Probably sometime between 1942 and 1943.
While a glowing report by several physicians prior to the campaign in 1932 that claimed that he was in terrific shape and that he didn't have major long lasting effects from polio was one that he bought, FDR was actually in fairly good health for much of the 1930s. When at the White House, he was swimming 3 or 4 times a week right up through 1941, and by and large his only real medical complaint until that point was that he had significant sinus problems. This was a major reason why Admiral McIntire became his physician; he was an ear, nose and throat specialist who like many prominent American naval officers of World War II had worked with FDR during his stint as Assistant Secretary 20 years earlier. (This was also why there was no air conditioning used in the White House during the administration; it severely aggravated FDR's condition.)
We only have intermittent records of FDR's blood pressure during the 1930s; his official governmental medical records were 'lost' after his death (almost certainly at the hand of McIntire), so what we have comes from other sources and it's hard to tell how often it was taken. That limited series tells a story of someone who was moderately hypertensive, but generally fell within what was considered acceptable ranges contemporaneously right up until the middle of 1941:
| Date | Blood Pressure |
|---|---|
| 1931 | 140/100 |
| 1935 | 136/78 |
| 1937 | 162/98 |
| 1940 | 178/88 |
| Feb 1941 | 178/102 |
| May 1941 | 150-160/80-85 |
This all generally falls into what we'd consider nowadays as the slow silent killer range of blood pressure, but the fact that his diastolic was still prehypertensive during multiple readings in May 1941 is a reasonably good indicator that he probably wasn't in awful shape at that point. Someone with his readings in modern times would probably be read the riot act by their physician all the way back in 1931 - pressured to give up smoking, watch sodium intake, increase exercise a bit, and be put on anywhere between one to three blood pressure medications until it fell into the normal range - and be monitored relatively closely afterwards. That said, absent any alarming labs or other test results, once treated by modern medicine 1941 era FDR would have been a patient kept under watchful vigilance by a primary physician rather than a cardiology referral.
Unfortunately, almost none of those lifestyle changes were known back then besides a marginal concern about exercise, nor would most medical professionals have even been all that concerned with the May 1941 numbers - blood pressure was expected to rise with age. Then comes the war. Roosevelt is often working 16 hour days especially before Jimmy Byrnes takes over the domestic front in 1943, ups his smoking, and stops swimming. We have no blood pressure readings available from between 1941 and when Howard Bruenn finally got brought in as his cardiologist in March 1944; whether or not McIntire just didn't take them or they got 'lost' in his medical file to hide damning details is unknown. Notwithstanding the very obvious physical signs of congestive heart failure that brought Bruenn into the picture, just on their own the 1944 numbers would have been frightening even to that era of physicians:
| Date | Blood Pressure |
|---|---|
| March 28 | 180/90 |
| March 31 | 220-230/120-130 |
| April | 190-220/98-116 |
| May | 205-240/105-130 |
| June | 186-220/88-120 |
His blood pressure isn't taken quite as often (or not recorded) as often after that, but the June ranges are fairly typical for the rest of his life, right up until the final day when he wakes up at 300/190, a number that isn't conducive to life. Incidentally, Bruenn tried to get him to start swimming again starting in the spring of 1944, but the one attempt shot his blood pressure up so much that he thought a heart attack was likely then and there so it was immediately discontinued.
The general gist of this is that that kind of hypertensive emergency doesn't appear overnight, so it's reasonable to guess that he went from hypertensive into congestive heart failure sometime in the preceding two years. Precisely when is unknown, but that lack of monitoring is the basis for a lot of well-deserved criticism of McIntire, and if another physician had been managing him he probably would have been alarmed sometime during those two years.
That said, while McIntire almost certainly committed malpractice even under the standards of the time, I do think there are a few things to be said in his defense. First, what generally doesn't get written about is that he wore two hats during the war, not just as FDR's physician but also at FDR's direction as Surgeon General of the Navy, which during the war years meant he was responsible for overseeing all hospitals and medical staff of the Navy, Marine Corps, and Coast Guard. While McIntire should have passed FDR's care on to another physician long before 1944 given his immense responsibilities for staff and policy, FDR also didn't want another one until his daughter Anna became genuinely alarmed at his condition and essentially forced McIntire to consult with Bruenn. Second, FDR was an immensely difficult patient to manage; he simply didn't want to ever get confirmation on how sick he was (the classic example is that Bruenn never got a single question from him about his medications or status and couldn't venture it to him) and even enjoyed playing games in that regard where he told Daisy Suckley that he'd figured out they weren't telling him the full story about how bad things were - and enjoyed that they didn't realize he had done so. Third, even if it still had been managed slightly better and he hadn't had a cerebral hemorrhage, his life expectancy from 1942 onwards would have likely been under a decade as his kidneys would have probably gone within a few years in any case given the long history of hypertension - along with dialysis not really being available for chronic conditions until the early 1960s.
There is good reason why the Framingham Heart Study started in 1948; FDR was just one of an epidemic of those in their 40s, 50s, and 60s who administered or fought the war and dropped dead at a relatively young age from a cardiovascular issue in the 1940s and 1950s. The amount of heart attacks and other cardiovascular issues leapt dramatically after the war, but it really was FDR's death in particular that set off the alarm bells and got the funding to begin research on the epidemiology. There's a famous newspaper war deaths list for April 12, 1945 which has at its top 'Commander in Chief, Roosevelt, Franklin D., Hyde Park, New York." With what we know today about what the war did to his health, that entry was far more apt than they realized at the time, but it's also appropriate to note that the research launched by it saved tens of millions of lives as a result.