With this study on the acute course of COVID-19 infections in PWS patients and the long-term impact on polysomnographic evaluation, we found subtle but significant and lasting changes detected by PSG. It was hypothesized that patients with PWS would be an at-risk group for a severe course of a COVID-19 infection because of the syndrome associated with obesity and respiratory dysfunction. However, in accordance with others [15, 16], we found only mild to moderate impairment directly associated with the infection. However, through analysis of PSG parameters and oxygen saturation, long-term lasting changes to the respiratory system were detected.

Similar to the two mentioned studies [15, 16], respiratory problems and breathing difficulties were only reported twice, tachypnoea occurred four times in our cohort. Fever was the most frequently observed symptom reported by caregivers in our study cohort. Despite the potential for lack of appropriate febrile response, which is often seen in PWS patients due to disturbed thermoregulation [22], we suggest that fever or raised body temperature is a reliable symptom of COVID-19 infection in PWS patients, as it is in healthy individuals [10, 23]. In general, in our study the reported clinical symptoms during the infection observed by caregivers did not differ from those already reported in studies on healthy children [10, 23] and is in accordance with other studies including children and adults with PWS [15, 16]. Thus, we assume that PWS patients have similar symptom profiles in comparison to the general population during the acute infection.

Although the severity of infection was only mild to moderate, and the caregivers reported a total remission of the symptoms in almost all participants, there were statistically significant differences in PSG at follow-up examination after an average of three months post-COVID-19 infection. Time of sleep spent in the target area of 95 to 100% oxygen saturation was significantly less in comparison to the examination before COVID-19 infection. The changes in the lowermost SpO2 and mean SpO2 were small, so the clinical relevance of these two changes remains unclear. However, this was clearly a directional trend since the majority of participants showed worsened lowest and mean SpO2 after COVID-19 infection. Moreover, there was a statistically significant shift to more frequent hypopnoea in the post-COVID-19 follow-up examination. This is in accordance with what has already been reported in non-PWS COVID-19 patients. Low oxygen saturation without dyspnoea during the infection and during the acute rehabilitation phase has been described as “happy or silent hypoxia” [24,25,26,27,28,29,30]. The reason for the discrepancy between decreased perception of oxygen deficiency and the absence of the feeling of dyspnoea is not fully understood [30]. However, reduced SpO2 during the COVID-19 infection seems to affect PWS patients, as it has been described in a previous study on children and adults with PWS [16]. Whittington et al. reported that “low measured oxygen saturation” was mentioned as a symptom, although it was not explicitly asked for in their questionnaire. They reported that it occurred in “less than 10%” of the patients [16, 31]. It remains unclear, whether reduced SpO2 would have been mentioned more frequently, if it had been studied in routine SpO2 measurements in the acute phase, or if low measured SpO2 would have been added in the dropdown menu of the survey. SpO2 was not measured during the acute infection in our patients, because no patients were hospitalisedduring their infection.

PWS patients might be an at-risk group for “happy hypoxia” in COVID-19 infection. In earlier studies, Arens R and Gozal D et al. already showed a non-existent or significantly attenuated hypoxic ventilatory response in patients with PWS that was independent of the degree of obesity [32]. In a further study, they also showed a blunted hypoxic arousal response during sleep in PWS patients compared to a non-PWS control group [33]. They concluded that the primary abnormality of ventilatory control in PWS patients involves the peripheral chemoreceptor pathways [31,32,33]. Thus, an abnormal ventilation control with insensitivity to O2 and CO2 levels might contribute to low measured SpO2 with absence of dyspnoea in PWS patients with COVID-19 infection.

However, because there is still a reduction of SpO2 during sleep after an average of three months’ time after the acute COVID-19 infection, we assume that reduced SpO2 during sleep might be a persisting problem since the acute infection. Further studies are required to prove this hypothesis.

Raised BMI and obesity are strong confounders for sleep related breathing disorders (SRBDs) in simply obese children and children with obesity and PWS [34]. Therefore, we wanted to rule out that an increase in BMI might be the reason for the deterioration of SpO2 and more frequent hypopnea during sleep after an average of three months after COVID-19 infection. Mean BMI-SDSPWS remained in the normal range, and the majority of our PWS patients were not obese. Additionally, BMI-SDS for healthy children remained in the normal range, despite a small but significant increase after COVID-19. However, multiple regression analysis of BMI SDShealthy did not show statistically significant influence on the results of the respiratory SpO2 parameters. Thus, a change in BMI SDShealthy could be due to PWS specific phenotypic and metabolic characteristics, as well as changes in nutritional and eating behaviour with increasing age [3, 22]. As these are not typical in children and adolescents without PWS, PWS syndrome specific BMI SDS are calculated [18].

Another possible explanation for our results might be alterations in pulmonary tissue or muscles of respiration. A study on adult non-PWS patients with OSA reported a significant increase in median CPAP pressure in their auto-CPAP treatment after COVID-19 infection [35]. Apnoea-hypopnoea index was slightly increased, but not significantly [35]. So, assuming there was no increased obstruction rate due to possible upper airway changes, e.g. persisting tonsil hyperplasia, this might imply some form of alteration in respiratory mechanics. We found no increase in obstructive sleep apnoea in our PWS patients after COVID-19 infection, but there was an increase in the number of hypopnoeas. More frequent shallower breathing due to alterations in respiratory mechanics might lead to a reduction of SpO2 measurement. This could be caused by changes in pulmonary tissue or muscles of respiration, or both.

To date, there have only been few studies in other paediatric patient populations comparing respiratory parameters in the same patients before and after COVID-19 infection and then comparing them with a non-COVID-19 control group. Mogensen et al. compared spirometric lung function tests in a population-based sample of young, healthy adults with and without asthma and a mild-to-moderate COVID-19 disease before COVID-19 infection with spirometric lung function tests after COVID-19 infection [36]. They found no evidence that an infection with COVID-19 resulted in impaired spirometric lung function [36]. In contrast, in a study by Soyak Aytekin et al. mean expiratory flow 25–75% values were significantly reduced after COVID-19 infection in some paediatric patients with asthma compared to prior lung function tests [37]. Thus, the low SpO2 values during sleep in our cohort might also be caused by a small airway dysfunction after COVID-19 [37].

Furthermore, a disturbed gas exchange after COVID-19 has already been discussed as a possible consequence of COVID-19 infection. Wu et al. reported in their study on 3-, 6-, 9-, and 12 months respiratory outcomes in adults without PWS who were hospitalized with severe COVID-19 infection, that diffusion capacity of the lungs for carbon monoxide was 77% of predicted at three months, 76% of predicted at six months, and increased to 88% of predicted at twelve months after hospital discharge [25]. The forced vital capacity also increased from 92% of predicted capacity after three months to 98% of predicted after twelve months [25]. As the diffusion capacity of the lungs for carbon monoxide improved with time [25], the question arises, whether the reduction in SpO2 during sleep in our study might be a consequence of a disturbed gas exchange and therefore, an improvement in SpO2 could also be expected with time. So far, our data does not indicate an improvement with time, as time since infection did not show an influence on PSG parameters in multiple regression analysis. As our study results are based on an average time span of three months after COVID-19 infection, further long-term studies are needed to better assess the time frame in which lowered SpO2 at sleep persists and to evaluate the potential resulting clinical implications of persisting increased sleep time in the lowered SpO2 range of 90–95% SpO2 after COVID-19 infection. For example, in a study by Lau et al.., an impact on verbal memory performance was associated with SpO2 nadir during sleep in children with OSA [38]. In their study, the group without OSA and better verbal memory performance showed a mean SpO2 nadir of 94.82% (range: 89–98) vs. 91.52 (range 76–97) in the OSA group with lower performance. The mean nadir in both groups was above 90%, once 92% and once 95%. So, it could be assumed that not only mean SpO2 < 90% might have clinical consequences, but also differences in the range of 90–100% with relatively more sleep time spent in lower intervals of 90–100%. Another possible clinical implication of reduced SpO2 during sleep is potential neurocognitive deficits, which are associated with sleep disordered breathing in children with sleep related breathing disorders [39]. Further studies are clearly needed to gain more data on clinical implications. Regular follow-up examinations by means of neurocognitive testing in combination with more frequent PSG are important to identify clinical implications after COVID-19 infection and should be considered for the regular care after COVID-19 infection.

Since cognitive disability and speech and language delay are known clinical manifestations in PWS [22], particular attention should be paid to the development of these parameters after COVID-19 infection, and possible treatment options should be discussed in time. CPAP, which is already used for PWS patients in the setting of OSA [40], could be considered as a possible treatment option in the case of persisting SpO2 below the optimal range during sleep after COVID-19 infection in PWS patients. According to a study by Turner et al., CPAP treatment had significant positive effects on working memory, long-term verbal memory, and short-term visuospatial memory in patients with OSA [41].

One limitation to our study is the possibility that some members of our control group did in fact contract COVID-19 during this time but remained asymptomatic and as such were not tested. Because of the retrospective study design, some symptoms the PWS patients experienced during their infection might not have been mentioned by the caregivers because of recall bias or because they have not been observed, due to a high pain threshold, which is a typical feature in people with PWS [1, 22].

The young age of our study cohort may have also contributed to a milder course of the disease [42,43,44].

Because of the low proportion of PWS patients with three or more vaccinations before their COVID-19 infection, no statistical analyses investigating the influence of vaccination status on course of COVID-19 infection and PSG parameters were performed.

Since we do not have a non-PWS control group, we cannot comment on whether the effect of COVID-19 infection on PSG parameters noted in this study apply to the general population. Whether a transfer of our results is possible should be investigated in future research. Moreover, we found only mild differences and the clinical implications remain unclear. However, all of our results are unidirectional and still apparent after approximately three months after COVID-19 infection, even if the majority of parents reported no persistent clinical overt symptoms. Further studies are needed to determine clinical significance and duration of the deterioration of PSG in patients with PWS.

The comparison between COVID-19 and the control group at post-examination showed a tendency towards more OSA in the COVID-19 group compared to the control group, but when analysing the COVID-19 subgroup, there were no statistically significant changes in HF and the number of OSA and hypopnoeas between pre- and post- COVID-19 examination. However, statistically significant differences were found when the genetic groups in the COVID-19 group were analysed separately. We observed that the deletion genetic subgroup had a statistically significant effect on an increased number of OSA, whereas the UPD genetic subgroup was associated with a reduced risk of higher HF and fewer hypopnoeas. There only have been few studies on the influence of genetics on cardiovascular and pulmonary risk in PWS patients so far. In an earlier study by Torrado et al.., more frequently present sleep disturbances were described in PWS patients with a deletion [45].

Moreover, a recent study by Cintra, Rocha et al. showed more frequent OSA in the deletion genetic group compared to the UPD group and 0% of the patients with an UDP had hypoventilation [46]. This is in accordance with the findings in our study and suggests that the genetic deletion may have a higher risk over time for sleep-disordered breathing and especially for OSA, whereas UPD might be associated with better progression over time. Future studies, with focus on the effect of genetics on cardiopulmonary parameters in a larger study cohort are of importance to prove this hypothesis.

The strengths of our study include the large patient population for PWS with COVID-19 infection, the large patient population without COVID-19 infection, and the assessment of longitudinal data. This allows us not only to compare changes pre- and post- COVID-19 for each participant individually, but also to compare the collectives, with and without COVID-19 infection. By regular follow-up examinations due to PWS, we were also able to exclude participants from whom only one valid examination was available or who were lost to follow-up. To the best of our knowledge, there are no prior studies in other paediatric study populations that directly compare longitudinal individual PSG evaluations before and after a COVID-19 infection.

Although on 4th May 2023 the World Health Organization declared COVID-19 to be an ongoing and established health issue which no longer constitutes a public health emergency of international concern [47], our study shows that we should continue to conduct studies on COVID-19 infection in PWS, in order to learn more about possible long-term consequences and possible treatment options to ensure optimal patient care.