Section 8—Clinical Relevance

This final section addresses the clinical implications of the nonthermal (mechanical) bioeffects of diagnostic ultrasound. As previously defined in Section 2, bioeffects are assumed to be mechanical when the temperature rise at the site of a biologic effect is < 1°C. An important source of mechanical bioeffects in biological tissues is believed to be acoustic cavitation. Acoustic cavitation may be defined as the interaction between a stabilized gas pocket (microbubbles) and an applied acoustic field, resulting in either transient (inertial) or stable cavitation. Gas bodies may occur naturally, as in the lung and intestine, may be induced from preexisting nuclei in tissues subjected to an appropriate acoustic pressure (renal effects of lithotripsy), or may be exogenously introduced as contrast agents.

8.1 Clinical Relevance: Mechanical Bioeffects in Tissues Known to Have Naturally Occurring Gas Bodies

From a clinical relevance perspective, the first obvious question is whether diagnostic ultrasound systems can generate specific ultrasound quantities known to reliably produce mechanical bioeffects in biological tissues. A number of investigators have theoretically suggested the generation of transient cavities in liquids by microsecond pulses of clinically relevant ultrasound (Flynn, 1982; Apfel, 1982), whereas others have experimentally demonstrated evidence for ultrasonically induced cavitation both in vitro and in vivo at peak acoustic negative pressures less than those found in some diagnostic ultrasound equipment (ter Haar and Daniels, 1981; Fowlkes and Crum, 1988; Roy et al, 1990; Holland et al, 1996).

The second obvious question is whether there are tissues or tissue conditions in the human fetus, neonate, child, or adult that make them susceptible to these known mechanical bioeffects produced by diagnostic ultrasound systems. There are only two organ systems in humans known to have naturally occurring stabilized gas pockets—the lung and the intestine. This is true after birth only because in the human fetal state (as is true for all mammals), the lung and intestine are both fluid filled. Although to date, there are no clinical or experimental mechanical bioeffects data available for human lung or intestine (exclusive of incidental extracorporeal shockwave lithotripsy anecdotal data), a number of investigators have clearly demonstrated the occurrence of lung hemorrhage in a variety of mammalian models, as well as hemorrhage in mouse intestine from diagnostic ultrasound (Child et al, 1990; Zachary and O’Brien, 1995; Holland et al, 1996). Bioeffects in tissues with naturally occurring gas bodies are extensively reviewed in Section 4 together with a discussion of the theoretical basis for their occurrence and available experimental proof.

Several additional questions can be posed if the answer to the previous two questions is in the affirmative. The first is whether such a mechanical bioeffect always occurs in susceptible tissue(s) when they are exposed to diagnostic ultrasound. Another question is whether susceptibility varies with fetal state. A third question is whether neonatal tissues are more susceptible than adult tissues, i.e., does susceptibility to mechanical bioeffects decline with advancing age after birth? Tissue properties that may serve as potential determinants of susceptibility to the mechanical (cavitational and noncavitational) bioeffects of diagnostic ultrasound are discussed in detail in Section 3. Incidental or deliberate exposure of the human lung occurs during a variety of specific ultrasound examinations, including routine and transesophageal echocardiography and chest wall, pleural space, diaphragm, and lung parenchymal evaluations; the intestine, however, is always exposed whenever an abdominal organ is examined and during most obstetrical and gynecologic studies. Using a variety of experimental models, a number of investigators have shown that the threshold acoustic pressure for lung hemorrhage across species is 1 MPa (Mechanical Index [MI] 0.5) with a weak correlation with frequency, pulse length, and total exposure time (Section 4) (Child et al, 1990; Raeman et al, 1996). At the present time, data are inconsistent as to whether there are age or species differences in the thresholds for lung hemorrhage, although animals may be grouped into those species with “thin” visceral pleura (mice, rats, rabbit, cats, dogs, and monkeys), which are more susceptible to lung hemorrhage, versus species with “thick” visceral pleura (sheep, pigs, humans, cattle, and horses) who may be relatively protected from mechanical bioeffects, as discussed in Section 3 (Frizzell et al, 1994; Zachary and O’Brien, 1995; Baggs et al, 1996; O’Brien and Zachary, 1996; Dalecki et al, 1997b, 1997a). Similarly, the threshold acoustic pressure for intestinal hemorrhage in the mouse is 2 MPa at 1 MHz as reported by a number of investigators (Miller and Thomas, 1994; Raeman et al, 1994; Dalecki et al, 1995a, 1995c). These data are discussed in detail in Section 4 for the interested clinical reader.

The fourth and fifth conditions that need to be met for clinical relevance are whether these mechanical bioeffects pose any biohazard and, if so, whether the biohazard is such that it outweighs the benefits of the information gained from the diagnostic ultrasound examination. In general, in all species studied, the focus of lung hemorrhage is always related to a pleural surface irrespective of the size of the hemorrhage that then appears to spread to a varying depth into lung parenchyma 2–4 mm. However, there are reports of hemorrhage seen in the contralateral lung (Tarantal and Canfield, 1994; Zachary and O’Brien, 1995; Holland et al, 1996). The results of various laboratory investigations of the microscopic location and possible pathogenesis of lung hemorrhage in published models are summarized in Section 3. When the physical properties of the human lung, including morphology and volume, are compared with the physical properties of the murine lung (the most commonly studied animal model for diagnostic ultrasound induced lung hemorrhage), it is intuitively apparent that the likelihood of any, much less clinically significant, lung hemorrhage during a routine diagnostic ultrasound examination is highly unlikely if not extremely improbable (O’Brien and Zachary, 1996). There are two specific theoretical situations in which the potential for lung hemorrhage exists during a clinical ultrasound examination. These are transesophageal echocardiography at any age and transthoracic echocardiography in the premature neonate, especially in the presence of exogenously administered surfactant (Stevenson et al, 1992; Tarantal, private communication; Holland et al, 1996). There are no data available, however, to support an increased rate of pulmonary hemorrhage even in these two potential high-risk conditions. In the case of the intestine, as noted, data are restricted to the murine model, in which the hemorrhagic lesions have always been noted to be submucosal, mucosal, and otherwise restricted to the lumen without any evidence of free blood in the peritoneal cavity. Although any segment of the intestine may potentially be exposed to diagnostic ultrasound during a typical clinical abdominal examination (including obstetrical and gynecological studies), the possibility of damage is highly unlikely in the presence of active peristalsis, transducer movement, and the relatively thick serosa, structural fibers, and quantity of longitudinal and circular smooth muscle in human intestine (see Section 3). As noted for the lung, the premature or term neonate may be more susceptible to intestinal hemorrhage, especially when the intestine is compromised by necrotizing enterocolitis or other pathologic conditions inhibiting peristalsis and promoting intraluminal and submucosal (intestinal wall) gas collections.

Finally, there are two special situations, that of the fetus and germ cells, in which the presence of a true biohazardous mechanical bioeffect of diagnostic ultrasound could be respectively devastating to the individual or to the species or both. With regard to the fetus, with lungs and intestines filled with amniotic fluid, to date there are no experimental data to suggest any lung or intestinal hemorrhage during in utero exposure even when exposed to acoustic pressures a magnitude greater than those known to cause lung hemorrhage in neonatal or adult mice (Hartman et al, 1990). In the case of germ cells (human ova) exposed to diagnostic ultrasound during a transvaginal or transabdominal (using a filled bladder as an acoustic window) gynecologic examination of the ovaries, it remains unclear whether appropriate stabilized gaseous nuclei even exist to initiate cavitation-related ultrasound bioeffects. An even more unlikely situation is that of the female fetus whose developing ova are exposed to diagnostic ultrasound during a routine obstetrical examination, in which the exposure would be so fleeting in a medium with no known preexisting gas nuclei to generate microbubbles that this is mentioned only for the sake of completion.

8.2 Clinical Relevance: Mechanical Bioeffects in Tissues Without Well-Defined Gas Bodies

Even in the absence of well-defined gas bodies, the preponderance of studies reporting nonthermal bioeffects of ultrasound were conducted under conditions that could not exclude bubble activity as a mechanism. Much of the research was conducted in vitro, under conditions generally believed to be much more conducive to bubble activity than would be found in vivo. Thus, the primary utility of in vitro studies is to provide information about ultrasound and biological mechanisms of action. In addition, the research was often conducted using ultrasound frequencies, pulse conditions, and dwell times not commonly used in diagnostic ultrasound. In many cases, these experimental ultrasound field conditions would be expected to enhance any bubble activity. For example, many bioeffects, including kidney tissue damage (Delius et al, 1988b, 1988c; Mayer et al, 1990; Weber et al, 1992) and effects on or near bone (Delius et al, 1995; Dalecki et al, 1997b), have been reported following exposure to extracorporeal shock wave lithotripsy sources. These studies provide information under extreme ultrasound field conditions that have been used to guide research for bioeffects under ultrasound field conditions typically used in diagnostic applications. Although the presence of cavitation nuclei and cavitation cannot be ruled out under in vivo conditions in tissues other than lung and intestine, no adverse bioeffects have been confirmed under exposure conditions relevant to diagnostic ultrasound.

However, cavitation may not be the only mechanical mechanism responsible for causing bioeffects. Displacement phenomena, radiation force effects, and macrostreaming may also play a role in some bioeffects. Some of the bioeffects that have been reported under exposure conditions relevant to diagnostic ultrasound, but for which the primary ultrasound mechanism of action is not believed to be cavitation, include thrombolytic effects (Nilsson et al, 1995; Tachibana and Tachibana, 1997), potentiation of drug therapy (Harrison et al, 1991; Saad and Hahn, 1992; Yumita et al, 1997), and stimulated fracture healing in bone (Pilla et al, 1990; Wang et al, 1994; Heckman et al, 1994). The ultrasound mechanism of action and, with the exception of bone fracture healing (Yang et al, 1996), the biological mechanism of action for these effects are not well understood. A greater understanding of these beneficial effects is needed to ensure that they do not have adverse consequences in biological systems or developmental stages other than those in which they have been studied.

8.3 Clinical Relevance: Mechanical Bioeffects in Tissues in the Presence of Gas-Carrier Ultrasound Contrast Agents

In vitro studies demonstrate that gas-carrier contrast agents (GCAs) are capable of significantly lowering the threshold acoustic pressures required for cavitation and may increase the extent of cavitation activity in fluid containing little or no cavitation nuclei (Holland and Apfel, 1990; Miller and Thomas, 1995b; Deng et al, 1996). These findings have potentially important implications for the mechanical bioeffects in vivo.

Much of the research on potential adverse bioeffects of GCAs has focused on blood, based on the intended use of GCAs. In vitro studies indicate that the percent hemolysis decreases as the hematocrit increases (Williams et al, 1991; Brayman et al, 1992; Carstensen et al, 1992), but the total number of cells lysed remains constant (Miller MW et al, 1995; Brayman et al, 1995). These results indicate that under worst case conditions, only a small percentage of the systemic erythrocytes potentially might be lysed. Results of in vitro and in vivo studies demonstrate a strong dependence of hemolysis on ultrasound frequency (Dalecki et al 1997d, 1997f; Brayman and Miller, 1997a; Brayman et al 1997b) and further suggest that ultrasound-induced hemolysis should not constitute a hazard for healthy individuals when using GCAs in current diagnostic imaging procedures.

It is generally accepted that mouse lung is perhaps the most sensitive tissue to hemorrhage under diagnostically relevant ultrasound exposure conditions. This sensitivity is due to physiology of the mouse lung and the presence of gas in the tissue. However, it has been demonstrated that the presence of GCAs does not increase the risk of mouse lung hemorrhage under diagnostically relevant ultrasound exposures (Raeman et al, 1997). On the other hand, rupture of muscle microvessels (Skyba et al, 1998) and intestinal hemorrhage (Miller DL and Geis, 1998d) have been reported in rodents injected with GCAs prior to ultrasound exposure. These results and clinical experience do not indicate that GCAs pose any unreasonable health hazard from vasculature damage under current diagnostic ultrasound procedures. However, because of the possibility of even limited vasculature damage and the unknown mechanism of action and health implications of some potential therapeutic applications under acoustic exposure levels comparable to diagnostic applications, caution should be exercised before extending the use of GCAs into obstetrical procedures.

Tissue damage has been demonstrated to be increased in the presence of GCAs in mouse lung and intestine tissues exposed to lithotripter fields. In addition, mice also showed extensive hemorrhages in the muscle, fat, mesentery, kidneys, stomach, bladder, and seminal vesicles (Dalecki et al, 1997b). It has also been reported that GCAs continue to supply cavitation nuclei for several hours after injection (Dalecki et al, 1997e). These studies suggest caution should be practiced in the use of GCAs before or during lithotripsy procedures.

GCAs have also been employed for therapeutic purposes, e.g., tumor treatment (with or without drugs) and thrombolysis (with or without drugs). These applications are still in the research phase of development and often employ ultrasound field conditions not used diagnostically. One of the concerns with the use of GCAs and ultrasound for the treatment of tumors is whether treatment may increase the prevalence of metastasis. Of the few studies exploring this possibility (Geldhof et al, 1989; Oosterhof et al, 1990, 1996; Hoshi et al, 1991), all have employed lithotripsy and all but one have been negative (Oosterhof et al, 1996). Nevertheless, the one positive study does raise the issue of whether the use of GCAs to image tumors may contribute to metastasis. A greater understanding of the mechanisms of action and clinical effects of these therapeutic uses of GCAs is needed to determine whether any potential adverse effects are relevant to diagnostic ultrasound procedures.

8.4 Mechanical Bioeffects in Tissues with Unintentional Gas Collections

Gas collections occur unintentionally in the body via natural phenomena, occupational conditions, and iatrogenic sources. While no controlled studies have been performed to address the interaction of diagnostic ultrasound with unintentional gas collections in the body, it is worthwhile to review the causes and locations of these collections in the context of this section.

Some examples of natural phenomenon that introduce gas into the body would include forceful exhalation, as is seen when playing a horn instrument, and paroxysmal coughing. These activities usually result in subcutaneous emphysema that resolves within 24 hours. Distraction of joints, as is seen when a parent picks a child up by the extremity, results in nitrogen gas formation within the joint. These are small collections that are transient. No case reports or formal studies are found in the medical literature to address the interactions of diagnostic ultrasound with either subcutaneous or intraarticular gas collections.

Occupational conditions that result in unintentional gas collections in the body include occupations to which extremes of pressure are applied to the human body, and then the body is returned to normal atmospheric pressure (commercial and recreational scuba divers, mountain climbers, aviation and space travelers, and hyperbaric chamber attendants). Venous gas bubbles have been detected in sport divers and commercial divers who use compressed air while diving (Boussuges et al, 1998; Eckenhoff et al, 1990). Breath-hold diving is not associated with venous gas emboli (Boussuges et al, 1997). Echenhoff et al (1990) concluded that 50% of humans can be expected to generate endogenous bubbles after decompression from a steady state pressure exposure of only 135 kPa (11 feet of sea water). This study also suggested that the endogenous bubble formation was due to preexisting gas collections within the body. The presence of venous gas emboli correlates with the occurrence of decompression sickness when divers return to atmospheric pressure. One mechanism hypothesized in postdiving for acute decompression illness and subclinical neurologic impairment is shunting of the emboli into the arterial circulation through a patent foramen ovale (Glen et al, 1995). Venous gas emboli are also reported in altitude-induced decompression sickness (Pilmanis et al, 1996).

In hyperbaric chambers, venous gas microemboli have demonstrated a significant reduction in transfer factor for carbon monoxide and reductions in arterial PO₂ levels on the average of 20 Torr when the dives were followed with air decompression versus oxygen decompression (Thorsen et al, 1995). Dujic et al (1993) concluded that the reduction in pulmonary diffusing capacity was observed in parallel with the appearance of venous bubbles and hypothesized that the bubbles cause pulmonary microembolization, which triggers a complex sequence of events. Complement activation does not appear to play a role in the sequence of events (Shastri et al, 1997).

Pulsed and continuous wave Doppler techniques are the common methods utilized in evaluation of the heart, subclavian vein, and transcranial circulation in recreational and commercial-diving experiments, as well as in hyperbaric chamber-diving experiments. While one study was performed to evaluate the use of Doppler echocardiography during human saturation dives and found it safe (Lafay et al, 1997), no studies have been performed to evaluate the safety of Doppler interrogation following the return to atmospheric pressure. On the other hand, since thousands of divers have been studied with Doppler technique, and no case reports of worsening of symptoms following the Doppler examination have been reported, this anecdotal evidence would support the fact that routine diagnostic Doppler sonography is reasonably safe.

The iatrogenic introduction of unintentional gas collections into the body is commonplace. Both gaseous and solid microemboli are associated with cardiopulmonary bypass surgery with 6%–34% of patients having persistent neurologic deficits or neuropsychological deterioration (Sotaniemi et al, 1986; Smith PL et al, 1986; Aris et al, 1986; Shaw et al, 1987; Roach et al, 1996). Gaseous emboli have also been noted in patients with artificial heart valves (Georgiadis et al, 1998), in patients on positive pressure ventilation (Weaver and Morris, 1998), and in almost any medical scenario in which a catheter is introduced into the body. The Doppler waveform has been characterized to indicate whether the emboli are gaseous or solid (Markus, 1993; Litscher et al, 1997). No research studies have been conducted to evaluate the number of emboli, the size of gas emboli, or the change in symptoms before and after administration of Doppler ultrasound energy. However, there are no case reports to indicate that changes occur at current diagnostic Doppler levels.

In conclusion, any situation that causes local or generalized pressure changes in the body and a subsequent return to normal atmospheric pressure can lead to the formation of unintentional gas collections. In many occupational and iatrogenic situations, Doppler sonography is the imaging modality of choice for documenting the presence of venous gas emboli. There are currently no case reports or controlled studies to indicate a change in the number of emboli, the size of the emboli, or the patient’s symptoms following the Doppler examination.