Fasting has real physiological consequences, and the traditions are less naive about this than outsiders often assume.
Every large tradition excuses the vulnerable, and not grudgingly. Islamic law treats fasting while ill or pregnant as wrong rather than merely permitted. Jewish law requires eating on the Day of Atonement where life is at risk, and a rabbi will rule on it if asked. The clinical literature on Ramadan is substantial and largely reassuring for healthy adults, while identifying genuine risks for people with diabetes, for those on medication schedules that assume regular food, and for outdoor workers in heat. Medical bodies in Muslim-majority countries publish guidance jointly with religious authorities, which is a useful instance of the two operating together rather than against one another.
The genuinely difficult cases sit at the end of life. The Jain vow of sallekhana ends in death by design, and is defended by its practitioners as the opposite of suicide precisely because it excludes despair and appetite; Indian courts have suspended a ruling against it and never resolved the question. Christian ascetic tradition contains extreme fasting, and historians of late medieval women's piety have argued at length over how far the recorded cases overlap with what would now be diagnosed as an eating disorder. That argument cannot be settled from the sources, and it should not be closed in either direction by assertion, neither by pathologising the women involved nor by refusing to consider it.
There is also a secular inheritance. Intermittent fasting and extended water fasts are now marketed as health protocols with the religious frame removed. They are not the same practice. The traditional forms are communal, calendrical and bounded by rules somebody else wrote, and that changes both their psychology and their risks.