Review Reports
- Rukmono Siswishanto 1,*,
- Detty Siti Nurdiati 1 and
- Dean Batrisha 1
- et al.
Reviewer 1: Anonymous Reviewer 2: Anonymous Reviewer 3: Vitor Engracia Valenti
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis systematic review addresses a timely and clinically relevant topic by evaluating wearable-derived heart rate variability (HRV) as a non-invasive marker of systemic inflammation and follows established reporting standards (PRISMA, PROSPERO, SWiM). The search strategy is comprehensive and the methodological approach is transparent; however, several limitations substantially restrict the strength of the conclusions. The evidence base is small and heterogeneous, with only 11 studies, predominantly observational designs, and wide variability in populations, inflammatory conditions, wearable devices, recording durations, and biomarkers. Reliance on vote counting and sign tests, while methodologically justified under SWiM, limits interpretability because studies are weighted equally regardless of sample size, quality, or precision, and no effect size estimation is provided. Associations beyond SDNN–CRP are inconsistent, often based on single studies, and lack standardized adjustment for confounders such as physical activity, medication use, and circadian influences. Moreover, the absence of formal diagnostic accuracy metrics, validated thresholds, and longitudinal outcome prediction limits clinical applicability. Overall, while the review provides a useful overview of emerging evidence, stronger emphasis on standardized HRV acquisition, external validation, quantitative synthesis where feasible, and clinically meaningful endpoints is necessary before wearable-derived HRV can be considered a reliable tool for inflammation detection.
Author Response
Diagnostics MDPI
Revision Round 1
# Reviewer 1
Comments 1
The evidence base is small and heterogeneous, with only 11 studies, predominantly observational designs, and wide variability in populations, inflammatory conditions, wearable devices, recording durations, and biomarkers.
Response 1
We agree that the current evidence base is limited in size and heterogeneous with respect to study design, populations, inflammatory contexts, wearable devices, recording durations, and biomarkers. To address this, we have revised the early Discussion section to explicitly frame the review as a synthesis of emerging evidence, rather than a definitive evaluation. We now clearly state that the findings should be interpreted in light of the limited number of studies and substantial methodological heterogeneity across the included literature.
Accordingly, this review should be interpreted as a structured synthesis of emerging evidence rather than a definitive evaluation, reflecting the limited number of available studies and the substantial heterogeneity in study populations, inflammatory contexts, and wearable methodologies. (Pg.13, Line 405)
Comments 2
Reliance on vote counting and sign tests, while methodologically justified under SWiM, limits interpretability because studies are weighted equally regardless of sample size, quality, or precision, and no effect size estimation is provided.
Response 2
We agree with the reviewer that reliance on vote counting and sign tests limits interpretability, as these approaches assign equal weight to studies regardless of sample size, precision, or methodological quality and do not provide estimates of effect magnitude or diagnostic performance. We have revised the Limitations section to explicitly acknowledge these constraints. The manuscript now states that the SWiM-based synthesis is intended to reflect the direction and consistency of associations only, rather than their strength, clinical relevance, or predictive value, and that quantitative pooling or effect size estimation was not feasible due to substantial heterogeneity in reporting formats, adjustment strategies, and outcome definitions across studies. These clarifications are intended to ensure appropriate interpretation of the findings.
Several limitations must be acknowledged. The SWiM-based vote counting approach applied in this review has inherent constraints, as all included studies contribute equally regardless of sample size, precision, or methodological quality, and the synthesis does not estimate effect magnitude, diagnostic performance, or predictive value, thus this review provides a lower level of certainty regarding overall strength of evidence [18]. These methods were selected because of substantial heterogeneity primarily in outcome metrics, study design, populations, wearable technologies, and analytical approaches preventing formal quantitative meta-analysis, making a SWiM-based synthesis the most appropriate approach for summarizing qualitative patterns of association despite its inferential limitations. Beyond these analytical constraints, many included studies had small sample sizes, increasing the risk of type II error, and the predominance of observational designs limits causal inference. In addition, key confounders known to influence HRV including medication use, circadian timing, physical activity levels, respiratory patterns, and individual physiological differences were inconsistently reported or adjusted for, which may have biased observed associations. Substantial heterogeneity in study populations, wearable device technologies, recording durations, biomarker assays, and analytical protocols further complicates cross-study comparisons and may contribute to inconsistent findings across inflammatory markers.(Pg. 14, line 464)
Comments 3
Associations beyond SDNN–CRP are inconsistent, often based on single studies, and lack standardized adjustment for confounders such as physical activity, medication use, and circadian influences.
Response 3
We agree with the reviewer and have revised the Discussion to address this point more explicitly. We now state that HRV associations with inflammatory markers other than CRP are inconsistent and frequently based on single studies, precluding robust inference. In addition, we emphasize that interpretation of HRV inflammation relationships is constrained by inconsistent reporting and adjustment for key confounders including physical activity, medication use, circadian timing, respiratory patterns, and comorbidities which reflects limitations of the underlying primary studies rather than the review methodology itself. These clarifications were added to guide appropriate interpretation of the findings. To address this, we added the following clarification to the Discussion:
In contrast, biomarkers such as fecal calprotectin and galectin-3 were each represented by single studies, and their reported associations should therefore be considered exploratory. Moreover, the apparent consistency observed for SDNN-CRP should be interpreted cautiously, as sign-test results were borderline and confidence intervals were wide, reflecting limited statistical precision and the small number of contributing studies. The diversity of study populations ranging from acute inflammatory responses to chronic metabolic or cardiovascular conditions together with the predominance of observational designs restricts causal interpretation. Inconsistent reporting and adjustment for key confounders, including physical activity, mediaction use, circardian timing, respiratory patterns, and comorbidities, further limit generalizability across clinical contexts.(Pg.13, line 412)
Comments 4
Moreover, the absence of formal diagnostic accuracy metrics, validated thresholds, and longitudinal outcome prediction limits clinical applicability.
Response 4
We agree with the reviewer that the absence of formal diagnostic accuracy metrics, validated thresholds, and longitudinal outcome prediction currently limits clinical applicability. To address this concern, we revised the Abstract, Discussion, and Conclusions to explicitly clarify the scope of inference. The manuscript now states that wearable-derived HRV cannot presently be considered a diagnostic or predictive tool for systemic inflammation and should instead be viewed as an exploratory or adjunctive physiological correlate. We further emphasize that formal diagnostic accuracy analyses, validated thresholds, and longitudinal outcome-based studies are necessary prerequisites before any clinical implementation can be considered.
Reviewer 2 Report
Comments and Suggestions for AuthorsMajor Limitations
1.The evidence base is small (11 studies) and highly heterogeneous in populations, devices, biomarkers, and recording protocols, substantially limiting internal validity and generalizability.
2.Reliance on vote counting and sign tests, while justified, inherently ignores effect size, study weighting, and precision, which weakens inferential strength.
3.Diagnostic accuracy—the stated clinical objective—is not formally evaluated (e.g., sensitivity, specificity, AUC), limiting the translational impact of the conclusions.
4.Several included studies have small sample sizes and observational designs with limited adjustment for confounders (e.g., medications, circadian effects, physical activity).
Methodological Concerns
.Inclusion of both acute experimental endotoxemia and chronic inflammatory conditions under a single framework may obscure fundamentally different autonomic–immune dynamics.
6.Language restriction (English and Indonesian only) introduces potential selection bias.
7.Some biomarkers (e.g., fecal calprotectin, galectin-3) are represented by single studies, yet are discussed alongside more established markers such as CRP.
Interpretation and Presentation
8.The conclusion that SDNN shows the “most consistent” association with CRP is reasonable but should be framed more cautiously given the borderline statistical significance and wide confidence intervals.
9.The manuscript would benefit from clearer separation between physiological association and clinical diagnostic utility.
10.Tables are comprehensive but dense; summarizing key findings in a simplified figure or hierarchy of evidence could improve readability.
11.To improve discussion section add "The Relationship Between Heart Rate Variability Parameters and Atrioventricular Nodal Reentrant Tachycardia. Koşuyolu Heart Journal, 2018, 21.2: 128-134."
Comments on the Quality of English LanguageShould be improved
Author Response
Diagnostics MDPI
Revision Round 1
# Reviewer 2
Comment 1
The evidence base is small (11 studies) and highly heterogeneous in populations, devices, biomarkers, and recording protocols, substantially limiting internal validity and generalizability.
Response 1
We agree that the evidence base is limited and heterogeneous. We have revised the Discussion to explicitly frame this review as a structured synthesis of emerging evidence rather than a definitive evaluation, and we emphasize the implications of heterogeneity for generalizability.
Accordingly, this review should be interpreted as a structured synthesis of emerging evidence rather than a definitive evaluation, reflecting the limited number of available studies and the substantial heterogeneity in study populations, inflammatory contexts, and wearable methodologies. (Pg.13, Line 405)
Comment 2
Reliance on vote counting and sign tests, while justified, inherently ignores effect size, study weighting, and precision, which weakens inferential strength.
Response 2
We thank the reviewer for this helpful suggestion. In response, we have revised the Discussion to explicitly link the acknowledged limitations of the SWiM-based vote counting approach with the rationale for its use. We now clarify that substantial heterogeneity across study designs, populations, wearable technologies, outcome definitions, and analytical approaches precluded formal quantitative meta-analysis, and that a SWiM-based synthesis represented the most appropriate method for summarizing qualitative patterns of association despite its inherent limitations on inferential strength.
Several limitations must be acknowledged. The SWiM-based vote counting approach applied in this review has inherent constraints, as all included studies contribute equally regardless of sample size, precision, or methodological quality, and the synthesis does not estimate effect magnitude, diagnostic performance, or predictive value, thus this review provides a lower level of certainty regarding overall strength of evidence [18]. These methods were selected because substantial heterogeneity primarily in outcome metrics, study design, populations, wearable technologies, and analytical approaches preventing formal quantitative meta-analysis, making a SWiM-based synthesis the most appropriate approach for summarizing qualitative patterns of association despite its inferential limitations. Beyond these analytical constraints, many included studies had small sample sizes, increasing the risk of type II error, and the predominance of observational designs limits causal inference. In addition, key confounders known to influence HRV including medication use, circadian timing, physical activity levels, respiratory patterns, and individual physiological differences were inconsistently reported or adjusted for, which may have biased observed associations. Substantial heterogeneity in study populations, wearable device technologies, recording durations, biomarker assays, and analytical protocols further complicates cross-study comparisons and may contribute to inconsistent findings across inflammatory markers.(Pg. 14, line 464)
Comment 3
Diagnostic accuracy—the stated clinical objective—is not formally evaluated (e.g., sensitivity, specificity, AUC), limiting the translational impact of the conclusions.
Response 3
We thank the reviewer for this comment. We would like to clarify that formal diagnostic accuracy evaluation was not a stated objective of this review. As revised, the manuscript consistently frames its aim as synthesizing association-based evidence between wearable-derived HRV and inflammatory biomarkers. To address this point explicitly, we now state in the Discussion and Conclusions that formal diagnostic accuracy analyses (e.g., sensitivity, specificity, AUC) represent a critical next step for future studies prior to clinical translation.
Throughout this review, associations between wearable-derived HRV and in-flammatory biomarkers are interpreted as physiological correlations rather than evidence of diagnostic performance. Establishing such biological associations represents a prerequisite for, but is conceptually distinct from the validation of HRV as a diagnostic or predictive tool. In our review, formal diagnostic metrics, such as sensitivity, speci-ficity, and AUC, were inconsistently reported across the included studies and were often unavailable. Accordingly, a formal evaluation of diagnostic accuracy was not feasible and the translational implications of these findings are intentionally framed within an exploratory context. (Pg.14 line 454)
Comment 4
Several included studies have small sample sizes and observational designs with limited adjustment for confounders (e.g., medications, circadian effects, physical activity).
Response 4
We agree and have emphasized that many included studies had small sample sizes and predominantly observational designs, with inconsistent adjustment for key confounders such as medication use, circadian timing, and physical activity. These limitations are now highlighted earlier in the Discussion and in the Limitations section.
In contrast, biomarkers such as fecal calprotectin and galectin-3 were each represented by single studies, and their reported associations should therefore be considered exploratory. Moreover, the apparent consistency observed for SDNN-CRP should be interpreted cautiously, as sign-test results were borderline and confidence intervals were wide, reflecting limited statistical precision and the small number of contributing studies. The diversity of study populations ranging from acute inflammatory responses to chronic metabolic or cardiovascular conditions together with the predominance of observational designs restricts causal interpretation. Inconsistent reporting and adjustment for key confounders, including physical activity, mediaction use, circardian timing, respiratory patterns, and comorbidities, further limit generalizability across clinical contexts.(Pg.13, line 412)
Methodological Concerns
Comment 5
Inclusion of both acute experimental endotoxemia and chronic inflammatory conditions under a single framework may obscure fundamentally different autonomic–immune dynamics.
Response 5
We thank the reviewer for this constructive suggestion. We have expanded the Discussion to explicitly address the implications of including both acute and chronic inflammatory states, noting that these contexts may reflect distinct autonomic–immune dynamics and that future studies or systematic reviews would benefit from stratifying analyses by inflammatory chronicity.
The diversity of study populations ranging from acute inflammatory responses to chronic metabolic or cardiovascular conditions together with the predominance of observational designs restricts causal interpretation. Inconsistent reporting and adjustment for key confounders, including physical activity, mediaction use, circardian timing, respiratory patterns, and comorbidities, further limit generalizability across clinical contexts. (Pg. 14, line 417)
Comment 6
Language restriction (English and Indonesian only) introduces potential selection bias.
Response 6
We thank the reviewer for this comment. As suggested, we have revised the Limitations section to more clearly emphasize that restricting inclusion to English and Indonesian publications may lead to an incomplete representation of the global evidence base, particularly from non-English- and non-Indonesian-speaking research communities.
Finally, restricting inclusion to English- and Indonesian-language publications may have excluded relevant studies published in other languages, introducing potential language-related selection bias. (Pg. 15, line 481)
Comment 7
Some biomarkers (e.g., fecal calprotectin, galectin-3) are represented by single studies, yet are discussed alongside more established markers such as CRP.
Response 7
We thank the reviewer for this helpful suggestion. In response, we have added explicit cautionary statements in both the Results and Discussion sections indicating that findings for biomarkers represented by single studies (e.g., fecal calprotectin and galectin-3) should be interpreted cautiously and considered preliminary, in contrast to the more consistent evidence observed for CRP.
In contrast, biomarkers such as fecal calprotectin and galectin-3 were each represented by single studies, and their reported associations should therefore be considered exploratory. Moreover, the apparent consistency observed for SDNN-CRP should be interpreted cautiously, as sign-test results were borderline and confidence intervals were wide, reflecting limited statistical precision and the small number of contributing studies. (Pg.13, line 412)
Interpretation and Presentation
Comment 8
The conclusion that SDNN shows the “most consistent” association with CRP is reasonable but should be framed more cautiously given the borderline statistical significance and wide confidence intervals.
Response 8
We thank the reviewer for this important point. In response, we have softened the language used to describe the SDNN–CRP association in the Discussion and Conclusions, replacing absolute phrasing with more cautious terms (e.g., “appears to show the most consistent trend”). We also explicitly acknowledge the borderline significance of the sign-test results and the wide confidence intervals, and we clarify that associations involving RMSSD were statistically non-significant. These revisions better align the narrative interpretation with the underlying statistical uncertainty.
Comment 9
The manuscript would benefit from clearer separation between physiological association and clinical diagnostic utility.
Response 9
We agree with the reviewer and have revised the manuscript to more clearly distinguish between physiological associations and clinical diagnostic utility. Throughout the Abstract, Discussion, and Conclusions, we now explicitly frame wearable-derived HRV as a physiological correlate of systemic inflammation rather than a validated diagnostic or predictive tool. Statements implying clinical applicability have been refined, and the absence of formal diagnostic accuracy metrics (e.g., sensitivity, specificity, AUC) is now emphasized to prevent overinterpretation.
Comment 10
Tables are comprehensive but dense; summarizing key findings in a simplified figure or hierarchy of evidence could improve readability.
Response 10
We agree with the reviewer. To improve readability and facilitate interpretation of the dense tabular data, we emphasized the use of effect direction plots and vote-counting summaries as simplified visual overviews of the consistency and direction of associations across studies (Supplementary Tables S4 and S5; Table 2). The table captions were also revised to clarify that these symbols support qualitative interpretation rather than quantitative or diagnostic inference.
Comment 11
To improve discussion section add "The Relationship Between Heart Rate Variability Parameters and Atrioventricular Nodal Reentrant Tachycardia. Koşuyolu Heart Journal, 2018, 21.2: 128-134."
Response 11
We agree with the reviewer and have revised the Discussion to incorporate evidence from cardiovascular autonomic disorders, including atrioventricular nodal reentrant tachycardia (AVNRT), to contextualize HRV as a marker of sympathovagal balance rather than a disease-specific diagnostic signal. The added text cites studies demonstrating altered HRV patterns in AVNRT despite the absence of structural heart disease, thereby reinforcing the interpretation of HRV as a physiological correlate of autonomic regulation (Page 14, line 436).
Evidence from cardiovascular autonomic disorders further supports the interpretation of HRV as a marker of sympathovagal balance rather than a disease-specific diagnostic signal. For example, studies in atrioventricular nodal reentrant tachycardia (AVNRT) demonstrate altered HRV patterns consistent with heightened sympathetic modulation, even in the absence of structural heart disease [35,36]. These findings underscore that HRV indices are sensitive to underlying autonomic state across diverse clinical conditions, reinforcing their role as physiological correlates of autonomic regulation rather than direct indicators of specific pathological processes.(Pg.14, line 436)
Reviewer 3 Report
Comments and Suggestions for AuthorsThe main question addressed by this systematic review is whether heart rate variability (HRV) derived from wearable devices is meaningfully associated with systemic inflammatory biomarkers, and whether wearable-derived HRV could plausibly serve as a non-invasive physiological indicator of systemic inflammation across different clinical contexts.
In my view, this is a relevant and well-defined research question. The authors clearly focus on commonly used HRV indices—particularly SDNN and RMSSD—and their associations with established inflammatory markers such as CRP, IL-6, and TNF-α. This focus is appropriate, as these markers are widely used in both clinical and experimental research, and they represent distinct aspects of inflammatory activity.
The topic itself is timely and important. Wearable technologies have rapidly expanded into both research and consumer health settings, often with claims that exceed the available evidence. While the association between HRV and inflammation has been explored previously using traditional ECG recordings, this review addresses a meaningful gap by focusing specifically on wearable-derived HRV. This distinction matters. Wearable devices differ substantially from standard ECG in terms of signal quality, recording duration, motion artefacts, and real-world conditions, and these factors can influence HRV metrics in non-trivial ways.
To the best of my knowledge, this appears to be among the first systematic reviews to explicitly synthesize evidence on wearable-derived HRV in relation to inflammatory biomarkers across diverse populations. In that sense, the manuscript makes a novel contribution.
Compared with prior reviews and meta-analyses examining HRV and inflammation more broadly, the added value of this study lies in several specific choices. First, the exclusive focus on wearable devices, rather than laboratory-based ECG, is a strength. Second, the differentiation between ECG-based and PPG-based wearables is particularly useful, as several validation studies (for instance, those comparing chest-strap ECG devices with wrist-worn PPG sensors during free-living conditions) have demonstrated meaningful differences in signal reliability. I found this distinction clinically and methodologically important. Third, the decision to apply a SWiM framework, rather than forcing a quantitative meta-analysis in the presence of substantial heterogeneity, strikes me as appropriate and responsible.
The identification of SDNN—especially when derived from ECG-based wearables—as the HRV metric most consistently associated with CRP is an interesting finding. However, I would personally urge some caution here. The number of contributing studies remains limited, and confidence intervals are often wide. While the direction of association appears consistent, the strength and robustness of this relationship are still uncertain.
One aspect I appreciated is that the authors do not overstate their findings. The manuscript repeatedly acknowledges heterogeneity, uncertainty, and current limitations. In a field where consumer-grade technologies are frequently overinterpreted, this restraint is both refreshing and necessary.
That said, there are several areas where the manuscript could be improved.
First, the concept of “systemic inflammation” would benefit from clearer operationalization. Across the included studies, inflammation appears to be defined in different ways—sometimes based on clinical diagnosis, sometimes on biomarker thresholds, and in other cases on experimental induction. A brief clarification of how these approaches were treated conceptually would improve clarity for the reader.
Second, confounding factors deserve more explicit attention. Variables such as medication use, physical activity levels, circadian rhythm, respiratory patterns, and comorbidities are inconsistently reported across studies, yet all are known to influence HRV. While this limitation is mentioned, I believe it should be emphasized more clearly, either in the methods or early in the discussion, as it directly affects the interpretation of HRV–inflammation associations.
Third, the inclusion of both randomized and observational study designs is reasonable, and I do not see this as a flaw per se. However, a short justification explaining how evidence from these different designs was interpreted together would strengthen methodological transparency.
Another point that could benefit from clarification concerns studies reporting multiple inflammatory biomarkers. In such cases, clearer description of the decision rules used to ensure independence in the sign-test analysis would help readers better understand the analytical choices and their implications.
Overall, the conclusions are largely consistent with the presented evidence and do address the main research question. The authors appropriately state that wearable-derived HRV—particularly SDNN from ECG-based devices—shows potential associations with systemic inflammation, especially CRP. Still, I would recommend slightly more cautious wording in a few places to avoid any implication of diagnostic or predictive utility. Given the predominance of observational designs, modest sample sizes, and substantial heterogeneity, the evidence currently supports association rather than clinical application.
I would therefore encourage the authors to emphasize more clearly that wearable HRV should, at present, be viewed as exploratory or adjunctive, rather than as a validated clinical tool.
The reference list is appropriate, relevant, and up to date. It includes foundational literature on HRV physiology, key studies on neuro-immune interactions, methodological guidance for systematic reviews and SWiM reporting, and recent wearable validation studies. I did not identify any major omissions.
The tables summarizing study characteristics are generally clear and informative, although minor formatting and consistency issues should be checked. The effect-direction plots are appropriate and useful within the SWiM framework; however, clearer captions that guide readers toward clinical interpretation—not only statistical direction—would improve their accessibility. The PRISMA flow diagram is appropriate and clearly presented.
In summary, this is a well-conducted and thoughtful review that addresses an important and timely question. With clearer conceptual definitions, more explicit discussion of confounding, and slightly more cautious interpretation of clinical implications, the manuscript would be further strengthened.
Author Response
Diagnostics MDPI
Revision Round 1
# Reviewer 3
Comment 1
The concept of “systemic inflammation” would benefit from clearer operationalization. Across the included studies, inflammation appears to be defined in different ways—sometimes based on clinical diagnosis, sometimes on biomarker thresholds, and in other cases on experimental induction. A brief clarification of how these approaches were treated conceptually would improve clarity for the reader.
Response 1
We thank the reviewer for this helpful suggestion. We have added a brief clarification to the manuscript describing how “systemic inflammation” was operationalized across included studies. Specifically, we now explain that inflammation was defined using clinical diagnoses, biomarker-based thresholds, or experimentally induced inflammatory models, which were treated as conceptually related but biologically heterogeneous manifestations of systemic inflammatory activity. This clarification was added to improve conceptual transparency for readers.
In this review, “systemic inflammation” was operationalized broadly to encompass clinical inflammatory conditions, biomarker-defined inflammatory states, and experimentally induced inflammatory responses. Included studies defined inflammation using established circulating biomarkers (e.g., CRP and cytokines), clinical diagnoses associated with inflammatory activity, or controlled experimental models such as endotoxemia. These approaches were treated as conceptually related but biologically heterogeneous manifestations of systemic inflammatory activity. (Pg.3, line 95)
Comment 2
Confounding factors deserve more explicit attention. Variables such as medication use, physical activity levels, circadian rhythm, respiratory patterns, and comorbidities are inconsistently reported across studies, yet all are known to influence HRV. While this limitation is mentioned, I believe it should be emphasized more clearly, either in the methods or early in the discussion, as it directly affects the interpretation of HRV–inflammation associations.
Response 2
We agree with the reviewer and have strengthened the Discussion to emphasize the impact of key confounding factors, including medication use, physical activity, circadian timing, respiratory patterns, and comorbidities. We now explicitly state that inconsistent reporting and adjustment for these variables constrains interpretation of HRV–inflammation associations and reflects limitations of the underlying primary studies rather than the review methodology. This point is highlighted earlier in the Discussion to guide appropriate interpretation.
In contrast, biomarkers such as fecal calprotectin and galectin-3 were each represented by single studies, and their reported associations should therefore be considered exploratory. Moreover, the apparent consistency observed for SDNN-CRP should be interpreted cautiously, as sign-test results were borderline and confidence intervals were wide, reflecting limited statistical precision and the small number of contributing studies. The diversity of study populations ranging from acute inflammatory responses to chronic metabolic or cardiovascular conditions together with the predominance of observational designs restricts causal interpretation. Inconsistent reporting and adjustment for key confounders, including physical activity, mediaction use, circardian timing, respiratory patterns, and comorbidities, further limit generalizability across clinical contexts.(Pg.13, line 412)
Comment 3
The inclusion of both randomized and observational study designs is reasonable, and I do not see this as a flaw per se. However, a short justification explaining how evidence from these different designs was interpreted together would strengthen methodological transparency.
Response 3
We thank the reviewer for this suggestion. We have added a brief justification in the Methods (Data Synthesis and Analysis) section clarifying that randomized and observational studies were synthesized together to assess the consistency of associations rather than to infer causal effects, with study design considered descriptively rather than hierarchically weighted. This clarification was included to improve methodological transparency.
Randomized and observational studies were synthesized together to assess the consistency of associations between wearable-derived HRV and inflammatory biomarkers rather than to infer causal effects, with study design considered descriptively rather than hierarchically weighted. (Pg.4, line 177)
Comment 4
Another point that could benefit from clarification concerns studies reporting multiple inflammatory biomarkers. In such cases, clearer description of the decision rules used to ensure independence in the sign-test analysis would help readers better understand the analytical choices and their implications.
Response 4
We thank the reviewer for this suggestion. We have clarified the decision rules in the Methods (Data Synthesis and Analysis) section, stating that when studies reported multiple inflammatory biomarkers, a single representative HRV biomarker comparison was selected per study for sign-test analyses to preserve statistical independence, with priority given to the primary or most clinically relevant biomarker reported. This clarification is intended to improve transparency regarding our analytical choices.
When studies reported multiple inflammatory biomarkers, a single representative HRV biomarker comparison was selected per study for sign-test analysis to preserve statistical independence, with priority given to the primary or most clinically relevant biomarker reported, with CRP prioritized when available. (Pg.5, line 193)
Comment 5
The authors appropriately state that wearable-derived HRV—particularly SDNN from ECG-based devices—shows potential associations with systemic inflammation, especially CRP. Still, I would recommend slightly more cautious wording in a few places to avoid any implication of diagnostic or predictive utility. Given the predominance of observational designs, modest sample sizes, and substantial heterogeneity, the evidence currently supports association rather than clinical application.
Response 5
We agree with the reviewer and have revised the Abstract, Discussion, and Conclusions to further soften clinical language and avoid any implication of diagnostic or predictive utility. The manuscript now explicitly emphasizes that wearable-derived HRV should currently be regarded as exploratory or adjunctive rather than a validated clinical tool, consistent with the predominance of observational evidence, modest sample sizes, and substantial heterogeneity.
Comment 6
The tables summarizing study characteristics are generally clear and informative, although minor formatting and consistency issues should be checked. The effect-direction plots are appropriate and useful within the SWiM framework; however, clearer captions that guide readers toward clinical interpretation—not only statistical direction—would improve their accessibility. The PRISMA flow diagram is appropriate and clearly presented.
Response 6
We thank the reviewer for this suggestion. We have revised the effect-direction plot captions to more clearly guide interpretation, emphasizing that the figures illustrate direction of association only and do not convey effect magnitude or clinical decision thresholds.
Table 1. General characteristics of studies included in present systematic review. (Pg7., line 261)
The symbols summarize the direction of reported associations and are intended to support qualitative interpretation rather than quantitative or diagnostic inference. (Pg.11, line 307)
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsI do not find the corrections made in the article to be positive. It would be appropriate to further improve the current version.
Comments on the Quality of English LanguageShould be improved
Reviewer 3 Report
Comments and Suggestions for AuthorsAll requested changes were made.