In this first detailed fast-track VATS study, we have shown that the perioperative time-course of several HRV indices is dependent on both what time of day they are measured, and on which day they are measured in relation to surgery. We found that indices measuring overall variability of the heart rate (SDNN and TP) were markedly reduced after VATS lobectomy, especially in the day- and nighttime, showing only slight signs of returning towards baseline at the end of the study 9 days after surgery. This cohort seemed to be a representative sample of the patients treated at our institution, as they were similar to a previous larger cohort [14].
A study in 117 patients found significantly higher preoperative values in HRV parameters reflecting increased parasympathetic activity in the 19 patients who developed PoAF following open pulmonary resection [15], in line with the association between parasympathetic stimulation and increased risk of atrial fibrillation [16]. In a comparative study of thoracotomy and thoracoscopy, Ivanovic et al [17] found fewer cases of PoAF in the thoracoscopy group but without HRV data.
Another study [18], comparing perioperative HRV in lobectomy by VATS (n = 112) or open thoracotomy (n = 112) did not find any differences in PoAF between the two techniques. However, they reported an unexpected near 2-fold increase in SDNN in the first 24 hours following both VATS and open surgery [18]. Our results are in contrast with this, as we saw a 30% drop from PRE1 to POD1 (Fig. 2 Day). There are several inconsistencies in the reported methodology in their paper, confounding interpretation. We have inquired the corresponding author about access to their results through several channels without reply so far, hindering interpretation and comparison between studies.
In our study, only two of 28 patients had PoAF, precluding statistical analysis on HRV associations.
While the most important reason for prolonged hospital stay after VATS is air leak, other key reasons are: pain, pneumonia, and PoAF [11, 14]. Preoperative HRV has shown promise in predicting pain, pneumonia, and PoAF in preliminary studies, in other surgical settings including thoracic surgery [1, 15], and has been associated with differing trajectories of molecular inflammatory responses following abdominal surgery, including higher CRP in the group of patients with lower short-term variation [19], suggesting an association between inflammation and HRV changes.
No studies have examined the correlation between HRV and postoperative outcome in VATS, but the results presented in this study will aid in designing such future procedure-specific trials. From our study it seems that SDNN or TP in the night- or daytime would be the HRV measures most sensitive to the surgical insult and continuous measurement across several days would be recommended. However, if one is interested in a stable measure across both time of day, and day in relation to surgery, LF/HF or DFA1 might be a suitable measure as they show little pre- and even perioperative variation.
In accordance with previous publications on VATS [20] we found an increase in orthostatic symptoms on the first postoperative day. Preoperative HRV measurement could potentially identify patients at risk of postoperative orthostatic intolerance, as shown for intraoperative hypotension [1, 2], but this hypothesis-generating study lack the power for such analysis.
We also saw an increase in non-orthostatic symptoms (pain, weakness, and fatigue) in the patients undergoing VATS from the POD1 as demonstrated in a recent study with perioperative assessment of physical activity [21], but our study has no power for analysis of HRV in relation to reported symptom scores.
The strengths of this study are that patients were followed continuously for a longer time both pre- and postoperatively than previous studies in VATS, that we assessed autonomic dysfunction through both subjective measures via the ODSS and objective HRV measures using more HRV indices than previously reported. Limitations are the small sample size, not allowing any secondary analysis of risk stratification, especially for PoAF, and the partially missing data in 4 patients. We did not control for time of surgery, and therefore had to discard the day recording on the DOS, and the study was performed in free-roaming subjects limiting comparison to shorter recordings under resting conditions.
In conclusion, for the first time in fast-track VATS, we have shown that HRV indices exhibit different responses to surgery, depending on both the day of recording in relation to surgery and what time of the day the recording was made. Secondly, that some indices exhibit marked circadian variation in the preoperative period. Total HRV was markedly reduced following surgery, whereas single, more specific measures were more stable. Reductions in total HRV coincided with increased orthostatic intolerance as well as with pain, weakness, and fatigue at least 9 days after surgery. These results are valuable in designing future HRV studies in VATS aimed at preoperative risk stratification.