Tuberculosis in the Americas: Current perspectives • 167 



reported, as arc abscesses on the external aspect of the thorax 

 (Hodgson et al. 1969; Yau and Hodgson 1968; Jaffe 1972; 

 Johnson and Rothstein 19.'S2; Rechtman 1929), 



Most clinical experience with rib tuberculosis follows the 

 pattern described by Tatelman and Drouillard ( 1953). These 

 authors report four types of tuberculosis of the rib: ( 1 ) cos- 

 tovertebral. (2) costochondral, (3) isolated body lesion, and 

 (4) multiple cystic foci. In each case, destructive processes 

 constitute the primary symptom, with the majority of the 

 lesions involving the body. These resorptive foci follow the 

 typical pattern for bone tuberculosis and are clearly dilTcrent 

 from the mild periostitic response described by Kelley and 

 Micozzi ( 1984). Certainly, the features described by Kelley 

 and Micozzi may indeed reflect tuberculosis. Persons en- 

 gaged in paleoepidemiological study should, however, also 

 be aware of the patterns typically taken by the resorptive foci 

 reptirted in the medical literature. 



Although Kelley and Micozzi (1984:386) correctly indi- 

 cate that tuberculosis is the most commonly observed inflam- 

 matory condition in the ribs, a differential diagnosis of rib 

 lesions attributable to tuberculosis must consider other forms 

 of pathology. For instance, Tatelman and Drouillard report 

 that "tuberculosis is second only to metastatic malignancy as 

 a cause of destructive lesions of the ribs" (1953:923). Thus, it 

 is clear that metastatic processes must be considered in devel- 

 oping a differential diagnosis of tuberculosis-like lesions in 

 ribs. 



In discussing differential diagnosis, Kelley and Micozzi 

 (1984:386) state: "Conditions to be considered in the differ- 

 ential diagnosis of skeletal tuberculosis are actinomycosis, 

 typhoid, pyogenic osteomyelitis and syphilis (Sinoff and Se- 

 gal 1975)." Sinoff and Segal's article, entitled "Tuberculous 

 osteomyelitis of the ribs: a case report," however, makes it 

 clear that their primary concern is with manifestations in the 

 rib. rather than "skeletal tuberculosis" in general, as sug- 

 gested by Kelley and Micozzi. The full quotation from Sinoff 

 and Segal reads: "The single most important differential di- 

 agnosis (in rib tuberculosis) is metastatic carcinoma, but 

 other possible diseases include: (i) actinomycosis, (ii) ty- 

 phoid or paratyphoid osteomyelitis, (iii) pyogenic os- 

 teomyelitis and (iv) syphilis" (1975:866). Thus. Sinoff and 

 Segal are focusing upon tuberculous manifestations of the rib 

 and emphasizing metastatic carcinoma. 



Kelley and Micozzi (1984:386) follow their listing of dis- 

 eases appropriate for differential diagnosis in "skeletal 

 tuberculosis" — actinomycosis, typhoid, pyogenic osteo- 

 myelitis, and syphilis — with the statement. "However, none 

 of these conditions commonly affects the ribs." This conclu- 

 sion is remarkable since Sinoff and Segal (1975) are clearly 

 concentrating upon forms of disease affecting the rih. And in 

 fact, one of the conditions most likely to produce just the type 

 of proliferative response described by Kelley and Micozzi is 

 actinomycosis. Tatelman and Drouilard (1953:932) describe 

 rib involvement in actinomycosis through direct extension 

 from lung and pleural involvement, emphasizing productive 



2agnb Paleopathology Symp 1988 



changes due to this condition. A classic description derives 

 from Edeiken (1981:792): "When the thoracic wall is af- 

 fected, the ribs may show destruction, although they usually 

 react with periosteal new bone formation and become thick 

 . . . which, although not diagnostic of actinomycosis, is 

 most suggestive." Clearly, actinomycosis and perhaps other 

 related disease processes must be considered when periostitic 

 reaction is observed on ancient ribs. Thus, while Kelley and 

 Micozzi may be correct in attributing the mild periostitis 

 observed in the Hamann-Todd series to tuberculosis, it is 

 clear that conditions other than tuberculosis can produce 

 periosteal reaction on ribs. 



The age profiles for the total Estuquina skeletal series and 

 the subsample with more than five observable thoracic/lum- 

 bar vertebrae (Table 3) do not differ significantly (Kolmo- 

 gorov-Smimov test, D„ „ = 0.06; D^^, = 0.1 1 1; p > .05). 

 For the purposes of further testing, the affected individuals 

 are compared with the vertebrally observable subsample. 

 The two subgroups, rib and nonrib. indicated in Table 4 do 

 not differ significantly in age structure (D^, „ = 92; D^^,^ = 

 142; p > .05). The total affected sample does, however, 

 differ significantly in mortality experience from the larger 

 Estuquina skeletal series (D,„ „ = 0.33; D^.„, = 0.29; p < 

 .01). This difference is influenced primarily by the elevated 

 numbers of affected individuals in the young adult and el- 

 derly adult years. The former include active cases, while the 

 latter comprise the more extreme and healed examples. Low 

 visibility and perhaps low prevalence of bony involvement 

 likely affect the figures for young juveniles. 



These results indicate that young adults are overrepre- 

 sented in the affected sample, as they were not in the series 

 reported by Allison et al. (1981). This apparent elevation of 

 young adult morbidity, whether the full series or only that 

 with more than five thoracic or lumbar vertebrae present is 

 used for comparison, is expected for a tuberculosis-like 

 pathology. The pattern is characteristic of the larger North 

 American samples reported in Table 1 . 



An unusual feature of the Estuquina series is, however, an 

 apparent skewed sex ratio. Although males are overrepre- 

 sented in the full sample (99 males, 71 females), the affected 

 adult sample presents a decidedly more extreme bias (19 

 males, 7 females). When the tubercular sample is subdivided 

 by area of involvement (Table 4), we find rib lesions in 8 

 males and 4 females, while among the remainder, 12 males 

 and only 3 females are affected. Two of the three females 

 with classic Pott's disease are older individuals with extreme 

 degeneration of the spine (M6-102la, M6-5390). A chi- 

 square compari.son indicates that sex ratio of affected individ- 

 uals is significantly different from the total sample (x^ = 

 4.346, p = .037). Partitioning the affected group by lesion 

 location demonstrates that the individuals presenting rib in- 

 volvement are not so biased toward males as those presenting 

 classic Pott's disease. The comparison for those with rib 

 lesions yields a X' of 0.377 (p = .539). The value for the 

 more "classic" examples is 3.204 (p < .073). 



