Do the Methods Justify the Conclusion?

readily available to clinicians. This was also a few years before we began to use mechanical ventilation for conditions other than po¬ liomyelitis, spinal cord injury, and kyphoscoliosis. Many surgeons could not duplicate Brantigan's mortality and morbidity. Although spirometry was slowly coming into general use, the clinical pulmo¬ nary function measurement was still in its early days and Brantigan presented no objective preoperative and postoperative data. Finally, the pulmonologists of the day were busy controlling the epidemic of tuberculosis and learning how to do so in the outpatient setting as sanitarium beds were closed apace. Thoracic surgeons were busy resecting residual "open negative" tuberculous cavities and solid

caseous

foci.

I did not mention Brantigan's name in my recent article in CHEST (1996; 109:540-48), because Brantigan's contribution was

to have the insight to recognize that reduction pneumoplasty would work for emphysema in the absence of a giant bullae. His formu¬ lation of the physiologic basis for operating on emphysematous pa¬ tients in the absence of bullae is slowly being confirmed as physi¬ ologic studies on lung volume reduction surgery are published. I have reviewed Brantigan's seminal contribution to reduction pneu¬ moplasty in another article.1 Since my review was on reduction pneumoplasty for bullous disease, I thought it inappropriate to present it there.

Gordon L. Snider, MD, FCCP, VA Medical Center,

Boston

Pneumocystis carinii pneumonia treated with corticosteroids, were also CMV positive, had a two times higher 3-month mortality than those who were CMV negative (p=0.08). This was not accounted for by differences in HIV helper cell count, P02, dura¬ tion of AIDS, or age. There are two possible interpretations for this

vere

who

finding. Steroids may further immunocompromise these HIV patients and allow a "colonizer" to become pathogenic. Alterna¬ tively, CMV infection may simply be a prognostic marker, which identifies patients likely to do poorly. In the study by Hayner et al,2 the retrieval of CMV was also associated with significantly greater 3- and 6-month overall mortality. To clarify the pathogenetic role of CMV in HIV-infected patients, what is needed, therefore, is a prospective intervention trial. A co¬ hort of HIV patients infected with CMV, based on BAL findings, would be randomized to receive ganciclovir therapy or placebo for 6 months. (Since we could expect to see differences in mortality by 6 months, as per the study by Hayner et al,2 this would seem to be an adequate treatment duration.) The lowest effective dose should be used to avoid drug toxicity. If there were a significant difference in observed mortality, then CMV infection is worth treating. If, on

the contrary, no difference were seen, then CMV would seem to be more of a marker of poor advanced immunosuppression state and thus an adverse prognostic marker. It is clearly of great importance to define the significance of CMV in this population, as larger proportions of AIDS-afflicted patients are surviving to later stages and one would expect a larger percentage of patients infected with CMV. Maria Antoniou, MD, and Charles K Chan, MD, FCCP,

Reference

technology assessment of surgical pro¬ cedures: the case of reduction pneumoplasty for emphysema. Am J Respir Crit Care Med 1996; 153:1208-13

University of Toronto, Toronto, Ontario, Canada

1 Snider GL. Health-care

Determining the Pathogenetic of Cytomegalovirus in Significance Patients With AIDS To the Editor:

Despite extensive research into the significance of identifying cytomegalovirus (CMV) in BAL fluid of patients infected with HIV,

clear understanding of the pathogenetic significance of CMV is still beyond our grasp. There are many reasons for this, foremost among which is the lack of a uniform definition of CMV pneu¬ monitis: should it be based on evidence of CMV infection (culture positivity) or disease (cytologic changes)? The latter seems to result in a low diagnostic yield, as seen in a study by Miles et al,1 in which, of 62 patients proved to have CMV infection by culture and immuno¬ fluorescence stain, only 5 had positive cytology for CMV. In a re¬ cent issue of CHEST, Hayner and colleagues2 found that the odds ratios for predicting mortality based on BAL evidence for CMV were comparably high whether CMV positivity was defined by culture or cytology. What criteria, therefore, should we be using? The high prevalence (28 to 52%)1>3 of CMV in patients with HIV potentially obscures its contribution to the high morbidity and mortality associated with HIV-related pulmonary complications. The added difficulty of distinguishing CMV disease from coinfection is highlighted by the results of an autopsy series,4 which dem¬ onstrated that 94% of 31 patients with CMV pneumonitis (positive culture and cytology) had coexisting pulmonary disease. The advent of steroid therapy for severe Pneumocystis carinii pneumonia has added a twist to this dilemma, which may result in a better understanding of the pathophysiology of CMV pneumoni¬ tis. In a recent article in CHEST, by Jensen et al,3 patients with se¬ a

References 1 Miles PR, Baughman RP, Linnemann CC. Cytomegalovirus in the bronchoalveolar lavage fluid of patients with AIDS. Chest 1990; 97:1072-76 2 Hayner CE, Baughman RP, Linnemann CC, et al. The relation¬

ship between cytomegalovirus retrieved by bronchoalveolar lav¬ age and mortality in patients with HIV. Chest 1995; 107:735-40 3 Jensen A-MB, Lundgren JD, Benfield T, et al. Does cytomega¬ lovirus predict a poor prognosis in Pneumocystis carinii pneu¬ monia treated with corticosteroids? a note for caution. Chest

1995; 108:411-14 4 Wallace JM, Hannah J. Cytomegalovirus pneumonitis in patients with AIDS: findings in an autopsy series. Chest 1987; 92:198-203

Do the Methods Justify the Conclusion?

Dopexamine and Splanchnic Blood Flow To the Editor: In a recent article in the journal, Maynard et al (CHEST 1995; 108:1648-54) describe the effects of dopexamine and dopamine on gastric intramucosal pH, monoethylglycinexylidide formation from lidocaine, and indocyanine green clearances in mechanically venti¬ lated critically ill patients. The authors conclude that low-dose dopexamine increases splanchnic blood flow as measured by gastric

intramucosal pH, monoethylglycinexylidide formation from lido¬ caine, and indocyanine green clearances. This conclusion was based on data as obtained from three comparable groups of patients: a group treated with dopamine (n=10), a group with dopexamine (n=10), and 1 control group treated with saline solution (n=5). We would like to make a comment on the methodology, for it remains CHEST / 110 / 3 / SEPTEMBER, 1996

863

unclear how the data of the control group were used. The authors in the control group." This suggests no state, "there were changes that data of the control group were not compared with the data of the study groups, but data of different time points were compared within the group. It is therefore of interest that the number of five patients is too low, and with nonparametric data it is hard to dem¬ any statistically significant change. Additionally, it is known that the vasodilator properties of dopexamine can induce a fall in BP, which can be corrected by the addition of extra fluids.1,2 However, the authors do not describe the fluid requiiements of the patients during the study group, although in our opinion, these data are of interest. Finally, we assume the given doses were in micrograms not mil¬ ligrams as stated, and that not all data in the tables were "after dopexamine." onstrate

Mark T Patten, MB, ChB, and Armand R.J. Girbes, MD, PhD,

Surgical Intensive Care Unit,

University Hospital Groningen,

Netherlands

References 1

Colardyn FC, Vandenbogaerde JF, Vogelaers DP, et al. Use of in patients with septic shock. Crit dopexamine hydrochloride Care Med 17:999-1003 1989;

2 Girbes ARJ, Van Veldhuisen DJ, Smit AJ. Nouveaux agonistes de la dopamine en therapie cardiovasculaire. Presse Med 1992;

21:1287-91

Patient Selection for

Uvulopalatopharyngoplasty

To the Editor: We read with interest the extensive review article on snoring by Victor Hoffstein (CHEST 1996; 109:201-22) but would like to clarify

few aspects from the surgical viewpoint. Dr. Hoffstein states that when patients are assessed using sleep nasendoscopy, soft palatal vibrations are always associated with circumferential narrowing of the velopharyngeal lumen in nonap¬ neic snorers and apneic patients who always collapse the same segment. The article by Croft and Pringle1 clearly states that the narrowing of thevelopharynx is only present in a proportion of nonapneic snorers and that in a large group of both nonapneic and apneic snorers there is multisegment collapse, which may or may not involve the soft palate. We believe this finding makes a uvulo¬ a

(UPPP) an illogical procedure to perform in palatopharyngoplasty this latter group of patients.2 There is now evidence to justify sleep nasendoscopy as an essential investigation prior to listing snorers for palatal surgery. Camilleri et al3 have shown that when patients with only palatal snoring (confirmed by sleep nasendoscopy) are compared with unselected snorers, success following UPPP rises from 61% cured (plus 27% better) to 94% cured (plus 6% better) in the endoscopic group. Although it is possible that patients with lesser degrees of tongue-base collapse may still benefit from palatal surgery, because of the lack of evidence to support surgery in this group, our clin¬ ical practice is to restrict UPPP to nasendoscopically confirmed snorers.

single-level palatal

Carney, MB, ChB, and Nick S. Jones, MB, BS, Department of Otolaryngology-Head 6- Neck Surgery, Queen's Medical Centre, A. Simon

Nottingham, United Kingdom

864

References 1 Croft

CB,

Pringle

M.

Sleep nasendoscopy: a technique

of

snoring and obstructive sleep apnea. Clin Oto¬ 16:504-09 laryngol 1991;Robinson Carney AS, PJ. Assessment and management of snor¬ ing: a surgical perspective. Br J Hosp Med 1995; 53:515-21 Camilleri AE, Ramamurthy L, Jones PH. Sleep nasendoscopy: what benefit to the management of snorers? J Laryngol Otol 1995; 109:1163-65 assessment in

2 3

To the Editor: Drs. Carney and Jones make two points. First, they object to my

of the word always, rather than sometimes, in stating that in nonapneic snorers the soft palate always \ibrates in conjunction with circumferential narrowing of the pharynx ("always" refers to "vibrates," not to "in conjunction"). Second, they draw attention to the fact that sleep nasendoscopy is a useful technique in identify¬ ing the site of collapse in nonapneic snorers, thus permitting bet¬ ter selection of patients for uvulopalatopharyngoplasty' and there¬ fore improving surgical success. I have no major disagreements with these points, but I wish to point out the following information. First, in describing palatal vibrations and circumferential collapse of the pharynx, I referred to the results of Croftand Pringle (Clin Otolaryngol 1991; 16:504-09) who found that "obvious palatal vibrations" were always present in snorers who had no observed episodes of airway collapse (group A patients). Since no results of polysomnography are given in that work, I could not assume that other patients (groups B and C), in whom palatal vibrations in conjunction with velopharyngeal collapse were not always present, were in fact nonapneic snorers. Second, snoring represents a diffuse, rather than localized, abnormality of the airway. Vibrations of the airway walls, which produce the snoring sound, occur when the appropriate conditions linking airway wall compliance, cross-sectional area, and flow are satisfied.1 This may occur anywhere in the airway between the na¬ sopharynx and laryngopharynx. Visualization of a single-isolated site of collapse within the airway, at a place easily accessible to a scal¬ pel or laser, is no guarantee that the operation will be successful. However, systematic testing of this hypothesis for a w^ell-defined group of nonapneic snorers, with proper objective and subjective assessment of snoring before and after surgery over extended fol¬ low-up time, is definitely worthwhile. use

Victor Hoffstein, MD, St. Michael's Hospital,

Toronto, Ontario, Canada

Reference 1

O. Theory and measurement of snores. J Appl Gavriely N, Jensen Physiol 1993; 74:2828-37

Video-Assisted Thoracoscopic vs "Maximal" Thymectomy for Thymectomy Myasthenia Gravis To the Editor: In

an

article entitled "Video-Assisted

Thoracoscopic Thymec¬

tomy for Myasthenia Gravis," Yim and colleagues (CHEST 1995; 108:1440-43) suggested that complete thymectomy can be achieved by this approach, although more investigations are needed to betCommunications to the Editor