Legal Briefs: Late-Onset Group B Streptococcus Meningitis: Should It Happen in a Newborn Intensive Care Unit?

Abstract

A 28-6/7-weeks'-gestational-age boy whose birthweight was 1,340 g was born to a 32-year-old G3P0 woman whose pregnancy was complicated by a short cervix and preterm labor. Two cervical cultures and 1 urine culture were positive for group B Streptococcus (GBS), the latest positive cervical culture taken 2 days before delivery. Antibiotics were not given despite preterm labor with a positive GBS culture and bacteriuria. The infectious disease (ID) consult retained by the plaintiff was critical of this omission. The mother was given magnesium sulfate for tocolysis and antenatal steroids. An emergency cesarean delivery was performed because of heart rate (HR) decelerations down to the 70s for 6 to 7 minutes. At the time of delivery, membranes were ruptured and clear fluid was found. Apgar scores were 91 and 105. The GBS status was documented “unknown” in the delivery and neonatal records and remained unknown throughout the infant’s hospital stay. The plaintiff ID expert was critical of the physicians caring for the infant because he felt they should have become aware of the maternal GBS status. The treating neonatologists pointed out that nothing would have been done differently for the infant with that knowledge because the infant’s blood culture was sent and he was given empiric antibiotic.Continuous positive airway pressure was started immediately after birth, but because respiratory distress syndrome developed, he was intubated and given poractant alfa (Curosurf; Cornerstone Therapeutics, Inc, Cary, NC). The infant was extubated a few hours later and remained on room air for 1 month, initially with nasal continuous positive airway pressure, then nasal cannula. The blood culture was negative, and complete blood cell (CBC) count was normal. Ampicillin and gentamicin were discontinued after 2 days. His baseline HRs averaged 146 beats per minute until caffeine was started. Then it was higher, averaging mostly 150 to 160 beats per minute. He had a few mild early apneic episodes attributed to the magnesium that the mother received. However, after the first couple of days, he had an unremarkable course. He received increasing feeds of breast milk and was a stable growing infant. After caffeine had begun, the baseline HR was slightly higher, averaging 150 to 160 seconds. A cranial ultrasound was normal on day 4. Although over 90% of the HR were less than 160 during the first month and although apneic, desaturation and bradycardic episodes were minimal and each time could be explained by an external event, on day 30 a significant and spontaneous bradycardic and desaturation event occurred. Although only a rare minimal gastric residual occurred during the previous 30 days, now a 10% residual occurred shortly after the bradycardic and desaturation event. The plaintiff neonatologist suggested that although a premature infant is entitled to have both apneic, desaturation and bradycardic episodes, as well as residuals on feedings, this was atypical for this infant who had not behaved in this way previously. Furthermore, on this day, day 30, the majority of the HRs were well over 160 beats per minute. The plaintiff neonatologist pointed out that increased vigilance was required at this point. The nurse retained by the defense said total vigilance was practiced because she was on a monitor. The defense attorney questioned the plaintiff neonatologist on how one would know there was not a heightened vigilance. The plaintiff neonatologist contended that there was absolutely no documentation in the nursing notes or mention in the physicians’ charting to acknowledge any awareness of the various “red flags.” These were signals that were different for this infant. The neonatologist asserted that a proactive approach is needed when an infant acts differently. That would have demonstrated a heightened sense of vigilance. The approach could include more frequent vital signs, blood pressure (BP) measurements, notification to the neonatologist of the events, a CBC count, C-reactive protein, or even potentially a blood culture. Standard of care did not necessarily require each and every one, but these were areas where a proactive approach could be done. But an awareness of these events being out of the ordinary was absent. The defense neonatologist disagreed by saying that premature infants behave this way; if something were unusual, the abnormality would have persisted and been continuous, not intermittent. All of the HRs would have been very high, and intolerance to feeds would have been 100% present. Furthermore, any abnormalities would have been very much worse and been simultaneously expressed by the infant. He also contended that tachycardia in a newborn is a persistent HR greater than 180 beats per minute.Later that day (day 30) the BP dropped to 53/16 with a mean of 31 mm Hg. The plaintiff neonatologist contended that not only was the diastolic pressure very, very low but that the mean was significantly lower than what this infant had ever had. He pointed out that very long stretches occurred where the BP was not taken at all. The most recent BP was over 24 hours previous and the next was 24 hours later. Moreover, the physician was not notified of this low BP. The defense nurse said that this low BP could be explained by a number of possibilities; the infant might have been eating or maybe it was taken from the leg. The HR was found to be 203 beats per minute during the time that the BP was low. The plaintiff neonatologist said that it was at this point that a septic evaluation, lumbar puncture (if tolerated), and CRP should have been performed and antibiotics started. Not to do an evaluation and start antibiotics was below the standard of care. The defense neonatologist disagreed. On day 31, the HRs continued to be high with an average of 170 beats per minute. Feeds were increased on morning rounds. Later that day, the infant suddenly developed 20 severe bradycardic episodes for which he received positive pressure ventilation. He was then placed on a nasal cannula. He had a CBC count and blood culture sent. Initially the thought was aspiration, but chest radiograph results were negative. Subsequently, a diagnosis of necrotizing enterocolitis was entertained, but abdominal radiograph results were negative. The white blood count was 5.0 × 103 u/L, the hematocrit level was 34% (.34), and the platelet count was 804 × 103 u/L. One hour after the severe bradycardias, amikacin and vancomycin were ordered. The plaintiff ID expert maintained that because the mother was colonized with GBS, infection with GBS should have been suspected and antibiotics that would better attack GBS should have been chosen. Furthermore, antibiotics that could penetrate the central nervous system (CNS) should have been chosen. An LP was not done, and the ever-present possibility of meningitis should have been entertained. He pointed out that amikacin is not effective against GBS, and vancomycin does not penetrate the CNS well. Three hours later, the infant was endotracheally intubated. Three and a half hours after the event, amikacin was given, and a half an hour later vancomycin was administered. Both the plaintiff ID and neonatologist were critical of this excessive time lag between symptoms consistent with infection and the institution of antibiotics. Six hours later, piperacillin-tazobactam (Zosyn; Wyeth Pharmaceuticals, Inc, Philadelphia, PA) was started. Both the plaintiff neonatologist and ID contended that if the early signs, such as the tachycardias, gastric residuals, and low BP, had been appreciated (36 to 48 hours before the event), more frequent vital signs, BPs, and a CBC count would have revealed that the infant was becoming sick. Timely intervention would have prevented this dramatic deterioration. The defense neonatologist disagreed. He declared that severe bradycardias ushered in the sickness and that the earlier signs that the plaintiff neonatologist referred to as “red flags” were unrelated. The plaintiff neonatologist and ID stressed that the infant was bacteremic 24 to 36 hours before the event. Furthermore, both maintained that a 3-½-hour delay between the event and administration of the antibiotics was clearly substandard. The defense neonatologist disagreed and pointed out that many things needed to be done including nasal cannula, radiographs, and blood tests. The plaintiff neonatologist contended only a blood culture was required before starting the antibiotics and that all the interventions could have and should have been performed simultaneously.Twelve hours after the blood culture was done, GBS grew. The plaintiff neonatologist and ID both contended that GBS infection was present for at least 1 day and a half before deterioration. The defense neonatologist said that GBS is fulminant and does not have a time period where warning signs are given, maybe a 30-minute period of bacteremia at most before deterioration. The plaintiff neonatologist and ID disagreed and pointed out that late-onset sepsis from GBS is typically insidious and presents exactly as it did in this infant. After the blood culture was found to be positive, breast milk was sent for culture. It grew GBS and 2 other organisms. The plaintiff ID expert thought that the GBS was transmitted to the child via the breast milk. He could not differentiate if it was from the skin and handling of the milk or from the ducts of the mother’s breast, but thought the former was more likely because the mother did not have mastitis or bacteremia. The plaintiff neonatologist was asked if the infant acquired the GBS from the breast milk, and he responded that he did not know because serotyping was not performed and the mother did not have mastitis, but it was possible. Asked if the mother should have been allowed to breastfeed, he answered yes. The plaintiff ID pointed out that he agreed that the infant should have been allowed to have breast milk, but that the mother should have been informed of her GBS status and should have been told that there were case reports linking mothers with GBS positive status with GBS positive breast milk and potentially late-onset sepsis. Plaintiff ID agreed that even with strict hand hygiene, it is not possible to eliminate all organisms on the skin. The environmental sources could also be contaminated. The plaintiff neonatologist said it was not possible to tell from where the GBS originated. On day 32 (1 day postbradycardic event), amikacin and vancomycin were discontinued, and ampicillin was started. On day 32, a repeat cranial ultrasound was normal. On day 34, a lumbar puncture was done. The cerebral spinal fluid had 2,200 white blood cells, 140,000 red blood cells, a protein of 431 mg/dL, and a glucose of less than 5 mg/dL; cerebro- spinal fluid culture was negative. Repeat blood cultures were positive for the GBS until the ampicillin was instituted.Over the next few days, the infant had a very stormy course including disseminated intravascular coagulopathy. He developed septic emboli to both feet. The mother insisted in her deposition that the doctor did a heel stick from that foot despite the nurse admonishing him and signs on the incubator saying not to do so. The mother said in her deposition that several of the nurses wore jewelry, and that she repeatedly complained that they did not use proper hygiene. Although the plaintiff neonatologist agreed that proper hand washing is vital, there was no evidence of departure in the medical records. He pointed out that it would be a jury call to determine the credibility of the mother’s testimony about proper hygiene.The defense neonatologist maintained that 11% to 15% of 28-week infants die before discharge from a NICU. Plaintiff neonatologist strongly disagreed stating that most survive. The defense neonatologist said that of those who survive, all are abnormal on follow-up. He further maintained that they all need some level of assistance for the rest of their lives. The plaintiff neonatologist said that they tend to do well unless they are born with anomalies or have some sentinel event around the time of labor and delivery or more rarely a major problem in the NICU. Subsequent imaging was performed demonstrating grade 3 and 4 intraventricular hemorrhages. The infant developed progressive posthemorrhagic ventriculomegaly requiring a ventriculoperitoneal shunt. He was made DNR (do not resuscitate) at 2 months of age and admitted to hospice care. He was subsequently discharged from hospice to his home. He had a several admissions for seizure control. At 5 years, he had spastic quadriparesis with superimposed right hemiparesis and hypotonic trunk, neck, and facial musculature. He had cortical visual impairment. He was nonverbal and nonambulatory. He had some communication skills, and his seizures were under control. After jury selection, a settlement agreement was reached.Three clinical patterns of GBS disease occur among young infants with GBS infection that are epidemiologically distinct and related to age at onset. Early-onset GBS (EOGBS) disease ranges from 0 to 6 days and usually occurs within the first 24 hours. Late-onset GBS (LOGBS) disease has a range of 7 to 89 days and typically occurs at 3 to 4 weeks. Late, late-onset disease occurs beyond age 89 days and usually occurs in very preterm infants requiring prolonged hospitalizations. Intrapartum antibiotic prophylaxis (IAP) has dramatically reduced EOGBS but has not changed the incidence of LOGBS. Although IAP does not prevent LOGBS, studies have revealed that intrapartum antibiotics are associated with delayed presentation of symptoms and milder LOGBS. The incidence of LOGBS in the United States is approximately 0.3 to 0.4 cases per 1,000 live births, nearly equal to that of EOGBS. A European study demonstrated a lower incidence overall and revealed that premature infants have a 6-time higher incidence of LOGBS compared with term infants (1.4 vs .24 per 1,000 live births). The case-fatality ratio in term infants ranges from 1% to 3% but is higher in preterm neonates (20% for EOGBS and 5% for LOGBS). Approximately 50% of EOGBS still afflicts term neonates and because the infection frequently manifests within 24 hours, the infants are frequently in the hospital. At least 50% of infants with LOGBS are born before 37 weeks’ gestation, but many are already at home when they develop LOGBS.Anecdotal case reports have suggested that breast milk is a possible source of LOGBS, especially if GBS mastitis is present, because particularly massive GBS inoculums may lead to heavy neonatal colonization, a known risk factor for invasive GBS disease. Many times the source of GBS is unclear in LOGBS. In the discussed case, the mother should have received IAP, but the infant did not develop EOGBS. The Centers for Disease Control and Prevention guidelines were not followed. Whether the infant would have benefited from IAP, one cannot be certain. There is no explanation of why the nurses and the obstetrician did not know or follow through with finding out the GBS status and bacteriuria.In this case, the infant should have been at an advantage because he was in a NICU on a continuous monitor. Unfortunately the signals that the infant sent out when he became 33 weeks’ gestation at 31 days old (intermittent tachycardia, hypotension, gastric residuals, apnea, and desaturation) were not appreciated. The bacteremia preceded the meningitis, as it generally does (except in cases of ventricular taps where organisms might be inadvertently introduced). The infant showed signs of bacteremia. Furthermore, the 3.5-hour delay in the antibiotic administration once the infant showed obvious signs of deterioration and the choice of antibiotics were inappropriate considering the GBS status of the mother. At some point during the bacteremic phase, the CNS became seeded. During the 3.5-hour interval between the “crash” and administration of antibiotics, the GBS was allowed to rapidly replicate unchecked. As clinical symptomatology is often subtle and nonspecific in the early phase of the disease (when intervention is the most effective), it is vitally important to notice the signals that these little ones are providing, and not delay in evaluation and treatment. Delays have important consequences in GBS infection. Once an infant is clearly symptomatic, the risk of injury and death increases considerably.One approach that has been suggested to help with early diagnosis of GBS infection is to evaluate beat-to-beat HR monitoring. Although this is appealing and hopefully will add another dimension to help care providers be more proactive, simply appreciating the vital sign abnormalities such as increase in HR, changes in ambient temperatures, new onset of apnea, desaturation and bradycardia, or new residuals provide a huge amount of input to care physicians and nurses. Ultimately GBS vaccination in conjunction with IAP may help to accomplish the ultimate goal of GBS disease elimination.