Paediatric reference values for the work rate-indexed systolic blood pressure response during exercise

An inadequate increase or a decrease in systolic blood pressure (SBP) during exercise may indicate significant cardiovascular disease.1–3 As the SBP normally increases during a physical stress test in parallel with the work rate (WR) increment, the change in SBP by the increase in WR in Watt (SBP/W-slope) was recently proposed as a physiologically relevant method to evaluate the exercise SBP response in adults.4 However, paediatric reference values for the SBP/W-slope are lacking worldwide. We therefore aimed to evaluate the paediatric SBP response in relation to WR to provide reference values for exercise stress testing in children. In a consecutive cohort study, data from paediatric clinical bicycle exercise tests at a single centre between 2005 and 2016 were included. The clinical database was cross-linked with the National Patient Register until October 2016, to obtain data on hospital in-/out-patient clinical data. This database and the exercise stress protocol have been detailed elsewhere.4–6 Only children who performed a technically adequate test until maximal exhaustion were included, while patients with conditions or medications known to be associated with an abnormal SBP response during exercise were excluded (Supplementary material online). The difference between the first and last SBP measurements was divided by the difference in WR between these measurements to calculate the SBP/W-slope. Associations between potentially predictive variables and SBP/W-slope and the maximal SBP (SBPmax), respectively, were explored using scatterplots and mean differences by age-, height-, weight-, and WR-based groups. Stepwise regression analysis was performed, starting with the covariates that were considered to have the strongest association with the log-transformed SBP/W-slope or SBPmax. The variables were kept in the model if statistically significant (P < 0.05) and if they increased the model’s adjusted R2 value by at least 0.02. Sensitivity analyses were performed by excluding patients not reaching a rating of perceived exertion ≥17, and by calculating the SBP/W-slope at the earliest time point of the SBPmax, e.g. if the last and 2nd last SBP measurements were identical (Supplementary material online). Statistical significance was defined as a two-tailed P-value <0.05. Statistical analyses were performed using R (v.3.5.3, R Foundation for Statistical Computing, Vienna, Austria). The study was approved by the regional Ethical Review Boards (Dnr 2012/379-31, 2018/141-31), who waived individual informed consent. In total, 304 subjects (42% girls) were included; mean ± (standard deviation) age 14.6 ± 2.6 years; body mass index 19.6 ± 2.5 kg/m2; peak heart rate 188.0 ± 8.9 beats/min; peak WR (WRpeak) 99.4 ± 14% of predicted. Detailed baseline characteristics and a flow chart of inclusion/exclusion are presented in Supplementary material online. We found a strong association between SBPmax and WRpeak for both sexes (Figure 1). Although mean SBPmax was higher in boys than girls (167 vs. 153 mmHg, P < 0.001), there was no difference between girls and boys in SBPmax when tabulated by WRpeak (Figure 1) and no difference in mean SBP/W-slope, 0.38 ± 0.14 vs. 0.36 ± 0.14, P = 0.22. The SBP/W-slope was weakly correlated to height, weight, and age (R2 0.001–0.02). Including age and anthropometric data in the regression analysis of the log-transformed SBP/W-slope yielded a very low final R2, and none of the variables were statistically significant (Supplementary material online, Table SC). Instead of a regression equation, sex- and age-neutral normative paediatric values for the SBP/W-slope and tabulated data for SBPmax at various WRs are presented in Table 1. (A) Association of WRpeak and SBPmax during exercise described by scatterplots for SBPmax and WR stratified by sex. (B) Mean SBPmax (95% confidence intervals) in male and female children, and percentile values (5th, 10th, 90th, 95th) for all subjects stratified by WR. (C) Mean SBP/W-slope (95% confidence intervals) in different groups based on height in male and female children. (D) Mean SBP/W-slope (95% confidence intervals) in different groups based on weight in male and female children. (E) Mean SBP/W-slope (95% confidence intervals) in different groups based on age in male and female children. In all panels, red circles denote female subjects, and green triangles denote male subjects. Paediatric sex and age-neutral normative values for SBP/W-slope and SBPmax at different workloads Data presented in percentiles (all ages, both sexes). SBPmax stratified by peak work rate. SBPmax, Maximal systolic blood pressure during exercise; W, Watt. Paediatric sex and age-neutral normative values for SBP/W-slope and SBPmax at different workloads Data presented in percentiles (all ages, both sexes). SBPmax stratified by peak work rate. SBPmax, Maximal systolic blood pressure during exercise; W, Watt. SBP in mmHg, WRpeak in W, and age in (integer) years. The standard errors of the estimate for the models were 15.3 for females and 15.0 for males. This study confirms that SBPmax is linearly associated with WRpeak and that WRpeak is the single-most-important variable for predicting the SBP response during exercise in children, in both sexes. In light of this, we provide three ways of evaluating whether the SBP response in paediatric bicycle exercise stress testing is normal: (i) novel paediatric reference values for the SBP/W-slope, (ii) normative reference values for SBPmax tabulated by WRpeak, and (iii) reference equations for SBPmax, which includes WR. Future studies are needed to evaluate the prognostic and diagnostic value of these three measures. We found the paediatric reference values for the SBP/W-slope to be lower, compared to recently published values for adults.4 Possibly by children having a better ability to lower peripheral resistance during exercise compared to older individuals.7 Interestingly, and in contrast to adult data,4 we found no difference in SBPmax between males and females when adjusting for WR—possibly explained by less differences between sexes in terms of muscle mass, height, weight, and physical performance in children compared to in adults. Thus, there is no need for sex- or age stratification of SBP/W-slope reference values in paediatric populations, making clinical application of these reference values straightforward. Our finding agrees with prior studies concluding that by relating SBPmax to WRpeak it is possible to make paediatric reference values independent of sex.8–11 These reference values can be applied to Scandinavian populations but are not necessarily generalizable to other populations. Limitations to our study include the relatively small sample size, and that the study population consisted of children who underwent clinical exercise testing. Also, we lack information on congenital heart disease beyond valvular heart disease and cardiomyopathy. However, we believe that by applying strict exclusion criteria, including only subjects free from relevant cardiovascular disease and medications, with normal exercise tests including a normal exercise capacity and an adequate heart rate increase,12 the impact of the potential bias of a clinical referral sample is small. In conclusion, the SBP response during exercise in children is closely related to WR, which should be considered in clinical interpretation of the SBP response in paediatric exercise tests. We present the first paediatric reference values for the SBP/W-slope as a feasible method for WR-indexing the SBP response in children. Supplementary material is available at European Journal of Preventive Cardiology online. An unrestricted grant from the Swedish Society for Medical Research and the Swedish Research Council (Dnr: 2019-02081 to M.E.); an unrestricted grant from the Scientific Committee of Blekinge County Council to V.E.; post-doctoral research grants from The Swedish Heart Lung Foundation (20200553), the Swedish Cardiac Society, the Royal Swedish Academy of Sciences (LM2019-0013), Women & Health Foundation, Region Kronoberg (8301), the Swedish Heart and Lung Association, Swedish Association of Clinical Physiology, and the Scandinavian Society of Clinical Physiology & Nuclear Medicine to T.L. An unrestricted grant from the county council of Östergötland to K.H. Remaining authors received no funding. Conflict of interest: none declared.