Er:YAG laser in hysterectomized women with stress urinary incontinence: a VELA retrospective cohort, non-inferiority study

Abstract

Objective: To test whether the erbium-doped yttrium aluminum garnet (Er:YAG) SMOOTH^® laser treatment efficacy on stress urinary incontinence (SUI) in hysterectomized patients is non-inferior to its efficacy in non-hysterectomized patients.

Methods: In this real-world, retrospective cohort study performed in Turkey, Croatia and Italy, we enrolled a consecutive sample of 35 hysterectomized and 34 non-hysterectomized patients with SUI. We used the Er:YAG SMOOTH^® laser (Fotona, Slovenia) with a wave length of 2940 nm. The primary outcome was median reduction of SUI symptoms measured by the International Consultation on Incontinence Questionnaire-Urinary Incontinence Short-Form (ICIQ-SF) with the non-inferiority margin defined as the minimum clinically important difference of ICIQ-SF (δ < 2.52 points).

Results: In hysterectomized patients, the ICIQ-SF was reduced by 5 points (95% confidence interval 3–8; p < 0.001), a reduction of 45% (95% confidence interval 36–67%). After adjustment for baseline ICIQ-SF and five covariates, the reduction of symptoms in the hysterectomized group was not inferior to the reduction in the non-hysterectomized group.

Conclusion: The Er:YAG SMOOTH^® laser treatment seems to improve the symptoms of SUI in hysterectomized women not clinically relevantly less than in non-hysterectomized women. It seems that the beneficial effect of Er:YAG SMOOTH^® laser treatment for SUI in hysterectomized women is time-limited.

摘要

目的：为了测定掺铒钇铝石榴石（Er：YAG）SMOOTHV^®激光治疗对子宫切除术患者压力性尿失禁（SUI）的疗效是否不次于其对子宫未切除患者的疗效。

方法：在土耳其、克罗地亚和意大利进行的这项真实世界回顾性队列研究中, 我们连续纳入了35例子宫切除术的SUI患者和34例子宫未切除的SUI患者。我们采用了波长为2940 nm的Er：YAG SMOOTHV^®激光器（Fotona, Slovenia）。主要结局是根据国际尿失禁咨询委员会尿失禁问卷简表（ICIQ-SF）测得的SUI症状中位数降低, 非劣效性界值定义为ICIQ-SF的最小临床重要差异（ẟ<2.52分）。

结果：接受子宫切除术的患者中, ICIQ-SF降低了5分（95％置信区间3-8；p <0.001）, 降低了45％（95％置信区间36-67％）。在对基线ICIQ-SF和五个协变量进行调整后, 子宫切除术组的症状减轻不次于非子宫切除组的症状减轻。

结论：Er：YAG SMOOTHV^®激光治疗似乎可以改善子宫切除术女性的SUI症状, 但在临床上并没有比非子宫切除的女性少。似乎, Er：YAG SMOOTHV^®激光治疗对子宫切除后女性SUI的有益作用是有时间限制的。

Keywords:

• Stress urinary incontinence
• hysterectomy
• Er:YAG SMOOTH^® laser

Introduction

Hysterectomy is one of the commonest gynecological surgeries. The incidence of urinary incontinence after hysterectomy is reported roughly as 10%¹. It is a controversial issue whether hysterectomy itself may cause or improve the symptoms of urinary incontinence. For example, the mean value of Valsalva leak point pressure was significantly lower in the group of patients who had a positive hysterectomy history, and patients in the hysterectomy group were significantly 6.3 times more at risk of severe stress urinary incontinence (SUI)^2,3. Another debatable subject arises from the type of hysterectomy (total vs. subtotal, abdominal vs. vaginal, open vs. laparoscopic) versus type of induced urinary incontinence^4,5. SUI is usually defined as involuntary urinary leakage by physical efforts⁶ and can be treated by both non-invasive and invasive procedures⁷. Although the patient’s adherence and the success rate are quite low in non-invasive procedures such as pelvic floor muscle exercise⁸, invasive procedures such as with transobturator tape and tension-free vaginal tape have quite satisfying success rates, but involve some serious complications^9,10. At the same time, urethral bulking agents are now an established alternative to mid-urethral slings, with a lower complication rate.

Laser energy induces neocollagenesis, elastogenesis and neoangiogenesis in the vaginal mucosa, supporting a destabilized urethra and bladder neck^11,12. Most of studies regarding laser treatment of SUI have been performed using the erbium-doped yttrium aluminum garnet (Er:YAG) SMOOTH^® system. Its efficacy and safety have been demonstrated by a randomized controlled trial¹³ and a number of observational studies^14–22. The Vaginal Er:YAG SMOOTH^® Laser Academy (VELA) is a recently founded international society focusing on the efficacy and safety issues of Er:YAG SMOOTH^® laser treatments, not only in SUI patients but also those with other genital disorders.

In the first collaborative study by VELA, our objective was to test whether the efficacy of the Er:YAG SMOOTH^® laser treatment on SUI, measured by the International Consultation on Incontinence Questionnaire-Urinary Incontinence Short-Form (ICIQ-SF), in hysterectomized patients is non-inferior to its efficacy in non-hysterectomized patients, with the non-inferiority margin defined as the minimum clinically important difference of ICIQ-SF (δ < 2.52 points)²³, and after adjustment for the baseline ICIQ-SF score, age, body mass index (BMI), parity, number of sessions, and duration of follow-up. Our hypothesis was that the Er:YAG SMOOTH^® laser treatment is not clinically relevantly less effective in hysterectomized then in non-hysterectomized patients suffering from SUI.

Methods

Study design

This multicenter, real-world, observational, retrospective cohort study was conducted between April 2018 and May 2019. The patients were enrolled in three obstetric/gynecology clinics: Istanbul Cerrahpaşa University, Istanbul, Turkey; Institute for Women’s Health, Zagreb, Croatia; and Pisa University Hospital, Pisa, Italy. After the intervention, participants were followed at 6-month intervals. The procedure was explained to all participants in detail, and they all signed the informed consent. All three centers are members of the VELA Society which defined the criteria, procedures, the common informed consent form and instruments for measuring outcomes in routine application of the Er:YAG SMOOTH^® laser treatment. Numeric codes were assigned to the participants before the analysis, so that patients’ identities were known only to their physicians. The study was conducted in accordance with the Declaration of Helsinki of 1975 as revised in 2013²⁴. The Institutional Ethics Committee and Reviewer Board from the Istanbul Cerrahpaşa University authorized the protocol (# 52825153-604.01.01-100626). This observational study was not registered at the public repository.

Target population

The target population was composed of 35 hysterectomized and 34 non-hysterectomized women treated with the Er:YAG SMOOTH^® laser for SUI. All the patients were submitted to pretreatment evaluation. Diagnosis of SUI was established on the basis of medical history, pelvic examination at rest and during a Valsalva manouver plus cough stress test. All the patients completed the ICIQ-SF. Patients were informed about the laser procedure and its potential adverse effects. All the interventions were performed in outpatient departments. During the intervention, patients were lying in a normal gynecological lithotomy position and received no anesthesia.

The exclusion criteria were: any neurologic disease, insulin-dependent diabetes mellitus, active urinary tract infection, hematuria, and undiagnosed vaginal bleeding. We enrolled the total available population by the order of patients’ arrival for the treatment.

Outcomes

The primary outcome

The primary outcome was the median reduction in SUI symptoms measured by the ICIQ-SF and adjusted for the baseline ICIQ-SF score, age, BMI, parity, number of sessions, and duration of follow-up. We calculated the reduction of symptoms from the baseline ICIQ-SF score to the lowest ICIQ-SF score achieved during the entire follow-up after the intervention. We defined the non-inferiority margin as the minimum clinically important difference of ICIQ-SF (δ < 2.52 points) as determined by Nyström et al.²³. The ICIQ-UI Short Form evaluates the self-reported and subjectively assessed frequency and severity of urinary incontinence and its impact on quality of life. Participants completed questionnaires before and after the intervention, and at check-up examinations every 6 months thereafter.

Secondary outcomes

The secondary outcomes were (1) percentage change in ICIQ-SF score after the intervention relative to the baseline score; (2) ‘maximum improvement time’ (MIT) defined as the patients’ reported point in the time-line just before the efficacy of the procedure started to decrease, with the time of intervention as the reference point; (3) ‘total improvement time’ (TIT), defined as the point in the time-line when the efficacy disappeared and returned to the state before the procedure, with the time of intervention as the reference point; (4) the percentage of patients experiencing treatment success, defined as ICIQ-SF ≤ 6, which indicates a successful patient-reported outcome on the Patient Global Impression of Improvement following surgical treatment of SUI²⁵.

Confounders and other variables

The potential confounders that we controlled as covariates in multivariable analysis of the differences between hysterectomized and non-hysterectomized women were age, parity and BMI. Additional potential effect-modifying variables that we controlled as covariates were number of Er:YAG SMOOTH^® laser sessions. Other variables used to describe the samples from the target populations were menopausal status, years in menopause, and type of hysterectomy. Pelvic examination, vaginal smear test and ultrasound were performed in all the patients.

Intervention

The Er:YAG SMOOTH^® laser with a wavelength of 2940 nm (SP Dynamis, Fotona, Slovenia) was used with the parameters previously described^17,22. The IncontiLase^® intravaginal application of ‘smooth’ mode was used with fluence of 10.0 J/cm²; four pulses per point; pulse duration of 250 ms; spot size 7 mm, and frequency of 1.6 Hz. The laser sessions were performed at monthly intervals. The total number of sessions was decided by the physicians in accordance with the manufacturer’s protocol, which recommends two sessions, but physicians were free to adapt the number of sessions according to the vaginal length and width, and severity of atrophic vaginitis. No anesthetics were used.

Statistical analysis

In the main analysis, we claimed non-inferiority if the upper limit of the one-sided 95% confidence interval (CI) of the median reduction of SUI in the hysterectomized group was within the margin of δ < 2.52 ICIQ-SF points from the median reduction in the non-hysterectomized group, after the adjustment for baseline ICIQ-SF score, age, BMI, parity, number of sessions, and the duration of follow-up. Introductory analysis of efficacy was performed by the Wilcoxon matched-pairs, signed-rank test within each study group. For the standardized effect size for the change within each group, we calculated r as r = z/(sqrt(2 × n) where z was the standardized Wilcoxon test statistic and n was the sample size. We calculated crude and adjusted medians (interquartile range; IQR) ICIQ-SF before and after the intervention, median reduction and median percentage change of SUI symptoms, median (IQR) of MIT and TIT, and the percentage of patients achieving treatment success, defined as the ICIQ-SF ≤ 6²⁵. We used the multivariable quantile and Poisson regression only to adjust the medians and CIs for the baseline ICIQ-SF score, age, BMI, parity, number of sessions and the duration of follow-up. Except for the CI, we did not statistically compare the two groups on secondary outcomes because we did not have a valid and reliable non-inferiority margin for these outcomes. We used Klovning’s stratification²⁶ of urinary incontinence only for descriptive purposes: very severe (ICIQ-SF score 19–21), severe (13–18), moderate 6–12, mild 1–5, and as no leakage if it was 0, but in the analysis we used the original ICIQ-SF values. No data were missing in any of the outcomes or potential confounders. We controlled the false-positive rate using the Benjamini–Hocberg procedure with the false-discovery rate (FDR) set in advance at < 10%. Statistical significance was defined as p < 0.05; CIs were calculated at 95%, one-sided in the non-inferiority test, and two-sided in all other cases. Statistical analysis was performed using the StataCorp. 2019 (Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC).

Results

We enrolled 35 previously hysterectomized women with SUI and 34 women with an intact uterus but having clinically relevant symptoms of SUI. The median (IQR) age of hysterectomized women was 62 (53–66) years with 15/32 (43%) having a severe or very severe baseline SUI (Table 1). Hysterectomized women were markedly older, more often postmenopausal, more often treated with three sessions, and followed up for a longer time (Table 1). In other characteristics, the two cohorts were comparable.

Table 1. Characteristics of 35 hysterectomized and 34 non-hysterectomized women.

	Hysterectomized (n = 35)	Non-hysterectomized (n = 34)
Age (years), median (IQR)	62 (53–66)	50 (45–55)
Parity, median (IQR)	2 (1–2)	2 (1–2)
Categorized parity, n (%)
0	2 (6)	2 (6)
1	15 (44)	14 (40)
2	11 (32)	12 (34)
≥ 3	6 (18)	7 (20)
Years since menopause, median (IQR)	12 (6–17)	1 (0–7)
Menopausal status, n (%)
premenopausal	1 (3)	16 (47)
postmenopausal	34 (97)	18 (53)
BMI (kg/m^2), median (IQR)	26 (24–28)	25 (22–28)
Categorized BMI, n (%)
normal (< 24.9)	15 (43)	17 (50)
overweight (25.0–29.9)	19 (54)	15 (44)
obese (≥ 30.0)	1 (3)	2 (6)
Type of hysterectomy, n (%)
none	0 (0)	34 (100)
TAH + BSO	28 (80)	0 (0.0)
TAH	5 (14)	0 (0.0)
VTH	1 (3)	0 (0.0)
STH + BSO	1 (3)	0 (0.0)
Number of laser sessions, median (IQR)	3 (2–3)	2 (2–3)
Number of laser sessions, n (%)
1	4 (11)	5 (15)
2	7 (20)	18 (53)
3	24 (69)	8 (24)
4	0 (0)	3 (9)
Follow-up (months), median (IQR)	24 (12–24)	17 (10–27)
ICIQ-SF at enrollment	12 (11–15)	14 (9–16)
Categorized ICIQ-SF at enrollment
mild (1–5)	2 (6)	3 (9)
moderate (6–12)	18 (51)	11 (32)
severe (13–18)	13 (37)	15 (44)
very severe (19–21)	2 (6)	5 (15)

Data are presented as number (percentage) of patients if not stated otherwise.

IQR, interquartile range; UI, urinary incontinence; SUI, stress UI; MUI, mixed UI; TAH, total abdominal hysterectomy; BSO, bilateral salpingo-oophorectomy; VTH, vaginal total hysterectomy; BMI, body mass index; ICIQ-SF, International Consultation on Incontinence Questionnaire-Urinary Incontinence Short-Form.

In the sample of hysterectomized women, the ICIQ-SF score was significantly lowered, from the median (IQR) of 12 (11–15) at baseline to 6 (4–7) points after the treatment (Wilcoxon matched-pairs signed-rank test, z = 5.06; p < 0.001; r = 0.60; FDR < 10%) (Figure 1 and Table 2). This median (95% CI) ICIQ lowering by 5 (3–8) points represented a 45% (95% CI 36–67%) decrease from the baseline values. It was above the minimum clinically relevant difference of 2.52 ICIQ-SF points. The median (IQR) MIT was 12 (6–18) months, and TIT was 18 (6–24) months. Overall, 18/35 (51%; 95% CI 34–69%) patients achieved treatment success defined as ICIQ-SF ≤ 6 points on ICIQ-SF.

Figure 1. The median ICIQ-SF scores in hysterectomized and non-hysterectomized women before and after the Er:YAG SMOOTH^® treatment. ICIQ-SF, International Consultation on Incontinence Questionnaire-Urinary Incontinence Short-Form.

Table 2. Treatment outcomes.

	Hysterectomized (n = 35)					Non-hysterectomized (n = 34)
	Baseline	After the intervention		Δ (95% CI)	Δ% (95% CI)	Baseline	After the intervention	Δ (95% CI)	Δ% (95% CI)
Unadjusted analysis
Primary outcome
ICIQ-SF	12 (11–15)		6 (4–7)	−5 (−8, −3)	−45 (−67, −36)	14 (9–16)	4 (1–10)	−7 (−10, −3)	−64 (−88, −35)
Secondary outcomes
MIT (months)			12 (6–18)				12 (6–17)
TIT (months)			18 (6–24)				13 (6–17)
Treatment success^a, n (%)			18 (51)				22 (65)
Adjusted analysis^b
Primary outcome
ICIQ-SF	12 (11, 13)		7 (5, 8)	−6 (−7, −5)	−47 (−59, −35)	14 (13, 14)	5 (4, 7)	−7 (−9, −6)	−65 (−78, −53)
Secondary outcomes
MIT (months)			10 (8, 12)				14 (12–16)	18 (16, 20)
TIT (months)			15 (13, 17)				17 (16, 20)
Treatment success^a, % (95% CI)			52 (34, 70)				64 (46, 81)

Data are presented as median (IQR) in the unadjusted, and median (two-sided 95% CI) in adjusted analysis if not stated otherwise.

ICIQ-SF, International Consultation on Incontinence Questionnaire-Urinary Incontinence Short-Form; IQR, interquartile range; MIT, maximum improvement time; TIT, total improvement time; Δ, median of absolute differences from baseline to after the intervention; CI, confidence interval; Δ%, median of relative differences from baseline to after the intervention relative to the baseline value.

^aTreatment success was defined as ICIQ-SF ≤ 6 points; ^bAdjusted for baseline ICIQ-SF score, age, body mass index, parity, number of sessions, and the duration of follow-up using the quantile regression for ICIQ-SF, MIT and TIT, and Poisson regression for treatment success.

The lowering of ICIQ-SF was significant and clinically relevant in the non-hysterectomized women as well (Wilcoxon matched-pairs signed-rank test, z = 4.93; p < 0.001; r = 0.60; FDR < 10%) (Table 2). The median (95% CI) ICIQ-SF lowering by 7 (3–10) points represented a 64% (95% CI 35–88%) decrease from the baseline values (Figure 1). The median (IQR) MIT was 12 (6–17) months, and TIT was 13 (6–17) months. Overall, 22/34 (65%; 95% CI 46–80%) patients achieved treatment success defined as ICIQ-SF ≤ 6 points.

After the adjustment for the baseline ICIQ-SF, age, BMI, parity, number of sessions and follow-up time, the median (95% CI) ICIQ-SF reduction after the intervention was −7 (−9, −6) in the non-hysterectomized group (Table 2). The margin of non-inferiority was a ICIQ-SF reduction of −4 ICIQ-SF points, calculated as the ICIQ-SF reduction in the non-hysterectomized group (Δ = −7) plus the non-inferiority margin (δ = 2.52). In the hysterectomized group, the adjusted reduction of ICIQ-SF was −6 (95% CI −7, −5) (Table 2), and the one-sided 95% CI was −5 as well. Therefore, the upper limit of the one-sided 95% CI of the median ICIQ-SF reduction in the hysterectomized group was within the margin of non-inferiority.

Discussion

Our study supported the hypothesis of non-inferiority of intravaginal Er:YAG SMOOTH^® laser treatment efficacy on the symptoms of SUI in hysterectomized women compared to its already proven efficacy in non-hysterectomized patients. Although many studies have confirmed Er:YAG SMOOTH^® laser efficacy in the treatment of SUI^13–22, as far as we are aware there are no published data on laser treatment of SUI in hysterectomized women.

There might be some important consequences of hysterectomy which might prevent the beneficial effect of laser treatment for SUI symptoms^3,4,27. First, section of sacrouterine and cardinal ligaments, in the course of total hysterectomy, may destabilize the pelvic floor. Second, the length of vagina is reduced, and third, the pelvic blood and nerve supplies to the bladder become impaired. Additionally, these factors could deteriorate sphincter function, resulting in insufficient improvement in hysterectomized patients treated with laser.

Since the ‘smooth’ mode of the Er:YAG laser is non-ablative and increases the local temperature to the stage that initiates neocollagenosis, neofibrogenesis and neoangiogenesis in the connective tissue of the vaginal mucosa^11,12, it may be expected that the scar tissue due to the hysterectomy may weaken the response to the Er:YAG SMOOTH^® laser. In contrast, Engin et al. proposed that non-ablative Er:YAG SMOOTH^® laser treatment improved the scar tissue and increased the collagen content in acne scar tissue²⁸ beyond the results of the other dermatological studies²⁹.

Also, in the treatment of burn scars, non-ablative fractional erbium lasers create columns of coagulated tissue composed of denatured collagen while leaving the epidermal layer intact³⁰. Taudorf et al. showed that non-ablative fractional laser provides long-term improvement of mature burn scars. Their post-treatment histology study showed an interwoven collagen with an overall increased similarity to normal unaffected skin³¹. In accordance with these findings, the Er:YAG SMOOTH^® laser may improve both the quality and the quantity of scar tissue affecting the upper part of the vagina after hysterectomy.

Our findings are in the accordance with the previously published data in non-hysterectomized women^16–20. Fistonic and Fistonic clearly showed that age, parity, BMI, and severity of SUI symptoms before the treatment were predictive factors for the efficacy of the Er:YAG laser treatment in non-hysterectomized women with SUI²⁰. Similarly, in a larger group of patients with urinary incontinence, it was shown that women with 3.6 kg/m² less value in BMI showed better results from Er:YAG SMOOTH^® treatment for SUI²². Our study was not adequately powered for the subgroup analysis of the effect of BMI in the hysterectomized patients, and future studies should address this issue because higher BMI increases the risk of SUI after hysterectomy³.

Results of this VELA collaborative study further indicate that more laser sessions are associated with a better response. Most of the studies, in non-hysterectomized patients, recommend two or three sessions at a month apart^13–21. Our previous study found that the higher pulse number in the total sessions per patient was associated with better improvement²², so the total laser energy expenditure during the sessions may also be a predictive parameter for the success of Er:YAG laser treatment of urinary incontinence (p = 0.059). MIT and TIT were significantly longer in the patients in the high-improvement group.

It is reasonable to compare this finding, as more energy stored in the vaginal mucosa induces a better effect.

Although the duration of improvement after the Er:YAG SMOOTH^® laser treatment on genitourinary symptoms of menopause was claimed to continue up to 18 months, or even to 24 months^29,32, the patients with SUI were followed up for a relatively short period of time in most of the studies^13–21. In our study, previously hysterectomized patients were followed up for a median of 24 months, and the median time to maximum reduction in ICIQ-SF was around 12 months, while the duration of the response was around 18 months in hysterectomized and 13 months in non-hysterectomized patients after the completion of Er:YAG SMOOTH^® laser sessions.

However, the beneficial effect of Er:YAG SMOOTH^® laser treatment on SUI is of limited duration. Therefore, during patient counseling, detailed information about the efficacy and safety of laser treatment for SUI in hysterectomized patients should be given in order to prevent unrealistic expectations. Periodic examinations and evaluations would reveal the need for further laser treatments.

There are several limitations of our study. The main limitation is a lack of randomized control groups not treated with the Er:YAG SMOOTH^® laser. This limitation increased the risk of unmeasured confounding effects and therefore poses a threat to the interval validity and generalizability of our findings. We could not speculate on the direction or magnitude of the so-caused bias. Second, the small sample size prevented a subgroup analysis, multilevel analysis or some other reliable control of the design effect, and affected the precision, reliability and generalizability of our findings, but not necessarily jeopardized the internal validity of the study. Third, we used only subjective, patient-reported outcomes in an open-label study with no control group. This may have increased the risk of measurement bias caused by the placebo effect, which probably acted against the null hypothesis of no effect of the Er:YAG SMOOTH^® laser on the symptoms of SUI, but we were not able to determine its magnitude. Fourth, we omitted to control and recode the time from hysterectomy to the intervention; this may have increased the heterogeneity of our hysterectomized group, while lowering the difference from the non-hysterectomized group. We could not control for the effect of this possible, unmeasured confounder, and future studies should explore the possible moderating effects of the time from hysterectomy on the efficacy of Er:YAG SMOOTH^® laser treatment of SUI, and to the severity of SUI itself. Lastl, the physicians were free to change the number of sessions from two, as recommended by the manufacturer’s protocol, according to the vaginal length and width, and severity of atrophic vaginitis. However, the physicians could be biased due to the hysterectomy and could have performed additional laser sessions arbitrarily. This possibly unnecessary increase in the number of sessions may increase the effect of Er:YAG SMOOTH^® laser treatment in hysterectomized patients. We tried to control this source of bias by adjusting the main analysis for the number of sessions. However, this adjustment could not solve the effects of the possibly arbitrary and unnecessary increase in the number of sessions, but just the effect of their overall larger number. So, future studies should compare the two targeted groups with more rigorously defined criteria for the adaptation of the number of sessions.

The uniqueness of our study was the assessment of the effects of the Er:YAG SMOOTH^® laser treatment in the population of hysterectomized women because the real-world data on this subject are missing.

Conclusion

The Er:YAG SMOOTH^® laser treatment seems to improve the symptoms of SUI in hysterectomized women not clinically relevantly less than in non-hysterectomized women. It seems that the beneficial effect of Er:YAG SMOOTH^® laser treatment for SUI in hysterectomized women is time-limited. These results are promising, and therefore prospective, multicenter, randomized, controlled trials are needed with sham and surgery control groups and more rigorous and objectively defined criteria for adapting the number of sessions.

Potential conflict of interest

The authors declare no conflict of interest. The authors alone are responsible for the content and writing of the paper.

Source of funding

Nil. The study was funded by the authors.