This study was conducted in strict accordance with the Cochrane Handbook of Systematic Reviews and Meta-analysis ver. 6.2 (2021) [19] and reported according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement guidelines [20]. In consultation with a research librarian, an electronic search was performed on August 22, 2021 in the following databases: MEDLINE (via PubMed), EMBASE, the Cochrane Library, and the ClinicalTrials.gov website to identify all published and indexed studies, and the gray literature comparing laser treatment vs. hemorrhoidectomy for grade II/III hemorrhoids. A repetitive and exhaustive permutation of the following Medical Subject Headings (MeSH) (expanded) terms were used: “laser,” “coagulation,” and “hemorrhoids.” The reference lists of relevant studies were manually searched to identify additional studies.

Both randomized controlled trials (RCTs) and non-RCTs were included if comparative outcomes were reported for patients undergoing upfront laser therapy or CH for grade II or III hemorrhoids. Single-arm, noncomparative studies were excluded. Non-English studies and English studies with no extractable data were excluded. Studies including grade IV hemorrhoids were excluded, as excisional hemorrhoidectomy currently remains the mainstay treatment method [4].

Studies were selected in 2 stages. First, 2 reviewers (IJYW, CHK) independently screened and assessed the studies for inclusion by their titles and abstracts. Full-text articles were subsequently obtained for review. The senior author served as the arbiter to resolve differences of opinion regarding the studies’ eligibility by consensus. The search strategy is shown in the PRISMA diagram (Fig. 1).

The outcomes of interest were divided into intraoperative and postoperative variables and extracted independently by 2 authors using a standardized proforma. The former included operative time and intraoperative blood loss. Postoperative variables included postoperative analgesia use, postoperative morbidity including bleeding, urinary retention, thrombosis, pain (measured via the visual analogue scale [VAS]), as well as the time of return to work/daily activities and longer-term symptom recurrence. Wherever reported, quality of life (QoL) scores were also assessed. In addition to the outcomes above, we extracted the following data from each study: first author, year, type of publication, age, sex, grade of hemorrhoids, and preoperative symptoms.

Meta-analyses were undertaken as per recommendations from the Cochrane Handbook [19]. Statistical analyses were performed using the RevMan 5.4 software (The Nordic Cochrane Centre), where pooling of weighted mean differences or standardized mean differences was conducted to generate summary statistics for continuous variables, while the risk ratio (RR) was used for dichotomous variables. Statistical heterogeneity was assessed using the I² statistic. A random-effects model was chosen when the I² statistic was greater than 50%, and a fixed-effects model otherwise. Results were reported with 95% confidence intervals (CIs), and a P-value of less than 0.05 was treated as statistically significant. Whenever variables were reported as median (range), the methods described by Hozo et al. [21] were implemented to convert them to the respective mean and standard deviation. To reduce heterogeneity, although subgroup and meta-regression analyses were considered, they were not performed given the small number of studies [19].

The Cochrane Risk of Bias tool was used to assess the study quality and risk of bias for RCTs based on aspects of selection, performance, detection, attrition, reporting, and other bias. For the quality of non-RCTs, the Newcastle-Ottawa Scale (NOS) [22] (for cohort studies) was utilized to assess the domains of patient selection, comparability of study groups, and outcome assessment. Publication bias was not assessed using the funnel plot and Egger regression test since there were fewer than 10 studies [19].

The systematic search across the various databases yielded an initial total of 260 publications. After removing duplicate publications and excluding those that did not fit the inclusion criteria based on title and abstract review, 15 publications remained and were reviewed in their entirety. Six publications were excluded for reasons stated in Fig. 1, while 9 [10, 12, 18, 23-28] were included after the final review, of which 3 were RCTs [10, 18, 27], 4 were prospective cohort studies [23-25, 28], and 2 were retrospective studies [12, 26]. Notably, of the 6 studies excluded, the decision to exclude that of Mohammed et al. [29] was largely based on a lack of clear data reporting and a high risk of reporting bias, despite a large sample size of 1,000 patients, which would have strengthened our study’s power and external validity.

The 9 studies included 661 patients with grade II or III hemorrhoids, of whom 336 underwent LH and 325 underwent CH. The mean age of the LH and CH cohorts ranged from 34.7 to 47.0 years and 33.7 to 49.0 years, respectively. Males predominated in both arms. For studies that reported preoperative symptoms, the most common symptom was rectal bleeding, followed by pain and itching. Regarding the surgical technique, all studies employed conventional Milligan-Morgan (MM) hemorrhoidectomy. Detailed data can be found in Table 1.

Of the 6 non-RCTs, all had a score of above 7 out of the maximum 9 on the NOS and were assessed to be robust methodologically, and of low risk of bias in terms of patient selection, comparability of study groups, and outcome assessment. Despite implementing surgical interventions, 3 RCTs had a low risk for performance and detection bias given the presence of blinding of participants, personnel, and outcome assessors (Supplementary Figs. 1, 2).

The LH group had a significantly shorter operative time (7 studies, n= 556) than the CH group (mean difference [MD], –12.65 minutes; 95% CI, –16.00 to –9.29 minutes; P< 0.001) (Fig. 2). In addition, blood loss (2 studies, n= 120) was significantly lower in the LH group than in the CH group (MD, –19.78 mL; 95% CI, –23.15 to –16.42 mL; P < 0.001) (Supplementary Fig. 3). Both meta-analyses were performed using fixed-effects models.

In terms of postoperative morbidity, the risks of bleeding (6 studies, n= 520; RR, 0.18; 95% CI, 0.12–0.28; P< 0.001) (Fig. 3), and urinary incontinence (5 studies, n= 316; RR, 0.30; 95% CI, 0.10–0.95; P= 0.040) were significantly lower in the LH group (Supplementary Fig. 4). The pooled incidence of postoperative thrombosis was 0.07% across 3 studies that reported this outcome. Patients in the LH group returned to work or daily activities (4 studies, n= 420) significantly earlier than patients in the CH group (MD, –11.81 days; 95% CI, –19.39 to –4.23 days; P = 0.002) (Supplementary Fig. 5).

Concerning postoperative pain, the use of analgesia (3 studies, n= 160) was significantly lower in the LH group than in the CH group (RR, 0.59; 95% CI, 0.42–0.81; P= 0.001) (Supplementary Fig. 6). Only VAS scores reported on postoperative day (POD) 1 were meta-analyzable. Patients in the LH group had significantly lower POD 1 VAS scores (5 studies, n= 445) than those in the CH group (MD, –2.09; 95% CI, –3.44 to –0.75; P = 0.002) (Fig. 4). This remained consistent in the VAS scores 1 week postoperatively (2 studies, n = 105; MD, –3.94; 95% CI, –6.36 to –1.52; P= 0.001) (Supplementary Fig. 7).

Four studies, including those of Shabahang et al. [27], Eskandaros and Darwish [28], Maloku et al. [24], and Hassan and El-Shemy [25] reported longer-term VAS scores. However, these were not meta-analyzable due to heterogeneity in reporting. Shabahang et al. [27] reported no significant difference in VAS scores at 6 months postoperatively, despite a reduction in pain intensity in both groups (1.20± 0.40 in LH group vs. 1.37± 0.58 in CH group; P= 0.174). Eskandaros and Darwish [28] demonstrated a significantly lower VAS score in the LH group than in the CH group at 4 weeks (0.0± 0.0 vs. 1.2± 0.7, respectively; P< 0.001) and 8 weeks postoperatively (0.0 ± 0.0 vs. 0.4 ± 0.5, respectively; P < 0.001). Maloku et al. [24] showed that VAS scores were significantly lower in the LH group than in the CH group at POD 30 (0.2± 0.1 vs. 0.8± 0.2, respectively; P< 0.001) and POD 60 (similar values as POD 30). Hassan and El-Shemy [25] also reported significantly lower VAS scores in the LH group than in the CH group up to POD 30. Abdulkarim et al. [26] assessed pain symptoms postoperatively between the LH and CH groups and classified the pain scores as mild (85.7% vs. 66.7%, respectively), moderate (4.8% vs. 20.0%, respectively), and severe (9.5% vs. 13.3%, respectively), but found no significant differences (mean pain score of 1.05 vs. 2.00, respectively; P= 0.277).

Five studies [10, 18, 23-25] reported longer-term outcomes. Alsisy et al. [23] demonstrated that there were no statistically significant differences in terms of bleeding, pain, itching, and recurrence at 3 months of follow-up. Similar observations were noted by Maloku et al. [24] at 2 months of follow-up, where the incidence of bleeding was similar between both groups. Hassan and El-Shemy [25] also reported similar pain scores between both groups at 6 months of follow-up.

Patient follow-ups at 12 months were conducted in both RCTs included in this meta-analysis [10, 18]. There were no significant differences in the proportion of patients who remained completely asymptomatic (RR, 0.89; 95% CI, 0.74–1.08; P = 0.240) (Supplementary Fig. 8). Risks of bleeding (RR, 1.00; 95% CI, 0.42–2.37; P< 0.999) (Supplementary Fig. 9), and prolapse (RR, 1.71; 95% CI, 0.70–4.20; P= 0.240) were comparable (Supplementary Fig. 10).

Poskus et al. [18] and Shabahang et al. [27] were the only 2 researchers that formally conducted a QoL assessment using the 36 Item Short Form (SF-36) questionnaire. The former showed that the 1-year general health evaluation score was higher in the LH group than in the CH group (mean [interquartile range]: 60 [43–60] vs. 58 [50–70], P= 0.023). Furthermore, when patients were asked to evaluate the operation using a VAS from 1 to 10 at the 1-year visit, LH was regarded as the best operation by patients [18]. Corroborative findings were seen in the trial by Shabahang et al. [27], where the LH group reported better overall QoL than the CH group (P= 0.037).