Blood pressure in children with attention deficit/hyperactivity disorder

Abstract

Objectives

Children with attention deficit/hyperactivity disorder (ADHD) are frequently treated with psycho-stimulant agents causing a modest but significant increase in blood pressure and heart rate. The objective of this study was to define blood pressure characteristics in children with ADHD treated with a variety of medications in a community setup.

Methods

Children registered at a large paediatric clinic in Calgary, AB with documented histories of ADHD were randomly contacted. Consenting participants had standardized office BP measurements, ambulatory blood pressure monitoring (ABPM) studies and were asked to complete the sleep disturbance scale for children (SDSC) questionnaire. Findings were compared with data from the Canadian Health Measures Survey (CMHS).

Results

Fifty-five children (47 males) aged 7 to 17 years (average 11.6 ± 2.5 years) with an average BMI z-score of −0.37 ± 1.22 completed the study. All children were medicated, the majority (82%), with various types of stimulant agents. Elevated office BP values were more prevalent than in the CMHS; >90th percentile in 5 (9.1%) and >95th percentile in 3 (5.5%). ABPM confirmed ‘white coat hypertension’ in 3 (5.5%), masked hypertension in 2 (3.6%) and nondipping in 28 (51%). The SDSC score suggested that 43 (78%) children had disturbed sleep. Logistic regression modelling indicated that nondipping correlated with disturbed sleep.

Conclusion

The ‘white coat’ phenomenon may be responsible for increased prevalence of elevated rest/office BP values in children with ADHD. Prevalent sleep ‘non-dipping’ in this population is associated with sleep disturbances but clinical significance of this finding requires further investigation.

BACKGROUND

Attention deficit/hyperactivity disorder (ADHD) is a common neurobehavioural childhood disorder that can continue into adulthood (1,2). A national population-based sample of US children found that 8.7% of children aged 8 to 15 years, an estimated 2.4 million, meet DSM-IV criteria for ADHD (1). Management of this disorder is based on pharmacotherapy, most frequently with a member of the psycho-stimulant medications family (3). The incidence of ADHD and use of medications for treatment, particularly psycho-stimulant agents, has steadily increased over the last few decades (4,5). Efficacy of stimulant medications in alleviating symptoms of ADHD has been well established since 1970s and over the last 40 years, many different drugs belonging to this group have been developed, the latest addition being long-acting preparations (3).

Despite their proven efficacy, stimulant medications have always been under intense scrutiny for potentially serious side effects such as insomnia, growth retardation, exacerbation of pre-existing psychiatric disorders, adverse cardiovascular effects and abusive use (6). Concerns regarding an associated increased risk for sudden death have not been substantiated by large observational studies (7). A modest but significant increase in blood pressure and heart rate are well known adverse effect of stimulants that have been documented in many retrospective and prospective studies focusing on safety of specific agents (8–18). Nevertheless, population-based prospective studies investigating the overall effect of drug therapy on blood pressure in children with ADHD are rare. The objective of this study was to define blood pressure characteristics in a sample of otherwise healthy children treated with a variety of medications indicated for ADHD and to compare prevalence of suspected hypertension in this cohort with the prevalence documented in the general Canadian paediatric population.

METHODS

We conducted an observational comparative study, with prospective measurement of BP in a sample of children with ADHD, currently undergoing pharmaceutical treatment, at a large community general paediatric clinic in Calgary, AB. Following an official request for collaboration, three consenting independent paediatricians practicing at the clinic, provided lists of all patients under their care with a documented diagnosis of ADHD. Patient names and contact information were compiled in a single list in random order generated by a computer algorithm. Candidates were contacted by phone in the order they appeared on the randomly-compiled list for preliminary consent as well as screening for eligibility to participate. Children aged 5 to 18 years with a documented diagnosis of ADHD and receiving ongoing treatment with any type of medication indicated for ADHD were eligible for inclusion. Interested subjects were scheduled for a clinic appointment during which a full medical history, physical examination and anthropometric measurements were performed, followed by initiation of a 24-hour ambulatory blood pressure monitoring (ABPM) test.

Children with pre-existing medical histories or physical exam findings suggestive of other chronic diseases (particularly chronic kidney disease, heart disease or diabetes mellitus) or taking any other prescribed medications were excluded from the analysis.

Office blood pressure was measured with the Carescape V100 (GE Healthcare) monitor three times, 2–5 minutes apart; an average of the last two measurement was recorded. Office blood pressure values percentiles were calculated and classified according to the fourth report on the diagnosis, evaluation and treatment of high blood pressure in children and adolescents (19). The ABPM was performed with the Welch Allyn ABPM 6100 monitor, with set parameters according to reference values provided by Wuhl et al. (20). The ABPM blood pressure index was calculated by dividing the average blood pressure for each patient by the 95th percentile blood pressure value specific for that patient, such that an index greater than 1.0 is equivalent with an average BP greater than the 95th percentile (21). The operative protocol for ABPM administration and interpretation of results were conducted according to Urbina et al. updated by Flynn et al. (22,23). BMI and BMI z-score were calculated based on height and weight recorded at the time of ABPM initiation. Classification of BMI was made according to the CDC method (http://www.cdc.gov/obesity/childhood/).

Given the risk for insomnia associated with stimulant treatment and the possible effect of insomnia on nocturnal blood pressure values, the sleep disturbance scale for children (SDSC) questionnaire was also administered and filled out by parents during the visit. The SDSC is a 26-item instrument for evaluating sleep among children aged 3 to 18 years. It is validated at a cut-off score of 39 to predict sleep disturbances in children, with 0.89 sensitivity and 0.74 specificity (24). Higher SDSC scores indicate increasing severity, with a cut-off score of 50 or higher marking ‘severe’ sleep disturbance.

To compare prevalence of suspected hypertension in children with ADHD with prevalence in the general paediatric Canadian population, we used office blood pressure values documented in our sample and available data from the Canadian Health Measures Survey (CHMS). The CHMS is a national, representative, cross-sectional, direct-measures survey of more than 15,000 Canadians aged 6 to 79 years who underwent anthropometric as well as blood pressure and heart rate measurements during single visits to mobile examinations centres across Canada. In the CHMS, BP was also measured with a calibrated oscilometric electronic device; however there were six measurements 1 minute apart with the average of the last five being documented. Data from two cycles (2007 to 2009 and 2012 to 2013) were used for comparison (25,26).

Descriptive statistics included calculation of means and proportions. Chi-square analysis was used to compare prevalence of blood pressure values in the prehypertension or hypertension range among children medicated for ADHD, with children from the CHMS survey representing the general Canadian paediatric population. Children were also grouped according to the presence or absence of ABPM abnormalities (mean ambulatory pressure > 95th percentile and/or sleep nondipping) and were categorized as having sleep disturbance, based on the SDSC score. A reverse stepwise-elimination logistic regression modelling process was used to assess the association between ABPM abnormalities and sleep disturbance in the ADHD population, controlling for confounders as appropriate.

RESULTS

The compiled randomized list of potential participants included 240 children with a recorded diagnosis of ADHD. Contact was attempted with 145 families, 19 of whom were not reached due to outdated/incorrect phone numbers. Of those reached, 26 were excluded since they were not taking any prescription medications. Two patients declined participation and three were excluded because of ongoing treatment for comorbid chronic diseases. Of the remaining 95 eligible candidates, 57 were able to coordinate an appointment and an ABPM test during the study period. Two children were excluded from the final analysis due to noncompliance with the ABPM test, leaving a sample of 55 participants. The study was terminated before all candidates were contacted due to exhaustions of available resources. More than 85% of children in the study sample were treated with a type of psychostimulant medication indicated for ADHD: 36% methamphetamine, 22% combination of dextroamphetamine and amphetamine, 22% lisdexamphetamine and 3% dextroamphetamine. The rest of the children (14%) were treated with the norepinephrine reuptake inhibitor, atomoxetine.

Descriptive statistics of participants’ demographics, BMI, proportion of children with abnormal office BP as well as available parallel data from the CMHS are shown in Table 1. Office BP percentiles of the subjects are graphically illustrated in Figure 1 and the distribution of ABPM parameters is shown in Figure 2. We observed a significantly higher prevalence of suspected prehypertension (P=0.031) and hypertension (P=0.033) in the study sample, compared to the Canadian paediatric population (Table 1). Blood pressure classification combining results of office BP values and ABPM tests as well as prevalence of sleep disturbances are listed in Table 2. The ABPM studies demonstrated that all subjects with abnormal office blood pressure values had ‘white coat’ hypertension or could not be classified. Interestingly, the ABPM studies indicated that more than half of the children in our ADHD sample (51%, n=28) experienced nondipping. Sleep, mean ambulatory blood pressure above the 95th percentile was observed in 3.64% of cases, and sleep blood pressure load greater than 25% was observed in 18.8% of children in the study. Forty-three participants (78%) scored more than 39 on the SDSC instrument indicating a high prevalence of sleep disturbances, with more than half (51%) of the study sample scoring in the severe sleep disturbance range. Children with ADHD and ABPM abnormalities (except nondipping) were 5.6-fold more likely to score above 39 > on the SDSC than those without ABPM abnormalities (OR 5.6, 95% CI: 1.25 to 24.95; P=0.02), when controlling for BMI, age and sex. Similarly, children who had nondipping on ABPM were at five times the odds of having a sleep disturbance (SDCS score > 39) than their peers who did not experience nondipping (OR 5.13, 95% CI: 1.1 to 23; P=0.03) controlling for age, sex and BMI.

Table 1.

Demographic characteristics and office BP of children with ADHD and children from the Canadian Health Measures Survey (CMHS)

	ADHD sample	CHMS 2007–2009	CMHS 2012–2013
N (Boys %)	55(85.5%)	1850 (51.2%)	1899 (50.1%)
Age range, years	7–17	6–17	6–19
Average BMI/BMI-z score	17.9*/−0.37	19.6*/0.49	20.3/n/a
Overweight and obese, %	7.3%**	26.5%**	25.6%
Prehypertension (single office visit)	9.1%†	2.2%	3%†
Hypertension (single office visit)	5.45%‡	0.8%	2%‡

Difference in proportions/means: *1.4 (95% CI: 15.3–16.9), P= 0.0004; **−0.192 (95% CI: −0.310–0.074), P=0.002; ^†0.043 (95% CI: 0.004–0.141), P=0.031; ^‡0.035 (95% CI: −0.005–0.114), P=0.033.

ADHD Attention deficit/hyperactivity disorder; BP Blood pressure.

Figure 1.

Calculated office blood pressure (BP) percentiles of children with ADHD. (A) Systolic BP percentiles; (B) Diastolic BP percentiles.

Figure 2.

Ambulatory blood pressure (BP) monitoring results distribution among children with ADHD. (A) Ambulatory BP index; (B) Ambulatory BP load.

Table 2.

Classification of office BP values in combination with ABPM results and frequency of sleep disturbances in children with ADHD

Classification based on a single office BP and ABPM results*	N=55 (100%)
Normal BP (Office BP 90th percentile, ABPM average <95th percentile, Load <25%)	39 (70.9%)
White Coat HTN (Office BP ≥95th percentile, ABPM average <95th percentile, Load <25%)	3 (5.45%)
Pre-HTN Office (Office BP ≥90th percentile, ABPM average <95th percentile, Load ≥25%)	0
Masked HTN (Office BP <95th percentile, ABPM average >95th percentile, Load ≥25%)	2 (3.64%)
Ambulatory HTN (Office BP >95th percentile, ABPM average >95th percentile, Load 25–50%)	0
Unclassified 1 (Office BP ≤90th percentile, ABPM average <95th percentile, Load ≥25%)	6 (10.9%)
Unclassified 2 (Office BP ≥90th<95th percentile, ABPM average <95th percentile, Load <25%)	5 (9.1%)
Non-dipping (Mean sleep/(Mean wake – 0.1*Mean wake) > 1.0)	28 (50.91%)
Sleep disturbance based on a score > 39 on the SDSC questionnaire	43 (78.2%)

*Based on Flynn et al. (22).

ABPM Ambulatory blood pressure monitoring; ADHD Attention deficit/hyperactivity disorder; BP Blood pressure; HTN Hypertension; SDSC Sleep disturbance scale for children

DISCUSSION

This small prospective observational study provides reassuring evidence regarding the risk for clinically significant hypertension in children medicated for ADHD. As anticipated, ADHD pharmacotherapy was associated with more prevalent elevated office blood pressure values compared to children in the general population. However, ABPM in these children demonstrated that this largely represents ‘white coat’ hypertension. Moreover, office BP values suggesting prehypertension were not confirmed as such, using the BP classification criteria combining office BP values with ABPM results. Systematic screening of participants with ABPM identified potentially clinically significant BP abnormalities that are not detectable by office BP measurements, such as the high prevalence of nondipping, which appeared to be well correlated with sleep disturbances. A high prevalence of nocturnal BP load was also detected by ABPM as well as two cases on masked hypertension.

This study was inspired by a previous retrospective observation, that in our centre, a significant proportion of children referred for investigation of suspected hypertension had a comorbid diagnosis of ADHD (27). The results of the current study offer a partial explanation for this observation. Access to the CMHS data provided a unique opportunity to compare this sample of children medicated for ADHD with an appropriate control group of healthy Canadian children who were also randomly recruited for anthropometric and BP measurements using comparable equipment and methodology, except for the total number of measurements (25,28). Prevalence of elevated BP values in children sampled by the CMHS was significantly lower than reported by many large retrospective studies, some of which found the prevalence of elevated blood pressure in children, on a single visit, to be as high as 15% and even higher among obese children (29–31). However, similar to most other studies, the CMHS also found overweight to be a strong risk factor for both prehypertension and hypertension. Interestingly, the higher prevalence of elevated office BP values in children with ADHD does not appear to be related to overweight as the average BMI of this group was significantly lower compared to the CMHS sample of healthy Canadian children.

Despite the small sample, results of this study suggest that the ‘white coat’ phenomenon is largely responsible for elevated office blood pressure in children medicated for ADHD. Studies involving otherwise healthy children suspected of hypertension based on office measurements, found white coat hypertension (WCH) in up to 50% of cases (32,33). In children, WCH was found to be associated with higher BMIs and higher left ventricular mass which was still present even after adjusting for body mass index while in adults WCH is considered as a risk factor for cardiovascular events (34,35). The significance of WCH in children medicated for ADHD, who seem to have a tendency for lower BMIs, is not known.

The clinical significance of the very high risk for sleep nondipping discovered by ABPM screening is also not well defined. In children with diabetes, sleep nondipping was associated with increased risk for BP related target organ damage while in adults it has been shown to be associated with all-cause mortality (36,37). Overweight and obesity have also been found to be associated with higher prevalence of nondipping in up to 42% of prepubertal and early pubertal severely obese children (37). The pathophysiologic mechanism of nondipping is likely multifactorial and currently poorly defined, however sleep disturbances and autonomic nervous system imbalance have been implicated (36). Both these mechanisms have been associated with ADHD and have also been shown to be influenced by psycho-stimulant agents (38–42).

The SDSC demonstrated the fact that sleep disturbances are prevalent in children medicated for ADHD and a statistical correlation with sleep nondipping and other sleep ABPM abnormalities was identified, but the study was not designed to prove or search for causality. Furthermore, as with WCH, the significance of nondipping in this population is not known and is likely to be different than in populations with multiple cardiovascular risk factors such as diabetes and overweight.

The study was limited by the relatively small sample of participating children, the convenience bias related to recruitment from a single clinic and lack of a well-matched control group of nonmedicated children with ADHD. The use of the CMHS cohort as control representing the general population was less than ideal in view of the variability in the number of BP measurement. Nevertheless, this was a relatively unique attempt to define BP at rest and on ambulatory monitoring in a random sample of children with ADHD treated in the community with a variety of medications indicated for ADHD. Previous prospective small studies with rigorous cross-over design have proven beyond doubt that stimulant medications treatment in children is associated with a statistically significant elevation in ABPM values (13,14). However, these studies that focused exclusively on the effect of one or two psychostimulant agents did not demonstrate that this effect is also clinically significant or is associated with increased prevalence of hypertension. The aim of the present study was to capture a ‘real life’ sample of children medicated for ADHD rather than focus on the effect of a specific drug. This approach has the potential to provide more accurate answers to general questions related to BP, in children with ADHD given an increasing variety of available medications and a well-documented tendency of physicians and patients to try different approaches and make frequent dosage changes while monitoring clinical improvement of symptoms (43,44).

In summary, we prospectively examined blood pressure values measured during a single screening office visit and by ABPM in a sample of children with ADHD from a large community paediatric clinic. Compared to screened children in the general Canadian population, this cohort demonstrated higher prevalence of office BP values in the hypertension or prehypertension range despite having a lower average BMI compared to children in the general Canadian population. Ambulatory monitoring did not confirm a high prevalence of clinically significant hypertension in these children but suggested more frequent occurrence of ‘white coat’ hypertension as well as a common loss of the typical circadian BP pattern of sleep dipping. These findings support current American Academy of Pediatrics recommendation for careful longitudinal follow-up of children medicated for ADHD, including more frequent office BP monitoring; however, their long-term clinical significance is unknown. Longitudinal long-term studies focusing on the effect of childhood initiated ADHD pharmacotherapy on BP and cardiovascular outcomes will be required to eliminate such knowledge gaps.

Conflict of Interest

The authors have no conflicts of interest relevant to this article to disclose.