30 Aug 2017

Management and controversies of classical Hodgkin lymphoma in pregnancy

Management and controversies of classical Hodgkin lymphoma in pregnancy

The goal of managing classical Hodgkin lymphoma (cHL) in pregnancy is to obtain good long-term outcomes for both the mother and fetus. Given the excellent outcomes outside of pregnancy, the goal of treatment should remain curative. There remains a tension and debate regarding the timing of chemotherapy, the curative nature of such treatment and the timing of delivery. Moreover, the aim during pregnancy should be to minimize fetal toxicity and optimize perinatal outcomes. The management of cHL within pregnancy was covered within the excellent recent British Committee for Standards in Haematology guidelines, but with necessary brevity. By reviewing the literature over the last 30 years, herein we discuss the options for management during each trimester. Critical organogenesis occurs between 2 and 8 weeks post-conception; during which time the immature fetus is vulnerable to cytotoxic exposure. We discuss the evidence for using ABVD (doxorubicin, bleomycin, vinblastine and dacarbazine) and single agent vinblastine in the first trimester. cHL presenting in pregnancy raises complex and difficult ethical dilemmas that can cause anxiety for patients, families and physicians. Decision-making must be multi-disciplinary and holistic, taking into account the patient’s wishes, psycho-social and religious beliefs and personal circumstances. Clear communication between the haemato-oncologist, medical obstetrician, nurse specialists, midwives and neonatologists is paramount to a successful outcome.

Diagnostic approach to Hodgkin lymphoma in pregnancyIn 2011, there were 459 new cases of classical Hodgkin lymphoma (cHL) in females aged between 15 and 50 years in the UK (http://www.cancerresearchuk.org/cancer-info/cancerstats/types/hodgkinslymphoma/incidence/uk-hodgkins-lymphoma-incidence-statistics). From this data, it is logical to surmize that a small proportion of these cases will present in pregnancy each year in the UK. The fundamental principle of the diagnostic evaluation of cHL in pregnancy requires balancing disease assessment and staging investigations with the risks to maternal and fetal wellbeing. Initial evaluation is similar to that of a non-pregnant patient with cHL. Histological confirmation should be obtained by excision or core biopsy of involved tissue, procedures that are generally safe in pregnancy. Laboratory investigations should include a full blood count, erythrocyte sedimentation rate (ESR), liver and renal function, bone profile, lactate dehydrogenase (LDH), albumin, serum protein electrophoresis and human immunodeficiency virus (HIV) testing. Physiological changes that accompany increasing gestational age can result in a raised ESR, low haemoglobin and raised alkaline phosphatase (ALP) (due to placentally-derived ALP). Results therefore should be interpreted in the context of the gestation (Broek & Letsky, 2001).
Staging and imaging modalities in pregnancy

Outside of pregnancy, the management of cHL follows well-established guidelines based on initial staging assessment and prognostic criteria using FDG-PET/CT. Accurate radiological staging is the basis for the selection of an appropriate therapeutic approach to maximize survival and minimize treatment-related morbidity (Nisce et al, 1986).Outside of pregnancy, there is also growing consensus that interim PET-based responses are predictive of outcome. Interim PET [usually after two cycles of ABVD (doxorubicin, bleomycin, vinblastine and dacarbazine)] has been shown to be a clear prognostic indicator of disease-free survival (Cerci et al, 2010; Gallamini et al, 2007; Hutchings et al, 2006) and, as such, testing of PET response-adapted treatment pathways have been described in recent retrospective and non-randomized analyses (Avigdor et al, 2010; Gallamini et al, 2011) and are the subject of ongoing clinical trials.Bone marrow involvement carries a worse prognosis and occurs in 4–16% of cHL (El-Galaly et al, 2012). It is generally accepted that FDG-PET/CT can accurately detect marrow involvement, which has abrogated the need for biopsy in most cases (El-Galaly et al, 2012). However, due to the paucity of safety data in pregnancy, FDG-PET/CT has generally been avoided. In a small number of case series, the total fetal dose from FDG-PET/CT has been estimated by taking into account the fetal FDG self-dose, dose coming from the maternal tissues and CT dose (Takalkar et al, 2011; Zanotti-Fregonara et al, 2010, 2012). This ranges from 11·1 to 21·8 mGy for various trimesters in pregnancy with no documented adverse pregnancy outcomes described.The advice remains to avoid ionizing radiation during pregnancy and therefore PET CT imaging is contraindicated. It is likely that PET-based, risk-adapted therapy outside of pregnancy will become a standard of care, but such an approach should be discouraged in pregnant patients.
Standard CT ScanningThe current accepted cumulative dose of ionizing radiation to the fetus is unclear. Although the incidence of gross congenital malformations is not increased at doses <200 mGy (Wagner & Applegate, 2008), there is a suggestion from the literature that doses of 10 mGy have been associated with childhood leukaemia, with a relative risk of 1·4 (Brent, 2009; Doll & Wakeford, 1997). The fetus may have an increased lifetime risk of malignancy depending on the dose of radiation received. For example, at 15 weeks’ gestation or older, a fetal dose of 50 mGy confers a lifetime attributed cancer incidence of 2% (Austin & Frush, 2012). In the third trimester, exposure to high dose radiation may deplete cell populations but will not result in gross organ malformations.A typical fetal radiation dose delivered by a single CT examination varies depending on the anatomical scanning zone. The approximate mean mGy fetal dose is <0·005 mGy for a head CT, 0·006 mGy for a chest CT, 8·0 mGy (maximum 49 mGy) for an abdominal CT and 25 mGy (maximum 80 mGy) for a pelvic CT (Brent, 2009). The risk of fetal radiation exposure relative to the area scanned needs considering if a CT scan is required. Although the dose delivered to a fetus during conventional CT varies according to the window examined and the gestation of the fetus, there are serious concerns over using this imaging modality during pregnancy.
Magnetic resonance imagingMagnetic resonance imaging (MRI) has been proposed as an alternative to FDG-PET/CT or standard CT scanning in staging cHL to avoid ionizing radiation. There are two main concerns that arise with the use of MRI in this patient cohort; firstly, whether its ability to detect marrow involvement of disease is comparable to FDG-PET/CT and secondly, its risk to the mother and fetus.There are no prospective studies comparing modern, advanced MRI techniques and FDG-PET/CT in detecting lymphomatous bone marrow involvement. In a meta-analysis by Wu et al (2012), MRI when compared with FDG-PET/CT appeared to be inferior in terms of specificity (75·9% vs. 90·3%, P < 0·0001). However, four of the eight MRI studies included in the analysis were performed more than 10 years prior and may not accurately reflect the technological improvements to MR imaging (Kwee & Nievelstein, 2011). Advances in MRI infrastructure, sequence acquisition protocols and the ability to perform whole-body imaging make MRI a valuable staging modality in cHL.The safety concerns relate to potential teratogenicity and fetal acoustic damage (Bulas & Egloff, 2013). Loud noise is produced by the MR scan due to rapidly oscillating coils, and this may adversely affect the developing fetal acoustic system. However follow-up studies on infants who had undergone antenatal MRI suggested that hearing is preserved (Baker et al, 1994).Concern remains regarding the potential effect of gadolinium contrast on the developing fetus. Studies in rats display bidirectional passage of gadolinium across the placenta. Gadolinium accumulates in the fetal kidneys and is excreted into amniotic fluid. Reabsorption and clearance is believed to occur by the fetus swallowing amniotic fluid, which results in gadolinium diffusion back across the placenta into the maternal circulation (Sundgren & Leander, 2011). Teratogenic effects have not been observed when gadolinium was inadvertently used in early pregnancy (Barkhof et al, 1992; Shoenut et al, 1993). In view of the above, the US food and Drug Administration (FDA) recommends that gadolinium is relatively contraindicated in pregnancy but may be used if its benefits outweigh the potential risks.

Overall, the use of MRI to stage cHL remains preferable to imaging using ionizing radiation although caution should be exerted in the first trimester of pregnancy. Patients should be adequately counselled regarding the potential risks and benefits.

Fetal development and timing of treatmentEarly gestation commences from the time of conception until somite formation (up to 3 weeks from conception), sometimes termed ‘the all or nothing stage’ (Koren & Lishner, 2010) as insult to the embryo in this phase results in extremes in outcome i.e. either death or normal survival. The embryo at this stage consists of totipotent cells that are capable of repair and recovery in response to insult. Therefore teratogen exposure limited to the presomitic stage usually does not cause congenital malformations. The embryo is most sensitive to radiation injury at this stage.The embryonic period, between weeks 2 and 8 after conception (4–10 weeks after the first day of the last period) is when critical organogenesis occurs. This is the period when the developing embryo is most sensitivity to teratogenic insult. Cells have begun to differentiate, which limits the regenerative potential following an insult. Teratogen exposure or high dose radiation during this period has the greatest likelihood of causing a structural anomaly, although as the embryo matures, it becomes less sensitive to teratogens as organogenesis nears completion.The fetal phase lasts from the end of the embryonic stage to term, when growth and functional maturation of formed organs and systems occurs. Teratogen exposure in this period may affect fetal growth (e.g., intrauterine growth restriction, IUGR) and the size or function of an organ, rather than cause gross structural anomalies (Koren & Lishner, 2010).The pregnancy complications related to management of cHL are largely determined by the gestational age at presentation. The timing of treatment delivery is critical as the effect produced by a teratogen depends upon the developmental stage in which the conceptus is exposed.

Treatment of early stage and late stage cHL outside pregnancyOutside of pregnancy, patients with newly diagnosed cHL are often managed according to whether the patient has early favourable, early unfavourable and late stage disease (Follows et al, 2014).The definition and management of early stage favourable cHL is outlined in Table 1. As described in the recent BCSH guideline, two cycles of ABVD followed by 20 Gy involved field radiotherapy (IFRT) is the standard of care in the UK for early favourable cHL. Early stage disease is considered unfavourable if any of the German Hodgkin Study Group (GHSG) or European Organization for Research and Treatment of Cancer criteria are met. The GHSG HD11 study (Eich et al, 2010) found that the failure-free survival was worse when four cycles of ABVD and 20 Gy IFRT was given when compared to four ABVD plus 30 Gy or four BEACOPP (bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, prednisolone) plus 20–30 Gy. Most treating UK clinicians consider four cycles of ABVD and 30 Gy IFRT the standard of care.

The prognosis of advanced stage disease (stage III–IV) has traditionally been determined by the international prognostic score (IPS) (Hasenclever et al, 1998). When first developed, the score predicted accurately a 5-year survival between 89% (0 factors) and 56% (≥5 factors). However, advances in treatment have improved prognosis for all groups and the index now has less discriminatory power.There is currently a lack of international consensus regarding the optimal first line treatment for advanced stage cHL. ABVD is favoured in the UK, some parts of continental Europe, Australia and the USA, whereas in Germany and other parts of Europe, BEACOPPESCALATED is first line. Viviani et al (2011) randomized 331 patients to six to eight cycles of ABVD or eight cycles of BEACOPP (four cycles of the dose-intense escalated regimen followed by four baseline dose cycles) in patients with stage IIB-IV disease. Localized radiotherapy as consolidation and autologous stem cell transplantation (ASCT) at relapse was permissive within the trial. The 7-year progression-free survival (PFS) was 85% after BEACOPP and 73% after ABVD (P = 0·004). More patients required salvage therapy and an ASCT post-ABVD but higher initial toxicity was observed with BEACOPPESCALATED. Due to the efficacy and curative potential of second line treatment, the 7-year overall survival (OS) was not significantly different [89% (BEACOPP) vs. 84% (ABVD) P = 0·39] although OS was not the primary endpoint. Standard therapy outside of clinical trials in most UK centres remains six cycles of ABVD.
Pharmacology of chemotherapy in pregnancyData collected over recent decades describes the use of many chemotherapeutics at different time points during pregnancy in solid tumours, haematological malignancies and non-malignant conditions. ABVD combines an anthracycline, a vinca alkaloid, an alkylating agent and a glycopeptide antibiotic.When considering the potential drug effects on the embryo or fetus, the fetal gestation, the gestation of proposed treatment, the drug and the dose of drug need to be considered. Teratogenicity refers to the potential for congenital malformations due to disruption of organogenesis, which largely occurs between weeks 2 and 8 after conception. Fetotoxicity refers to the potential for alternations in the structure or function of an organ system that is normally formed. This can occur from the time of organogenesis until term. Any drug of high lipid solubility, i.e., loosely bound to plasma protein, and of low molecular weight will cross the placenta more effectively (Doll et al, 1989). It is not fully understood why two fetuses exposed to the same medication at the same gestation may experience different outcomes.

Alkylating agents in pregnancyAlkylating agents are key components in multi-agent chemotherapy and have been used in pregnancy in all three trimesters. Fetal malformations have been described following cyclophosphamide in the first trimester. These include toe atresia, ocular anomalies, low-set ears, cleft palate, oesophageal atresia, arm deformity and abnormal inferior vena cava development. However, data is mixed, with other reports describing how 16 other fetuses were exposed during the first trimester with no anomalies noted (Doll et al, 1989; Reynoso et al, 1987; Toledo, 1971). In two series of 15 cases of exposure to busulphan, one developed pyloric stenosis and the other unilateral renal agenesis. None of eight fetuses exposed in the first trimester developed abnormalities (Boros & Reynolds, 1977; Earll & May, 1965).Historical cases of fetal mortality or malformations have often been exposed to alkylators in combination with radiotherapy or other chemotherapy. Despite this, there is some evidence of safety in the second and third trimesters, with only a small proportion of fetal fatalities (Cantini & Yanes, 1984; Warkany et al, 1959).
Anthracyclines in pregnancyAnthracyclines are typically large molecular weight chemotherapeutics. It is postulated that they are less liable to cross the placenta due to their size, their hydrophilic molecular properties, and that they are substrates for placental P-glycoprotein (Smit et al, 1999). The lipid structure of doxorubicin, epirubicin and idarubicin differ and so do their relative safety profiles in pregnancy. Idarubicin and epirubicin exposure in pregnancy has been limited given the concerning outcomes in fetuses exposed. For example, in 17 historical cases exposed to either of these anthracyclines, there were four fetal/neonatal deaths, and two cases of fetal cardiomyopathy (Achtari & Hohlfeld, 2000; Giacalone et al, 1999; Peres et al, 2001; Siu et al, 2002). This concern has been challenged recently by data from breast cancer patients treated in pregnancy and from pre-clinical animal models suggesting that epirubicin may have a similar toxicity and pharmacological placental transfer profile to doxorubicin in the second and third trimester, although use in the first trimester remains associated with poor fetal outcomes (Mir et al, 2008a).In contrast, case series published demonstrate the relative safety of daunorubicin (O’Donnell et al, 1979; Reynoso & Huerta, 1994; Reynoso et al, 1987; Volkenandt, 1988) and doxorubicin, with the latter the preferred agent given the more extensive safety data published in pregnancy. Doxorubicin use has been described in well over 200 fetal exposures (Germann, 2004). There are rare examples of fetal anomalies, although concurrent radiotherapy and chemotherapy were typically also used in these cases. Other complications described include IUGR, intrauterine death (IUD) and second trimester miscarriage; (Ebert et al, 1997; Karp et al, 1983; Murray, 1984; Toledo, 1971; Zemlickis et al, 1992) although again, multi-agent therapy was given, including anti-metabolites that are now considered strictly contraindicated in pregnancy (Brenner et al, 2012).
Vinca alkaloids in pregnancyVinca alkaloids display high protein binding properties in the plasma, and as such have a lower placental transfer compared to other agents (Doll et al, 1989). A case series of vinca alkaloid exposure in pregnancy bears this out, with little teratogenicity displayed. In one case series of 29 exposures, a single neonate had an atrial septal defect and absent radii and fifth finger digits following multiagent chemotherapy (including cytarabine and doxorubicin) (Ebert et al, 1997). Over 100 cases of vinca alkaloid exposures are reported in pregnancy, with malformations typically only noted following first trimester exposure. Examples of fetal or neonatal deaths typically involve exposure to multiagent chemo-radiotherapy (radiotherapy, doxorubicin and prednisolone in case one; prednisolone and epirubicin in case two, ifosfamidine and dactinomycin in case three and 6-mercaptopurine, methotrexate, cytarabine and prednisolone in case four) (Cantini & Yanes, 1984; Fernandez et al, 1989; Karp et al, 1983; Peres et al, 2001).
SearchesThe authors performed searches of Medline, Current Contents, PubMed, and references from relevant articles using the search terms ‘Hodgkin(s) lymphoma’ ‘Hodgkin(s)’ ‘pregnancy’ and ‘haematological malignancy’, or ‘lymphoma’. Abstracts and reports from meetings were included only when they related to previously published work. Table 2 summarizes the fetal and maternal outcomes of the all the key reported data of cHL case series and individual case reports within the published literature over the last 30 years. Understandably, there have been no randomized studies performed, and most case series are retrospective.

Treatment options for cHL in pregnancycHL is one of the commonest cancers presenting in pregnancy, occurring in approximately 1 in 1000 to 1 in 6000 deliveries (Pentheroudakis & Pavlidis, 2006). Up to 3% of new presentations occur in pregnancy (Bachanova & Connors, 2013; Lishner et al, 1992) and as such there is a limited evidence base from which to guide decision-making, although new data has become available in recent years. Given the excellent treatment outcomes outside of pregnancy, the goal of therapy within pregnancy must remain curative. Pregnancy does not alter the natural history of cHL, fetal development or pregnancy outcomes per se (Avilés et al, 1991; Barry et al, 1962; Gelb et al, 1996). There are many examples of excellent outcomes for both mother and baby.Management decisions at any time point in pregnancy require the guidance of a multidisciplinary team. Clear communication between the haemato-oncologist, medical obstetrician, clinical nurse specialists, midwives and neonatologists familiar with the effects of maternal lymphoma and chemotherapy on the neonate is paramount to a successful outcome. It is critical to understand that lymphoma presenting in pregnancy can raise complex and difficult ethical dilemmas, which can be often anxiety-provoking for patients and treating physicians. At all times, clinical decision making must be multidisciplinary and holistic, taking into account the natural history of the disease, the patient’s wishes, ethical, psycho-social and religious beliefs and personal circumstances.
Management in the second and third trimester and fetal outcomesA key decision to make when managing a new presentation during the second and third trimester is whether chemotherapy is necessary pre-delivery. This requires careful consideration regarding the disease stage, the gestation of pregnancy, the anatomical sites involved (e.g. pelvic disease), the clinical condition of the mother, and the pace of radiologically-defined tumour growth. Some data suggests early and occasionally late-stage cHL can be safely observed with preserved outcomes for neonates and the mother. This is particularly true if presentation is near term where delivery can be expedited and treatment with standard therapy postpartum can be initiated with little delay (Bachanova & Connors, 2013; Evens et al, 2013; Gelb et al, 1996; Thomas et al, 1976), although there is no gold standard management for the second and third trimester and patients presenting with advanced cHL typically require treatment. There is increasing evidence that the use of ABVD in the second and third trimester is safe and does not result in inferior short-term outcomes when compared with deferring therapy. Seventeen patients treated with ABVD (12 in the second trimester, five in the third trimester) all delivered normal healthy infants with no malformations, spontaneous abortions, or neonatal intensive care admissions (Evens et al, 2013). There were no differences noted in this study of 40 cHL patients managed in pregnancy between rates of induction of labour (35% vs. 45%), preterm delivery (39% vs. 38%), lower segment caesarean section (LSCS) (20% vs. 18%) or infants <10th percentile size for gestational age (10% vs. 29%) when antenatal treatment was compared to deferral. The study did not present data on long-term neurocognitive childhood outcome. Other small series also provide relatively reassuring data regarding the use of ABVD in the second and third trimester. Case series of up to 19 patients (totalling over 70 cases when the series reported by Evens et al (2013) is included) have been treated with ABVD with generally good outcomes. One case of HIV-associated cHL was treated successfully with three cycles of ABVD alongside zidovudine, lamivudine and granulocyte colony-stimulating factor (G-CSF) (Klepfish et al, 2000) and another with MOPP (mechlorethamine, vincristine procarbazine and prednisolone)/ABV for two cycles plus azidothimidine with normal delivery (Okechukwu & Ross, 1998). One mother treated with ABVD during the second trimester had a stillbirth of twins at 26 weeks. A case of syndactyly requiring correction and one other minor malformation were noted in two infants who were otherwise healthy. This was considered unrelated to chemotherapy as the fetuses were not exposed during organogenesis. Otherwise, all other infants were healthy at birth (Anselmo et al, 1999; Avilés et al, 1991; Cardonick & Iacobucci, 2004; Cardonick et al, 2010; Dilek et al, 2006; Fanale et al, 2012; Klepfish et al, 2000; Lishner et al, 1992). Although not all series have long-term follow-up assessing infant outcome, a series of nine infants were all developing normally aged 5 years (Cardonick & Iacobucci, 2004) and 19 infants were developing normally at 0–10 years (Cardonick et al, 2010) having been exposed to ABVD in the second trimester. Outcomes for mothers and infants exposed to MOPP, ABV, AVD and MOPP/AVD are described in Table 2. Although there is minimal experience of dacarbazine alone in pregnancy, it is well recognized now that responses and outcomes outside pregnancy are inferior if this agent is omitted from ABVD (Ogura et al, 2013) and its inclusion is therefore considered critical.There are no reports documenting the safety of the doxorubicin, vinblastine, mechlorethamine, etoposide and prednisone (Stanford V protocol), BEACOPP or BEACOPPESCALATED. There is therefore no evidence to guide the safe use of these regimens in pregnancy, but given the high alkylator dose embedded within the regimens, their use cannot be considered desirable in pregnancy.
Management in the first trimesterAs discussed, critical organogenesis occurs between weeks 2 and 8 after conception, during which time the immature fetus is particularly vulnerable to cytotoxic exposure. The management of cHL in the first trimester is controversial, with a number of options having been explored in the literature. These include deferral of therapy until the second trimester, termination of pregnancy with immediate initiation of therapy, ABVD (or other regimens) whilst continuing pregnancy, and single agent vinblastine as a bridging therapy.Deferral of therapy whilst continuing with the pregnancy remains a reasonable approach as long as the stage, anatomical sites, clinical condition of the patient and disease tempo facilitate this. Termination of pregnancy is most typically considered for medical reasons when advanced stage, aggressive disease is causing severe and/or life threatening symptoms. Decision-making in this difficult scenario should be made with robust multidisciplinary team input and careful counselling of the patient.

ABVDPerhaps most controversially, ABVD has been used to treat cHL in the first trimester. Avilés and colleagues have collected considerable data over the last 40 years from a Mexican registry on the outcomes of all haematological malignancies treated with chemotherapy in pregnancy (Avilés & Neri, 2001; Avilés & Niz, 1988; Avilés et al, 1991, 2012). The registry has followed up both maternal and infant outcome over many years, and as such has impressively mature data to examine. From 1975 to 2008, of 15 750 cases of haematological malignancy, 143 female patients were pregnant and treated with combination chemotherapy. The most recent update from this group was published in 2012 and included data on 54 newborn infants whose mothers had received potentially curative chemotherapy during the first trimester (Avilés et al, 2012). Nineteen mothers with newly diagnosed cHL were treated with combination chemotherapy (12 ABVD; 5 MOPP/ABVD; 2 MOPP) with curative intent. All infants were delivered with a median birth weight of 3·125 kg (range 2·8–3·65 kg) and were alive at a median of 25·3 years, with normal physical, psychological, neurological, cardiac and educational status. No secondary cancers were observed in the infants and the mothers had a 20-year PFS of 89%. Alives et alare clearly convinced that ABVD in the first trimester is safe and beneficial for both the mother and the fetus.No cases of definite ABVD-induced first trimester teratogenicity could be found within the published literature. Although in addition to describing the case of the still birth of twins at 26 weeks following ABVD in the second trimester, Dilek et al (2006) also presented a single case of ABVD exposure within the first trimester leading to IUGR and partial agenesis of a metacarpal bone and hypoplasia of two phalanges. However, the fetus was exposed to three cycles of COPP (cyclophosphamide, vincristine, procarbazine and prednisone) pre-conception followed by two cycles of ABVD in the first trimester (Dilek et al, 2006).The relative lack of data regarding the safety of ABVD has led to inconsistent recommendations. Some authors have displayed an understandably cautious stance on the use of ABVD in the first trimester (Azim et al, 2010a; Pereg et al, 2007) whereas others argue it is likely to be safe for fetal development across all trimesters (Avilés et al, 2012; Bachanova & Connors, 2008, 2013).
Single agent vinblastineGiven this relative lack of safety data, other groups have investigated the use of single agent vinblastine as a form of bridging strategy to control the disease during pregnancy enabling definitive treatment to start postpartum or during later trimesters. Single agent vinblastine has demonstrated a >75% overall response rate with minimal toxicity. Interval dosing several weeks apart have been used and considered adequate for disease control. The British Columbia Cancer Agency (BCCA) group have used this approach for more than two decades in six patients with no teratogenicity and good long term fetal and maternal outcomes (Bachanova & Connors, 2013). Outside of this small series there are very few other cases of single agent vinblastine use. Two patients treated across the three trimesters had normal deliveries (female; 2·3 kg at 40 weeks and male; 3·6 kg at 35 weeks) (Wiebe & Sipila, 1994), and two patients treated in the second and third trimesters in three pregnancies delivered normal full term infants (Nisce et al, 1986). Vinblastine exposure in the first trimester has been associated with hydrocephalus and a spontaneous miscarriage at 6 weeks in two separate cases (Mulvihill et al, 1987). The significance of this is unclear given the 20% background rate of spontaneous miscarriages in the first trimester.
Relapsed cHL presenting in pregnancyCases of cHL relapsing in pregnancy are very rare. Clinicians often advise patients not to become pregnant during early follow-up post-chemotherapy (particularly the first two years when relapses typically occur). Some regimens, such as BEACOPP, can reduce fertility considerably. Despite this, cases of relapsed disease have been published, although only a single case treated with a regimen containing typical salvage chemotherapy has been documented (Peres et al, 2001). The few cases of relapsed cHL are summarized in Table 3. There is minimal data therefore on the salvage combinations typically used in cHL with combinations of platinum chemotherapy, etoposide, gemcitabine, and cytarabine commonly used as the backbone.

Salvage agentsCisplatin has been used in pregnancy in a number of solid tumours, typically lung cancer and ovarian cancer (Azim et al, 2010b; Morice et al, 2012). In the largest series, 36 cases received cisplatin in combination with additional chemotherapy (Mir et al, 2008b). Two fetal malformations were noted, although any direct causal relationship was unclear. Of interest, platinum-DNA adducts were detected in the neonates exposed. Gemcitabine is a nucleoside analogue often used in relapsed cHL. Toxicity has only been seen in supra-therapeutic doses in animal studies but gemcitabine use in pregnancy is rare and conclusions are hard to draw (Wiesweg et al, 2014). The topoisomerase inhibitor etoposide has recently been linked to genomic instability and KMT2A (previously termed mix-lineage leukaemia, MLL) rearrangement following embryonic exposure (Bueno et al, 2011). Given this, and the limited clinical experience, most clinicians would advise against its use in this setting. Cytarabine has been associated with limb deformities, particularly when used within the first trimester (Cardonick & Iacobucci, 2004). As discussed, data in cHL is very limited with these agents and any decisions made in this setting would require extrapolation of data from other settings. Patients in this situation should have their management individualized according the earlier principles discussed.
Brentuximab vedotin (BV)This agent is currently subject of a global front line clinical trial, which, if positive, may lead to its use becoming standard of care in some countries in the front line setting. In vitro data suggests that the binding of the active toxin in BV, monomethyl auristatin E (MMAE) to human serum plasma proteins ranged from 68% to 82%. MMAE is not thought likely to displace or to be displaced by highly protein-bound medicines. In vitro, MMAE was a substrate of P-glycoprotein and was not an inhibitor of P-glycoprotein at clinical concentrations. MMAE has a relatively high molecular weight of 718 (g/mol). Despite this relatively encouraging profile, pre-clinical safety data has shown that BV caused embryo-foetal lethality in pregnant female rats. At present there are no patients that have been exposed to BV reported in the literature, and its use would require careful communication of the unknown risks with the potential benefits.Pregnancy specifics related to HL treatment

AnaemiaWhen managing cHL in the pregnant patient, the clinician should be mindful of the effect of pregnancy on the disease and vice versa, in order to optimize outcomes for both (see Fig 1). Reversible causes of anaemia, such as iron deficiency, should be adequately treated to optimize the maternal baseline Hb and fetal well-being.

Dating ScanIn addition to accurate staging of disease, accurate gestation is paramount to guiding treatment options. The obstetric team should be involved at the earliest possible opportunity and dating ultrasound imaging obtained promptly.
Venous thrombosiscHL and pregnancy are both independent risk factors for venous thromboembolism (VTE). The relative risk of VTE in pregnancy is increased four to sixfold and this is increased a further fivefold in the puerperium (Heit, 2005; Pomp et al, 2008). Cancer increases the risk of VTE by more than fourfold (Heit et al, 2000). While receiving chemotherapy, cancer patients have a seven-fold risk of developing VTE as compared with other patients without cancer (Mandalà et al, 2010) and outcomes are poor in those with malignancy-induced thrombosis. Patients with cancer-related VTE are more likely to experience complications; both recurrent VTE and increased bleeding risk on anticoagulation. It is therefore important to consider the specific risk for an individual presenting with cHL in pregnancy. It is useful to refer to a validated VTE risk assessment tool, such as that published by the Royal College of Obstetrician and Gynaecologists (https://www.rcog.org.uk/globalassets/documents/guidelines/gtg37areducingriskthrombosis.pdf) to guide decisions regarding antenatal and postnatal thromboprophylaxis.
Bleomycin toxicity and the timing of neutropeniaIf ABVD is commenced antenatally, the mother should be monitored closely for respiratory complications resulting from possible bleomycin toxicity and exacerbated by diaphragmatic splitting with advancing pregnancy. The timing of chemotherapy cycles is important; one should avoid giving chemotherapy too close to the delivery date, which results in not only maternal, but also fetal neutropenia. Planned delivery at a centre with a high-risk pregnancy unit is strongly recommended.