Ok, I've been doing some research on the original question.
Life expectancies for all social classes and virtually all societies prior to 1800 seem to be fairly constant. Infant mortality rate runs from 15-20%. Only about half of all person's born survive to age 15. After age 15, whether by the simple expedient of anyone living through childhood probably having a high constitution, or whether by having developed a more robust immune system, the average person could - depending on where in the world he lived - expect to live between to between 43 and 52. Anyone above those ages could be considered 'old' - and would probably show it in his frame, teeth, and face.
Note that I've broken life expectancy out into two groups. Infant mortality is so high that it skews the numbers. Saying that the average life expectancy was only 26-28 skews the numbers.
One of the interesting things is that there doesn't seem to be much advantage - at least in terms of how long you live - in being wealthy. Between the greater risk of dying in battle, dying from disease while on campaign, dying in a duel, and the greater likelihood that you lived in a crowded unhygenic community that would immediately transmit any disease to every member of the city or household, being noble was of debateable advantage. Some studies have suggested that nobles lived shorter lives than thier peasant counterparts. Another interesting thing is that prior to the Enlightenment era, there doesn't seem to be a marked relationship to technology level and life expectancy. There does seem to be something of a relationship between caloric intake and life expectancy, but this isn't quite the same thing. For example, its estimated that the Aztec citizens had higher caloric intakes (and possiblely longer lives) than the Spainish citizens that conquered them.
On the other hand, lest you think I'm romanticizing the primitive, the lower the tech level the greater percentage of the deaths appear to be related to violence. In medieval Europe, only about 1% of the adult population died violent deaths. In nomadic hunter gatherer tribes, the numbers appear to go as high as 20%. Apparantly no society is quite as prone to total war as loosely organized family bands, and if you would allow me to go off on a tangent I might theorize that the purpose of governments is to insulate the majority of the group from the depredations of war.
Anyway, back on topid, if we assume that by slaves we mean 'adult slaves', and we assume that slaves aren't under normal circumstances living much less long than thier masters, then that suggests that each year conservatively 3.57% of the slaves die. After 10 years, 69.5% of the slave population will still be living slaves. This assumes of course that the practice of slavery is only moderately harsh.
Getting numbers for harsh slavery is complicated by two things. First, we have almost no data on slaves of antiquity. Second, the one harsh slave system that we have good data on is post 1800 and hense some of our assumptions are invalidated. As best as I can tell, the consequences of a being a slave in the antebellum south is that you were forced to continue to live a life of medieval misery and squallor while your white master's were increasingly enjoying the advantages of education, medicine, science, and modern agriculture. Hense, while white child mortality was falling by a significant degree, and white life expectancy rising significantly for the first time in history, black slave life expectancy remained as bad or worse as medieval France in the bad years - an average of just 22 years. Most of that appears to be due to infant mortality - probably as a result of overworked mother's recieving insufficient nutrition during prenancy - and I've not seen good estimates of how long adult slaves could expect to live - though I did find that 10% of the slave population was over 50, which appears to suggest medieval life spans. White life expectancy had risen from a probable average of 30 or so to the low 40's.
So, just as wild extrapolation, let's say that harsh slave conditions result in doubling of the death rate across the board. That's probably not a good assumptions, since it seems more likely that the harshness of slavery impacts the young disproportionately hard, but it's the best I can do. That yields a estimate that each year, under harsh conditions, 7.14% of all slaves die. After 10 years of such hardship, only 47.7% of the slaves will still be alive.
Now, you seem to want numbers on even harsh fantasy slavery, such as the life of the galley slaves in Ben Hur. Keeping in mind that in the real world, no ancient empire used slaves in galleys (that practice didn't start up until the 15th century or so, and it wasn't in warships that I'm aware of), if you really want numbers for slavery as a death sentence, you could double the numbers a few more times - even though they stop making economic sense. At double the estimate for real world harshness, you get only 21% of the slaves still alive after 10 years. Double it again, and only 3.5% of the slaves are alive after 10 years.
Those kind of numbers only make since in unusual circumstances - early years of the sugar boom in South America, for example. Slaves could be purchased cheaply from African tribes who were warring on thier neighbors, sent a short distance across the sea to Spainish colonies to labor under extremely harsh conditions, and the profit was high enough to justify the blood. Or they could be used for actual slavery as a death sentence, say for Gladiators or convicts laboring in the 'salt mines', or whatever.