What's Actually Going On
Most spinal infection is haematogenous, meaning it arrives through the bloodstream. Germs travelling in the blood lodge in the well-supplied bone at the edge of a vertebra (the endplate), then spread across the poorly-supplied disc into the vertebra next door. That's why the disc and both neighbouring vertebrae are usually infected together, the pattern called spondylodiscitis. Left unchecked the infection erodes bone, can collapse a vertebra, and can track backwards into the spinal canal to form a pocket of pus (an epidural abscess) that squeezes the nerves.
Staphylococcus aureus is the most common culprit, but the list is long: other bacteria, resistant strains like MRSA, fungi, and tuberculosis among them. That's exactly why a sample has to be taken to identify the germ, rather than guessing an antibiotic, because the wrong guess treats the wrong bug.
How to Identify It
There is no test on the treatment couch that diagnoses a spinal infection. It's a blood-test-and-scan diagnosis. The signs that should trigger that workup:
- Back pain that does not behave mechanically: constant, present at rest, worse at night, not eased by position. This is the core tell.
- Fever or feeling systemically unwell (present in only a share of cases, so its absence does not reassure).
- Raised inflammatory markers on a blood test ESR / CRP — most useful screen
- A confirming MRI scan imaging reference standard
- A risk factor for germs in the blood (injecting drug use, recent infection or heart-valve infection, diabetes, weakened immunity, dialysis, recent spinal procedure).
What it is not. A mechanical strain eases with rest, follows a load-and-recover pattern, and comes with normal blood markers and no fever. A spinal infection shares the "sinister back pain" picture with a spinal tumour, and blood tests plus a scan are what separate them. The point of this card is that the infection is the one that mimics a strain most convincingly early on.
Why no sensitivity or specificity figures are shown: this run found no clean diagnostic-accuracy numbers for individual signs, and the literature is heterogeneous enough that inventing them would be dishonest. The action is refer for bloods and a scan, not lean on a test statistic.
The Debate
Where the guidance is still moving
Older view: long antibiotic courses (12 weeks or more) are needed for pyogenic vertebral osteomyelitis.
Newer evidence: 6 weeks is no worse than 12 in most confirmed, uncomplicated cases (Bernard 2015, N=351; supported by a 2025 network review). A genuine, evidence-based shortening.
Older default: conservative management, surgery late.
Newer view: a conditional shift toward earlier surgery in selected patients (deficit, instability, abscess, failed therapy). Patient selection does the work; the antibiotics-first core still holds for the uncomplicated case.
Older habit: start antibiotics on suspicion; repeat MRI to confirm cure.
Newer view: take a sample before antibiotics in the stable patient, then follow the blood markers rather than routinely re-scanning someone who is improving (IDSA 2015).
Honest Limitations
The evidence isn't physiotherapy evidence
Almost all of it is infection-medicine, radiology, and spinal-surgery evidence, built in patients who already reached hospital. There is no physiotherapy trial for spinal infection because physiotherapy isn't the treatment. This card's value sits upstream of any hands-on work: recognise and refer.
It's mostly observational and heterogeneous
Outside the single antibiotic-duration trial, most of the literature is retrospective case series, and even the definition of the disease varies between studies. That's why this card gives directions and triggers, not precise percentages.
The mimic works in real clinics
Spinal infection is diagnosed late precisely because early on it looks like ordinary back pain to anyone not screening for the red flags. A busy first-contact clinician is exactly where that miss happens.
The Nuance
Most uncomplicated spinal infections are treated successfully with antibiotics alone, and the course can be as short as 6 weeks when the case is confirmed and uncomplicated. Surgery isn't a default, it's reserved for nerve damage, instability, a compressing pocket of pus, blood poisoning, or a failure of medical treatment. The danger here isn't usually the treatment, it's the delay to diagnosis: a new nerve deficit recovers poorly even after surgery, so the window to catch it is short. None of this is a physical-therapy decision. The physical therapist's whole job is to get the patient into this pathway early.
