Collected patient data included primary data originating from patient-reported outcomes (PRO) comprising the Audit of Diabetes-Dependent Quality of Life (ADDQoL), the Diabetes Treatment Satisfaction Questionnaire (DTSQ), and the Hypoglycemia Fear Survey-II (HFS-II) worry subscale, and relevant medical data collected on individual case report forms, including comorbidities and vital signs, disease-related variables at inclusion, current diabetes pharmacotherapy, significant non-drug therapies for diabetes, concomitant medication, laboratory values obtained prior to study visit, and the treating physician’s specialty.

According to the WHO, Body Mass Index (BMI) is used to classify a subject under one of the following categories: BMI < 18.5 kg/m² indicates underweight, 18.5 ≤ BMI < 25 kg/m² is normal range, 25 ≤ BMI < 30 kg/m² indicates overweight and BMI ≥ 30 kg/m² indicates obesity.

The ADDQoL questionnaire evaluates patients QoL and is composed of two items, the first relating to the patient’s general QoL and the second relating to the effect of diabetes on patients’ QoL, followed by a series of 19 questions. The DTSQ questionnaire measures treatment satisfaction and its score values can range from 0 to 36, where 0 indicates complete dissatisfaction with the treatment and 36 full satisfaction. The HFS-II worry subscale questionnaire measures the worrying and fear of diabetes patients from hypoglycemia events.

The sample size was calculated according to the standard deviation expected for the DTSQs, which is the questionnaire having shown the highest variability [1]. Under the hypothesis of normality, the formula used to estimate the required sample size for one primary outcome measure is: N = (2 × 1.96 × SD/width of 95% CI) 2. According to this formula, a sample size of 314 patients per country is required to estimate the mean score obtained for DTSQs (SD = 6.1) with a required precision set to 1.35 using a two-sided 95% CI. Assuming that approximately 30% of enrolled patients would not be evaluable for the primary outcome measures, 410 patients per country were planned to be enrolled in this study. Analysis was done on the reference population of both countries, defined as all patients included in the study who signed the ICF and who fulfilled eligibility criteria. The means of quantitative variables were compared using Student T-test or Wilcoxon-Whitney test depending on the normality of the data. The association between two qualitative variables was tested using the Chi-square test or the Fisher’s exact test. Qualitative variables were described in terms of number of observed values, number of missing data, frequencies and percentages. Quantitative variables were described in terms of number of observed values, number of missing data, mean, standard deviation (SD), median, lower and upper quartile, minimum and maximum. Differences were considered statistically significant when P value was <0.05. No adjustment for multiplicity was performed. Mixed model linear regression was used to provide the estimates of the mean difference between outcomes (ADDQol, DTSQs, HFS-II worry subscale scores), according to treatment subgroups, diabetes control, or comorbidities. The models were adjusted for site and for other variables in the model (patient-level variables, physician-level variables). All analyses were conducted using SAS (version 9.4; SAS Institute).

Investigators have recruited 825 patients who signed the ICF in 91 centers in Lebanon and Jordan. Only 694 (84.1%) patients were eligible for analysis (reference population, Figure 1).

Patients’ age ranged from 22 to 87 years with an average of 57.0 ± 11.0 years. Most of the patients (68.4%) aged between 40 and 60 years, whereas 26.4% of them were older than 65 years and 5.2% younger than 40 years. Other socio-demographic and lifestyle-related data as well as clinical characteristics are reported in Table 1. Over 75% of patients received education about T2DM and

its complications mostly twice per year. Education and awareness material were mostly provided by the treating physician (498 [92.1%] patients), followed by nutritionists, nurses and pharmacists.

Most patients were either overweight (237 [39.5%]) or obese (335 [48.4%]). Average systolic blood pressure (BP) was 131.0 ± 14.7 mmHg and average diastolic BP was 79.4 ± 9.0 mmHg, indicating an overall well-controlled BP in T2DM patients. Renal function was also assessed and serum creatinine levels showed an average of 97.3 ± 106.1 µmol/L for 573 patients. Most of the patients (85.2%) had their serum creatinine levels below 106.1 µmol/L, and only 14 patients (2.4%) above 353.7 µmol/L. Values for urinary albumin levels as well as the albumin-creatinine ratio were not conclusive for the reference population as they were only collected in 117 and 162 patients, respectively, showing a suboptimal screening for albuminuria by physicians. Patients’ lipid profile and liver function tests are displayed in Table 2.

The majority of patients had a family history of diabetes, mostly from their maternal side (62.7% out of 582 patients). Patients had had diabetes for less than five years (32%), for five to nine years (25%), and for over nine years (43%). Most of the patients (70.1%) practiced SMBG at a rate of 15.7 ± 14.1 times per month. At least one hypoglycemia event within the last year prior to the study visit was reported by 201 (29.0%) patients and the average number was 8.3 ± 14.1 hypoglycemia events per patient per year. All 694 patients of the reference population were taking anti-diabetes medications, 510 (73.5%) received exclusively oral antidiabetics (OADs) and 28 (4.0%) received only injectable medications, while 156 (22.5%) received a combination treatment of OADs and injectable medications. Figure 2 summarizes treatment pattern of patients on oral medications; over 40% of patients on OADs were treated with a combination of two medications (Figure 2(A)), mainly metformin combined with either sulfonylurea or with a dipeptidyl peptidase 4 (DPP-4) inhibitor (Figure 2(B)).

Injectable medications mainly included insulin and only one patient was treated with a glucagon-like peptide 1 (GLP-1) receptor agonist. Table 3 displays

DPP-4, dipeptidyl peptidase-4; GLP-1, glucagon-like peptide; OADs, oral antidiabetics; SGLT2, sodium-glucose cotransporter 2.

details of the treatment patterns followed by all patients of the reference population. It is worth noting that 20 patients (2.9%) were prescribed non-drug therapies, mainly herbal preparations.

Over 80% of patients were also being treated for other conditions. Most concomitant medications were HMG CoA reductase inhibitors (72.4%), angiotensin receptor blockers (39.3%), platelet aggregation inhibitors (32.9%) beta-blocking agents (21.2%), and angiotensin-converting enzyme inhibitors (13.0%).

Physicians set targets for HbA1c values, on a case-by-case basis, ranging from 5.0% (31 mmol/mol) to 10.0% (86 mmol/mol). The average HbA1c target set by physicians was of 6.8% ± 0.5% (51.3 ± 5.4 mmol/mol) for all patients. Physician-set targets were reached by 268 (38.6%) patients, and 245 (35.3%) patients reached the absolute HbA1c target of 7.0% (53 mmol/mol). The three main reasons for failing to reach HbA1c targets were poor adherence to diet and/or physical exercise recommendations (85.9%), poor adherence to SMBG practice (48.8%), and failure of the current antidiabetic treatment regimen (38.7%). On average, HbA1c values decreased by 0.5% ± 1.5% (5.2 ± 17.0 mmol/mol) in the period extending between 12 months and one month prior to the study visit. Upon comparing HbA1c target achievement among treatment subgroups, results showed that 43.0% of patients treated with OADs reached the HbA1c target set by their physicians, whereas 28.9% of those treated with a combination of OADs and injectable medications achieved their target, and only 14.3% of patients on injectable medications were able to achieve the HbA1c target set by their physicians. The difference between these treatment subgroups was statistically significant (P < 0.001).

To evaluate the QoL and treatment satisfaction of patients, PRO scores were determined according to the ADDQoL and DTSQ questionnaires as well as the HFS-II worry survey. Out of the 694 patients in the reference population, 8 (1.2%) did not complete the ADDQoL questionnaire, 15 (2.2%) did not complete the DTSQ questionnaire, and 19 (2.7%) did not complete the HFS-II worry survey. Average weighted impact score for the ADDQoL questionnaire ranged for 685 patients from −8.8 to 0.5 with a mean of −3.1 ± 1.9. The range boundaries of the score values of the ADDQoL questionnaire are −9 to +3, where −9 indicates utter dissatisfaction and +3 indicates full satisfaction. The score of the ADDQoL first item ranged from −3.0 to 3.0 with a mean of 1.1 ± 1.1, and the score of the second item from −3.0 to 1.0 with a mean of −1.7 ± 0.9. Patients’ answers to these two items are detailed in Table 4. Overall, around 25% of patients described their current QoL as below good, and close to 90% thought that if they did not have diabetes, their QoL would improve to different extents.

Patients scored a mean of 27.6 ± 6.9 (with a range of 1 to 36) on the DTSQ, which evaluates treatment satisfaction. DTSQ questionnaire score values can range from 0 to 36, where 0 indicates complete dissatisfaction with the treatment and 36 full satisfaction. Therefore, an average score of 27.6 ± 6.9 suggests that patients’ remained not fully satisfied about their treatment scheme. When patients were asked in the DTSQ questionnaire about the frequency at which they felt their blood glucose levels were unacceptably high or unacceptably low, the average scores of their answers were 2.8 ± 1.8 and 2.0 ± 1.8, respectively, on a scale from 0 to 6, 0 indicating “none of the time” and 6 indicating “most of the time”.

The average score of the HFS-II worry subscale was 20.5 ± 17.8, on a scale going from 0 (no worry whatsoever) to 72 (extreme worry). The calculated score shows that patients remain far from having no worries at all and are still afraid of hypoglycemia.

PRO scores were then compared among different treatment patterns, HbA1c target achievement status and comorbidities. Mean differences between these groups were calculated after adjusting for patients’ characteristics. Table 5 compares the QoL, treatment satisfaction, and worry of hypoglycemia of patients on injectable and OADs and those on injectable medication to patients on OADs only. Compared to orally treated patients, a statistically significant decrease was noted in ADDQoL scores for the subgroup of patients treated with injectable

ADDQoL: Audit of Diabetes-Dependent Quality of Life; QoL: Quality of Life.

ADDQoL: Audit of diabetes dependent quality of life, DPP-4: dipeptidyl peptidase 4, DTSQ: Diabetes treatment satisfaction questionnaire, HFS-II: Hypoglycemia fear survey-II, OAD: Oral antidiabetics, PRO: Patient-reported outcome.*, P < 0.01; **, P < 0.001.

medication (−1.1) (P = 0.0031) and for those treated with OADs and injectable medications (−0.7) (P < 0.001), irrespective of patients’ characteristics or physicians’ specialty. As far as treatment satisfaction was concerned, a statistically significant decrease (−1.7) was noted in DTSQ scores of patients treated with OADs and injectable medications compared to patients treated with only OADs (P = 0.0097). Concerning HFS-II worry survey scores, they were higher in patients treated with injectable medication only (+7.81) when compared to those treated with OADs (P = 0.0053).

Among patients treated with OADs, ADDQoL scores showed that those treated with a combination of metformin and DPP-4 inhibitors, scoring −2.7 ± 1.7, were more satisfied than those treated with metformin and sulfonylureas, scoring −3.2 ± 1.5 (P = 0.039). The DTSQ questionnaire also showed a better QoL reported by patients treated with metformin and DPP 4 inhibitors who scored 28.7 ± 6.1 compared to those treated with metformin and sulfonylureas who scored 25.6 ± 7.2 (P = 0.002), and it was clear that the perception of hyperglycemia frequency was lower when patients were treated with metformin and DPP-4 inhibitors (P = 0.007). HFS worry survey revealed results pointing in the same direction as the previous questionnaires and showing that patients treated with metformin and DPP4-inhibitors attributed a lower effect of diabetes on their QoL than those treated with metformin and sulfonylureas, visible by respective scores of 15.2 ± 15.9 and 20.4 ± 19.0. However, this difference was not statistically significant (P = 0.060).

On the other hand, ADDQoL questionnaire, DTSQ questionnaire and HFS-II worry subscale scores were all significantly different in patients achieving their HbA1c target compared to those who did not achieved it (−0.6 [P = 0.0016], −2.5 [P < 0.001] and +3.3 [P = 0.0071], respectively). However, comorbidities did not seem to majorly impact PRO scores. Indeed, only three conditions are suggested to have a certain effect on patients’ perceptions. Genitourinary infection compromised patient satisfaction, with a decreased DTSQ score (−2.0; P = 0.02), diabetic retinopathy was associated with enhanced anxiety, reflected by an increase in HFS-II worry scale survey score (+3.3; P = 0.04), and chronic kidney disease reduced QoL, as revealed by a decrease in the ADDQoL score (−0.4; P = 0.04).