Effects of weight loss on QTc in people with obesity: a systematic review and meta-analysis

Abstract

Background and aims

Overweight and obesity have been found to exhibit a statistically significant increase in corrected QT interval (QTc), a major contributing factor to sudden death. However, the influence of widely used weight loss strategies including diet, exercise, anti-obesity drugs, and bariatric surgery on QTc remains inconsistent. Therefore, the present systematic review and meta-analysis aim to quantitatively analyse and evaluate the effect of weight loss on QTc in obese patients after diet control with exercise intervention and anti-obesity drugs, as well as bariatric surgery.

Methods

Twenty randomised controlled trials (RCT) and observational studies were included in the meta-analysis on the effects of weight loss on QTc. The fixed-effects model was employed in the RCTs, and the random-effects model was employed due to the presence of statistical heterogeneity among observational studies. Subgroup analysis was conducted to understand the differences in distinct weight loss methods and follow-up time.

Results

Overall, the QTc of people with obesity after weight loss was shorter than that before (mean difference (MD) = 21.97 ms, 95% confidence interval (CI) = 12.42, 31.52, p < .0001). Subgroup analysis restricted to seven included studies whose intervention was diet control with exercise showed a decrease of QTc with statistical significance (MD = 9.35 ms, 95%CI = 2.56, 37.54, p = .007). In the remaining 11 studies, bariatric surgery was the weight loss method. The results also showed a shortening of QTc after surgery, and the difference was statistically significant (MD = 29.04 ms, 95%CI = −16.46, 41.62, p < .00001). A statistically significant difference in QTc shortening at 6 months compared to pre-operation values was further observed (MD = −31.01 ms, 95%CI = −2.89, −59.12, p = .03). The shortening of QTc at 12 months of follow-up was also significantly different from that before surgery (MD = 36.47 ms, 95%CI = 14.17, 58.78, p < .00001). Moreover, the differences became more pronounced as the follow-up time extended.

Conclusions

We demonstrate that weight loss links to a shortened QTc, without considering the means of weight loss. Bariatric surgery has been found to result in a greater reduction in QTc.

Keywords:

• Weight loss
• obesity
• QTc
• diet control
• exercise
• anti-obesity drugs

Introduction

Overweight and obesity, an alarmingly increasing global public health issue, drive a myriad of metabolic perturbations and serve as a risk factor for multiple chronic diseases. Obesity has been identified as a significant predictor of sudden cardiac death in the Framingham Heart Study [1]. A recent Mendelian randomisation study has confirmed that genetically predicted BMI was associated with a longer QTc [2]. Electromechanical dispersion in long QT syndrome is usually characterised by prolonged ventricular repolarisation [3], which is a major contributing factor to sudden death [4,5]. Most studies assessing ventricular repolarisation in people with obesity utilise QTc as diagnostic probes, while QTc represents correction of the QT interval for heart rate, usually calculated using Bazett’s formula [6].

A meta-analysis conducted by Sjaak Pouwels, aggregated data from seven studies and focused on obesity patients associated with electrocardiogram (ECG) abnormalities and arrhythmias found surgeries including vertical banded gastroplasty (VBG), Roux-en-Y gastric bypass (RYGB) surgery and sleeve gastrectomy yielded to a significantly decreased QTc of −33.6 ms (95%CI = −49.8 to −17.4) [7]. However, this meta-analysis only focused on the influence of bariatric surgery on QTc with no other means of weight loss involved. Whether the most widely used methods of weight loss in everyday life including diet control, physical activity, and anti-obesity drugs, would pose a beneficial effect of QT and thus sudden death remains poorly evaluated.

For diet control, a study has shown that with a low-calorie diet (810 kcal per day) for rapid weight loss of 8 weeks, QTc slightly decreased without statistical significance [8]. The ketogenic diet (KD) is a high-fat, low-carbohydrate, adequate-protein diet that is an emerging way of lose weight in recent years. The association between KD and prolonged QT interval, life-threatening ventricular arrhythmias, and sudden death remains controversial [9,10] and without any systematic evaluation [11].

Physical activity brings benefits to the cardiovascular system, in addition to lose weight [12]. Regarding those who lose weight by exercise, studies found that moderate-intensity continuous exercise or high-intensity intermittent exercise showed different influences on QTc [13].

The impact of new anti-obesity drugs, including semaglutide, liraglutide, tirzepatide and orlistat, on QT remains uncertain. Semaglutide is a GLP-1 analogue approved for weight loss and, at the same time, used for type-2 diabetes. Randomised controlled trials (RCTs) have confirmed that semaglutide does not prolong the QTc [14,15]. Moreover, sibutramine should not be ignored. The Sibutramine Cardiovascular Outcome Trial studies have demonstrated a significant increase in QT dispersion in patients administered sibutramine, and its cardiac arrests adverse effect has been reported in many other countries. Sibutramine was first banned for weight loss [16,17], and then withdrew in Europe [18].

Both RCT and observational study are included in our meta-analysis. Nevertheless, no investigation has been performed to comprehensively explore the influence of different weight loss methods on QTc. Therefore, the present systematic review and meta-analysis aimed to quantitatively analyse and evaluate the effect of weight loss on QTc in people with obesity after diet control with physical activity, and anti-obesity drugs, as well as bariatric surgery.

Materials and methods

Inclusion criteria: (1) Type of study: RCT and observational studies published in English; (2) subjects: obese adult patients with ≥18 years of age and BMI ≥30 kg/m²; (3) interventions: diet control; exercise; anti-obesity drugs, bariatric surgery, including VBG, RYGB surgery and sleeve gastrectomy; and (4) indicator for further observation: QTc, QTd, and BMI.

Exclusion criteria: (1) QTc data missing; (2) no specific description of QTc was provided; only change trend was noted, and no specific coordinate value was derived; (3) study indicators were QT and QTd, but heart rate was not provided to convert to QTc; (4) study indicator was rQT of Holter ECG, which cannot be converted to QTc.

Search strategy: An online literature search was conducted on several databases, including PubMed, Embase, Cochrane Library, Clinical Trials, Web of Science, CNKI, and WanFang Database, from their inception until October 2023.

The following English search terms were used: weight loss, obesity, adiposity tissue, bariatric surgery, QTc interval, diet control, exercise, and anti-obesity drugs. The string used for the literature search using the following keywords was modified for each database: (((((QT) AND (QTc)) AND (Obesity)) OR (Weight loss)) OR ((((Diet control) OR (Exercise)) OR (Anti-obesity Drugs)) OR (Bariatric surgery))) AND ((Randomized Controlled Trail) OR (Observational study)). The corresponding search strategies were modified for different databases, and other journals were manually searched to trace the references of the included literature (Figure 1). The protocol of this systematic review was not registered in advance.

Baseline data extraction and evaluation of study quality

The authors LKW and NH individually screened and selected studies based on title and abstract. After primary selection, the authors reviewed the full text of the selected studies and determined their eligibility for inclusion based on the established selection criteria. The data extracted from the studies included information on author names, publication year, type of study, sample size, age, intervention measures, BMI, gender QT interval, QTc, QTd, potassium level, serum glucose, and follow-up time. The methodological quality of the observational studies was evaluated using the standard Newcastle–Ottawa Scale (NOS) for non-randomised trials (Figure 2). The NOS assigns a maximum of nine points for the least risk of bias in three domains: (1) selection of study groups (four points), (2) comparability of groups (two points), and (3) assessment of exposure and outcomes (three points) for case–control and cohort studies, respectively. Therefore, stars were assigned to each quality item and served as a quick visual assessment. Studies with the highest quality received up to nine stars [19]. The risk of bias of RCT was evaluated according to the Revised Cochrane risk-of-bias tool for randomised trials, including seven aspects: (1) random sequence generation (selection bias), (2) allocation concealment (selection bias), (3) blinding of participants and personnel (performance bias), (4) blinding of outcome assessment (detection bias), (5) incomplete outcome data (attrition bias), (6) selective reporting (reporting bias), and (7) other bias.

Disagreements between the two authors were resolved by discussion with each other and with the senior author (LY) until a consensus was reached. Cohen’s kappa score was calculated to determine the level of agreement (between LKW and NH).

Statistical analysis

RevMan 5.4, provided by the Cochrane Collaboration Statistical software, was used for statistical analysis. Mean difference (MD) was used to determine the statistical effects of continuous variables with the same unit. A 95% confidence interval (CI) was calculated for each effect size, and I² was calculated to assess heterogeneity between the studies. A sensitivity analysis was conducted for results with I² > 50% and p < .10. In all tests, p values of <.05 were considered statistically significant.

The χ² test was used to test the heterogeneity of the included studies. For p > .1 or I² ≤ 50%, the fixed-effects model was used; otherwise, the random-effects model. Subgroup analysis was performed based on the possible sources of heterogeneity.

Electrocardiogram definitions

The QT interval, ranging up to 440 ms, is measured from the first deflection of the QRS complex to the end of the T-wave at the isoelectric line. Heart rate-corrected QT intervals (QTc) are denoted as QTc and are calculated using various correction formulae. Bazett’s formula was frequently employed in the included studies. The QTd is defined as the difference between the longest and shortest QT intervals with a 12-lead ECG [20].

Literature screening process and results

A total of 12,106 English studies were initially retrieved from the databases. Twenty studies were finally included after browsing their title and abstracts and searching for the full text (Table 1). Eleven studies used bariatric surgery, seven studies used diet control with exercise, and two RCTs used semaglutide as an intervention.

Figure 1. Process of literature screening.

Figure 2. Assessment of the methodological quality of the included studies. S1: representativeness; S2: selection; S3: ascertainment; S4: demonstration; C1: comparability; O1: outcome selection; O2: outcome follow-up; O3: adequacy. For criteria points S1–S4, O1–O3, it is possible to achieve one star; for criteria point C1, it is possible to achieve two stars. *The study suffices in this criteria point. **The study suffices in these two criteria points.

Table 1. Characteristics of the included studies.

Study Country	Type of study	Number/gender/BMI	Intervention	Serum glucose	Access to QTc	Potassium level	Excluded criteria	Follow-up time	Primary outcome measures	Results
1. Bezante et al. [21] Italy	Observational study	55 patients/age, 39 ± 11 y/42 females/BMI 45 ± 11 kg/m^2	Bariatric surgery	102 ± 41.4 mg/dL	Bazett’s formula (QTc_QT/_RR)	Not given	Previous diagnosis of coronary artery disease, known cardiovascular disease, atrial fibrillation, known cardiac arrhythmias, cancer, or renal dysfunction. No drugs affecting QT were used	Between January 2001 and July 2002. 1 month,6 months, and 12 months after surgery	QTc and/or QTd, BMI	Pre-max QTc of 446 ± 28 ms and a mean QTd of 52 ± 20 ms, Post-QTc min (ms): 1 month: 387 ± 22, 6 months: 387 ± 21, and 12 months: 385 ± 21. Post-QTd (ms): 1 month: 32 ± 14, 6 months: 32 ± 12, and 12 months: 35 ± 12 (6 months after surgery)
2. Mukerji et al. [22] U.S.A.	Observational study	39 patients/BMI 43.8 ± 2.3 kg/m^2/31 women and 8 men/mean age 37 ± 8 years	Bariatric surgery (vertical banded gastroplasty (VBG))	Not mentioned	Bazett’s formula	Serum potassium, magnesium, and calcium levels were obtained in the fasting state on the same day as the electrocardiograms and echocardiograms.	Underlying organic heart disease, heart failure, current or previously treated hypertension, electrolyte disorder, or valvular heart disease. No drugs or factors affecting QT were used	Not given	QTc, BMI	Pre-BMI 42.8 ± 2.1, QTc 428.7 ± 18.5, QTc dispersion (ms) 44.1 ± 11.2, post-BMI 31.9 ± 2.2, QTc (ms) 410.3 ± 11.9, QTc dispersion (ms) 32.2 ± 3.3
3. Alpert et al. [23] U.S.A.	Observational study	Total 67 patients: 28 with and 39 without HF, BMI ≥40 kg/m^2, all patient, 46.2 ± 5.3 years	Bariatric surgery (vertical banded gastroplasty in most)	Diabetes mellitus patient, n = 20 (30)	Bazett’s formula	Serum potassium, magnesium, and calcium levels were obtained in the fasting state on the same day as the electrocardiograms and echocardiograms.	Any cardiac arrhythmia on a 12-lead electrocardiogram (including premature beats and atrial fibrillation). Patients with primary pulmonary disease, AMI, hypertension, electrolyte disorder, and valvular heart disease. No drugs or factors affecting QT were used	5.0+–0.6 months after surgery	QTc and QTc dispersion, left ventricular (LV) mass/height^2	Pre-all patient QTc (ms): 433.4 ± 19.0, have HF 440.0 ± 17.5, no HF 428.7 ± 18.7, QTc dispersion (ms): all patient 50.3 ± 12.5, have HF 58.9 ± 8.2, no HF 44.0 ± 11.3, Post-all patient QTc (ms): 408.7 ± 9.8, have HF 406.4 ± 4.5, no HF 410.5 ± 2.0; QTc dispersion (ms): all patient 41.0 ± 11.5, have HF 53.4 ± 2.4, no HF 44.0 ± 11.3 32.0 ± 3.4
4. Ibisoglu et al. [24] Turkey	Observational study	58 patients	Bariatric surgery	Diabetes mellitus; n (%): pre-op 14 (24.1%), post-op (6 months) 3 (5.2%)	12-lead ECG	Not given	Coronary artery disease and heart failure. No drugs or factors affecting QT were used	Pre-op and 6 months post-op.	Heart rate, PR, QT max, QTc max, QTc min	QTmax (ms) pre-: 409.26 ± 35.60, post-6 months 392.40 ± 32.2; QT min (ms), pre-358.7 ± 35.49, post-6 months: 360.12 ± 33.38; QTc max (ms), pre-446.01 ± 30.89, post-6 months 407.64 ± 28.76, QTc min (ms), pre-395.45 ± 30.87, post-375.36 ± 28.49, cQTd (ms), pre 50.56 ± 7.12, post 6 months 32.28 ± 6.92
5. Grasser et al. [25] Switzerland	Observational study	49 patients/35 females	Roux-en-Y gastric bypass (RYGB) surgery	Glucose, mmol L^−1, pre: mean ± SD: 6.1 ± −2.3, post: mean ± SD: 4.8 ± 0.9	Bazett’s equation (QTc = QT/RR0.5), Fridericia’s equation (QTc = QT/RR0.33), Framingham’s equation (QTc = QT + 0.154 (1000 − RR)) and Hodges’s equation (QTc = QT + 1.75 (heart rate − 60)).	Not given	A known heart disease and any kind of arrhythmia or bundle blocks in ECG recordings	Before and 12 months after	QTinterval (QTc)	QTc (ms), pre: 429 ± 21; post-398 ± 23
6. Sarmiento-Cobos et al. [26] U.S.A.	Observational study	81 patients/females 67.9% (n = 55)/age 55.07 ± 14.17 years	Bariatric surgery (80% sleeve gastrectomy, 20% gastric bypass surgery)	No full text	No full text	No full text	No drugs or factors affecting QT were used	1 year	QT and [QTc] intervals	QTc: pre 438.7 ± 29, post 426.8 ± 25.3
7. Gul et al. [6] Turkey	Observational study	48 patients	Bariatric surgery	Glucose (mg/dL, mean ± SD) baseline: 112.62 ± 38.95, 6 months: 109 ±36.84	Bazett’s formula	Potassium (mmol/L): baseline 4.4 ± 0.3, 6 months: 4.6 ±0.29	Age > 45 years, coronary artery disease, heart failure with decreased left ventricular ejection fraction (EF < 50%) or requiring medical treatment, moderate or severe valvular heart disease, chronic kidney disease, suspicion of active infection or malignancy during admission, chronic obstructive pulmonary disease, connective tissue disorders, electrolyte imbalance, type I diabetes mellitus and hypertension for at least 1 year. No drugs or factors affecting QT were used	6 months	BMI, QTc, QTc-d	BMI, pre: 45.74 ±5.60, post: 6 months: 31.71 ±4.49, QTc pre: 395.87 ±18.00, post 6 months: 402.33 ±18.29. QTc-d pre: 48.60 ±9.98, post 6 months: 31.56 ±6.98.
8. Al-Salameh et al. [20] U.S.A.	Observational study	28 patients	Bariatric surgery (sleeve gastrectomy)	Three had diabetes	Bazett’s formula	Not given	Not given	3 months	Corrected QT (QTc),BMI	BMI pre: 45.27 ±6.09 kg/m^2, post 3 months: 38.32 ±5.19 kg/m^2 QTc pre: 427 ±18.6 ms, post 3 months: 398.6 ±15.5 ms.
9. Papaioannou et al. [27] Greece	Observational study	17 patients	VBG	Not mentioned	Bazett’s equation	Potassium (mmol/L) preoperative: 4.4 ± 0.4, postoperative: 4.5 ± 0.4.	Abnormal serum electrolyte values, cardiovascular problems, or long-term pharmaceutical therapy that might affect QTc	8–10 months	QTc, BMI	BMI pre-surgery: 49.7 ± 10.5, post-surgery: 36.6 ± 8. QTc (ms) pre: 428 ± 14, post-surgery: 393 ± 26.
10. Seshadri et al. [28] U.S.A.	Observational study	47 patients/BMI ≥35 kg/m^2	Randomised to either a carbohydrate-restricted diet, n = 25 (low-carbohydrate group: carbohydrate intake to ≤30 g/day) or a fat- and calorie-restricted diet, n = 22 (conventional group: 500 calorie deficit daily, with ≤30% of these calories derived from fat)	Provided insulin resistance	Bazett’s formula	Not mentioned	?	For 6 months	QTc, BMI	Baseline QTc (ms) low carbohydrate 431 ± 24, conventional (low fat) 426 ± 24.6 after diet control: −12.60 ± 22.56, +0.09 ± 21.19
11. Drigny et al. [13] Canada and France	Observational study	65 patients/age 53 ± 9 years	Moderate-intensity continuous exercise (MICE, n = 30) training, high-intensity interval exercise training (HIIT, n = 35)	Fasting glycaemia (mmol/L): MICE baseline: 5.8 ± 0.89, 9 months 5.76 ± 1.16; HIIT: baseline 5.33 ± 0.77, 9 months 5.47 ± 0.73	Bazett’s formula	Not mentioned	?	9 months	QT dispersion = QTd, standard deviation of QT = sdQT, relative dispersion of QT = rdQT, QT corrected dispersion = QTcd), BMI, waist circumference	MICE: QTd (ms) before 56 ± 18, 9 months 51 ± 20. QTcd (ms) before: 65 ± 26, 9 months: 57 ± 23. BMI before: 32 ± 6, after: 31.2 ± 5.9, HIIT: QTd (ms) before: 38 ± 9, 9 months: 34 ± 9; QTcd (ms) before: 41 ± 10, 9 months: 5: 41 ± 11. BMI before: 35.7 ± 5.1, after: 34 ± 4.8
12. Vedel-Larsen et al. [8] Denmark	Observational study	26 patients	Weight loss intervention (low-calorie diet (810 kcal per day, Cambridge Weight Plan Products^®) for 8 weeks. The dietary energy distribution was 46% from carbohydrates, 41% from proteins, and 13% from fat. Daily intake of protein was at least 43.2 g, and intake of the essential fatty acids, linoleic and linolenic acid, was 3 g and 0.4 g, respectively. Dietary fibre intake was 7.2 g per day at the minimum)	Fasting glucose (mmol/L): week 0; 5.46 ± 0.10, week 8: 5.08 ± 0.10	12-lead median ECGs	Potassium (mmol/L): week 0: 3.82 ± 0.05, week 8: 3.76 ± 0.06	Diabetes or other known illnesses that could affect glucose or lipid metabolism, pregnancy, and breastfeeding	8 weeks	QT, QTcF (Fridericia-corrected QT-interval), body weight	QT (ms): week 0: 398 ± 5, week 8:408 ± 5; QTcF (ms): week 0: 416 ± 3, week 8: 417 ± 3, NS; body weight (kg): week 0: 101 ± 3, week 8: 89 ± 25
13. Pontiroli et al. [29] Italy	Observational study	116 patients/20men, 96 women/age 45.1 ± 0.61 years	Bariatric surgery	Insulin and blood glucose level	Bazett’s formula	Mentioned but not given	12 diabetic patients	1 year	QTc	QTc (ms) before: 415 ± 40.7, 1 year: 401 ± 44.81
14. Russo et al. [30] Italy	Observational study	100 patients/42men, 58 women/age 50.65 ± 7.9 years	Bariatric surgery: bilio-intestinal bypass	Diabetes already excluded	Bazett’s formula	Not given	With a history of hypertension (systolic and diastolic blood pressure >140/90), diabetes or impaired glucose tolerance (IGT), anaemia, electrolyte imbalance, valvular heart disease, heart failure, coronary artery disease, connective tissue disorders, left bundle branch block or atrioventricular conduction abnormalities on ECG, previous arrhythmic episodes, hepatic, renal, thyroid diseases or sleep disorders	12 months	QTd, QTc	QTc (ms): pre: 464 ± 28.9, post: 400 ± 25.4, QTd (ms): pre: 60.9 ± 23, post: 40.5 ± 12.7
15. Gupta et al. [31] USA	Observational study	63 patients/mean age 42 years	Liquid protein diet therapy	–	Not given	–	Not given	6 months	QTc, QTd	QTc: before 423 ± 13.7, after 415 ± 12.7 (ms)
16. Mshui et al. [32] Japan	Observational study	36 patients/18 men, 18 women/age28 ± 9, women 33 ± 14 year/BMI men 35 ± 5,women 38 ± 6 kg/m^2	Behavioural therapy and VLCD: 370 kcal/D	Before 90 ± 22.1 mg/dL	Bazett’s formula	Before 4 ± 0.2 mg/dL, after 4 ± 0.3 mg/dL	Not given	5 years	QTc	QTc: before 445 ± 32, after 434 ± 28 (ms)
17. Pietrobelli et al. [33] USA	Observational study	30 patients/15 men, 15 women/age men 35.9 ± 7.1 years, women: 32.9 ± 7.5 year/BMI men: 29.9 ± 1.5,women: 30.0 ± 2.0 kg/m^2	1120 kcal/d (23% from fat) + 50 g protein/d	–	Not given	–	Patients with prolonged QTc were initially excluded.	8 days	QTc	QTc: before 414 ± 28, after 404 ± 25 (ms)
18. Seyfeli et al. [34] Turkey	Observational study	30 patients/24 women, 6 men/age 44 ± 8 year/BMI 42 ± 5 kg/m^2	–	–	Not given	–	–	12 weeks	QTc	QTc: before 417 ± 18, after 410 ± 17(ms)
19. Demmel et al. [14] Germany	RCT	76 patient/60% males/mean age 38.2 year/mean BMI 25.1 kg/m^2	Subcutaneous semaglutide 1.5 mg	Fast plasma glucose decrease with semaglutide (−0.33 ± 0.39 mmol/L)	Bazett’s correction, Fridericia’s correction	Not given	(1) Known or suspected hypersensitivity to trial products or ECG electrodes; (2) participation in another clinical trial 90 days before screening; (3) a QTcF interval (the Fridericia heart rate (HR) corrected interval) [13] of 450 ms; (4) hypotension (systolic blood pressure (SBP) of 90 mmHg or diastolic BP (DBP) of 50 mmHg) or hypertension (SBP 140 mmHg or DBP 90 mmHg) or HR of 50 beats per minute (bpm) or 90 bpm; (5) a history of seizures, epilepsy, syncope, cardiac arrest, cardiac arrhythmia or Torsade de Pointes, atrioventricular block or structural heart disease; (6) a family history of long QT syndrome; (7) a family history of sudden cardiac death.	0–48 h after last dose	QTc	ΔQTc: 0 h −3.2 (−6.6, 0.3), 12 h −3.4 (−7.0, 0.3)
20. Granhall et al. [15] Germany	RCT	46 patients (males)/age 18–55 year/BMI 18.5–28 kg/m^2	Oral semaglutide 1.2–3.6 g	Not given	Fridericia’s correction	Not given	Blood pressure outside range of 90–139 mmHg for systolic blood pressure or 50–89 mmHg for diastolic blood pressure. heart rate (HR) outside the range 45–89 beats per minute (bpm), smoking, history of major gastric surgery	0–12 h after last dose	QTc	ΔQTc: 0 h −3.2 (−6.3, 4.3), 12 h −3.1 (−6.1, 4.1)

Results

Features of the included study

A total of 1020 patients were included. All of the observational studies defined BMI ≥ 30 kg/m² as obese. Most of the observation indicators of the included studies were QTc before and after weight loss; only in one collaborative study the observation indicator was QTd. Therefore, this study failed to perform quantitative consolidation and only its results were described (that is also the study used exercise as intervention) [13]. The weight loss interventions included diet control with exercise, anti-obesity drugs, and bariatric surgery.

Furthermore, among the seven included studies with diet control with exercise as the method of weight loss, one study from the United States specifically described diet control using a low-carbohydrate diet and provided the composition ratio of carbohydrate and fat content in the diet [28]. In another study from Denmark, the diet followed a low-calorie diet (810 kcal per day, Cambridge Weight Plan Products^®) and met all recommendations for daily intake of vitamins and minerals [8]. Except for one study that chose behavioural therapy and VLCD: 370 kcal/D as the method [32], none of the other studies chose an extreme diet for weight loss.

Semaglutide is approved for weight loss and, at the same time, also be used for type 2 diabetes. We included two RCTs concerning semaglutide used for diabetes on the influence of QTc [14,15]. Their endpoint was the time point matching from baseline QTc to 12 h after last dose.

Among the 11 studies employing surgical weight loss methods, three studies implemented VBG [22,23, 27], one study utilised RYGB [25], and one study employed sleeve gastrectomy [20], another study combined both sleeve gastrectomy and gastric bypass surgery [26]. The remaining studies did not specify the surgical procedures used.

Meta-analysis

For RCTs, we use the tool provided by the Cochrane Handbook to find the standard deviations (SDs) from related information, such as CIs. All studies presented a low risk of bias as assessed by the Cochrane Collaboration tool for assessing risk of bias (Figure 3), and were deemed high quality by the GRADE system. Little heterogeneity was observed in the group comparisons; therefore, fixed effects model was used. The intervention was the semaglutide. It was shown that the semaglutide using time matching did not bring a difference on ΔQTc (MD = −0.06 ms, 90%CI = −6.77, 6.65) (Figure 4).

Figure 3. Risk of bias graph for the included RCTs.

Figure 4. Forest plot of two RCTs of semaglutide using time matching, showing mean difference values for ΔQTc between treatment group and placebo from baseline to 12 h after last dose. Overall mean difference was negative and suggesting treatment with semaglutide at steady state, did not result in QTc prolongation. CI: confidence interval; SD: standard deviation.

For observational studies, the random-effects model was used because of statistical heterogeneity among all the 18 included studies (p < .00001, I² = 97%). The intervention methods considered were diet control with exercise and bariatric surgery. Overall, the QTc of people with obesity after weight loss was shorter than that before (MD = 21.97 ms, 95%CI = 12.42, 31.52, p < .0001) (Figure 5).

Figure 5. Forest plot of 18 studies of people with obesity use diet control with exercise and bariatric surgery as the intervention, showing mean difference values for QTc between before and after weight loss. Overall mean difference was positive and significant suggesting that the weight loss shows a decrease in QTc. CI: confidence interval; SD: standard deviation.

Subgroup analysis was conducted because of differences in weight loss methods and follow-up time among the included studies. At first, studies were divided into groups according to the method of weight loss. Subgroup analysis was performed for the seven included studies whose intervention was diet control with exercise. The results, however, showed a shortening of QTc in the patients (MD = 9.35 ms, 95%CI = 2.56, 37.54, p = .007) (Figure 6) [8, 28, 31–35]. In the remaining 11 studies, surgery was the method used for weight loss. The results showed a shortening of QTc after bariatric surgery, and the difference was statistically significant (MD = 29.04 ms, 95%CI = −16.46, 41.62, p < .00001) (Figure 7) [6, 20–27, 29,30]. As considerable heterogeneity persisted in this subgroup (I² = 96%), we sought to determine whether differences in follow-up time among the included studies contributed to this variation. To explore this possibility, we conducted a subgroup analysis based on postoperative follow-up time. Our findings indicated a statistically significant difference in QTc shortening at 6 months compared to pre-operation values (MD = −31.01 ms, 95%CI = −2.89, −59.12, p = .03). The shortening of QTc at 12 months of follow-up was also significantly decreased from that before surgery (MD = 36.43 ms, 95%CI = 14.17, 58.78, p = .001). Moreover, the differences became more pronounced with the extension of follow-up time, as shown in Figures 8 and 9.

Figure 6. Forest plot of seven studies of people with obesity use diet control with exercise as the intervention, showing mean difference values for QTc between before and after weight loss. Overall mean difference was positive and suggesting that the weight loss shows a decrease in QTc. CI: confidence interval; SD: standard deviation.

Figure 7. Forest plot of 11 studies of people with obesity use bariatric surgery as the intervention, showing mean difference values for QTc between before and after weight loss. Overall mean difference was positive and significant suggesting that the weight loss shows a decrease in QTc. CI: confidence interval; SD: standard deviation.

Figure 8. Forest plot of four studies of people with obesity showing mean difference values for QTc 6 months after bariatric surgery. Overall mean difference was positive and significant suggesting that the weight loss shows a decrease in QTc. CI: confidence interval; SD: standard deviation.

Figure 9. Forest plot of five studies of people with obesity showing mean difference values for QTc 12 months after bariatric surgery. Overall mean difference was positive and significant suggesting that the weight loss shows a decrease in QTc. CI: confidence interval; SD: standard deviation.

Sensitive analysis

Sensitivity analysis was performed on the combined results of each index, and the combined effect size did not change significantly after the sequential removal of references one at a time, indicating the robustness of our results.

Discussion

A shortened QTc is known to reduce malignant ventricular arrhythmias and even the incidence of sudden death [7]. European Heart Journal has reported that the minimal clinically important difference (MCID) of QTc is 20 ms (SD = 15 ms) [36,37]. Our analysis revealed that weight loss resulted in a significant shortening of QTc in people with obesity. Although the mechanism is not fully understood, a possible mechanism is associated with insulin resistance and hyperinsulinaemia, which mediate abnormal ventricular repolarisation in people with obesity, resulting in prolonged QTc [38]. In recent years, several in vivo and in vitro studies have shown that insulin can regulate Na/K-ATPase on the surface of cell membrane, thus, affecting the ventricular repolarisation process and leading to the prolongation of QTc [39–41]. Moreover, people with obesity often have ventricular hypertrophy, which is also associated with repolarisation [23]. Another mechanism is that hyperinsulinaemia can cause hypokalaemia, which in turn prolongs the QTc interval [42]. The loss of weight in people with obesity not only improves insulin resistance but also reduces ventricular hypertrophy, thereby shortening QTc.

Seven studies with diet control along with exercise were separately analysed by us. Because several studies have demonstrated that exercise may improve the QT dispersion [43,44], our meta-analysis employed diet control with exercise as an intervention to lose weight, which shows an improvement to the QTc. However, it is difficult to determine which of these therapeutic interventions was more responsible for influencing QTc.

As for the new anti-obesity drugs, semaglutide is a GLP-1 analogue approved for weight loss and, at the same time, also used for type 2 diabetes. In all the included RCTs, type 2 diabetes was randomised to be treated with semaglutide to placebo. Although the subjects in these two RCTs were diabetic patients, the dose of semaglutide used was not 2.4 mg once a week for weight loss. However, this result is still very valuable in the use of semaglutide as an anti-obesity drug.

It is worth noting that the observed significant decrease in QTc following weight loss is due to the benefits conferred by bariatric surgery. Subgroup analysis showed that [6, 20–27] regardless of the type of surgery performed or the length of follow-up time, QTc was significantly shortened after weight loss, showing the robustness of results. Compared with the previous meta-analysis published back in 2021 [7], our updated meta-analysis involved more people with obesity by including the latest study of bariatric surgery and QTc thus greatly increased statistical power. Our study, despite reporting consistent findings with the previous meta-analysis, provides narrower 95 CIs, indicating an improved accuracy.

The present study has certain limitations, including (1) the fact that out of the 20 studies examined, only two were completely RCTs, while the remaining 18 were observational studies with some extent of randomisation. The included studies were of average quality, and the total number of cases was small (1020); (2) the follow-up time of the included studies was inconsistent, with the shortest being 3 months and the longest being 12 months; (3) considerable heterogeneity was observed in the included studies, which may be attributed to differences in sample size, baseline characteristics of the patient population, and trial duration; (4) some important indicators were not considered in the original study. For example, the levels of neuroendocrine hormones were not followed up in each independent study; this is a critical aspect as some neuroendocrine hormones can affect QTc, such as plasma insulin level, insulin resistance, and the measurement of sympathetic excitability.

Conclusions

In this present study, we demonstrate that weight loss, without considering the means, links to a shortened QTc. People with obesity can choose a more widely used method to lose weight, although bariatric surgery has been found to result in a greater decrease in QTc. Moreover, this systematic review is worth considering the potential benefits of adding high quality observational studies in meta-analysis to complement RCTs for the evaluation of outcome measures, especially after surgery.

Author contributions

Ye Zhu had full access to all of the data in the study and took responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Ye Zhu, Ying Li. Acquisition, analysis, or interpretation of data: Kaiwei Li, Han Nie, and Ying Li. Drafting of the manuscript: Ying Li and Xia Jiang. Critical revision of the manuscript for important intellectual content: Ying Li and Xia Jiang. Statistical analysis: Ying Li, Cheng Tan. Administrative, technical, or material support: Rui Shi, Cheng Tan. Study supervision: Ying Li and Ye Zhu. All authors reviewed the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The datasets used and analysed during the present study are available from the corresponding author upon reasonable request.

Additional information

Funding

No funding.