Which patients are more likely to experience compensatory hyperhidrosis after endoscopic thoracic sympathectomy: a meta-analysis and systematic review

Study design and protocol

This systematic review and meta-analysis were conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (PRISMA). The study protocol was prospectively registered with the International Prospective Register of Systematic Reviews under the registration number CRD42024592389, ensuring transparency and minimizing the risk of bias in study selection and data extraction. The protocol outlined the objectives, inclusion and exclusion criteria, search strategy, data extraction methods, and statistical analysis plan.

Literature search strategy

A comprehensive literature search was conducted in the PubMed, EMBASE, and Web of Science databases, covering studies published from their inception to September 11, 2024. The search strategy was developed in consultation with a medical librarian to maximize both sensitivity and specificity in identifying relevant studies. The primary search terms included: (“compensatory sweating”). To ensure a broad yet targeted search, additional terms such as (“endoscopic thoracic sympathectomy” OR “hyperhidrosis surgery” OR “ETS”) and (“hyperhidrosis” OR “excessive sweating”) were incorporated. The search strategies for each database were as follows:

For PubMed, the search formula was: (“facial hyperhidrosis” OR “sweating disorder” OR “excessive sweating” OR “facial hyperhidrosis” OR “sweating disorder” OR “excessive sweating” OR “facial blushing” OR “flushing” OR “blushing” OR “flushing”) AND (“endoscopic thoracic sympathectomy” OR “thoracic sympathectomy” OR “ETS” OR “sympathetic surgery” OR “surgical treatment”). For EMBASE, the search formula was: ((‘facial hyperhidrosis’:ti,ab OR ‘sweating disorder’:ti,ab OR ‘excessive sweating’:ti,ab OR ‘facial blushing’:ti,ab OR ‘flushing’:ti,ab OR ‘blushing’:ti,ab) AND (‘endoscopic thoracic sympathectomy’:ti,ab OR ‘thoracic sympathectomy’:ti,ab OR ‘ETS’:ti,ab OR ‘sympathetic surgery’:ti,ab OR ‘surgical treatment’:ti,ab)). For Web of Science, the search formula was: (“facial hyperhidrosis” OR “sweating disorder” OR “excessive sweating” OR “facial hyperhidrosis” OR “sweating disorder” OR “excessive sweating” OR “facial blushing” OR “flushing” OR “blushing” OR “flushing”) AND (“endoscopic thoracic sympathectomy” OR “thoracic sympathectomy” OR “ETS” OR “sympathetic surgery” OR “surgical treatment”).

The references of relevant articles and systematic reviews were manually screened to identify any additional studies that met the inclusion criteria.

Study selection and eligibility criteria

The study selection process followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Two independent reviewers (Zhi-yi Lin and Hong Xiao) conducted a two-stage screening process: an initial review of titles and abstracts, followed by a full-text review of articles that met the preliminary inclusion criteria. Any disagreements between the reviewers were resolved through discussion, or by consulting a third reviewer (Min Lin) if consensus could not be reached.

Inclusion criteria: (a) studies involving participants diagnosed with hyperhidrosis who underwent ETS; (b) reported the number of patients with CH or SCH, evaluation of the relationship between any risk factors and CH or SCH; (c) study designs including cohort studies, case-control studies, cross-sectional surveys, or controlled trials.

Exclusion criteria: (a) non-human studies; (b) review articles, case reports, studies with fewer than 30 participants, and studies lacking complete data; (c) patients underwent endoscopic thoracic clipping or computed tomography guided percutaneous thoracic sympathetic radiofrequency thermocoagulation.

All citations were managed using EndNote X9 (Clarivate Analytics, Philadelphia, PA, USA) to facilitate the removal of duplicates and streamline the selection process. The selection process, including the number of studies identified, screened, excluded, and included, is detailed in a PRISMA flow diagram (Fig. 1).

Data extraction

Two independent reviewers (Zhi-yi Lin and Hong Xiao) extracted data using a standardized form, which was pilot-tested on a subset of studies. The extracted information included study characteristics (authors, year of publication, country, study design), the sample size, the number of patients with CH or SCH, any type of related risk factors (such as age, sex, smoking history, BMI), intervention characteristics (surgical technique, level of sympathectomy). Discrepancies were resolved through discussion, or with input from a third reviewer (Min Lin). The final dataset was cross-verified for accuracy and completeness. Detailed data extracted from these studies are presented in Tables 1 and 2.

Quality assessment

The quality of included studies was assessed independently by two reviewers using appropriate tools: For cohort and case-control studies, the Newcastle-Ottawa Scale (NOS) was applied. Quality assessments considered selection bias, exposure/outcome assessment, confounding, and statistical reporting. Disagreements were resolved through discussion or consultation with a third reviewer. Any study assessed as having a high risk of bias excluded from the final analysis. The quality of each study was taken into account when interpreting the results of the meta-analysis, and sensitivity analyses were conducted to test the robustness of conclusions based on the inclusion of lower-quality studies. The quality analysis results of the article are presented in Tables 1 and 2.

Statistical analysis

The meta-analysis was conducted using Review Manager version 5.3 (Cochrane Collaboration, Oxford, UK) and StataSE 16.0 (StataCorp, College Station, TX, USA). A random-effects model was employed to calculate event rates and their corresponding 95% confidence intervals (CIs), in order to account for potential heterogeneity across studies. Statistical heterogeneity was assessed using the I² statistic and Cochran’s Q test, with an I² value greater than 50% or a P-value less than 0.10 indicating significant heterogeneity. Additionally, subgroup analysis and meta-regression were performed to explore potential sources of heterogeneity, such as differences in patient demographics or surgical techniques. Publication bias was evaluated visually using funnel plots, and statistically using Egger’s test. Sensitivity analysis was carried out by sequentially excluding individual studies to test the robustness of the overall results, ensuring that the findings were not overly influenced by any single study. The study selection process adhered to the PRISMA guidelines. A PRISMA flow diagram was provided to illustrate the details of study identification, screening, eligibility, and inclusion.

Eligible studies

In our initial literature search, we identified 1,554 articles. After removing duplicates, screening titles and abstracts, and reviewing full texts, a total of 10 articles met the inclusion criteria and were included in this meta-analysis (Masarwa et al., 2024; Adhami & Bell, 2023; Alkosha et al., 2023; Woo et al., 2022; Carvalho et al., 2022; Toolabi et al., 2022; Moon et al., 2020; Kara et al., 2019; Leiderman et al., 2018; Rodríguez et al., 2008). Among these 10 articles, six studies focused on the factors influencing the occurrence of compensatory sweating after sympathectomy, while three studies investigated factors influencing severe compensatory sweating rates post-sympathectomy. The remaining one article addressed both aspects (Tables 1 and 2). Figure 1 illustrates the PRISMA flowchart in the meta-analysis paper.

Quality assessment and publication bias

Each article’s quality was assessed using the Newcastle-Ottawa Scale, with a median quality score of six points falling within a range of six to seven points. Publication bias was evaluated using Egger’s test. Among the seven articles investigating CH occurrence rates, Egger’s test results showed no evidence of publication bias (P > 0.05). However, for the four articles focusing on SCH occurrence rates, Egger’s test indicated the presence of publication bias (P < 0.05). Even after applying the trim-and-fill method, the P-value remained below 0.05, indicating the robustness of the results, which were unaffected even with the addition of virtual studies.

Occurrence of compensatory hyperhidrosis

A total of seven studies reported on 2,566 patients regarding the occurrence of compensatory hyperhidrosis. Due to high heterogeneity (I² = 97%; Fig. 2), the random effect model was used for analysis, and the results showed that the incidence of CH was 0.62 (95% CI [0.51–0.72]). After sensitivity analysis, no sources of heterogeneity were found. The subgroup analysis showed that the incidence of CH in studies from Asia was higher than that in other continents (0.74 vs. 0.46, P = 0.0007; Fig. S1), which may be one source of heterogeneity. Subgroup analysis based on the definition of CH revealed that the incidence of CH defined using the HDSS was significantly lower than that defined by subjective assessment, suggesting a potential source of heterogeneity (Fig. S2). However, subgroup analysis with varying follow-up durations did not reveal any significant differences between the groups (Fig. S3). We further conducted a meta-regression based on whether the study originated from Asia, different definitions of CH, and varying follow-up times, yielding p-values of 0.055, 0.041, and 0.284, respectively. These results similarly indicate that differing definitions of CH may contribute to discrepancies in the reported incidence of CH. We compared the clinical parameters of patients with CH to those without CH to determine if there were any differences.

We compared the ages of patients with CH and those without CH; all 2,566 patients had reported ages. The random effects model showed that patients with CH were older than those without CH, with a WMD of 1.92 (95% CI [0.74–3.11], p = 0.002; Fig. 3A), indicating that older patients may be more likely to develop CH after undergoing ETS. This difference was particularly significant in non-Asian populations (Fig. S4). However, subgroup analyses based on varying definitions of CH (Fig. S5) and different follow-up durations (Fig. S6) did not reveal any significant age differences between the groups. A total of 939 male patients and 1,628 female patients underwent ETS surgery, but no correlation was found between gender and the occurrence of CH (P = 0.82; Fig. 3B). Five articles reported on a total of 764 patients regarding family history, with 144 patients having a family history and 620 patients without; due to low heterogeneity (I² = 0%), a fixed effects model was used, and no differences were found in the occurrence of CH with or without a family history (P = 0.45; Fig. 3C). Sensitivity analysis for the above comparisons revealed no sources of heterogeneity.

Four studies investigated the smoking history of patients, with I² = 89% (Fig. S7). Sensitivity analysis indicated that one article (Alkosha et al., 2023) might be a source of heterogeneity, so after excluding this article, a meta-analysis was conducted on the remaining three studies, involving a total of 355 patients, with I² = 70%. Among these, 57 patients had a smoking history, while 298 did not. The random effects model suggested that the rate of CH occurrence after surgery was significantly higher in patients with a smoking history compared to those without (P < 0.0001; Fig. 3D).

Four studies reported BMI data for a total of 2,069 patients. Due to high heterogeneity (I² = 97%), a random effects model was used. The results indicated that patients with a higher BMI were more likely to develop CH after ETS, with a WMD of 2.07 (95% CI [0.96–3.18], p = 0.0003; Fig. 3E). We performed a subgroup analysis to determine whether the studies targeted Asian populations, but we did not find a source of heterogeneity (Fig. S8). Besides, subgroup analyses based on varying definitions of CH (Fig. S9) and different follow-up durations (Fig. S10) did not reveal any significant BMI differences between the groups. Only two studies reported the season in which surgery was performed for a total of 268 patients. The random effects model showed no significant difference in the incidence of CH between surgeries conducted in spring and summer compared to those in autumn and winter (P = 0.84; Fig. 3F). Four studies recorded whether 2,019 patients had plantar hyperhidrosis before surgery, and the random effects model indicated that the occurrence of CH was not related to plantar hyperhidrosis (P = 0.31; Fig. 3G). Sensitivity analysis for the above comparisons revealed no sources of heterogeneity.

Occurrence of severe compensatory hyperhidrosis

Four studies involving a total of 1,618 patients reported on the occurrence of severe CH. Due to high heterogeneity (I² = 96%), a random effects model was used for the analysis. The results indicated that the incidence of SCH was 0.23 (95% CI [0.12–0.34]; Fig. 4). Sensitivity analysis did not reveal any sources of heterogeneity. Subgroup analysis based on the study’s origin in Asia, varying definitions of SCH, and different follow-up durations revealed no significant differences between the subgroups (Fig. S11). We subsequently conducted a meta-regression considering the study’s origin in Asia, different definitions of SCH, and follow-up periods, yielding p-values of 0.712, 0.937, and 0.803, respectively, indicating that these three factors are not sources of heterogeneity.

All four studies recorded the age of patients. Due to high heterogeneity, a random effects model was used for the analysis. The results indicated that there was no significant statistical difference in age between patients with and without SCH, with a WMD of 3.16 (95% CI [−0.39–6.72], p = 0.08; Fig. 5A). However, subgroup analysis revealed a notable difference in patient age between studies from Asia and those from non-Asia, while no significant age differences were observed among patients defined by varying definitions of SCH or different follow-up durations (Fig. S12). In terms of gender, there were 669 male patients and 949 female patients, but there was no statistical relationship between the occurrence of SCH and gender (P = 0.11; Fig. 5B). Two articles reported data on family history from 488 patients, with 266 having a family history and 222 not having one. The fixed effects model showed no difference in the occurrence of CH based on family history (P = 0.25; Fig. 5C). Sensitivity analysis for the above comparisons did not reveal any sources of heterogeneity.

All four studies recorded the BMI of patients. Due to low heterogeneity (I² = 0), a fixed effects model was used, and it was found that patients with SCH had a higher BMI compared to those without SCH, with an SMD of 1.20 (95% CI [1.01–1.39], p < 0.0001; Fig. 5D). Subgroup analysis based on the study’s origin in Asia, differing definitions of SCH, and varying follow-up periods found no significant differences in BMI between the subgroups (Fig. S13).