Variation in Management of Cutaneous Lumbosacral Findings in Newborns

This study was conducted through the Better Outcomes through Research for Newborns (BORN) Network, which includes clinicians from 120 newborn nurseries in 35 states in the United States. Approximately 400 000 neonates are cared for at these sites annually (∼10% of 4 million live births in the United States). Approval was obtained from the institutional review board at a university-affiliated hospital in the western United States.

An electronic survey was sent to each member of the BORN Network. The electronic survey included 18 photographs of lumbosacral findings, each with a brief case description, which were selected by one author to represent the range of lumbosacral findings seen in neonates. Photographs were obtained with permission from parents or under Creative Commons license. Respondents were asked to select a management plan for each case from choices that included routine care, watchful waiting, imaging, or subspecialty consultation. Regarding the selection of imaging tests, respondents were asked which management step he or she would take at birth. Options for imaging included radiograph, ultrasound, or MRI, and options for subspecialty consultation included pediatric urology, dermatology, or neurosurgery. Respondents were given the option to select other subspecialty consults or imaging not listed and write in comments. Information collected on respondents included self-reported years of clinical experience, percent of clinical time spent taking care of neonates <1 month of age, ease of access to imaging, and subspecialty consults and demographics. Each member of the BORN Network received an anonymous electronic link and up to 6 e-mail reminders to complete the survey.

We defined expectant management to include watchful waiting and/or routine care and intervention to include imaging and/or subspecialty consultation. For each case, we categorized agreement as “good” (>90%), “modest” (70% to 90%), and “poor” (<70%) on the basis of the proportion of agreement for expectant management or intervention, whichever was more commonly selected. We categorized years in clinical practice in 5-year intervals and percent full time equivalent (FTE) in quartiles as assessed in the administered survey. We used χ² analyses to compare the proportion choosing expectant management (routine care or watchful waiting) across dichotomous variables including board certification in pediatrics, years in clinical practice, clinical FTE spent caring for neonates ≤1 month of age, location of clinical practice (level I or II nursery versus other), respondents’ self-reported race (white versus person of color), and respondents’ reported ease of access to imaging modalities or consultants. Analyses were performed by using SPSS (IBM SPSS Statistics, IBM Corporation).

Of 407 clinicians in the BORN Network, 206 (51%) clinicians completed the survey. Characteristics of study participants are shown in Table 1. Respondents’ perception of the ease of obtaining imaging (radiograph, MRI, spinal ultrasound) or subspecialty consultation (dermatology, neurosurgery, urology) is shown in Table 1. Barriers to high-quality imaging reported by respondents included lack of pediatric radiology to read imaging and lack of skilled technicians to perform the studies. For MRI specifically, respondents reported difficulty with access to MRI (especially for respondents not in a children’s hospital), concerns over needing sedation, concerns regarding insurance coverage, and concerns around a complex scheduling process. Respondents’ perception regarding barriers to subspecialty consultations included limited or no availability of pediatric subspecialists in their institution. The majority of respondents (n = 180; 87.4%) reported their management decisions were not influenced by hospital policies, departmental or subspecialty consultation expectations, or local practice habits.

The percent agreement among survey respondents to the initial management of neonates with lumbosacral findings is shown in Fig 1. Respondents were in >90% agreement in the initial management approach of 8 of 18 cases. Of respondents who chose imaging and/or subspecialty consultation for any of the 18 cases, 53% chose to obtain a spinal ultrasound as the first initial step, 18% chose neurosurgery evaluation, and 13% chose MRI. Anomalies of the gluteal crease, flat vascular macules, and a small concentration of coccygeal hair had the lowest proportion of agreement among clinicians regarding management. For responses “watchful waiting” or “routine care,” respondents were asked to write in what circumstance(s) would prompt a different management step. These responses included (1) change in physical examination (increasing size of lesion, abnormal hip examination, drainage) or neurologic examination over time, (2) persistence of lumbosacral finding with no resolution, (3) asymmetry in extremities or muscle bulk, (4) changes with development, (5) urinary or bowel changes, or (6) family history of motor or musculoskeletal problems. In these listed circumstances, respondents stated they would consider imaging or subspecialty consultation.

In the χ² analysis, we compared the proportion of respondents choosing expectant management (routine care or watchful waiting) across dichotomous variables. We found statistically significant differences in the proportion of respondents choosing observation compared to intervention in 7 lumbosacral findings (P < .05). Specifically, for the deviated gluteal crease (case A) and faun tail nevus cases, a larger proportion of respondents who selected pediatrics as a specialty chose an intervention (P = .019 and P = .005, respectively). In the cases of hemangioma and coccygeal tag >5 mm + deviated crease, a larger proportion of respondents who identified as white selected an intervention (P = .04 and P = .02, respectively). For the small duplicated gluteal crease, respondents with increasing years in clinical practice chose an intervention (P = .04). In 2 cases, those of coccygeal hair and deviated gluteal crease (case A), there was a difference in respondents selecting intervention by clinical FTE with neonates <1 month of age (P = .02 for both cases).

Among 18 cases, there was an association between respondents’ report of “easy” access to and selection of ultrasound in 1 case (faun tail nevus P = .05). Cases for which there was an association between respondents’ report of easy access to and selection of MRI for initial management were the deep dimple (P = .05), lipoma (P = .05), faun tail nevus (P = .03), and coccygeal skin tag <5 mm (P = .03). Regarding consultations, cases with an association between respondents’ report of easy access to neurosurgery consultation and selection of neurosurgery consultation as an initial management plan included coccygeal tag >5 mm (P = .004), interrupted duplicated gluteal crease (P = .05), and coccygeal tag >5 mm + deviated crease (P = .03). There were no associations between selecting radiographs, dermatology consultations, or urology consultations and reported easy access to those specific imaging modalities or consultations.

Our results demonstrate a lack of consistency in clinical management of common neonatal lumbosacral findings. Of the 18 cases presented in our survey, there was good agreement regarding management for only 8 of 18 cases; 6 (33%) of the 18 cases had poor agreement (<70% agreement on management approach). These results are consistent with a recent international survey on the management of skin stigmata among pediatric neurosurgeons, which also revealed a lack of consensus across management of simple dimple, deviated gluteal fold, and discoloration.⁷  Together, these findings highlight the need for more robust evidence to guide clinical decision-making in these situations.

Our study identified areas both of consensus and of variation in management. Most respondents chose spinal ultrasound (53%) as the initial step in managing a concerning lumbosacral finding. Of note, researchers have shown that the sensitivity and specificity of ultrasound for diagnosis of OSD are high at 86.5% and 92.0%, respectively, when performed by an experienced operator.⁸  MRI, the third most frequent management step chosen in this study (13%), is considered the definitive imaging modality for spinal cord anomalies in the literature, with a sensitivity of 95.6% and specificity of 90.9%.⁸  Although some respondents reported concern about the need for sedation with MRI, it is important to note that sedation may not be required for MRI in neonates. However, given the wide range of OSD prevalence associated with various lumbosacral findings, controversy remains regarding which lumbosacral findings warrant either further investigation with ultrasound or MRI or any evaluation at all.

For the small portion of respondents whose management was guided by a specific text, guideline, or policy, the most commonly cited references were a clinical report by Dias and Partington⁵  and a review article by Zywicke and Rozzelle.⁶  The former stratified various cutaneous anomalies into low-, intermediate-, and high-risk categories but made no recommendations on management of each risk group.⁵  The latter proposed an algorithm whereby only single sacrococcygeal dimples <5 mm with no other associated cutaneous findings are observed and all other findings, including gluteal cleft anomalies, warrant imaging and/or neurosurgical referral.⁶ 

In our study, there is good consensus on the need for imaging or consult for lipoma, coccygeal tag >5 mm, faun tail nevus, and hemangioma, suggesting that most clinicians consider these high-risk lesions. On the other hand, the large majority agreed that simple dimple, lumbosacral hair, and small duplicated gluteal crease are benign with no need for imaging or consult. This is consistent with a study revealing the incidence of spinal abnormalities detected on spinal ultrasound in 5166 patients with sacral dimples (3.4%) to be nearly identical to the incidence of abnormal findings in children without sacral dimples (4.5%).⁹  There is controversy in the literature regarding whether the depth of dimple matters, as reflected in our data in which respondents only had moderate agreement on management of deep dimples.^5,6,9,10 

Anomalies of the gluteal crease had the lowest proportion of agreement among clinicians regarding management. Interestingly, anomalies of the gluteal crease are the most commonly encountered findings in well neonates, with 1 study reporting an incidence of 24.8%.¹  Given the low incidence rate of OSD at 0.04%, they are likely too common to be considered high risk.^1,4  However, some believe all gluteal cleft anomalies other than dimples warrant further evaluation.⁶  This disparity in opinion is well reflected in our study; 4 of 5 cases of gluteal cleft abnormalities garnered only poor to modest agreement.

This study has several important limitations. Our survey was completed by 51% of the BORN membership. Practice patterns of survey respondents might differ from those of nonrespondents, and the responses of BORN members might differ from those of clinicians who are not BORN members. Respondents were not asked to provide the reasoning behind their management selections. However, we did collect data on barriers to imaging or consults. Survey respondents relied on photographs and case vignettes only and were not able to do a physical examination of the neonate’s lumbosacral region. The size, resolution, or color of the image on the respondents’ screens could have altered the appearance of the finding and may have affected their initial management choice. The survey was not a previously validated tool and the definition of good, modest, and poor agreement was the result of discussion and consensus among the study investigators given that agreement could not be <50%. Respondents were not asked if they would perform imaging tests in sequence, such as starting with ultrasound and proceeding to MRI if an abnormality was detected. Our analysis is limited to evaluating the self-reported opinions of respondents without obtaining their actual behavior from another source and without measuring an effect of decisions on outcome. Therefore, we cannot draw direct conclusions on best practices in our study. Although our findings suggest that there might be some differences in the management of lumbosacral findings by specialty, years of clinical experience, clinical time spent in the care of neonates, and race of the provider, our numbers were small, and these findings warrant further study. Nonetheless, our study provides a meaningful snapshot of the current management preferences for lumbosacral lesions among clinicians who routinely take care of neonates.

Our results reveal a wide variation in management of common neonatal lumbosacral findings such as moderate deviation and duplication of the gluteal crease, vascular macules with and without dermal melanosis, and coccygeal hairiness. This variation in management is consistent with the lack of evidence regarding the positive and negative predictive values of these findings for the diagnosis of OSD and underscores the need for further scientific investigation in this area. A long-term prospective study of healthy neonatal populations could provide guidance to clinicians regarding the preferred course of action and thus potentially lead to more informed clinical decision-making in this area.

Administrative support was provided by Beth A. King of the BORN Network.