Manage emerging and pandemic infectious disease in pregnancy

In one line

Pregnancy is an immunological and physiological state that converts an ordinary respiratory or emerging viral infection into a potentially lethal one, so the consultant task is to recognise the pregnant woman as a high-risk host, treat aggressively and early, vaccinate proactively, and refuse to let the historical reflex of excluding pregnant women from trials and treatment leave them unprotected in the next outbreak.

Mechanism & pathophysiology

The pregnant woman is not simply a non-pregnant patient who happens to be carrying a fetus; she is a different host, and three coupled changes explain why she fares worse with respiratory and emerging infection.

The first is immunological. To tolerate a semi-allogeneic fetus, maternal adaptive immunity is remodelled rather than suppressed — a shift in the T-helper balance away from Th1 cell-mediated responses towards Th2 and regulatory phenotypes, with altered natural-killer-cell and complement regulation at the maternal–fetal interface. That remodelling is protective of the pregnancy but blunts the containment of intracellular pathogens, viruses among them. It is the reason influenza, varicella, hepatitis E and a string of emerging viruses behave more aggressively in pregnancy than the same virus in the same woman a year before or after. It is not global immunodeficiency — the febrile, antibody-producing machinery still works — but it is a measurable tilt that widens the window in which a virus can establish and disseminate before it is controlled.

The second is respiratory mechanics and oxygen economics. The gravid uterus splints the diaphragm and reduces functional residual capacity by roughly a fifth at term, so the oxygen reservoir that buffers a sick adult through an apnoea or a deteriorating pneumonia is already spent. Simultaneously, oxygen consumption rises by 20% or more to supply the fetoplacental unit and the increased maternal metabolic rate. A mother therefore starts any respiratory insult with a smaller reserve and a higher demand — she desaturates faster, tolerates less, and decompensates with a steepness that catches teams accustomed to the resilience of young non-pregnant adults. Compensatory respiratory alkalosis is the physiological baseline of pregnancy, so a "normal" or rising carbon dioxide in a breathless gravid woman signals exhaustion, not stability.

The third is the procoagulant state. Pregnancy is a physiological hypercoagulable condition — raised fibrinogen and factors VIII, IX and X, reduced protein S, acquired activated-protein-C resistance and impaired fibrinolysis — evolved to limit postpartum haemorrhage. Layered onto an infection that itself activates coagulation through endothelial injury and inflammation, this baseline tips towards thrombosis. COVID-19 made the point unmistakable: the disease is as much a thrombo-inflammatory endotheliopathy as a pneumonia, and pregnancy supplies the second hit. Venous thromboembolism, placental thrombotic pathology, and the microvascular injury that drives a pre-eclampsia-like picture all follow from the same convergence.

A fourth axis is the placenta itself as a target organ, and it explains why some emerging infections injure the fetus out of all proportion to maternal illness. The placenta is both a barrier and a route. For most respiratory viruses, including SARS-CoV-2, it is a competent barrier — direct fetal infection is rare, and the fetal harm is overwhelmingly indirect, mediated through maternal hypoxia, high fever (itself teratogenic in the first trimester and a driver of preterm labour later), placental thrombotic and inflammatory injury, and iatrogenic preterm delivery of a sick mother. For a smaller group of pathogens the placenta is permissive and the virus is directly neurotropic or fetotropic: Zika crosses and targets neural progenitor cells, cytomegalovirus and rubella cross and disseminate, and the consequence is a congenital syndrome rather than a maternal pneumonia. Reading where a new pathogen sits on this barrier-versus-route spectrum — does it threaten the mother, the fetus, or both, and by what mechanism — is the single most useful early question, because it dictates whether the clinical priority is maternal resuscitation, fetal surveillance and counselling, or both at once.

These mechanisms are general, and that generality is the point. Whatever the next novel pathogen is, the pregnant host will tend to present late (her warning signs disguised by the normal physiology of pregnancy — a resting tachycardia, a raised respiratory rate, a lower baseline blood pressure all read as "pregnant"), progress fast (small reserve, high demand), and thrombose. Pandemic preparedness in obstetrics is therefore a structural discipline, not a disease-specific one: the lesson of one outbreak is the template for the next, and a maternity service that has thought through surge triage, oxygen logistics, antiviral and corticosteroid protocols, thromboprophylaxis, infection prevention and the ethics of including pregnant women is ready for a pathogen it has never seen. The outbreaks that built the modern obstetric understanding of this host form a coherent lineage rather than a list: the 1918 H1N1 and 1957 H2N2 influenza pandemics first quantified the excess maternal mortality; the 2009 H1N1 pandemic re-proved it in the modern era and established early antiviral treatment; the 2015–16 Zika epidemic redefined what mild maternal illness can do to a fetus; the West African Ebola outbreak exposed the lethal combination of a haemorrhagic infection and a haemorrhage-prone organ; and SARS-CoV-2 from 2020 assembled, in real time, the cleanest evidence base and the sharpest ethical argument the field has. Each taught a transferable rule, and a consultant who can recite the rules can manage the next pathogen on the morning it arrives.

Assessment

The first move with any unwell pregnant woman in an outbreak is to assume the pregnancy is amplifying the illness and to stage her against that expectation rather than against a non-pregnant baseline.

• Recognise the camouflaged deterioration. Use an obstetric early-warning chart (MEOWS) with pregnancy-adjusted trigger thresholds; the single most sensitive sign of a deteriorating respiratory infection is a rising respiratory rate, which is also the sign most often dismissed because tachypnoea is mild and the woman looks deceptively well. A respiratory rate above 20–22 sustained, any new oxygen requirement, or a saturation below 94% on air in a pregnant woman with a respiratory virus is a regional/tertiary-referral trigger, not a watch-and-wait.
• Quantify reserve, not just the snapshot. Oxygen saturation, respiratory rate, work of breathing and trend over hours; a venous or arterial blood gas where available — a normalising or rising carbon dioxide in a tachypnoeic woman is impending failure. Lactate, full blood count, renal and liver function, and coagulation including D-dimer (interpreted against the elevated pregnancy reference range, not the non-pregnant cut-off).
• Establish the pathogen and the differential. Targeted molecular testing (SARS-CoV-2, influenza A/B PCR, respiratory viral panel where available); chest imaging should not be withheld for pregnancy — a chest radiograph delivers a negligible fetal dose and CT pulmonary angiography is justified when pulmonary embolism is a real possibility, because a missed maternal PE is the greater harm. In an emerging-disease context, take a structured exposure and travel history and notify according to the national outbreak case definition early; in South Africa the National Institute for Communicable Diseases is the route for novel-pathogen confirmation and reporting.
• Stage the host's modifiers — the SA layer. HIV status and, if positive, CD4 count and viral suppression; tuberculosis co-infection; anaemia; diabetes; gestation and fetal wellbeing. An immunosuppressed mother may not mount the fever and leucocytosis that would otherwise flag severe infection, so a blunted inflammatory picture in an unwell low-CD4 woman is a reason for more concern, not reassurance, and lowers the threshold for escalation.
• Assess the fetus as a second patient, separately. Maternal hypoxia and high fever both threaten the fetus; continuous monitoring at viable gestations, a steroid and magnesium plan if preterm delivery becomes likely, and a clear understanding that fetal compromise can be the first visible sign of maternal decompensation — but that stabilising the mother almost always comes before delivering the fetus, because a hypoxic mother cannot be rescued by emptying the uterus.

The major pathogens — what changes the plan

The generic high-risk-host approach is the frame; the specific pathogen then redirects it.

• COVID-19 is a maternal disease with a thrombotic and endothelial signature. The danger is a respiratory deterioration that outpaces the team's expectation, a pre-eclampsia-like picture that can mimic or coexist with true pre-eclampsia (the distinction matters because delivering for presumed pre-eclampsia that is actually COVID endotheliopathy may not help), and venous thromboembolism. The fetus is largely a bystander harmed by maternal hypoxia and prematurity; vertical transmission is low.
• Influenza is the prototype. A pregnant woman with an influenza-like illness in the season is a treat-on-suspicion patient, and the threat is a primary viral or secondary bacterial pneumonia with rapid respiratory decompensation. The third trimester carries the highest mortality. The fetal risk is again indirect — high fever and preterm labour.
• Zika inverts the usual worry: the mother is mildly unwell or asymptomatic while the fetus is at major risk. A pregnant woman with a compatible exposure (travel to or residence in a transmission area, or a sexual partner with exposure) and a rash, arthralgia or conjunctivitis needs serological/molecular testing and, regardless of symptom severity, serial fetal neurosonography looking for microcephaly, ventriculomegaly, intracranial calcification and the wider congenital syndrome. Counselling — including the option of termination within the South African legal framework where severe fetal abnormality is confirmed — is part of the plan.
• Ebola and the viral haemorrhagic fevers present as a febrile, then haemorrhagic and multi-organ illness in a returning traveller or outbreak contact. The obstetric specifics are brutal: maternal and fetal mortality are very high, fetal loss is near-universal, and the products of conception — amniotic fluid, placenta, fetus — carry enormous viral loads and are a direct infection hazard to staff. Management is high-level isolation, barrier precautions for any obstetric intervention, and outbreak-protocol care; this is a notifiable, public-health-led emergency the moment it is suspected.
• A novel, unidentified pathogen is managed on the principle, not the name: treat the pregnant woman as high-risk, isolate and protect staff with appropriate PPE, notify and test through the National Institute for Communicable Diseases, support the failing organ, and commit early to including her in whatever treatment and prophylaxis the response offers.

Management

Organise the response immediate → ongoing → long-term, and let the most important principle govern every tier: in a pandemic the pregnant woman is the high-risk patient who is, by reflex, under-treated — the discipline is to give her the same evidence-based therapy as anyone else, adjusted for pregnancy, not a diluted version of it.

Immediate — resuscitate and treat the host as high-risk.

• Oxygenation first. Target maternal saturations ≥94%; escalate through nasal oxygen, high-flow nasal oxygen and non-invasive ventilation early, and do not delay intubation in a tiring gravid woman whose reserve is already small. Position (including prone positioning of the awake or ventilated pregnant patient, with appropriate uterine support) and conservative fluid balance to avoid pulmonary oedema in a leaky, low-oncotic-pressure circulation.
• Pathogen-directed antiviral, started early and empirically. For influenza, oseltamivir 75 mg twice daily for five days should be given to any pregnant or recently postpartum woman with suspected influenza without waiting for confirmation — the benefit is time-critical and the evidence that early treatment prevents ICU admission and death is strong. The same "treat on suspicion, do not wait for the test" logic applies to whatever directed therapy exists for a novel pathogen; a pregnant woman is a priority for early antiviral access, not an afterthought.
• Corticosteroids for the unwell COVID-19 mother — but matched to her two indications. Dexamethasone reduces mortality in patients needing oxygen or ventilation; in pregnancy the practical regimen is prednisolone 40 mg orally daily (or intravenous hydrocortisone) for the maternal indication, switching to dexamethasone or betamethasone if antenatal fetal lung maturation is also required at a preterm gestation, since those cross the placenta whereas prednisolone is largely metabolised by it. Steroids are for the hypoxic mother; giving them to a well, non-oxygen-requiring woman confers no benefit and potential harm.
• Thromboprophylaxis as default, not as an afterthought. Any pregnant woman admitted with COVID-19 or another significant infection warrants prophylactic low-molecular-weight heparin (e.g. enoxaparin, weight-adjusted) unless she is actively bleeding or about to deliver, and a therapeutic anticoagulation decision the moment thromboembolism is suspected. The procoagulant convergence described above makes this one of the highest-value, lowest-cost interventions available.

Ongoing — the fetus, the timing of delivery, and the unit.

• Delivery is for an obstetric indication or to rescue maternal respiratory failure, not as a treatment for the viral illness itself. Most women recover without delivery; emptying the uterus does not cure the pneumonia and exposes a critically unstable woman to surgery. The exceptions are real and must be acted on decisively: a deteriorating mother in whom delivery will meaningfully improve maternal ventilation, established fetal compromise at a viable gestation, or a spontaneous-labour/obstetric emergency. Where preterm delivery is anticipated, give antenatal corticosteroids for the fetus and magnesium sulphate for neuroprotection on the usual gestational criteria.
• Vertical transmission of SARS-CoV-2 is low, and a maternal diagnosis is not in itself an indication for caesarean or for separating mother and baby. Mode of delivery follows obstetric indications. Breastfeeding is encouraged with respiratory hygiene — the milk carries protective antibody, and the documented neonatal infections track respiratory contact, not the milk.
• Escalate level of care proactively. The pregnant woman who needs more than simple oxygen belongs in a high-care or ICU setting with obstetric, intensive-care, anaesthetic and neonatal input. In the South African public service, this is a referral and bed-availability problem as much as a clinical one, and surge planning — designating maternity high-care capacity, oxygen supply, and transfer pathways before the next wave — is part of the management, learned painfully across the COVID-19 waves.

Long-term — vaccination, follow-up and the next outbreak.

• Maternal vaccination is the single most powerful long-term tool, and it works twice: it protects the mother from severe disease and, through transplacental antibody, protects the infant in the vulnerable first months before the child can be immunised. Seasonal influenza vaccine is recommended in every pregnancy, in any trimester; COVID-19 vaccination is recommended in pregnancy and the safety data are now extensive and reassuring. The same passive-protection logic underpins the established maternal pertussis and the newer maternal RSV strategies, and is the template for any future maternal vaccine against an emerging pathogen.
• Follow up the survivors. Post-COVID functional impairment, post-ICU physical and psychological sequelae, and — for emerging neurotropic infections such as Zika — long-term neurodevelopmental surveillance of the exposed infant.

The South African and LMIC dimension

The same illness plays out differently in a resource-limited, high-HIV-burden setting, and the consultant answer must name those differences rather than recite a high-income protocol.

• The host is sicker to start. A large fraction of pregnant women carry HIV, tuberculosis or both; advanced HIV blunts the inflammatory response that flags severe infection and worsens the trajectory of every co-infection, so a low-CD4 mother with a respiratory virus is a higher-acuity patient than the same infection in an immunocompetent woman. HIV is part of the assessment of every unwell pregnant woman in an outbreak, not an aside.
• Capacity, not just clinical judgement, decides outcome. Oxygen supply, high-care and ICU beds, ventilators and trained staff are the rate-limiting steps when a wave hits, and the COVID-19 pandemic exposed how quickly maternity high-care can be overwhelmed. Surge planning — designating maternity high-care capacity, securing oxygen logistics, pre-agreeing referral and transfer pathways from district to regional to tertiary level, and protecting obstetric ICU access against competing demand — is a managerial form of clinical care that the consultant is expected to drive, not merely observe.
• Access equity is the recurring injustice. Vaccines, antivirals and novel therapeutics reach LMIC populations later and in smaller quantities, and South Africa's COVID-19 vaccine rollout to pregnant women was both delayed relative to high-income countries and constrained by which products were available. The defensible position names the evidence-based intervention and the deliverable one, advocates for earlier and fairer access, and refuses to treat a pregnant woman's lower place in a global supply queue as a clinical reason to expect less of her care.
• Infection prevention must not deprioritise maternity. PPE, isolation capacity and staff protection in labour wards and theatres are essential; an outbreak that strips a maternity unit of protective equipment converts every delivery into a transmission and a staff-loss event, and obstetric services are too often treated as lower priority than general medical or critical-care areas.

Guidelines compared

The bodies broadly agree on the high-risk-host framing and the recommendation to vaccinate; the historically interesting differences are in how confidently and how early each moved, and how each handled the absence of pregnancy-specific trial data.

The substantive recent shift to flag is the move from cautious permissiveness to active recommendation for COVID-19 vaccination in pregnancy. Early-2021 statements were hedged ("may be offered if benefits outweigh risks") precisely because the registration trials had excluded pregnant women; by 2022, accumulated observational safety and effectiveness data converted that into a positive recommendation across RCOG, ACOG/CDC, WHO and SA. South Africa's published policy lagged the high-income recommendations by some months and was constrained by which vaccines were available, a concrete instance of the access inequity discussed below.

The evidence & the controversy

The COVID-19 pandemic produced, in real time, the cleanest demonstration ever assembled of the pregnant-host principle and of the harm done by excluding pregnant women from research and rollout — and it is the controversy worth being able to argue.

The severity signal is consistent and quantified. The PregCOV-19 living systematic review showed pregnant women with COVID-19 were more likely than non-pregnant women of reproductive age to need intensive care and invasive ventilation, and — comparing pregnant women with and without the disease — that infection raised the odds of maternal death, ICU admission and preterm birth. The INTERCOVID cohort, by enrolling matched uninfected controls concurrently across 18 countries, removed much of the confounding and found a striking excess of pre-eclampsia, severe infection, ICU admission and a markedly elevated maternal-mortality risk in diagnosed women. Reassuringly, the same datasets show low vertical transmission, which reorients the clinical worry away from the fetus-as-infected and towards the mother-as-host and the fetus-as-bystander to maternal hypoxia and prematurity. Influenza tells the identical story a decade earlier: in the 2009 H1N1 pandemic pregnant women were grossly over-represented among the deaths, the risk concentrated in the third trimester, and — the actionable finding — early oseltamivir sharply reduced ICU admission and death, which is exactly why we treat on suspicion now.

The therapeutic controversy is mostly resolved into nuance. Dexamethasone saves lives in hypoxic COVID-19 but the trial excluded pregnant women, so its use in pregnancy is a defensible extrapolation rather than a directly evidenced fact — and the sensible obstetric translation distinguishes the maternal steroid (prednisolone/hydrocortisone, placentally metabolised) from the fetal-lung-maturation steroid (betamethasone/dexamethasone, placentally transferred), giving the right drug for the right indication. The genuinely live argument is about inclusion and equity. Pregnant women were excluded from the COVID-19 vaccine registration trials — Cochrane's synthesis states plainly that no trials included pregnant women — so the world rolled vaccines out to everyone except the group at highest risk of dying, then spent a year generating the observational safety data that should arguably have informed the policy from the start. That data, when it came, was unambiguous: vaccination in pregnancy is not associated with congenital anomaly, miscarriage, preterm birth or other adverse outcomes, is associated with fewer stillbirths, and protects the infant against hospitalisation through transplacental antibody. The ethical critique — that the default reflex of "protecting" pregnant women by excluding them in fact exposes them and their babies to greater harm — has become a mainstream position, captured in the call to make pregnant women the "PROTECTED, not excluded" default of pandemic research. This is the topical, examinable controversy: the next pandemic's vaccine and antiviral trials should include pregnant women by design, with appropriate pharmacovigilance, and a defensible answer says so and grounds it in the broader research-ethics principle that excluding a vulnerable group is not the same as protecting it (the consent, autonomy and justice framework in clinical-ethics-consent).

The emerging and vector-borne examples extend the same reasoning into different organ systems and a sharper equity dimension. Zika rewrote the rule that mild maternal illness means a safe fetus: the Rio cohort found that despite trivial maternal symptoms, ZIKV infection produced adverse fetal outcomes in nearly half of pregnancies — microcephaly being only the visible tip of a congenital syndrome that includes ventriculomegaly, intracranial calcification, ocular and arthrogrypotic features — and, crucially, the harm was not confined to first-trimester exposure. Ebola and the viral haemorrhagic fevers taught the bleakiest lesson of all: maternal mortality and fetal loss approached the catastrophic, the products of conception are intensely infectious (a hazard to staff and a source-control problem), and pregnant women were initially excluded from both the experimental therapeutics and the ring-vaccination programmes during the West African outbreak — the starkest version of the exclusion harm. Beyond these named examples sits the standing principle for any novel outbreak — mpox, a future pandemic influenza reassortant, an unknown "Disease X": a low threshold to treat the pregnant woman as high-risk, structured surveillance and notification, infection prevention and PPE that does not deprioritise maternity staff, and a pre-committed intent to include pregnant women in the response rather than default to omitting them. The detail of any one pathogen will be new; the host, and the obligation, will not.

Landmark trials & key evidence

Exam traps & red flags

• Treating the pregnant woman as low-risk because she "looks well." Tachypnoea is the warning, the saturation falls late, and a young gravid woman compensates impressively until she crashes. A rising respiratory rate is the trigger to escalate.
• Waiting for a positive test before starting oseltamivir. In suspected influenza, early empirical oseltamivir is the intervention that prevents ICU admission and death; delay past four days multiplies the ICU risk severalfold.
• Withholding chest imaging or CTPA "to protect the fetus." Fetal doses are negligible; a missed maternal pneumonia or pulmonary embolism is the real harm. Investigate the mother properly.
• Giving the wrong steroid for the indication. Prednisolone/hydrocortisone for the maternal COVID indication (placentally metabolised); betamethasone/dexamethasone only when fetal lung maturation is also wanted — and no steroid at all for a well, non-oxygen-requiring woman.
• Delivering to "treat" the maternal infection. Emptying the uterus does not cure the pneumonia; delivery is for obstetric indications or to rescue maternal ventilation in a deteriorating woman, and a hypoxic mother is stabilised before, not by, delivery.
• Forgetting thromboprophylaxis. The procoagulant pregnant state plus a thrombo-inflammatory infection is a high VTE risk; LMWH is a default for the admitted infected mother.
• Caesarean or mother–baby separation for maternal COVID-19 alone. Vertical transmission is low; mode of delivery follows obstetric indications and breastfeeding is encouraged with respiratory hygiene.
• Assuming a mild maternal Zika illness means a safe fetus, or that only first-trimester exposure matters. Nearly half of infected pregnancies had adverse outcomes despite mild symptoms, across all trimesters.
• Declining vaccination in pregnancy on a "no trial data" reflex. The observational safety data are now extensive and reassuring; vaccination protects mother and infant, and the registration-trial exclusion is an argument for fixing future trials, not for withholding a recommended vaccine.
• Missing the immunosuppressed host. In an HIV-positive or otherwise immunosuppressed mother a blunted fever and white count understate severity; lower the threshold to escalate, not raise it.

Evidence anchors

• Allotey J, et al. — PregCOV-19 living systematic review, BMJ 2020;370:m3320
• Villar J, et al. — INTERCOVID multinational cohort, JAMA Pediatr 2021;175:817-826
• RECOVERY Collaborative Group — Dexamethasone in hospitalised COVID-19, N Engl J Med 2020
• Ding C, et al. — COVID-19 vaccination in pregnancy safety meta-analysis, Front Public Health 2023;11:1044031
• Halasa NB, et al. — Maternal mRNA vaccination and infant protection, MMWR 2022;71:264-270
• Graña C, et al. — Efficacy and safety of COVID-19 vaccines, Cochrane Database Syst Rev 2022;12:CD015477
• Siston AM, et al. — 2009 pandemic H1N1 in pregnant women, JAMA 2010;303:1517-25
• Brasil P, et al. — Zika virus infection in pregnant women in Rio de Janeiro, N Engl J Med 2016
• Chimukuche RS, et al. — COVID-19 vaccine policy for pregnant women in South Africa, Vaccines 2022;10:2077
• RCOG — Coronavirus (COVID-19) infection in pregnancy, Version 16.0 (15 December 2022); multidisciplinary guidance (RCOG/RCM/RCPCH/PHE/PHS).
• South Africa NDoH / SAHPRA — COVID-19 vaccination offered in pregnancy at antenatal and postnatal facilities; seasonal influenza vaccine recommended in pregnancy.