OBJECTIVE: To investigate the relation of prior intracervical procedures with very preterm birth. DESIGN: A population-based case-control study. SETTING: The study was conducted in Australia between 2002 and 2004. SAMPLE: Three hundred and forty-five women having a medically indicated and 236 having a spontaneous singleton birth between 20 and 31 weeks of gestation and 796 women selected randomly from all those giving birth at ≥37 weeks of gestation. METHODS: Interview data were analysed using logistic regression. MAIN OUTCOME MEASURE: Very preterm birth. RESULTS: Very preterm birth was significantly associated with having any intracervical procedure [adjusted odds ratio (AOR) 2.07; 95% confidence interval (CI) 1.6-2.7], in particular curettage associated with abortion (AOR 1.80; 95% CI 1.2-2.6). Assisted reproductive technology procedures were significantly associated with medically indicated very preterm birth (AOR 3.07; 95% CI 1.8-5.3) and treatments for precancerous cervical changes were significantly associated with spontaneous very preterm birth, as follows: conization/cone biopsy (AOR 3.33; 95% CI 1.8-6.2) and cauterization/ablation (AOR 2.27; 95% CI 1.4-3.8). Suction aspiration for abortion, abnormal Pap smear without treatment and abortion without instrumentation were not associated with very preterm birth. CONCLUSIONS: Intracervical procedures are associated with very preterm birth. Notably, curettage rather than any other procedure associated with abortion appears to be implicated in the risk. The introduction of infection during cervical procedures may be the common link with risks found. Changing clinical practice in the management of abortion and human papillomavirus vaccination may lead to lowering the risks of very preterm birth.