Maternal mortality in a rural Tanzanian hospital: fatal Jarisch-Herxheimer reaction in a case of relapsing fever in pregnancy

Introduction

Relapsing fever is a spirochaetal infection caused by one of the species of Borrelia. Transmission can be either louse borne (epidemic) or tick borne (endemic). The louse borne relapsing fever (LBRF) is caused by B. recurrentis and the tick borne relapsing fever (TBRF) can be caused by at least 20 Borrelia species. ¹

Relapsing fever is reported throughout the world. It is mainly endemic in sub-Saharan Africa, with epidemics in the Horn of Africa, but also occurs in the Americas, the Middle East, Mediterranean Europe and parts of Asia. In Tanzania the most common species is B. duttoni, but other species have also been reported. ¹

Transmission of a Borrelia infection is mainly through contact with a tick or louse bearing the spirochaetes. Body lice become infected by feeding on humans with spirochaetaemia. Transmission takes place if the lice are crushed and their blood contaminates open human skin. The ticks become infected by feeding on rodents or other small mammals and transmit if they take human blood meals. People often do not notice the bite as the ticks and louse feed painlessly for a short time at night.

Sometimes the disease is acquired through accidental inoculation of infected blood or by transplacental transmission. ² There is no human-to-human transmission by aerosol, saliva, urine, faeces or sexual contact.

After entering the bloodstream with the infected body fluid of a louse or a tick, spirochaetes multiply every 6–12 h. If the concentration reaches 10⁵–10⁸ spirochaetes per mL of blood, the spirochaetes leave the bloodstream and enter the brain, eye, heart, liver and other organs where it causes swelling of the endothelial cells, microvascular leakage, perivascular mononuclear cell infiltrates, micro-abscesses and haemorrhages. The severity of the disease correlates with the number of spirochaetes in the blood. ³ Relapses occur as a result of the antigenic variation by the spirochaetes.

Relapsing fever can be detected by: the use of microscopy; blood cultures; and polymerase chain reaction (PCR). Thin blood smears with Giemsa or Wright stain can detect spirochaetes easily if the concentration of microorganisms is greater than 10⁵/mL of blood. The optimum time for taking a blood sample is during the first fever attack when spirochaetaemia is at its highest. Thick blood smears are slightly more sensitive and can detect spirochaetes with a concentration greater than 10⁴/mL of blood. PCR is the most sensitive method. ⁴ In Dodoma rural district PCR identified twice the number of cases of relapsing fever compared with thick blood smears. ¹

Patients with relapsing fever present with a sudden onset of high fever during spirochaetaemia. Fever can be accompanied by chills, sweats, muscle and joint pain and nausea. Every first episode ends with a crisis with: rigors; rising fever; rising blood pressure; and a rising pulse. This is followed after 15–30 min by profuse sweating, a falling temperature and hypotension, lasting for a few hours. If untreated, it is during this period that most mortality occurs. After the crisis an afebrile period follows and one or more relapses occur. The longer the afebrile periods lasts, the shorter and milder are the relapses. Severe complications include meningitis and myocarditis.^5,6

Jarisch-Herxheimer reaction

The Jarisch-Herxheimer reaction (JHR) resembles bacterial sepsis and can occur after the initiation of antibiotics, such as penicillin or tetracycline, for the treatment tick-borne relapsing fever. Typically, the death of the bacteria and the associated release of endotoxins occur faster than the body can remove the substances. It usually manifests within a few hours of the first dose of antibiotic as: fever; chills; rigor; hypotension; headache; tachycardia; hyperventilation; vasodilation with flushing; myalgia (muscle pain); and anxiety. The intensity of the reaction indicates the severity of inflammation. The incidence of a JHR in TBRF in the Dworkin series was 54%. ⁷ A review of six small studies showed less JHR after using penicillin compared to after the use of tetracycline in LBRF. There was no difference in mortality. ⁸

Relapsing fever and pregnancy

Only a few articles have been published about relapsing fever in pregnancy. The prevalence is reported to be 7.5% in pregnant women in central Tanzania, an endemic area. ¹ In pregnant women complications of relapsing fever range from maternal anaemia and placental dysfunction to increased abortion rates, low birth weight and high perinatal and maternal mortality. ² Barclay showed that more than 80% of the women admitted with relapsing fever delivered their babies during a febrile period. ⁵ There was also a positive correlation between the grade of spirochaetaemia, the likelihood of delivery and neonatal mortality. There is evidence that the spirochaetes can cross the placenta and lead to neonatal relapsing fever.

Jongen et al. showed a higher density of spirochaetes during pregnancy. ² Also, there was an increased risk of giving birth during a TBRF attack and increased perinatal mortality due to preterm birth and low birth weight. Perinatal death, however, was not correlated with the degree of spirochaetaemia.

Case presentation

In rural northern Tanzania, a 19-year-old woman was admitted late in the night with headache, fever, general body malaise and vomiting of 2 days duration. Earlier that day she had delivered a female baby at home, that, according to her family, was doing well. On admission the woman had a 38℃ fever with stable vital signs. Diagnosed as a malaria case she was admitted to the hospital and given quinine. The urgent blood slide revealed a high number of spirochaetes of the species B. duttoni. No malaria parasites were observed. Early in the morning she was given 1 MU procaine penicillin fortified intramuscularly. Several hours later she developed a JHR. She was found sitting in bed with a tachypnoea of 60 breaths/min, a blood pressure of 90/20 mmHg and a pulse rate of 120 beats/min. She no longer had a fever. She was transferred to the intensive care unit (ICU) but she died 30 min later after a cardiac arrest and an unsuccessful resuscitation.

The baby developed an unknown illness 6 days after delivery. She was admitted to the hospital for 3 weeks and received intravenous antibiotics. The cause remained unknown. The baby recovered and was discharged in good health. Unfortunately we were not able to retrieve the hospital file.

Discussion

In African hospitals Borrelia infections, like malaria, are easy to diagnose via thick and/or thin blood smears. Unfortunately, health care workers are often focusing on malaria, even when the incidence of relapsing fever is high and the incidence of malaria is low as in Haydom.^1,9 Bacteria is killed through the actions of degranulation of neutrophils, an increase in inflammatory cytokines during the period of exacerbation and the initiation of a respiratory burst. Oxidizing agents are produced by the respiratory burst which plays a role in killing the bacteria. Hyperventilation, seen during the JHR, is a respiratory compensation mechanism responding to the lowering oxygen saturation levels.

All drugs that can eliminate Borrelia spirochaetes can induce a JHR. In severe cases, deaths have been reported during this reaction. Penicillin produces a delayed and milder reaction but the spirochaetaemia is more persistent and clears more slowly. Tetracycline has the advantage over penicillin as it causes a significantly shorter hypotensive phase. Erythromycin appears to act as rapidly as tetracycline and should be used in children and pregnant women.

Erythromycin given to shorten the duration of hypotension and oxygen to compensate for the extra consumption during the JHR seem to be the preferred treatment for high risk pregnant women with high counts of Borrelia spirochaetes in their blood. Complications can be expected in patients with a high density of spirochaetes. ¹⁰

The treatment of choice in TBRF is tetracycline or doxycycline. If tetracyclines are contra indicated, oral erythromycin or intravenous penicillin or ceftriaxone can be given. There is no evidence of resistance of relapsing fever against antibiotics. In TBRF the infection can reside in the brain and can be a reactivated on immunosuppression. As antibiotics cannot cross the blood-brain barrier, Larsson suggests using extended treatment instead of single dose treatment. ¹¹ In our case, the woman was treated with procaine penicillin, according to the hospital protocol. This is a defendable choice as most studies comparing penicillin and tetracycline found satisfactory results for penicillin given intramuscularly. ⁵ The study by Guerrier suggests an advantage in using penicillin which gives a lower rate of JHR in low risk patients ⁸ (low spirochaetaemia, PCR+, Borrelia index below 1; see Figure 1). Penicillin given on the first day to prevent JHR and tetracycline or erythromycin (pregnant women) given for 5 days, seems to be the best approach. OPEN IN VIEWER

A JHR is more common in LBRF and severe in high risk patients (high spirochaetaemia in blood). This reaction is likely to have been the cause of death in our patient.

In high risk pregnant women with high counts of Borrelia spirochaetes in their blood, giving erythromycin to shorten the duration of hypotension and in addition oxygen to compensate the extra consumption during the JHR, seems to be the preferred treatment. In high spirochaetaemia complications can be expected as in bacterial sepsis. ³

Based on previous experience with a case of fatal JHR ¹² and other published reports, most high risk patients with very high Borrelia load in the blood smear (BI = ‘uncountable’) will die. ¹³ Intensive care with and intravenous (IV) drip, erythromycin and oxygen could make the difference.

If a patient develops JHR, treatment should be largely supportive by an IV-drip (for the hypotension) and oxygen (for respiratory burst, oxygen dependent bacterial killing). The reaction should not be mistaken for an allergic drug reaction. Antibiotic treatment should be continued, preferably with erythromycin. Untreated disease can result in a worsening of symptoms, cardiac and central nervous system involvement and, ultimately, death.

The infection of the newborn could have been due to Borrelia infection. The incubation time is 3–10 days. If so, the newborn could have been infected in utero, either while passing through the birth canal or via a tick bite received at home.

Conclusion

In African hospitals Borrelia infections are, like malaria, easy to diagnose via thick and/or thin blood smears. Our case illustrates the importance of proper diagnosis, prompt treatment and close observation after starting the treatment. The newborn should be monitored closely in order to exclude signs of transplacental infection or started early with treatment.

Like in malaria, patients with relapsing fever can be divided into low risk and high risk patients, based on the spirochaetaemia in the blood. Complications or death can be expected in high risk patients. Low risk patients with a low Borrelia index can be treated at the outpatient department or by observation for only a day. High risk patients need admission and, if possible, intensive care.