Neonatal Infections: A Global Perspective

Infection as A Cause of Neonatal Death

The majority of infection-related neonatal deaths are thought to be caused by bacterial sepsis and meningitis, respiratory infection, neonatal tetanus, diarrhea, and omphalitis. Very little is known about viral infections in the newborn period in LMICs. Neonatal deaths caused by infection may occur early in the neonatal period, in the first 7 days of life, and are usually attributable to infection acquired during the peripartum process. Late neonatal deaths, those occurring from 8 to 28 days of life, are most commonly caused by acquisition of pathogens from the environment in which the vulnerable newborn is placed.

In LMICs, because about half of births and most neonatal deaths occur at home and are not attended by medical personnel, deaths are underreported, and information on cause of death is often incomplete. Remarkably few published studies worldwide present detailed surveillance data on numbers of births and neonatal deaths and on probable causes of death. Although hospital-based studies are important for accurately determining causes of morbidity and mortality, they may not reflect what is happening in the community and, because of selection bias, may not be representative of the population. A recent review summarized 32 community-based studies that were published from 1990 to 2007. Infection-specific mortality was found to range from 2.7 per 1000 live births in South Africa to 38.6 per 1000 live births in Somalia. Overall, 8% to 80% (median, 36.5%; interquartile range, 26%-49%) of all neonatal deaths in developing countries were found to be attributable to infections. However, significant data gaps exist, especially from low-resource countries. There is a need for carefully conducted population-based studies that assess the number and causes of neonatal deaths resulting from bacterial and viral infections in LMICs.

In the absence of better data, global estimates for causes of neonatal deaths have been derived through statistical modeling, extrapolating from evidence available from several countries at different levels of development and neonatal mortality rates. According to these estimates, infections are the second largest cause of neonatal mortality, accounting for 27.5% of all neonatal deaths; sepsis, pneumonia, and meningitis together account for 22.5% of neonatal deaths, whereas tetanus and diarrhea account for 2.5% each. This translates to 0.8 million neonatal deaths caused by infections, most of which can be averted with appropriate prevention and management. Using available data, it is estimated that between 252,000 to 552,000 neonatal deaths resulting from sepsis, pneumonia, or meningitis occur in developing countries each year ( Table 2-1 ). The range is large because of the imprecision of available data.

Incidence of Neonatal Sepsis, Bacteremia, and Meningitis and Associated Mortality

Hospital and community-based studies from LMICs have been reviewed recently to determine the incidence of neonatal sepsis, bacteremia, and meningitis; the case-fatality rates (CFRs) associated with these infections; and the spectrum of bacterial pathogens in different regions of the world. The vast majority of studies did not distinguish among maternally acquired, community-acquired, and nosocomial infections. Population-based studies from developing countries have reported clinical sepsis rates ranging from 49 to 170 per 1000 live births. The majority of studies reported sepsis-associated case fatality rates greater than 30%. A carefully conducted recent population-based surveillance study from Mirzapur, a rural part of Bangladesh, attempted to capture all births and all cases of sepsis in a well-defined population through active, household-level surveillance. The incidence of clinically suspected neonatal infection was approximately 50 per 1000 live births. However, improvements continue to be made in the clinical algorithm used in Integrated Management of Newborn and Childhood Illness (IMNCI) to identify newborns with clinically suspected serious infections. Further refinements will likely lead to lower estimates when clinical signs of lower specificity, such as fast breathing, are removed.

Information on incidence rates of neonatal bacteremia (sepsis confirmed by isolation of bacteria from the blood) from developing countries is extremely limited. The median incidence of blood culture confirmed sepsis was 16 per 1000 live births in developing countries, among 18 studies reviewed recently. Berkley and colleagues reported a bacteremia rate of 5.5 per 1000 live births in rural Kenya, most likely an underestimate because only infants presenting to their referral hospital from the surrounding catchment area were included, and no active case finding through community surveillance was conducted. In Mirzapur, Bangladesh, active population-based, household-level newborn illness surveillance detected an incidence rate of bacteremia of 3.0 per 1000 person-neonatal periods. Comparative figures for early-onset neonatal bacteremia reported in the United States range from 0.8 to 1 per 1000 live births. Population-based surveillance studies with a focus on early birth detection in home-delivered babies and both bacterial and viral etiology of serious infections in young infants are currently underway in Bangladesh, India, and Pakistan and will provide more reliable estimates of incidence of sepsis and bacteremia in South Asian newborns.

Very few studies on neonatal meningitis were available to evaluate incidence and CFRs by region. The incidence of neonatal meningitis ranged from 0.33 to 7.3 per 1000 live births (average, 1 per 1000 live births), with CFRs ranging from 13% to 59% .

Bacterial Pathogens Associated with Infections in Different Geographic Regions

Historical reviews from developed countries have demonstrated that the predominant organisms responsible for neonatal infections change over time. Prospective microbiologic surveillance is therefore important to guide empirical therapy and identify potential targets for vaccine development, identify new agents of importance for neonates, recognize epidemics, and monitor changes over time. Moreover, the organisms associated with neonatal infection are different in different geographic areas, reinforcing the need for local microbiologic surveillance. In areas where blood cultures in sick neonates cannot be performed, knowledge of the bacterial flora of the maternal genital tract may serve as a surrogate marker for organisms causing early-onset neonatal sepsis, meningitis, and pneumonia. The vast majority of studies on the causes of neonatal sepsis and meningitis are hospital reviews that include data on infants born in hospitals as well as those transferred from home or other facilities.

A recent review highlighted the scarcity of data on pathogens associated with neonatal sepsis and meningitis in LMICs. This review found 63 studies published between 1980 and 2007 that reported etiologic data from LMICs. The review also included findings from the Young Infant Clinical Signs Studies and community-based data from Karachi. Only 12 of these studies focused on community-acquired infections. In most of the remaining studies, it was difficult to determine whether infections were of maternal origin or were hospital- or community-acquired. Because of insufficient information provided, assumptions of community acquired infections were made if this was implied by the study setting. Therefore the possible inclusion of some nosocomial infections cannot be ruled out. Also, the infants’ ages at the time of infection were not always specified. The studies varied in the detail with which culture methods were presented.

Table 2-2 gives further details about the distribution of organisms by geographic region. The review found 19 studies that reported etiologic data for the entire neonatal period. In the aggregated data of these studies, the ratio of gram-negative to gram-positive organisms was 1.6:1, and Staphylococcus aureus, Escherichia coli, and Klebsiella spp. collectively caused almost half of all infections. This pattern was consistent across all regions except Africa, where gram-positive organisms were predominant because of higher frequency of S. aureus, Streptococcus pneumoniae, and Streptococcus pyogenes .

Forty-four studies presented the etiology of early-onset neonatal sepsis in LMICs; all were facility-based studies. One fourth of all episodes of early-onset neonatal sepsis were caused by Klebsiella spp., 18% by S. aureus , 15% by E. coli , 7% by group B streptococci (GBS), and 12% were caused collectively by Acinetobacter spp. and Pseudomonas spp. The overall ratio of gram-negative to gram-positive organisms was 2:1. However, in African countries, the ratio of gram-positive to gram-negative organisms was 1:1, with a larger proportion of infections caused by S. aureus and GBS . Pseudomonas spp. and Acinetobacter spp. were found to be more common in East Asian, Pacific, and South Asian countries. S. aureus was relatively uncommon in East Asia and Latin America compared with other regions.

The review also found 11 studies that reported etiologic data on community-acquired infections occurring between 7 and 59 days of life. Almost half of the isolates in this age group were from the large World Health Organization (WHO)-sponsored multicenter Young Infant Study conducted in the early 1990s in four developing countries: Ethiopia, The Gambia, Papua New Guinea, and the Philippines. The ratio of gram-negative to gram-positive organisms in this group was 0.8:1, with higher proportions of Salmonella spp. , Haemophilus influenzae, S. pneumoniae, and S. pyogenes, compared with the first week of life. Additional data on etiology of serious infections in the first 2 months of life is forthcoming from the second Young Infant Clinical Sign Study.

Although data are limited, studies involving home-delivered babies or babies from maternity hospitals and rural referral hospitals found gram-negative organisms to be more than three times as common as gram-positive organisms (ratio of 3:1 among home births, 3.5:1 among rural referral hospitals). Three gram-negative bacteria ( E. coli, Klebsiella spp., and Pseudomonas spp.) accounted for 43% to 64% of all infections, and the gram-positive S. aureus accounted for 8% to 21% of all infections. Among babies born at home, gram-negative organisms were responsible for 77% of all neonatal infections. In Mirzapur, Bangladesh, among home-born newborns identified through population-based household surveillance, half of all culture-proven episodes of suspected sepsis were due to gram-negative organisms, including Klebsiella spp., Pseudomonas spp., Acinetobacter spp., and Enterobacter spp. Among gram-positive cultures, S. aureus was the most common isolate, responsible for one third of all positive cultures.

The ongoing study evaluating pathogens of neonatal infections in South Asia (ANISA study) will provide valuable information on etiology of infections in low resource community settings.

Antimicrobial Resistance in Neonatal Pathogens

Rising rates of antimicrobial resistance among common pathogens involved in neonatal infections are being observed in LMICs. However, limited published information is available on antimicrobial resistance patterns among neonatal pathogens from community settings where a large proportion of births take place at home. A recent review identified only 10 studies during 1990 to 2007, including two unpublished works, that contributed resistance data from community settings in low- and middle-income countries, primarily regarding Klebsiella spp., E. coli, and S. aureus . Compared with data from hospital settings, resistance rates were lower in community-acquired infections. Among E. coli, greater than 70% of isolates were resistant to ampicillin, and 13% were resistant to gentamicin. Among Klebsiella spp., all were resistant to ampicillin and 60% to gentamicin. However, resistance to third-generation cephalosporins was uncommon, and methicillin-resistant S. aureus (MRSA) occurred rarely. Another recent review identified 19 studies from 13 different countries on resistance patterns from community settings. The study showed high rates of resistance of gram-negative bacteria and S. aureus . However, it is unclear how many of these infections were truly maternally or community acquired. Additional data on antimicrobial resistance patterns of neonatal pathogens encountered in home-delivered babies are needed to develop evidence-based guidelines for management. The ANISA study mentioned earlier will provide much-needed information from South Asia on antimicrobial resistance among neonatal pathogens among home-born infants.

By contrast, several studies from hospitalized infants in hospitals from LMICs show alarming antimicrobial resistance rates among neonatal pathogens in hospital nurseries. A large recent review showed that greater than 70% of neonatal isolates from hospitals of LMICs were resistant to ampicillin and gentamicin—the recommended regimen for the management of neonatal sepsis. Resistance was also documented against expensive second- and third-line agents; 46% of E. coli and 51% of Klebsiella spp. were resistant to the third-generation cephalosporin cefotaxime. Equally disturbing was the high prevalence of MRSA isolates, especially in South Asia, where they comprised 56% of all isolates. Pan-resistant Acinetobacter spp. infections are also now widely reported. In these resource-constrained settings, many multidrug-resistant pathogens are now unfortunately untreatable, and high mortality rates are observed.

Nosocomial Infections

Hospitals in LMICs are ill-equipped to provide hygienic care to the vulnerable newborn infant. A recent review of the rates of neonatal infections among hospital-born babies in LMICs found rates to be 3 to 20 times higher than those observed in industrialized countries. Moreover, a high proportion of infections in the early neonatal period were due to Klebsiella spp., Pseudomonas spp., and S. aureus , rather than organisms typically associated with the maternal birth canal, suggesting acquisition from the hospital environment, rather than the mother. Overall, gram-negative rods were found to be predominant, comprising 60% of all positive cultures from newborns. Klebsiella spp. were found to be the major pathogens, present in 23% of cases, followed by S. aureus (16.3%) and E. coli (12.2%).

High nosocomial infection rates observed among hospital-born babies in LMICs are attributable to lack of aseptic delivery and hand hygiene; lack of essential supplies, such as running water, soap, and gloves; equipment shortages; lack of sterilization facilities; lack of knowledge and training regarding adequate sterilization; overcrowded and understaffed health facilities; and inappropriate and prolonged use of antibiotics.

Acute Respiratory Infections/Pneumonia

The onset of pneumonia in neonates may occur early (acquired during birth from organisms that colonize or infect the maternal genital tract) or late (acquired later from organisms in the hospital, home, or community). Although only a few studies of the bacteriology of neonatal pneumonia have been performed, the findings suggest that organisms causing pneumonia are similar to those that cause neonatal sepsis. The role of viruses and agents such as Bordetella pertussis in neonatal pneumonia, especially in LMICs, remains unclear. Recent studies from developed countries suggest that viruses, including respiratory syncytial virus, parainfluenza viruses, adenoviruses, and influenza viruses, contribute to respiratory morbidity and mortality, especially during epidemic periods (G.L. Darmstadt, unpublished data). Maternal influenza vaccination during pregnancy in Bangladesh reduced febrile respiratory illnesses in their young infants by one third, compared with infants of mothers not receiving influenza vaccine, suggesting an important role for influenza viruses in neonatal acute respiratory infections (ARI).

Because of similarities in presentation, pneumonia in neonates is very difficult to differentiate from neonatal sepsis or meningitis, and all three diseases are often grouped under one category and treated similarly. Therefore assessing the true burden of neonatal respiratory infections is very difficult. A respiratory rate greater than 60 per minute in an infant younger than 2 months has been proposed as a sensitive sign of serious illness and possible pneumonia by the WHO, but concerns about low specificity resulting from conditions such as transient tachypnea of the newborn and upper respiratory infections remain to be addressed. In a review of the causes of neonatal mortality, Liu and colleagues estimated that 325,000 neonatal deaths caused by pneumonia occur annually in LMICs. In a carefully conducted community-based study in rural India, published in 1993, Bang and associates determined that 66% of pneumonia deaths in the first year of life occurred in the neonatal period.

It is difficult to determine the incidence of neonatal ARI/pneumonia in LMICs because many sick neonates are never referred for medical care. In a large community-based study of ARI in Bangladeshi children, the highest incidence of ARI was in children younger than 5 months. In the study by Bang and associates, there were 64 cases of pneumonia among 3100 children (incidence of 21/1000), but this finding underestimates the true incidence because it was known that many neonates were never brought for care. A community-based study conducted by English and colleagues in Kenya found the incidence of pneumonia to be as high as 81 per 1000 for children younger than 2 months. The risk of pneumonia and of ARI-related death increases in infants who are LBW and/or malnourished and in those who are not breastfed. In a study of LBW infants in India in which infants were visited weekly and mothers queried about disease, there were 61 episodes of moderate-to-severe ARI among 211 LBW infants and 125 episodes among 448 normal-weight infants. Although 33% of episodes occurred in LBW infants, 79% of the deaths occurred in this weight group.

Management of pneumonia in neonates follows the same principles as neonatal sepsis because the syndrome is difficult to distinguish clinically from sepsis. Trials are underway in Nigeria, Kenya, Democratic Republic of Congo, and Pakistan evaluating the efficacy of therapy in young infants who present with fast breathing as their sole clinical sign of illness.

Diarrhea

Although diarrheal diseases are important killers of children younger than 1 year, most deaths resulting from diarrhea during infancy occur in the second 6 months of life. Worldwide, only 1% of deaths in the neonatal period are attributed to diarrhea. The high prevalence of breastfeeding in the first month of life in LMICs most likely protects breast-fed newborns from diarrhea.

Kotloff and colleagues studied the etiology and burden of moderate-to-severe diarrhea in Kenya, Mali, Mozambique, The Gambia, Bangladesh, India, and Pakistan. The investigators enrolled 4029 infants over a period of 3 years and compared them with 4878 matched control subjects. Four pathogens were significantly associated with moderate-to-severe diarrhea: rotavirus, Cryptosporidium , Shigella, and enterotoxigenic E. coli . Rotavirus was the most common agent, with an incidence of 7 episodes per 100 child-years during infancy. Black and colleagues performed community studies of diarrheal epidemiology and etiology in a periurban community in Peru. The incidence of diarrhea was 9.8 episodes per child in the first year of life and did not differ significantly by month of age (0.64-1.0 episode per child-month). Mahmud and colleagues prospectively followed a cohort of 1476 Pakistani newborns from four different communities. Eighteen percent of infants evaluated in the first month of life (180/1028) had diarrhea.

Although home births still account for many of the births in LMICs, those born in hospitals are at risk for nosocomial diarrheal infections. Aye and associates studied diarrheal morbidity in neonates born at the largest maternity hospital in Rangoon, Myanmar. Diarrhea was a significant problem, with rates of 7 cases per 1000 live births for infants born vaginally and 50 per 1000 for infants delivered by cesarean section. These differences were attributed to the following: infants born by cesarean section remained hospitalized longer, were handled more by staff and less by their own mothers, and were less likely to be exclusively breastfed.

Rotavirus is one of the most important causes of diarrhea among infants and children worldwide, occurring most commonly in infants aged 3 months to 2 years. In LMICs, most rotavirus infections occur early in infancy. There are few reports of rotavirus diarrhea in newborns. It appears that in most cases, neonatal infection is asymptomatic, and that neonatal infection may protect against severe diarrhea in subsequent infections. Neonates are generally infected with unusual rotavirus strains that may be less virulent and may serve as natural immunogens. Exposure to the asymptomatic rotavirus I321 strain, in particular, has been shown to confer protection against symptomatic diarrheal episodes caused by rotavirus among neonates.

The rate of infection among neonates, however, may be more common than was previously thought. Cicirello and associates screened 169 newborns at six hospitals in Delhi, India and found a rotavirus prevalence of 26%. Prevalence increased directly with length of hospital stay. More recently, Ramani and associates found the prevalence of rotavirus among neonates with gastrointestinal symptoms to be as high as 55% in a tertiary hospital in southern India. Gladstone and colleagues studied a cohort of 373 children in India and found 56% were infected with rotavirus by 6 months of age. Rotavirus was identified in 15.2% of all diarrheal episodes. The high prevalence of neonatal infections in India (and perhaps in other low-resource country settings) could lead to priming of the immune system and have implications for vaccine efficacy. Several of the community-based studies reviewed earlier present data on diarrhea as a cause of neonatal death. In these studies, diarrhea was responsible for 1% to 12% of all neonatal deaths. In 9 of the 10 studies, 70 of 2673 neonatal deaths (3%) were attributed to diarrhea. Whereas diarrhea is more common in infants after 6 months of age, it is also associated with morbidity and, in some cases, mortality, for neonates in LMICs.

Omphalitis

In LMICs, aseptic delivery techniques and hygienic cord care have markedly decreased the occurrence of umbilical infection or omphalitis. Furthermore, prompt diagnosis and antimicrobial therapy have decreased morbidity and mortality if omphalitis develops. Omphalitis continues to be an important problem, however, where clean delivery and hygienic cord care practices remain a challenge, particularly among the world’s 60 million home births, which account for nearly half of all births, as well as for many facility-based births in low-resource settings. The necrotic tissue of the umbilical cord is an excellent medium for bacterial growth. The umbilical stump is rapidly colonized by bacteria from the maternal genital tract and from the environment. This colonized necrotic tissue, in close proximity to umbilical vessels, provides microbial pathogens with direct access to the bloodstream. Thus invasion of pathogens via the umbilicus may occur with or without the presence of signs of omphalitis, such as redness, pus discharge, swelling, or foul odor.

Omphalitis is associated with increased risk of mortality. Omphalitis may remain a localized infection or may spread to the abdominal wall, the peritoneum, the umbilical or portal vessels, or the liver. Infants who present with abdominal wall cellulitis or necrotizing fasciitis have a high incidence of associated bacteremia (often polymicrobial) and a high mortality rate.

Limited data are available on risk factors and incidence of umbilical infections from LMICs, especially from community settings. Overall, incidence of omphalitis in hospital-based studies has ranged from 2 to 77 per 1000 hospital-born infants, with the CFR ranging from 0% to 15%. Mullany and colleagues defined clinical algorithms for identification of umbilical infections and reported a 15% incidence of mild omphalitis, defined as the presence of moderate redness (<2 cm extension of redness onto the abdominal skin at the base of the cord stump) and a 1% incidence of severe omphalitis, defined as severe redness with pus, among 15,123 newborn babies identified in rural Nepal through community-based household surveillance. In Pemba, Tanzania, 9550 cord assessments in 1653 infants identified an omphalitis rate ranging from 1%, based on a definition of moderate to severe redness with pus discharge, to 12.0%, based on the presence of pus and foul odor. Mir and colleagues recently studied the burden and etiology of omphalitis from a community setting in Karachi, Pakistan, with a high proportion of unskilled home deliveries, and found the incidence of omphalitis to be 21%, with 2% of cases associated with sepsis.

A key risk factor for development of omphalitis in the community included topical applications of potentially unclean substances (e.g., mustard oil). Hand washing by the birth attendant with soap provided in the clean delivery kit, consistent hand washing by the mother, and the practice of skin-to-skin care reduced the risk of omphalitis.

Some information on microbiologic etiology of omphalitis from LMICs is available. Over a 2-year period, Güvenç and associates identified 88 newborns with omphalitis at a university hospital in eastern Turkey. Gram-positive organisms were isolated from 68% of umbilical cultures; gram-negative organisms were isolated from 60%, and multiple organisms were cultured in 28% of patients. Airede studied 33 Nigerian neonates with omphalitis. Aerobic bacteria were isolated from 70%, and anaerobic bacteria were isolated from 30%. Sixty percent of the aerobic isolates were gram-positive organisms, and polymicrobial isolates were common. Faridi and colleagues in India identified gram-negative organisms more frequently than gram-positive organisms (57% vs. 43%), but S. aureus was the single most frequent isolate (28%). In a study from Papua New Guinea, umbilical cultures were performed in 116 young infants with signs suggestive of omphalitis. The most frequently isolated organisms were group A β-hemolytic streptococci (44%), S. aureus (39%), Klebsiella spp. (17%), E. coli (17%), and Proteus mirabilis (16%). In infants with both omphalitis and bacteremia, S. aureus , S. pyogenes , and Klebsiella pneumoniae were isolated from both sites. In Thailand, postdischarge follow-up cultures from 180 newborns yielded a positive culture in all cases, mostly commonly for Klebsiella spp. (60%), E. coli (37%), Enterobacter spp. (32%), and S. aureus . In Oman, cultures from 207 newborns with signs of omphalitis yielded a positive culture in 191 cases; 57% were positive for S. aureus , 14% for E. coli, and 10% for Klebsiella spp.

Community-based data on the etiology of omphalitis in LMICs are scarce, particularly from African settings. Two recent studies in South Asia reported the etiology of omphalitis in community settings. Mir and colleagues in Pakistan identified S. aureus as the most common pathogen (52%), followed by S. pyogenes (18%), GBS (10%), Pseudomonas spp. (8.9%), Aeromonas spp. (3.2%), and Klebsiella spp. (2%). However, Mullany and colleagues identified gram-negative organisms as more commonly causing colonization of the umbilical stump among newborns who received dry cord care in Bangladesh. Gram-negative organisms were isolated in 76.3% of all swabs, compared with 55.2% of swabs that yielded gram-positive organisms (mainly S. aureus ). Among gram-negative organisms, E. coli was the most common organism (43%), followed by K. pneumoniae (34%) and Pseudomonas spp. (25%).

The method of caring for the umbilical cord after birth affects both bacterial colonization, time to cord separation, and risk for infection and mortality. Hygienic delivery and postnatal care practices, including hand washing and clean cord care, are important interventions to reduce risk of omphalitis and death. Clean birth kits, which package together items such as a sterile blade, sterile cord tie, and soap, are promoted in many settings, especially for home births, although evidence for impact of birth kits on reducing rates of omphalitis and neonatal mortality is limited.

During a study of pregnancy in a rural area of Papua New Guinea, Garner and colleagues detected a high prevalence of neonatal fever and umbilical infection, which were associated with the subsequent development of neonatal sepsis. They designed an intervention program for umbilical cord care that included maternal health education and umbilical care packs containing acriflavine spirit and new razor blades. Neonatal sepsis was significantly less frequent in the intervention group. Mullany and colleagues demonstrated a 75% reduction (95% CI, 47% to 88%) in severe umbilical cord infections and a 24% reduction (95% CI, −4% to 55%) in all-cause neonatal mortality in a large ( n = 15,123) community-based trial of 4% chlorhexidine cord cleansing, applied once daily for 8 of the first 10 days of life, compared with dry cord care. In infants enrolled within the first 24 hours of life, mortality was significantly reduced by 34% (95% CI, 5% to 54%) in the chlorhexidine cord cleansing group. In a third study arm, soap and water did not reduce infection or mortality risk, compared with dry cord care. Chlorhexidine treatment delayed cord separation by about 1 day. However, this was not associated with increased risk of omphalitis. Soofi and colleagues in Pakistan and Arifeen and colleagues in Bangladesh also found that use of 4% chlorhexidine application resulted in significant reduction of omphalitis and neonatal deaths.

A WHO expert review panel convened in September 2012 reviewed the evidence for use of chlorhexidine for cord care in low-income countries. The panel recommended chlorhexidine for routine cord care in home-delivered infants. However, formal guidance on this from WHO is awaited. Because of lack of sufficient evidence from hospital settings, clean cord care is still recommended for hospital-born infants, although it is acknowledged that antiseptics might benefit infants in settings where harmful substances are traditionally applied.

Tetanus

Neonatal tetanus, caused by Clostridium tetani, is an underreported “silent” illness. The disease may go unrecognized because it attacks newborns in the poorest countries of the world in the first few days of life, often while they are still confined to home, because of a high and rapid CFR (85% untreated) and because of poor access to medical care. The surveillance case definition of neonatal tetanus is relatively straightforward, that is, the ability of a newborn to suck at birth and for the first few days of life, followed by inability to suck starting between 3 and 10 days of age, spasms, stiffness, convulsions, and death.

Neonatal tetanus is a completely preventable disease. It can be prevented by immunizing the mother before or during pregnancy and/or by ensuring a clean delivery, clean cutting of the umbilical cord, and proper care of the cord in the days after birth. Clean delivery practices have additional benefits: prevention of other maternal and neonatal infections, in addition to tetanus. Tetanus threatens mothers as well as babies, and tetanus-related mortality is a complication of both induced abortion and childbirth in unimmunized women. Immunization of women with at least three doses of tetanus toxoid vaccine provides complete prevention against both maternal and neonatal tetanus.

The Maternal and Neonatal Tetanus Elimination Initiative of the United Nations Children’s Fund (UNICEF), the WHO, the United Nations Population Fund (UNFPA), and other partners, established in 1999, has led to the vaccination of hundreds of millions of women of childbearing age against tetanus, either through vaccination campaigns or during routine antenatal care (ANC) visits. Between 2000 and 2013, 31 countries, 19 of 35 states in India, and 29 of 33 provinces in Indonesia eliminated tetanus. An estimated 74% of women of childbearing age in developing countries are now adequately protected from tetanus, associated with marked and rapid declines in global deaths attributed to tetanus, from an estimated 146,000 in 2000 to 58,000 in 2010. Only an estimated 1% of global neonatal deaths are now attributed to tetanus. Progress continues, and the elimination of maternal and neonatal tetanus remains a global goal.

Ophthalmia Neonatorum

Ophthalmia neonatorum, defined as purulent conjunctivitis in the first 28 days of life, remains a common problem in many LMICs. The risk of infection in the neonate is directly related to the prevalence of maternal infection and the frequency of ocular prophylaxis. Infants born in areas of the world with high rates of sexually transmitted diseases (STDs) are at greatest risk.

Data on incidence and bacteriologic spectrum from specific countries are limited. Although a wide array of agents are cultured from infants with ophthalmia neonatorum, Neisseria gonorrhoeae (the gonococcus) and Chlamydia trachomatis are the most important etiologic agents from a global perspective and share similar mechanisms of pathogenesis. Infection is acquired from an infected mother during passage through the birth canal or through an ascending route. Clinical examination alone cannot distinguish infection caused by one etiologic agent from infection caused by another; each produces a purulent conjunctivitis. However, gonococcal ophthalmia may appear earlier and is typically more severe than chlamydial conjunctivitis. Untreated gonococcal conjunctivitis may lead to corneal scarring and blindness, whereas the risk of severe ocular damage is low with chlamydial infection. Without ocular prophylaxis, ophthalmia neonatorum will develop in 30% to 42% of infants born to mothers with untreated N. gonorrhoeae infection and in approximately 30% of infants exposed to Chlamydia .

A 5-year study from Iran showed S. aureus to be the major organism responsible for ophthalmia neonatorum. Similar predominance of S. aureus has been reported from Argentina and Pakistan. The reasons for these differences in etiology are not well understood, and data from the lowest-resource countries are not available.

Strategies to prevent or ameliorate ocular morbidity related to ophthalmia neonatorum include (1) primary prevention of STDs; (2) antenatal screening for and treatment of STDs, particularly gonorrhea and Chlamydia infection; (3) eye prophylaxis at birth; and (4) early diagnosis and treatment of ophthalmia neonatorum. For developing countries, eye prophylaxis soon after birth is the most cost-effective and feasible strategy in settings where STD rates are high. Eye prophylaxis is used primarily to prevent gonococcal ophthalmia. Primary prevention of STDs in LMICs is limited, although promotion of condom use has been successful in reducing STDs in some countries. Screening women at prenatal and STD clinics and treatment based on a syndromic approach (i.e., treat for possible infections in all women with vaginal discharge without laboratory confirmation) is cost-effective but may lead to overtreatment of uninfected women and missed cases.

Eye prophylaxis consists of cleaning the eyelids and instilling an antimicrobial agent into the eyes as soon after birth as possible. The agent should be placed directly into the conjunctival sac (using clean hands), and the eyes should not be flushed after instillation. Infants born both vaginally and by cesarean section should receive prophylaxis. Although no agent is 100% effective at preventing disease, the use of 1% silver nitrate solution (introduced by Credé in 1881) dramatically reduced the incidence of ophthalmia neonatorum. This inexpensive agent is still widely used in many parts of the world. The major problems with silver nitrate are that it may cause chemical conjunctivitis in up to 50% of infants, and it has limited antimicrobial activity against Chlamydia . In LMICs where heat and improper storage may be a problem, evaporation and concentration are particular concerns. Although 1% tetracycline and 0.5% erythromycin ointments are commonly used and are as effective as silver nitrate for the prevention of gonococcal conjunctivitis, these agents are more expensive and unavailable in many parts of the world. Moreover, silver nitrate appears to be a better prophylactic agent in areas where penicillinase-producing N. gonorrhoeae (PPNG) is a problem.

The ideal prophylactic agent for low-resource settings would have a broad antimicrobial spectrum and also be available and affordable. Povidone-iodine is an inexpensive, nontoxic topical agent that is potentially widely available. Recent studies suggest that it may be useful in preventing ophthalmia neonatorum. A prospective masked, controlled trial of ocular prophylaxis using 2.5% povidone-iodine solution, 1% silver nitrate solution, or 0.5% erythromycin ointment was conducted in Kenya. Of 3117 neonates randomized to receive a study drug, 13.1% in the povidone-iodine group versus 15.2% of those who received erythromycin and 17.5% in the silver nitrate group developed infectious conjunctivitis ( P < .01). The high rates of infection in this study despite ocular prophylaxis are striking. Although there was no significant difference among agents in prevention of gonococcal ophthalmia (1% or less for each agent), povidone-iodine was most effective in preventing chlamydial conjunctivitis. A 2003 study by the same group compared prophylaxis with one drop and with two drops of the povidone-iodine solution instilled in both eyes at birth in 719 Kenyan neonates. No cases of N. gonorrhoeae infection were identified. Double application did not change the rates of infection with C. trachomatis (4.2% and 3.9%). Although the antimicrobial spectrum of povidone-iodine is wider than that of the other topical agents and antibacterial resistance has not been demonstrated, published data on the efficacy of povidone-iodine against PPNG are not yet available. Of note, 2.5% povidone-iodine might also be useful as an antimicrobial agent for cord care, which is of relevance in the prevention of omphalitis (see earlier discussion). Another trial in Iran compared the efficacy of topical povidone-iodine versus erythromycin as prophylactic agents for ophthalmia neonatorum, compared with no prophylaxis. Among 330 infants studied, ophthalmia neonatorum developed in 9% of neonates receiving povidone-iodine, 18% of neonates receiving erythromycin, and 22% of the neonates receiving no prophylaxis. Further studies are needed on the safety and efficacy of povidone-iodine in LMICs.

The frequency of practice of ocular prophylaxis in LMICs is unknown. In consideration of the high rates of STDs among pregnant women in many low-resource settings, eye prophylaxis is an important blindness prevention strategy. For infants born at home, a single dose of antimicrobial agent for ocular prophylaxis could be added to birth kits and potentially distributed to trained birth attendants during ANC, although more information about the feasibility and acceptability of this approach is needed. The strategy of ocular prophylaxis is more cost-effective than early diagnosis and appropriate treatment. Furthermore, in areas of the world in which access to medical care is limited and effective drugs are scarce or unavailable, it may be the only viable strategy.

No prevention strategy is 100% effective. Even with prophylaxis, 5% to 10% of infants will develop ophthalmia. All infants with ophthalmia must be given appropriate treatment, even if they received prophylaxis at birth. A single dose of either ceftriaxone (2–50mg/kg intravenously [IV] or intramuscularly [IM], not to exceed 125mg) or cefotaxime (100mg/kg, IV or IM) is effective therapy for gonococcal ophthalmia caused by both PPNG and non-PPNG strains. Gentamicin and kanamycin also have been shown to be effective therapeutic agents and may be more readily available in some settings. Rarely, gonococcal infection acquired at birth may become disseminated, resulting in arthritis, septicemia, and even meningitis. Neonates with disseminated gonococcal disease require systemic therapy with ceftriaxone (25-50mg/kg once daily) or cefotaxime (25mg/kg IM or IV twice daily) for 7 days for arthritis or sepsis or 10 to 14 days for meningitis. If a lumbar puncture cannot be performed and meningitis cannot be ruled out in an infant with evidence of dissemination, the longer period of therapy should be chosen. Infants with chlamydial conjunctivitis should receive a 2-week course of oral erythromycin (50 mg/kg/day in four divided doses). After the immediate neonatal period, oral sulfonamides may be used.

Human Immunodeficiency Virus (HIV) Infection

The Joint United Nations Programme on HIV/AIDS (UNAIDS) and the WHO estimate that in 2011 approximately 34 million people worldwide were infected with HIV and new infections were occurring at a rate of approximately 2.5 million per year. Most HIV infections occur in LMICs; greater than 90% of those infected live in sub-Saharan Africa, Asia, Latin America, or the Caribbean. Women are particularly vulnerable to HIV infection; worldwide, approximately 50% of cases occur in women. The proportion of women infected with HIV has increased in many regions, with women representing approximately 58% of HIV infections in sub-Saharan Africa. An estimated 330,000 children were infected with HIV in 2011, mostly by maternal-to-child transmission, either in utero, at the time of delivery, or through breastfeeding.

Because HIV increases deaths among young adults, the acquired immunodeficiency syndrome (AIDS) epidemic has resulted in a generation of AIDS orphans. As of 2011, about 17.3 million children younger than 18 years have lost one or both parents to AIDS, with the vast majority in sub-Saharan Africa. It is well known that maternal mortality increases neonatal and infant deaths, independent of HIV infection. Global estimates for 2011, including the number of people living with HIV infection/AIDS, the number newly infected, and total AIDS deaths, are presented in Table 2-3 .