acceptable for some low-benefit indications.

The mature conclusion is therefore:

Systemic fluoroquinolones are causally associated with tendon injury, including tendon rupture, particularly involving the Achilles tendon. Risk is increased in susceptible populations such as older patients and those receiving systemic corticosteroids, and tendon injury can occur during treatment or after treatment has ended. The clinical association is well established, while the precise molecular mechanism remains incompletely understood. The signal illustrates why pharmacovigilance should quantify risk, identify susceptible populations, distinguish clinical evidence from mechanistic hypotheses, and translate the totality of evidence into proportionate risk minimisation rather than relying on either spontaneous reports or mechanistic speculation alone.

51. Lessons for Signal Evaluators

This signal provides several reusable lessons.

Lesson 1 — A recognisable phenotype matters

A tendon rupture is more informative when the clinical pattern and anatomical site fit the suspected adverse reaction.

Lesson 2 — Timing matters

A short treatment course does not imply that an adverse effect must be immediate.

Lesson 3 — Patient susceptibility matters

Age, corticosteroids, renal impairment and transplantation modify risk.

Lesson 4 — Do not confuse disproportionality with relative risk

A large ROR is not a direct estimate of patient-level risk.

Lesson 5 — Mechanistic evidence is supportive

A plausible mechanism strengthens the causal interpretation but should not be presented as established when it remains uncertain.

Lesson 6 — Class effects require evidence

Shared pharmacology helps, but clinical and regulatory evidence across substances should be considered.

Lesson 7 — Risk minimisation should be targeted

The appropriate response was not simply "never use fluoroquinolones."

It was to restrict use where benefit is low, identify higher-risk patients and strengthen warnings.

Lesson 8 — Benefit-risk is indication-specific

The acceptability of the same adverse-event risk can differ substantially depending on the clinical need for the medicine.

References

1. European Medicines Agency. Quinolone- and fluoroquinolone-containing medicinal products - referral. European Medicines Agency.

2. European Medicines Agency. Fluoroquinolone and quinolone antibiotics: PRAC recommends new restrictions on use following review of disabling and potentially long-lasting side effects. 2018.

3. European Medicines Agency. Disabling and potentially permanent side effects lead to suspension or restrictions of quinolone and fluoroquinolone antibiotics. 2018.

4. van der Linden PD, Sturkenboom MCJM, Herings RMC, et al. Increased risk of Achilles tendon rupture with quinolone antibacterial use, especially in elderly patients taking oral corticosteroids. Arch Intern Med. 2003;163:1801-1807.

5. Daneman N, Lu H, Redelmeier DA. Fluoroquinolones and collagen associated severe adverse events: a longitudinal cohort study. BMJ Open. 2015;5:e010077.

6. Morales DR, Slattery J, Pacurariu A, Pinheiro L, McGettigan P, Hargreaves S. Relative and absolute risk of tendon rupture with fluoroquinolone and concomitant fluoroquinolone-corticosteroid therapy: population based nested case-control study. Clin Drug Investig. 2019;39:205-213.

7. Arabyat RM, Raisch DW, McKoy JM, Bennett CL. Fluoroquinolone-associated tendon-rupture: a summary of reports in the Food and Drug Administration's adverse event reporting system. Expert Opin Drug Saf. 2015;14:1651-1658. doi:10.1517/14740338.2015.1085968.

8. Shu Y, Zhang Q, He X, Liu Y, Wu Y, Chen Y. Fluoroquinolone-associated suspected tendonitis and tendon rupture: a pharmacovigilance analysis from 2016 to 2021 based on the FAERS database. Front Pharmacol. 2022;13:990241.

9. Kim GK. The risk of fluoroquinolone-induced tendinopathy and tendon rupture: what does the clinician need to know? J Clin Aesthet Dermatol. 2010;3:49-54.

10. Lewis T, Cook J. Fluoroquinolones and tendinopathy: a guide for athletes and sports clinicians and a systematic review of the literature. J Athl Train. 2014;49:422-427.

11. Guzzinati S, et al. Biological and chemical changes in fluoroquinolone-associated tendinopathies: a systematic review. Br Med Bull. 2019;130:39-49. doi:10.1093/bmb/ldz006.

12. European Medicines Agency. Guideline on good pharmacovigilance practices (GVP) Module IX — Signal Management.

13. European Medicines Agency. Guideline on good pharmacovigilance practices (GVP) Module V — Risk Management Systems.

14. European Medicines Agency. Relevant historical and current European product information for fluoroquinolone-containing medicinal products concerning tendon injury and tendon rupture.