Diabetes is associated with a variety of pulmonary disorders beyond its associations with sleep apnea discussed in the previous commentary. The Fremantle Diabetes Study in Western Australia showed that persons with type 2 diabetes (T2D) had 5%–12% reduction in vital capacity and expiratory flow rates with evidence of worsening over a 7-year period of observation; in follow-up, each 10% reduction in the forced expiratory volume in 1 s (FEV1) was associated with a 13% increase in mortality controlling for age, sex, duration, hypertension, retinopathy, neuropathy, albuminuria, and coronary disease.¹ This relationship extends to prediabetes. Among 2332 initially nondiabetic persons in Malmö, Sweden, homeostatic model assessment of insulin resistance (HOMA-IR) was elevated at 10-year follow-up in approximately one third of those in the lowest quartile of baseline forced vital capacity (FVC), and each 10% greater FVC was associated with a 10% lower likelihood of IR and with 10% lower likelihood of diabetes, adjusted for age, smoking, and body mass index (BMI), while cardiovascular disease (CVD) events significantly increased among those persons with FVC below the median who had developed insulin resistance.² Meta-analyses of 39 studies of 1274 persons with T1D and 1353 controls and of 66 studies of 11 134 persons with T2D and 48 377 controls both showed 6%–10% lower pulmonary flow rates than those among controls.^{3, 4} The UK Biobank Study compared 372 093 nondiabetic persons (glycosylated hemoglobin [HbA1c] < 5.7), 53 378 with prediabetes (HbA1c 5.7–6.4), and 27 209 with diabetes (HbA1c ≥ 6.5), adjusting for factors including age, sex, ethnicity, BMI, education, and cigarette and alcohol use; those with prediabetes and diabetes had a 10-year chronic obstructive pulmonary disease (COPD) relative risk of 1.18 (1.13–1.24) and 1.35 (1.24–1.47), respectively, with preexisting diabetes associated with a ~1.2-fold greater risk than that among those diagnosed at the time of study entry, and with COPD-specific mortality nearly doubled among those with diabetes diagnosed ≥7 years prior to study entry.⁵ In a study of 48 nondiabetic persons, 68 with prediabetes, 29 newly diagnosed T2D, and 110 with long-term T2D, symptoms of breathlessness were present in none, 3%, 10%, and 15% of the respective groups, along with reduction in FVC and pulmonary diffusing capacity.⁶

It is, then, to be expected that a variety of lung diseases occur in association with diabetes. Asthma prevalence is more than doubled with diabetes, with greater severity and with evidence of airway inflammation and hyperresponsiveness. COPD also shows greater severity among people with diabetes, in part related to obesity and to systemic inflammation. People with idiopathic pulmonary fibrosis (IPF) and with pulmonary hypertension (PH) are more likely to have diabetes, and although it is not clear that their prevalences are greater among people having diabetes, there is evidence that diabetes is associated with greater severity of these conditions as well.⁷ COPD is associated with many risk mediators in common with those of diabetes, including cigarette use, aging, hypertension, dyslipidemia, and insulin resistance, and there is evidence that COPD is associated with low-grade systemic inflammation and to hypercoagulability, and that exacerbations of COPD may be related to heart failure and to coronary heart disease, with elevation in troponin and in basic natriuretic peptide (BNP) often accompanying such episodes.⁸ A meta-analysis comparing 25 180 persons with IPF to 73 434 controls showed the likelihood of diabetes to be 1.54-fold greater in the former group, although evidence of diabetes as a risk factor for IPF was not found.⁹ Part of the difficulty may be in underdiagnosis, with a study of 53 persons with T2D and unexplained exertional dyspnea showing lower peak oxygen uptake during exercise and higher mean pulmonary artery pressure.¹⁰ The association of IPF severity with diabetes may be related to oxidative stress, to endothelial dysfunction, to inflammation, to obesity, to diaphragmatic muscle dysfunction, to autonomic neuropathy, and to albuminuria.^{11, 12} Insulin resistance has been proposed as underlying the relationship between diabetes and PH, with consequent inflammation, dyslipidemia, and endothelial dysfunction leading to adverse pulmonary vascular remodeling and to right ventricular dysfunction.¹³ In the Fremantle Diabetes Study, analysis of the subset of persons with T2D who had echocardiograms obtained for clinical reasons showed 6.4% and 2.6% prevalence of estimated right ventricular systolic pressure >30 and >40, respectively, and over 6.6-year follow-up an additional 9.2% and 5.0% developed this degree of PH, suggesting that the risk of PH is approximately 40% greater among persons with diabetes.¹⁴ A study of patients with PH having versus not having diabetes showed that the former had higher BNP, shorter 6-min walking distance, more dyspnea, higher pulmonary artery pressure, and shorter survival,¹⁵ and a study of 110 563 persons with newly diagnosed PH in the US national Veterans Health Affairs database showed that more than one third had diabetes, which was associated with a 1.21-fold greater mortality risk.¹⁶ Genetic association studies in a population of 14 861 persons with echocardiogram-measured pulmonary pressure and right ventricular function suggest both T2D and obesity to be associated with greater levels of tricuspid regurgitation and right ventricular systolic pressure, suggesting both to play roles in the development of PH.¹⁷

A number of approaches to diabetes treatment may have pulmonary benefit. It should be noted that epidemiologic study findings need to be regarded with skepticism, a caution which was recently raised in analysis of potential biases affecting the evidence that metformin may lead to reduction in cancer and/or to improved outcome of persons being treated for cancer.¹⁸ Bearing in mind this caveat, we can review several interesting reports. In a study of 3599 adults with both IPF and T2D, controlling for age, sex, race/ethnicity, residence region, year, medications, oxygen use, smoking status, healthcare use, and comorbidities, those treated with metformin had 54% lower mortality and 18% lower risk of hospitalization.¹⁹ In a study of 350 536 persons with new-onset diabetes not having COPD, pioglitazone treatment was associated with lower likelihood of development of COPD, particularly among those with longer duration of pioglitazone treatment.²⁰ Thiazolidinediones are, however, complex agents to administer to people with cardiac and/or pulmonary disease, with a study of hospitalizations among 402 153 persons with both T2D and COPD showing those treated with a thiazolidinedione having >50% increase in likelihood of CVD and of heart failure events, as well as increased risks of development of bacterial pneumonia, of requirement for noninvasive positive pressure ventilation during hospitalization, and of development of lung cancer.²¹

In the monocrotaline-treated rat PH model, the sodium-glucose co-transporter-2 inhibitor (SGLT2i) empagliflozin reduced right ventricular and pulmonary artery pressures, decreased right ventricular hypertrophy and fibrosis, and was associated with improved survival.²² A 12-week trial of 65 persons with heart failure and an implanted pulmonary artery pressure sensor revealed that those randomized to receive empagliflozin had a reduction in pulmonary artery diastolic pressure.²³ In an epidemiologic study of persons with diabetes and COPD using the UK Clinical Practice Research dataset, those treated with a SGLT2i had a 41% reduction in COPD exacerbations requiring hospitalization in comparison to those treated with a sulfonylurea, although moderate exacerbations treated on an outpatient basis were not more common.²⁴

The greatest degree of improvement appears to be seen with the use of glucagon-like peptide 1 receptor agonists (GLP-1RA). GLP-1 receptor expression in the lungs is greater than that in most other tissues, and administration of GLP-1RA reduces the pulmonary response to inhaled allergens in a mouse model and decreases bronchial hyperresponsiveness and inflammatory markers in animal models.²⁵ In the epidemiologic study using the UK Clinical Practice Research dataset, severe and moderate COPD exacerbations were reduced by 41% and by 48%, respectively.²⁴ A similar study from the Mass General Brigham dataset showed the greatest reduction in both severe and moderate COPD exacerbations with GLP-1RA treatment, with SGLT2i appearing to be associated with benefit as well.²⁶ A third study compared the risk of chronic lower respiratory disease exacerbations with GLP-1RA versus dipeptidyl peptidase IV inhibitor (DPP-IVi) in 4150 and 12 540 persons, respectively, finding a 48% reduction with GLP-1RA in likelihood of hospitalization and a 30% reduction in total pulmonary exacerbations.²⁷ A meta-analysis of 28 randomized controlled trials of GLP-1RA involving 77 485 participants showed a 14% decrease in likelihood of overall respiratory disease with a 34% lower likelihood of pulmonary edema and a 14% lower likelihood of bronchitis.²⁸

In summary, there is convincing evidence that diabetes is associated with a decrease in pulmonary function. A variety of mechanisms appear to be involved, including insulin resistance, endothelial dysfunction, inflammation, and effects of obesity. PH, pulmonary fibrosis, asthma, and COPD all have adverse associations with diabetes. A variety of diabetes treatment approaches appear of benefit in treating these pulmonary conditions, with experimental evidence in animal studies and randomized controlled trials of benefit of SGLT2i, and with strong epidemiologic evidence of benefit of the GLP-1RA class.

除了之前评论中讨论的与睡眠呼吸暂停的关系之外，糖尿病还与多种肺部疾病有关。西澳大利亚弗里曼特尔糖尿病研究表明，2 型糖尿病 (T2D) 患者的肺活量和呼气流速降低了 5%–12%，并且在 7 年的观察期内有恶化的证据；在随访中，在控制了年龄、性别、病程、高血压、视网膜病变、神经病变、蛋白尿和冠心病的情况下，1秒用力呼气量（FEV1）每减少10%，死亡率就会增加13%。¹这种关系也适用于糖尿病前期。在瑞典马尔默的 2332 名最初非糖尿病患者中，胰岛素抵抗稳态模型评估 (HOMA-IR) 在 10 年随访中升高，其中约三分之一处于基线用力肺活量 (FVC) 最低四分位数的患者中，并且根据年龄、吸烟情况和体重指数 (BMI) 进行调整后，用力肺活量每增加 10%，患 IR 的可能性就会降低 10%，患糖尿病的可能性就会降低 10%，而这些人中心血管疾病 (CVD) 事件显着增加FVC 低于中位数的人已出现胰岛素抵抗。²对 1274 名 T1D 患者和 1353 名对照者的 39 项研究以及针对 11 134 名 T2D 患者和 48 377 名对照者的 66 项研究进行的荟萃分析均显示，肺流速比对照者低 6%–10%。^{3, 4}英国生物银行研究比较了 372 093 名非糖尿病患者（糖化血红蛋白 [HbA1c] < 5.7）、53 378 名糖尿病前期患者（HbA1c 5.7–6.4）和 27 209 名糖尿病患者（HbA1c ≥ 6.5），并调整了包括年龄、性别、种族、体重指数、教育程度以及吸烟和饮酒情况；糖尿病前期和糖尿病患者的 10 年慢性阻塞性肺病 (COPD) 相对风险分别为 1.18 (1.13–1.24) 和 1.35 (1.24–1.47)，而既往患有糖尿病的患者的风险约为 1.2 倍在进入研究时诊断的患者中，在进入研究前 7 年以上诊断为糖尿病的患者中，COPD 特异性死亡率几乎翻了一番。⁵在一项针对 48 名非糖尿病患者、68 名糖尿病前期患者、29 名新诊断的 T2D 患者和 110 名长期 T2D 患者的研究中，各组中无呼吸困难症状、3%、10% 和 15% FVC 和肺弥散能力降低。⁶

因此，可以预见多种肺部疾病的发生与糖尿病有关。糖尿病患者的哮喘患病率增加了一倍多，且严重程度更高，并且有气道炎症和高反应性的证据。慢性阻塞性肺病在糖尿病患者中也表现得更为严重，部分与肥胖和全身炎症有关。患有特发性肺纤维化 (IPF) 和肺动脉高压 (PH) 的人更容易患糖尿病，尽管尚不清楚糖尿病患者中这些疾病的患病率是否更高，但有证据表明糖尿病与这些疾病的严重程度有关条件也。⁷ COPD 与许多与糖尿病相同的风险介质相关，包括吸烟、衰老、高血压、血脂异常和胰岛素抵抗，并且有证据表明 COPD 与低度全身炎症和高凝状态相关，并且病情加重COPD 的发生可能与心力衰竭和冠心病有关，肌钙蛋白和碱性钠尿肽 (BNP) 升高通常伴随此类发作。⁸一项对 25 180 名 IPF 患者与 73 434 名对照者进行比较的荟萃分析显示，前一组患者患糖尿病的可能性是前者的 1.54 倍，尽管没有发现糖尿病是 IPF 危险因素的证据。⁹部分困难可能在于诊断不足，对 53 名患有无法解释的劳力性呼吸困难的 2 型糖尿病患者的一项研究显示，运动期间峰值摄氧量较低，平均肺动脉压较高。¹⁰ IPF 严重程度与糖尿病的关联可能与氧化应激、内皮功能障碍、炎症、肥胖、膈肌功能障碍、自主神经病变和蛋白尿有关。^{11, 12}胰岛素抵抗被认为是糖尿病和 PH 之间关系的基础，随之而来的炎症、血脂异常和内皮功能障碍导致不良的肺血管重塑和右心室功能障碍。¹³在弗里曼特尔糖尿病研究中，对因临床原因进行超声心动图检查的 T2D 患者子集进行的分析显示，估计右心室收缩压 >30 和 >40 的患病率分别为 6.4% 和 2.6%，且随访时间超过 6.6 年。另有 9.2% 和 5.0% 的人患有这种程度的 PH，这表明糖尿病患者患 PH 的风险大约高出 40%。¹⁴一项针对患有糖尿病的肺动脉高压患者与未患有糖尿病的患者的研究表明，前者的 BNP 较高、6 分钟步行距离较短、呼吸困难较多、肺动脉压力较高且生存期较短，¹⁵美国国家退伍军人健康事务数据库对 110 563 名新诊断的肺动脉高压患者进行的一项研究表明，超过三分之一患有糖尿病，这与 1.21 倍的死亡风险相关。¹⁶对 14 861 人进行的遗传关联研究通过超声心动图测量了肺动脉压和右心室功能，结果表明 T2D 和肥胖均与较高水平的三尖瓣反流和右心室收缩压相关，表明两者在PH值。^{17 号}

许多糖尿病治疗方法可能对肺部有益。应该指出的是，流行病学研究结果需要持怀疑态度，最近在分析影响二甲双胍可能导致癌症减少和/或改善癌症治疗者预后的证据的潜在偏差时提出了这一警告。¹⁸牢记这一警告，我们可以回顾一些有趣的报告。在一项针对 3599 名患有 IPF 和 T2D 的成年人的研究中，控制了年龄、性别、种族/民族、居住地区、年份、药物、氧气使用、吸烟状况、医疗保健使用和合并症，接受二甲双胍治疗的患者死亡率降低了 54%住院风险降低 18%。¹⁹在一项对 350 536 名未患有慢性阻塞性肺病 (COPD) 的新发糖尿病患者的研究中，吡格列酮治疗与较低的患慢性阻塞性肺病 (COPD) 的可能性相关，尤其是那些接受吡格列酮治疗时间较长的患者。²⁰然而，噻唑烷二酮类药物对于患有心脏和/或肺部疾病的患者来说是一种复杂的药物，一项对 402 153 名患有 T2D 和 COPD 患者的住院研究显示，接受噻唑烷二酮类治疗的患者患 CVD 和 COPD 的可能性增加了 50% 以上。心力衰竭事件的发生率以及细菌性肺炎发生、住院期间需要无创正压通气以及肺癌发生的风险增加。²¹

在野百合碱治疗的大鼠 PH 模型中，钠-葡萄糖协同转运蛋白 2 抑制剂 (SGLT2i) 恩格列净可降低右心室和肺动脉压力，减少右心室肥厚和纤维化，并与改善生存相关。²²对 65 名患有心力衰竭并植入肺动脉压力传感器的患者进行的一项为期 12 周的试验显示，随机接受恩格列净治疗的患者肺动脉舒张压有所降低。²³在一项使用英国临床实践研究数据集对糖尿病合并慢性阻塞性肺病患者进行的流行病学研究中，与接受磺酰脲类治疗的患者相比，接受 SGLT2i 治疗的患者需要住院治疗的慢性阻塞性肺病急性加重的情况减少了 41%，尽管中度急性加重是在门诊治疗的基础并不常见。²⁴

使用胰高血糖素样肽 1 受体激动剂 (GLP-1RA) 似乎可以看到最大程度的改善。肺部的 GLP-1 受体表达高于大多数其他组织，在小鼠模型中给予 GLP-1RA 可以降低肺部对吸入过敏原的反应，并降低动物模型中的支气管高反应性和炎症标志物。²⁵在使用英国临床实践研究数据集的流行病学研究中，重度和中度 COPD 急性加重分别减少了 41% 和 48%。²⁴来自麻省总医院 Brigham 数据集的一项类似研究显示，GLP-1RA 治疗可最大程度地减少重度和中度 COPD 急性加重，而 SGLT2i 似乎也与获益相关。²⁶第三项研究分别对 4150 人和 12540 人使用 GLP-1RA 与二肽基肽酶 IV 抑制剂 (DPP-IVi) 进行了慢性下呼吸道疾病恶化的风险比较，发现 GLP-1RA 使住院和治疗的可能性降低了 48%。肺部症状总加重减少 30%。²⁷对涉及 77 485 名参与者的 28 项 GLP-1RA 随机对照试验进行的荟萃分析显示，总体呼吸系统疾病的可能性降低了 14%，肺水肿的可能性降低了 34%，支气管炎的可能性降低了 14%。²⁸

总之，有令人信服的证据表明糖尿病与肺功能下降有关。似乎涉及多种机制，包括胰岛素抵抗、内皮功能障碍、炎症和肥胖的影响。肺PH、肺纤维化、哮喘和慢性阻塞性肺病都与糖尿病有不良关联。多种糖尿病治疗方法似乎对治疗这些肺部疾病有益，动物研究和随机对照试验的实验证据表明 SGLT2i 有益，并且强有力的流行病学证据表明 GLP-1RA 类有益。