Neisseria gonorrhea is an exclusively human pathogen that causes a spectrum of disease ranging from asymptomatic mucosal infection to bacteremia with fatal meningitis. The gonococcus is of special concern for pregnant women because of a possibly increased potential for disseminated disease,¹ an association with preterm rupture of membranes and preterm delivery,^2,3 and the risks of neonatal infection.^4,5 Gonorrhea still remains a significant disease in the United States among pregnant patients. It is estimated that the nearly one million new cases of gonorrhea reported annually represent only approximately one half of all cases.⁶

Although manifestations of gonorrhea in men were recognized in writings of the ancient Chinese, Egyptians, Romans, and Greeks, gonococcal morbidity in women and its relation to male infection was not recognized until recent times. Neisser first observed the gonococcus in stained smears of urethral exudate in 1879,⁷ making it the second reported bacterial pathogen (the first was Bacillus anthracis). The gonococcus was first isolated by culture in 1882. Effective antimicrobial therapy for N. gonorrhoeae became available with the introduction of sulfonamides in the 1930s.

Organism

N. gonorrhoeae is an aerobic gram-negative diplococcus. The gonococcus is encapsulated and non-spore-forming. Media enriched with heated blood (chocolate agar) and other supplements (e.g., glucose, ferric ions, thiamine pyrophosphate) are used to culture gonococci. Starch is often added to the culture to adsorb fatty acids, heavy metals, and other substances that tend to inhibit growth. The antimicrobials vancomycin, colistin, and nystatin are placed in the media to inhibit other bacteria, particularly for genital cultures. Carbon dioxide is required to initiate gonococcal growth. N. gonorrhoeae is very temperature sensitive; growth is optimal at 35°C. Plates should be warmed before inoculation, and cultures should never be refrigerated.

Gonococci can be identified by several steps. Gram stains identify the organism as a gram-negative diplococcus. N. gonorrhoeae colonies stain a deep purple with the addition of dimethyl-p-phenylenediamine (oxidase test) because of the presence of oxidase produced by the organism. Sugar utilization tests also should be used to differentiate N. gonorrhoeae from other Neisseria species found in humans. Of the nine species of Neisseria indigenous to man, only N. meningitidis and N. gonorrhoeae are pathogenic. N. gonorrhoeae utilizes glucose, but it is unable to utilize maltose, sucrose, or lactose, thus distinguishing it from the other Neisseria species.

Several methods have been used to type strains of gonococci. The first method is based on the colonial morphology of the gonococcus on transparent solid media. Typing is based according to opacity and colony size. Opaque colonies contain surface proteins (Protein II) that promote their adherence to neutrophils and other gonococci.⁸ In contrast, transparent colonies lack protein II and do not adhere to neutrophils; they also resist being killed by serum. Small-colony transparent variants (types 1 and 2) are seen most commonly on primary isolation, and these are the virulent strains of the organism.⁹ Transparent colonies account for most isolates from blood and fallopian tubes. Pili are the hairlike appendages that assist in attachment and interfere with phagocytosis.¹⁰ Pili are usually present only on types 1 and 2 cells. On repeated subculture, large-colony, opaque types 3 and 4 tend to be formed: These are avirulent and lack pili.

More stable methods of differentiating subtypes of gonococci include serogrouping and auxotyping. Serogrouping is based on the antigenic structure of outer-membrane Protein I, which is antigenically stable within a particular strain.¹¹ Three serogroups (WI, WII, WIII) and several serotypes of gonococci can be identified with this method. Auxotyping utilizes variations in specific nutritional requirements between strains for typing. Serogrouping and auxotyping are somewhat complicated, but they are of use in epidemiologic studies of gonococci.¹² Unfortunately, typing and gonococcal virulence factors have not been used to study infections among pregnant women.

Variations in antibiotic susceptibility have important implications for modern treatment schemes. Relative or low-level penicillin resistance first became apparent in the 1950s, when strains with minimal inhibitory concentrations (MICs) greater than 0.018 μg/mL of penicillin were noted.¹³ Increasing penicillin resistance continued into the 1970s, but the trends were reversed in industrialized countries by widespread use of higher-dose treatment regimens.¹⁴ This low-level chromosomally mediated resistance can be selected by the use of low-dose treatment regimens. The recent outbreak of chromosomally mediated, nonpenicillinase-producing, relatively highly resistant (MIC = 2 to 4 μg/mL) strains underscores the need for testing of gonococcal antibiotic susceptibility among patients who have failed therapy.¹⁵ A second pattern of gonococcal susceptibility is the absolute resistance to penicillin caused by plasmid-mediated β-lactamase production. These penicillinase-producing N. gonorrhoeae (PPNG) have been common in several overseas geographic areas for many years, but their incidence has increased dramatically in the US since 1980.¹⁶ Of concern is the possibility that the plasmid can be transmitted to other bacteria, including Haemophilus influenzae and N. meningitidis, which are specific neonatal pathogens. No gonococcal strains resistant to ceftriaxone, cefixime, ciprofloxacin, or ofloxacin, the currently recommended regimens by the Centers for Disease Control, have been reported in the US.¹⁷

Epidemiology

The gonococcus has shown remarkable adaptability throughout the past 40 years. Gonorrhea increased markedly during and shortly after World War II until the introduction of penicillin therapy. The incidence then decreased until 1957, when gonorrhea rates began to rise.¹⁸ From 1962 to 1975, the incidence rose dramatically such that in 1965, gonorrhea became the most widely reported communicable disease in the US, a position it still maintains. Although the incidence of gonorrhea has declined slowly since 1975, there are still roughly one million new cases reported annually in the US. Reported cases are believed to represent between one third and two thirds of actual cases.⁶

The prevalence of gonorrhea in pregnant patients varies widely from 1% to greater than 7.3%.^19,20 Several social and demographic variables have been associated with an increased risk of gonorrhea acquisition: The highest rates of infection are among single, nonwhite, poor teen-agers.²¹ More than 80% of cases occur among persons less than 30 years old. By 1982, infection rates for 15- to 19-year-old women surpassed those for 20- to 24-year-old women.²² Nonwhite women are 15 times more likely to have a reported case of gonorrhea than white women.²³ Being unmarried and of low socioeconomic status are both independently correlated with an increased prevalence of gonorrhea in women.²³ Previous gonococcal infection is another significant risk factor: Up to 40% of pregnant women with gonorrhea previously had gonorrhea.²⁴ Women with gonorrhea also frequently have other concurrent sexually transmitted infections, particularly Trichomonas vaginalis. This epidemiologic data helps to identify the group most likely to benefit from gonorrhea screening. In addition, because approximately 40% to 50% of women with gonorrhea also are infected with Chlamydia trachomatis, this group would also benefit from chlamydial screening.²⁵

Pathogenesis

Data on N. gonorrhoeae acquisition have been gleaned from epidemiologic studies. Men have a 20% to 30% chance of contracting infection from a single exposure to an infected woman.²⁶ Women have a 60% to 90% risk of acquiring infection during each exposure to an infected man.²⁷ Both the higher chance of transmission to women per exposure and the high frequency of asymptomatic infections among women provide a large reservoir of infected pregnant women. Vaginally delivered neonates have a 30% to 35% chance of developing gonococcal conjunctivitis or positive orogastric aspirate cultures from infected mothers.²⁸ Gonococcus has even been isolated from the conjunctiva²⁹ and oropharynx³⁰ of infants delivered by cesarean section if rupture of the membranes occurred before the operation. Although uncommon in modern hospitals today, epidemics of infant gonococcal infection secondary to fomite transmission were not uncommon in nurseries before effective gonococcal diagnosis and therapy became available.

Gonococci selectively infect columnar and transitional epithelial cells. By using pili and possibly by outer-membrane Protein II, the organism first attaches to the epithelial cell. After attachment, gonococci penetrate into the host cell by endocytosis. Once intracellular, the gonococci destroy the host cell, enabling the organism to reach the subepithelial connective tissue, and potentially, the bloodstream. Gonococci also produce proteases,³¹ phospholipases,³² and elastases and contain cytotoxic lipopolysaccharide (LPS);³³ all of these factors probably aid in the spread of infection.

The role of the host immune response in preventing gonococcal infection has not been well defined. Several properties of the gonococcus seem to protect it from usual host defense mechanisms. As mentioned, the presence of pili on the gonococcus interferes with phagocytosis by neutrophils.¹⁰ Immunoglobulin A (IgA) in the cervix is believed to prevent pathogenic organisms from ascending into the uterus. However, endometrial ascent is aided in pathogenic Neisseria species (e.g., N. gonorrhoeae and N. meningitidis) by their production of an IgA protease³¹ that cleaves to and inactivates IgA. (Commensal Neisseria species do not produce IgA protease.)

IgM antibodies against gonococcal LPS, probably due to previous exposure to antigenically similar Neisseria species, can be found in the sera of persons with no history of previous gonococcal infection.³⁴ These antibodies bind to LPS and complement to form a bactericidal C5–9 complex. Many gonococcal strains exhibit stable serum resistance to this bactericidal mechanism, especially those isolated from the blood and synovial fluid of patients with disseminated gonococcal infection.^35,36 Serum resistance seems to be due to an IgG antibody that binds to gonococcal outer-membrane protein III, preventing effective formation of the IgM-LPS-complement complex.

Host flora may influence gonococcal growth. T. vaginalis can phagocytize gonococci; lactobacilli, Staphylococcus epidermidis, and Candida albicans can inhibit their growth. Other species may stimulate gonococcal growth. Physical barriers, particularly the dense cervical mucus plug present in pregnancy, impede the gonococcus from reaching the endometrium. In addition, after the 12th week of pregnancy, the decidua capsularis fuses with the decidua parietalis, obliterating the endometrial cavity and obstructing the intraluminal route of spread. Much more needs to be learned about the host-gonococcal interaction and its implications for vaccine production and disease prevention.

Gonococcal Infections

Gonococcal infections in women take several forms. Most of the positive cultures are from asymptomatic women. Although women comprise approximately 40% of reported gonorrhea cases, they nevertheless account for the majority of well-recognized serious sequelae, including salpingitis and infertility. Less well-recognized complications associated with pregnancy, such as prematurity and neonatal infection, further increase the proportion of serious sequelae occurring among pregnant women. Although gonococcal endocarditis is a rare complication, a fatal case in a pregnant woman was recently reported.³⁷

SALPINGITIS.

The risk of acquiring N. gonorrhoeae salpingitis in pregnancy is considerably lower than the 10% to 20% risk reported among nonpregnant women with gonorrhea.³⁸ Higher progesterone levels of pregnancy cause the cervical mucus to be less permeable to motile sperm and possibly to cervical and vaginal microorganisms, and mucus may also provide some immunity to infection. As mentioned, the fusion of the chorion with the endometrial decidua beyond 12 menstrual weeks obliterates the uterine cavity and markedly reduces the chance for intraluminal spread of organisms. Although well-documented cases of N. gonorrhoeae salpingitis have been reported during pregnancy, this is a rare occurrence. Thus, other causes of pelvic pain should be excluded before the diagnosis of gonococcal salpingitis is made. In most women, N. gonorrhoeae salpingitis is diagnosed in the first trimester of pregnancy; many of these women have had endometrial salpingeal infection previously or shortly after conception. In the second or third trimester, abdominal and adnexal tenderness due to gonococcal infection can be noted, but frank salpingitis is very rare. Gonococcal infection of paracervical lymphatics or vessels, or both, often accounts for the peritoneal signs of infection at this stage of pregnancy. A cervical Gram's stain and gonococcal culture should be obtained in pregnant women with unexplained signs of acute abdomen.

DISSEMINATED GONOCOCCAL INFECTION.

Disseminated gonococcal infection (DGI) arises when gonococci invade the bloodstream as a result of genital tract, pharyngeal, or rectal mucosal infection. DGI develops in approximately 1% of patients with mucosal infections. DGI is more common in women than men. Studies have variably reported an increased risk of DGI among pregnant women (up to 40%).¹ Among nonpregnant women, DGI usually occurs within 5 days before menses. These findings are consistent with the hypothesis that more invasive transparent gonococcal colony types predominate in the presence of high progesterone levels. DGI gonococcal strains also tend (1) to comprise certain auxotypes with specific growth requirements¹²; (2) to be highly susceptible to penicillin; (3) to be of the WI serogroup; and (4) to be resistant to host serum bactericidal activity.³⁹ Therefore a variety of organisms and host factors may contribute to disseminated infection. In fact, none of the studies reporting an increased rate of DGI in pregnancy has been controlled adequately for the increased rate of testing for gonococcal disease of the pregnant versus nonpregnant patient. It remains unclear whether pregnant women are uniquely more susceptible to DGI than nonpregnant women, or whether they are simply more likely to have the disease recognized.

Signs and symptoms of DGI do not change with pregnancy. The initial bacteremic stage is accompanied by fever (usually between 38°C and 39°C), malaise, anorexia, and occasionally, shaking chills. Skin lesions, the most common manifestation of DGI, begin as small erythematous macules and evolve into pustular lesions. These lesions usually occur on the arms and legs, especially near the small joints of the hands and feet, and may be either symptomatic or asymptomatic. Skin lesions rarely appear on the face or trunk. New lesions may continue to appear for several days. Tenosynovitis occurs frequently, usually involving the hands and feet. Severe pain on motion of the affected tendon is seen, and swelling, erythema, and tenderness over the tendon is present on exam. Pus is usually not found within the tendon sheath, and the gonococcus rarely can be isolated from this site. In DGI, migratory polyarthralgias are more common than septic arthritis. Between one half and two thirds of the septic arthritis in American adults is caused by the gonococcus.⁴⁰ Arthritis usually occurs later in the course of DGI than skin lesions and tenosynovitis, and it most commonly involves the knees. Other frequent sites include the elbow, ankle, and metacarpophalangeal and interphalangeal joints, although any joint may be involved. When monoarthritis or oligoarthritis is present, joint fluid must be obtained for analysis. In DGI, the fluid has a high white blood cell count (usually >30,000 cells/mL with mostly polymorphonuclear leukocytes) and a low glucose level. Other findings uncommonly associated with DGI include pharyngitis, mild hepatitis, endocarditis, meningitis, pericarditis, pneumonia, and osteomyelitis.

The diagnosis of DGI depends on a high clinical suspicion for the disease. DGI should be considered in the presence of tenosynovitis, monoarthritis or oligoarthritis, or unexplained pustular skin lesions of the arms and legs. Once the diagnosis is suspected, appropriate mucosal cultures (e.g., from cervix, rectum, pharynx, urethra) should be obtained because these will be positive in 80% of patients.¹ Blood cultures will be positive in only 25% of patients with DGI, and only in the early bacteremic stage.¹ Joint fluid cultures or Gram's stains, or both, are positive in only 15% to 20% of cases.¹ Treatment is discussed below.

SYMPTOMATIC GONOCOCCAL DISEASE.

Before the initiation of routine N. gonorrhoeae prenatal screening, only symptomatic gonococcal infections were detected during pregnancy. In one study,⁴¹ symptoms of gonococcal infection found in 13 of 1800 unscreened pregnant women included septic abortion (n = 13), premature rupture of membranes (n = 8), cervicitis (n = 8), postpartum fever (n = 3), first-trimester salpingitis (n = 2), arthritis (n = 1), infants with ophthalmia neonatorum (n = 2). Although studies have not demonstrated an increased incidence of uncomplicated first-trimester abortion in women with gonococcal cervical infection, the risk of postabortal endometritis is believed to be increased among women with N. gonorrhoeae.⁴² These studies of symptomatic disease caused by the gonococcus helped to increase the use of routine prenatal screening for all pregnant women.

Routine prenatal culture for gonorrhea became widespread in the late 1960s, after the introduction of selective Thayer-Martin medium in 1964 and after several studies indicated a substantial prevalence of asymptomatic gonorrhea. Several groups began retrospective studies of the incidence of pregnancy complications in asymptomatic women with versus without gonorrhea. An increased incidence of several complications among previously asymptomatic patients with intrapartum N. gonorrhoeae was reported. Prematurity rates among women with gonorrhea ranged from 12% to 40%^5,43 (Table 1) compared with an overall prematurity rate in the US of approximately 8%. In some studies, a statistically significant high prematurity rate was found in women with versus without gonorrhea, but this finding has not been consistent throughout the literature. Edwards and co-workers,⁴³ who reported the highest prematurity rate, found the increased prematurity rate only in the group with positive intrapartum cultures. Amstey and Steadman,⁴⁴ found high prematurity rates even in women with first-trimester gonorrhea who were treated and remained culture-negative throughout the remainder of the pregnancy. In contrast, Charles and associates,²⁰ found no significant difference in outcome between women treated early in pregnancy for gonorrhea and those with negative cultures throughout pregnancy. Stoll and colleagues,⁴⁵ found no difference in birth weight and gestational age at delivery between patients with and without gonorrhea during pregnancy.

TABLE 1. Pregnancy Outcome with Positive N. Gonorrhoeae Cultures

There has also been an association between gonorrhea and increased perinatal mortality,^43,44 premature rupture of membranes,^43,44,46 and maternal^42,44,46 morbidity. Edwards and co-workers,⁴³ reported an increase in premature or prolonged rupture of membranes only in the group with positive intrapartum cultures. Amstey and Steadman⁴⁴ found these relationships even among women who had been previously treated for gonorrhea during this pregnancy, whereas Israel and colleagues,⁵ found no increase in maternal morbidity among women previously treated for gonorrhea. Data from these retrospective studies suggests that the presence of N. gonorrhoeae in the cervix during the third trimester may be a risk factor for prematurity, premature rupture of membranes, prolonged membrane rupture, and maternal febrile morbidity.

Recent prospective studies have yielded more consistent results. In a case-control study of preterm rupture of membranes, N. gonorrhoeae was present in 13% of patients with preterm membrane rupture and in no control subjects.⁴⁷ However, because all of the patients in the study had been screened and, if positive, treated for gonorrhea earlier in pregnancy, specific risk calculations were not performed. Two studies from Africa that included screening but not treatment of N. gonorrhoeae in pregnancy showed an increased risk of preterm birth and low birth weight among women with N. gonorrhoeae infection.^2,3 Elliott and associates² found a nearly threefold increased risk of preterm birth related to gonorrhea in pregnancy, independent of other risk factors for prematurity. Donders and co-workers³ found a sixfold increased risk of prematurity with N. gonorrhoeae infection, again independent of other risk factors. Thus, screening and treatment for N. gonorrhoeae infection in pregnancy is important to prevent maternal, fetal, and neonatal complications.

All women should be cultured for N. gonorrhoeae on their first prenatal visit, and those with positive cultures should be treated with an appropriate antibiotic regimen. Frequent follow-up cultures should be obtained from this group because 11% to 30% of pregnant women who are initially culture-positive will have recurrent gonorrhea after therapy.⁴⁸ This high recurrence rate occurs despite the treatment of male partners. In addition, women at high risk for infection (e.g., a history of gonorrhea or other sexually transmitted disease [STD] before pregnancy; young, nonwhite, single, and nulliparous women; women with other STDs identified during the current pregnancy) also should have repeat gonorrhea cultures obtained in the third trimester.

NEONATAL INFECTION.

The prevention of neonatal infection is another reason to diagnose and treat maternal gonococcal infections during pregnancy. Ophthalmia neonatorum, the first recognized form of neonatal ophthalmic infection, was described in 1881 by Hirschberg and Krause who identified the gonococcus in smears of conjunctival exudate from infected neonates. Crede introduced topical silver nitrite for the prevention of ophthalmia neonatorum in the same year.⁴⁹ Ophthalmia neonatorum from all causes occurred in 9% of infants born in the US before the use of silver nitrite prophylaxis; one half of these cases were believed to be caused by N. gonorrhoeae.⁴⁹ With prophylaxis, the rate of infectious conjunctivitis dropped to approximately 0.3% of infants^28,50 (Table 2). Before the widespread neonatal eye prophylaxis, one-quarter of the blindness in the US was attributed to gonococcal ophthalmia neonatorum. Largely because of neonatal eye prophylaxis and the availability of antibiotic therapy, only 0.1% of blindness was attributed to ophthalmia neonatorum in 1959.⁵¹ N. gonorrhoeae remains a frequent cause of neonatal conjunctivitis and blindness in developing countries, where the prevalence of the organism is high and antenatal screening is uncommon. Today, conjunctivitis is caused by other factors, including chemical irritation from silver nitrate; conjunctival infection from C. trachomatis, S. aureus, Haemophilus species, Streptococcus pneumoniae, enterococcus, or herpes simplex virus is more common than gonococcal conjunctivitis.

TABLE 2. Risk Estimates of Gonococcal Ophthalmia Neonatorum After Various Types of Prophylaxis

Despite its infrequency, N. gonorrhoeae causes the most serious conjunctival inflammation, and it should always be considered in the differential diagnosis of this syndrome. Signs of gonococcal conjunctivitis usually develop 2 to 5 days after birth, although more indolent infections, possibly due to partial suppression by prophylaxis, may appear in infants up to 3 weeks of age. Occasionally, gonococcal ophthalmia may appear a few hours after birth in babies born after prolonged rupture of membranes. Signs of infection include a purulent conjunctival exudate and edema of both lids. The infection is usually bilateral. If not treated promptly, the infection can spread to the cornea, causing either ulcerations with permanent scarring or a perforation with the formation of anterior synechiae, anterior staphyloma, panophthalmitis, and even loss of the eye. Systemic spread had been reported with fatal results. Early antibiotic therapy causes prompt healing without sequelae. Because of the potentially devastating effects, gonococcal ophthalmia should be suspected in any infant less than 1 month old with conjunctivitis, and a Gram's stain and culture should be performed to exclude N. gonorrhoeae infection. Ceftriaxone 25 to 50 mg/kg as a single intravenous or intramuscular dose not to exceed 125 mg is indicated for proven or suspected gonococcal ophthalmia neonatorum¹⁷ (Table 3). In addition, the eyes should be irrigated with saline to remove exudate.

TABLE 3. Treatment of Gonococcal Infections in Infants

Gonococcal arthritis is the most common, but still rare, manifestation of systemic gonococcal infection in the newborn.⁴ Symptoms usually appear in infants 1 to 4 weeks old and are nonspecific, initially including fever, irritability, and poor feeding. This stage is followed by erythema, warmth, swelling, and tenderness over the affected joint. Frequently, no antecedent history of conjunctivitis or other gonococcal infection exists. Dissemination occurs from an infected scalp electrode site or rectal, pharyngeal, urethral, or vaginal colonization. Unlike adult gonococcal arthritis, neonatal disease frequently involves multiple joints, and large joint involvement predominates. Skin lesions are seldom, if ever, associated with gonococcal arthritis in infants. The diagnosis is established by gram-positive N. gonorrhoeae stains or by culturing of N. gonorrhoeae from purulent fluid aspirated from the infected joint, or by a combination of these two methods. Blood and cerebral fluid cultures should be performed as well. Recommended therapy is listed in Table 3.¹⁷ Drainage of the infected joint by the use of either arthrocentesis or open procedures may also be necessary. Surgical drainage is mandatory in cases of gonococcal infection of the hip joint.

Gonococcal septicemia can develop after prolonged rupture of membranes, especially among premature infants. In one study, N. gonorrhoeae was the third most common pathogen, following E. coli and Group B Streptococcus, isolated from the gastric aspirate of neonates with suspected sepsis.⁴⁶ Gonococcal meningitis is a rare neonatal infection that is usually associated with ophthalmia neonatorum. Gonococcal endocarditis has not been documented in the neonatal period. Therapy for gonococcal septicemia is the same as for arthritis (see Table 3). If gonococcal meningitis is present, therapy should be continued for 10 to 14 days. Other less common sites of neonatal gonococcal infection include the rectum, vagina, pharynx, nares, umbilical stump, urethra, and scalp electrode site. Systemic infection may develop from any of these sites. Infants of mothers with untreated gonorrhea at delivery should be treated with a single dose of ceftriaxone.¹⁷ The parents of neonates with gonococcal infection should be evaluated for infection and treated accordingly.

Diagnosis

In women, the standard method of diagnosing gonococcal infection remains a culture plated on antibiotic-containing selective media. Several studies have shown that a single endocervical culture on selective medium detects 80% to 90% of cervical infection. Duplicate cervical culture specimens increase the yield of positive cultures by 10% to 20%. Because pharyngeal or rectal cultures are positive in only 5% of women with a negative cervical culture, the decision to culture noncervical sites such as the pharynx, urethra, and rectum should be individualized depending on the history, symptoms, and resources available. Gram's stains are too insensitive to replace the culture in women, especially those with a low risk of infection.⁵² Enzyme immunoassay techniques for the rapid diagnosis of gonococcal disease have lacked consistent sensitivity and specificity for use in nonpregnant and pregnant women, particularly among populations with low gonococcal prevalence.^53,54 Gram-stained smears and enzyme immunoassay techniques have not been evaluated in the diagnosis of extragenital gonococcal infections, although their sensitivity and specificity is expected to be even lower in these settings. Fluorescent antibodies against gonococcal antigens in smears also have not proved useful for diagnosis of uncomplicated infections or for screening purposes.²⁷ Recent studies have indicated a high sensitivity and specificity for a DNA probe assay for the detection of cervical N. gonorrhoeae in pregnant and nonpregnant women with a prevalence of infection of at least 2%.^55,56 The assay can be run in the laboratory in 2 hours. If confirmed in further studies of pregnant women with a prevalence of infection less than 2%, this assay could replace cultures for the diagnosis of N. gonorrhoeae. Pending these studies, N. gonorrhoeae detection in women should utilize cervical cultures either plated immediately onto selective media or inoculated into appropriate transport systems. As previously discussed, all pregnant women should be cultured on their first prenatal visit, and women at high risk for gonococcal infection should be cultured again during the third trimester.

Maternal Treatment

Treatment regimens for N. gonorrhoeae have been revised frequently because of the remarkable adaptation of the organism to antibiotic pressure. Low-level, chromosomally mediated penicillin resistance has been apparent since soon after the introduction of penicillin, with ever-increasing doses required for cure.¹³ Plasmid mediated, absolute penicillin resistance secondary to β-lactamase production has become an increasing problem in the US since the introduction of PPNG strains in 1976.¹⁶ More than twice as many PPNG cases were reported to the Centers for Disease Control in the first nine months of 1985 compared with the same period in 1984.¹⁶ All 50 states in the US reported PPNG cases in 1985. Because of the increasing numbers of PPNG cases, penicillin regimens are no longer recommended for treatment of gonorrhea. Specific treatment regimens are listed in Table 4.¹⁷

TABLE 4. Treatment of Gonococcal Infections in Pregnancy

In summary, gonococcal infections remain a significant concern in pregnant and nonpregnant patients alike. Routine prenatal cultures in the first trimester are mandatory, as is rapid treatment of infected patients and partners to prevent maternal, fetal, and neonatal complications. Repeat cultures should be done on patients at high risk for gonococcal infection, especially those with previous positive cultures in this pregnancy. Prevention remains the best treatment: Infected persons must be treated rapidly to prevent disease spread, and intense efforts must be made toward contact tracing and the development of a vaccine.

Syphilis is the third most common reportable communicable disease in the US, exceeded only by gonorrhea and varicella.⁵⁷ Despite the marked decrease in the incidence of syphilis after the introduction of penicillin, physicians still need to be aware of its protean manifestations, and they must consider syphilis in the differential diagnosis of many syndromes. Primary and secondary syphilis increased remarkably in the late 1980s from a population rate of 11 cases/100,000 in 1985 to a rate of 20/100,000 in 1990. Rates have been declining since then, but still remain approximately 10/100,000.⁵⁷ This increase in cases has been attributed to the increase in crack cocaine use and associated prostitution.^58,59

A syphilis epidemic raged in Europe in the late 1400s, leading many historians to assume that the disease was introduced to Europe by Columbus's sailors on their return from the West Indies (the Colombian theory).⁶⁰ Introduction of the disease to a nonimmune population could account for the increased severity that was manifested in the epidemic. An alternate theory (the pre-Colombian theory) holds that syphilis developed as a sexually transmitted variant of cutaneous treponemal infections originating in Central Africa, which spread rapidly through Europe during the many wars of the late 15th century. Whatever its origins, syphilis received its name in 1530 from the poet Fracastorius who described the shepherd, Syphilis, afflicted with the disease.⁶¹

Organism

Syphilis is caused by the organism Treponema pallidum, a member of the order Spirochaetales. Other pathogenic members of this order cause yaws, pinta, and endemic syphilis.⁶² Several other members of the genus Treponema are nonpathogenic and are widespread in nature, including normal human mouth flora. Differentiation between the virulent treponemes by morphologic, chemical, and immunologic studies is still not possible, although pathogenic strains are easily differentiated by the syndromes they cause.⁶² Pathogenic groups can be distinguished from nonpathogenic groups by DNA hybridization.⁶³ Because of its small size (6 to 15 μm), T. pallidum is not visible by standard light microscopy, but it can be seen with a darkfield microscope. On darkfield examination, the organism appears wavelike from its spiral structure and moves with a rotary motion.⁶⁴ T. pallidum has only recently been cultured in vitro, and much work remains to be done on the characterization of treponemes for potential vaccine production. Unfortunately, there is no good animal model for the study of all stages of syphilis.⁶² Rabbits manifest disease similar to primary and secondary syphilis in humans, but not even primates have shown characteristics of tertiary syphilis.

Pathogenesis

T. pallidum initially enters from mucosal or cutaneous lesions of an infected person through skin or mucous membranes abraded during sexual activity. Treponemes do not seem to penetrate unbroken, keratinized epithelium. The inoculum necessary for infection is unknown, but because of the organism's long replication period (30 hours), the smaller the inoculum, the longer it will take for symptoms to appear. The organisms do, however, appear to reach local lymph nodes within 30 minutes after skin inoculation. Viable treponemes attach to cells, after which polymorphonuclear leukocytes accumulate in the area of infection.⁶⁵ Both B and T lymphocytes accumulate in the area.^66,67 Activated macrophages ingest treponemes.⁶⁸ By the time a chancre appears, most patients have evidence of antibody formation to the organism. Although both humoral and cell-mediated immunity to T. pallidum have been demonstrated, the host response and factors enabling the organism to persist and multiply have not been well characterized. The treponemes may actually suppress host immunity to facilitate their own survival.

Epidemiology

Epidemiologic studies indicate that approximately one third of persons who are sexually exposed to syphilis will contract the disease.⁶⁹ Fomite transmission is rare because T. pallidum is easily destroyed by drying, heat, and soap. Neonates most commonly acquire an in utero infection from an infected mother, although rarely they have become infected from contact with a vaginal lesion at birth. Transmission of syphilis from blood transfusions is rare because serologic testing is performed on all blood. The rare occurrence of syphilis incubating in the donor blood at the time of donation going on to cause infection in the recipient has been reported,⁷⁰ but this possibility is further reduced by the fact that blood usually is stored for more than 72 hours at 5°C.⁷¹

Despite the continued exquisite sensitivity of T. pallidum to penicillin, syphilis remains prevalent. The lowest number of reported cases occurred in the mid-1950s, but subsequently decreased funding and control measures has led to an increase in cases. As mentioned, a marked increase in cases occurred in the late 1980s related to crack cocaine use and associated prostitution.⁵⁷ A concomitant increase in cases of congenital syphilis also has been seen, increasing from approximately 2200 cases in 1985 to more than 4000 cases in 1991.⁵⁷ The increase was most striking in Florida, Texas, California, and New York City and was associated with maternal drug use and lack of prenatal care.^72,73 As with gonorrhea, estimates are that only one in three cases are actually reported.⁶ Of women with infectious syphilis, 80% are in the 15to 34-year-old group, which is the range of the peak reproductive years.⁵⁷ In fact, the incidence curve of congenital syphilis parallels that of female infectious syphilis with a lag of 1 year.

Several groups are at especially high risk for contracting syphilis. As with other STDs, syphilis occurs most commonly in young, single, nonwhite persons living in urban communities.⁷⁴ The risk of syphilis is increased among women who have no prenatal care and who are drug users.^72,73 Although some of the preponderance of STDs in this population is due to reporting bias (these persons tend to attend public health clinics, which have more complete reporting than private offices), they remain a group requiring frequent syphilis screening. There is also an increased prevalence of syphilis among homosexual men. As with women, these men often remain asymptomatic and unknowingly infectious during the primary stage because the chancre is in the rectum. Because 10% to 35% of homosexual men seen in STD clinics report having female as well as male sexual contacts,⁷⁵ this group now represents an important source of infection leading to congenital syphilis. Pregnant women in high-risk groups should have repeat serologic screening for syphilis in the third trimester in addition to their first-visit testing. All patients diagnosed with syphilis should be counseled and tested for HIV infection.

Clinical Syndromes

PRIMARY SYPHILIS.

Approximately 30% of sexual contacts of persons with primary or secondary syphilis will become infected.⁶⁹ A chancre, the manifestation of primary syphilis, typically develops at the site of treponemal inoculation 3 weeks (range, 10 to 90 days) after exposure.⁶² Classically, the chancre is a single, clean-based, painless ulcer with indurated margins. In up to 40% of cases, multiple lesions appear, and atypically appearing lesions actually are more common than the classically described chancre.⁷⁶ In women, chancres may go unnoticed on the cervix or in the vagina. Five percent of lesions are extragenital (e.g., on the mouth, anus, breast, rectum, fingers). Extragenital lesions are especially likely to be atypical. Chancres heal spontaneously without therapy in 2 to 8 weeks. Unilateral or bilateral painless regional lymphadenopathy usually occurs 1 week after the chancre. Both the ulcer and the adenopathy may be painful if secondary bacterial infection occurs.

SECONDARY SYPHILIS.

Manifestations of secondary syphilis typically develop about 6 weeks after the appearance of the chancre, which still could be present. The skin lesions of secondary syphilis are typically papulosquamous, but they can take various forms. The rash is usually widespread and, unlike most skin diseases, often involves the palms and soles. Patchy alopecia may accompany the skin lesions; hair regrows without treatment. Highly infectious, erosive lesions of the mucous membranes called mucous patches may appear in the oral, nasal, or genital mucosa. Other systemic symptoms common in this stage include malaise, low-grade fever, sore throat, headaches, hoarseness, anorexia, and arthralgias. These nonspecific symptoms are usually mild and transient. Nontender, generalized lymphadenopathy is nearly always present. Less common symptoms include arthritis, bursitis, osteitis, myositis, uveitis, and meningitis. Hepatosplenomegaly occurs in a minority of cases. With this wide variety of manifestations, it is easy to see why syphilis masquerades as a large number of other diseases. Common laboratory findings include anemia, leukocytosis, and an increased erythrocyte sedimentation rate.⁷⁷ Laboratory evidence of hepatitis⁷⁸ and mild nephrotic syndrome⁷⁹ may be seen. Systemic symptoms usually resolve quickly without treatment; skin lesions heal in 2 to 3 weeks, often with hyperpigmentation. Skin and mucous membrane lesions of secondary syphilis may relapse for up to 4 years in untreated infection. The diagnosis of secondary syphilis often depends on serologic testing. Darkfield examination should be done on moist, nonoral lesions (e.g., condylomata lata) when present. Oral lesions can give false-positive results because of nonpathogenic treponemes indigenous to the mouth. When darkfield examination is inconclusive or impossible to obtain, a presumptive diagnosis of secondary syphilis can be made by the combination of clinical presentation and positive serologic tests (for a detailed discussion, see section on Diagnosis, below). Virtually all patients with secondary syphilis will have reactive treponemal and nontreponemal tests.⁸⁰ Of patients with secondary syphilis, 1% to 2% may have negative nontreponemal test results because of the prozone reaction. This phenomenon occurs as a result of an overwhelming excess of anticardiolipin antibody in the patient's serum which interferes with antigen-antibody binding and flocculation. The excess antibody effect can be overcome by serum dilution: A serum dilution procedure should be requested when secondary syphilis is strongly suspected but initial results are negative. Serum dilution should be considered for routine use among pregnant women in areas of high syphilis prevalence.^81,82

LATENT SYPHILIS.

After the lesions of secondary syphilis resolve, untreated patients pass into the latent stage of syphilis. Latent syphilis is diagnosed on the basis of a consistent history or serologic studies, or both, in the absence of clinical signs and symptoms of syphilis. A careful history and physical examination as well as cerebrospinal fluid studies should be performed to rule out cardiovascular or neurosyphilis before the diagnosis of latent syphilis is made. Because mucocutaneous relapses are most likely to occur within the first year after infection, latent syphilis is divided into early latency (up to 1 year from onset of infection) and late latency (more than 1 year from onset of infection).

TERTIARY SYPHILIS.

Tertiary syphilis develops in approximately 15% to 25% of untreated patients with acquired syphilis.^83,84 Gummas, granulomatous lesions with necrotic centers secondary to obliterative endarteritis, may develop in the skin, bones, cartilage, or other body organs. Gummatous involvement of the heart, brain, or liver can be fatal. Vasculitis of the vasa vasorum of the great vessels can lead to aneurysm formation, aortic valvular insufficiency, and heart failure. Cardiac and aortic involvement usually is detected 10 to 30 years after the initial infection. Central nervous system (CNS) involvement and symptoms can develop from 1 to more than 30 years after the initial infection. Neurologic syndromes include syphilitic meningitis, general paresis, and tabes dorsalis. Tertiary syphilis is uncommon in pregnant women.

Diagnosis

The diagnosis of syphilis depends on a high index of suspicion for the disease, either from the clinical presentation or from the presence of epidemiologic risk factors. The differential diagnosis of genital ulcers includes herpes simplex, furuncle, chancroid, granuloma inguinale, lymphogranuloma venereum, secondary infection, and carcinoma. Darkfield examination must be performed on all suspect genital or extragenital lesions (Table 5). A positive darkfield microscopic examination is the only definitive diagnostic test for syphilis and at times the only method of early diagnosis. Material for this examination is obtained from the serum exudate of skin or mucosal lesions or from aspiration of regional lymph nodes. The material is anaerobically sealed under a cover slip with petrolatum until examined. Darkfield examination may be the only way to diagnose primary syphilis for the 2 weeks after the chancre appears, when serologic tests can be negative. Negative darkfield examination from a highly suspicious lesion should be repeated daily for at least 2 days. Serologic testing also should be done on patients with genital ulcers. Direct fluorescent antibody testing is now available as a substitute for the darkfield exam. For this test, specimens are obtained in the same fashion, allowed to dry on a slide, stained with fluorescein-labeled antibody to T. pallidum, and viewed with a fluorescent microscope.

TABLE 5. Diagnosis of Adult Syphilis

Serologic tests for syphilis can be divided into two types: nontreponemal and treponemal (see Table 5). Nontreponemal tests are designed to detect reagin, a heterogeneous group of antibodies present in the serum of patients with syphilis and certain other conditions. The antigen used in these tests is a cardiolipin-lecithin antigen, which is a normal component of human tissue; for the tests, it usually is derived from beef heart. The nontreponemal tests include the Venereal Disease ResearchLaboratory (VDRL) test, the unheated serum reagin (USR) test, the rapid plasma reagin (RPR) card test, and the automated reagin test (ART). Nontreponemal tests are easy and inexpensive to perform, and all have similar sensitivities and specificities. Nontreponemal tests can be quantitated, and serial titers can be used to follow the patient's response to therapy. Test titers are not interchangeable, however, and the same quantitative nontreponemal test should be performed for a given patient. False-positive results of nontreponemal tests can occur transiently after acute febrile illnesses or immunizations or during pregnancy. Chronic false-positive results can occur among patients with autoimmune diseases and intravenous drug abuse. False-positive tests usually have a low titer (<1:8). False-negative results can occur with the prozone phenomenon (mentioned above).⁸²

The standard treponemal test is the fluorescent treponemal antibody absorption test (FTA-ABS), which is designed to detect antibody directed against pathogenic treponemes. The serum is diluted with a sorbent to bind antibody directed against nonpathogenic treponemes and then is layered on a slide to which T. pallidum has been fixed. Fluorescein-labeled antibody against human Ig is added. Fluorescence indicates that antitreponemal antibodies in the patient's serum have bound to the treponemes. Newer, less expensive treponemal tests include the hemagglutination treponemal test for syphilis (HATTS) and the microhemagglutination assay for T. pallidum (MHA-TP).⁸⁵ Specific IgM T. pallidum antibody tests have been developed, but they remain less reliable and more difficult to perform than more traditional tests.⁸⁶ Treponemal tests remain positive after therapy, and specific titers are not useful for following disease activity. Treponemal tests can be false-positive in connective tissue diseases, pregnancy, and some acute and chronic infections (e.g., mononucleosis, leprosy). A treponemal test will be positive with any nonvenereal form of treponemal disease (e.g., yaws, pinta).

A summary of syphilis serologic testing is provided in Table 5. Note that treponemal tests are more likely than nontreponemal tests to be positive in primary and latent stages. Thus, patients with suspected syphilis should have both tests performed. Virtually all serologic tests are positive in the secondary stage, but sensitivity decreases in the tertiary stage. There are uncertainties about the effectiveness of antibiotic therapy for CNS infection (see section on Therapy, below). Therefore, all women with late latent or tertiary syphilis should undergo a lumbar puncture, as should women with CNS symptoms in the secondary and early latent phase.

Syphilis in Pregnancy

The manifestations of syphilis are not different, stage for stage, between pregnant and nonpregnant women. All pregnant women should have a serologic test for syphilis on the first prenatal visit. Women at higher risk for syphilis infection, such as pregnant women who already have another STD, are single, or are using drugs during the pregnancy, should have a repeat syphilis serology in the third trimester. Women with nonimmune hydrops or stillbirth should also undergo syphilis testing. Darkfield examination of the amniotic fluid for spirochetes also may be helpful in the evaluation of stillbirth when syphilis is suspected,⁸⁷ but probably adds little to the evaluation of a patient with syphilis and a live fetus. Women diagnosed with syphilis during pregnancy should have an ultrasound assessment of the fetus for stigmata of congenital syphilis and for later comparison to evaluate fetal growth. The outcome of pregnancies complicated by syphilis according to maternal disease stage is shown in Table 6. In a recent report of 56 cases of syphilis during pregnancy, only 7 of whom received therapy during pregnancy, 34% were stillborn and the mean gestational age at delivery was 32.3 weeks.⁷²

TABLE 6. Outcome of Pregnancy with Maternal Syphilis

Drug treatment of syphilis in pregnancy is discussed in detail below (see section on Therapy). Pregnant women with syphilis are at high risk for the Jarisch-Herxheimer reaction (JHR), which can include preterm contractions, decreased fetal movement, and fetal heart rate decelerations. Therefore, consideration should be given to hospitalizing pregnant women who are more than 20 weeks' gestation and who have primary or secondary syphilis for initial therapy. Two thirds of women in whom a JHR developed during therapy had decreased fetal movement and contractions.⁸⁸ In addition, acute increases in Doppler systolic: diastolic ratios have been reported after therapy for early syphilis.⁸⁹ Symptoms begin 2 to 8 hours after therapy and subside by 16 to 24 hours.⁸⁸

Congenital Syphilis

Congenital syphilis continues to occur, although it is totally preventable with early and adequate prenatal care. In a recent review of infected infants, 60% of mothers had no prenatal care, laboratory or physician error occurred in 32%, and in only 8% was treatment failure implicated.⁹⁰ The frequency and manifestations of congenital syphilis in the newborn depend on the stage of maternal syphilis during pregnancy and gestational age at onset of primary infection. Historically, it was believed that T. pallidum did not cross the placental barrier before 18 weeks' gestation because (1) before this point in pregnancy, neonates of treated mothers showed no manifestations of congenital syphilis; and (2) T. pallidum was not identified in the placenta. In 1976, Harter and Benirschke⁹¹ identified treponemes in abortus material from women with untreated syphilis in the first trimester. The absence of lesions from syphilis before 18 weeks' gestation is now believed to be due to the inability of the fetus to mount an inflammatory response before 18 to 20 weeks. This “grace period” for the formation of syphilitic lesions underscores the need for first-trimester screening and prompt treatment of infected women. The later in pregnancy that treatment is begun, the greater the chance that the child will develop stigmata of congenital syphilis. The fetus is more likely to become infected from mothers with primary or secondary syphilis (50%) than from those with latent disease (10 to 40%) (see Table 6).⁹² The longer the duration of untreated maternal syphilis, the lower the probability of congenital syphilis.

Congenital syphilis, as an adult disease, is a systemic process consisting of vasculitis with resultant necrosis and fibrosis. Inflammation is most marked around small vessels, with resultant obliterative endarteritis. All lesions except the primary chancre present in adults with acquired syphilis can be seen in congenital syphilis. Hematogenous passage through the placenta can lead to widespread fetal lesions. The spectrum of congenital syphilis ranges from spontaneous abortion and stillbirth through prematurity (see Table 6), to severe neonatal disease and death, to asymptomatic late disease detected only by positive serologic tests.⁹² Neonates become infected by blood-borne organisms, and all neonatal disease manifestations are more common during the early stage of maternal disease, in which bacteremia is more common than in latent maternal infection.

The placenta in congenital syphilis may appear normal on gross examination, or it may be paler, thicker, and larger than normal.⁹³ Any unusually large placenta warrants microscopic examination and a repeat maternal serologic test for syphilis in the mother. Histologic examination reveals focal villitis with enlarged, clubbed villi and increased connective tissue stroma around the capillaries.⁹⁴ Diffuse fibrosis also may be present. Silver stains will demonstrate T. pallidum in the placenta. Perivascular sclerosis and inflammation of the umbilical cord also may be seen.⁹⁵ Congenital syphilis should be included in the differential diagnosis of stillbirth. Typically, a stillborn with congenital syphilis appears macerated, with a collapsed skull and protuberant abdomen. Vesicular or bullous skin lesions may contain large numbers of treponemes. Hepatosplenomegaly and extramedullary hematopoiesis are present. On autopsy, inflammation and fibrosis may be seen in the lungs, gastrointestinal tract, pancreas, meninges, and bones. Renal damage seen on autopsy seems to be due to immune complex deposition with and without complement fixation. Hydrops seen in congenital syphilis is probably caused by a combination of anemia from hemolysis, hypersplenism, and decreased red cell production in the bone marrow and from hypoproteinemia due to hepatitis.

Congenital syphilis may be first suspected in the presence of abnormal prenatal ultrasound findings. Serologic testing of the mother for syphilis should always be included in the evaluation of a fetus with nonimmune hydrops.^96,97 Other ultrasonographic findings in fetuses with congenital syphilis without hydrops have included hepatomegaly, increased placental thickness, dilated small bowel, and abdominal calcifications.^98,99,100

Congenital syphilis in live-born infants is divided into early disease and late disease, depending on whether manifestations appear before or after 2 years of age. Early disease may be apparent at birth in severely affected infants. These infants often are premature and small for gestational age. They may be hydropic, and 50% have hepatosplenomegaly. These infants are obviously sick at birth and usually have hemolytic anemia and infectious skin lesions, called syphilitic pemphigus. Cutaneous extramedullary hematopoiesis causes a “blueberry-muffin” rash. The mortality rate of severely ill infants with such manifestations at birth is more than 50% despite therapy.¹⁰¹ Because infants with congenital syphilis and their mothers may have negative serologic tests at the time of delivery when the maternal infection was recent, all infants presenting with skin lesions or fever should have serologic tests for syphilis performed.¹⁰² The vesicular or bullous skin eruptions containing spirochetes may first appear several weeks after birth. Desquamation may occur and may be generalized or limited to the palms and soles. More than one half of these infants will have snuffles, a persistent rhinitis characteristic of infants with congenital syphilis. This purulent nasal discharge is highly infectious and may become blood-tinged if the mucosa ulcerates. The nasal cartilage can become eroded, leading to “saddle-nose” deformity. Extension of the infection to the throat can produce a hoarse or aphonic cry. Generalized lymphadenopathy also is found in approximately 50% of infants with early congenital syphilis. Because generalized lymphadenopathy, especially when it involves epitrochlear nodes, is uncommon in early infancy, syphilis should always be considered in the differential diagnosis of this presentation. Hair growth in these infants is often sparse, and alopecia, especially of the eyebrows, is typical. Infectious mucous patches may be observed on the tongue and palate.

Of infants with early congenital syphilis, 70% to 80% have radiographic evidence of disease.¹⁰³ Osteochondritis is the most common manifestation. Pseudoparalysis of Parrot occurs when periarticular swelling and bone tenderness on active or passive motion limits motion of the affected extremity. Because the bone changes include both proliferative and destructive processes, radiographs demonstrate increased density alternating with rarefaction. Osteochondritis appears on radiographic studies at least 5 weeks after initial bone infection; periostitis appears at least 16 weeks after infection. In cases of second-trimester infection, the diagnosis of congenital syphilis can be made in utero on the basis of radiographic findings. A skeletal survey should be performed on any infant suspected of having congenital syphilis.

As noted, anemia is common in infants with congenital syphilis. The anemia can be normochromic, normocytic, or macrocytic, and the Coomb's test is negative. Smears may contain numerous erythroblasts, which can be confused as a sign of immune hydrops. In some cases, the bone marrow is replaced with syphilitic granulation tissue, although the most common cause of the anemia is autoimmune hemolysis. Leukocytosis is present in 70% of syphilitic infants; significant thrombocytopenia, probably an autoimmune phenomenon, is seen in 30%.¹⁰³

Approximately one half of babies with congenital syphilis who are tested are found to have abnormal cerebrospinal fluid.¹⁰³ Typically, the fluid shows lymphocytosis (>100 cells/mL), increased protein, and a positive VDRL. Symptoms appear in 5% to 15% of infected infants (usually 3 to 6 months old) who have not received treatment. Symptoms vary and include acute syphilitic meningitis, cerebrovascular accident, hydrocephalus, cranial nerve palsies, seizure disorders, and impaired development. The pathogenic process common to all manifestations is meningovascular T. pallidum infection with resultant inflammation and scarring.

Other organ systems are involved less commonly in early congenital syphilis. Renal and ocular manifestations and, occasionally, syphilitic pneumonitis, chronic diarrhea, or myocarditis may occur in early congenital syphilis.

Late congenital syphilis, corresponding to tertiary syphilis in the adult, is diagnosed after 2 years of age. Eight percent of children with congenital syphilis are diagnosed in this stage, and often as late as puberty. Serology is always positive during this late stage. The classically described syndromes of late congenital syphilis are rarely seen today, probably because of partial treatment of syphilis with antibiotics prescribed for other childhood infections. Two groups of lesions are seen in late congenital syphilis: (1) stigmata, which are secondary to prior inflammation at a critical developmental stage and have since healed with scarring and ongoing inflammatory processes; and (2) Hutchinson's triad, which includes Hutchinson's teeth (small canine teeth and widely spaced, notched incisors), interstitial keratitis, and eighth nerve deafness. Other stigmata are listed in Table 7.

TABLE 7. Manifestations of Late Congenital Syphilis

The diagnosis of congenital syphilis is often more difficult than the diagnosis of adult disease and requires a high index of suspicion. The diagnosis of syphilis in an infant is classified as confirmed or presumptive, depending on several factors (Table 8):¹⁰⁴

TABLE 8. Case Definition of Congenital Syphilis in an Infant Less Than 2 Years of Age

1. Confirmed case:Treponema pallidum identified in specimen from nonoral lesion placenta, or autopsy material by darkfield microscopy, fluorescent antibody, or other specific stain
   
2. Presumptive case:
   1. Infant born to a mother with untreated or inadequately treated syphilis (nonpenicillin therapy or adequate penicillin regimen <30 days before delivery) or
      
   2. Infant or child with a reactive treponemal test for syphilis and one or more of the following:
      1. Signs of congenital syphilis on exam including hepatosplenomegaly, characteristic skin rash, condyloma lata, snuffles, jaundice, pseudoparalysis, or edema secondary to nephrotic syndrome
         
      2. Signs of congenital syphilis on long bone radiograph (metaphyseal dystrophy, osteitis, periosteal reaction)
         
      3. Reactive cerebrospinal fluid (CSF) VDRL
         
      4. Elevated CSF cell count or protein without other cause
         
      5. Quantitative nontreponemal serologic titers fourfold or more above the mother's, drawn at birth
         
      6. Reactive FTA-ABS-19S-IgM antibody test
         
      
      
   
   

A positive nontreponemal test should be confirmed with a treponemal test. Although both types of tests may be positive because of passive placental transfer of maternal IgG antibody, the infant should be treated unless adequate maternal therapy during pregnancy can be documented. If repeat treponemal tests remain positive at 3 to 6 months in treated infants, these infants are considered to have had congenital syphilis. If the nontreponemal test remains positive at 6 months, therapy should be repeated because treatment was inadequate. If maternal treatment during pregnancy is considered adequate and reinfection is not a possibility, the asymptomatic infant can be followed with serial nontreponemal test titers. If the positive test is due solely to maternal antibody, the titer should decrease fourfold within 2 months and disappear by 6 months. Total infant IgM levels and IgM-FTA-ABS tests have been proposed to detect maternal antibody in the serologic diagnosis of congenital syphilis, but neither has proved adequately sensitive or specific for sole use.¹⁰³ A careful clinical and laboratory evaluation of any infant suspected of having congenital syphilis will often help to strengthen the diagnosis. Cerebrospinal fluid evaluation should be done because at least 50% of infected infants will have pleocytosis or a positive CSF VDRL, or both.¹⁰³ Radiographic studies will provide evidence of congenital syphilis in 50% to 80% of cases by 6 to 12 weeks of life. The diagnosis of congenital syphilis is very unlikely, but not impossible, with negative serologic tests.¹⁰⁴ Diagnosis as soon as possible after birth is important because infants who are asymptomatic at birth and who receive effective therapy within 3 months are spared the stigmata of the disease.

Therapy

Penicillin remains the standard therapy for syphilis. Although T. pallidum cannot be cultured and tested for antibiotic susceptibility, clinical response suggests that resistance to penicillin has not developed over the last 50 years.¹⁰⁵ Organisms must be growing and dividing in order for the penicillin to be effective. Because the replication time for T. pallidum is long (30 hours in early disease; longer in latent stages), therapeutic levels of penicillin are required for longer periods of time than for other bacterial infections. Penicillin serum levels of greater than 0.03 μg/mL are required for at least 2 weeks to eliminate T. pallidum. Longer regimens are necessary for treatment of neurosyphilis or latent syphilis of long duration. Penicillinase-resistant penicillins are not indicated for therapy of syphilis alone. Newer penicillins such as ureidopenicillins have not been studied. First- and second-generation cephalosporins are less effective than penicillin G but more effective than tetracycline or chloramphenicol. These cephalosporins also cross the blood-brain barrier poorly compared with penicillin. Third-generation cephalosporins have not been well studied for treatment of syphilis. Aminoglycosides, clindamycin, and spectinomycin are not effective against T. pallidum. Erythromycin and tetracyclines are effective for syphilis therapy in penicillin-allergic individuals, although these drugs are less active than penicillin G and should not be used to treat syphilis in pregnancy.¹⁰⁶ Treatment schedules are given in Table 9.¹⁷

TABLE 9. Treatment of Syphilis in Pregnancy and Infancy

Penicillin-allergic pregnant women present management dilemmas. Erythromycin failure occurs in 10% of nonpregnant women with early syphilis. Maternal treatment with erythromycin leads to inadequate fetal therapy because fetal levels are 10% of maternal levels.¹⁰⁷ Tetracycline cannot be used in pregnancy. Therefore, penicillin-allergic pregnant women with syphilis should be considered for oral penicillin desensitization followed by intravenous penicillin therapy. This procedure is well-described elsewhere,^17,108,109 but it should be supervised by physicians familiar with its use. Usual therapeutic doses of penicillin are then given. In two series of patients treated with this protocol,^108,109 several had urticaria or pruritus during therapy, but none had serious reactions. Treatment with this protocol allows adequate fetal therapy for syphilis during pregnancy and helps prevent sequelae of disease.

In summary, syphilis remains an important STD and perinatal pathogen. A high suspicion for this disease must be maintained to ensure adequate diagnosis and treatment. All pregnant women should be screened with a nontreponemal test on their first prenatal visit, and patients with positive tests must receive rapid follow-up evaluation and therapy. To prevent reinfection and further spread of the disease, all sexual partners must be contacted and treated as well. Careful clinical and serologic follow-up of both the mother and baby is required for at least 2 years. In women with initially negative tests who are in groups at high risk for syphilis, testing should be repeated in the third trimester. Because a large percentage of infants with congenital syphilis are born to women without prenatal care, congenital syphilis represents a preventable failure of prenatal health care.

1. Holmes KK, Counts GW, Beaty HN: Disseminated gonococcal infection. Ann Intern Med 74: 979, 1971

2. Elliott B, Brunham RC, Laga M et al: Maternal gonococcal infection as a preventable risk factor for low birth weight. J Infect Dis 161: 531, 1990

3. Donders GG, Desmyter J, De Wet DH et al: The association of gonorrhoea and syphilis with premature birth and low birthweight. Genitourin Med 69: 98, 1993

4. Kohen DP: Neonatal gonococcal arthritis: Three cases and review of the literature. Pediatrics 53: 436, 1974

5. Israel KS, Rissing KB, Brooks GF: Neonatal and childhood gonococcal infections. Clin Obstet Gynecol 18: 143, 1975

6. Fleming WL, Brown WN, Donohue JF et al: National survey of venereal disease treated by physicians in 1968. JAMA 211: 11, 1970

7. Kampmeier RH: Identification of the gonococcus by Albert Neisser. Sex Transm Dis 5: 71, 1981

8. Swanson J: Cell wall outer membrane variants of Neisseria gonorrhoeae. In Brooks GF, Gotschlich EC, Holmes KK et al (eds): Immunobiology of Neisseria gonorrhoeae, pp 130–137. Washington, DC, American Society of Microbiology, 1978

9. Kellogg DS, Peacock WL, Deacon WE et al: Neisseria gonorrhoeae: I. Virulence genetically linked to clonal variation. J Bacteriol 85: 1274, 1963

10. McGee ZA et al: Relationship of Pili to colonial morphology among pathogenic and nonpathogenic species of Neisseria. Infect Immun 15: 594, 1977

11. Johnston KH, Holmes KK, Gotschlish EC: The serological classification of Neisseria gonorrhoeae I. Isolation of the outer membrane complex responsible for serotype specificity. J Exp Med 143: 741, 1976

12. Knapp JS, Holmes KK: Disseminated gonococcal infections caused by Neisseria gonorrhoeae with unique nutritional requirements. J Infect Dis 132: 204, 1975

13. Martin JE, Lester A, Price EV, Schmale JD: Comparative study of gonococcal susceptibility to penicillin in the United States, 1955-1969. J Infect Dis 122: 459, 1970

14. Jaffe HW, Biddle JW, Thornsberry C et al: National gonorrhea therapy monitoring study: In vitro antibiotic susceptibility and its correlation with treatment results. N Engl J Med 294: 5, 1976

15. Faruki H, Kohmescher RN, McKinney WP, Sperling PF: A community-based outbreak of infection with penicillin-resistant Neisseria gonorrhoeae not producing penicillinase (chromosomally mediated resistance). N Engl J Med 313: 607, 1985

16. Penicillinase-producing Neisseria gonorrhoeae—United States, Florida. MMWR 35:12, 1986

17. Centers for Disease Control: 1993 Sexually transmitted diseases treatment guidelines. MMWR 42(RR-14):1, 1993

18. Barnes RC, Holmes KK: Epidemiology of gonorrhea: Current perspectives. Epidemiol Rev 6: 1, 1984

19. Spence MR: Gonorrhea in a military prenatal population. Obstet Gynecol 42: 223, 1973

20. Charles AG, Cohen S, Kass MB, Richman R: Asymptomatic gonorrhea in prenatal patients. Am J Obstet Gynecol 108: 595, 1970

21. Centers for Disease Control and Prevention. Summary of notifiable diseases, US, 1994. MMWR 43:10, 1995

22. Gonorrhea—United States, 1983. MMWR 33:361, 1984

23. Pedersen AHB, Bonin P: Screening females for asymptomatic gonorrhea infection. Northwest Med 70: 255, 1971

24. Holmes KK, Buchanan TM, Adam JL et al: Is serology useful in gonorrhea? A critical analysis of factors influencing serodiagnosis. In Brooks GF, Gotschlich EC, Holmes KK et al (eds): Immunobiology of Neisseria gonorrhoeae, pp 370–376. Washington, DC, American Society of Microbiology, 1978

25. Christmas JT, Wendel GD, Bawdon RE et al: Concomitant infection with Neisseria gonorrhoeae and Chlamydia trachomatis in pregnancy. Obstet Gynecol 74: 297, 1989

26. Holmes KK, Johnson DW, Trostle HJ: An estimate of the risk of men acquiring gonorrhea by sexual contact with infected females. Am J Epidemiol 91: 170, 1970

27. Thin RNT, Williams DA, Nicol CS: Direct and delayed methods of immunofluorescent diagnosis of gonorrhea in women. Br J Vener Dis 47: 27, 1971

28. Rothenberg R: Ophthalmia neonatorum due to Neisseria gonorrhoeae: Prevention and treatment. Sex Transm Dis 9 (suppl 2): 187, 1979

29. Thompson TR, Swanson RE, Weisner PJ: Gonococcal ophthalmia neonatorum: Relationship of time of infection to relevant control measures. JAMA 228: 186, 1974

30. Nickerson CW: Gonorrhea amnionitis. Obstet Gynecol 48: 815, 1973

31. Plant AG, Gilbert JV, Cirtenstein MS, Capra JD: Neisseria gonorrhoeae and Neisseria meningitidis: Extracellular enzyme cleaves human immunoglobulin A. Science 190: 1103, 1975

32. Morse SA, Cacciapuoti AF, Lysko PG: Physiology of Neisseria gonorrhoeae. Adv Microbiol Physiol 20: 251, 1979

33. Perry MB, Diena BB, Ashton FE: Lipopolysaccharides of Neisseria gonorrhoeae. In Roberts RB (ed): The Gonococcus, pp 285–302. Wiley, New York, 1977

34. Glynn AA, Ward ME: Nature and heterogenicity of the antigens of Neisseria gonorrhoeae involved in the serum bactericidal reaction. Infect Immun 2: 162, 1970

35. Rice PA, Kasper DL: Characterization of serum resistance of Neisseria gonorrhoeae that disseminates: Roles of blocking antibody and gonococcal outer membrane proteins. J Clin Invest 70: 157, 1982

36. Schoolnik GK, Buchanan TM, Holmes KK: Gonococci causing disseminated gonococcal infection are resistant to the bactericidal action of normal human sera. J Clin Invest 58: 1163, 1976

37. Bataskov KL, Hariharan S, Horowitz MD et al: Gonococcal endocarditis complicating pregnancy: A case report and literature review. Obstet Gynecol 78: 494, 1991

38. Platt R, Price PA, McCormack WM: Risk of acquiring gonorrhea and prevalence of abnormal adnexal findings among women recently exposed to gonorrhea. JAMA 250: 3205, 1983

39. Knapp JS, Thornsberry C, Schoolnik GA et al: Phenotypic and epidemiologic correlates of auxotypes in Neisseria gonorrhoeae. J Infect Dis 138: 160, 1978

40. Sharp JT, Lidsky MD, Duffy J, Duncan MW: Infectious arthritis. Arch Intern Med 139: 1125, 1979

41. Sarrel PN, Pruett KA: Symptomatic gonorrhea during pregnancy. Obstet Gynecol 32: 670, 1968

42. Burkman RT, Tonascia JA, Atienza MF, King TM: Untreated endocervical gonorrhea and endometritis following elective abortion. Am J Obstet Gynecol 126: 648, 1976

43. Edwards LE, Barrada MI, Hamann AA, Hakanson EY: Gonorrhea in pregnancy. Am J Obstet Gynecol 132: 637, 1978

44. Amstey MS, Steadman KT: Asymptomatic gonorrhea and pregnancy. J Am Vener Dis Assoc 3: 14, 1976

45. Stoll BJ, Kanto WP, Glass RI, Pushkin J: Treated maternal gonorrhea without adverse effect on outcome of pregnancy. South Med J 75: 1236, 1982

46. Hansfield HH, Hodson WA, Holmes KK: Neonatal gonococcal infection: I. Orogastric contamination with Neisseria gonorrhoeae. JAMA 225: 697, 1973

47. Alger LS, Lovchik JC, Hebel JR et al: The association of Chlamydia trachomatis, Neisseria gonorrhoeae, and group B streptococci with preterm rupture of membranes and pregnancy outcome. Am J Obstet Gynecol 159: 297, 1988

48. Jones DED, Brame RG, Jones CP: Gonorrhea in obstetric patients. J Am Vener Dis Assoc 2: 30, 1976

49. Howe L: Crede's method for prevention of purulent ophthalmia in infancy in public institutions. Trans Am Opthalmol Soc 8: 52, 1987

50. Hammerschlag MR, Cummings C, Roblin PM et al: Efficacy of neonatal ocular prophylaxis for the prevention of chlamydial and gonococal conjunctivitis. N Engl J Med 320: 769, 1989

51. Hatfield EM: Causes of blindness in school children. Sight Sav Rev 33: 318, 1963

52. Lossick JG, Smeltzer MP, Curran JW: The value of the cervical gram stain in the diagnosis and treatment of gonorrhea in women in a venereal disease clinic. Sex Transm Dis 9: 124, 1982

53. Schachter J, McCormack WM, Smith RF et al: Enzyme immunoassay for diagnosis of gonorrhea. J Clin Microbiol 19: 57, 1984

54. Thomason JL, Gelbart SM, Sobieski VJ et al: Effectiveness of gonozyme for detection of gonorrhea in low-risk pregnant and gynecologic populations. Sex Transm Dis 16: 28, 1988

55. Panke ES, Yang LI, Leist PA et al: Comparison of gen-probe DNA probe test and culture for the detection of Neisseria gonorrhoeae in endocervical specimens. J Clin Microbiol 29: 883, 1991

56. Hosein IK, Kaunitz AM, Craft SJ: Detection of cervical Chlamydia trachomatis and Neisseria gonorrhoeae with deoxyribonucleic acid probe assays in obstetric patients. Am J Obstet Gynecol 167: 588, 1992

57. Summary of notifiable diseases, United States 1992. MMWR 41:3, 1993

58. Rolfs RT, Goldberg M, Sharrar RG: Risk factors for syphilis: Cocaine use and prostitution. Am J Public Health 80: 853, 1990

59. Farley TA, Hadler JL, Gunn RA: The syphilis epidemic in Connecticut: Relationship to drug use and prostitution. Sex Transm Dis 17: 163, 1990

60. Hudson EH: Treponematoses and anthropology. Ann Intern Med 58: 1037, 1963

61. Truffi M: Hieronymous Frascator's Syphilis: A Translation in Prose. 2nd ed. St. Louis, Urologic and Cutaneous Press, 1931

62. Turner TB, Hollander DH: Biology of the treponematoses. WHO Monogr Ser 35: 1, 1957

63. Fieldsteel AH, Miao RH: Genetics of treponema. In Schell RF, Musher DM (eds): Pathogenesis and Immunology of Treponemal Infections, pp 39–56. New York, Dekker, 1982

64. Clarkson KA: Technique of darkfield examination. Med Tech Bull 7: 199, 1956

65. Musher DM, Hague-Park M, Gyorkey F et al: The interaction between Treponema pallidum and human polymorphonuclear leukocytes. J Infect Dis 147: 77, 1983

66. Lukehart SA, Baker-Zander SA, Seu S: Characterization of lymphocyte responsiveness in early experimental syphilis: II. Nature of cellular infiltration and Treponema pallidum distribution in testicular lesions. J Immunol 124: 454, 1980

67. Soltani K, Aronson IK, Brickman F, Lorincz AL: Detection by direct immunofluorescence of antibodies to Treponema pallidum in cutaneous infiltrates of rabbit syphilomas. J Infect Dis 138: 222, 1978

68. Lukehart SA, Miller JN: Demonstration of the in vitro phagocytosis of Treponema pallidum by rabbit peritoneal macrophages. J Immunol 121: 2014, 1978

69. Schroeter AL, Turner RH, Lucas JB, Brown WJ: Therapy for incubating syphilis: Effectiveness of gonorrhea treatment. JAMA 318: 711, 1971

70. Risseeuw-Appel IM, Kothe FC: Transfusion syphilis: A case report. Sex Transm Dis 10: 200, 1983

71. Turner TB, Bauer JH, Diseker TH et al: The viability of T. pallidum in refrigerated whole blood and in dessicated serum. Trans Assoc Am Phys 56: 106, 1941

72. Ricci JM, Fojaco RM, O'Sullivan MJ: Congenital syphilis: The University of Miami/Jackson Memorial Medical Center experience, 1986-1988. Obstet Gynecol 74: 687, 1989

73. Minkoff HL, McCalla S, Delke I et al: The relationship of cocaine use to syphilis and human immunodeficiency virus infections among inner city parturient women. Am J Obstet Gynecol 163: 521, 1990

74. National Center for Health Statistics, US Department of Health and Human Services: Vital Statistics of the U.S., 1977. Vol I, Natality. Washington, DC, US Government Printing Office, 1981

75. Judson F, Penley KA, Robinson ME, Smith JK: Comparative prevalence rates of sexually transmitted diseases in heterosexual and homosexual men. Am J Epidemiol 112: 836, 1980

76. Thin RN: Early syphilis in the adult. In Holmes KK, Mardh P-A, Sparling PF, Wiesner PJ et al (eds): Sexually Transmitted Diseases, 2nd ed, pp 221—230. New York, McGraw Hill, 1984

77. Fowler W: The erythrocyte sedimentation rate in syphilis. Br J Vener Dis 52: 309, 1976

78. Feher J, Somogyi T, Timmer M, J'ozsa L: Early syphilitic hepatitis. Lancet 2: 896, 1975

79. Bhorade MS, Carag HB, Lee HJ et al: Nephropathy of secondary syphilis: A clinical and pathological spectrum. JAMA 316: 1159, 1971

80. Rudolph AH: The serologic diagnosis of syphilis. Consultant 16: 49, 1976

81. Spangler AS, Jackson JH, Fiumara NJ, Warthin TA: Syphilis with a negative blood test reaction. JAMA 189: 87, 1964

82. Berkowitz K, Baxi L, Fox HE: False-negative syphilis screening: The prozone phenomenon, nonimmune hydrops, and diagnosis of syphilis during pregnancy. Am J Obstet Gynecol 163: 975, 1990

83. Gjestland T: The Oslo study of untreated syphilis: An epidemiologic investigation of the natural course of syphilitic infection based on a restudy of the Boeck-Bruusgaand material. Acta Derm Venereol 35 (suppl 36): 147, 1955

84. Rockwell DH, Yobs AR, Moore MB Jr et al: The Tuskegee study of untreated syphilis. Arch Intern Med 114: 792, 1964

85. Jaffe HW, Larsen SA, Jones OG: Hemagglutination tests for syphilis antibody. Am J Clin Pathol 70: 230, 1978

86. Ijsselmuiden OE, VanDer Sluis JJ, Mulder A et al: An IgM capture enzyme linked immunosorbent assay to detect IgM antibodies to treponemes in patients with syphilis. Genitourin Med 65: 79, 1989

87. Wendel GD, Maberry MC, Christmas JT et al: Examination of amniotic fluid in diagnosing congenital syphilis with fetal death. Obstet Gynecol 74: 967, 1989

88. Klein VR, Cox SM, Mitchell MD et al: The Jarisch-Herxheimer reaction complicating syphilotherapy in pregnancy. Obstet Gynecol 75: 375, 1990

89. Lucas MJ, Theriot SK, Wendel GD Jr: Doppler systolic-diastolic ratios in pregnancy complicated by syphilis. Obstet Gynecol 77: 217, 1991

90. Mascola L, Pelosi R, Blount JH et al: Congenital syphilis: Why is it still occurring? JAMA 252: 1719, 1984

91. Harter CA, Benirschke K: Fetal syphilis in the first trimester. Am J Obstet Gynecol 124: 705, 1976

92. Fiumara N, Fleming WL, Downing JG, Good FL: The incidence of prenatal syphilis at the Boston City Hospital. N Engl J Med 247: 48, 1952

93. Walter P, Blot P, Ivanoff B: The placental lesions in congenital syphilis. Virchows Arch 397: 313, 1982

94. Qureshi F, Jacques SM, Reyes MP: Placental histopathology in syphilis. Hum Pathol 24: 779, 1993

95. Knowles S, Frost T: Umbilical cord sclerosis as an indicator of congenital syphilis. J Clin Pathol 42: 1157, 1989

96. Hallak M, Peipert JF, Ludomirsky A et al: Nonimmune hydrops fetalis and fetal congenital syphilis. J Reprod Med 37: 173, 1992

97. Barton JR, Thorpe EM Jr, Shaver DC et al: Nonimmune hydrops fetalis associated with maternal infection with syphilis. Am J Obstet Gynecol 167: 56, 1992

98. Nathan L, Twickler DM, Peters MT et al: Fetal syphilis: Correlation of sonographic findings and rabbit infectivity testing of amniotic fluid. J Ultrasound Med 2: 97, 1993

99. Satin AJ, Twickler DM, Wendel GD Jr: Congenital syphilis associated with dilatation of fetal small bowel. J Ultrasound Med 11: 49, 1992

100. Hill LM, Maloney JB: An unusual constellation of sonographic findings associated with congenital syphilis. Obstet Gynecol 78: 895, 1991

101. Hira SK, Bhat GJ, Patel JB et al: Early congenital syphilis: Clinico-radiologic features in 202 patients. Sex Transm Dis 12: 177, 1985

102. Dorfman DH, Glaser JH: Congenital syphilis presenting in infants after the newborn period. N Engl J Med 323: 1299, 1990

103. Fiumara NJ: Syphilis in newborn children. Clin Obstet Gynecol 18: 183, 1975

104. Centers for Disease Control: Guidelines for the prevention and control of congenital syphilis. MMWR 37(suppl 1):1, 1988

105. Idsoe O, Guthe T, Willcox RR: Penicillin in the treatment of syphilis: The experience of three decades. Bull WHO 47 (suppl): 1, 1972

106. Schroeter AL, Lucas JB, Price EV, Falcon VH: Treatment for early syphilis and reactivity of serologic tests. JAMA 221: 471, 1972

107. Philipson A, Sabath LD, Charles D: Transplacental passage of erythromycin and clindamycin. N Engl J Med 288: 1219, 1973

108. Sullivan TJ, Yecies LD, Shatz GS et al: Desensitization of patients allergic to penicillin using orally administered B-lactam antibiotics. J Allergy Clin Immunol 69: 275, 1982

109. Wendel GD, Stark BJ, Jamison RB et al: Penicillin allergy and desensitization in serious infections during pregnancy. N Engl J Med 312: 1229, 1985