Diabetes and the Risk of Long-term Post-COVID Symptoms

The world is suffering a dramatic situation of catastrophic proportions due to the worldwide spread of the coronavirus disease 2019 (COVID-19) caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) (1). SARS-CoV-2 seems to disproportionately impact people with preexisting medical comorbidities (e.g., diabetes, hypertension, or cardiovascular conditions) (2). In fact, diabetes is associated with higher mortality, more severity, and worse progression of COVID-19 (3). This association is based on the hypothesis that the ACE2 receptor (4,5) and the proinflammatory response (i.e., cytokine storm) usually seen with SARS-CoV-2 infection are overexpressed in individuals with diabetes (6).

Health care professionals are also in front of a second pandemic related to SARS-CoV-2, the “long-haulers,” that is, people experiencing symptoms after the acute phase (post-COVID symptoms) far longer than would be expected (7). A recent preprint meta-analysis found that almost 80% of COVID-19 survivors exhibit post-COVID symptoms; however, most studies have not considered the role of medical comorbidities (8). One study suggested that patients with diabetes suffered from acute post-COVID symptoms (e.g., 4 weeks after the infection) such as pain or sleep disturbances (9). No previous study has investigated whether patients with diabetes are at risk for developing long-term post-COVID symptoms. This study investigated the association of diabetes with long-term post-COVID symptoms in hospitalized COVID-19 survivors.

The current multicenter case-control study included patients hospitalized for SARS-CoV-2 infection (ICD-10 code) during the first wave of the pandemic (from 1 March to 31 May 2020) of three public hospitals of Madrid (Spain). From all individuals with a positive diagnosis of SARS-CoV-2 by real-time RT-PCR and consistent radiological findings who were hospitalized during the first wave of the pandemic in these three hospitals (total, 8,500), a sample of 600 patients from each hospital was randomly selected for this study. From the selected patients, those with a medical history of diabetes prior to hospitalization were included as case subjects. The diagnosis of diabetes (as reflected on medical records) was based on fasting plasma glucose or A1C criteria according to the American Diabetes Association (10). Additionally, two matched subjects without preexisting diabetes per each case subject were recruited as control subjects. Each control subject was matched by age and sex. If more than two control subjects per case subject were available, the selection was done randomly. The study design was approved by the local Ethics Committee of all hospitals (HUIL/092-20, HUF/EC1517, HSO25112020). All participants provided informed consent before data were collected.

Clinical and hospitalization data were collected from hospital medical records. Participants were scheduled for a telephone interview by trained health care professionals. Patients were asked to report the presence of symptoms after hospitalization and whether symptoms persisted at the time of the study. All participants were systematically asked about the following list of post-COVID symptoms in a standardized fashion: dyspnea, fatigue, chest pain, headache, anosmia, ageusia, cough, palpitations, diarrhea, cognitive blunting/brain fog, pain, or memory loss, but they were free to report any further symptom that they considered relevant.

The Hospital Anxiety and Depression Scale (HADS) and the Pittsburgh Sleep Quality Index (PSQI) were used to assess anxiety/depression symptoms and sleep quality, respectively, because both can be adequately administered by telephone interview (11). Briefly, the HADS includes an anxiety symptoms subscale (HADS-A, seven items, 21 points) and a depressive symptoms subscale (HADS-D, seven items, 21 points) (12). We considered the cutoff scores recommended for the Spanish population (HADS-A ≥12 points; HADS-D ≥10 points) for determining the presence of anxiety and depressive symptoms, respectively (13). The PSQI evaluates the quality of sleep over the previous month throughout 19 self-rated questions assessing different sleep aspects (14). Questions are answered on a 4-point Likert-type scale (0–3), and the sum is transformed into a global score (0–21 points), where higher scores are indicative of worse sleep quality. A total score ≥8.0 points suggests poor sleep quality (14).

We also took the following items from the Functional Impairment Checklist (FIC), a disease-specific tool used for evaluating the functional consequences of SARS (15): dyspnea at rest, dyspnea on exertion, generalized fatigue, limitations occupational activities, limitations social/leisure activities, limitations in basic activities of daily living, and limitations in instrumental activities of daily living.

The statistical analysis was conducted with Stata 16.1 statistical software (StataCorp LLC, College Station, TX). The McNemar and paired Student t tests were applied to compare proportions and means between patients with and without diabetes. Multivariable conditional logistic regression models were constructed to identify the variables associated with the presence of diabetes. Adjusted odd ratios (OR) or incident rate ratios (IRR) with their 95% CI were calculated.

All data generated or analyzed during this study are included in the published text. The data sets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.

From 1,800 hospitalized patients during the first wave of the pandemic and randomly selected, 145 presented the diagnosis of diabetes. In addition, 290 age- and sex-matched individuals without diabetes were recruited. All patients with diabetes were under controlled insulin intake by their medical doctor before being infected. In general, no differences in symptoms at hospital admission were seen, except for myalgia, which was more prevalent in individuals with diabetes (Table 1). A significantly greater proportion of patients with diabetes reported comorbid hypertension, obesity, and migraine compared with patients without diabetes (χ² = 51.373, P < 0.001).

Participants were assessed a mean of 7.6 (SD 0.6) months after hospital discharge. From the total sample, just 85 (19%) were completely free of any post-COVID symptom at the follow-up period. The number of post-COVID symptoms was similar (IRR 1.06, 95% CI 0.92–1.24, P = 0.372) between groups with diabetes (mean 2.0, SD 1.3) and without diabetes (mean 1.9, SD 1.5). The most prevalent post-COVID symptoms were fatigue, dyspnea on exertion, and musculoskeletal pain (Table 1). No differences were observed in the presence of fatigue (OR 1.45, 95% CI 0.93–2.25, P = 0.101), dyspnea on exertion (OR 0.97, 95% CI 0.64–1.47, P = 0.886), and musculoskeletal pain (OR 0.951, 95% CI 0.76–1.18, P = 0.367) between patients with or without diabetes (Table 1). Further, no between-group differences in anxiety symptoms (OR 1.30, 95% CI 0.77–2.20, P = 0.320), depressive symptoms (OR 1.31, 95% CI 0.79–2.17, P = 0.294), or poor sleep quality (OR 1.34, 95% CI 0.89–2.03, P = 0.156) were found (Table 2).

Table 2 summarizes functional limitations during daily life activities in both groups. From the total sample of 435, 196 (45%) experienced at least one functional limitation with daily living activities, without between-group differences (OR 1.07, 95% CI 0.71–1.62, P = 0.728). A total of 64 patients (14.7%) experienced limitations with occupational activities, 159 (36.5%) limitations with social/leisure activities, 100 (23%) limitations with instrumental activities of daily living, and 132 (19.4%) limitations with basic activities of daily living. No between-group differences existed in limitations with occupational activities (OR 0.73, 95% CI 0.40–1.35, P = 0.319), limitations with leisure/social activities (OR 1.34, 95% CI 0.87–2.06, P = 0.189), limitations with instrumental activities of daily living (OR 1.05, 95% CI 0.67–1.65, P = 0.818), and limitations with basic activities of daily living (OR 1.04, 95% CI 0.63–1.71, P = 0.866) (Table 2).

Identification of patients at a higher risk of developing post-COVID symptoms is crucial. In this case-control study, we investigated the association of diabetes with long-term post-COVID symptoms in a cohort of hospitalized COVID-19 patients. Our study showed that diabetes was not a risk factor for post-COVID symptoms when assessed an average of 7 months after discharge. Further, diabetes was not associated with differences in limitations with daily living activities after hospital discharge.

The prevalence of diabetes in patients with COVID-19 is 8% (95% CI 6.0–11) (16) in agreement with our data (145 of 1,800 [8.05%]). Nevertheless, we should recognize that our sample was older (70 years) than previous data (mean age 50 years) (17). The fact that our patients were older could explain the higher frequency of comorbidities (e.g., obesity or hypertension) that also influence the course of COVID-19 (2). Independently of the presence of other comorbidities, individuals with diabetes did not exhibit different long-term post-COVID symptoms compared with those without diabetes.

We observed that fatigue, dyspnea on exertion, and musculoskeletal pain were the most prevalent long-term post-COVID symptoms in patients with diabetes, with a prevalence of 66.2%, 53.8%, and 44.8%, respectively. Our findings were similar to those previously reported by Akter et al. (9) for pain (43.8%) 4 weeks after. Other symptoms observed by Akter et al. (9) were concentration loss (28.8%), anxiety/depression (25.4%), sleep disorders (34.4%), memory loss (24.7%), and hair fall (13%). Again, we reported similar prevalence rates of memory loss (23.4%), anxiety/depressive levels (21%), or sleep problems (37.9%) but lower rate of concentration loss (8.3%). It is important to consider that Akter et al. (9) analyzed post-COVID symptoms in an acute post-COVID phase, whereas our study is the first one analyzing symptoms at a long-term post-COVID phase (18).

Our study has some limitations. First, patients were followed up by telephone and not face-to-face.

Second, only hospitalized patients were included. Surprisingly, the number of patients needing intensive care unit admission in our study (6.2%) was smaller than that previously reported (20–30%) in a recent multinational study (19). This is probably because patients included in the current study were required to survive to discharge to be included.

Third, we did not collect measures of COVID-19 disease (e.g., inflammatory biomarkers or serum levels of proteins).

Fourth, we collected data cross-sectionally; therefore, the exploratory nature of the study needs to be confirmed in longitudinal studies.

Finally, we did not control medication intake for comorbidities (e.g., hypertension or obesity), which could have an interaction. However, Almeida-Pititto et al. (2) found that none of the 40 studies included in their meta-analysis reported data on diabetes or hypertensive medications.

In conclusion, current and previous evidence suggest that diabetes seems to play a more relevant role during the acute phase of COVID-19 rather than for the development of post-COVID symptoms in previously hospitalized COVID-19 survivors. Future longitudinal studies are needed to further confirm these assumptions.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. All authors contributed to the study concept and design. C.F.-d.-l.-P. and V.H.-B. conducted the literature review and did the statistical analysis. C.G., J.T.-M., M.V.-A., and S.P.-C. recruited participants. J.A.A.-N. supervised the study. All authors collected data and contributed to interpretation of data. All authors contributed to drafting the paper, revised the text for intellectual content, and read and approved the final version of the manuscript. C.F.-d.-l.-P. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.