Aplastic anemia in two consecutive pregnancies: obstetric and anesthetic management

Accepted Manuscript Case report Aplastic anemia in two consecutive pregnancies: Obstetric and anesthetic management E. Riveros-Perez, A. Hermesch, L. Barbour, J. Hawkins PII: DOI: Reference:

S0959-289X(17)30270-4 http://dx.doi.org/10.1016/j.ijoa.2017.08.010 YIJOA 2614

To appear in:

International Journal of Obstetric Anesthesia

Received Date: Revised Date: Accepted Date:

30 June 2017 28 August 2017 30 August 2017

Please cite this article as: Riveros-Perez, E., Hermesch, A., Barbour, L., Hawkins, J., Aplastic anemia in two consecutive pregnancies: Obstetric and anesthetic management, International Journal of Obstetric Anesthesia (2017), doi: http://dx.doi.org/10.1016/j.ijoa.2017.08.010

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Aplastic anemia in two consecutive pregnancies: Obstetric and anesthetic management E. Riveros-Perez MD1, A. Hermesch MD PhD2, L. Barbour MD MSPH3, J. Hawkins MD4. 1

Assistant Professor. Department of Anesthesiology and Perioperative Medicine. Medical College of Georgia at Augusta University. Correspondence: 1120 15th Street. Augusta, GA. 30912. [email protected] 2 Instructor/Fellow Ob/Gyn Maternal Fetal Medicine. University of Colorado School of Medicine 3 Professor of Medicine and Obstetrics and Gynecology. University of Colorado School of Medicine 4 Professor. Department of Anesthesiology. Director of Obstetric Anesthesia. University of Colorado School of Medicine Conflicts of interest: None Financial Disclosures: None

Abstract Aplastic anemia is a serious condition occasionally coexisting with pregnancy. This pathological process is associated with significant maternal and neonatal morbidity and mortality. Obstetric and anesthetic management are particularly challenging, and treatment requires knowledge of pathophysiologic mechanisms in order to provide safe care to this group of patients. We describe the successful obstetric management and labor analgesia of a patient with diagnosis of aplastic anemia in two consecutive pregnancies. Keywords: Aplastic Anemia, Obstetric Anesthesia, Thrombocytopenia, Neuraxial Anesthesia Introduction Aplastic anemia (AA) is a life-threatening, albeit rare disease, with serious implications for pregnancy.1 It is defined by the presence of pancytopenia with hypocellular bone marrow in the absence of an abnormal infiltrate or bone marrow fibrosis.2 There are limited available data about its natural history in pregnancy and there are no reported cases of recurrence with successful management in consecutive pregnancies. Here we present the first reported case of aplastic anemia in a patient who despite deterioration of her medical condition, successfully carried two consecutive pregnancies with favorable obstetric and neonatal outcomes, and in whom, labor analgesia was adapted to her condition with excellent patient satisfaction. Anesthetic, hematologic and obstetric care during each pregnancy is presented. Therapy during the peripartum period is also discussed.

Case presentation A 22-year-old patient who had two prior pregnancies (one spontaneous abortion and one living child), presented to the high-risk obstetric service in 2014 at 24weeks gestation. The patient had a complex medical history characterized by an episode of what appeared to be a selflimited hepatitis during treatment with isoniazid for a positive tuberculosis skin test 4 years prior. One year later the patient developed an episode of fulminant hepatitis diagnosed with liver biopsy. Infectious and autoimmune causes were ruled out and the etiology was unclear (isoniazid versus idiopathic). She received treatment with steroids and intravenous immunoglobulin (IVIG). After a full clinical recovery, the patient returned two years later with a clinical picture of Herpes Zoster. During her workup, pancytopenia was diagnosed. At that time, her platelet count was 20x109/L, white cell count was 3x109/L and hemoglobin was 10.7 g/dL. A bone marrow biopsy revealed trilineal hypoplasia with 5 to 10% cellularity and absence of an infiltrative process. Iron stores were absent and flow cytometry did not show any monotypic Bcell or abnormal T-cell populations. The granulocyte and erythroid maturation appeared progressive. She had a normal spleen on ultrasound and normal levels of vitamin B12, folate and thyroid-stimulating hormone (TSH). Although anti-thymocyte globulin and cyclosporine were discussed as well as bone marrow transplant, due to social and insurance issues, the patient was unable to proceed with any treatment. Heavy menses were her only manifestation of bleeding before pregnancy and no blood product transfusion was necessary. Immediately prior to this pregnancy, her platelet count remained at approximately 20x109/L, but during the second trimester, her platelet count decreased to 13x109/L. Further, she developed symptomatic anemia with a hemoglobin of 6.8 g/dL requiring transfusion of two units of packed red blood cells (pRBC). Transfusion of five units of platelets increased her count to 55x109/L. During the remainder of her pregnancy, platelet transfusions were avoided to try to prevent alloimmunization. Induction of labor was scheduled at 36 weeks because of the significant maternal risk of bleeding during a protracted non-monitored first stage of spontaneous labor. On admission, her platelet count nadired to 8x109/L, without a significant response to transfusion of 2 units of platelets. She received two doses of oral dexamethasone

for fetal lung maturity before delivery due to concerns about giving intramuscular injections of betamethasone. Coagulation tests including D-dimer and fibrinogen were normal. With the presumption of platelet alloimmunization, she received 2 units of cross-matched platelets resulting in a count of 22x109/L. This prompted a second transfusion of two units of crossmatched platelets achieving a count of 71x109/L. Labor was induced using misoprostol and oxytocin.

The anesthetic alternatives were discussed with the patient, including systemic analgesia and general anesthesia for eventual cesarean delivery, recognizing the effect of anesthetics on uterine tone and peripartum hemorrhage. The risks of spinal hematoma if a neuraxial technique were to be used were also explained to the patient. Intravenous patient-controlled analgesia (IV PCA) with fentanyl was started, with a loading dose of 50 mcg, followed by bolus injection of 20 mcg with lockout time of five minutes without basal rate, according to institutional protocol. Vaginal delivery was uneventful with a platelet count of 52x109/L. The newborn was small-for-gestational-age weighing 2152 g and had Apgar scores of 9 and 10 at one and five minutes respectively. For long acting reversible contraception (LARC), a levonorgestrel intrauterine device was placed immediately after delivery of the placenta.

Postpartum analgesia consisted of regularly scheduled oral acetaminophen 650 mg and oral oxycodone 10 mg as required. This regimen provided adequate pain control with a maximum visual analogue scale for pain of 4 out of 10 (mean 3/10). Hematology was consulted during her pregnancy. Options given to the patient after delivery included contacting the National Institutes of Health in order to receive an interventional study therapy for aplastic anemia. This was not feasible for the patient because of cost and travel away from her family. She was also offered anti-thymocyte globulin and cyclosporine during her hospitalization after delivery, but she declined because of concerns for cost. The patient was discharged on postpartum day four with a platelet count of 30x109/L, not requiring further transfusion. Three weeks later, the patient was seen in the emergency room due to vaginal bleeding with evidence of anemia with a hemoglobin of 6.5 g/dL and a platelet count of

13x109/L. One unit of pRBC and one unit of cross-matched platelets were transfused to a hemoglobin of 7.3 g/dL, a platelet count of 32x109/L and resolution of hemorrhage. Partly due to her lack of insurance, the patient received infrequent follow-up by a hematologist in her community but experienced no further bleeding complications. However, she later had her levonorgesterel intrauterine device removed because of irregular heavy vaginal bleeding. The patient returned two years later for her first prenatal visit at 20 weeks of pregnancy, complaining of dizziness and fatigue accompanied by a hemoglobin of 8.1 g/dL, and a platelet count of 21x109/L. Her high-risk pregnancy was closely observed and did not require any platelet transfusions until 32 weeks. At that time, she was admitted due to an episode of epistaxis and a fall in platelets to 9x109/L, which was unresponsive to transfusion of one unit of cross-matched platelets. Given the possibility of an existing component of immune-mediated destruction of platelets and/or erythrocytes, a trial of prednisone 1 mg/kg/day was administered with no significant response. In addition, the patient underwent a trial of HLAmatched platelet transfusion achieving a level of 24x109/L. The patient was discharged and readmitted at 36 weeks for induction of labor, given the critical need to have a controlled and highly monitored delivery with hematology, interventional radiology and the blood bank readily available. In preparation for induction, two units of pRBC were transfused to a hemoglobin of 10.1 g/dL. IVIG at a dose of 0.4 mg/Kg/day until delivery and tranexamic acid infusion were started. Interventional Radiology service was notified and placed on standby, and HLA-matched platelet transfusions were initiated after the start of uterine contractions, with four units administered, achieving a peak level of 103x109/L. Analgesia was provided with fentanyl IV PCA using the institutional protocol.

Vaginal delivery was uncomplicated and the infant weighed 2792 g with Apgar scores of 9 and 9 at one and five minutes respectively. Uterine contractility was adequate and oxytocin infusion was continued for 24 hours. An immediate postpartum tubal ligation with clips was proposed to the patient, in order to take advantage of the peak platelet count, but she refused. The patient was discharged from the hospital on her third postpartum day without signs of bleeding or anemia. Her discharge hemoglobin was 11.2 g/dL and her platelet count was 103x109/L. A

levonorgestrel-releasing intrauterine system was placed immediately after delivery of placenta. Follow up appointments with hematology and obstetrics were provided. Postpartum pain management was identical to the regimen administered during her prior delivery with excellent patient satisfaction. Discussion Our patient with aplastic anemia had two successful, successive pregnancies after the diagnosis was made. Due to thrombocytopenia in each pregnancy, neuraxial analgesia for labor pain was avoided. Analgesia was provided using intravenous PCA fentanyl. Aplastic anemia is an infrequent but serious entity when encountered in pregnancy, placing the mother and her fetus at risk for a poor outcome.3 Unfortunately, more than 50% of cases are idiopathic and a clear etiology is identified only in one-third of cases. This means that treatment of a specific cause is not commonly possible. In our case the cause was not clearly established, even though there are reports of fatal AA associated with antituberculosis therapy.4 The diagnosis of AA during pregnancy can be made either before or during gestation but in both cases it is associated with significant fetal and maternal morbidity and mortality.3 Fetuses from mothers with AA often develop intrauterine growth restriction, while the most common maternal complications are hemorrhage and sepsis.5 Although a potential causal relationship between pregnancy and AA remains a matter of debate, fertility appears to be unaffected so women with AA may become pregnant. In these cases, complications such as preterm birth, intrauterine death, stillbirth and spontaneous miscarriage have incidences ranging between 12 and 33%.6,7 Maternal morbidity from preeclampsia, heart failure, postpartum hemorrhage and infection approaches 25%.7 With severe maternal thrombocytopenia, fetal and maternal morbidity and maternal hemorrhagic complications are higher with nearly 75% requiring transfusion.8 Because of the consistent finding of thrombocytopenia in pregnant patients with AA, the involvement of the anesthesiologist in the peripartum period is paramount. In our patient with AA, a trial of prednisone was given, but not surprisingly the response was poor and she required red cell and platelet transfusions throughout her pregnancies. In her

third pregnancy, the response to platelet transfusions was adequate, but during her subsequent pregnancy, transfusion of HLA-matched platelets became necessary due to platelet transfusion refractoriness. A significant fraction of HLA-matched platelet transfusions in AA patients is often unsuccessful in increasing the platelet count. In contrast, some mismatched transfusions are sometimes effective due to a high degree of serologic cross-reactivity.9 In cases of intraoperative bleeding in an AA patient who has demonstrated a poor response to HLAmatched platelets, it is justifiable to transfuse large amount of pooled random donor platelets in hopes of finding a suitable match.

Because of the possible implications of thrombocytopenia in these patients, anesthesia should be consulted early in pregnancy. A comprehensive discussion of the different anesthetic techniques should take place. Alternative analgesic options for labor pain were explored with our patient focusing on non-neuraxial techniques. Non-pharmacological methods for labor analgesia include childbirth education, emotional support, aromatherapy, biofeedback, acupuncture and hypnosis. None of these techniques has been rigorously studied, and their efficacy is affected by individual and cultural factors.10 Therefore recommendations to use these methods as sole analgesic options, cannot be made.

Systemic analgesia with inhaled nitrous oxide and intravenous opioids are other non-neuraxial options. Few rigorous clinical studies have addressed labor analgesia with nitrous oxide and this technique is the focus of research to evaluate efficacy, safety, as well as patient satisfaction.11 Our patient refused to use nitrous oxide for labor. Opioid labor analgesia may be associated with significant side effects and lower efficacy compared to neuraxial analgesia. Both remifentanil and fentanyl have been used for labor analgesia, usually as a patientcontrolled intravenous technique. Remifentanil has a shorter context-sensitive half-life but has been associated with hypoxic episodes and cardiac and respiratory arrest during labor.12 Therefore, IV-PCA remifentanil administration requires close monitoring of respiratory variables. In a retrospective study comparing IV-PCA fentanyl with no analgesia, labor was

prolonged in the fentanyl group (possibly related to patients with longer labors requiring analgesia), but there was excellent patient satisfaction and neonatal outcome.13

Although neuraxial techniques are the first choice for labor analgesia and for operative delivery in obstetric patients, thrombocytopenia in patients with AA may contraindicate their use. There is no consensus as to the platelet count required to provide a safe neuraxial block and no evidence to support administering platelet transfusions to raise the platelet count in order to provide neuraxial anesthesia. While platelet counts greater than 50X109/L are usually acceptable to perform a cesarean delivery, 14 there is no indication that this platelet count is safe for a neuraxial block. The risk of a spinal hematoma following epidural and spinal anesthesia is approximately 1:168,000 and 1:200,000 respectively.15,16 In a review of 61 cases of spinal hematoma after a neuraxial technique between 1906 and 1984, Vandermeulen et al. found that 68% of those cases were receiving heparin and only four cases had thrombocytopenia.17 In a retrospective review of 47 cases with non-preeclamptic thrombocytopenia (platelet count <100x109/L), Tanaka et al. found that neuraxial anesthesia was provided to 91.9% of those with a platelet count of 80-99x109/L and to 48.1% of those who had a platelet count of 50-79x109/L with no neurological complications.18 The majority of these patients had autoimmune thrombocytopenic purpura or gestational thrombocytopenia. Given the non-homogenous etiology of thrombocytopenia in retrospective studies and case reports published to date, it is not possible to recommend a specific platelet count prior to doing a neuraxial block. For our patient, neuraxial techniques for labor analgesia were considered contraindicated by the anesthesia team. Following discussion with our patient, we chose to use intravenous patient-controlled analgesia with fentanyl on both occasions. We recognize that the decision to utilize a neuraxial technique in the context of thrombocytopenia may differ among practitioners.

Multimodal techniques are generally recommended for postpartum and post-cesarean delivery analgesia. A combination of neuraxial or systemic opioids with non-opioid medications reduces the side effects of different techniques, used in an isolated fashion.19 Postpartum pain

management is challenging in patients with AA, as neuraxial techniques and use of nonsteroidal anti-inflammatory drugs (NSAIDs) may be contraindicated. Because of the relative contraindications to NSAIDs and neuraxial blocks, we chose to provide postpartum analgesia with regularly scheduled acetaminophen and oxycodone, as required for breakthrough pain.

In conclusion, aplastic anemia rarely improves in pregnancy and requires a comprehensive multidisciplinary team approach that plans for a number of possible complicating scenarios. Transfusions should be managed conservatively to avoid alloimmunization. Fetal growth and well-being should also be monitored throughout pregnancy, and a vaginal route is the preferred mode of delivery. A comprehensive anesthetic plan should directly address the degree of cytopenias, infectious and bleeding complications, systemic effects of anemia, side effects of therapy and fetal distress or well-being. A “safe” platelet count threshold to perform a neuraxial block is lacking and the anesthetic technique should be chosen based on an individualized assessment. Effective contraception is critical, and long-acting reversible contraception or tubal ligation should be strongly considered given the likelihood of worsening disease progression or relapse in pregnancy.

References 1. Kaufman DW, Kelly JP, Issaragrisil S, et al. Relative incidence of agranulocytosis and aplastic anemia. Am J Hematol 2006;81:65-7. 2. Camitta BM, Storb R, Thomas ED. Aplastic anemia (first of two parts): pathogenesis, diagnosis, treatment, and prognosis. N Engl J Med 1982;306:645-52. 3. Deka D, Malhotra N, Sinha A, Banerjee N, Kashyap R, Roy KK. Pregnancy associated aplastic anemia: maternal and fetal outcome. J Obstet Gynaecol Res 2003;29:67-72. 4. Williams CK, Aderoju EA, Adenle AD, Sekoni G, Esan GJ. Aplastic anaemia associated with anti-tuberculosis chemotherapy. Acta Haematol 1982;68:329-32. 5. Ohba T, Yoshimura T, Araki M, et al. Aplastic anemia in pregnancy: treatment with cyclosporine and granulocyte-colony stimulating factor. Acta Obstet Gynecol Scand 1999;78:458-61. 6. Kwon JY, Lee Y, Shin JC, Lee JW, Rha JG, Kim SP. Supportive management of pregnancyassociated aplastic anemia. Int J Gynaecol Obstet 2006;95:115-20. 7. Bo L, Mei-Ying L, Yang Z, Shan-Mi W, Xiao-Hong Z. Aplastic anemia associated with pregnancy: maternal and fetal complications. J Matern Fetal Neonatal Med 2016;29:1120-4.

8. Shin JE, Lee Y, Kim SJ, Shin JC. Association of severe thrombocytopenia and poor prognosis in pregnancies with aplastic anemia. PLoS One 2014;9:e103066. 9. Schiffer CA. Management of patients refractory of platelet transfusion. Leukemia 2001;15:683-5. 10. Wong CA. Advances in labor analgesia. Int J Womens Health 2010;1:139-54. 11. Richardson MG, Lopez BM, Baysinger CL, Shotwell MS, Chestnut DH. Nitrous Oxide During Labor: Maternal Satisfaction Does Not Depend Exclusively on Analgesic Effectiveness. Anesth Analg 2017;124:548-553. 12. Van de Velde M, Carvalho B. Remifentanil for labor analgesia: an evidence-based narrative review. Int J Obstet Anesth 2016;25:66-74. 13. Miyakoshi K, Tanaka M, Morisaki H, et al. Perinatal outcomes: intravenous patientcontrolled fentanyl versus no analgesia in labor. J Obstet Gynaecol Res 2013;39:783-9. 14. Hahn JS, Chung KS, Lee SJ, et al. Surgical intervention in patients with aplastic anemia. Yonsei Med J 1992;33:173-82. 15. Ruppen W, Derry S, McQuay H, Moore RA. Incidence of epidural hematoma, infection, and neurologic injury in obstetric patients with epidural analgesia/anesthesia. Anesthesiology 2006;105:394-9. 16. Moen V, Dahlgren N, Irestedt L. Severe neurological complications after central neuraxial blockades in Sweden 1990-1999. Anesthesiology 2004;101:950-9. 17. Vandermeulen EP, Van Aken H, Vermylen J. Anticoagulants and spinal-epidural anesthesia. Anesth Analg 1994;79:1165-77. 18. Tanaka M, Balki M, McLeod A, Carvalho JC. Regional anesthesia and non-preeclamptic thrombocytopenia: time to re-think the safe platelet count. Rev Bras Anestesiol 2009;59:142-53. 19. Carvalho B, Butwick AJ. Postcesarean delivery analgesia. Best Pract Res Clin Anaesthesiol 2017;31:69-79.

Aplastic Anemia with Deterioration in a Subsequent Pregnancy: A Review of Anesthetic and Obstetric Considerations Highlights • Aplastic anemia is a serious condition occasionally coexisting with pregnancy • Obstetric and anesthetic management are challenging requiring an interdisciplinary approach • Conservative transfusion management is critical to prevent alloimmunization • No threshold platelet count for neuraxial anesthesia has been established • Anesthetic technique must be evaluated on a case to case basis