Symptoms of Cerebrospinal Shunt Malfunction in Young Children: A National Caregiver Survey

Abstract

Objective: This study aimed to describe shunt malfunction symptoms in children ≤5 years old. Results: In a national survey of 228 caregivers, vomiting (23.1%), irritability (20.8%), and sleepiness (17.2%) were the most frequent symptoms of malfunction. These symptoms also occurred in over 1/3 of “false alarms” experienced by 75% of respondents. Compared with malfunctions, irritability (OR = 1.39, 95% CI [1.05, 1.85], p = 0.022) and fever (OR = 2.22, 95% CI [1.44, 3.44], p < 0.001) were more likely false alarms. Caregivers counseled about “most” symptoms were more confident detecting malfunctions than those informed of “some” (p = 0.036). The majority of caregivers (85%) first contacted a neurosurgeon with concerns about malfunction, followed by neurologists (22%) and family/friends (19%). Most (85%) struggled to differentiate malfunction from regular development. Conclusions: Vomiting, irritability, and sleepiness were the most common symptoms of shunt malfunction and false alarms for children ≤5 years. Most caregivers reported challenges differentiating malfunctions from their child's development.

Introduction

Over the past two decades the number of pediatric cerebrospinal shunts placed annually has nearly doubled.^1,2 The increasing utilization of pediatric shunts correlates with the growing number of children surviving prematurity and other congenital conditions. A substantial number of those children will experience a shunt malfunction in infancy or early childhood; cerebrospinal shunts in children have failure rates of 30–40% at one year and 50% at two years.^3,4 Patients under 6 months old⁵ and those with intraventricular hemorrhage⁶ are at even higher risk. Detecting shunt malfunction in a timely manner is essential because untreated shunt failure can result in brain injury and mortality rates of 2% or higher^7–9 in a population of children already at high risk for neurodevelopmental impairment and comorbidities.

Infants and young children cannot reliably self-report symptoms of shunt malfunctions, leaving home caregivers with full responsibility for detecting possible shunt failure. Additionally, early childhood is characterized by constant developmental changes, evolving baseline behaviors related to sleep, eating, and activity, and frequent minor infections, illnesses and injuries. Despite the unique challenges of detecting shunt malfunction in this context, and despite the substantial number of shunt malfunctions that occur during early childhood, data focused on this population are limited to studies performed two decades ago. Those decades have coincided with expanding surgical approaches to pediatric hydrocephalus and a greater diversity of underlying medical complexity among infants and young children receiving shunts.^10–12 It is important to assess whether the presentation of shunt malfunction in infants and young children is changing in this era.

In 2021, we published results from a small national cohort of caregivers whose child was previously diagnosed with neonatal hydrocephalus. Multiple caregivers reported that their child's symptoms of shunt malfunction differed from what clinicians had told them to expect; this finding was most pronounced among caregivers whose child was an infant or young child.¹³ Any potential mismatch between clinician counseling and a child's presentation may reflect the reality that few studies have focused solely on malfunction signs and symptoms in infants and young children. The Pediatric Shunt Design Trial is among the most contemporary and complete descriptions of shunt malfunction in early childhood.¹⁴ In this 2001 cohort of 431 young patients, factors like decreased consciousness and wound erythema occurred rarely but were always indicative of shunt failure. More common factors like irritability, nausea, vomiting, and bulging fontanelle were predictive of shunt malfunction soon after shunt insertion (median time 4.6 months) but became less predictive 9 or more months after insertion.

Most studies of signs and symptoms of shunt malfunction include children from infancy to young adulthood, making it difficult to tailor counseling for families of the youngest patients. Importantly, there is a lack of detailed data from families, even though families are the first line adjudicators of whether a child is experiencing a sign or symptom of concern.

Here we aimed to characterize the signs and symptoms of shunt malfunction in children ≤5 years old and the caregiver experience of identifying and seeking care for shunt malfunction. We report results of a national survey of caregivers of children ≤5 years old with shunts. This contemporary population of infants and young children contributes important information about the initial signs and symptoms of shunt failure, to reinforce clinician counseling and family education. Family education is a key aspect of improving child outcomes following cerebrospinal shunt placement, as improved family education is associated with reduced mortality related to shunt malfunction.^8,9,15

Methods

English-speaking caregivers of children ≤5 years of age were recruited in collaboration with the Hydrocephalus Association. An anonymous cross-sectional survey was offered over a predetermined 6 week timeline from 9/23/21-11/1/21 over various Hydrocephalus Association platforms including the website, via email listservs, and public social media platforms (Facebook, Instagram, Twitter). The survey was created by content experts in pediatric neurology, neonatology, and pediatric palliative care. The survey was pilot tested by board members from the Hydrocephalus Association, including a parent member, and then edited accordingly based on feedback before finalizing. Survey data was collected and managed using REDCap electronic data capture tools^16,17 hosted by the Children's Hospital of the King's Daughters.

Survey questions included a comprehensive list of shunt malfunction signs or symptoms that was created to be inclusive of the symptoms noted by Garton¹⁴ and Ackerman¹⁵; caregivers were asked to indicate which of the signs or symptoms their child exhibited during their first, and when applicable, subsequent shunt malfunctions. In addition, caregivers were asked targeted questions regarding sources of education about shunt malfunctions and what actions they took when they suspected a shunt malfunction (See supplemental material).

All statistical tests were performed using SPSS.26 (Chicago, IL).¹⁸ Continuous variables are presented as mean, standard deviation, median, 25^th and 75^th percentiles. Categorical variables are presented as frequency and percentage. The Kruskal-Wallis test was used to compare continuous variables. Chi-square test was used to compare categorical variables. All statistical tests were two-sided and p < 0.05 was considered statistically significant.

The Eastern Virginia Medical School (IRB #21-07-EX-0175) approved the study; caregivers gave consent on REDCap before beginning the survey.

Results

Participant Characteristics

A total of 240 caregivers responded to the survey; 12 records were incomplete or truncated due to exclusion criteria (eg current age of child). 228 complete caregiver responses were analyzed, representing 228 children. Table 1 describes the characteristics of caregivers: most were white females, were the mother of the child with the shunt, were employed full time and had at least some college education.

Table 1.

Caregiver Characteristics.

Caregiver Characteristics (n = 228)	Median (IQR) or N (%)
Median age at survey(years)	34 (30 − 38)
Relationship to child
Parent	223 (98.2)
Grandparent	2 (0.9)
Other	2 (0.9)
Gender
Female	203 (89.8)
Male	23 (10.2)
Other	0 (0.0)
Current Employment
Stay at home caregiver	86 (35.1)
Employed part time	33 (13.5)
Employed full time	111 (45.3)
Part time student	5 (2.0)
Full time student	5 (2.0)
Other	5 (2.0)
Highest Level of Education
Doctorate	8 (3.5)
Masters	46 (20.3)
College	143 (63.0)
High School or less	30 (13.2)
Self-identified Race
Caucasian	173 (70.6)
African American/Black	17 (6.9)
Hispanic	18 (7.3)
Asian	9 (3.7)
American Indian/Alaska Native	1 (0.4)
Other	16 (6.5)

Table 2 describes the characteristics of the children. Most received their shunt when they were <6 months old, and almost half had a confirmed shunt malfunction before 6 months of age.

Table 2.

Child Characteristics.

		Total (n = 228)
		N (%)
Age	< 1 year	20 (8.8)
	1 year	25 (11.0)
	2 years	48 (21.0)
	3 years	36 (15.8)
	4 years	52 (22.8)
	5 years	47 (20.6)
Age when first shunt placed	<6 months old	199 (87.3)
	6 months – 1 year	21 (9.2)
	>1 year old	8 (3.5)
Age at first shunt malfunction	< 6 months	101 (45.9)
	6 months- 1 year	51 (23.2)
	>1 year	68 (30.9)
Caregiver report of reason for first shunt malfunction	Piece Broke	25 (11.0)
	Infection	29 (12.7)
	Clog	111 (48.7)
	I don't know	19 (8.3)
	Other	54 (23.7)
Episodes of possible shunt malfunction in the last 5 years	0	10 (0.04)
	1–2	69 (30.3)
	3–5	76 (33.3)
	6 − 10	35 (15.4)
	>10	38 (16.7)
Number of confirmed shunt malfunctions in last 5 years	0	29 (12.7)
	1 − 2	135 (59.2)
	3 − 5	42 (18.4)
	6–10	17 (7.5)
	>10	5 (2.2)

Presenting for Shunt Evaluation

When worried that their child might have signs or symptoms consistent with a shunt malfunction, just under half (45%) of caregivers first contacted their neurosurgeon's on-call line. They were as likely to call family or friends for advice as to contact their pediatrician (18% vs 20%, respectively). Over half of families (55%) took their child to the emergency department for evaluation, while 27% went to their neurosurgery clinic.

Table 3 describes the frequency of caregiver-reported signs and/or symptoms during any shunt malfunction up through five years of age. The signs and/or symptoms of vomiting (23.1%), irritability (20.8%), and sleepiness (17.2%) were most common. Most children presented with ≥ 2 signs/symptoms, but 15% of shunt malfunctions reportedly presented with a single sign or symptom. A “full fontanelle” and leaking shunt were the most commonly reported isolated signs/symptoms.

Table 3.

Symptoms of Shunt Malfunction in Children ≤5 Years Old.

Symptom	Any malfunction (N = 221*) n (%)	Presented as isolated symptom (N = 221) n (%)
Vomiting	51 (23.1)	3 (1.4)
Irritable	46 (20.8)	1 (0.5)
Sleeping more than usual	38 (17.2)	2 (0.9)
Poor eating/drinking	28 (12.7)	2 (0.9)
Full soft spot(fontanelle)	25 (11.3)	5 (2.3)
Decreased consciousness	22 (10)	1 (0.5)
Nausea	22 (10)	0 (0)
Change in eye movements	22 (10)	1 (0.5)
Head swelling	18 (8.1)	4 (1.8)
Headache	15 (6.8)	1 (0.5)
Pain	14 (6.3)	0 (0)
Other	8 (3.6)	6 (2.7)
Dizziness	7 (3.2)	0 (0)
Seizures	7 (3.2)	2 (0.9)
Fever	6 (2.7)	0 (0)
Stiff neck	4 (1.8)	0 (0)
Redness around shunt	4 (1.8)	0 (0)
Change in walking	3 (1.4)	0 (0)
Leaking shunt	2 (0.9)	5 (2.3)
New stuttering	1 (0.5)	0 (0)
Change in talking	1 (0.5)	0 (0)
New bedwetting	0 (0)	0 (0)

* = 7 respondents did not check any symptoms.

For those caregivers whose children had at least 2 shunt malfunctions, we asked them to report their child's symptoms in the first versus later malfunctions (Table 4). The signs and/or symptoms of vomiting (47.4%), irritability (47.4%), sleepiness (45.2%), poor eating and/or drinking (32.5%), and full fontanelle (30.3%) were reported in at least one third of initial shunt malfunctions. Those same signs and/or symptoms, in addition to nausea (34.3%), decreased consciousness (33.0%), and change in eye movements (32.4%) were reported for at least one third of subsequent malfunctions. Most (86%) caregivers reported that their child's first and subsequent malfunctions had at least one sign or symptom in common, with vomiting (47.2%), irritability (42.6%), and sleepiness (35.2%) most likely to occur during >1 malfunction.

Table 4.

Frequency of Symptoms in First Malfunction, Later Malfunctions, and “False Alarms”.

Symptom	First malfunction (N = 108) n (%)	Later malfunctions (N = 108) n (%)	False Alarms (N = 108) n (%)
Vomiting	51 (47.4)	69 (63.9)	83 (45.6)
Irritability	51 (47.4)	66 (61.1)	84 (46.5)
Sleeping more than usual	49 (45.2)	54 (50.0)	68 (37.7)
Poor eating/drinking	35 (32.5)	44 (40.7)	48 (26.8)
Full soft spot(fontanelle)	33 (30.3)	33 (30.6)	11 (6.1)
Decreased consciousness	28 (25.9)	36 (33.0)	20 (11.0)
Nausea	26 (23.7)	37 (34.3)	43 (24.1)
Change in eye movements	25 (22.8)	35 (32.4)	10 (11.0)
Head swelling	23 (21.1)	26 (24.1)	8 (4.4)
Headache	17 (15.4)	36 (33.3)	36 (20.2)
Other	17 (15.4)	17 (15.7)	8 (4.4)
Sleeping less than usual	6 (5.7)	14 (13.0)	13 (7.0)
Pain	13 (11.8)	25 (22.2)	17 (9.2)
Fever	10 (8.8)	22 (20.4)	43 (24.1)
Seizures	8 (7.5)	17 (15.7)	17 (9.2)
Leaking shunt	8 (7.5)	7 (6.5)	2 (0.9)
Change in walking	7 (6.6)	9 (8.3)	10 (5.7)
Redness around shunt	7 (6.6)	12 (11.1)	4 (2.2)
Dizziness	5 (4.8)	10 (9.3)	8 (4.4)
Change in talking	4 (3.5)	9 (8.3)	6 (3.5)
New stuttering	2 (1.9)	2 (1.9)	3 (1.8)
New bedwetting	1 (0.4)	3 (2.8)	4 (2.2)
Stiff neck	4 (3.5)	8 (7.4)	5 (2.6)

A minority (15%) of caregivers reported that their child had “other” symptoms at the time of their shunt malfunction. The most common “other’ symptoms were behavior change (ie “not acting himself”), altered skin color, low heart rate, a distended scalp vein, and chronic symptoms that ebbed and flowed for weeks to months. For 5 caregivers, the “other” symptoms described overlap with those on the list (eg “crying” vs “irritability,” “leaking incision sites” vs “leaking shunt”).

The majority (75%) of caregivers had experienced at least one “false alarm,” worrying their child had a shunt malfunction when there was not one. Table 4 also shows the frequency of symptoms during “false alarms.” The same three symptoms that were most common during confirmed shunt malfunctions were also most common in false alarms: vomiting (45.6%), irritability (46.5%), and sleepiness (37.7%). Compared with actual shunt malfunctions, irritability (OR = 1.39, 95% CI [1.05, 1.85], p = 0.022) and fever (OR = 2.22, 95% CI [1.44, 3.44], p < 0.001) were more likely to be a false alarm. In this survey there were no caregiver-reported symptoms that were statistically more likely to be associated with true malfunction as opposed to a “false alarm.”

Parent Preparation to Detect Shunt Malfunction

Most caregivers reported that they received counseling and education about how to detect a shunt malfunction in multiple locations within their health system: a neonatal or pediatric intensive care unit (ICU) (68%), an outpatient clinic (65%), a general pediatric ward (46%), and an emergency department (42%). The vast majority (95%) reported this counseling/education from a neurosurgeon. Under half learned about shunt malfunction from an ICU clinician (44%), pediatrician (30%), or neurologist (22%). Notably, one third of respondents reported learning from other parents, and over two-thirds reported finding their own resources for education about shunt malfunctions.

When asked, “How many of your child's symptoms had health care providers told you to expect?” slightly over half (55%) of caregivers said “most” while a minority responded “few” (12%) or “none” (8%). Caregivers who were counseled about “most” of the signs and symptoms were significantly more confident in their ability to detect a shunt malfunction than were those who were informed about “some” symptoms (p = 0.036).

Approximately half of caregivers reported that the difficulty of detecting shunt malformations remained the same over time (52%), while under one-third (30%) reported that their child's shunt malfunctions became easier to detect over time (Table 5). Notably, detecting shunt malfunctions became harder for a number of caregivers once their child's anterior fontanelle closed.

Table 5.

Why Detecting Shunt Malfunctions Became Easier Versus Harder*.

Easier
We recognize the signs	He had the same symptoms each time. Vomiting became his number one symptom. We knew that he's not textbook-- he never really projectile vomited, never had sundowning eyes. But we knew seizures--breakthrough seizures were a first sign of something wrong.
Child getting older	Older and can verbalize. It's easier to know when my child is not being his normal self.
Trusted my gut	Because I didn't hesitate and just assumed it was shunt rather than second guessing like the first time.
Spoke up faster	Each time I called the on-call line for our neurosurgeon's office, and each time we had a visit with the neurosurgeon and/or his amazing NP, I learned what questions to ask myself and which answers lead to which actions to take.
Harder
Fontanelle closed	With the fontanelle closed, that's one less sure sign that the shunt isn't working. Babies cannot easily express their discomfort, so as a parent it's all guesswork until they can speak. Once his soft spot closed it's harder to know if it's a shunt malfunction or regular illness. It feels like it could be harder now that his fontanelle has closed. That was the main red flag, there were no other obvious signs there was a malfunction.
Confused with developmental changes	All the symptoms of shunt failure mimic the symptoms of all sorts of very typical issues that can arise in toddlers (teething, sleep regressions). When your toddler is nonverbal, it all becomes a guessing game, which is terrifying.
Symptoms change	Symptoms changed as he got older.
Lack of symptoms	Because she showed little or no symptoms, I still didn't know what to look out for. Because of her lack of symptoms, I’m constantly worrying and have made trips to hospital just to reassure me that everything is ok.
Paranoid	I’m nervous every time she's not well. I immediately think shunt even when though it's not. It has me on edge constantly.

 * = selected caregiver quotes

The majority (85%) of caregivers reported it was either “very” or “somewhat” difficult to differentiate symptoms of shunt malfunction from their child's regular development.

Discussion

Vomiting, irritability, and sleeping more than usual were the most common symptoms of initial and subsequent shunt malfunction for children ≤ 5 years old in our cohort. These same signs were also the most prevalent in subsequent malfunctions and were the most common triggers of evaluations that revealed no shunt malfunction (“false alarms”). In the most recent comparable cohort of young children with shunts, published by Garton et al in 2001,¹⁴ increased head circumference (18%), bulging fontanelle (14.6%), irritability (13%) and nausea/vomiting (12%) were noted as the most common signs of shunt failure. Data from older children suggest that shunt malfunction after early childhood is more likely to present with lethargy or shunt site swelling.⁷ It is notable in our cohort that 15% of shunt malfunctions reportedly presented with “other” symptoms; “not acting him/herself” was the most common “other” sign. In addition, 14% of shunt malfunctions among infants and young children in our cohort were reported to present with a single sign or symptom; no note is made in the Garton study of how often symptoms were clustered versus isolated for individual patients.

Parent-reported signs and symptoms in this cohort of young children are different from those reported by Garton et al and are different than those reported for older children. Notably, the most common signs or symptoms completely overlapped with the most common triggers of false alarms. Vomiting, irritability, and sleeping changes are common in all infants and young children, including those without shunts, related to typical developmental changes and illnesses of early childhood. That non-specific symptoms are unreliable in distinguishing pediatric shunt malfunction from false alarms was recently highlighted by Razmara et al¹⁹ and our data suggest that this challenge may be particularly problematic for the youngest patients.

Taken together, these data underscore that symptom tracking alone is not adequate to detect shunt malfunction in infants and young children. Expanded access to objective measures of intracranial pressure and potential malfunctions is critical. In addition to decreasing radiation exposure, MRI is superior to CT at detecting shunt malfunction.^20,21 Recent data demonstrates that despite a nearly 20% increase in emergency department visits between 2006–2017 among children with shunts, overall fewer children received surgical interventions because of access to MRIs.²² Caregivers can be encouraged to bring their children to attention with concerning symptoms, as evaluation will result in neither excess radiation exposure nor unnecessary interventions. Additionally, rapid MRI is increasingly available²³ though associated with a longer time to imaging (53 min) and a longer emergency department length of stay (52 min).²⁴ Machine learning also shows potential to combine clinical, radiologic, surgical, and shunt-design factors to improve detection of shunt malfunction.²⁵

Continuing to evolve objective tools is essential, but we also must help “first responder” caregivers prepare for their essential role in bringing children to medical attention for neuroimaging and for detailed medical exams. In a recent large study of >9000 children of all ages who received a shunt revision, peritonitis, papilledema and oculomotor palsies were the most accurate indicators of shunt malfunction.¹⁹ The utility of these signs is somewhat limited by the fact that they can only be detected by medical professionals, once a caregiver has raised a concern.

Our national sample confirms that neurosurgeons are doing the bulk of family counseling and preparation for shunt malfunctions. The caregivers that received the most counseling felt the most prepared. However, the majority (85%) of this well-educated cohort still found it difficult to distinguish young childhood development or minor illnesses from a shunt malfunction. A minority felt that detecting malfunctions got easier over time; many felt it got more difficult as the fontanelle closed. Other studies conflict about whether parent accuracy in detecting shunt malfunctions increases²⁶ or decreases²⁷ as parents gain more experience. Not quite half of the caregivers in our study contacted their neurosurgeon on-call line when worried, the remainder were as likely to call a family or friend as to call their neurologist. This suggests ongoing opportunities to strengthen family preparation for when to worry about shunt malfunction, and what to do when they are worried. Neonatal neurocritical care programs, which focus on the provision of brain-focused care and partnership with neurosurgical teams, may be uniquely suited to meet this need if infants receive their first shunt in the hospital.^28–30

Approximately half of caregivers reported that their infant or young child had signs and/or symptoms that they had not been counseled about previously. We cannot confirm whether these caregivers did receive this information; however, information delivery was nonetheless not salient enough to meaningfully guide caregiver decision-making. These results suggest an opportunity for educational methods to complement clinician counseling. The Agency for Healthcare Research and Quality (AHRQ) advocates the teach-back method to bridge differing levels of health literacy³¹; with this method, the clinician delivers information to a caregiver and asks them to “teach back” what they heard, using their own words. This strategy might overcome the challenge of caregivers not understanding that their child's sign and/or symptom (eg “crying”) is the same one their clinician counseled them about (eg “irritability”).

Caregivers frequently interacted with other healthcare providers (ICU clinicians, neurologists, pediatricians), but report they rarely received counseling about shunt malfunctions from these providers. This may represent a missed opportunity for enhanced involvement of neurologists, who often see patients with hydrocephalus alongside neurosurgeons. It may be that other clinician types, like primary care clinicians, would benefit from continuing medical education about cerebrospinal shunts and the signs and/or symptoms of malfunction, so that they can reinforce caregiver counseling. Many medical certification boards (eg American Academy of Pediatrics³²), for example, are now requiring online learning; a module about shunts/shunt malfunctions delivered in this setting could reach a national group of physicians. Caregivers frequent multiple sites within healthcare systems, from inpatient units to outpatient clinics to emergency departments. Distributing educational materials in these locations may be useful for both clinicians and caregivers, eg via quick response (QR) codes linked to computerized content, a strategy with growing evidence for effectiveness in health education.³³

There are also opportunities to leverage the finding that caregivers commonly reached out to other families (33%) and non-clinical resources (69%) to learn about shunt malfunction. Multiple studies confirm that families of children with medical complexity connect with each other and educate each other via social media. When this peer mentorship occurs via non-monitored platforms (eg Facebook^34,35), it is difficult to assess information quality. Caregivers should be provided with options for high-quality resources. The Hydrocephalus Association is a non-profit organization with several family supports, including information about symptoms of shunt malfunction,³⁶ a helpline caregivers can call for resources, and a peer support network³⁷ that pairs parents with trained parent volunteers to share experiences and supports relevant to pediatric hydrocephalus. High-quality information sources are especially important for caregivers whose infants or young children are having their first malfunctions. Figure 1 summarizes these opportunities to reinforce education about shunt malfunction for caregivers and clinicians.

Figure 1.

Education Opportunities for Shunt Malfunction “First Responders”: Caregivers and Healthcare Providers.

We recognize several limitations to this study. The data presented here about the signs and symptoms of shunt malfunction come from caregiver report and can certainly have limitations, although Watkins et al noted that caregivers commonly have greater accuracy in this context than practitioners or hospitals.³⁸ The children were all ≤5 years old but at different ages at the time of the study, so not all contribute 5 years of data. The finding that the signs and/or symptoms of initial shunt malfunctions differed somewhat from subsequent ones suggests room for studies with the power to isolate the signs and/or symptoms of these distinct groups. The caregivers were mostly well-educated, white females. Potential disparities in shunt care,⁴ as in other areas of medicine, demand broader research with heterogeneous populations.

Conclusions

Vomiting, irritability, and/or sleeping more than usual were the presenting symptoms of shunt malfunction for children ≤5 years old in this national cohort. These same symptoms were just as likely to be false alarms. Despite advances in objective measures of shunt malfunction, caregivers at home remain the “first responders” for determining when there is a concern. Nearly all caregivers struggled to distinguish young childhood development or minor illnesses from a shunt malfunction and only a minority felt that detecting malfunctions got easier over time. Future interventions should leverage these opportunities to reinforce education of families of young children with shunts.