BACKGROUND
Intradialytic hypotension (IDH) is the most common complication of haemodialysis (HD). A purely nadir-based definition of IDH (nadir systolic blood pressure [SBP] of <90 mmHg or <100 mmHg, depending on the pre-dialysis SBP) was most strongly associated with mortality; whereas, definitions consisting of symptoms, as outlined in current guidelines, were not.1 The research team recently showed that the vast majority of IDH episodes (98%) are asymptomatic.2 Furthermore, in clinical practice, blood pressure (BP) is routinely measured twice per hour, while additional readings are only performed when symptoms arise. Hence, it is conceivable that the true incidence of IDH might be severely underestimated with the current routine BP-monitoring regimen.
AIMS
As intradialytic BP-drops may induce injury of vital organs, the primary goal is to assess the occurrence of IDH when BP is monitored continuously using a finger cuff (FC).
Yet, first, it appeared essential to validate intradialytic BP measurements using an FC by assessing the level of agreement with the standard oscillometric brachial arterial pressure (oBAP).
METHODS
Forty patients with HD were cross-over randomised to four dialysis modalities, each for 2 weeks: standard HD, cool HD, low-volume haemodiafiltration, and high-volume HDF. BP was measured every 15 minutes with an oBAP, and also every 20 seconds with a non-invasive FC device (ClearSight, Edwards Lifesciences, Irvine). For validation, the same arm sequential BP measurement method was applied. Paired BP measurements were compared with the Bland-Altman method. Bias was defined as the mean difference (MD) between oBAP and FC (oBAP minus FC), and 95% limits of agreement as the MD ±1.96 standard deviation The IDH analysis was restricted to FC measurements with an acceptable level of agreement, defined as a MD of ≤10 mmHg,3 and their corresponding 15-minute interval.
RESULTS
Seventy-five percent of participants were male, the mean age was 69.7 (±13.5 years), and median dialysis vintage was 3.0 years (interquartile range: 1.0–5.8). Diabetes mellitus and cardiovascular disease were present in 48% and 73% of participants, respectively. In 144 dialysis sessions, 33.4% of the FC measurements were missing, of which 43.8% were unexplained.
Validation
The Bland-Altman analysis of 1,221 paired BP measurements revealed a mean bias ± limits of agreement of 13±45 mmHg for SBP and 13±26 mmHg for diastolic BP.
IDH analysis
Twenty-seven percent of the paired BP measurements exhibited acceptable accuracy, resulting in the inclusion of 17,940 FC and 412 oBAP measurements. The FC detected 439 IDH episodes that had a median duration of 40 seconds (two measurements), with 1,703 out of 17,840 measurements meeting the IDH criteria. In contrast, during the same period, only 20 IDH episodes (of 412 measurements) were detected by the oBAP method.
CONCLUSIONS
Continuous BP-monitoring revealed an IDH occurrence which was ±22 times higher than 15-minute interval BP measurements, indicating that the majority of IDH episodes are missed with the current routine regimen.
As the FC showed poor accuracy and precision compared to the oBAP method, this method is, in fact, inappropriate for intradialytic BP-monitoring in patients undergoing HD.