Clinical overview
The cardiotocograph (CTG) is the most widely used — and most widely misused — tool in intrapartum care. It records the fetal heart rate (FHR) against uterine activity, and is meant to be a screening test for fetal hypoxaemia developing in labour. Used well, it identifies the fetus that is decompensating in time to act. Used badly, it generates a torrent of false positives that drive unnecessary caesareans, while occasionally lulling clinicians into missing genuine deterioration. The registrar examiner is rarely testing whether you can name a deceleration; they are testing whether you can interpret a whole trace in clinical context and convert that interpretation into a safe, timely decision.
Interpretation is therefore a higher-order skill (HOTS): it demands that you read the four features systematically, classify the trace, weigh it against the clinical picture (gestation, meconium, sepsis, growth restriction, oxytocin, stage of labour), decide on an action, and — critically — recognise when the CTG is no longer reassuring enough to wait. CTG never stands alone. It is one input alongside the partogram, the clinical risk assessment that selected the monitoring method (see fetal-monitoring-methods), and your understanding of fetal oxygenation physiology (see contractions-fetal-oxygenation). In South Africa, where intrapartum hypoxia and birth asphyxia remain a leading avoidable contributor to perinatal death in the Saving Mothers and Babies (NCCEMD/Perinatal Problem Identification Programme) data, disciplined CTG interpretation and timely escalation are core registrar competencies.
Core knowledge
Why the fetal heart rate tells us about oxygenation
The healthy fetal heart is driven by competing autonomic inputs. Sympathetic tone tends to raise the rate; parasympathetic (vagal) tone, maturing with gestation, lowers it and generates the beat-to-beat irregularity we see as variability. A well-oxygenated fetus with an intact, awake central nervous system produces a trace with normal baseline, normal variability and accelerations. Hypoxia and acidaemia blunt this autonomic richness: variability falls, accelerations disappear, and — depending on the mechanism — characteristic decelerations and baseline shifts emerge. The CTG is thus an indirect window onto cerebral oxygenation, and its features must always be read as a pattern over time, not as isolated events.
The physiological substrate matters because it explains the rules. Each contraction transiently reduces intervillous perfusion; a fetus with good reserve tolerates this, while one with reduced reserve (placental insufficiency, growth restriction — see placental-insufficiency-response and intrauterine-growth-restriction) decompensates. Cord compression triggers a vagal reflex (variable decelerations); head compression in the second stage does likewise; uteroplacental insufficiency produces the gradual, contraction-mirroring late deceleration that is the most ominous reflex pattern.
The four features
Every CTG is interpreted on four features, then classified. Define each precisely:
- Baseline rate — the mean FHR over ~5–10 minutes, excluding accelerations and decelerations. Standard normal teaching is 110–160 bpm. Below 110 is bradycardia; above 160 is tachycardia. Baseline tends to be at the higher end preterm.
- Variability — the bandwidth of beat-to-beat fluctuation around the baseline. Normal is conventionally 5–25 bpm. Reduced variability (<5 bpm) sustained over time is concerning; it may also reflect fetal sleep (usually <40 min) or maternal sedation/opioids. A sinusoidal pattern (smooth, regular oscillation, absent variability) is pathological and suggests severe anaemia or hypoxia.
- Accelerations — transient rises ≥15 bpm for ≥15 seconds (classic term threshold). Their presence is reassuring; their absence in an otherwise normal trace is of uncertain significance in labour.
- Decelerations — transient falls in FHR, classified by timing and shape:
- Early — shallow, symmetrical, mirror the contraction (nadir with peak); benign, vagal from head compression.
- Variable — abrupt drop and recovery, varying in timing/shape; usually cord compression. Concerning ("atypical/complicated") features classically include slow return to baseline, loss of variability within the deceleration, biphasic ("W") shape, and prolonged duration.
- Late — gradual onset, nadir after the contraction peak, gradual recovery; signify uteroplacental insufficiency and are the most worrying reflex pattern. Note the Nov 2025 NICE NG229 update: a late deceleration is now classed as a RED feature in its own right — regardless of duration or whether it is repetitive (the old "repetitive/periodic" qualifier was removed because late decelerations correlate with fetal hypoxia). So even an isolated late deceleration pushes the trace toward the pathological/urgent category and prompts escalation — do not wait for them to become repetitive.
- Prolonged — a fall lasting >2 minutes. Beyond ~3 minutes this is an acute event demanding immediate action.
Uterine activity
The "toco" half of the trace records frequency and (with intrauterine pressure catheters) strength of contractions. Tachysystole is >5 contractions in 10 minutes averaged over 30 minutes; it reduces the recovery time between contractions and is a frequent, correctable cause of an abnormal FHR — especially with oxytocin or prostaglandins. The NICE NG229 emphasis on tachysystole (>5 in 10) and contractions lasting >2 minutes as triggers for review reflects this: too-frequent contractions are a common, reversible driver of fetal compromise.
Classification
Modern guidance moves away from rigid three-tier "normal/suspicious/pathological" matrices toward continual risk assessment — looking at the whole picture and the trend, and reviewing whenever risk changes (NICE NG229, Fetal monitoring in labour, 2022). It is still useful to anchor your description in the FIGO consensus categories:
| Category | Interpretation | Action principle |
|---|---|---|
| Normal | All four features reassuring | Continue; no intervention for the trace itself |
| Suspicious / non-reassuring | One feature non-reassuring | Conservative measures + closer review; find a reversible cause |
| Pathological | ≥1 abnormal feature / lack of variability with decelerations | Urgent senior review; conservative measures; plan expedited birth if no improvement |
| Acute / "need for urgent intervention" | Prolonged deceleration/bradycardia, acute event | Immediate response — exclude reversible cause, prepare for delivery |
The categories are a communication tool, not a substitute for judgement. A "suspicious" trace in a septic, meconium-stained, growth-restricted fetus on oxytocin is far more worrying than the same trace in a healthy term labour.
Assessment
Before you read a single squiggle
Confirm you are interpreting the right thing. Two recurring errors are reading the maternal heart rate as the fetal trace (a fatal pitfall when the "FHR" tracks maternal tachycardia and accelerations coincide with contractions/pushing) and misjudging paper speed. Standardise to the paper speed and scale your unit uses — South African units commonly run at 1 cm/min while many UK references assume 3 cm/min; decelerations and variability look different at different speeds, and you must read the trace in the units in which it was recorded. Never silently assume a speed. Check the date/time, the maternal observations, and that fetal and maternal signals are distinct (palpate the maternal pulse against the trace if in doubt).
A systematic reading routine
Read every CTG the same way, every time, out loud if needed. A widely taught mnemonic structure is Dr C Bravado:
- Define Risk — why is this woman on a CTG? Low-risk labour does not need continuous monitoring; the trigger to start it (meconium, pyrexia ≥38°C, hypertension, bleeding, oxytocin, growth restriction, abnormal intermittent auscultation) frames how worried to be.
- Contractions — frequency in 10 minutes; tachysystole?
- Baseline RAte.
- Variability.
- Accelerations.
- Decelerations — type, depth, duration, recovery, relation to contractions.
- Overall impression + plan.
Then synthesise: classify the trace, state the most likely physiology, name a reversible cause if present, and commit to a timed action.
Figure I17.1 — Systematic CTG readout: risk, uterine activity, four FHR features and a timed overall plan.
Interpreting the trend, not the snapshot
Hypoxia in labour usually evolves. The classic deteriorating sequence is the appearance of decelerations, then a rising baseline (sympathetic compensation), then loss of variability, and finally a terminal bradycardia. A trace that is steadily worsening over an hour is a different clinical animal from a transient dip that recovers. Document the time you reviewed, who reviewed with you, and what you decided — a buddy/fresh-eyes review is recommended for any non-reassuring trace, and is invaluable both clinically and medicolegally.
Adjuncts and their limits
CTG has poor specificity, so adjuncts have been sought to reduce unnecessary intervention. Fetal scalp blood sampling (FBS) for lactate/pH was traditionally used to confirm acidaemia before acting on a pathological trace; access, contraindications (suspected fetal bleeding disorder, maternal HIV/HBV/HCV transmission concerns, prematurity), and practical delay limit its use, and many South African units do not offer it. Fetal scalp stimulation producing an acceleration is a reassuring bedside test. Computerised CTG with the Dawes-Redman criteria belongs to the antenatal setting, not intrapartum interpretation. Where adjuncts are unavailable — common in district and many regional SA settings — the threshold to escalate or deliver on the CTG and clinical picture alone is necessarily lower; do not wait for a test you cannot perform.
Management
Interpretation is only worth something if it drives action. Management flows directly from classification.
Normal trace
Continue monitoring; document. Do not intervene on the trace — interventions for a normal CTG are unnecessary and risk iatrogenic harm.
Suspicious / non-reassuring trace — conservative measures first
The single most useful management move is to look for and correct a reversible cause. Standard conservative measures:
- Reposition the mother (left lateral) to relieve aortocaval compression and any cord compression.
- Stop or reduce oxytocin if tachysystole or a non-reassuring pattern is present — uterine hyperstimulation is a leading reversible cause.
- Correct maternal hypotension (e.g. after regional analgesia) with fluids ± vasopressor per protocol.
- Treat maternal pyrexia/sepsis — give fluids, find a source, start antibiotics; maternal fever raises fetal metabolic demand.
- Exclude acute events — vaginal examination to check progress and exclude cord prolapse, abruption, or scar dehiscence.
- Consider tocolysis (e.g. a single dose of a beta-agonist or other agent per local protocol) to relieve hyperstimulation while you decide.
Review again at a defined interval. If features improve, continue; if they persist or worsen, escalate.
Pathological trace
Call for senior/obstetric review immediately, continue conservative measures, and plan expedited birth if there is no rapid improvement — operative vaginal birth if the criteria are met and birth is imminent (see instrumental-delivery), otherwise caesarean section (see safe-caesarean-technique). The decision-to-delivery interval for fetal compromise is conventionally targeted at 30 minutes for a Category-1 ("immediate threat to life") caesarean, and units audit against this; in practice the relevant standard is "as fast as is safely achievable", and you should activate the team early rather than wait.
Acute event — the emergency drill
A prolonged deceleration / acute bradycardia (FHR persistently low, typically >3 minutes) is an obstetric emergency. Act, do not deliberate:
- Call for help — summon senior obstetric, midwifery, anaesthetic and neonatal/theatre teams; declare an emergency.
- STOP oxytocin immediately.
- Position left lateral; examine vaginally at once to exclude cord prolapse (if found, elevate the presenting part, fill the bladder, knee–chest/Trendelenburg and move to immediate delivery) and to assess progress.
- Resuscitate the mother — IV fluids, correct hypotension, give facial oxygen if hypoxaemic, consider acute tocolysis for hyperstimulation.
- If the FHR recovers within ~3 minutes and a reversible cause is corrected, you may continue with close observation. If it does not recover, proceed to immediate delivery — operative vaginal birth if safely achievable and the criteria are met, otherwise emergency (Category-1) caesarean.
- Prepare for neonatal resuscitation — ensure a resuscitation-trained provider is present at birth (see neonatal-resuscitation and initiation-of-respiration).
Always treat a sustained, unrecovering bradycardia as the real thing while you exclude the catastrophic causes — abruption (antepartum-haemorrhage), uterine rupture (uterine-rupture, especially in a VBAC) and cord prolapse — any of which mandates immediate delivery and maternal resuscitation in parallel.
South African context
- Continuous CTG requires a functioning machine, paper, and a clinician who reads it. In many district hospitals intermittent auscultation (Pinard or Doppler) per the National Integrated Maternal and Perinatal Care Guideline (NDoH, 2024) is the standard for low-risk labour, with CTG reserved for risk factors and referral. Know the indications to escalate the level of care: a pathological CTG in a setting without theatre access is a reason to activate transfer early, because the decision-to-delivery clock keeps running.
- HIV is highly prevalent; FBS is generally avoided where it risks vertical transmission, lowering the threshold to act on the trace alone. Confirm maternal status and the unit's policy.
- Birth asphyxia/intrapartum hypoxia is a recurrent avoidable theme in NCCEMD/PPIP perinatal mortality findings; the commonest contributory failures are not recognising an abnormal trace, misreading maternal for fetal heart rate, and delay in acting — exactly the interpretive failures this objective targets.
Red flags / pitfalls
- Maternal heart rate masquerading as fetal. If the "FHR" is suspiciously low/normal with accelerations coinciding with pushing, palpate the maternal pulse against the trace and consider a fetal scalp electrode. Missing this can hide a dead or severely compromised fetus.
- Treating the trace, not the woman. A CTG abnormality is a prompt to examine, find a reversible cause and reassess — not an automatic ticket to theatre. Conservative measures resolve many "pathological" traces.
- Ignoring the clinical context. The same trace means different things in a growth-restricted, meconium-stained, septic, or post-dates fetus, or one on oxytocin. Always integrate risk.
- Sinusoidal pattern dismissed. A true sinusoidal trace is pathological (severe anaemia — e.g. feto-maternal haemorrhage, rhesus disease; severe hypoxia) and warrants urgent senior assessment, not reassurance.
- Reduced variability over-attributed to sleep. Fetal sleep cycles are usually <40 minutes; sustained reduced variability beyond that, or with other abnormal features, is hypoxia until proven otherwise.
- Tachysystole left uncorrected. Failing to stop/reduce oxytocin in the face of >5 contractions in 10 minutes with an abnormal FHR is a classic, avoidable error.
- Wrong paper speed / scale. Decelerations and variability are read differently at 1 cm/min vs 3 cm/min — interpret in the units recorded.
- No fresh-eyes review and poor documentation. Non-reassuring traces should be reviewed with a colleague and decisions timed and signed; absence of this is both a clinical and medicolegal liability.
- Waiting for an unavailable adjunct. Do not delay delivery awaiting FBS/lactate where it cannot be done — escalate on the CTG plus clinical picture.
Evidence anchors
- NICE NG229 — Fetal monitoring in labour (2022). The current standalone guidance: continual risk assessment, simplified categorisation, and explicit triggers for review (tachysystole >5 in 10, contractions >2 minutes, meconium, maternal pyrexia ≥38°C). The primary interpretive framework for this objective.
- NICE NG235 — Intrapartum care (2023). Care of healthy women and babies in labour; frames who needs continuous CTG versus intermittent auscultation and the management of conservative measures and escalation.
- FIGO consensus guidelines on intrapartum fetal monitoring. Source of the normal/suspicious/pathological feature definitions and the four-feature framework used internationally.
- WHO Labour Care Guide (2020). Supersedes the partograph in many settings and structures intrapartum surveillance, including FHR documentation, alongside CTG where used.
- South African National Integrated Maternal and Perinatal Care Guideline (NDoH, 2024), NDoH. The SA source of truth for intrapartum monitoring, indications for continuous CTG versus intermittent auscultation, and referral/level-of-care pathways.
- Saving Mothers and Babies / NCCEMD and PPIP reports. SA perinatal-mortality surveillance; intrapartum hypoxia/birth asphyxia recurs as an avoidable contributor, with monitoring-interpretation and delay-in-action failures repeatedly identified.
Note on hedged facts: the numerical thresholds in this chapter — baseline 110–160 bpm, variability 5–25 bpm, acceleration ≥15 bpm/≥15 s, deceleration timing definitions, tachysystole >5 in 10, the >2–3-minute prolonged-deceleration/bradycardia threshold, the <40-minute sleep-cycle teaching, and the 30-minute Category-1 decision-to-delivery target — are standard intrapartum-monitoring teaching consistent with FIGO/NICE NG229. They are presented as conventional thresholds; confirm exact local cut-offs against the current NICE NG229 and the SA 5th-edition guideline before applying clinically. Fetal scalp blood sampling availability and HIV-related contraindications vary by unit and were stated cautiously.