Real-World Comparison of Oral Versus Injectable Semaglutide for the Reduction of Hemoglobin A_1C and Weight in Patients with Type 2 Diabetes

Abstract

Background:

No head-to-head comparisons of semaglutide formulations currently exist in the literature. In practice, many may think that oral and injectable semaglutide formulations are interchangeable, although there is currently limited real-world data to determine whether this is accurate.

Objective:

The purpose of this study was to determine the effect of oral versus injectable semaglutide on hemoglobin A_1C (HbA_1C) and weight in patients with type 2 diabetes (T2D).

Methods:

This was a retrospective single-center review of adult patients who had a diagnosis of T2D and were treated with oral or injectable semaglutide between November 1, 2019, and July 31, 2022. Primary outcome was a comparison of changes in HbA_1C (%) and weight (kg) from baseline to 6 months between patients receiving oral versus injectable semaglutide, stratified according to highest dose received. Secondary outcomes included frequency of dose reductions and discontinuations, achievement of clinical goals, and presence of an embedded clinical pharmacist at patients’ primary care office.

Results:

A total of 105 patients were included. Patients experienced mean decreases in HbA_1C and weight from baseline to 6 months of −1.75% (P < 0.001) and −3.64 kg (P = 0.015), respectively, in the oral semaglutide group and −1.35% (P < 0.001) and −5.26 kg (P < 0.001), respectively, in the injectable semaglutide group. When directly comparing semaglutide formulations, oral semaglutide demonstrated a 0.4% greater numerical reduction in HbA_1C (P = 0.523) and injectable semaglutide demonstrated a 1.62-kg greater numerical reduction in weight (P = 0.312). Adverse events (AEs) occurred more frequently with oral semaglutide than with injectable semaglutide (16.7% vs 4.9%). Discontinuation due to AEs was more common with oral semaglutide.

Conclusion:

In this study, patients with T2D who received oral semaglutide demonstrated greater reductions in HbA_1C, whereas those treated with injectable semaglutide had greater reductions in weight, although there were no statistically significant reductions in HbA_1C or weight between the 2 formulations. Rates of AEs and discontinuation were more common in the oral semaglutide group.

Introduction

The American Diabetes Association estimates that ~37.3 million Americans have been diagnosed with diabetes mellitus, with the large majority of these diagnoses (90%-95%) being type 2 diabetes (T2D). Persistent hyperglycemia can cause a wide range of complications, including micro- and macrovascular complications such as diabetic nephropathy, neuropathy, coronary artery disease, peripheral arterial disease, and stroke. However, multiple medication classes have been approved for the treatment of T2D that have demonstrated efficacy for preventing these complications.^1,2 Guideline recommendations include the use of glucagon-like peptide 1 receptor agonists (GLP-1 RAs) with demonstrated cardiovascular benefits as an option for first-line therapy in those with atherosclerotic cardiovascular disease or with indicators of high cardiovascular risk. ³

Since Food and Drug Administration approval in 2014, once-weekly injectable semaglutide has been used for the treatment of adults with T2D and remains one of the most commonly prescribed agents in the GLP-1 RA class given its efficacy, once-weekly dosing, and coverage on many insurance formularies. ⁴ In September 2019, oral semaglutide was approved for adults with T2D based on the results of the PIONEER trials, and it became the first oral option available in the GLP-1 RA medication class. ⁵ A retrospective study conducted by Chowdhury et al ⁶ on the use of oral and injectable semaglutide use in clinical practice in the United Kingdom demonstrated changes in hemoglobin A_1C (HbA_1C) of −1.77% and −1.90% for oral and injectable semaglutide, respectively, in addition to average changes in weight of −9.0 kg and −7.2 kg, respectively. When comparing the 2 formulations, Chowdhury et al ⁶ found no statistically significant differences in terms of change from baseline in HbA_1C and weight. A randomized study by Davies et al ⁷ was conducted to compare the effect of oral semaglutide with placebo and injectable semaglutide in patients with T2D but did not directly compare the two formulations. The PIONEER and SUSTAIN trials have individually compared oral and injectable semaglutide with placebo, and results demonstrated the average changes in HbA_1C versus placebo for oral and injectable semaglutide were −1.1% and −1.56%, respectively, in addition to average changes in weight of −2.3 kg and −4.53 kg, respectively.^5,8 Clinical trials have demonstrated that injectable semaglutide reduces the risk of major cardiovascular events, but the same benefit was not seen with oral semaglutide.^{9 -11} Additionally, multiple systematic reviews have concluded that maximum dose injectable semaglutide is more effective at lowering HbA_1C compared with oral semaglutide.^12,13 To date, limited real-world data have been available that directly compare formulations to determine whether oral semaglutide has comparable outcomes to its injectable predecessor. With the increasing popularity of GLP-1 RAs and recent shortages with expanded indications for patients with obesity, our study aimed to provide real-world efficacy and safety data regarding the use of these 2 formulations in clinical practice when determining appropriate treatment options for patients with T2D. ¹⁴

Methods

Study design and population

This study was a retrospective single-center observational review of adult patients within a large academic health system who had a diagnosis of T2D and were treated with oral semaglutide at doses of 7 or 14 mg daily or injectable semaglutide at doses of 0.5 or 1 mg weekly between November 1, 2019, and July 31, 2022. This study met Federal and university criteria for exemption by the institution’s research subjects review board.

All adult patients (≥18 years of age) with an active prescription, defined as refilling at least once, presuming that they were on the medication for at least 1 month confirmed via refill history provided in the electronic medical record, for oral semaglutide at doses of 7 or 14 mg daily or injectable semaglutide at doses of 0.5 or 1 mg weekly and available baseline HbA_1C concentrations and weight values within 3 months prior to starting semaglutide therapy were included. To allow for at least 6 months of follow-up to adequately evaluate primary and secondary outcomes, data were collected through a date of last follow-up of January 31, 2023. There were no exclusions based on use of other concomitant antihyperglycemic medications by patients during the study period. Patients with no documented HbA_1C concentration and weight values within 26 weeks (±6 weeks) following initiation of semaglutide, those who were treated with any medication(s) or supplement(s) for the purpose of weight loss (eg, other GLP-1 RAs, topiramate, phentermine, topiramate-phentermine, amphetamines, benzphetamine, phendimetrazine, diethylpropion, bupropion-naltrexone, or orlistat) during the study period, or those with gaps in refill history (>60 days between fills) indicating potential nonadherence to semaglutide were excluded.

Data collection and study outcomes

Patients were identified from an electronic medical record report identifying an active prescription for oral or injectable semaglutide during the study period. Data-collection variables included patient demographics, baseline HbA_1C concentration and weight, and study outcome measures. Study data were collected and managed using Research Electronic Data Capture (REDCap), a secure web-based software platform. To mitigate the risk of inconsistency or potential errors in the interpretation and categorization of data, data collection was completed by only 2 investigators using the standardized data-collection tool and detailed data dictionary.

The primary outcome was to compare change in HbA_1C concentration (%) and weight (kg) from baseline to 26 weeks (±6 weeks) between patients receiving oral versus injectable semaglutide, stratified according to highest dose received. Each patient served as their own control, and baseline HbA_1C concentration and weight were defined as the patient’s most recent value within 3 months prior to the initiation of semaglutide. Patients were placed into 1 of 4 groups (ie, oral 7 mg, oral 14 mg, injectable 0.5 mg, and injectable 1 mg) based on the highest dose and which formulation of semaglutide they received during the study period and stayed in the assigned group regardless of whether or not the semaglutide dose was reduced.

Secondary outcomes included describing the incidence of and reason(s) for semaglutide dose reductions or therapy discontinuation, percentage of patients achieving an HbA_1C concentration of <7% and/or a clinically significant weight loss (≥5% of their total body weight) from baseline to study completion, and presence of an embedded clinical pharmacist at the patient’s primary care office. Subgroup analyses were conducted to determine relationships between various baseline characteristics and the ability to achieve desired clinical outcomes such as an HbA_1C concentration of <7% and/or weight loss of ≥5% of total body weight.

Statistical analysis

All study data were analyzed by the primary investigator and coinvestigators. Simple descriptive statistics appropriate for nonnormally distributed data (ie, median and interquartile range [IQR]) were used to summarize patient demographics and clinical measures. Given the retrospective nature of the study, there was a finite number of patients that could be included, and therefore, we did not perform a power calculation prior to the study start date. The Wilcoxon signed-rank test was used to evaluate changes in HbA_1C concentration and weight from baseline to various time points. The χ² independence test was used to evaluate the association between exploratory subgroups, achievement of an HbA_1C concentration of <7%, and weight reduction of ≥5% of total body weight at any point. Tests were 2 tailed, and a P value of <0.05 represented statistical significance. All statistical analyses were performed using Jamovi version 2.3 (Jamovi, Sydney Australia).

Results

A total of 605 patients were screened, 222 met the inclusion criterion, and 117 met the exclusion criteria (Figure 1). The most common reasons for exclusion were no available follow-up HbA_1C or weight data at 26 weeks (89%) or missing or inconsistent refill history (38%). Therefore, 105 patients were included in the final analysis, 12 patients in the oral 7-mg group, 11 patients in the oral 14-mg group, 34 patients in the injectable 0.5-mg group, and 48 patients in the injectable 1-mg group.

Figure 1.

Patient enrollment flowchart.

HbA_lC, hemoglobin A_lC; SG, semaglutide; Inj, injectable

^aSome patients met multiple exclusion criteria.

The study population was primarily white (74.3%) males (54.3%) with a median age of 60 years (IQR, 49-67) years. Concomitant antihyperglycemic medications included metformin (76.2%), insulin (52.4%), sodium-glucose cotransporter 2 (SGLT-2) inhibitors (39%), and sulfonylureas (27.6%). Patients had a median HbA_1C concentration of 8.4% (IQR, 7.4%-9.4%) and median weight of 106.9 kg (IQR, 88.5-117 kg) at baseline. Semaglutide therapy was prescribed most by the patients’ primary care provider (61%), and 22.9% of the included patients had an embedded clinical pharmacist practicing comprehensive medication management at their primary care provider’s office. Baseline characteristics are listed in Table 1.

Table 1.

Baseline characteristics.

Characteristic	Oral SG (n = 23)	Inj SG (n = 82)	Total (n = 105)	χ2 value	P value
Age (y), median (IQR)	59 (55-64)	60 (48-68)	60 (49-67)	—	0.692
Sex, n (%)
Male	12 (52.2)	45 (54.9)	57 (54.3)	0.0529	0.818
Female	11 (47.8)	37 (45.1)	48 (45.7)
Race, n (%)
White	17 (73.9)	61 (74.4)	78 (74.3)	3.28	0.194
Black	6 (26.1)	13 (15.9)	19 (18.1)
Other	0 (0)	8 (9.8)	8 (7.6)
Ethnicity, n (%)
Not Hispanic or Latino	23 (100)	72 (87.8)	95 (90.5)	3.10	0.212
Hispanic or Latino	0 (0)	4 (4.9)	4 (3.8)
Unknown	0 (0)	6 (7.3)	6 (5.7)
Insurance type, n (%)
Medicare	7 (30.4)	28 (34.1)	35 (33.3)	2.01	0.366
Commercial	8 (34.7)	37 (45.1)	45 (42.9)
Medicaid	8 (34.7)	17 (20.7)	25 (23.8)
Other antidiabetic agents, n (%)
Metformin	16 (69.6)	64 (78.0)	80 (76.2)	0.713	0.399
Insulin	7 (30.4)	48 (58.5)	55 (52.4)	5.69	0.017
SGLT-2 inhibitor	9 (39.1)	32 (39.0)	41 (39.0)	0.00	0.993
Sulfonylurea	6 (26.1)	23 (28.0)	29 (27.6)	0.0346	0.852
Thiazolidinedione	2 (8.7)	5 (6.1)	7 (6.7)	0.195	0.659
None	2 (8.7)	3 (3.7)	5 (4.8)	1.00	0.316
Other	4 (17.4)	17 (20.7)	21 (20.0)	0.001	0.973
Use of GLP-1 prior, n (%)	5 (21.7)	21 (25.6)	26 (24.8)	0.144	0.704
Length of GLP-1 therapy prior, n (%)
0-24 weeks	1 (4.3)	3 (3.7)	4 (3.8)	0.101	0.750
25-48 weeks	1 (4.3)	2 (2.4)	3 (2.9)	0.434	0.510
>49 weeks	3 (13.0)	16 (19.5)	19 (18.1)	0.538	0.463
Baseline weight (kg), median (IQR)	110 (82.8-124.4)	106.6 (90.3-116)	106.9 (88.5-117)	–	0.427
Baseline BMI (kg/m2), median (IQR)	35.7 (28.5-41.5)	35.7 (31.8-40.3)	35.7 (31.7-40.5)	–	0.410
Baseline HbA1C (5), median (IQR)	8.2 (7-10.2)	8.4 (7.6-9.2)	8.4 (7.4-8.4)	–	0.601

Oral SG, oral semaglutide; Inj SG, injectable semaglutide; GLP-1, glucagon-like peptide; T2D, type 2 diabetes; SGLT-2, sodium-glucose cotransporter-2; HbA_1C, hemoglobin A_1C; Baseline, 3 months prior to initiation of therapy

Note. “Other” antidiabetic agents also include glinides, α-glucosidase inhibitors, and dipeptidyl peptidase-4 inhibitors. “Other” race includes Asian, Other, and Unknown.

Treatment with oral semaglutide (7- and 14-mg doses) resulted in a 0.4% greater reduction in HbA_1C concentration at 26 weeks (P = 0.523), whereas treatment with injectable semaglutide (0.5- and 1-mg doses) resulted in a 1.62-kg greater reduction in weight at 26 weeks (P = 0.312; Figures 2 and 3). Specifically, patients treated with oral semaglutide experienced mean decreases in HbA_1C concentration and weight from baseline to 26 weeks of −0.85% (P < 0.003) and −2.8 kg (P = 0.77), respectively, in the 7-mg group and −2.43% (P = 0.003) and −3.70 kg (P = 0.102), respectively, in the 14-mg group. Patients treated with injectable semaglutide experienced mean decreases in HbA_1C concentration and weight from baseline to 26 weeks of −0.95% (P < 0.001) and −4.1 kg (P < 0.001), respectively, in the 0.5-mg group and −1.70% (P < 0.001) and −6.24 kg (P < 0.001), respectively, in the 1-mg group. At 26 weeks, the median percent reductions in weight from baseline were 3.74% and 4.02% in the oral and injectable semaglutide groups, respectively.

Figure 2.

HbA_1C outcomes with oral and injectable semaglutide: (A) change in HbA_lC concentration from baseline to 26 weeks ± 6 weeks for combined oral and injectable semaglutide; (B) change in HbA_lC concentration from baseline to 26 weeks ± 6 weeks for oral (7 and 14 mg) and injectable (0.5 and 1 mg) semaglutide.

SG, semaglutide; HbA_1C, hemoglobin A_lC; Inj, injectable

Figure 3.

Weight outcomes with oral and injectable semaglutide: (A) change in weight from baseline to 26 weeks ± 6 weeks for combined oral and injectable semaglutide; (B) change in weight from baseline to 26 weeks ± 6 weeks for oral (7 and 14 mg) and injectable (0.5 and 1 mg) semaglutide.

SG, semaglutide; Inj, injectable

Overall, 6 patients (5.7%) experienced an adverse effect (AE) related to semaglutide during the study period, 2 patients (16.7%) in the oral group and 4 patients (4.9%) in the injectable group. The most common AEs were nausea (50%) and diarrhea (16.7%). Semaglutide discontinuation occurred in 4 patients (3.8%) during the study period due to gastrointestinal side effects, 1 patient (8.3%) in the oral group and 3 patients (3.7%) in the injectable groups. No patient required dose reductions of semaglutide, and no deaths were reported during the study period. Adverse effects and reasons for discontinuation are listed in Table 2.

Table 2.

Adverse effects and trends in dose reduction.

Variables	Oral SG 7 mg (n = 12)	Oral SG 14 mg (n = 11)	Inj SG 0.5 mg (n = 34)	Inj SG 1 mg (n = 48)	Total (n = 105)
Experienced ≥1 AE, n (%)	2 (16.7)	—	3 (8.8)	1 (2.1)	6 (5.7)
Time to first AE (d), median (IQR)	145 (108-183)	—	185 (183-294)	293 (293-293)	203 (179-275)
Type of AE, n (%)
Nausea	1 (8.3)	—	1 (2.9)	1 (2.1)	3 (2.9)
Diarrhea	0 (0)	—	1 (2.9)	0 (0)	1 (0.95)
Dyspepsia	0 (0)	—	0 (0)	0 (0)	0 (0)
Allergic reaction	0 (0)	—	0 (0)	0 (0)	0 (0)
Other	1 (8.3)	—	0 (0)	0 (0)	1 (0.95)
Unknown	0 (0)	—	1 (2.9)	0 (0)	1 (0.95)
Discontinued therapy, n (%)	1 (8.3)	1 (9.1)	7 (20.6)	3 (6.3)	12 (11.4)
Time to discontinuation (d), median (IQR)	220 (220-220)	177 (177-177)	327 (247-541)	364 (327-476)	322 (217-498)
Reason for discontinuation, n (%)
GI side effects	1 (8.3)	0 (0)	2 (5.9)	1 (2.1)	4 (3.8)
Injection-site reaction	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
Hypoglycemia	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
Pancreatitis	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
Cost	0 (0)	0 (0)	1 (2.9)	0 (0)	1 (0.95)
Improved HbA1C	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
Allergic reaction	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
Other a	0 (0)	1 (9.1)	4 (11.8)	2 (4.2)	7 (6.7)
Unknown	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)

AE, adverse effect; GI, gastrointestinal

^aReason was not specified in patient’s chart.

At the end of the study period, 13 of 23 patients (56.5%) in the oral group and 45 of 82 patients (54.9%) in the injectable group achieved an HbA_1C concentration of <7%. However, 17 of 23 patients (73.9%) in the oral group and 63 of 82 patients (76.9%) in the injectable group did achieve an HbA_1C concentration of <7% at any point during the study period. A baseline HbA_1C concentration of <8.5% was associated with achieving an HbA_1C concentration of <7% at any point during the study period in both the oral and injectable groups, respectively (P = 0.043 and P = 0.001; see online Supplementary Table S1). Also, patients who did not receive insulin were more likely to achieve an HbA_1C concentration of <7%. No other variables such as baseline body mass index (BMI), baseline weight, patient sex, sulfonylurea use, or a pharmacist embedded in the clinic were associated with achievement of an HbA_1C concentration of <7% in univariate exploratory analyses.

Additionally, at the end of the study period, 9 of 23 patients (39.1%) in the oral group and 45 of 82 patients (54.9%) in the injectable group achieved clinically significant weight loss (≥5% of total body weight). However, 11 of 23 patients (47.8%) in the oral group and 54 of 82 patients (65.9%) in the injectable group did achieve a weight loss of ≥5% at any point during the study period. Additionally, a baseline HbA_1C concentration of ≥8.5% was associated with achieving a clinically significant weight loss during the study period in the oral group but not in the injectable group (P = 0.044; see online Supplementary Table S2). No other variables, such as baseline BMI, baseline weight, patient sex, insulin use, sulfonylurea use, or a pharmacist embedded in the clinic, were associated with achievement of clinically significant weight loss in univariate exploratory analyses.

Discussion

In this retrospective observational review of patients with T2D, our results demonstrated that patients treated with either oral or injectable semaglutide experienced reductions in HbA_1C concentration and weight at 26 weeks. When directly comparing between semaglutide formulations, there were greater reductions in HbA_1C concentration in the oral group and greater reductions in weight in the injectable group, but these results were not statistically significant. To our knowledge, this is one of the first real-world studies directly comparing oral and injectable semaglutide.

The design of this study was unique given that we included patients on other concomitant antihyperglycemic medications, with the goal of assessing outcomes in a way that is applicable to a realistic clinical setting. Our primary outcome evaluated changes at 26 weeks, which is similar to the retrospective study conducted by Chowdhury et al, in addition to the majority of studies in the PIONEER trial series for oral semaglutide, but it differs from the SUSTAIN trial series for injectable semaglutide, which assessed outcomes at 30 weeks or longer. The 26-week endpoints evaluated in this study add unique data to the injectable semaglutide literature and also may help to guide real-world treatment decisions that are likely to be made at 3- and 6-month time intervals as opposed to 12 months. In addition, our study population was similar demographically to those included in the PIONEER and SUSTAIN trial series, with most patients being white males with a baseline HbA_1C concentration of ≥8% and baseline BMI of ≥32 kg/m². Our study also was similar demographically to the retrospective study conducted by Chowdhury et al in terms of the baseline HbA_1C concentration of ≥8% and baseline BMI of ≥32 kg/m², although most of the patients in this study were female.

Patients in our study who were treated with oral semaglutide saw greater numerical reductions in HbA_1C concentration at 26 weeks compared with those treated with injectable semaglutide, but these results were not statistically significant. Although the differences were not significant, it was unexpected to see a larger numerical reduction in HbA_1C concentration in the oral group because this is inconsistent with other available literature demonstrating larger reductions in HbA_1C concentration with injectable semaglutide. ¹² The average HbA_1C concentration reduction seen in our study among patients treated with oral semaglutide was larger than what was reported with oral semaglutide (14 mg) in the PIONEER trials (–1.8% vs −1.1%, respectively). In contrast, the average HbA_1C concentration reduction shown in our study in the injectable semaglutide group was slightly less than what was demonstrated with injectable semaglutide (1 mg) in the SUSTAIN trial but was relatively comparable (–1.4% vs −1.56%, respectively).^5,8 Our study results were inconsistent with the current literature due to the larger numerical reductions that were seen in the oral semaglutide group, with most of the larger HbA_1C concentration reductions among patients treated with oral semaglutide driven mainly by the 14-mg group versus the 7-mg group (–2.43% vs −0.85%, respectively). Although, when comparing the results of our study with those of the retrospective study conducted by Chowdhury et al, the average HbA_1C concentration reductions seen in the oral semaglutide group were comparable overall. The number of patients in our study in the oral semaglutide 14-mg group likely contributes to the unexpectedly large numerical reduction in HbA_1C concentration. Additionally, patients in our study treated with 14-mg daily oral semaglutide had the highest baseline HbA_1C concentration and the largest percentage of those on other antihyperglycemic medications, specifically SGLT-2 inhibitors (63.3%), which likely contributed to the more robust reduction in HbA_1C concentration.^15,16

In total, a lower percentage of patients in our study achieved an HbA_1C concentration of <7% in both the oral and injectable semaglutide groups (56.5% and 54.9%, respectively) at the completion of the study compared with those in the PIONEER and SUSTAIN trials (72.9% and 73%, respectively). However, when compared with patients who achieved an HbA_1C concentration of <7% at any point during the study, our results are similar to those reports. Still, it is possible that the baseline HbA_1C concentration differences seen between our study (8.4%) and the PIONEER and SUSTAIN trials (~8%) may have influenced these differences.^5,8 Due to the retrospective nature of our study and small sample size compared with the PIONEER and SUSTAIN trials, there are likely several confounders that we were unable to control for that contributed to the smaller percentage of patients achieving HbA_1C concentration goals. We are unable to determine whether patients treated with oral or injectable semaglutide would have experienced continued reductions in HbA_1C concentration if they were treated for >26 weeks.

In exploratory analyses, we identified that not receiving insulin was associated with the achievement of an HbA_1C concentration of <7% at any point during the study period. These results are different from those of other published analyses, possibly due to the low sample size in our study. Specifically, an analysis published by Gamble et al demonstrated a larger reduction in HbA_1C concentration when insulin was added to a patient’s current antihyperglycemic regimen, although this finding did not reach statistical significance (relative risk [RR] = 1.10; 95% CI, 0.80-1.52). ¹⁷ The PIONEER 8 and SUSTAIN 5 trials demonstrated that semaglutide in addition to insulin resulted in greater HbA_1C concentration reductions (–1.2% and −1.8%, respectively) than insulin alone (–0.1%).^18,19 Furthermore, the PIONEER 1 and SUSTAIN 1 trials demonstrated slightly lower numerical reductions in HbA_1C concentration (–1.1% and −1.6%, respectively) in those taking either oral or injectable semaglutide without insulin compared with the HbA_1C concentration reductions seen in PIONEER 8 and SUSTAIN 5 trials in patients taking concomitant semaglutide and insulin, further underscoring the likelihood that patients are more likely to achieve HbA_1C concentration goals when insulin is part of their antihyperglycemic regimen.^18,19 We found a baseline HbA_1C concentration of <8.5% associated with achieving an HbA_1C concentration of <7% in the oral and injectable semaglutide groups, respectively, at any time point during the study in exploratory analyses. This is different from the results reported in a retrospective study conducted by Chowdhury et al that found significantly greater improvements in HbA_1C concentration in patients with a higher HbA_1C concentration at baseline versus those with a lower HbA_1C concentration at baseline for both formulations of semaglutide. ⁶ There were small patient numbers in both analyses, and further studies should be performed to investigate these findings.

Patients in our study who were treated with injectable semaglutide saw greater numerical reductions in weight at 26 weeks compared with those treated with oral semaglutide. The average weight loss seen in our study across both formulations was slightly greater than what was reported in the PIONEER and SUSTAIN trials but was relatively comparable. In total, a higher percentage of patients in our study achieved clinically significant weight loss (≥5% reduction in total body weight) in both the oral and injectable semaglutide groups (39.1% and 54.9%, respectively) compared with what was seen in the PIONEER and SUSTAIN trials (34.1% and 41%, respectively).^5,8 These differences may be due to the fact that the median baseline weight and BMI in our study were notably higher (106.9 kg and 35.7 kg/m²) than in the PIONEER (88 kg and 32 kg/m²) and SUSTAIN (92 kg and 33 kg/m²) trials, which may have allowed for further reduction in weight throughout our trial period.^5,8 This theory is also similar to what was seen in the oral semaglutide group in the retrospective study conducted by Chowdhury et al that demonstrated greater numerical improvement in weight in those with a higher baseline BMI. ⁶ Additionally, our study included a large proportion of patients who were also being treated with metformin (76.2%) and/or an SGLT-2 inhibitor (39%), which are likely to have contributed to additional weight loss. This theory is supported by data from multiple large meta-analyses that showed that the addition of metformin and SGLT-2 inhibitors to existing therapies contributed to additional weight loss in patients with T2D.^15,16

Given the robust amount of literature demonstrating injectable semaglutide’s ability to provide significant weight loss, it was our expectation that treatment with injectable semaglutide would result in more weight loss than its oral counterpart. Results of our study were similar to those of a 2019 meta-analysis conducted by Nuhoho et al that demonstrated that greater reductions in total body weight were seen with injectable semaglutide compared with those seen in patients treated with oral semaglutide. ¹² Due to the fact that our study only assessed outcomes up to 26 weeks, we are unable to determine whether patients treated with either formulation of semaglutide would have experienced continued weight loss over a longer duration of time.

The rate of reported AEs in our study was much lower than in the PIONEER and SUSTAIN trials (5% vs 54.9% and 60%, respectively).^5,8 In this study, 2 patients (8.7%) treated with oral semaglutide, both of whom received the 7-mg daily dose, and 4 patients (4.9%) treated with injectable semaglutide, 3 of whom were receiving injectable semaglutide 0.5 mg and 1 of whom was receiving injectable semaglutide 1 mg, reported experiencing >1 medication-related AEs during the study period. Our results are similar to those of a 2021 systematic review that reported similar rates of AEs in patients treated with either oral or injectable semaglutide. ²⁰ Additionally, this review found that the proportion of patients with premature discontinuation due to AEs was higher in patients taking oral semaglutide than in those taking injectable semaglutide, which is similar to the results of our study. Patients in this study had been treated with the initial doses of oral and injectable semaglutide for at least 1 month, essentially serving as a run-in period that may have eliminated many patients who did experience or would have experienced medication-related AEs. Additionally, our study relied on AEs documented in the medical record, whereas both the PIONEER and SUSTAIN trials conducted formal adverse drug reaction screenings, which could explain the large difference in reported AEs and contribute to the absence of AEs seen within the oral semaglutide 14-mg group. Similar to the retrospective study conducted by Chowdhury et al, in addition to the PIONEER and SUSTAIN trials, gastrointestinal side effects were the most commonly reported AEs in our study. No patients required dose reductions in either the oral or injectable groups, and only 4 patients (3.8%) discontinued therapy due to an in this study, with the highest rate of discontinuation observed in the oral semaglutide 7-mg group (8.3%).

The key strength of this study is that our patient population and study design are representative of a real-world clinical setting, including patients with various baseline HbA_1C concentration, weight, and BMI values who are taking other concomitant antihyperglycemic medications and have a history of prior GLP-1 RA and/or insulin use. Additionally, our study collected data at a consistent duration of follow-up for each patient and adds a unique follow-up timeframe of 26 weeks, which is currently missing from the injectable semaglutide literature.

Despite the strengths of our study, there also were limitations. Our single-center study was retrospective in nature and thus depended on accurate documentation in the electronic medical record. Because this study period took place during the COVID-19 global pandemic and in-person visits were limited, the ability to obtain outcome measurements such as HbA_1C concentration and weight values was affected, therefore limiting the sample size due to a lack of patients with available follow-up data. Primary outcomes were predefined, but due to the number of statistical tests completed, it is acknowledged that the chance for type I error is high, and this should be considered as hypothesis generating only. Due to the small sample size, we believe that this limited the ability to establish statistical significance when comparing change in HbA_1C concentration and weight between formulations and likely contributed to the unexpectedly large numerical HbA_1C concentration decrease seen in the oral semaglutide 14-mg group. Data were not collected for patients who may have started on additional antihyperglycemic agents during the study period, which likely increased the number of confounding factors impacting weight and/or HbA_1C concentration values. Although most of the data for outcomes were assessed manually, all data were collected by 2 researchers to ensure consistency and reduce the chance of error. Lastly, outcomes assessing semaglutide 2-mg once-weekly injection were not included in this study due to the approval of the 2-mg dose occurring toward the end of our study timeframe.

In the future, large-scale real-world studies are needed to confirm our findings for patients treated with oral or injectable semaglutide and potentially compare treatment effects of each formulation within the same patient population. Moreover, a study assessing the real-world impact of the semaglutide 2-mg once-weekly injection on HbA_1C concentration and weight should be considered to determine whether additional HbA_1C concentration and weight reductions occur in comparison with the doses of oral and injectable semaglutide discussed in this study. This is of increasing clinical relevance given the growing evidence that suggests a benefit in targeting higher percent weight reductions from baseline (ie, >15%) for improved patient outcomes. ²¹

Conclusion

When directly comparing the oral and injectable formulations of semaglutide, there were no statistically significant differences in change in HbA_1C concentration and weight from baseline to 26 weeks between the 2 groups. However, patients treated with oral semaglutide experienced numerically larger reductions in HbA_1C concentration, whereas patients treated with injectable semaglutide had numerically larger reductions in weight. No significant safety concerns were identified with either formulation of semaglutide. The real-world clinical setting and inclusion of patients being treated with additional antihyperglycemic medications make the results of this study relevant to a broad patient population and add necessary data to guide clinicians when selecting between oral and injectable semaglutide.