Sir, Association of atrial septal defect (ASD) and pulmonary stenosis (PS) with bidirectional shunt in pregnancy is a very rare entity. Death can occur anytime during pregnancy or labour in these patients, precipitated by altered cardiovascular physiology of pregnancy.[1] A 25-year-old primi, an unbooked case with 38 weeks of gestation was posted for emergency caesarean section (CS) for cephalopelvic disproportion. In the second trimester, assessment of breathlessness (NYHA Grade III) with two-dimensional echocardiography revealed an ostium secundum type of ASD (8 mm × 10 mm) with predominantly right to left shunt (Qp: Qs - 0.85), severe PS (Maximum velocity [Vmax ]/peak pressure gradient [PGmax ] - 4.8 m/s/70.4 mm Hg), mild tricuspid regurgitation (Vmax /PGmax - 2.7 m/s/30.5 mm Hg), mild pulmonary regurgitation (Vmax /PGmax - 2.1 m/s/17.58 mm Hg), right ventricular hypertrophy (right ventricle diastolic wall thickness 6 mm) and normal biventricular function with ejection fraction of 60%. In the operation room, oxygen saturation (on room air) was 90-95%, pulse rate was 114/min regular and noninvasive blood pressure was 150/90 mmHg. The patient was premedicated with injection ranitidine (50 mg) and injection metoclopramide (10 mg) and colloid was started at 50 ml/h. Injection ceftriaxone 1 g was administered for antibiotic coverage. The initial central venous pressure was 13 cm H2O and was maintained between 12 and 14 cm H2O. Under aseptic precautions, an 18-gauge epidural catheter was inserted at L3 -L4 epidural interspace. A total volume of 20 ml local anaesthetic solution (13 ml of 2% lignocaine with adrenaline (5 μg/ml) and 1 ml of sodium bicarbonate and 1 ml (50 mg) of tramadol diluted with 5 ml saline) was administered through the epidural catheter in 5 ml increments for achieving a block height up to T4 dermatome. At the time of skin incision, injection ketamine (50 mg) and injection midazolam (1 mg) was administered in titrating doses. CS proceeded uneventfully, and a healthy male baby was delivered with Apgar score of 8 and 9 at 1 and 5 min, respectively. A bolus dose of 0.4 mg intravenous methylergometrine was administered slowly, and a slow infusion of oxytocin 10 units was started. The surgery lasted for about 60 min, and the blood loss was 500 ml. A total of 450 ml colloid and 100 ml oxytocin-normal saline infusion was given. Post-operative analgesia was provided with epidural top-up, whenever the patient complained of pain or visual analogue score was ≥4. At 48th h of follow-up, she had maximal scores for post-operative pain relief and overall satisfaction. She was discharged on the 7th post-operative day and referred to cardiology department for further management. The mortality in patients with ASD and severe PS with right to left shunt in pregnancy is unknown,[2] as many unsuccessful cases are not reported. In our case, although the patient was having reversal of shunt, she was asymptomatic and had no signs of chronic cyanosis. This may be due to hyperdynamic circulation of pregnancy producing increased pulmonary gradients and the restrictive ASD causing inadequate mixing of blood to produce clinical cyanosis.[3] Epidural anaesthesia (EA) was planned because of its slower onset and less and delayed haemodynamic variations compared to spinal and combined spinal-epidural technique.[4] General anaesthesia (GA) poses clear risks related to pulmonary vascular resistance and magnitude of the shunt. As the patient was in labour, we had no sufficient time for further cardiac evaluation. Pulse oximeter is used to assess the degree of right to left shunt during the peri-operative period.[5] Our principle was to maintain the cardiac output and to avoid the fall in systemic vascular resistance (SVR). We optimised the SVR based on mean arterial pressure to avoid the increase in shunt magnitude. Oxytocin bolus can cause direct vasodilatation and reduce SVR with a compensatory increase in heart rate and cardiac output in pregnant women. Use of methylergometrine lowered the use of vasoconstrictors[6] and added to haemodynamic stability. Colloids provide greater haemodynamic stability by remaining in the intravascular space for longer time and avoid volume overloading; this helps to decrease the magnitude of shunt and right ventricular failure.[7] The use of lower dose of lignocaine with adrenaline, bicarbonate and tramadol provided a more rapid onset and superior quality block to touch.[8] We did not institute any thromboprophylaxis measures; EA itself can reduce risk of perioperative venous stasis and embolism. To conclude, EA can be safely and effectively used for CS in pregnant patients having pulmonary vascular disease with reversal of shunts provided the SVR is maintained throughout the surgical procedure. Major haemodynamic complications and risk of embolism during GA can be avoided if EA is properly planned and administered. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest.