Does endoscopic thoracic sympathectomy improve the quality of life of patients with primary hyperhidrosis? A single center retrospective review

Aims

The aims of this study are to evaluate the QOL of patients with PHH who had undergone ETS compared to their QOL pre-operatively, and to determine whether CS would affect the QOL of patients post-operatively.

Study design

This study was conducted from January to July 2019 at the National Heart Institute (IJN) in Kuala Lumpur which was a five-year retrospective review of the QOL of patients who underwent bilateral endoscopic thoracic sympathectomy (BETS) for the treatment of PHH at the National Heart Institute of Malaysia [Institut Jantung Negara (IJN)] from January 2014 to December 2018. With the approval of the IJN Research Ethics Committee (IJNREC), all patients’ medical records were screened to ensure the completion of data. A complete medical record would include the proper and adequate documentation of patient’s demographics, preoperative consultation record, medical and surgical record, laboratory results and postoperative follow-up notes. All eligible patients were contacted through phone calls for the administration of the questionnaire, which comprises of preoperative and postoperative scoring.

Inclusion criteria: bilateral endoscopic thoracic sympathectomy cases; completed medical records; consented for an interview.

Exclusion criteria: missing data from the medical record; medical condition that resulted in secondary hyperhidrosis; redo endoscopic thoracic sympathectomy; failure to give consent for an interview

This constituted a total of 62 patients who all had moderate to severe PHH prior to surgery, involving more than one location.

Ethical statement

Ethical approval was obtained from both the IJN Research Ethics Committee (IJNREC/207/2017) and Monash University Human Research Ethics Committee (MUHREC/9214). The study was also registered with the National Medical Research Register (NMRR-17-3133-39469). Verbal consent to participate in the study was obtained over the telephone. This method of consent was approved by the Research Ethics Committees, as set out in the study protocol submitted to them.

All research procedures were done in accordance with the ethical regulations set by the IJN Ethics Committee (IJNEC), Monash University Human Research Ethics Committee (MUHREC) and it abides with the Helsinki Declaration revised in 2013.

Surgical techniques

All patients were assessed for the severity of hyperhidrosis before surgery. Once the anaesthetic assessment is completed, elective surgery of BETS was scheduled. All BETS were performed by two different surgeons with the same level of experience and using similar techniques. All surgeries were performed in a standardized manner using the same equipment under general anaesthesia. None of the surgeries was converted into open surgery. In our institute, BETS was conducted in such a way that two incisions were made for the access of ports and thoracoscope. The first incision was made around the mid-axillary line over the fifth intercostal space. After deflating the lung of the operative side, the pleural space was entered by blunt dissection. Then, a second incision was made around the submammary fold under endoscopic guidance. For each side of the surgery, two ports of either 5mm or 10mm were used with a 30 degree thoracoscope. Once the ganglion was identified, the sympathetic chain would be interrupted using excision method. The same procedure was then repeated on the contralateral side at the same level of interruption in all cases. All specimens excised were sent for histopathology examination for the confirmation of ganglion excision. Patients were hospitalised for a total duration of three to four days to monitor for postoperative complications such as pneumothorax and wound management. All patients were subjected to chest X-rays which were reviewed by surgeons before discharge.

For this study, all surgical and inpatient notes were analysed and recorded including the total operative time, days of hospital stay, pain score, postoperative complications, and medications used.

Follow-up

All except one patient had at least one follow-up after surgery. About 63% of them attended follow-up twice and one of them had it thrice. During follow up, patients were assessed for symptom resolution, occurrence and severity of CS, pain, wound healing as well as other problems resulting from surgery.

Questionnaire

All questionnaires were administered during the period of January until July 2019. The questionnaire used for this study was the validated Health Related Quality of Life (HRQOL) Questionnaire in PHH which is a modified questionnaire based on studies by Milanez de Campos et al²⁵ and Keller and colleagues²⁴, which was designed to specifically evaluate preoperative and postoperative improvement of QOL in patients with hyperhidrosis. After obtaining the demographic data from the case records, the patients were interviewed via telephone on the severity of their PHH and also any possible recurrence, and on CS (Extended data: Appendix 1⁴⁶). This is followed by the second set of questionnaires that assessed the impact of QOL in four major domains, namely the functional, social, personal and psychological domains (Extedned data: Appendix 2⁴⁶).

The questionnaire starts with a single question of ‘In general, how would you rate your quality of life before and after treatment?”, in which patients will be asked to rate from 0 to 10, with 0 being the worst and 10 being excellent. This was followed by a total of 29 questions with 17 questions on functional domain, 3 questions on social domain, 2 for personal domain and 7 questions on psychological domain. Within the functional domain, the questions were further divided into 9 questions that focus on palmar sweating, 5 questions on feet sweating, and 3 questions on axillary sweating. The final total score of QOL can therefore possibly range from 0 to 290. In this study, QOL was assessed in three major forms that are: a) general QOL, b) QOL score as per domain and c) total QOL score of all domains, as illustrated in Figure 1.

Figure 1. Structure to the interpretation of the Health Related Quality of Life.

Furtermore, patients were asked to rate their overall satisfaction level from 0 being the worst, 5 being neutral, to 10 being excellent, and the score was further categorised into very unsatisfied (0-1), unsatisfied (2-4), neutral (5), satisfied (6-8) and completely satisfied (9-10). For the purpose of this study, there were additional open-ended questions after each section of the domain to address all potential issues that might be missed from the questionnaire. A good mix between closed-ended and limited number of open ended-questions allows for a smooth flow of interview while adequately addressing patients’ concern. All questions were asked in a simple and clear manner without medical jargons to avoid ambiguity or misunderstanding. At the end of the interview, we also asked if patients regretted the decision to undergo surgery.

Data collection

All complete patients’ medical records were reviewed by a single investigator to record relevant data. Data collection was performed by a single investigator throughout the study in order to reduce potential bias secondary to inconsistent administration of questionnnaire and data handling. Prior to interview, a separate reference list with only patients’ name and contact details was generated. This ensured the blinding of investigator from the level of interruption that patients had underwent, therefore reducing investigator bias. Before the interview, a template of conversation was created for a good conversation flow. This template included a standard introduction to the identity of investigator, purpose of the project, assurrance of patients’ confidentiality, and ended by obtaining patients’ informed consent. In addition, pilot tests were run with ten volunteers to gain feedback to improve the process of interview. These were patients from our previous cohort of study on ETS¹⁷ and no changes were made to the questionnaire after the pilot study.

Prior to the interview, patients were asked to choose their language of preference, between Malay or English for better communication. This was done on a background of an investigator who is fluent in both languages. Good rapport was built by asking daily questions to make patients feel more at ease in expressing themselves. In order to reduce performance bias, all patients were reassured that there will be no coercion on their response and honest feedback is greatly appreciated for the benefit of scientific research. To reduce recall bias, all patients were informed at the beginning of the phone call that the interview would take an average of 30 minutes; hence, appointment will be made at another time if patients were occupied at that particular moment. Besides, for each of the questions asked, patients were given adequate time to recall without any prompts. This was strictly adhered to across all interviews to minimise recall bias. When asked to rate the QOL from a scale of 0 to 10 with regards to hyperhidrosis, a standard explanation was given to all patients that was: 0 represents the worst QOL you could imagine, while 10 represents excellent QOL, and 5 being a neutral response.

Statistical analysis

Data entry and analysis were performed using IBM® SPSS Statistics version 25.0. Descriptive statistics were reported as percentages for discrete variables, continuous data were reported as means with standard deviations (SDs) for parametric data and medians with interquartile ranges (IQRs) for non-parametric data. Internal consistency of questions was tested separately for each domains using Cronbach’s Alpha reliability test. Questions from each subset of functional domain and psychological domain demonstrate excellent internal consistency. Personal domain demonstrates poor consistency and hence may impose a limitation to our analysis (Table 1). Tested variables were first explored for its normality, outliers and linear relationship and those which met the assumptions will be analysed using parametric analysis whereas those that did not will be analysed using non-parametric analysis. Correlations between factors affecting the occurrence of CS were assessed using Spearman’s rho or Pearson’s correlation. Coefficient of more than 0.7 will be considered significant. Differences between categorical variables were analysed using Fisher’s exact test or chi-squared test, depending on the assumptions. Differences for independent continuous variables were analysed using independent t-test or Mann-Whitney test for parametric and non-parametric data, respectively. The differences in QOL score rated by patients before and after operation were compared using the paired t-tests for normally distributed data, or Wilcoxon signed-rank test for skewed data. As the improvement of QOL in certain domains was not normally distributed, all comparisons of QOL improvement in each domains were done using the Wilcoxon test. The value of significance was taken at p<0.05 for all the analysis.

Number of cases included

A total of 62 patients were included based on the criteria of complete medical records and having BETS done between January 2014 and December 2018. One patient was excluded for having undergone the surgery twice. All relevant medical records were screened thoroughly to ensure that there was no underlying medical condition that may be a confounder to secondary hyperhidrosis such as anxiety disorder, diabetes, or hyperthyroidism. The final diagnosis of PHH in each patient was confirmed through consultation record after the exclusion of possible secondary causes of hyperhidrosis. Thirteen patients were lost to contact, and two patients did not consent to the phone interview for personal reasons. Finally, 46 patients (response rate: 74.2%) consented to the telephone interview, and all of them completed the HRQOL questionnaire (Figure 2).

Figure 2. Flow chart demonstrating the process of patients’ recruitment

Demographics

The studied population comprised of 30 (65.2%) and 16 female (34.8%) patients. In total, 36 of patients were ethnic Malays (84.8%), 6 were Chinese (13%) and 1 was Indian (2.2%), which appropriately represented the racial distribution in Malaysia. None of the patients had any medical conditions that might predispose them to secondary hyperhidrosis. Blood tests were done on all patients to ensure there were no cases of thyroid dysfunction, or diabetes. Upon first consultation, no patient reported any anxiety disorder which might worsen the severity of sweating.

The age of the study population ranged from 13 to 55 years old, with 84.8% of the study population being younger than 30 years old. The distribution was positively skewed with a median of 20.5 years old. The BMI of the study population is normally distributed with a mean of 23.02 kg/m² (SD=4.04). When classified into categories, 1 (2.2%) patient was underweight, 25 (54.3%) were normal, 16 (34.8%) were overweight and 4 (8.7%) were obese (Table 2).

Hyperhidrosis profile

Location of hyperhidrosis

Only two patients had sweating from a single location (4.3%), as most of them had hyperhidrosis in two locations (73.9%), with the remaining reported PHH involving more than two locations (21.7%). Combined palmar-plantar hyperhidrosis was present in 73.9% of the study population, with the rest distributed in all different possible combinations, as summarised in Table 3. Broken down to each location and its individual reported frequencies, the palm was the most commonly reported location by all patients (100%), followed by the soles of the feet [43 (93.5%) patients].

Severity

All patients experienced moderate to severe sweating from the palms and soles, which prompted them to seek surgical treatment. Reportedly, 45.7% had moderate palmar hyperhidrosis while 54.3% had severe condition. None had mild sweating prior to surgery. Some described dribbling of sweat from hands and feet, which embarrassed and impaired their daily activities. The majority of patients described worsening of sweating with hot weather and relieved by staying in cool condition. A small proportion of patients described worsening of sweating by stress. In some cases, sweating was persistent even under cool conditions, leading to a very poor QOL before operation.

Level of sympathectomy

In our centre, majority of BETS were performed at a level of T2-T3 (78.3%), the frequency reported is summarised in Table 3. A T2 level has been well recognised to increase the risk of CS and other complications if interrupted^26,27. Hence, we further categorise patients into T2-involved and T2-spared group to study the difference between both groups.

Occurrence of compensatory sweating

Although the definition of CS remains vague, it is essentially a condition of increased sweating at a previously normal location, to an extent of being noticed by patients. In our study, CS was reported in up to 41 (89.1%) of patients. We further assessed these patients for the location(s) involved and its severity. Data on the onset of CS first becoming noticeable was also recorded. When assessing the locations involved in CS, every patient was asked regarding the presence of CS at locations such as the face, axilla, trunk, abdomen or groin, and lower limbs. They were then grouped into patients who had CS from one location, two locations and more than two locations. In total, 38 (92.7%) patients reported having CS on the body, followed by lower limbs (48.8%). CS at the abdominal and groin region was reported by 10 (24.4%) patients, and axillary region by 4 (9.8%) patients. We noticed that 64.3% of patients who experienced CS first noticed the sweating within the first month of the surgery. Among these, 31.7% noticed it as early as during the first week of surgery. Less than a quarter (13.0%) noticed CS more than six-months after the surgery. Patients with CS were also asked to rate the severity of sweating as mild, moderate or severe according to the pre-set standard definition (Table 4). The majority of them had moderate CS (41.5%) or severe CS (43.9%). Only 14.6% of them described mild sweating and this information is summarised in Table 4.

Evolution of compensatory sweating

The majority of patients reported no improvement in the intensity of CS. Only four (9.8%) patients reported complete resolution of CS over time, and three (7.3%) reported gradual improvement with time. In those who claimed that CS had resolved, one described having CS on their trunk for around seven-months to a year, before it progressively resolved. The other two patients descibed having CS for more than a year which slowly improved and then resolved. There was a patient who described having CS at multiple locations including the trunk, abdomen and groin, and thigh but CS from the thigh had resolved after it occurred for a week, even though CS from other locations remained.

Patient satisfaction

In general, there was great satisfaction reported by patients with a median score of 7 (IQR=4). There was an ascending trend of frequency from the worst to excellent as shown below (Table 5). In this study, we took the value of 5 as neutral as explained to all patients before the administration of questionnaire.

Compensatiry sweating and patient satisfaction

We did not find any significant difference in satisfaction between the group of patients with CS and without CS (p=0.219) (Table 6).

We noticed that comparing patients with mild to moderate (non-severe) CS with those who had severe CS, the satisfaction reported was highly and significantly lower in the latter group (p=0.001) (Table 7).

Quality of life Improvement

General QOL

In our study, 37 (80.4%) patients reported improvement in the general QOL after surgery, whereas 5 (10.9%) reported no changes with general QOL and 4 (8.7%) reported worse QOL after surgery. All four patients who reported worse general QOL after surgery had severe CS involving more than one location. It was demonstrated in this study that the improvement in the general QOL of patients after surgery was highly statistically significant (p<0.001) (Table 8).

QOL improvement by domain

For comparison of QOL improvement in each of the four domains, the mean changes in score were adjusted to give an average improvement of score per question. It can be observed in Figure 3 below that all domains have demonstrated an increment in the median score for postoperative QOL, although it was not prominent in the axilla (p=0.087). Besides, the functional domain of hands had shown the greatest improvement as compared to other domains with average increment of 5.86 per question (p<0.001). This was followed by the social domain, which also showed substantial improvement with an average increment of 5.12 per question (p<0.001).

Figure 3. Boxplot illustrating the summation of quality of life score from each domain for both preoperative and postoperative period

Summary of improvement in QOL

Table 9 summarizes the preoperative and postoperative median score for each question with its respective p value. It was noted that all domains had significant improvement in QOL score except for the axillary domain.

Table 10 summarizes the median score for each domain and its respective average increment in score per question and p values. This is followed by Table 11 that describes the change in QOL in each domain.

Total sum of QOL

We also summed up the total score from all domains to compare the changes before and after surgery. Of all patients, only two (4.3%) had a reduction in the total QOL score, whereby the rest of them (95.7%) reported improvement in total QOL. By comparing the differences between the preoperative and postoperative total QOL score using paired t-test, there was a statistically significant difference (p<0.001), with a mean difference of 113.54 (SD=70.79, 95% CI=95.52 - 134.57) (Table 12).

Factors affecting QOL

To study the factors associated with QOL after surgery, selected variables were initially tested for their correlation with general and total QOL score using Spearman’s and Pearson’s correlation. It was noted that the severity of CS and number of location(s) involved in CS were significantly associated with both general and total QOL (Table 13).

It is interesting to note that when comparing the postoperative general QOL and total QOL between patients with CS and without CS, no significant difference was noted (p=0.303 and p=0.167 respectively) (Table 14). However, when we compared the general QOL reported by patients with non-severe CS and those with severe CS, there was a significantly lower general QOL reported by the latter (p<0.001) (Table 15). We also performed an analysis between the severity of CS with the total QOL score. By running an independent t-test, the results were consistent with the above analysis in which the mean total QOL score was noted to be significantly lower in the group with severe CS (Mean=217.17 vs 184.22, p=0.008) (Table 15).

Post-hoc Bonferroni was performed to improve the accuracy of such comparison. It was found that there was still a significant association between two factors (p=0.022) with a significant difference in mean total QOL reported between patients with moderate and severe CS (p=0.019) (Table 16).

Another significant finding is with regard to the number of locations of CS. By using Kruskal-Wallis test, we noticed a significantly lower general QOL reported in patients with CS involving more than two locations, as compared with those involving one location or two locations (p=0.008). Post-hoc Dunn test had demonstrated, in particular, a significant difference in median score between patients with one location and more than two locations (p=0.024) (Table 17). As illustrated in Figure 4, we can see an obvious downward trend of the median general QOL score from ‘one location’ box to ‘two locations’ and to ‘more than two locations’.

Figure 4. Simple boxplot illustrating the median of general QOL score postoperative as categorised by number of location(s) of CS.

Similarly, we also ran a test to study the association between number of CS location(s) and total QOL. By using individual t-test, patients with more than two locations of CS reported significantly lower total QOL score as compared to those with one location of CS, with a mean difference of 39.48 (95% CI=4.44 - 74.52, p=0.029) (Table 18).

Limitations

Given the relative rarity of this surgery with less than 15 cases performed per year at our institute, we believe a longer duration of such retrospective study is more effective in capturing a higher number of patients. However, there were a few limitations in this study. First of all, the study design requires inclusion of telephonic follow-up. As with any research that involves telephone follow-up, reaching all participants was a difficult and tedious process because of changed or disconnected numbers and hence a few patients were lost. In addition, since the data was pre-recorded, there may be inaccuracy and inadequacy with regards to the history and examination notes. Furthermore, as data was collected from a single centre, this may potentially introduce selection bias. In addition, as the QOL questionnaire was applied postoperatively, there was also a risk of recall bias. We have taken a few steps, as mentioned in Methods, that were carried out and strictly adhered in order to reduce recall bias. Additionally, Cronbach alpha reliability test was run and showed poor internal consistency for the questions in the personal domain, which might affect the analysis that involve total QOL score. Finally, as this study also looked into the occurrence of CS and factors affecting it, a small sample size in the group of patients without CS has limited our analysis related to the occurrence of CS.

Extended data

Harvard Dataverse. Replication Data for: Does endoscopic thoracic sympathectomy improve the quality of life of patients with primary hyperhidrosis? A single center retrospective review, https://doi.org/10.7910/DVN/REQ7Y7⁴⁶

This project contains the following extended data:

Data are available under the terms of the Creative Commons Zero “No rights reserved” data waiver (CC0 1.0 Public domain dedication).