We tried to apPly the two−compartment model to the metal)olic process of intramuscular inj ection of Amikacin. On its way we came across an evidence, 、 which suPports our hypothesis・This evidence is Verified by some other drugs. 1°L・tthe. c・ncentrati・n。f・・Ub・t鋤・・in.b1・Od・t t h・urs・fter it・ ad㎞inistrati㎝be O rt). Then according. to the two’−co叩artment mode1, we have 〔1・1) σ「t)一・1・rp(λユt)チ・2・Mλ2妨 f°「th・・ang・⇔㌔〆砲ere・i qnd Xi「i ・=ヱ・2)are c・n・t飢t・・λヱ・λ2 i・ assumed. Then, h being a given constant, the fbllowing difference equation is derived from (1.1). . .
(1・2) 0「t≠2hノーShO「t・h)チPバ0・
where we set (1・3) Sh=e・rp rλヱhノ・・arp’rλ2h)飢d Ph 一 ・up[Cxヱ+λ2周・ Sh and ph are the sum and the product of two roots of a quadraticequatlon
蕊)discr三,5蕊・=一゜;。h,
Let the notation be indexed by i for the i−th individua1. This index may be dropPed, if there is no fear of confusiop・ Acc・rdi㎎t・・u・h)rP・thesi・・th・,c・rre1・ti・n. b・tween S玩飢d P厄鵬t be high・Ass㎝・th・t th・.・・rrel・ti・n・・effi・i・nt・h b・tween these tw・ parameters be≠ヱ. Then the straight lines(1・5) 〃−Shi・’ 一 Phis‥」・2・…・・
are
@concu「rent・or there exists a fixed point (αh,βh) sudh that (1・6〕 βh = ・・hShi 一・ Phi・ Fig. 1 is drawn for.the first example in. theヒnext section.261
262
M.MASUYAMA
5 Fig. Replacing h by 2h, we should obtain by our hypothesis the relation(1・7) β2ゲ・2施厄一P2hi・.’ .
Since there exist algebraic relations ・(…〕 ・、バ・膓…ら2−・、h…h
by・h・d・f血・・i・n〔…〕,repre・en・血・ph2 t・・f・・d͡・dra・i・f皿・・i・n・麺 Sh, we obtain the following relations.〔…〕 ・2バ・h2−・h・孤d・2、一・h2・
輪・he c∞・dina・…f・h・previ・u・ fixed p・i・t are gi・・n by r・h.・h2ノ・ Referring to the equation (1.4),we conclude that one root of it isin・・p㎝・・…f・h・血・司孤d h・nce・h−・、h・
Si五ce・㌍1h・1d・・nly ・pP・・x迦・t・1y, thi・m・孤・th・t th・血・ng− individual variability of one root is sma11. By the way, we notice that when th・di・cr血m孤・Dh i・n・g・・i・…h・n 1・hl<<・一一%iS d迦…i・n・ess−一, suggestj丑g that it owes most probably to errors in measurement. Before proceeding to the next sect‡on,、we note that even though two 「lmuom va「iables Xl and X2 are㎜tually independent, their sum S and their P「oduct P are not independent. .we obtain the spurious correlation coefficient (1.11〕 ,一。。vrs.・Pノ/』r5ノ.。副pノ]1/2. 20 We set h = ヱrhrノ. Example 1. Intramuscular inj ection of 200mg Ami](acin. (Pig. 1 G 2〕 The following table is made of the data given in the reference [2]. Subj ect l 2 3
4510 78Qン
Olつ乙111
5 1.264 0.582 ヱ.167 −0.295 1.297 0.597 −0.175 0.349 0.322 0.326 2.002 0.ヱ85 P O.423 −0.028 0.333 −0.598 0.449 −0.042 −0.554 −0.ヱ95 −0.139 −0.238 0.969 −0.320 al O.6267 0.6699 0.6397 0.6606 0.66ヱ9 0.6493 0.5671 0.6774 0.8ヱ93 0.6657 1. a2# −0.0447 0.4971 −0.9347 −0.0636 −0.8369 −0.3003 −0.245ヱ ー0.3514 ヱ.2827 −0.4807 rDl=−0・094ノ a)1=−0・114) hrst=ヱ∼4
Tableβヱ.㌘S。e惣;el。lsg2,隠。W隠t,12,51撰ぷ;、1’68°
The c・rrelati・n・・effi・i・・t b・tween S ・nd P is equal t・・ヱ=0・9978 f・r ・−il2・ubj ect・・lf・w・・h・1ve at・・nd・m th・・amPl・・th・n w・・btqip. ・、 一・・9994 f°・・ubj ect・#ヱNfl 6・and ・ヱー0・9979 f・r・Ubj ect・#7∼fl i8・H・nce th・ correlation is significant. If we assume th…d・r ・」>92 b・加een tw・r・・ts・th・n w・・btain Ml =0・700ヱ・M2 =0・0547・γ2=0・0298・V2 =.0・2918 ・nd P=0・2604・H・nce our high correlation is not attributable to「the spurious one. . The c・・re1・ti・n・・effi・i・・t・2 b・tween S2・・d P2 i・equ・1 t・0・9850・ Exa皿ple 2. Intra皿uscular inj ection of ヱ20∼ηg Tobramycin [.5].. (Fig. 3〕#91=0・6638・・ヱー0・0632∫
z2=−0・1578・s2 = 0.6220.264
M.MASUYAMA
1.0 0.5 0.0 一〇.5 0.0 1.OFig.2
2.0 F 10 0 P. Tobramycin 0 10Fig.3
205 6.31 4.6ヱ 2.96 1.72 ヱ.085 0.323 6 6.24 4.ヱ6 2.75 1.3ヱ 22.429 ヱ4.5ヱ2 Table 2. rヱ = 0・9999 and r2 = 0・999 998・ Exanlg)1e 3. Naldixic acid 〔250∼ng tablet)[2]. (Fig. 4) n=6. r = 0.9990・ ヱ L。t f。u。。bservati。。, at’.t.. 6.8.24. and 32h「s b。 。n。1y、ed。、 if th・y were・bse・v・d・t・qu・1 t血・int・rva1・th・n we・bt・in ・1 一 O・ヱ1・H・nce if the model is not valid, we obtain smaller pヱ. ア2=0・99997・ Exalnple 4. Metronidazole (251mg sugar−coated pi11) [2]. 〔Fig. 5) n= 10. r = 0.9995. r = 0.999 97. ヱ 2 勺︹odu柏×パ㊥HdZ Subj ect 1 2