Adverse pregnancy outcomes following syphilis treatment in pregnancy in the UK

Introduction

Syphilis in pregnancy remains a worldwide problem with high rates of vertical transmission leading to congenital syphilis and other severe adverse pregnancy outcomes (APOs). The World Health Organisation estimates that untreated active syphilis in pregnancy results in rates of 25% second-trimester miscarriage or stillbirth, 13% pre-term birth before 32 weeks gestation, 11% neonatal death and 20% congenital syphilis amongst surviving infants. ¹ The main risk of vertical infection is from untreated early syphilis or from new infection in pregnancy. However, late syphilis can also be transmitted^2,3 and it is not always possible to accurately time syphilis infection. Penicillin treatment given early in pregnancy is highly effective at reducing syphilis-related APOs. ¹

With antenatal screening in the UK the incidence of congenital syphilis is extremely low. However, congenital syphilis is now on the rise again, and interim results from a collaboration between Public Health England and the British Paediatric Surveillance unit estimates an incidence of 0.0136 per 1000 live and stillbirths in 2010 and 0.0025/1000 in 2011. ⁴ There are few published data on adverse outcomes of syphilis in pregnancy (other than congenital syphilis) from countries with effective antenatal screening and good maternity services.

Like other centres in the UK, our clinic is managing increasing numbers of pregnant women with positive syphilis serology. We performed a service evaluation of management of syphilis in pregnancy at the Leeds Centre for Sexual Health between 2005 and 2012, to determine whether best practice was being achieved. In particular we aimed to evaluate whether women with active syphilis were experiencing severe APOs, and if so what service improvements could be made.

Methods

Women with positive syphilis serology in pregnancy managed in Leeds, with an expected outcome between 2005 and 2012, were identified from the NHS Infectious Diseases in Pregnancy Screening Programme in England. Demographic information, gestation at diagnosis, stage of syphilis, serology titres, treatment given and pregnancy outcome were obtained from the sexual health and maternity records. The women were divided into three groups: Group 1: untreated syphilis requiring treatment in the current pregnancy; Group 2: reported, but unconfirmed, adequate treatment prior to the current pregnancy (these women were also treated in the current pregnancy); Group 3: syphilis confirmed as adequately treated prior to the current pregnancy with no evidence of reinfection.

Syphilis screening in the UK is by enzyme immunoassay (EIA) for Treponema pallidum IgG and IgM, followed by treponemal antibody testing using Treponema pallidum particle agglutination assay (TPPA) and non-treponemal testing using rapid plasma reagin (RPR) or venereal disease research laboratory (VDRL) test if positive. ⁵

The stage of syphilis was classified according to UK guidelines. Positive serology in the absence of symptoms or signs is termed latent disease. Within the first two years of infection it is classed as ‘early latent’ syphilis, after two years it is considered ‘late latent’ syphilis. ⁶ Where it is not possible to accurately time the length of infection classification of late latent syphilis is made.

Severe APOs were defined as stillbirth (an infant born dead after 24 completed weeks gestation), second-trimester miscarriage (between 14 to 24 weeks gestation), very pre-term delivery (before 32 weeks gestation), neonatal death (within first 28 days of life) and confirmed congenital syphilis (defined as abnormal physical examination consistent with congenital syphilis, or a serum non-treponemal titre four-fold higher than that of the mother, or a positive dark field/polymerase chain reaction (PCR) test of body fluid). The Centers for Disease Control and Prevention (CDC) data reporting definition for possible congenital syphilis was used, defined as an infant whose mother had untreated or inadequately treated syphilis at delivery (including where the mother received recommended treatment less than four weeks before delivery). ⁷ Minor APOs were birth weight below 2.5 kg, pre-term delivery between 32 to 37 weeks gestation, polyhydramnios and non-syphilitic placental abnormalities.

Comparisons of age, ethnicity, pre-treatment RPR level, gestation at diagnosis and treatment, rates of APO and time between treatment and pregnancy outcome between women in Groups 1 and 3 were performed using the Mann Whitney U test, Chi square test and Fisher’s exact test as appropriate.

Results

Sixty cases of positive syphilis serology in pregnancy were identified in 42 women; three were excluded (one moved to another centre during pregnancy, two underwent termination of pregnancy) leaving 57 cases.

In all, 24 cases of untreated syphilis required treatment in the current pregnancy (Group 1), seven cases had reported, but unconfirmed, previous treatment and so were also treated in the current pregnancy (Group 2) and 26 cases had adequate treatment prior to current pregnancy and no evidence of reinfection (Group 3) so were not treated in this pregnancy. None had symptoms of syphilis.

The mean age was 28, 34 and 28 years for Groups 1, 2 and 3, respectively, with 71%, 43% and 61% being of black ethnicity and the median (range) pre-treatment RPR titres being positive at neat dilution (not detected to 1:128) for Group 1, positive at neat dilution (not detected to 1:8) for Group 2, and not detected (not detected to 1:4) for Group 3. There were no differences in age and ethnicity between Groups 1 and 3 (age p = 0.5, Mann Whitney U test, ethnicity p = 0.49, Chi square test).

The median time from positive syphilis blood test to starting treatment was four weeks in Group 1 and five weeks in Group 2. Within Group 1, the median time from initial blood test to treatment was three weeks for those who experienced APOs and five weeks for those who did not (see Table 1). Gestation at treatment ranged from 14 to 37 weeks. All were treated according to national guidelines for their stage of disease, ⁶ with either benzathine or procaine penicillin intramuscularly (IM). None had a recorded Jarisch–Herxheimer reaction. None of the women in Group 3 were treated in pregnancy as all had documented completed treatment prior to conception.

OPEN IN VIEWER

Table 1.

Characteristics of women in Group 1 experiencing severe APOs compared with those who did not.

	With severe APO (n = 5)	Without severe APO (n = 19)	p Value
Demographics
Median age (range) in years	29 (26–41)	28 (19–41)
Mean age in years	31	28	0.35  a
Black ethnicity	80%	63%	0.63  b
Early stage syphilis	20%	16%	1.00  b
Pre-treatment RPR
Median	Positive (neat)	Positive (neat)
Range	ND–1:64	ND–1:128
RPR ≥ 1:16	20%	16%	1.00  b
Timing of diagnosis in weeks
Median gestation at diagnosis	14	15
Mean (range) gestation at diagnosis	16.2 (11–27)	17.3 (7–36)	0.76  a
Median gestation at treatment	17	20
Mean (range) gestation at treatment	18.6 (14–28)	21.4 (13–37)	0.38  a
Median time between treatment and pregnancy outcome	3	18
Mean (range) between treatment and pregnancy outcome	5.2 (0–14)	16.2 (2–28)	0.006  a

Four babies were given treatment with 10 days of benzylpenicillin due to birth within four weeks of treatment of the mother (normal physical examination, serological non-treponemal tests available for three out of four cases showing titres less than four-fold maternal titre), fulfilling the CDC data reporting definition of possible congenital syphilis. ⁷

Four women in Group 1 were diagnosed with early latent syphilis (infection confirmed to be within the previous two years ⁶ ). One is described in case three below. A second had delayed booking and had incomplete treatment starting at 36 weeks due to difficulty engaging with services. At term delivery placental abnormalities of patchy acute inflammation, small infarcts and congestion of the maternal surfaces were noted (minor APO). The mother and infant were re-treated after delivery. The remaining two cases of early syphilis infection had no recorded APOs.

The remaining 20 women in Group 1 were classed as late latent syphilis. There was no history of stillbirth in the group, but one woman had a previous second-trimester miscarriage at 20 weeks, and another had a pre-term delivery and neonatal death at less than 37 weeks (no further information available as both were outside the UK). Four women had severe APOs, two second-trimester miscarriages, one pre-term birth at 28 weeks and one stillbirth at 32 weeks. One woman with late latent syphilis had polyhydramnios but the pregnancy outcome was not affected and the link to syphilis is uncertain. There were no cases of term congenital syphilis or term neonatal death.

There were therefore five severe APOs in total in Group 1 (5/24, 21%). These are discussed in detail below. These were all in HIV-negative women who were non-smokers, and no other obstetric or medical risk factors for APOs were found.

There were no severe APOs in Group 2 or 3, no minor APOs in Group 2 and 3/26 (12%) minor APOs in Group 3 consisting of three spontaneous pre-term births between 32 and 37 weeks. All three of those women had other risk factors for pre-term delivery (smoking, cervical incompetence with cerclage and poorly-controlled maternal diabetes with obesity).

The rate of severe APOs in those treated for the first time in the current pregnancy (Group 1) at 21% was significantly higher than the 0% in Group 3 fully treated prior to pregnancy (p = 0.02; Fisher’s exact two-tailed test).

The age, ethnicity, syphilis results, timing of diagnosis, treatment and pregnancy outcome of women in Group 1 in relation to severe APOs are shown in Table 1. There were no differences between those experiencing severe APOs versus those who did not, but the numbers are small. Of women in Group 1 with a pre-treatment RPR ≥ 1:16, 50% (two of four) experienced an APO compared with 25% (five of 20) of those with lower titres, but this was not significant (p = 0.55, Fisher’s exact test). However, there was a significant reduction of 11 weeks in the mean time between treatment and pregnancy outcome in those who experienced a severe APO versus those who did not, reflecting that all the severe APOs occurred at or before 32 weeks gestation.

Details of the severe APOs from Group 1:

Cases one and two: Two women aged 27 and 41 from sub-Saharan Africa experienced second-trimester miscarriages at 21 weeks gestation. One was within four weeks of initiation of maternal treatment with benzathine penicillin IM weekly for three weeks, thus fulfilling the CDC clinical case definition of possible syphilitic stillbirth. ⁸ Case one was in a 27-year-old woman from the Congo with a history of one first-trimester miscarriage two years earlier but no syphilis serology from that previous pregnancy. Late latent syphilis was diagnosed on screening bloods at 11 weeks (highest pre-treatment RPR 1:2) and three once-weekly treatments of benzathine penicillin IM were started at 14 weeks. A dating scan was normal but an anatomy scan at 20 weeks identified oligohydramnios with growth restriction. There was spontaneous vaginal delivery at 21 weeks. Post-mortem showed acute chorioamnionitis with foetal vascular reaction. The maternal placenta was noted to be highly irregular with no convincing clot or crater. Autoantibody screen and clotting screen on the mother was negative. Case two was an Angolan woman aged 41 with one previous child (aged seven years). Syphilis was diagnosed on screening bloods at 12 weeks (highest pre-treatment RPR ‘positive at neat dilution’) and three once-weekly treatments with benzathine penicillin IM were started at 17 weeks. Anomaly scan was normal at 20 weeks, but spontaneous vaginal delivery followed a presentation with pain and bleeding at 21 weeks. Post-mortem showed acute necrotising chorioamnionitis with foetal inflammatory response and growth of mixed anaerobic organisms on placental swab. Autoantibody screen and clotting screen on the mother was negative.

Case three: A 26-year-old white British woman had a spontaneous pre-term birth at 25 weeks gestation with subsequent neonatal death. Obstetric history included four normal pregnancies with term deliveries. Her early syphilis was diagnosed at antenatal screening at 14 weeks, with a pre-treatment RPR of 1:64. Her current male partner was also diagnosed with early syphilis and so infection during pregnancy or close to conception was suspected. She was treated with one benzathine penicillin IM injection at 18 weeks gestation. At 20 and 23 weeks she presented with per vagina (PV) bleeding; high vaginal swabs for infection were negative. She had spontaneous rupture of membranes and was commenced on erythromycin at 24 weeks, but went on to deliver a live infant at 25 weeks (birth weight 650 g) that died three days later. Histology showed changes consistent with placental syphilis of acute chorioamnionitis, acute intervillositis and foetal vessel vasculitis. No neonatal post-mortem was performed. The infant’s syphilis serology was EIA-positive, TPPA-positive (1:20560), IgM negative, RPR not detected (closest maternal RPR 1 in 32).

Case four: A 29-year-old woman from Angola had a pre-term birth at 28 weeks with spontaneous rupture of membranes. Obstetric history revealed three previous term pregnancies outside the UK with no known history of syphilis testing. Syphilis was diagnosed on antenatal screening at 27 weeks and classed as late latent in view of there being no previous serology. Treatment was initiated a week later with daily procaine penicillin, but after six days she suffered spontaneous rupture of the membranes and delivered a live infant. The mother and infant were both treated after delivery. This fulfills the CDC clinical case definition of possible congenital syphilis. ⁷ The infant had no recorded serological tests.

Case five: A 31-year-old woman from the Gambia had a stillbirth at 32 weeks. She had one previous second-trimester miscarriage outside the UK but no previous syphilis serology. Late latent syphilis was diagnosed on booking bloods at 14 weeks and at 16 weeks she was given three once-weekly benzathine penicillin injections. An anatomy scan was normal at 20 weeks. The infant was delivered stillborn at 32 weeks following presentation with no foetal movement. At post-mortem no features of congenital syphilis were found. A placental swab had no bacterial growth, and placental examination revealed subchorionic inflammation but no chorioamnionitis. There was no evidence of abruption and no other cause for the stillbirth was identified.

Discussion

Although the numbers in this review are small, our local experience indicates that treatment of syphilis in pregnancy prevents term congenital syphilis and term neonatal death, but up to one-fifth of pregnancies may be affected by second-trimester miscarriage, very pre-term delivery and stillbirth. The rates of stillbirth (4%) and very pre-term birth (8%) in Group 1 are considerably higher than national UK background rates (0.5% for stillbirth, ⁹ 3.3% for very pre-term birth ¹⁰ ). Although we cannot conclusively show causality with syphilis we found no severe APOs in those women adequately treated for syphilis prior to pregnancy who would have similar risk factors for other causes of APOs. Also, our high rate of APOs is in keeping with the 21% recently reported from North East England where miscarriages, stillbirth and cases of congenital syphilis occurred. ¹¹

The uptake of antenatal syphilis screening in Leeds during 2012 was 99.7% and hence lack of testing is not a barrier to appropriate care locally. Leeds has dedicated screening midwives and good links between sexual health, obstetric services and paediatric services with clear pathways for referral and a monthly Multidisciplinary Team (MDT) meeting to record all local pregnancy and infant outcomes.

The major weakness of this review is the small number of cases, as syphilis remains a relatively rare infection in women in the UK. Also, APOs are often multifactorial and it is not certain that syphilis was the single underlying cause in these cases.

National guidelines state all patients should be ‘given a detailed explanation of syphilis, including the long-term implications for themselves and their partners/families’. ⁶ We would suggest based on our findings that care should be taken to counsel pregnant women that treating their syphilis will reduce the risk of severe APOs but, depending on their stage of syphilis and the gestation when treated, it may not prevent them completely.

There were some missed opportunities identified in our pathways. Late booking of the pregnancy with subsequent late diagnosis on screening bloods was a contributing factor to delayed treatment in case four (UK National Health Service guidance recommends screening bloods at a booking appointment between 8 and 12 weeks gestation). In addition, the majority of women in Group 1 had English as a second language and 50% required the services of a translator. These may be areas where increased community education would be of benefit.

Given the known severe APOs, which can occur with untreated syphilis, and the knowledge that foetal and placental damage can be prevented by treating before 18 weeks, ¹ we suggest that treatment of syphilis in pregnancy should be viewed with the same priority as prevention of mother-to-child transmission of HIV. Rapid referral pathways from antenatal screening with multidisciplinary management are essential for both, and a combined syphilis and HIV pregnancy MDT works well in our experience and has been suggested by others. ¹¹ This is also in line with the World Health Organisation’s suggested integrated approach to the elimination of mother-to-child transmission of both HIV and syphilis. ¹²

Current literature and policy for the developing world rightly focuses on treatment in pregnancy to reduce congenital syphilis. ¹³ However, there also needs to be accurate information on rates of other severe APOs following treatment, particularly in countries with effective screening where congenital syphilis is rare. Our data, and those of others, ¹¹ suggest the rate of APOs may be high although we cannot conclusively prove that syphilis was the single cause. Accurate data on the pregnancy outcomes of larger numbers of pregnant women with syphilis in countries such as the UK are needed through collaboration between maternity and sexual health services nationally. Public Health England and the British Association for Sexual Health and HIV are investigating options for such collaboration.