In this real-world study of 94,827 French smokers seeking help to quit in a SCS, one-month continued abstinence rates were 33% among diabetic smokers and higher among non-diabetic smokers at 38% (p < 0.001). Consistently, evidence suggests that diabetic smokers are less likely to quit smoking when compared to non-diabetic smokers [11,12,28]. Moreover, diabetic smokers are less likely to be active in self-care or to comply with diabetes care recommendations [28]. This trend is reflected in the present study, in which diabetic smokers were less likely to seek help to quit on their own through SCSs compared to their non-diabetic counterparts (26% versus 46%; p < 0.001). Furthermore, diabetic smokers in this study were more frequently referred to a SCS after hospital contact than non-diabetic smokers (55% versus 34%; p < 0.001). This may be related to the higher prevalence of comorbidities among diabetic smokers, including cardiovascular risk factors, CVDs, and respiratory diseases, as well as heavier smoking patterns compared to non-diabetic smokers. Diabetic smokers were also significantly older than their non-diabetic counterparts in the current study (mean age, 54 versus 46 years; p < 0.001). This difference is in line with expectations, given that the mean age at diabetes diagnosis is 50 years [29]. Hence, smokers with diabetes might, on average, be older than those without diabetes. Of note, an earlier analysis of the CDTnet registry revealed that smokers referred to French cessation services were heavier smokers than smokers in the general population [30].

Our study highlighted several predictors influencing smoking abstinence in diabetic smokers, with or without CVDs, including employment status, prior quit attempts, low nicotine dependence, high confidence in quitting, prescription of combined NRT or varenicline as smoking cessation treatments, and multiple follow-up consultations. These results are in line with the findings of an online cross-sectional survey involving 201 smokers and ex-smokers in Australia, which highlighted the crucial role of follow-up and support in smoking cessation management as well as the effectiveness of varenicline prescription in succeeding to quit [31]. In alignment with prior research [32], we found that diabetic smokers are more likely to experience socioeconomic disadvantages, in terms of employment and education, compared to non-diabetic smokers. The impact of socioeconomic factors on smoking cessation is well-established, with individuals from lower socioeconomic backgrounds facing challenges in successfully quitting smoking, irrespective of their diabetes status [12,14,33–35]. Socioeconomic factors can influence access to resources, social support, and the ability to cope with stress, all of which can impact the success of smoking cessation efforts [12,33]. Hence, to reduce inequity in smoking cessation outcomes, multifaceted interventions are required at individual, community, and population levels, recognising the wider context of socioeconomically disadvantaged smokers and the interplay between socioeconomic status, diabetes, and smoking [12,33].

Within our study population, individuals with ≥3 prior attempts to quit smoking exhibited an approximately 60% increased likelihood of achieving smoking abstinence, regardless of diabetes status or the presence of CVDs. Smoking cessation is a chronic and dynamic process that often produces a complex back-and-forth pattern of quit attempts [36,37]. Indeed, the estimated average number of quit attempts expected before achieving long-term tobacco abstinence is 6 [38]. The most effective way to achieve smoking abstinence among diabetic smokers is to combine both behavioural and pharmacologic therapies [1,39]. Behavioural interventions are particularly important to improve self-confidence to quit [40], which has been found in our study to have a significantly positive impact on smoking abstinence. There is evidence of a robust dose-response relationship, wherein more intensive behavioural and psychological treatments, characterised by higher amounts of contact time and more sessions, yield a greater odds of sustained smoking cessation [9]. This relationship is further underscored in the present study, with the number of follow-up consultations emerging as the most impactful factor on smoking abstinence. Notably, the OR exceeded 3 for both diabetic and non-diabetic smokers, reaching 5.29 in diabetic smokers without CVDs who attended ≥7 follow-up consultations at a SCS.

The recommended first-line pharmacologic therapies for smoking cessation in both diabetic and general populations include combined NRT, varenicline, and bupropion [1,39,40]. In our study population, bupropion was rarely used (0.2%), potentially due to the fact that bupropion is not reimbursed in France [1]. Combined NRT and varenicline are considered similar in efficacy, and are regarded as the two most effective smoking cessation aids currently available [40]. Notably, combined NRT was frequently prescribed in our study, with a usage rate of 45% versus up to 17% for a single form of NRT. The rationale for combined NRT is that long-acting transdermal nicotine patches can provide a stable level of nicotine necessary to achieve and sustain cessation, while short-acting oral forms of NRT can deal with emergent nicotine cravings [1]. NRT prescription practices in French cessation services align with the literature, which has demonstrated an increased benefit on abstinence with combined NRT versus single-form NRT [41,42]. Importantly, our study discourages the use of oral nicotine substitute as the sole smoking cessation medication, as it was associated with a 30% lower likelihood of smoking abstinence among diabetic smokers.

Other negative factors for smoking abstinence among diabetic smokers in the current study included a history of MI/angina, chronic bronchitis/COPD, a history of depression, anxiety symptoms, and cannabis use within the last 30 days. In particular, the only factor that differed significantly between diabetic and non-diabetic smokers was the presence of MI/angina, highlighting the need for comprehensive smoking cessation management in diabetic smokers with MI/angina. Remarkably, earlier studies have also shown that smokers with pre-existing heart or respiratory disease were less likely to quit smoking [11,12]. One plausible explanation is that the motivation to quit smoking tends to be higher when individuals perceive a clinical trigger event as life-threatening. This motivation may be less pronounced in those who have already experienced a cardiac or respiratory event in the past [11,12]. In addition, weight gain after quitting is a common concern for individuals with diabetes and CVDs, as it may lead to poor glycemic control and increases the risk of complications [12,43,44]. Interestingly, among non-diabetic smokers in our study, having a BMI ≥ 25 kg/m² was associated with a 12% higher likelihood of successful smoking cessation (95% CI, 1.04–1.20; p = 0.002). This suggests that individuals with higher BMI are less concerned about weight gain associated with quit attempts [11]. However, this association was not observed among diabetic smokers. Nonetheless, in individuals with diabetes, the cardiovascular health and mortality benefits of smoking cessation outweigh concerns about weight gain [45]. Therefore, a proactive approach is needed to manage weight gain among diabetic individuals attempting to quit smoking, especially those who smoke heavily or are prone to binge eating [45].

Smoking quit rates are also generally lower in individuals with mental health conditions, since such conditions can complicate smoking cessation efforts by affecting motivation, coping mechanisms, and overall well-being [12,28,40]. Similarly, another analysis of CDTnet revealed that recent cannabis users in France were less likely to quit smoking [14]. This underscores the potential influence of concurrent substance use, like tobacco and cannabis, on smoking behaviour and suggests that addressing multiple substance use is crucial for successful smoking cessation [46].

This study has strengths and limitations. As in any observational study, causality cannot be established, and unmeasured confounding factors could influence the observed associations. Additionally, due to the large sample size, some statistically significant results with small absolute differences should be interpreted cautiously. Moreover, our study population consisted of treatment-seeking patients attending ≥2 visits to a SCS. These results may not apply to the general diabetic population of smokers. Furthermore, although all collected information was validated by the SCS staff before being registered in CDTnet, reliance on self-reported data may introduce recall bias. Another limitation of our study is the inability to include racial or ethnic data as predictors, due to French legal restrictions that prohibit the collection of information directly revealing racial or ethnic origins. Nevertheless, our study is strengthened by the inclusion of every region in France, including overseas territories, in the CDTnet registry, which ensures a geographically diverse and nationally representative sample. In addition, smoking cessation was based on validated abstinence through CO measurement at each follow-up visit, offering a reliable assessment compared to self-reported cessation.