Was there ever a vaccine candidate that showed negative side effects many months or years after the patient was injected?

Question

The current prognosis for a COVID-19 vaccine to become available is "12-18 months" and a big part of the reason for such a long timeline is the fact that scientists want to ensure that the vaccine is safe enough. But did this ever happen in practice? An article on CNN mentions the fiasco of the 1976 vaccine:

In 1976, President Gerald Ford's administration reacted at speed to a novel swine flu outbreak, ignoring the World Health Organization's words of caution and vowing to vaccinate "every man, woman and child in the United States." After 45 million people were vaccinated, the flu turned out to be mild. Worse, researchers discovered that a disproportionately high number of the vaccinated -- roughly 450 in all -- had developed Guillain-Barré syndrome, a rare disorder in which the body's immune system attacks the nerves, leading to paralysis. At least 30 people died. Upon discovery of the risk, the program was terminated in late 1976. A crush of lawsuits against the federal government followed.

However this looks like a very rare complication, only causing problems for 1 in 100,000 patients. And even that complication could've been detected quickly with enough volunteers:

About a week after getting the swine flu shot, she recalled, “I was so weak I couldn’t push down the toaster button.” She spent a month in the hospital, paralyzed from the neck down, before gradually recovering.

So has there ever been a vaccine trial satisfying the following conditions?

1. Significant side effects were detected
2. Those side effects appeared in otherwise healthy candidates after more than 3 months since they were injected with the vaccine
3. These side effects were something other than birth defects (those obviously take up to 9 months to show up)

If not, what causes scientists to be so cautious about testing a new vaccine quickly?

Answer 1

How long can it take to figure it out if a vaccine gives bad side effects? How about 20 years? Because deciding if the observed side effects are caused by a vaccine or not is not actually trivial:

all three of Sabin’s OPV strains were approved for use in the US, and in 1961-62 they replaced IPV for routine immunization against poliomyelitis.

As soon as OPV was used in mass immunizations in the US, cases of vaccine-associated paralysis were described. Initially Sabin decried these findings, arguing that temporal association of paralysis with vaccine administration was not sufficient to implicate OPV. He suggested that the observed paralysis was caused by wild-type viruses, not his vaccine strains.

A breakthrough in our understanding of vaccine-associated paralysis came in the early 1980s when the recently developed DNA sequencing methods were used to determine the nucleotide sequences of the genomes of the Sabin type 3 vaccine, the neurovirulent virus from which it was derived, and a virus isolated from a child who had developed paralysis after administration of OPV. The results enumerated for the first time the mutations that distinguish the Sabin vaccine from its neurovirulent parent. More importantly, the genome sequence of the vaccine-associated isolate proved that it was derived from the Sabin vaccine and was not a wild-type poliovirus.

We now understand that every recipient of OPV excretes, within a few days, viruses that are more neurovirulent that the vaccine strains. This evolution occurs because during replication of the OPV strains in the human intestine, the viral genome undergoes mutation and recombination that eliminate the attenuating mutations that Sabin so carefully selected by passage in different hosts.

From 1961 to 1989 there were an average of 9 cases (range, 1-25 cases) of vaccine-associated paralytic poliomyelitis (VAPP) in the United States, in vaccine recipients or their contacts, or 1 VAPP case per 2.9 million doses of OPV distributed (illustrated). Given this serious side effect, the use of OPV was evaluated several times by the Institute of Medicine, the Centers for Disease Control and Prevention, and the Advisory Committee on Immunization Practices. Each time it was decided that the risks associated with the use of OPV justified the cases of VAPP. It was believed that a switch to IPV would lead to outbreaks of poliomyelitis, because: OPV was better than IPV at protecting non-immunized recipients; the need to inject IPV would lead to reduced compliance; and IPV was known to induce less protective mucosal immunity than OPV.

And yeah whether the benefits outweigh the risks of severe albeit seldom-encountered side effects is a balancing matter:

After the WHO began its poliovirus eradication initiative in 1988, the risk of poliovirus importation into the US slowly decreased until it became very difficult to justify routine use of OPV. In 1996 the Advisory Committee on Immunization Practices decided that the US would transition to IPV and by 2000 IPV had replaced OPV for the routine prevention of poliomyelitis. As a consequence VAPP has been eliminated from the US.

Yes, yes, I can see the objections already that with the current state of biology/medicine we'd figure it out faster now. YMMV, i.e. it's down to "expert opinion" whether we could completely avoid a repeat of VAPP.

Answer 2

Historically, vaccines eg 1955 polio vaccine had side effects due to manufacturing issues allowing live virus to appear in the vaccine in the Cutter company’s vaccine. In 1976, a swine flu vaccine was shown to have a risk of 1 in 100,000 of Guillaine-Barre Syndrome, a transient but serious neuromuscular syndrome. The 1998 rotavirus vaccine was shown to prevent serious life-threatening diarrheal illness in infants, but later studies suggested a risk of a rare form of bowel obstruction called intussusception. The recommendation for RotaShield vaccine was withdrawn.

These side effects generally appeared within weeks, months or a few years of the vaccine administrative.

Source: https://www.cdc.gov/vaccinesafety/concerns/concerns-history.html

Over time this led to significant increase in safety oversight including the FDA VAERS (vaccine adverse event reporting system). Many long-term vaccine safety studies and monitoring systems have been built since then.

For example, the Kaiser vaccine study center looked at hundreds of thousands of MMRV vaccine doses, for example. They found no long term side effects, but in this case the goal was to understand side effects relative to MMR & V(aricella) vaccines given separately versus together.

Source: https://www.webmd.com/children/vaccines/news/20150107/long-term-study-finds-measles-vaccines-safe

Long term side effects can be a complication of any medical intervention, such as medications or devices also, as well as compared to the risk of untreated disease not prevented or not treated. As a result, scientists and clinicians weigh the risks, benefits and alternatives and ideally partner with patients, parents and caregivers in informed consent. Where information has not yet been gathered about long-term safety — or durability of the effect — large long term safety studies have generally been required, and are also possible to detect with such systems as the FDA Sentinel system launched in 2008 and other similar safety systems available now (2020) that were not previously available. These systems permit longer range, larger population monitoring of both safety and effectiveness of vaccines and other medical treatments, and further detail by subgroups who may be at higher risk of side effects, or who may be at higher chance of benefiting from vaccines or medications.