Local incidence of Jarisch–Herxheimer reaction in pregnancy following penicillin treatment for syphilis: A case series

Abstract

Background

The literature suggests that the Jarisch–Herxheimer (J-H) reaction following antimicrobial treatment of syphilis is common and may precipitate uterine activity. Local practice is to transfer syphilitic parturients beyond gestational age of viability from rural locations to a tertiary care centre for treatment. Study objectives were to delineate local incidence and risk factors for the J-H reaction among pregnant women receiving treatment for syphilis.

Methods

A retrospective chart review was conducted on pregnant women diagnosed with syphilis and treated during pregnancy at a tertiary care centre between 2012 and 2018. J-H reaction was defined as having ≥1 of the following symptoms within 24 hours of antibiotic treatment: fever (temperature ≥38°C), clinical description of a painful or itchy skin lesion, headache, hypotension (systolic blood pressure <90 mmHg), uterine contractions, or fetal heart rate decelerations. Descriptive statistical analysis was performed with mean and median used as measures of central tendency for continuous and categorical data, respectively.

Results

Fifty-eight charts were eligible for inclusion. Mean maternal age was 25.1 (SD 5.6) years, and mean gestational age was 20.4 (SD 9.5) weeks when syphilis was diagnosed. One patient (1/58, 1.7%) met J-H reaction criteria. Mean gestational age at delivery was 37.1 (SD 3.4) weeks. One stillbirth (1.7%) was identified.

Conclusions

The J-H reaction is less common at our centre than the literature suggests. Further research is important to identify risk factors associated with J-H reaction to optimize resource allocation in the context of treatment of syphilis during pregnancy.

Background

Infection with syphilis during pregnancy is a cause for heightened concern due to the risk of transplacental transfer and infection of the fetus. If untreated, it can result in devastating health consequences for the fetus and infant, including prematurity, hepatomegaly, skeletal abnormalities, and fetal demise (1). The risk of congenital infection has been reported as 70%–100% for untreated primary and secondary syphilis, 40% for early latent untreated syphilis, and less than 10% for late latent untreated syphilis (1,2).

The rates of syphilis have been increasing across Canada. What is unique in Manitoba and the prairie provinces is that many of the cases are occurring in reproductive-aged women, resulting in a resurgence of congenital syphilis (3). In Manitoba, the rate of syphilis infections has been increasing from 9.2 per 100,000 in 2014 to a rate of 57.9 per 100,000 in 2018. In 2014, women accounted for only 13% of infectious syphilis, but by 2018, women accounted for 48% of new cases (3). In 2020, Winnipeg had a total of 19 confirmed cases of congenital syphilis, including 6 stillbirths and 9 probable cases (personal communication with Dr Pierre Plourde, Medical Officer of Health, Winnipeg Regional Health Authority, Winnipeg, Manitoba, Canada).

The Society of Obstetricians and Gynaecologists of Canada (SOGC) issued a statement that enhanced screening for sexually transmitted infections (STIs), including syphilis, should be considered at multiple time points in pregnancy (3). Universal antepartum screening for syphilis infection is widely recommended because treatment of antepartum maternal syphilis with appropriate antibiotics attenuates adverse outcomes in both the mother and the child (1). During pregnancy, penicillin is the only recommended antibiotic effective in treating maternal disease, preventing transmission to the fetus, and treating established fetal disease. Additionally, there have been no clinically relevant penicillin-resistant strains of Treponemal pallidum identified. Therefore, pregnant patients diagnosed with syphilis who have an allergy to penicillin should receive desensitization followed by penicillin therapy (1).

While penicillin is an excellent treatment for syphilis, it may precipitate the Jarisch–Herxheimer (J-H) reaction. The J-H reaction is an acute febrile reaction accompanied by headache, myalgia, rash, and hypotension and can also precipitate uterine contractions, preterm labour, and/or non-reassuring fetal heart rate tracings in pregnant women (1). These symptoms are thought to result from the release of treponemal lipopolysaccharide from dying spirochetes and an increase in circulatory cytokine (1). As a precaution for the risk of preterm labour, pregnant women in our catchment area (which includes remote communities in Nunavut, northern Manitoba, and Northwestern Ontario), who have syphilis and who are beyond the gestational age of viability (typically 23+0 at our centre) have historically been transferred from the periphery to a tertiary care centre to receive penicillin treatment. The rationale for this practice is based on the considerable geographical distance compounded by often inclement weather impacting the management of patients in referring peripheral sites prompting transfer to sites with the resources to handle preterm birth. Physicians at our centre have anecdotally observed the incidence of J-H reaction among their treated patients to be less than the incidence quoted in the literature (approximately 40%) (4). Formal quantification of the incidence of J-H reaction as well as risk factors associated with J-H reaction in our clinical setting could assist in more cost-effective resource allocation, specifically where patient transfer for treatment after fetal viability is concerned. The primary objective of this study is to delineate the local incidence of the J-H reaction following the administration of benzathine penicillin G for treatment of syphilis in pregnancy.

Methods

This study was a retrospective chart review of all women diagnosed and treated for syphilis during pregnancy at a single tertiary care centre in Winnipeg, Manitoba, between 2012 and 2018. The Cadham Provincial Laboratory Information and Management Systems (LIMS) was used to identify pregnant women with positive syphilis serology (treponemal and non-treponemal tests). Clinical data were extracted from hospital records, including maternal age, gestational age of diagnosis, clinical stage of syphilis, type of antibiotic used and dose, and timing of J-H reaction. In accordance with the provincial testing algorithms, the diagnosis of syphilis was made based on the traditional algorithm (non-treponemal screening followed by treponemal confirmation) for cases prior to August 2013 and based on the reverse algorithm (treponemal screening followed by non-treponemal confirmation) for cases after August 2013. Stage of syphilis was determined by a pediatric infectious disease specialist (author AAL) based on the information available in the hospital record. All prior syphilis testing results for residents of Manitoba are available through an online clinical database in our province (eChart) to inform clinical decision making; however, prior results are not always available for out-of-province cases transferred to Winnipeg for management. In cases where the clinical stage of syphilis could not be ascertained from the hospital record, the clinical stage was listed as unknown. Charts in which a historical syphilis infection was noted without evidence of reinfection were excluded from this case series.

In our province, clinical staging and management are informed by the Manitoba Syphilis Protocol. Specifically, sonographic assessment is performed for all cases of maternal syphilis. All cases of primary, secondary, and early latent (<1-year duration) syphilis in pregnancy are treated with 1–2 weekly doses of benzathine penicillin G 2.4 million units intramuscular injections (IM) with a strong culture among providers and public health officials to give 2 doses. All cases of late latent (>1-year duration) syphilis in pregnancy are treated with 3 weekly doses of benzathine penicillin G 2.4 million units IM (5). The duration of fetal monitoring following administration of benzathine penicillin G for treatment of syphilis during pregnancy is not standardized but determined by the most responsible physician caring for the patient. During the study period, the patient was monitored for 2–6 hours and then discharged from hospital but educated about J-H reaction and advised to stay within the city, self-monitor, and re-present for concerning symptoms.

J-H reaction was defined as having any one of the following within 24 hours of antibiotic treatment: fever (temperature ≥38°C), clinical description of a painful or itchy skin lesion (ie, rash), headache, hypotension (defined as systolic blood pressure <90 mmHg), uterine contractions (as identified by clinical staff using tocometry) or fetal heart rate (FHR) decelerations documented by clinical staff (staff at our centre are trained in FHR interpretation based on consensus guidelines of Fetal Health Surveillance authored by the SOGC) (6).

Descriptive statistical analysis, including calculation of the frequency of J-H reaction, was performed using SAS (v9.4, SAS Institute, Cary, NC, USA). Descriptive statistics were performed using mean and median for continuous data and frequencies for categorical data. This research was approved by the Manitoba Research Ethics Board (H2018:039).

Results

A total of 90 charts were reviewed for inclusion. Of these, 32 were excluded for reasons including previous syphilis infection treated prior to current pregnancy (n = 6), being diagnosed post-partum (n = 6), medical abortion (n = 1), or lack of information in the chart due to delivery being outside of Winnipeg and lack of documentation (n = 19). Of the 58 charts that met inclusion criteria, the mean maternal age was 25.1 (SD 5.6) years and mean gestational age at diagnosis of syphilis was 20.4 (SD 9.5) weeks. The mean gestational age of delivery was 37.1 (SD 3.4) weeks. Half of the patients were diagnosed with an infectious stage of syphilis, either primary, secondary, or early latent (Table 1). All patients received benzathine penicillin G 2.4 million units IM, and pre-treatment venereal disease research laboratory (VDRL) titres were recorded (Table 2). Of these, 16/58 patients were beyond 23 weeks of viable gestational age. This cohort includes both local cases and those transferred for treatment.

Table 1:

Patient characteristics

Characteristic	Mean (SD), or %
Mean maternal age, y	25.1 (5.6)
Mean gestational age at diagnosis, weeks	20.4 (9.5)
Mean gestational age at delivery, weeks	37.1 (3.4)
Median pre-treatment VDRL	16
Syphilis stage
Primary	19.0
Secondary	6.9
Early latent	25.9
Late latent	3.4
Unspecified	44.8*
Patients from out of province	27.6

* No documentation of what stage of syphilis the patient has is recorded in the chart and no description of clinical signs or symptoms

VDRL = Venereal disease research laboratory

Table 2:

Patient VDRL titres prior to treatment with 2.4 million units of benzathine penicillin G

Pre-treatment VDRL titre	No. (%) of patients with VDRL titre; n = 58
1:1	2 (3.4)
1:2	3 (5.2)
1:4	2 (3.4)
1:8	10 (17.2)
1:16	13 (22.4)
1:32	9 (15.5)
1:64	7 (12.1)
1:128	5 (8.6)
1:256	1 (1.7)
Unspecified*	6 (10.3)

* In cases of unspecified stage of syphilis, 3 doses of benzathine penicillin G were required to be considered adequately treated

VDRL = Venereal disease research laboratory

One patient (1/58) met the criteria for defining J-H reaction following syphilis treatment. The patient with symptoms of J-H reaction was a 17-year-old female who was diagnosed with syphilis at 26 weeks gestation. The clinical stage of syphilis could not be determined from the chart. Her pre-treatment VDRL was 1:16, and she was administered 2.4 million units IM of penicillin. Two hours following treatment, she experienced uterine activity that was detectable on tocometry; this resolved within 10 minutes, and she had no other J-H reaction symptoms. The patient had no documented ultrasound findings suggestive of congenital syphilis. This patient went on to have a healthy spontaneous vaginal delivery at 38 weeks gestation.

One patient was a 26-year-old female with poor prenatal care who was diagnosed with syphilis at approximately 27 weeks gestation. Pre-treatment VDRL was 1:16, and she received one dose of 2.4 million units IM of penicillin. She then failed to return for subsequent treatments and repeat bloodwork. Her second dose was administered at 30 weeks gestation after she presented to the hospital with preterm premature rupture of membranes (PPROM) and was admitted for antepartum hemorrhage. Her VDRL titres were 1:8 at this time. Approximately 25 hours after administering her second dose of penicillin, FHR decelerations were noted, and delivery was expedited via Caesarean section. Given that symptoms occurred outside of the designated window of 24 hours following penicillin administration and the concomitant diagnosis of PPROM, this patient was not considered to have had a J-H reaction.

One patient in the study was a 20-year-old female with known secondary syphilis found to have a stillbirth at 29 weeks gestation due to placental abruption. Her pre-treatment VDRL was 1:8 dilutions, for which she received 3 weekly doses of benzathine penicillin G 2.4 million units IM during her second trimester of pregnancy, 2 months prior to the abruption. This patient had a known history of gestational hypertension and used cocaine the day prior to the pregnancy loss. Pregnancy loss was attributed to placental abruption in the context of known risk factors, including cocaine use and gestational hypertension.

Owing to the small number of cases meeting J-H criteria, it was not feasible to undertake statistical analysis to assess for correlation between disease characteristics (eg, clinical stage, gestational age, and pre-treatment VDRL) and risk for J-H reaction following treatment.

Discussion

Although the literature reports a 40%–45% incidence of J-H reaction among pregnant women following treatment for infectious syphilis, our findings were inconsistent with this (7,8). Based on the information gathered in our study, the incidence of J-H reaction is lower than reported in the literature. Our study found an incidence of 1.7%.

The J-H reaction during pregnancy has not been extensively described, but a few cohorts are available in the literature for comparison. A prospective study by Klein et al actively followed 33 admitted pregnant patients being treated for syphilis for 24 hours following treatment. They reported a 45% incidence of J-H reaction following treatment and noted that the cases of J-H reaction were concentrated in patients with primary and secondary syphilis. The most common symptom of J-H reaction reported in this prospective analysis was fever, uterine contractions, and decreased fetal movement (8). A retrospective study by Myles et al reported a 40% incidence of J-H reaction among a cohort of 50 pregnant patients treated for syphilis at a hospital in Chicago, Illinois, USA. At the centre where this study was being performed, patients were routinely admitted and monitored for 24 hours following treatment for syphilis. This group found no association between J-H reaction and gestational age and found that their rate of J-H reaction was comparable for those with infectious and latent syphilis. Out of the 31 available FHR tracings in their study, 42% developed regular uterine contractions, and 12 developed variable decelerations with spontaneous resolution within 24 hours of penicillin administration (7).

While both studies report higher rates of J-H reaction, the contrasting rates of J-H reaction between our article and these reports may be multifactorial. Clinical characteristics, including the stage of syphilis, pre-treatment VDRL titre, and gestational age have been suggested to modify the likelihood for J-H reaction following treatment. Specifically, Klein et al found that J-H reaction clustered within cases of primary and secondary syphilis compared to late latent syphilis. Only 1/4 of our cohort had definitive primary or secondary syphilis compared to the reports by Klein et al and Myles et al where 60% and 50% of their cohorts had either primary or secondary syphilis, respectively. As such, the lower incidence of J-H reaction in our cohort compared to previous reports may, in part, be due to important differences in clinical characteristics.

The lower rate of J-H that we report herein may also be related to important methodological differences between these reports and our own. Specifically, in both Klein et al and Myles et al, all patients were admitted to hospital and monitored for 24 hours following their initial dose of benzathine penicillin. This is not typical practice at our centre, where patients are discharged to the local community and educated to re-present for symptoms of concern for adverse reactions. As such, we may not have identified more minor changes defining features of J-H reaction that may have been noted with continuous medical observation. For example, while Klein et al identified fever as the most common feature of J-H reaction in their cohort, the mean temperature was 38.1°C with a maximal fever of 38.4°C in one patient, and Myles et identified a mean rise in temperature of 0.76°F. Certainly, while our definition of J-H reaction is largely comparable to that of other reports in the literature, important differences in length and intensity of active monitoring are likely to explain the vastly different reports in the incidence of J-H reaction following treatment for syphilis in pregnancy. While these important differences in clinical characteristics and methodology may contribute to the vastly different rate of J-H reaction observed, it follows—given the self-resolving nature of the J-H reaction—that future research should explore not only the rate of occurrence of J-H reaction by classic definition but also clinically important J-H reaction (eg, symptomatology in sufficient severity to generate clinical concern for fetal well-being and to justify intensified monitoring or therapeutic intervention).

Limitations

The retrospective nature of our study is both a strength and a limitation. We were able to retrospectively analyze detailed hospital records pertaining to patient demographics and specific information such as stage of syphilis, gestational age of diagnosis, gestational age of treatment, response to treatment, and overall clinical course. This additional information provided a substantial amount of context to the raw data on which our findings are based. Moreover, the retrospective design allowed for a review of cases of a long period (2012–2018). This time frame helped ensure that the results and conclusions drawn from our study are consistent with the burden of syphilis in our province. However, this methodology was limited by the available documentation. Notably, one quarter of our cohort came from out of province and, as such, treating physicians were not able to easily call upon previous results through the provincial clinical database nor record it in the medical chart. This contributed significantly to the degree of missing data within this dataset, most notably for the stage of syphilis. Given that many patients were seen at the hospital only for treatment of their syphilis infection and went on to deliver in other obstetrical sites, pregnancy outcome data were also missing for a large portion of the dataset.

Based on the high rates of J-H reaction reported in the literature, researchers have explored avenues for generating cost-saving guidelines. Rac et al recommend that all viable pregnancies receive their first dose of benzathine penicillin G in a labour and delivery department under continuous fetal monitoring for at least 24 hours and that the remaining doses can be administered in an outpatient setting (9). Limitations notwithstanding, the data that we contribute herein suggests a need to continue to study and characterize J-H reaction in pregnancy. While strict criteria are important, further research into the most clinically meaningful features of J-H reaction as well as risk factors for J-H reaction constitute important knowledge gaps. A better ability to predict both the occurrence and clinical impact of J-H reaction would assist in better resource allocation for treatment of syphilis during pregnancy.

Conclusions

Based on the information gathered in our study, the perceived incidence of J-H reaction may be lower than reported in the literature when patients are managed as outpatients following their treatment for syphilis during pregnancy. Further research is important to identify risk factors associated with J-H reaction as well as to qualify clinically significant features of J-H reaction to optimize resources allocation in the context of treatment of syphilis during pregnancy. Given the burden on disease in our province and the centralization of most high-risk obstetrical care, we consider our centre to be an opportune site to conduct further prospective research to better define the contemporary Canadian experience with syphilis during pregnancy.

Ethics Approval:

This study was approved by the University of Manitoba Research Ethics Board (H2018:039).

Informed Consent:

N/A

Registry and the Registration No. of the Study/Trial:

N/A

Funding:

No funding was received for this work.

Disclosures:

The authors have nothing to disclose.

Peer Review:

This article has been peer reviewed.

Animal Studies:

N/A