HELLP Syndrome: Hemolysis, Elevated Liver Enzymes, and Low Platelets

DR STONE: Mrs N is a 32-year-old businesswoman from California. I became familiar with her case 2½ years ago when I was a postdoctoral fellow at the University of California, San Francisco, and she presented to her local physician with right-upper–quadrant pain. She has graciously agreed to join us at Hopkins today for Grand Rounds.

Mrs N, was the right-upper–quadrant pain the first symptom of your illness?

PATIENT: No. For 2 or 3 weeks before that, I had noticed some swelling around my eyes and over my brow bone. My eyes looked puffy. On Friday, the day before the pain started, my husband and I went away for the weekend. During the 2-hour drive, I realized that my ankles were swollen. To relieve the swelling, I put my feet up on the dashboard of the car. After about 30 minutes, the definition of my ankles had returned but my knees and upper legs had become very swollen. They stayed swollen for the rest of the weekend.

DR STONE: Did you notice any changes in your urine around that time? Was there more or less urine, a change in color, or any other difference?

PATIENT: None that I recall.

DR STONE: Please tell us about the abdominal pain.

PATIENT: On Saturday, I developed some pain near my ribs on the right side. I thought I had pulled a muscle in my rib cage. I couldn't get comfortable in bed, despite trying several positions. At about 3:00 in the morning, I took a bath to try to relax. Lying in the bathtub didn't help, so I did some stretching exercises on the floor. That didn't help, either.

DR STONE: Did the pain remind you of heartburn?

PATIENT: No.

DR STONE: How long did the pain last?

PATIENT: For several hours. Then it subsided on its own. It was gone by morning.

DR STONE: Did you have any history of gallbladder problems before you got this right-upper–quadrant pain?

PATIENT: No.

DR STONE: Had you ever been jaundiced or had problems with your liver?

PATIENT: No.

DR STONE: In addition to your leg swelling and right-upper–quadrant pain, you had one other symptom.

PATIENT: On Saturday evening, my husband and I were playing backgammon. My eyes seemed to be playing tricks on me: I saw extra dots on his dice and thought he was cheating. When I looked again, I realized I was wrong and he was actually playing fairly. I attributed my vision problems to dim lighting in the room.

DR STONE: Mrs N had a history of migraine headaches, treated with ibuprofen and sumatriptan. At the time of this illness, she had not taken either medication in months. She had had the usual childhood illnesses and had undergone a tonsillectomy when she was 5 years old. At 12 years of age, she had been admitted to her local hospital with a kidney problem that neither she nor her mother could recall well. She had stayed overnight, undergone a procedure that was probably an intravenous pyelogram, and been discharged the next day. Since then, she has not had any known kidney problems. She was taking no medications at the time of her presentation and had no drug allergies. She is of Greek ancestry but was born and grew up in the United States. Her family history is remarkable only for her father's death from melanoma when he was 68 years old. She had no family history of connective tissue disease. She did not smoke, she drank only socially, and she had not used illicit drugs.

Mrs. N, when did you decide to see a doctor about your symptoms?

PATIENT: When I returned home on Sunday night, I was so exhausted that I made an appointment to be seen on Monday.

DR STONE: Two important pieces of information have been intentionally omitted from the history so far. The first is that Mrs N was 32 weeks' pregnant at the onset of this illness. The second is that she is my wife, and her real name is "M."

The next afternoon I took M to the obstetrician. She was feeling much better. The right-upper–quadrant pain was gone. Her blood pressure, taken by the nurse's aide, was normal. She had a little pedal edema, but the physician's assistant and I considered that consistent with a 32-week pregnancy. Her deep tendon reflexes were within normal limits.

Feeling completely reassured, we were about to leave the doctor's office when the physician's assistant, in an act of thoroughness for which we will forever be grateful, took M's blood pressure again. It was actually 160/110 mm Hg. We also learned that the urine dipstick had shown proteinuria (4+). M appeared to have preeclampsia. She had all 3 cardinal features: hypertension, proteinuria, and edema.

A diagnosis of preeclampsia would mean, at least, that M would be on bed rest for the 6 remaining weeks of her pregnancy. We were sent to the local hospital for routine laboratory tests and further observation of M's blood pressure. When the test results returned, the obstetrician informed us that M needed to be delivered immediately. Her hematocrit was 0.26, low even for the anemia of pregnancy. Her aspartate aminotransferase (AST) level was 512 U/L (reference range, 0-35 U/L), her alanine aminotransferase (ALT) level was 420 U/L (reference range, 0-35 U/L), her lactate dehydrogenase level was 718 U/L (reference range, 122-220 U/L), and her total serum bilirubin level was 55 µmol/L (3.2 mg/dL) (reference range, 5-21 µmol/L [0.3-1.2 mg/dL]). Most of the bilirubin was indirect, suggesting hemolysis. But the real surprise was that M's platelet count was 69,000 ×10⁹/L. It was only then that the true diagnosis struck me: M had the HELLP syndrome.

In 1954, Pritchard and colleagues¹ published what was probably the first description of the syndrome that we now call HELLP. They described 3 women, all at the same stage of pregnancy as M, who had intravascular hemolysis, thrombocytopenia, and other hematologic abnormalities. Serum transaminase levels were not measured in those days, so their elevation was not part of the original description. It was only in 1982 that Weinstein² coined the syndrome's catchy acronym. The "H" stands for hemolysis, the "EL" for elevated liver enzymes, and the "LP" for low platelets.

The HELLP syndrome is one of the hypertensive disorders of pregnancy, which also include preeclampsia and eclampsia. In the United States, these conditions are among the top 3 causes of maternal mortality. Worldwide, excluding unsafe terminations of pregnancy and ectopic pregnancy, they are perhaps the leading causes of obstetrical death. Obstetricians continue to debate the precise relationship among these 3 disorders. Some experts consider HELLP to be merely a severe form of preeclampsia³ ; others believe that HELLP and preeclampsia are separate disorders with overlapping features.^{4 - 5} Preeclampsia, characterized by hypertension, proteinuria, and edema, complicates 3% to 4% of all pregnancies in the United States.⁶ When women with these features develop seizures, they are said to have eclampsia, a term derived from the Greek word for "lightning." The HELLP syndrome may follow either preeclampsia or eclampsia. Between 4% and 12% of patients with severe preeclampsia or eclampsia develop HELLP.⁷ But as we will see later, HELLP is not always preceded by either condition. Estimates of HELLP's frequency are confounded by referral bias (many patients are sent to tertiary care centers), but a reasonable estimate is about 1 in 1000 pregnancies.

In 1916, the German physician Paul Zweifel called preeclampsia and the other hypertensive disorders of pregnancy "disorders of theories" because much heat but little light had been shed on their pathogenesis. The past 80 years have been hotter, but little brighter. One current theory is that these disorders begin in the placenta and begin to end when the placenta is removed. Endothelial cell activation and "leakiness" of blood vessels also play central roles. In the hypertensive disorders of pregnancy, the uteroplacental unit is believed to become dysfunctional, perhaps because the uterine spiral arteries fail to adapt to pregnancy.^{8 - 9} As a consequence, in the late second or early third trimester, the placenta may become ischemic, triggering the release of still uncharacterized factors that damage the endothelium. Dysfunction of the trophoblasts and endothelial cells may lead to activation of platelets, to generation of vasoconstrictors such as thromboxane A₂, serotonin, and endothelin, and to lowered levels of vasodilators such as prostacyclin and nitric oxide.^{10 - 12} This cascade of events leads to preeclampsia and, in some patients, to the HELLP syndrome.¹³

The HELLP syndrome has 3 general diagnostic criteria. Most patients have microangiopathic changes, evident on a peripheral blood smear. The fragmented red blood cells (schistocytes) are indistinguishable from those in thrombotic thrombocytopenic purpura (TTP), hemolytic uremic syndrome (HUS), and other microangiopathic hemolytic anemias. As a result of the hemolysis, patients with HELLP typically have elevated serum bilirubin and lactate dehydrogenase levels but low serum haptoglobin levels. Some patients, however, have little or no evidence of hemolysis. Experts have described women with all features of the disease except hemolysis as having "ELLP."

The second criterion for HELLP is elevated liver enzymes. The particular enzyme levels measured have varied from study to study, but a reasonable guideline is an AST level more than 3 SDs above the norm for a given laboratory.^{14 - 15}

The third criterion is thrombocytopenia (a platelet count <100000×10⁹/L). The thrombocytopenia results from the consumption of peripheral blood platelets as they aggregate at sites of endothelial damage. Bone marrow biopsy specimens show excessive megakaryocytes; the marrow resembles that in immune thrombocytopenic purpura. The thrombocytopenia of HELLP is occasionally severe, but both the thrombocytopenia and the hemolysis are typically milder than those of TTP/HUS.

The liver's histopathologic appearance in HELLP reflects the peripheral blood findings. The hepatic sinusoids contain hyaline deposits of a fibrinlike material, reminiscent of the fibrin deposits found in small blood vessels in TTP/HUS. These fibrinlike deposits eventually obstruct hepatic blood flow, leading to hepatocellular necrosis in the periportal regions.¹⁶ Hepatic infarction explains the right-upper–quadrant pain that affects 90% of patients with HELLP. Referred pain from the liver can produce atypical shoulder, neck, and other upper body pains that are sometimes the presenting symptoms of HELLP. In a small minority of patients, tears in the liver surface lead to subcapsular hematomas.

Other features of HELLP include malaise, nausea, and vomiting, as well as headaches caused by the hypertension. Patients with advanced disease may have seizures.

The diagnosis of HELLP is often delayed because the syndrome is easily confused with other conditions, ranging from the benign to the life-threatening: viral syndromes, regional musculoskeletal problems, gastroenteritis, peptic ulcer disease, gallbladder disorders, hepatitis, pyelonephritis, kidney stones, systemic lupus erythematosus, immune thrombocytopenic purpura, other microangiopathic hemolytic anemias, and acute fatty liver of pregnancy.¹⁷ Diagnosing HELLP is notoriously difficult in the 15% of patients who have neither hypertension nor proteinuria.^{14 - 15}

The HELLP syndrome can cause multiple organ system complications. One of the most dreaded is hepatic rupture, affecting perhaps 1% of patients; this is obviously a surgical emergency.¹⁸ Some patients with HELLP develop disseminated intravascular coagulation (DIC), but it is possible that DIC occurs only in patients with other obstetrical complications, such as abruptio placenta. Unlike patients with DIC, those with HELLP typically have normal prothrombin and partial thromboplastin times and fibrinogen levels. However, some experts maintain that if more sophisticated tests were performed for DIC, such as measurement of antithrombin III levels, most patients would be shown to have low-grade DIC.³

The multi-organ dysfunction in HELLP can lead to acute tubular necrosis and renal failure. Preeclampsia is associated with glomerular endotheliosis, whose pathologic hallmark is a thickening of the basement membranes¹⁹ ; a similar renal lesion may account for the proteinuria in HELLP. With proper supportive care, most patients fully recover kidney function.

Other complications of HELLP are pulmonary edema and the adult respiratory distress syndrome.

Unfortunately, HELLP can be fatal to both mother and fetus.

The obstetricians and nurses prepared M for an emergency cesarean section. Her thrombocytopenia (and unknown rate of platelet decline) contraindicated the placement of an epidural catheter, so she underwent general anesthesia. As I made anxious telephone calls to inform our families, I consoled myself with 2 mistaken beliefs that, I realize now, were attempts to reassure myself that everything would turn out fine. My first myth was that HELLP resolves as soon as the baby is delivered. On the contrary, the syndrome can continue, and, in 30% of patients, HELLP begins only after delivery from a seemingly normal pregnancy.⁴

My second myth was that HELLP is a disease only of the mother, and that babies typically fare well. The maternal mortality is 1% to 4%.²⁰ The perinatal mortality is considerably higher: probably 10% to 20%.¹⁵ Because HELLP usually begins well before term, the most frequent causes of perinatal mortality are complications of prematurity.

Before M went to the operating room, she and I were told that our baby had a 50% chance of needing respiratory assistance after birth. Fortunately, within a few long seconds after our daughter was delivered, the neonatologist pronounced her "a keeper." She required only a little supplemental oxygen by nasal cannula.

M's condition, however, continued to deteriorate over the next few days. In particular, her blood pressure rose even higher, requiring treatment with nifedipine. M, can you tell us what the days after delivery were like?

Patient: I was really out of it for the first few days. The magnesium sulfate I was given to prevent seizures made me feel terrible, and I was getting narcotics for pain. During the rare moments when I could think clearly, I realized that I hadn't been able to see my baby yet. I wasn't sure I was being told the whole truth when I was told that she was OK. And despite reassurances from my husband and the staff, I was afraid I was going to die. I didn't see my baby until she was 48 hours old, and then only for a few minutes. My first reaction was, "I really did have a baby." It was several more days before I was able to start bonding with her, after I felt better and she was able to leave the intensive care unit for more than a few minutes.

During the first 2 or 3 days after the delivery, one of my main problems was extreme sensitivity to light. I had to wear sunglasses and keep the room very dark. On the rare occasions when I opened my eyes, I discovered that I had double vision and even tunnel vision. I also thought that the room I was in was brown; after I recovered, I realized it was pink.

I learned something else through the nurses' early checks. In addition to measuring my blood pressure, they tested my reflexes regularly and always described them as "brisk." That sounded like a good thing to me, but in fact what I had was clonus, apparently a bad thing.

DR STONE: Yes, clonus is a very bad thing! And M's clonus was remarkable, easily 6 or 7 beats, almost sustained, a frightening illustration of her neurologic dysfunction.

On the morning of the third hospital day, as I walked past the nurses' station on M's ward, I overheard a nurse saying into the telephone, "I feel like I'm sitting on a time bomb." M's platelet count had continued to fall after delivery, and that morning it had reached 19000×10⁹/L. The physicians wanted to give M a platelet transfusion. Now I was confronted with a dilemma. So far, I had played the role of husband rather than physician, but I recalled that one must use extreme caution in giving blood products to patients with some forms of microangiopathic hemolytic anemia. In TTP, for example, platelet transfusions may do nothing more than add fuel to the fire, and they are strongly contraindicated.²¹ Ultimately, I yielded to the obstetricians' greater experience and consented to the platelet transfusion. After M received 6 units of platelets, her platelet count rose to 72000×10⁹/L, and she had no adverse sequelae. Since then, I have learned that, for unknown reasons, blood products can be given with relative impunity in the HELLP syndrome.^{4 ,22} Once M's platelets stabilized, she and I both began to believe that she was going to recover.

M came home from the hospital after 6 days and the baby 3 days afterward. Both did well. A few months later, a paper was published describing a reversible posterior leukoencephalopathy syndrome associated with acute hypertensive encephalopathy.²³ Among the patients were 3 women with a variety of neurologic and visual symptoms secondary to eclampsia. Their hypertension led to brain edema and striking changes in the white matter of the occipital lobe. The changes were completely reversible upon correction of the hypertension. Although brain edema is not a uniform finding in the hypertensive disorders of pregnancy,²⁴ such changes probably accounted for M's neurologic and visual symptoms.

There are 2 strategies for treating the HELLP syndrome. After Weinstein coined the acronym in 1982, he and others^{2 ,14} advocated immediate delivery of the fetus, regardless of gestational age, because delivery is the ultimate treatment. This strategy remains the standard of care for treating the hypertensive disorders of pregnancy. More recently, some experts have advocated trying to delay delivery by even a few days if the fetus is extremely immature (<26 weeks' gestation).^{5 ,25} This approach is extremely controversial and should be attempted only at tertiary care centers. Most obstetricians are reluctant to delay delivery for more than 48 hours.

From the moment of diagnosis, all patients with HELLP require strict bed rest, primarily in the left lateral decubitus position, which has long been known to ameliorate the hypertension of pregnancy. Magnesium sulfate is given to help prevent convulsions. Blood pressure control and meticulous fluid management are both crucial. If mothers require antihypertensive medication, most obstetricians favor either hydralazine or a β-blocker. Sodium nitroprusside is useful for refractory hypertension if delivery is imminent; for example, the drug may be given during the induction of general anesthesia to blunt the hypertensive response to intubation. Prolonged use, however, can lead to cyanide toxicity in the fetus. Intravenous fluids (colloid and crystalloid) can be used to counteract the plasma volume contraction that accompanies HELLP, but excessive administration of fluids exacerbates edema and worsens the patient's overall condition.

In addition to blood pressure, the mother's deep tendon reflexes, renal function, and serum transaminases are followed closely. The fetal heart rate is monitored by Doppler ultrasound. Corticosteroids may be given to enhance the fetus' lung development. Even thrombocytopenic women can have amniocentesis to gauge fetal lung maturity, but amniocentesis seldom alters management.

The fetus is delivered under the most controlled circumstances possible. If the conditions of the mother and fetus have remained stable, the mother is given a trial of labor. Many women have normal vaginal deliveries. If the mother or fetus becomes unstable, a cesarean section is performed. Blood products are given as needed, and obstetricians use them fairly liberally. Although their value is unproven, plasma exchange²⁶ and dexamethasone²⁷ are both considerations for critically ill women with prolonged postpartum HELLP. With good supportive care, the great majority of women recover completely. Most babies also tolerate well the stressful circumstances of their birth.

There is little information on the likelihood of the HELLP syndrome recurring. Although 1 study reported an increased risk of preeclampsia in later pregnancies, the frequency of recurrent HELLP syndrome was only about 3%.²⁸ Now that M and I have 2½ years of perspective on her illness, the relatively low risk of recurrence encourages us to consider a second pregnancy.

In summary, all physicians should know that a cardinal symptom of the HELLP syndrome is right-upper–quadrant pain. Clinicians examining pregnant women in a primary care or subspecialty setting should have a low threshold for ordering a complete blood count, urinalysis, and liver function tests, even if patients' complaints are nonspecific. If there is any suspicion of the HELLP syndrome, the physician should examine the peripheral blood smear. Finally, pregnant women need regular, accurate blood pressure measurement. A blood pressure of 140/90 mm Hg, normal in most nonpregnant patients, may indicate serious disease in pregnant women.