To Dye or Not to Dye: A Randomized, Clinical Trial of a Triple Dye/Alcohol Regime Versus Dry Cord Care

METHODS

Study subjects were recruited at BC Women’s Hospital, a tertiary-level teaching hospital and referral center for the Province of British Columbia, where 7000 deliveries take place annually. Women were recruited by nurses in the single-room maternity care modules or postpartum wards. Before commencing the study, approval was received from the University of British Columbia Clinical Ethics Committee and the BC Women’s Hospital Research Review Committee. Infants were considered eligible if they had not had antibiotics since birth, if the mother had a Vancouver address, and if the mother was fluent in English, Cantonese, or Mandarin. Infants were excluded if they had to be admitted to the level III nursery. At BC Women’s, 35% of clients are of Chinese descent. Consent forms were translated into Chinese script, checked by a second translator for accuracy and cultural context, then back-translated and revised as necessary. After obtaining written consent, study subjects were randomized by means of consecutively numbered opaque envelopes within which a sticker indicated either standard cord care (2 applications of triple dye within 4 hours of birth followed by alcohol swabbing 3 times daily) or dry care (spot cleansing skin in the periumbilical area whenever it became soiled with feces. Mothers were instructed to use any mild nonperfumed soap and water, wipe the area dry with a cotton swab or cloth, and let it continue to air dry). In both groups, infants received routine daily bathing with mild bar soap and water while in hospital. Randomization was stratified on the clinical area (single room maternity care, postpartum module, or level II nursery) in which the infant was residing. The sticker was put on the infant’s cot identifying which arm of the study the infant was in. Primary nurses were responsible for either applying triple dye and teaching mothers to cleanse the umbilical stump with alcohol twice daily, or teaching mothers to clean the area with soap and water cleansing as necessary after soiling with stool. At BC Women’s, standard eye care at birth consists of prophylactic application of erythromycin ointment to the conjunctiva.

The primary outcomes were omphalitis and S aureus-induced conjunctivitis or skin infection. Omphalitis was defined as erythema (redness, swelling, and/or warmth) of the abdominal skin in the periumbilical region, extending beyond 5 mm from the umbilicus. Conjunctivitis was identified when redness involved the whole length of the conjunctiva and the attending physician had cultured the discharge and it had been found to be positive for S aureus. Skin infection was identified by a diagnosis of bullous impetigo with cultures positive for S aureus. Secondary outcomes included rates of colonization of the cord by S aureus and other pathogens.

We were unable to undertake a priori sample size calculations because we did not know the baseline rate of staphylococcal infection in our institution. Based on resources available to us in our research grant, we set a target enrollment rate of 800 subjects, 400 in each arm of the trial.

Enrolled study subjects were identified by the hospital/community liaison nurse. Hospital-based liaison nurses identify all new mothers in the hospital on a daily basis and forward information about them to community health nurses who then provide a postpartum home visiting program. Community health nurses are mandated to contact new mothers within 24 hours of hospital discharge. Most visits take place on the second or third day postdischarge. As much as possible, nurses are assigned to mothers on the basis of a common language; when this is not possible, professional translators are used. After identification of study subjects in our study, by means of stickers on the infant cots, the liaison nurses attached a study data form to the information routinely sent to the community health nurse who would be caring for that patient in the community. Community health nurses completed the study form at the time of the home visit and then faxed them back to the study coordinator at the hospital. The form documented observations of the periumbilical region, including redness, swelling, exudate, and odor; presence of skin lesions; and appearance of conjunctiva, including redness and discharge. The date of the visit, number and ages of siblings in the home, number of persons living in the home, and method of infant feeding were documented as well. Nurses noted if they had contacted the infant’s attending physician related to their observations of the cord and periumbilical area, conjunctivitis, or impetigo.

Mothers were given individually labeled swabs to take home with them from the hospital. The community health nurse obtained a swab during the home visit from around the circumference of the umbilical area. The swab was transported in Stuart’s medium, which is stable, unrefrigerated, for 48 hours. Community health nurses sent these cultures from their own health unit by courier to a central office where they were forwarded on the same day, Monday to Friday, by courier to BC Children’s laboratory. On weekends, swabs were refrigerated and stored at the health unit until Monday, when they were sent to the lab.

Specimens were plated to a 5% sheep blood agar plate, a chocolate plate, a MacConkey plate, and an anaerobic Brucella plate. A thioglycollate broth was also inoculated. The aerobic plates were incubated for 16 to 18 hours in 5% CO₂ at 35°C. These plates were then read at 18 hours and reincubated for an additional 24 hours. The anaerobic plate was incubated for 48 hours in an anaerobic jar at 35°C. The broth culture was incubated in O₂ at 35°C for 48 hours and checked daily for growth. Organisms were identified using standard microbiologic techniques. Laboratory personnel doing cultures were blinded as to exposure group.

At 2 to 3 weeks of age, mothers were telephoned by the research coordinator and asked whether their infant had been seen by a physician for a nonroutine visit and whether concerns about infection had been identified. Nurses fluent in Cantonese made phone calls to mothers speaking Chinese languages. If there had been contact with a physician related to concerns about infection, a data form was faxed to the physician’s office for documentation of the physician’s observations, diagnosis, and treatment. In the event that a mother/infant did not have a home visit or a telephone, the physician was contacted by the study nurse for the appropriate information. Nurses and physicians could not be blinded as to the exposure status of the infant because of the purple staining of the cord among infants randomized to receive triple dye.

Study groups were compared with respect to demographic and pregnancy-related variables using the χ² statistic, or when expected cell frequencies were <2, the Fisher exact test for categorical variables and the t test for continuous variables. Outcomes were compared in a similar fashion. The type I (α) error was set at 0.05.

RESULTS

A total of 766 subjects were enrolled in the study. Four subjects in the triple dye group and 3 in the dry care group were not compliant with study protocols, but their outcomes were retained in an intention-to-treat analysis. Study groups were similar with respect to mean maternal age, marital status, ethnicity, parity, and method of delivery (Table 1). Newborns in study groups were breastfeeding exclusively or in combination with formula in similar proportions. Length of newborn stay in hospital was not different among study groups. At the time of the home visit, the age of the newborn, method of infant feeding, proportion of newborns with siblings in the home, and number of persons per household was similar among study groups (Table 1).

Enrollment of study subjects took place evenly among most postpartum units (20.5%, 22.3%, 21.1%, 26.1%) with fewer on a module that only had 4 postpartum beds (8.9%) and in the level II observation nursery (1.4%). Location of study subjects was not different among study groups.

Follow-up was achieved by at least 1 method among 344 subjects (89.6%) in the triple dye group and 356 subjects (92.9%) in the dry care group. Two hundred eighty-seven mothers (74.7%) in the triple dye group and 309 mothers (80.9%) in the dry care group received home visits. The median number of days postdelivery was 3 in both groups with 96% of infants being visited between 2 to 5 days in either group.

One infant in the dry care group was diagnosed as having omphalitis. This newborn was diagnosed at 55 hours of age (while still in hospital) on the basis of erythema and edema extending 6 mm from the umbilicus (Fig 1). The pregnancy had been uncomplicated and the infant was born by cesarean section at 38 weeks’ gestation with Apgars of 8 and 9 at 1 and 5 minutes. The newborn examination at birth was normal. An initial culture before commencing antibiotics grew α-hemolytic streptococcus and coagulase-negative staphylococcus. Blood cultures and CBC (complete blood count) were normal. The infant was treated with intravenous cloxacillin and cefotaxime with resolution of the cellulitis. The infant was discharged at 3 days of age on a 10-day course of oral Keflex. At 1 month of age, a second course of Keflex was commenced for persistent erythema on the superior aspect of the cord with resolution of the erythema thereafter.

• Download figure
• Open in new tab
• Download powerpoint

Fig 1.

Erythema around the umbilical cord stump in the newborn with omphalitis receiving dry cord care.

Computing 95% confidence intervals around our rate of 1 case in 382 births (.0026 or 2.6/1000) our estimate of the true population rate is in the range of 0 to 8.57 cases per thousand. It is noteworthy that in the 4 months after our study in which the hospital protocol changed to dry cord care exclusively, another 2 cases of omphalitis were diagnosed. This follow-up rate, 2 in 2300 births or 0.85/1000, is not statistically significantly different to our study rate of 2.6/1000.

At the initial home visit, more infants in the dry care group had exudate from the umbilical cord stump (7.4% vs 0.3%; P < .001) and a foul odor (2.9% vs 0.7%; P = .04; Table 2). There were no differences in rates of observation of redness or warmth of the cord among study groups. There were no differences in rates of observations of discharge from the conjunctiva. Groups were marginally different with respect to redness of the conjunctiva (0% vs 1.4%; P = .05), but there were no swabs of the conjunctiva taken by physicians. Nurses visiting infants in the dry care group did not contact the physician more frequently because of concerns related to potential infections than the triple dye group. A perusal of reasons for calling physicians revealed that 3 infants in the triple dye group had redness or discharge from the eye and in the dry care group, 2 infants had discharge from their eyes, and 3 had redness surrounding the umbilical cord. One infant in each group was treated with ophthalmic antibiotic drops/ointment.

Swabs were obtained from 281 (73.1%) of infants in the triple dye group and 308 (80.6%) in the dry care group. Bacteriology findings are reported in Table 3. Colonization with Escherichia coli (34.2% vs 22.1%), coagulase-negative Staphylococcus (69.5% vs 50.5%), S aureus (31.3% vs 2.8%), group B Streptococcus (11.7% vs 6.0%), and bacteria in the “other” group (17.5% vs 6.8%) was significantly more common in the dry care group. There were no differences in rates of colonization with Gram-negative enteric organisms, enterococcus species, or other streptococcal species. Neither group contained infants colonized with Streptococcus A. Infants in the dry care group were significantly less likely to have no growth (1.3% vs 15.7%). There were no statistically significant differences in distributions of bacteria among the different clinical areas participating in the study.

Telephone contact was made with 228 (59.3%) of mothers in the triple dye group and 218 (57.1%) in the dry care group (Table 4). Research nurses tried to contact subjects to a maximum of 5 times, including during evenings and weekends. Reasons for not completing the telephone interview were documented. Nineteen percent of clients were not home and messages were left but were not returned. Among 5.2% of clients, the phone was no longer in service or the phone number was incorrect. A language for which the study did not have translation available accounted for 1.0%. Among an additional 16.3%, no one answered the phone and there was no answering machine. The distribution of barriers to phone contact did not differ between the groups.

At the time of the follow-up phone call, between 2 and 3 weeks after birth, there were no differences in rates of mother’s reports of redness, warmth, exudate, or foul odor associated with the cord stump. One mother in each group reported bullous lesions (raised watery rash). Neither of these rashes were observed by the community health nurse. The mother of the infant in the triple dye group did not contact her infant’s physician and the rash had resolved at the time of the phone call. The umbilical cultures grew Enterococcus faecalis, coagulase-negative Staphylococcus, and α-hemolytic streptococcus. The infant in the dry care group had been taken to a walk-in clinic and antibiotic ointment prescribed. This infant was then seen by a pediatrician who discontinued the cream and did not prescribe additional treatment. Bacteriology for this infant indicated heavy growth of S aureus and coagulase-negative Staphylococcus. Significantly more mothers in the dry care group (2.8% vs 0%) stated that their infant’s physician had mentioned concerns about infection to them compared with none in the triple dye group. Among these 6 infants, 4 were told to apply alcohol to the cord and 2 to use polysporin ointment.

There were no differences in reported rates of mothers contacting physicians in regard to concerns about infection. Among the 7 mothers in the triple dye group who contacted physicians, 2 were concerned about conjunctivitis and/or discharge from the eye, and the remaining 5 were related to redness, odor, or discharge in the umbilical region. One newborn had antibiotic ointment prescribed for the eyes and 1 had antibiotic cream prescribed for the periumbilical area. In the dry care group, all 8 calls related to concerns about the umbilical cord. One mother was advised to use alcohol, and 1 was prescribed polysporin to apply to the periumbilical region.

Among nonroutine visits to physicians, 5 in the triple dye group were related to discharge from the eye; 2 of these were treated with antibiotic drops. One was for discharge from the cord, which was not treated. One other was for reasons unrelated to study outcomes. In the dry care group, 4 visits were related to discharge, warmth, redness, and/or foul odor around the umbilical stump and 1 for red conjunctiva. One cord stump was treated with antibiotic cream, and the conjunctivitis was treated with antibiotic eye drops. All other visits were for reasons unrelated to study outcomes, including milk allergies, respiratory infections, jaundice, and fever. There were no differences in the proportions of infants given antibiotics in the first 2 to 3 weeks of life among groups. Among 7 infants receiving antibiotics in the triple dye group, 4 were for eye discharge, 1 was for an ear infection, 1 for fever of unknown origin, and 1 for odor emanating from the umbilical stump. In the dry care group, 6 infants were given antibiotics; 3 were for discharge/odor from the umbilical stump, 1 for conjunctivitis, 1 for fever of unknown origin, and 1 for a respiratory infection.

DISCUSSION

Our pilot study demonstrates that the risk of serious infection arising from the untreated umbilical cord is <0.9%. A 1997 review has estimated the incidence of omphalitis among hospitalized infants to be 0.7% or 7/1000.⁹ Eighty-four infants were identified with a discharge diagnosis of omphalitis at Children’s Hospital in Los Angeles between 1967 and 1985.¹⁰ The most common bacteria cultured was S aureus. Seven infants developed NF and of these, only 1 survived. All developed a coagulopathy, and 3 had seizures. The mortality rate overall for omphalitis was 7% in this series.

A cohort study from Stanford University Hospital reported on various methods of cord care after all infants were cultured and followed for 2 postnatal visits by their primary pediatric care physicians.¹¹ Reports of pustular dermatitis, paronychia, and omphalitis were accepted as staphylococcal infections, and 50% of such reported infections were cultured and documented as being attributable to coagulase-positive staphylococci. Conjunctivitis was accepted as staphylococcal infection only when this organism was cultured from the conjunctival discharge. Follow-up was achieved on 94% of infants discharged from the nursery. Colonization with coagulase-positive staphylococci in groups using hexachlorophene or bacitracin was <10%; among 2 groups using only soap and alcohol, 16% to 31%. Rates of pustular dermatitis ranged from 0.3 to 1.8 in the hexachlorophene and soap/bacitracin groups compared with 5.4 to 7.3 in the soap and alcohol groups; for paronychia-cellulitis 0.2 to 0.4 compared with 0.9 to 1.3; and for conjunctivitis 0 to 0.4 compared with 0.8 to 1.3. All comparison groups consisted of 130 to 400 subjects.

A 1985 Norwegian study also cultured bacterial samples from the nose and umbilicus at 5 to 6 days and followed infants until 6 weeks of life, comparing whole body soap wash with bactericidal agents.¹² Colonization rates for the control group (whole body soap wash) were 91% for S aureus, 23% for Escherichia coli, and 10% for group G Streptococcus. The infection rate for this group was reported as 10.8%; half acquired within the nursery and half after discharge. Infections (pemphigus, paronychia, conjunctivitis, umbilical infection) were almost primarily caused (96%) by S aureus.

The relationship between infection and colonization cannot be addressed in our study because only 1 infection was diagnosed. If such a relationship exists, then our findings of much higher rates of colonization in the dry care group for certain groups of pathogens, such as staphylococci and E coli, are cause for concern. Stark and Harrisson,¹³ in their 1992 series of 370 infants with dry cord care, reported that 44 (12%) of infants developed S aureus infections. The infections consisted of septic blisters (n = 27), conjunctivitis (n = 9), or umbilical infections (n = 5). There was a significant difference in infection rate (P < .001) between those neonates with heavy (confluent) growth and either no growth or less than confluent growth.

Our study suggests that omphalitis remains a clinical entity and that there is potential risk in discontinuing bacteriocidal treatment of the umbilical cord stump. Cessation of bactericidal care of the umbilical stump must be accompanied by vigilant attention to the signs and symptoms of omphalitis. In our study, some infants were treated after discharge with local antibiotics in response to developing symptoms of infection The initial presentation of NF is one of foul-smelling umbilical discharge, erythema, and induration around the umbilicus, with rapid progression to frank gangrene.⁴ Death often occurs within 24 to 72 hours. A review of 103 neonates with omphalitis in Oman reported a rate of NF in 14 or 13.5% among infants with omphalitis. All were full-term newborns of normal birth weight. Risk factors included home birth and maternal infection, 2 factors that were not present in our study. Although the most common organism cultured was S aureus, some newborns were colonized only by Gram-negative organisms. The changing landscape of microbes within institutions and the community may dictate a case-specific approach to signs and symptoms of omphalitis and other infections. Areas in which there is a higher prevalence of organisms causing NF, neonatal tetany, or bacterial multiresistance may not be suitable for dry cord care. Mandatory reporting of omphalitis may be an appropriate method of monitoring incidence rates of omphalitis in response to changing practices in management of the newborn umbilical cord.