Diabetes Mellitos at tbc Massachusetts GeneralHospital From 1824 to 1898. A Study

of the Medical Records.

Presented to the Section on Practice of Medicine, at the FortyminthAnnual Meeting of the American Medical Association held at

Denver, Colo., June 7-10,1898.

BY REGINALD H. FITZ, M.D.,and ELLIOTT P. JOSLIN, M.D.

BOSTON, MASS.

REPRINTED FROM THEJOURNAL OF THE AMERICAN'TSKQ^UAT'J^NQj^ATW^I

JULY 4S. 1898j

CHICAGO.:American Medical Association Press.

1898.

DIABETES MELLITUS AT THE MASSACHUSETTS GENERAL HOSPITAL FROM 1824

TO 1898. A STUDY OF THE MEDI-CAL RECORDS.

BY REGINAL H. FITZ, M.D.and ELLIOTT P. JOSLIN, M.D.

The examination of the records of the cases of sac-charine diabetes treated in the wards of the Massa-chusetts General Hospital during the past seventy-four years was begun for the purpose of seekingevidence with regard to the pancreatic origin of thisdisease. (See Yale Medical Journal, 1898, iv, 275.)Since several hundred volumes were to be searched itseemed desirable to obtain at the same time informa-tion upon other points which might contribute to thestatistics of diabetes. Facts of interest in diagnosisand treatment were elicited as the study progressedand this paper is offered rather as a historical sketchof the progress of a disease drawn from the originalrecords than as an attempt to advance in any way ourknowledge of the subject. Since diabetes mellituswas first sharply differentiated from diabetes insipi-dus early in the present century, the hospital recordsmay be considered to be nearly as old as is our knowl-edge of the disease in question.

During these seventy-four years the total numberof cases of diabetes mellitus treated in the medicalwards was 172, but of these, 9 were readmitted, mak-ing 181 separate entries. The following table showsthe variation in the number of such patients duringsuccessive periods of years and their proportion to

2

the total number of cases treated during the sameperiods.

It appears from this table that within the pastthirteen years as many cases of diabetes were admit-ted to the hospital as in the previous sixty-one yearsand the per cent, in the past thirteen years in propertion to the total number of hospital entries has in-creased fourfold over that of the first fifteen years.This fact is not due to any recent increase in thenumber of beds open to this class of patients, nor,presumably to any considerable excess in the fre-quency of this disease in the region from which thehospital receives its patients. Whatever may be itscause this occurrence is to be regarded as evidence ofthe wholesome tendency of diabetics to place them-selves under careful medical supervision for suchlength of time as shall permit suitable observationsto determine tne severity and controllability of thedisease, that the future conduct of the patient mayaccordingly be regulated.

Of the 172 cases, 127 (or disregarding fractions,74 per cent.) were males and 45 (or 26 per cent.) werefemales. The oldest patient was 73 years of age andthe youngest 5 years, the average age of 161 patients

Period.Total num-

berofpatients.

Number ofcases ofdiabetes.

Per cent

From 1824 to 1840 5,328 7 .13From 1840 to 1855 6,136 16 .26From 1855 to 1870 8,660 24 .27From 1870 to 1885 11,690 39 .38From 1885 to 1898 16,085 86 .53

Age. Number ofcases. Per cent.

From 0 to 10 years 3 1.8From 10 to 20 years 15 8.8From 20 to 30 years 41 24.1From 30 to 40 years 43 25.3From 40 to 50 years 29 17.1From 50 to 60 years 28 16.5From 60 to 70 years •

• . . 11 6.4

3

being 33.4 years. The average age among 40 femalepatients was 39.1 years, being higher than that in 121males in whomit was 31.4 years. The proceeding tablerepresents the time of life at which the disease began.

Diabetes thus is shown to be a disease predomi-nantly affecting the middle third of life, though spar-ing neither the child of five nor the seniorof seventy-three.

But one of the patients was a negro, a further illus-tration of the rarity of diabetes in this race, althoughTyson (Practice of Medicine, 1896,751) has met withseveral instances. Ninety-seven of the patients livedin the United States, and 83 claimed this country astheir birthplace. Thirty-nine of the patients wereborn in Ireland, 11 in the British Provinces, 9 inEngland, 4 in Scotland, 7 in Germany, 2 in Swedenand lin Denmark. Forty-three of the patients weremechanics, 23 servants, 17 laborers and 13 farmers.Of merchants there were 5, and of clerks 7, Therewere 5 mariners and 4' teamsters. Although four ofthe patients were students there were but two profes-sions represented, a clergyman and an artist beingincluded among the entries. There were single or afew representatives of several other occupations, andconspicuous for its absence was that of liquor dealer.

In 42 cases the question of an inherited tendencyto the disease was raised and this factor was stated tohave been present in 10 patients and absent in 32. Inone case the father died of possible diabetes and thisdisease was the cause of the death of a brother and hisdaughter. The brother of one diabetic patient fellupon the ice, striking his head and died of diabetesfour days later. In three of the cases the symptomsof diabetes came on within a few weeks after severeinjury.

One of these patients while working in a mill in Julyso strained his back by lifting a heavy weight that hewas unable to accomplish much afterward on accountof supposed weakness of the spine. In August theappetite became increasedand in December he entered

4

the hospital for treatment of diabetes. The secondwas buried in d, sand bank three months beforeentrance and was taken out unconscious. Six weekslater thirst made its appearance as the first symptomindicative of the onset of the diabetes. The thirdfell fourteen feet and struck on the back of the head.He was unable to work for a fortnight. Very soonafterward he began to have great thirst, hunger, andfrequent micturition of large quantities of urine. Thedisease began in a female patient one month aftervery hard work in caring for a sick sister. The pos-sibility of contagion was entertained in the case of aservant whose mistress was suffering from a severevariety of diabetes.

Post-mortem examinations were made in 15 of the47 fatal oases. As is usual in this disease no charac-teristic lesions were found. The liver generally wasnormal in appearance although sometimes enlargedand in one instance it was rather small. The kidneysas a rule were normal, were enlarged in one instance,congested and with fat drops in the cortex in another,enlarged and hemorrhagic in a third, enlarged andinfiltrated with round cells in a fourth. The stomachwas large in two cases and Brunner’s glands werefound “developed” in a case under the care of HenryI. Bowditch in 1848. The condition of the pancreaswas not mentioned in seven of the cases. This glandwas stated to be normal in five and small in threeinstances. In one of the three its tissues were notremarkable, in another there was a moderate infiltra-tion of fat tissue around the head of the gland, and inthe third the weight was‘recorded as an ounce and ahalf, the section showing a coarsely granular surface.

No attempt has been made to note the relative fre-quency and the severity of the various symptoms.Particular attention has been paid, however, to therecords of the condition of the urine, since they fur-nish a valuable commentary of the progress of ourknowledge concerning the methods of studying thissecretion.

In the earliest volumes of the records especial im-portance was attached to the measurements of thequantity of urine passed as compared with the quan-tity of liquids ingested, since formerly it was the beliefthat patients excreted more urine than they drankliquid. As an illustration the following table is cop-ied from therecord of a case.

TREATED IN 1861.

Date. Water drank—Pints. Urine voided—Pints.

Specificgravity.

April 28 7XA4J|23 1030

April 29 9+ 1030April 30 4 11 1034May 1 3 8 1033May 2 4 9 ?

May 3 4 10 1027May 4 4^ 10 1025May 5 About 4^| 5 1025May 6 About 6 10 1027May 7 About 7% 11 1025May 8 Out of house. 6 1025May 9 Out of house, 5. 9 1030May 10 Notknown. 6 1030May 11 Notrecorded,patient went

frequently to stool.May 12 6A. 10 1033May 13 4^ 10 1027May 14 7 11 1025May 15 7 10 1024May 16 6A 10May 17 6, bread out. 10 1027May 18 7 10 1030May 19 8 10 1030May 20 7 12 1030May 21 8/4 14 1031May 22 8 18 1031May 23 7A 14 1033May 24 10 17 1031May 25 7. bread renewed. 16May 26 16 1031May 27 14 15 1031May 28 16 1033May 29 8 16 1034May 30 10 1032May 31 10 1032June 1 m 12 1030June 2 6M 9 1031June 3 7 12 1030June 4 7H 9 1030June 5 11 1030June 6 e'A 12 1025June 7 7A 12 1026June 8 10 1023June 9 9 14 1022June 10 18 1022June 11 11 18 1020June 12 11 18 1020June 13 Out of the house. 1021

6

It is noted, for instance, in Vol. i, that a patientwith diabetes simplex from Dec. 2 to Dec. 11, 1821,ingested forty-five pints and seven gills, but excretedforty-five pints and four gills. The first patient whois recorded to have had diabetes mellitus entered thehospital in 1824, and during the first three days ofhis stay is credited with drinking twenty and one-half pints of fluid, and is charged with passingtwenty-five pints of urine.

This doctrine of an excess of outgo of urine overincome of fluids appears in 1847 in Watson’s Practiceof Physic (3d American Edition, p. 869) as a state-ment that the “quantity of urine secreted and voidedis sometimes enormous, far more than could be sup-plied by the quantity of fluid taken as a drink.”

Twenty years later, Harley in his work on diabetessays (p. 50), “diabetic patients generally pass moreliquid than they take, about one-fifth or one-fourthmore.”

Lauder Brunton in 1879 (Reynold’s System ofMedicine, 1879, p. 887) opposes this view in the fol-lowing words: “It has been said that the quantityof urine excreted by the patients is greater than thatof their beverages, but tliis can hardly be the case,and the observations which seem to support this viewhave probably been made for too short a time, theapparent excess being most likely derived from waterlodged in the system.”

The difficulty of eradicating this erroneous idea isindicated by the fact thata recent graduate of the Har-vard Medical School and of the Massachusetts GeneralHospital, seriously asserted that “ it is perfectly wellknown that a person passes more urine than he drinksfluids.” Such a view is based probably upon a mis-interpretation of the physiologic teaching that morefluid is eliminated from the body in respiration andin secretion from the various glands than is taken inas liquid. In this connection it is interesting to notethat a metabolism experiment in diabetes was madeat the hospital in 1845. For it is then recorded that

thirty-six times; voided thirty-six pints. ... Itcontained 3 per cent, of sugar, corresponding to 227grains in one pint.” Only six of 166 patients passedover 300 ounces in one day. The highest sp. gravityrecorded is 1054, although repeatedly it was foundabove 1045. The highest percentage of sugar men-tioned was 12.5 per cent. In many instances therewas present between 9 and 10 per cent, of sugar.

The finding of sugar in the urine was determinedin the first instance until 1851 by means of the taste.

These rude methods of testing the saccharine con-dition of the urine were supplemented in 1827 byevaporating the specimen. It is recorded “ urineyesterday sixteen ounces, . . . seven ounces thismorning, . , . sixteen being boiled about two re-mains (sic), resembling molasses, of saccharine smelland taste.” In the following year, “ urine seven pints,natural appearance, without sediment, not givingsyrup on evaporation.” It is stated also that eightounces of urine yielded one ounce of syrup, and at

9

It seems worthy of reproduction as an illustration ofthe scientific methods employed even fifty years agoin the investigation of disease.

Boston, May 12, 1846,Dr. J. B. S. Jackson.

Dear Sir:—The microscopic examination which I have madeat yourrequest, of the sugar from diabetic urine has affordedsome results worthy of mention. The specimens of saccharineurine placed in my hands were from two diabetic patients nowin the hospital. My first attempt was to obtain the sugar in acrystallized state by the evaporation of successive portions ofthe urine on slips of glass, both spontaneously and by the aidof a gentle heat. But the results obtained in this way wereunsatisfactory from the difficulty with which this variety ofsugar crystallizes, and the following process was substituted.The urine was evaporated over a salt water bath, to the con-sistence of a thick syrup, and drops of this, spread very thinlyon slips of glass, were left to spontaneous evaporation. In thecourse of a few days the liquid disappeared, leaving on theglass little hemispherical grains of a white color (none of themas large as the head of a pin), either isolated or collected intomammillated crusts. With the urine of the male patient, Pet-tigrew, two or three days were sufficient for this purpose ; thatof the female patient, Gould, which appeared to contain lesssugar in proportion to the other constituents, required morethan a week’s exposure.

The objects thus procured were prepared for the microscopeby covering them with Canada balsam. Being now examinedwith moderate powers, each little grain resolved itself into acompact group of fibers radiating from a central point, manyof their ends projected beyond the mass, but exhibited no dis-tinct crystalline faces. This radiated structure, with a finesilky lustre, becomes distinctly visible to the naked eye in thelarger grains, when they are covered with the balsam. Theaid of the polarizing attachment to the microscope was nowcalled in to decide the question as to the crystalline structureof the grains, when the rich and varied colors exhibited in thepolarized light left no doubt of their really being groups ofradiating acicular crystals. Among these masses of sugarthere were noticed a few groups of more distinct crystal, notarranged like them in a stellate manner, and presentingentirely different appearances in polarized light. These aredoubtless some of the saline constituents of the urine. Asdiabetic sugar is classed with that variety, to which from itsabundance in the juice of ripe grapes the name of grapesugar is given, it became of interest to compare the appear-ance of the two under the microscope. Grape sugar is presentin large quantity in honey, and is the variety formed fromstarch or from cane sugar by long continued boiling with dilute

10

sulphuric acid. This variety is entirely distinct in its charac-ter from the common or cane sugar. A portion was preparedfrom honey and dissolved in alcohol, the solution, evaporatedto a syrup and prepared for the microscope in the same manner as the diabetic urine, crystallized in groups of radiatingfibers like the diabetic sugar, and presented similar appear-ances under the microscope, both with and without the polar-izing attachment. An alcoholic solution of the diabetic sugarwas also prepared, which crystallized in the same manner;a similar solution of cane sugar afforded no radiated groups,but distinct prismatic crystals which gave fine colored ringsin polarized light. Under the microscope the urine of bothpatients (before evaporation) is seen to contain great numbersof colorless globules, often connected in chains or groups ofhalf a dozen or more. I made a cursory examination for urea,by evaporating an ounce of the urine from each patient to onequarter and adding an equal volume of nitric acid. Pearlycrystalline plates of nitrate of urea soon made their appear-ance in both, showing urea to be present in considerableamount, but the quantity of urine operated upon was toosmall to allow of a correct determination of the amount.

Yours respectfully, John Bacon, Jr.

Although Moore’s test had been made known in1844, the first mention of its use in the hospital wasin 1851, when the following record appears “glucosestrongly marked on the addition of heat and liquorpotassee.”

In 1852 Dr. Bacon determined, by what means it isnot stated, that the per cent, of sugar in a given urinewas 4 per cent, and two years later the fermentationmethod was employed for ascertaining the percentage.

The breath of diabetic patients first attracted atten-tion in 1835, when a man entered the hospital with anote from his physician stating that it had a “pecu-liar sour smell.” The observation made of thispatient while in the hospital was of “a curious smellin breath which he has had since entrance.” In threecases the breath is stated to have the odor of newmilk. In 1846 it is noted of one of these “odor ofbreath considered by some of aromatic sweetness.”Other expressions are in 1846 “odor of boiled milk;”in 1848 “ward distinctly tainted with smell peculiarto diabetes,” “fruity odor;” in 1869 “sweet chloro-form-like smell in neighborhood of patient’s bed.”

11

That theacetone odor of the breath then was attract-ing general attention is apparent in 1847 from thestatement made in Watson’s Practice (op. cit. p. 868)that “according to Dr. Front the scent somewhatresembles that of sweet hay, but to my nose it is morelike the smell of certain apples or rather of an applechamber.” He attributes the peculiar odor to thesugar. In 1866 in Harley on Diabetes (op. cit.) thereis no mention of acetone or of diacetic acid, althoughFetters (Prager Vierteljahrsch. 1857,Iv) several yearsbefore had attributed the coma of diabetes to theformer agent.

Acetone was sought for in the breath of 19 oases,and in the urine of 24 cases. Its presence wasindicated by the odor in 16 of the former, andby chemical tests in 14 of the latter series. Itis interesting to note in connection with the resultof recent investigations on the increase of acetone inconsequence of a change to an absolute proteid diet,that the records state in 1894 that the “acetone smellhas been more since disappearance of sugar than atany other time.” Ten days later, however, withoutany change in diet there had been no return of sugarand no smell to the breath.

The ferric chlorid reaction, which, when positive,is considered generally to be evidence of the presenceof diacetic acid, was sought for in 22 oases and foundin 13, Albuminuria was present in the urine in 67cases and was absent in 60 cases. Casts were foundin 86 cases.

In four instances only is there a record of the exam-ination of the feces. In 1844 a patient reported thathis “food frequently passes away in lumps undigested”and fat is reported as absent in the stools of threepatients.

In six cases there was tinnitus aurium and twopatients had sciatica. Sexual power is noted as absentin 9 oases and present in 4. The catamenia were pres-ent in 11 cases and absent in 9. Jaundice is men-tioned to have occurred in five cases, and three of

12

these patients died in the hospital, one of pneumoniaand one of atrophied pancreas. Syphilis is stated tohave existed in one case. Diabetes and myxedemacoexisted in one patient and in another diabetic therewas a tumor of the thyroid body. Three of thepatientssuffered from gangrene of the toe, a fourth had a gan-grenous slough of the tissues of the leg, and a fifthdied with a carbunoular-like inflammation of the thigh.Three of the patients had abscesses, one being of thehip, another of the wrist, both cases ending fatally.The third abscess was of the jaw, but the patient grewbetter. In one case there was a round ulcer of thetoe, which improved under treatment.

The records afford some evidence concerning theduration of the disease, which in one case is stated tohave existed for twelve years, while three patients hadsuffered but two weeks. Of the latter, one died, onewas relieved and one unrelieved. The average dura-tion was one year and a half.

The length of the illness in the fatal cases is indi-cated in the following table:

These cases, necessarily those of grave diabetes,thus ended fatally in nearly three-fourths of thenumber in which this point could be determinedwithin one year after the disease made itself manifest,and within two years in nearly seven-eighths of theentire series. That diabetes may prove fatal at theend of a longer period of time, from 6 to 7 years asin one instance, is attributable to the well known factthat cases of mild diabetes may become suddenlysevere, and that the symptoms of this disease may fora while disappear to return at a later period. Anillustration of the latter condition, one which may be

DUEATION. NUMBER OF CASES.From 0 to 1 year 27From 1 to 2 yearsFrom 2 to 8 yearsFrom 3 to 4 years 1From 4 to 5 years 0From 5 to 6 years 1From 6 to 7 years

Unknown

47

13

called recurrent diabetes, is to be found in the recordof a patient who was in the hospital in 1862. InMarch, 1861, he suffered from cough with abundantyellow expectoration. In the latter part of the monththe thirst increased and became excessive; the patientdrank freely and passed large quantities of pale urine.The appetite failed and there was progressive debility.At the end of four months the quantity of urine be-came normal. The patient remained apparently welluntil the following February, when excessive thirst,polyuria and debility returned, and it was observedthat the urine left a sticky and sparkling deposit oncloth. Anotherinstance of recurrence is noted in 1866.Two years previous to the entrance of this patient heobserved that an abnormally large quantity of urinewas being passed in the twenty-four hours and thatthe thirst was excessive. These and other symptomscontinued for three months and then disappearedwithout treatment. At the end of eighteen months,however, theyreturned, accompanied by general weak-ness and faintness. It is recorded that the housephysician tasted the urine of this patient.

Of the 172 patients 47 or 27 per cent. died. Themortality among the 127 male patients w Tas 39 or 30.7percent.; of the 45 female patients 8 died, a mortalityof 17.7 per cent. The death rate among the malesthus proved nearly twice as high as among the females.The following table illustrates the variation in therate of mortality during successive periods of years.

It has previously been remarked that from 1824 to1885 there were treated 86 cases, the same number ofcases as in the period from 1885 to 1898. It is cer-

Period. No. ofcases. No. Of

deaths.Percentageof deaths.

From 1824 to 1840 7 1 14From 1840 to 1855 16 7 44From 1855 to 1870 24 5 20From 1870 to 1885. . 39 11 28From 1885 to 1898 86 23 27From 1824 to 1898 172 47 27

14

tainly noteworthy that in each of these intervals therewas an average of 27 per cent, of fatal cases, whichagain was the average mortality rate of the casestreated throughout the entire period from 1824 toto 1898. It might be thought from the inspection ofthis table that the method of treatment was superiorfrom 1824 to 1840 and from 1855 to 1870 and that inthe intervening years it was inferior. The informa-tion on this point later given affords no justificationfor such an inference.

Although it appears from the study of the recordsthat most cases of grave diabetes die within a year ortwo of the onset of the disease and that 27 per cent,of all cases are fatal, some evidence has been obtainedwith regard to the prognosis in the individual case.Heredity appears to have less importance than isusually assigned to it. Of ten cases in which hered-ity may have been of etiologic significance threedied, while of thirty-two cases in which heredity hadno influence eight died. Thus the percentage ofdeaths was merely a trifle higher in those cases inwhich an inherited tendency to diabetes may haveplayed a part than in those in which no such influ-ence existed.

The time of life at which the patient suffers fromthe disease appears to be of greater prognostic im-portance, as may be inferred from the following table:

Thus about one-third of the patients sufferingfrom diabetes during the first ten years of life andbetween the ages of 20 and 40 years die, and nearlyone-half of the cases between 60 and 70 years prove

Age. No. ofCases.

Deaths. Per cent.

Prom 0 to 10 years 3 1 33From 10 to 20 years 15 3 20From 20 to 30 years 41 14 34From 80 to 40 years 43 13 30From 40 to 50 years 29 4 14From 59 to 60 years 28 5 18From 60 to 70 years 11 5 45

170 45

15

fatal. The number of cases of diabetes in youngchildren, however, is so small as not to be of muchvalue in generalizing. The high rate of mortalityafter the age of 60 is due, possibly, only indirectly tothe diabetes, the patients perhaps being unable toresist the effects of other disease, even if mild, in thepresence of so grave an affection.

Thirty-eight of the 47 fatal cases died comatose, thecoma lasting one day in 10 cases, two days in 14 andthree days in 5 patients. In the remaining 9 casesthe duration of the coma was not to be determined.The frequency of the coma in relation to the age ofthe patient is indicated in the following table:

The suddenness of its onset in certain instances isapparent from the record of a patient in 1828 who wasable to leave the hospital for a few hours, at the endof which he became comatose and died. In 1846another patient absented himself from the hospitalwithout leave and became intoxicated; two days laterhe was comatose and died at the end of anothertwo days. On the other hand, a third patient whilein the hospital became intoxicated, had no coma andlived.

It is to be observed that of twenty fatal cases ofdiabetes there was present chronic nephritis in six,pulmonary tuberculosis and carbuncle or gangreneeach in three, myocarditis in two and pneumonia,pleurisy, cerebral thrombosis, alveolar abscess andatrophied pancreas each in one. A patient with dia-betes suffered at the same time from acute catarrhalcholangitis, acute rheumatism and chronic bronchitis.

The study of the records of these cases affords some

Age. No. ofcases.

Deathsfrom coma.

Per cent.

From 0 to 10 years 3 1 S3From 10 to 20 years 15 3 20From 20 to 30 years 41 12 29From 30 to 40 vears 43 9 21From 40 to 50 i ears 29 3 10From 50 to 00 years 28 4 14From 60 to 70 years 11 3 27

16

evidence relating to the prognostic value of the pres-ence of acetone and diacetic acid. Although the odorof acetone was present in the breath of 12 out of 14fatal cases, that is, in more than 85 per cent., and inthe urine of 6 out of 9 fatal cases, namely 66 percent.,in which the records show that attention was directedto this point, it was present also in the breath of 4out of 5 non-fatal cases, 80 per cent., and in the urineof 8 out of 15 non-fatal cases, 58 per cent. It is evi-dent from this comparison that the presence of anacetone odor in the breath or in the urine throwsbut little light upon the severity or mildness of thedisease.

The ferric chlorid reaction was present in 6 out of8 fatal oases and in one-half of the 14 non-fatal oasesin which it was sought for. This evidence, so far asit goes, therefore, is in support of the view that thisreaction has a somewhat serious significance.

It is interesting to note that in the earliest days ofthe hospital the improvement of the patient’s condi-tion was based upon an increase in weight.

It is to be remembered that John Rollo (Notes of aDiabetic Case, 1897) first ascertained that an animaldiet not only decreased the quantity of urine, butalso diminished the amount of sugar. The diet,therefore, from 1824 to 1840 was based evidently uponthis discovery and the first diabetic patient in thehospital was fed on “animal food and bread in smallquantities. Only one potato (sic) at dinner.” Latermolasses, beer and apples were excluded from his diet.In 1827 a patient’s fare consisted of two eggs, at ameal, meat, milk, butter and salt. Such a diet, how-ever, was not exclusive, for mention is made that vege-table food and sugar were not to be taken, but riceand milk porridge were allowed. Toast water andbroth were permitted, and crackers were given insteadof bread. Thirst was quenched with chamomile teaor balm tea, and lemons also served for this purpose.

The unlimited use of liquids was not permitted, forone patient was allowed only a “teacup of liquid at

17

meals.” Another was given, “balm tea as little as willsuffice,” and it is stated of a third that he “thinks hehas not drank all his allowance.”

The first treatment the first patient received was awarm bath, and as the quantity of urine appeared tobe less after the bath had been taken it was orderedto be repeated every other day. Warm salt bathsalso were used.

The advantages of exercise were appreciated at anearly period, for one patient in 1881 was advised toride horseback, and another was recommended to takeas much exercise as his strength would permit.

Alcohol was given as early as 1828, when the phy-sician in charge ordered, “let her take some brandyand water, taking one teaspoonful of brandy in hot orcold water, repeating it once in 3 or 4 hours if theeffects are grateful.” There is nothing in the recordto show that its action was displeasing to the patient.The recent investigations of von Noorden (“DieZuckerkranheit und Ihre Behandlung,” 1895, 146),show that this treatment has a sound scientific basissince one gram of alcohol is the equivalent ofseven calories, while the same quantity of proteidsand carbohydrates is equivalent only to four calories.

This patient was treated also with opium and “byerror took following in A. m.

ft Infus. Sennse Comp., 3iij.”On the next day he “was uncomfortable in head

with some depression of spirits.” During this patient’sstay in the hospital he was ordered powdered rhubarband the infusion of rhubarb and magnesia; a trial wasmade also of uva ursi andof the sodium and potassiumtartrate.

During the first fifteen years the treatment in gen-eral consisted of opium, either alone or with camphorand aloes. Cathartics were used liberally, and fre-quent mention is made of rhubarb, aloes, colocynth,senna, calomel, magnesium sulphate, castor oil andcroton oil. Enemata were employed freely, some ofthem being especially designated “active.” One

18

patient who complained of nausea promptly receiveda dose of ipecac.

Among other drugs in use during this period werequinin and dilute nitric acid, the decoction of sumacand powdered guiac.

Blisters and cupping often were used, and onepatient “thinks opium and blisters have been fol-lowed by the greatest improvement.” More severecounter-irritation was ordered sometimes, as appearsfrom the record, “apply caustic issue on each side ofthe spine.” In one case this application proved tobe directly injurious.

During the period of high mortality, that is from1840 to 1855, the following dietary is ordered: “Leanmeat, with a small quantity of stale, dry or toastedbread, avoiding all fatty, farinaceous and saccharinearticles. For drink, cold water and weak tea.” Theuse of liquids continued to be restricted, althoughslippery elm tea now came into use. Alcohol wasnot given regularly, and blisters no longer wereapplied. Leeches appeared to serve in the place ofthe latter and of issues for it is recorded, “Leach (sic)bites now ulcerated.”

Opium continued to be given, but not so largely asprevious to 1840. Cathartics were prescribed withfreedom, and the liberal use of calomel is suggestedby the frequent records of spongy gums and myrrhmouth washes. Iron now was used in the form ofthe tincture of the chlorid or the carbonate. Otherremedies employed were argentic nitrate, ammoniumhydrosulphuret, ammonium carbonate, conium, hyos-cyamus and valerian, aromatic sulphuric acid andcreosote. During this period the use of yeast wastried by McGregor (Watson’s Practice, p. 872), whostated that it made the patients feel as if they were“on the eve of being blown up.” A patient in thehospital was ordered to “take yeast 3ss. in milk terdie” This quantity was increased to an ounce dailyand then was resumed and omitted at intervals dur-ing a month, the yeast being taken five-sixths of the

19

time. The effect was not that mentioned by Watson,but a sense of constant nausea was the chief result-ing complaint.

From 1855 to 1870 especial attention was paid tothe selection of a suitable bread, and bran bread, stalebread, brown dyspepsia bread, toasted dyspepsiabread, ship biscuit and toasted ship bread were or-dered. Butter, cabbage, cheese and greens wereallowed, but shellfish and liver were excluded.Liquids at times were restricted, and cocoa was givensomewhat frequently. Alcohol was ordered occasion-ally, and one patient, notwithstanding milk punchwas prescribed p. r. n., remained in the hospital buta single day, at the end of which he was dischargedunrelieved. Opium was but little used, counter-irri-tants were discontinued and there was but little va-riety in the use of drugs. One patient received ahalf pint of iron rust water at dinner. To anotherin 1866 was given a half ounce of cod liver oil threetimes daily. As “ she likes it and don’t have enough,”the dose was increased to three ounces three timesdaily. This maximum dose was reduced in conse-quence of the production of loose movements of thebowels.

From 1870 to 1885 gluten bread was used, but wasdiscovered to contain 83 per cent, of starch. Patientswere advised “ to drink as little water as possible,”but as much as four quarts of skim-milk were givendaily. Opium and codeia were used frequently, butalcohol, blisters, baths and exercise were not advisedas a part of the routine treatment. On the contrary,a large number of drugs were employed, including

tier’s solution, bromid of arsenic, antimony, bis-

405). On the day after his admission he was giventhirty-three ounces of glycerin and died comatose onthe same day. The coma was accompanied by spasms,to relieve which ether was given, but the house phy-sician apparently feared that the death might beattributed to the effects of the ether.

In the period since 1885, Soya bean bread has beenadded to the dietary, and fats are used with freedom.The use of water is frequently, by no means always,restricted, and the use of skim-milk gradually haslessened. Alcohol, baths and exercise, are not madematters of routine. Opium and codeia are more oftenused thanbefore, and trials have been made with potas-sium citrate, lithium carbonate, jambul, salol, salicin,antipyrin and urethan. Saccharin occasionally is used.Up to this period comatose patients had been treatedsolely with stimulants and diuretics, and brandy,nitroglycerin and oxygen were given frequently atthe onset of coma. The use of solutions of sodiumohlorid or sodium bicarbonate in enema, subcutane-ous injection or by transfusion was begun and eightpatients were thus treated. Although these patientsdied, as had those otherwise treated, it appears from therecords that in certain instances life was somewhatprolonged. The patient’s mind usually became clearerafter the administration of the saline, and one per-son was aroused to such a degree that he is recorded tohave said, “ Oh! doctor, I’m glad you did that; I havenot felt so well for a long time.”

The practical outcome of this examination of therecords of cases of diabetes treated at the Massa-chusetts General Hospital during the past seventy-four years can be stated in a very tew words.