Introduction

Coronavirus disease 2019 (COVID-19) has been declared a pandemic and global emergency by the World Health Organization (WHO). It is an airborne disease caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), leading to a range of respiratory infections from mild symptoms to severe conditions such as pneumonia or acute respiratory distress syndrome (ARDS) [1].

Around 20% of severe cases are transferred to the intensive care unit (ICU) where invasive mechanical ventilation (IMV) is often required for approximately 10 days [2, 3]. While these vital methods are crucial for patient survival, they can cause lung injury, oxidative stress and a reduction of antioxidant enzymes, leading to accelerated telomere shortening and immune senescence [4,5,6]. This accelerated oxidative stress, in combination with normal aging and other health conditions, leads to telomere shortening. When telomeres reach a critical threshold in length, it can trigger senescence and apoptosis depending on the cell type. These factors also contribute to immune senescence, altering the function of the immune cells and increasing the risk of infection and age-related diseases [7].

Moreover, virus-induced senescence directly affects the length of lymphocyte telomeres, mostly impacting older adults who are at a higher risk of severe COVID-19-related complications [8, 9]. There is an association between shorter telomere length and the need for hospitalization, greater severity of COVID-19, and even mortality [10,11,12]. Schneider et al. recently linked telomere shortening to an increased mortality risk in respiratory diseases, with a Hazard Ratio of 1.4 [13]. A recent study also found that individuals who recovered from SARS-CoV-2 infection had shorter telomeres around one year after recovery, and this shortening was associated with higher severity of the disease [14]. However, while telomere shortening during ICU stay has been observed due to reasons other than COVID-19 [15], to date, there are no longitudinal studies evaluating telomere shortening or recovery in patients admitted to the ICU. Therefore, this study aimed to assess the change in telomere length in patients admitted to the ICU at least one year after hospital discharge.

Materials and methods

Design and study population

We performed a longitudinal study of COVID-19 patients admitted to the ICU at the Tajo University Hospital (Aranjuez, Madrid) and Infanta Cristina University Hospital (Parla, Madrid) from August 2020 to April 2021, who were followed for at least 12 months from hospital discharge. Patients were diagnosed by laboratory confirmation (RNA detection by PCR or serology-based methodology) or by consideration of clinical manifestations compatible with COVID-19. The strengthening the reporting of observational studies in epidemiology (STROBE-ID) checklist was followed (Supplemental File 1). The study protocol was approved by the Ethics Committee of the Institute of Health Carlos III (PI 33_2020-v3) and the Investigation Committee of both hospitals.

Clinical data and samples

Epidemiological and clinical variables were collected from medical records with research electronic data capture (REDCap) [16]. Peripheral blood samples were obtained in ethylenediaminetetraacetic acid (EDTA) tubes upon admission to the ICU and at least one year after discharge. DNA was isolated with the commercial kit “ReliaPrep™ Blood gDNA Miniprep System” from PROMEGA, according to the manufacturer’s instructions. DNA concentration and quality were determined using NanoDrop™ One/OneC (Thermo Scientific™). All samples were then diluted into 10 ng/uL working aliquots.

Telomere relative quantification

Relative telomere length (RTL) quantification was carried out by monochromatic multiplex real-time quantitative PCR (MMqPCR) assay as previously described [12]. Each MMqPCR plate included a standard curve generated by serial dilution (1:3) of a pool of DNA extracted from plasma samples of 20 healthy individuals. The standard curve was run in duplicate for each run, along with a duplicated negative control. Every experimental sample was tested in triplicate. Any samples with a coefficient of variation (CV) greater than 0.15 were reanalyzed, and samples that still had a CV greater than 0.15 were discarded (see Supplemental File 2 for extended information).

Study variables

The outcome variable was the change of RTL from baseline to one-year visit. This change was analyzed using the RTL ratio (one-year RTL / baseline RTL). The secondary outcome was the RTL at one-year visit.

The study factors included were: (a) Hospitalization variables: (i) ICU length of stay (ICU-LOS); (ii) need for IMV and its duration; (iii) prone position during the first 7 days of ICU stay. (b) Radiologic alterations in the one-year visit for the subgroup of patients who required IMV: presence of fibrosis on the X-ray image.

Statistical analysis

For the descriptive analysis, the McNemar test for categorical variables and the Wilcoxon test for continuous variables were used.

Regarding RTL measurement, outlier identification was performed using the interquartile range (IQR) method. Correlation between RTL and age was performed using Spearman correlation.

For telomere size differential analysis, we performed the Mann–Whitney test for transversal analysis and the Wilcoxon test for longitudinal analysis between paired samples. We used Generalized Linear Models (GLMs) with gamma distribution to study the association between the baseline RTL, RTL ratio, and one-year RTL values (y, dependent variables) with the need for IMV, IMV duration, ICU-LOS, prone position, and pulmonary fibrosis development at the end of the follow-up (x, independent variables). This test provides the arithmetic means (AMR) ratio and its significance level. GLMs were adjusted by the most relevant covariates by using a stepwise method (forward), where at each step, the covariates were considered to enter according to the lowest Akaike information criteria (AIC) for that specific model. The covariates included were age, gender, body mass index (BMI), sequential organ failure assessment (SOFA) score at admission, PCR plate, and follow-up time (months) when GLMs included variables from the one-year visit. For the association analysis with radiological alterations, we analyzed the differences in epidemiological and clinical variables between groups (no alterations and fibrosis), and significant variables were included as covariables in the model.

Statistical software R (v.4.2.2) was used to perform the statistical analysis. All p-values were two-tailed, and the statistical significance was p ≤ 0.05.

Results

Patient characteristics

Clinical and epidemiological characteristics of the 49 patients according to the baseline and one-year time (after a median of 13.8 months) are shown in Table 1.

Table 1 Clinical and epidemiological characteristics of COVID patients at admission ICU and one year after discharge
Full size table

Overall, the median age was 60 years; 71.4% were male, and more than 80% were Caucasian. In addition, 42.9% had hypertension, 55.1% obesity, and 37.4% diabetes. Regarding hospitalization variables, 73.5% required IMV, patients had a median ICU-LOS of 12 days and IMV duration of 14 days, and 38.8% needed a prone position (first 7 days). At baseline, no patient was vaccinated but received the first dose before the one-year visit.

In addition, we found a negative but not significant physiological correlation between RTL and age at baseline (r = − 0.242; p = 0.094) and significant at one-year visit (r = − 0.310; p = 0.030) (Supplemental File 3).

Evolution of RTL according to hospitalization variables

Firstly, we analyzed the epidemiological and clinical differences between patients who experienced an increase in RTL and those who had a decrease from baseline to one year. We found that patients who shortened their RTL showed significantly higher body temperature (p = 0.036), longer ICU LOS (p = 0.008), a greater need for IMV (p = 0.002), and all of them developed pulmonary fibrosis by the end of the follow-up period (p = 0.019). It's worth noting that there were no differences in terms of age or the presence of comorbidities (Supplemental File 4).

ICU-LOS While no differences were found at baseline, prolonged ICU-LOS was related to RTL differently. Thus, we observed that only those patients with ICU-LOS > 12 days showed significantly shortened telomeres size (p < 0.001), while patients with an ICU-LOS < 12 days showed similar RTL (p = 0.380) during follow-up (Fig. 1A).

Fig. 1
figure 1

A Evolution of relative telomere length (RTL) in COVID-19 patients, stratified by length of stay (LOS) in the intensive care unit (ICU), need for invasive mechanical ventilation (IMV), and need for prone position, from baseline to one-year visit (12 months since hospital discharged). B Differences in RTL, one-year visit, in COVID-19 patients who required IMV and had X-ray images, comparing patients without radiological alterations and with fibrosis. Statistics: Data represent the crude means and 95% confidence interval for each group of patients. P-values were calculated using the Mann–Whitney test for transversal analysis and the Wilcoxon test for longitudinal analysis between paired samples. R and p-value in the scatter plot were calculated by a Spearman’s correlation. Statistical significance was determined as p ≤ 0.05. Abbreviations: RTL, Relative telomere length; ICU, Intensive care unit; IMV, Invasive mechanical ventilation

Full size image

Need for IMV and its duration No differences in RTL were found at baseline, but those patients who required IMV significantly reduced their RTL during follow-up (p < 0.001). So, those patients admitted to the ICU but who did not receive IMV significantly increased their RTL during follow-up (p = 0.012) (Fig. 1A) (Supplemental File 5). Besides, one year after discharge, patients without IMV showed higher RTL than those with IMV, although this difference did not reach significance p = 0.068).

Need for a prone position The requirement for a prone position suggests a more severe and worse evolution of ARDS. Therefore, we analyzed the RTL between patients with and without needing a prone position during the first 7 days. While no differences were observed at baseline, those patients who required a prone position significantly reduced their RTL during follow-up (p = 0.005). On the other hand, patients who were not placed in the prone position did not modify their RTL during follow-up (Fig. 1A).

Radiologic abnormalities at one-year visit

For the subgroup of patients who required IMV and had an X-ray image at the one-year visit (n = 33), 63.3% had no radiological alterations, 9.1% had interstitial infiltrates, and 27.3% had fibrosis. Regarding the clinical and epidemiological characteristics of extreme phenotypes (fibrosis and no radiological alterations), we only observed significant differences in the presence of arterial hypertension at baseline (p = 0.035) and IMV days during ICU stay (p = 0.043), so both variables were included in the model (Supplemental File 6).

Regarding the association analysis, patients with fibrosis showed a higher RTL shortening at the one-year visit (p = 0.007) than those without radiological abnormalities. Additionally, these differences remained significant at the one-year visit, where patients who developed pulmonary fibrosis showed significantly lower RTL (p = 0.040) (Fig. 1B).

All the significant results for each comparison were confirmed by multivariate GLMs adjusted by the covariates selected by a stepwise method (Supplemental File 5).

Discussion

In this longitudinal study, we found that patients with longer ICU-LOS and those who required IMV had a greater shortening of the RTL during the first year after hospital discharge. In addition, patients who developed pulmonary fibrosis showed a deep reduction of RTL during the first year after hospital discharge, showing lower RTL at the one-year visit than those who did not have pulmonary fibrosis. To our knowledge, our study is the first one showing an association between RTL shortening and clinical variables of hospitalization after more than one year of follow-up in COVID-19 patients admitted to the ICU.

Telomere length shortens with each cell division and, when it reaches a critical length (dysfunctional telomeres), senescence and apoptosis are elicited [17]. Telomere shortening has been associated with numerous diseases, including age-related diseases, stress-related conditions, cancer, and bacterial and viral infections [18, 19]. In the context of these diseases, inflammation or oxidative stress induces accelerated telomere shortening [20]. However, this molecular process can be reversed in some cell types due to the presence of telomerase enzyme, which is responsible for maintaining telomere length [21]. In our study, we could not measure telomerase, so we could not evaluate the cells’ capacity to reverse telomere shortening.

In the context of COVID-19, complications of the disease are more likely to occur in older individuals, suggesting that age-related molecular pathways contribute to its severity [22]. In this regard, several studies have shown a link between shorter telomeres and an increased need for hospitalization, greater disease severity, and even mortality in COVID-19 patients [10,11,12]. However, few studies have investigated the changes in telomere length after patients are discharged from the hospital. Retuerto et al. found that, after a median of 14 months of follow-up, 35% of study participants experienced telomere shortening. Although they did not find an association between RTL and clinical outcomes in the hospital, this might be due to the heterogeneous nature of their study population, which included varying levels of disease severity. However, the patients who had a more severe disease also exhibed the greatest telomere shortening [14]. The impact and, consequently, the implications of telomere shortening in COVID-19 patients admitted to the ICU remain unknown.

In this study, patients with longer ICU-LOS experienced a shortening of RTL during follow-up. This could be attributed to the critical exhaustion of the immune system after an extended ICU stay, where telomeres would reach a critical size, compromising the capacity of telomerase to restore telomere length. Previous studies have demonstrated significant changes in leukocyte telomere size in critically ill patients during the ICU stay [15, 23]. However, to our knowledge, no studies have investigated the impact of ICU-LOS on RTL after hospital discharge. Our findings highlight the importance of follow-up for patients with prolonged ICU-LOS to monitor for future senescence-related complications.

Patients requiring IMV had a longer ICU-LOS and a worse prognosis [24]. Overall, we found that, in critically ill COVID-19 patients, the need for IMV was associated with telomere shortening during follow-up, possibly reflecting lung damage and loss of functionality. To our knowledge, few studies have addressed this issue. Only Liang et al. observed telomere shortening in ICU patients requiring IMV, but their evaluation was conducted only two days after ICU discharge [23]. Therefore, additional studies with longer follow-up periods are needed to elucidate the impact of IMV on RTL reduction and its association with post-COVID syndrome in patients admitted to the ICU.

In severe patients with worse evolution of ARDS, a prone position is recommended to improve oxygen saturation and reduce lung damage [25, 26]. A prone position significantly reduces the rate of tracheal intubation and mortality [27], although there are also studies with contradictory results [28]. Moreover, among intubated patients, those who respond to the prone position in terms of oxygenation show better survival than non-responders [29]. Regarding evolution after discharge, our study revealed that COVID-19 patients who required a prone position experienced telomere shortening during follow-up. To our knowledge, these findings are the first to associate the prone position with reduced telomere size in COVID-19 patients during the first year after hospital discharge. Further studies with a longer follow-up are needed to determine whether telomere size evolves to more critical levels in this subgroup of patients.

COVID-19 patients often develop fibrotic pulmonary sequelae [30,31,32]. There is also an association between telomere shortening in immune cells and the development of lung diseases, such as idiopathic pulmonary fibrosis [7, 33,34,35]. In our study, while no significant differences in RTL were observed at baseline, patients who experienced RTL shortening during follow-up developed lung fibrosis. Similarly, patients with fibrotic pulmonary sequelae had shorter RTL at the one-year visit than those without radiological abnormalities. To our knowledge, this is the first study assessing the RTL and fibrosis development during follow-up in ICU patients. Still, these findings are in concordance with previous studies in hospitalized patients. McGroder et al. showed that patients with shorter telomere size at hospitalization were more prone to develop fibrotic abnormalities in the lung four months later, but the telomere length dynamics were not addressed [36]. Using a different approach (Quantitative telomere Fluorescence in situ Hybridization Q-FISH), Martínez et al. found loss of alveolar type II (ATII) cellularity and shorter telomeres in these cells in post-COVID-19 lung cancer patients. Authors suggested that ATII cell telomere reduction may trigger lung fibrosis development in post-COVID-19 patients [37], although these results were obtained in a cancer background, and studies on non-lung cancer COVID-19 patients are needed. Similarly, Mulet et al., in a recent prospective study [38], described that this attrition could also be evaluated on peripheral blood cells, as COVID-19 patients with radiological signs of fibrosis had significant telomere shortening one year after leaving the hospital. However, this study involved only 19 patients, and the authors did not specify how they measured telomere length. Retuerto et al. also observed that patients who experienced peripheral blood leukocyte telomere shortening after discharge were more prone to show persistent radiographic abnormalities. However, this study was conducted during the first pandemic wave, when the hospital admission criteria were exceptional.

Our study is the first to confirm these findings in ICU patients, with a homogeneous group characterized by similar severity and clinical strategies. However, further studies with longer follow-up periods are needed to establish the role of RTL shortening on fibrotic pulmonary sequelae and vice versa.

Limitations

Several issues must be considered to interpret our data correctly. The sample size was limited as only patients with the one-year visit were included in the study. This may have limited the statistical power of the analysis, and some significant associations may not have been detected. However, despite its preliminary nature, the statistical analyses correctly addressed this issue. It is important to note that other potential factors potentially impacting telomere shortening, like telomerase activity, could not be measured. Additionally, it is unclear whether the viral genotype or the viral inoculum, which may impact the disease severity, could have also played a role in influencing telomere attrition. This study also has several strengths, such as the fact that all patients were recruited during the second and third pandemic waves, where the ICU admission criteria were similar, and the selected method of measuring telomere length. We measured RTL in whole blood, and its determination could be transferred to routine clinical practice. This study addressed a topic that has been scarcely studied until now in critical patients because the recruitment and follow-up of these patients are arduous. In addition, it should be noted that we analyzed RTL before initiating any mechanical ventilation.

Conclusions

In conclusion, we found that patients with longer ICU stays, need for IMV, and fibrosis development shortened telomere size in the first year after hospital discharge. In addition, patients who required IMV and developed pulmonary fibrosis showed shorter telomeres at the one-year visit. The findings of this preliminary study suggest that these patients need extensive monitoring as they are at higher risk of developing cellular senescence-related and chronic respiratory complications.