A Composite Biomarker Panel as a Highly Informative and Reliable Tool for Predicting Septic Complications
The retrospective analysis of septic complications, which was made in recent years, allowed researchers Deutchman and Tracey  to make a surprising observation. Among all patients who survive after severe sepsis, in the next 5 years develop serious complications like myopathy, neuropathy, immunodeficiency and others, which leads to death of up to 75% of patients. Another important scientific discovery is the proof that lethal outcomes from sepsis are triggered not only by bacteremia, which lasts long enough, but also by the production of a tumor necrosis factor (TNF). Neutralization of TNF by monoclonal antibodies saved infected baboons that had a huge amount of bacteria in their blood flow from septic death [2,3].
In spite of all research activities, the search for good prognostic immune biomarkers for generalization of infections continues. First of all, it depends on achievements in diagnosis and that’s why such a panel should be based on adequate understanding of its etiology and pathogenesis of complications occurring. We talk about immune markers because sepsis is primarily an infectious complication, so disorders in the immune system in response to foreign substances, or so-called antigens, are considered to be the basement of pathogenesis of primary complications. It is equally important that incorrect functioning of the immune system – progressive deficiency on one side and hyperfunction on the other – is the crucial background aggregated with other factors that causes development of sepsis.
Among different types of sepsis, burn sepsis has the most dramatic clinical manifestation with secondary ambustial immunodeficiency of an adaptive specific immune response and excessive activation of intrinsic non-specific immunity. For this reason it is possible to discover general alterations in parameters of the immune system, which make such infectious complications worse and are specific for the disease [4,5]. There is a huge importance for early diagnosis in the case of burn complications progress because any hesitation in therapeutic intervention for even a few hours could be fatal, because infections develop rapidly and it is necessary to slow down or inhibit the growth in time . In this regard, the immune system is considered to be the most prompt indicator to represent and predict burn disease severity.
Unfortunately, there is no conclusive solution for the opportune detection and prediction of septic complications right now. Currently, enhanced studies are taking place all around the world and over 20 approaches have been published in this field.
Materials and Methods
Detection of Promising Biomarkers
We have read a lot of reports conducted in the field of biomarker discovery and have tried to detect ones with high diagnostic and prognostic capabilities for septic complications. They are:
soluble form of phagocytes’ membrane receptor sCD14- ST – presepsin 
• soluble IL1β and its IL1β-converting enzyme (caspase 1) 
• triggering receptor expressed on myeloid cells TREM- 1 of immunoglobulin superfamily and its genetic biomarkers rs2234246 and rs2234237 [9,10,11]
• transmembrane soluble CD163-molecule from mononuclear phagocytes [10,12]
• many different variants of endogenous non-coding small RNA (sRNA) not larger than 22 nucleotides long – microRNA-150 , sRNA-297 and sRNA-574, sRNA- 146a and sRNA-223 , sRNA-15a and sRNA-16 .
• polymorphism of TLR4 299 genetic mutant on phagocytes .
• procalcitonin, the intermediate product of calcitonin formation; lipocalin-2 (lcn2), also known as NGAL (neutrophil gelatinase–associated lipocalin) [17,18].
• protein C, important physiological inhibitor of hemopexis with anticoagulative activity
It should be noted that the entire set of these markers has not been appreciated for its diagnostic value, but according to researchers, all the markers have great promises in clinical practice. Within this framework the research of N.Shapiro et al. is of particular interest : 9 various markers of sepsis were enrolled during retrospective study in hundreds of patients, and it emerged that only 3 of them – lipocalin, protein C, soluble IL1β receptor – had high prognostic ability in sepsis lethality within three days, whereas 6 others (procalcitonin was at once removed as not enough informative) – D-dimer, macrophage migration inhibitory factor (MIF), C-reactive protein (CRP), tumor necrosis factor (TNF), peptidoglycane recognition protein (PGRP), brain natriuretic peptide (BNP) turned out to be not sensitive enough to provide a forecast. Analogous multiplex methods were proved by other researchers [20,21].
From the very beginning of our studies of septic complications, we believed that only a comprehensive immunoassay allows immediate identification of the specific parts of the immune system that suffered from burns. Based on detected immune markers it is immediately clear whether the efficiency of immune preparations we used is adequate or not. A very important fact is that the analysis requires a minimum amount of material, and just 1-3 hours are needed to receive the information about the immune state. This makes it possible to control the growing deterioration of the patient and gives us the time required to prescribe appropriate therapy. More than once, immune analysis has helped us choose an appropriate therapeutic approach in our practice.
Starting with 7 simple markers from different families of the immune system (see. below), we allow creation of a multiplex panel of markers for the prognosis of septic complications of burns, with a combinatorial multiplex panel of biomarkers that provides not only a high accuracy of prediction of sepsis, but also a possibility of diagnostic, prognostic and predictive monitoring capabilities for the severe complications of burns.
General Characteristics of the Burn Patients
Immune analyses were performed in 45 burn patients with septic complications. Patients were selected according to the following criteria adopted by the A.V. Vishnevsky Institute of Surgery:
1. sudden deterioration of general condition;
2. high hectic fever, chills;
3. the divergence of temperature and pulse;
4. progressive dysfunction of the cardiovascular system, the development of kidney and liver failure, despite treatment;
5. the deterioration of burn wounds, atrophic granulation, secondary necrosis;
6. progression of anemia (Hb ≤ 90 units);
7. lymphopenia <109 cells \ l, neutrophilia, a shift of the ratio of blood cells to the left;
8. ESR > 60 mm \ hour; 8. high bacterial contamination of wound;
9. detection on the tissue biopsy the association of ≥ 4-5 types of microorganisms;
10. persistent bacteremia, detection of the association of the microorganisms in the blood cultures;
11. development of immunosuppression – deficit of the phagocytosis, T and B lymphocytes, IgG, an increase or decrease in spontaneous chemiluminescence.
The compare group involved 40 persons with burns, who were in inpatient treatment at A.A. Vishnevsky Institute of Surgery but without sepsis diagnosis. There were also 16 healthy one-time donors undergoing an immunoassay. General characteristics of burned patients is given in table 1.
Table 1. General characteristic of burn patients.
This table shows that age distribution was close, but the group with septic complications in rate of burn surface 31 > 60 % was the most severe.
Materials and Methods
Phenotypic analysis of lymphoid and phagocytic cells (neutrophils and monocytes) was performed by the method of flow cytometry with monoclonal antibodies (mAbs) from “BD Biosciences, Becton, Disckinson and Co.” (USA) by a flow cytometr BD FACSCalibur, using 50 mcl of human whole blood with EDTA with subsequent exposure by mAbs (CD3, CD4, CD8, CD16, CD21, CD25, CD4 + CD25 +, CD14, CD11b, CD54, CD56, CD64, CD95, HLA-DR \ CD3, CD3 + CD56 +), FITC-marked (fluorescein isothiocyanate) or double marked by FITC \ PE (phycoerythrin) or PE only. Then, detection of the immunoregulatory index CD4+\CD8+ and removal of erythrocytes by lysis buffer was performed. Additionally, for the identification of the populations of leukocytes, granulocytes or monocytes expressing the various FITC-labeled markers, they were high-lighted by the respective mAbs CD45, CD66B or CD14, which were PE-labeled. During the phenotypic analysis as it was described in  blood formula was evaluated, and also the leucocytes shift index (LSI) and leukocyte index of intoxication (LII) were determined.
Analysis of the immunoglobulin isotypes (IgG, IgA, IgM) was performed using monospecific antiserum CJSC RPA “Sintek” (Russia) with the turbidimetric method by a semiautomated biochemical analyzer Screen Master Plus by Hospitex Diagnostics S. A. (Switzerland) at 340nm wavelength.
Oxygen metabolism of neutrophils was analyzed by the chemiluminescence method (CL) with luminometer L-1251 (LKB, Sweden) enhanced by the addition of 1-106 М luminol and 25mM lucigenin , induced by zymosan opsonized with serum of healthy donors in 10mcl whole blood in 30- 40 min, evaluating in mV\100 phagocytes ( in conversion to neutrophils). Oxygenation in the presence of luminol supports determination of intracellular generation of a reactive oxygen intermediate . In the case with presence of lucigenin only, extracellular generation of superoxide oxygen anion was detected .
Collected data was assessed by double statistic analysis. First, we determined the relative number of patients with deviating immune parameters between groups of significant differences with 2-3 levels of immunodeficiency or immune stimulation according to an established statistic method . Second, we have selected such patients due to assured statistic differences in cell subpopulations, and then additionally calculated the accuracy of the numerical disparities of patients in chosen groups by χ2 criteria with Р<0,05 .
Results and Discussions
Immune Diagnosis of Sepsis
Based on the earlier obtained results about changes in the immune system after burns , we confirmed the objectivity, high rapidity and informative nature of a conducted system. There was also a task on the development of risk prediction of infection generalization and its change-over into a septic process. This was aided by the rapidness of our methods that require a minimal amount of material and only 1-3 hours to receive all the information about the immune status of the organism.
We used a lot of analysis of the immune status for more than 20 phenotypic and functional markers on three types of cells – lymphocytes, granulocytes (neutrophils) and monocytes – in order to select the most informative types for the development of a predictive pattern for the septic complications in burn disease.
Observations showed that after one to two days, a deep deficiency of total lymphocytes starts to develop (Figure 1).
Figure 1. The relative number of patients with statistically significant deficiency of total lymphocytes, natural killer cells and HLADR+ monocytes in different groups of burn patients compared with healthy persons (P <0.01).
It was found that if we use statistically significant deficient content of Lc that is less than 9.3% it would lead to the fact that the group of patients with sepsis consists of 67%, and patients with burns but without sepsis consists only of 21% (Figure.1), i. e. the difference is 3.19 times and has high significance (χ2 = 18,44, P <0,001). At the same time, the percentage of patients with very severe burn disease progression with septic complications and total deficiency of lymphocytes less than 5% was 29%, and without sepsis, only 2% (Figure. 2).
Figure 2. The relative number of patients with statistically significant deficiency of total lymphocytes and natural killer in different groups of patients with severe disease progression compared with healthy persons (P <0,002 and <0.001).
The difference was even more appreciable than before – by 14.5 times with high accuracy (χ2 = 9,28, P <0,002), that is in severe burn patients the differences between the groups of patients with sepsis and without it are more expressed.
Clear differences between the groups of burn patients with and without sepsis could be identified by the deficit of natural killer cells (Figure 1). The relative number of patients with a significant shortage of killers less than 5% in the presence of sepsis was 72% and only 14% without sepsis, so that the difference was by 5.14 times (χ2 = 26,65, P <0,001). In severe burn patients with sepsis the higher deficit was registered in 34% of patients with sepsis, and only 3% without sepsis (Figure. 2). In this case, the difference was by 11.33 times with high accuracy (χ2 = 11,62, P <0,001). That is, maintenance of natural killer cells was observed in significantly more patients with sepsis in comparison with burns without sepsis. The assessment of sepsis is therefore possible with a deficit of killers less than 5%, and even more so with a deficit less than 3%.
Also, among patients with burns, a deficiency of IgG was observed. Thus burn sepsis patients with a protein content <6 g\L appeared 56%, and without sepsis only 7% (Figure 3), the χ2 = 20,39, P <0,001. In severe burn disease progression of IgG deficit was even more significant, i.e. concentration less than 4 g\L was observed in 33% of patients in the group with septic complications (χ2 = 15,1, P <0,001), while there weren’t burn patients without sepsis and deficiency of IgG in the whole group (Figure. 3).
Figure 3. The relative number of patients with significant deficiency of IgG in different groups of burn patients compared with
healthy persons (P <0,001).
Thus, burn patients with sepsis with such low concentrations of immunoglobulin were 8 times greater. That is, even with reduction of immunoglobulin concentrations to <6 g \ l there is a very high probability of the diagnosis of sepsis, and the difference in the number of patients with a deeper deficit <4 g \ l in severe burn patients is further increased. Quite often IgG completely disappeared from circulation in patients with deep burns, thermal damage of the respiratory tract, severe mental stress, and of course with development of septic complications.
Intensity of endogenous intoxication of leukocytes expressed by one of the most important indexes – LII (see. above).
It was found that in 56% of patients with burn sepsis the value of LII was more than 4 units (Figure. 4), for burns without sepsis – 15%, when the difference was 3.73 times (χ2 = 14,86, P <0,001).
Figure 4. Development of Endogenous Intoxication In Burn Patients (P <0,001 and <0,002).
At much higher values of LII, more than 8 units, reflecting the phase of decompensation (Figure. 4), the figures were respectively 27% and 3%, and the difference has amounted 9 times (χ2 = 8,23, P <0,002). Thus, there is a high probability of the diagnosis and prognosis of sepsis in the values of LII more than 4 units, and the differences between groups of patients is further increased at the terminal values of LII more than 8 units.
We also used HLA-DR + MON in the diagnosis of sepsis. Despite the fact that this marker indicates the development of the inflammatory process, and the more it is pronounced the more reduction of HLA-DR+ MON is observed, it is also a good marker for burns without sepsis. However, in sepsis this index is reduced in a significantly larger number of patients (Figure 1).
Reducing the relative amount of HLA-DR + Mon to less than 50% was registered in 53% of patients with burns, compared with 26% of patients without burns, the difference between groups was 2.04 and it was approved by high statistical significance (χ2 = 7.0, P < 0,01). Consequently, if the content of HLA-DR + Mon is less than 50% it would also be used in the prediction and diagnosis of sepsis.
It is well known that in various inflammatory processes, surgical interventions, infections and burns, there is a shift to the left of the ratio of blood cells and an increase of the amount of stab neutrophils. Therefore, an attempt was made to use this indicator in the diagnosis and prognosis of septic complications of burns (Figure 5).
Figure 5. The relative number of patients with a significant increase of the stab neutrophils amount (compared with healthy persons) in different groups of patients with burns (* P <0,001).
The results show that only the contents of band neutrophils more than 21% are observed in 75% of patients with burn sepsis and only 26% of burns without sepsis, and it should be noticed that the difference between groups is also very pronounced – 2.88 times coupled with high statistical significance (χ2 = 26, 91, P <0,001). With increase of band neutrophils in a lesser degree (in other bands), the number of patients with a high content of band neutrophils in the group of patients with burns without sepsis even exceed these values in patients with sepsis (Figure 5). Therefore, when assessing the shift of the blood formula > 21% band neutrophils, this criterion may also have a high degree of probability and can be used in the diagnosis and prognosis of sepsis in burn patients.
Expression of high-affinity granulocytes Fcγ-receptor (CD64 + granulocytes) reflects the state of activation of the cells, especially in the development of various types of inflammation, infection, trauma, and so on. The content of CD64 + granulocytes is increased very rapidly in various pathologies [28,29], including as we have found in burns . As it turned out, the number of CD64 + granulocytes in burn injuries when associated with septic complications reaches much higher values than for burns without sepsis, in the most significant way it is expressed with maximum values of the CD64+ granulocytes amount constituting 90-100% of the cells (Figure 6).
Figure 6. The relative number of patients with a statistically significantincrease of granulocytes with expression of high affinity Fcγ-receptor (CD64 + Gr) and leukocytes in the different groups of burn patients compared with health control group (* P <0,001).
Such a concentration of granulocytes was observed in 73% of patients with septic complications of burns and 33% of burns without sepsis (Figure 6). The difference was highly significant (χ2 = 14,05, P <0,001), or 2.21 times.
Note that in daily practice, we conditionally denote the marker CD64 + as a “septic” one and consider it as extremely informative in the development of inflammatory diseases, generalization of infection, sepsis, surgical trauma, and so on.
It is well-known that higher information of leukocytic reactions occurs quickly after the development of inflammation, injury, infection, tissue destruction, etc., and is widely used for diagnostic purposes. For this reason to evaluate the prognostic septic complications of burns using this marker was difficult enough; all the more so, in septic complications often normal or decreased white blood cell count is observed . However, after dividing patients into groups based on various level of leukocytosis in our studies, we got an opportunity to use this parameter as an extra marker of burn sepsis (Figure 6).
It has been found that the number of patients with burns complicated by sepsis would not differ from the number of patients with burns without sepsis, if groups of patients were formed with a content of 10.6 – 18 billion\L leukocytes (Figure 6). However, when using a leukocytosis range within 18-25 billion\L there were statistically significant differences between groups, since in that case sepsis burns were detected in 22% of patients and only 2% in patients without sepsis, and the difference was 11 times. Despite the small percentage of these patients, high leukocytosis of more than 18-25 bln\L could still be used as an extra criteria of sepsis, but only in combination with other ones.
In the process of monitoring the burn patients with septic complications and without them, we noticed that the average value of the oxygen metabolism index in the group (which was assessed by the analysis of the chemiluminescence of neutrophilic phagocytes) and peak values of the index were higher in the presence of septic complication than without it. Moreover, burns covering over 50% of body surface associated with sepsis in 23 patients accompanied by a significant increase of the average intensity of chemiluminescence to 41,6 mV / 100 phagocyte as compared with 16 patients with sepsis and burn areas of 20-49% (19,61 mV / 100 phagocytes). At the same time oxygen metabolism of the burn patients without sepsis burns at the area of more or less than 50% of the total body surface was not different from control values – respectively 19.32 and 16,93 mV / 100 phagocytes. Thus, though not a highly significant key marker of sepsis, oxygen metabolism of phagocytes is additional strong evidence (together with others!) in the diagnosis of sepsis and its prognosis.
Generally, it should be noted that the single parameters of the immune status do not allow prediction of sepsis, because they work only in pairs. This shows the overall picture of the prediction and the diagnosis of sepsis based on the analysis of statistically significant deviations (according to our classification of degrees of stimulation and deficiency of the immune system , which is shown in Figure 7.
Figure 7. The count of statistically significant deviations (marked *) of immune parameters in patients with burns of various groups.Legend: 1- leukocytes, 2- stab neutrophils, 3- lymphocytes, 4- Leukocyte intoxication index, LII, 5- NK-cells, 6- HLA+ monocytes, 7-IgG, 8- chemiluminescence, 9- CD64- granulocytes.
As it showed in Figure 5 immune indicators (marked with asterisks) can be used for prediction and diagnosis of septic complications in burns. These indicators are:
• a lack of total lymphocytes
• natural killer cells
• HLA-DR + monocytes
• hyper-activation of endogenous intoxication index LII
It is like an “unconditional” marker for diagnosis. The introduction of specific numerical restrictions on other immune parameters, increases the number of informative key indicators (marked with **) to 7 (Figure 8).
• the deficit of total lymphocytes
• natural killer cells
• HLA-DR + monocytes
• hyperactivation of CD64 + granulocytes
• stab neutrophils
• index of endogenous intoxication LII
Figure 8. Occurrence of statistically significant deviations (marked **) of immune indicators (with additional limits) in burn patients with different groups.
Legend: 1- leukocytes, 2- stab neutrophils (>21%), 3- lymphocytes, 4- leukocyte intoxication index, LII, 5- NK-cells, 6- HLA+ monocytes, 7-IgG, 8- chemiluminescence, 9- CD64- granulocytes (>90%).
Thus, what is presented and discussed in this article data allows us to offer the basic formula of the diagnosis and prognosis of septic complications of burn disease:
Formula of the diagnosis and prognosis of sepsis in burn disease:
• Lymphocytes (<9,3%)
• endogenous intoxication index (> 4
• Natural killer cells (<5%)
• CD64 + granulocytes (90 – 100%)
• HLA-DR + monocytes (<50%)
• band neutrophils (> 21%)
• IgG (<6 g \ l)
↓↓However, in severe burn patients proposed formula already reflects some deeper changes in the immune system:Formula of the diagnosis and prognosis of septic complications of burn disease in its severe progression:
• Lymphocytes (<5%)
• endogenous intoxication index (> 8
• Natural killer cells (<3%)
• CD64 + granulocytes (90 – 100%)
• HLA-DR + monocytes (<25%)
• band neutrophils (> 21%)
• IgG (<4 g\L)
Discussion of the results
To determine the “specificity” and “sensitivity” of the proposed overall immune formula the matched immune and clinical analysis of septic complications was determined in burn patients. We determined the terms of establishing the immune clinical diagnosis of sepsis simultaneously, while establishing clinical diagnosis after the clinical diagnosis of sepsis. In the latter case, the immunological diagnosis was made after clinical diagnosis in a day. In fact, we decide the question regarding the prediction of septic complications of burns with the help of immune approaches (Table 2).
The table shows that in 24 burn patients with sepsis, immunological diagnosis was made preliminarily, before the establishment of a clinical diagnosis. In 33.33% of cases it was made 1-2 days prior to clinical diagnosis of sepsis, 41.7% in 3-4 days, 16.7% in 4-6 days and in two cases (8.3%) in about 2 weeks. Clinical diagnosis of sepsis was confirmed immunologically in 7 patients in the same day and after five days.
But the appropriate balance between inflammation and anti-inflammation seems to be fundamental to a better understanding of the sepsis progression. Several mediatorselective anti-inflammatory drugs were utilized for a brief period, but were considered disappointing. While an exaggerated inflammatory response is deleterious, the timing and appropriateness of therapy may be key to patient outcomes. To date, more than 100 RCTs (Phases I–III) looking at therapies that can modulate inflammation have been conducted, with contradictory results. With more than 200 potential targets, it is worrisome that there are no definitive therapies to improve the outcome in septic patients.
Additional considerations should be taken into account. First, sepsis is a complex process with a high degree of heterogeneity. . The host immune response in sepsis varies between patients and in the same patient over time . Moreover, recent studies have shown that immune response kinetics differ according to the underlying type of infection, therefore the strategy of using the same biomarkers in all sepsis cases may not be optimal [32,33,34]. Second, therapeutic interventions are often not given in an appropriate dosage or duration. In addition, the timing of interventions affects therapeutic modulation, so time of diagnosis is crucial.
Suggested approaches are able to immediately give the doctor information about the presence of sepsis. Regarding possible current prognosis of sepsis development, they can provide a reliable monitoring of the effectiveness of immunotropic therapy and allows rapid change of its tactics, which is extremely important. This possibility has been demonstrated repeatedly over the clinical practice at the A.V. Vishnevsky Institute of Surgery, when we had to perform up to 12 tests of the immune status and correct the current therapy in time.
The fact of the matter is that an immunoassay allows specialists to immediately identify the specific parts of the immune system that are violated by burns using a specific targeted immune correction. Applied immune markers make it immediately visible if the immune drugs efficiency is enough or insufficient. This fact can help estimate the developing deterioration of the patient, and allows prescribing on time the necessary targeted treatment.
It is important that the analysis require a minimal amount of material and only 1-3 hours to receive all the information about the immune status of the patient. This feature is often used during an acute problem of immunological analysis and the conclusion is needed to be provided to the council of physicians in tight time frames. More than once the immune analysis was one of the determining factors in the choosing of a therapeutic approach.
1. Daniel WK Kao, Joseph P Fiorellini. An interarch alveolar ridge relationship classification. Int J Periodontics Restorative Dent. 2010, 30(5): 523-529.
4. Studer S, Naef R, Scharer P. Adjustment of localized alveolar ridge defects by soft tissue transplantation to improve mucogingival esthetics: A proposal for clinical classification and evaluation of procedures. Quintessence Int. 1997, 28(12): 785- 805.
5. Khojasteh A, Morad G, Behnia H. Clinical importance of recipient site characteristics for vertical ridge augmentation: a systematcic review of literature and proposal of a classification. J Oral Implantol. 2013, 39(3): 386-398..
7. Bel JC, Court C, Cogan A, Chantelot C, Piétu G et al. Unicondylar fractures of the distal femur. Orthop Traumatol Surg Res. 2014, 100(8): 873-877.
9.Lepley CR, Throckmorton GS, Ceen RF, Buschang PH. Relative contributions of occlusion, maximum bite force, and chewing cycle kinematics to masticatory performance. Am J Orthod Dentofacial Orthop. 2011, 139(5): 606-613.
12. Li Z, Kuhn G, von Salis-Soglio M, Cooke SJ, Schirmer M et al. In vivo monitoring of bone architecture and remodeling after implant insertion: The different responses ofcortical and trabecular bone. Bone. 2015, 81: 468-477.
13. Kopp S, Kuzelka J, Goldmann T, Himmlova L, Ihde S, Modeling of load transmission and distribution of deformation energy before and after healing of basal dental implants in the human mandible. Biomed Tech (Berl). 2011, 56(1): 53-58.
15. Donsimoni JM, Dohan D. Les implants maxillo-faciaux à plateaux d’assise: Concepts et technologies orthopédiques, réhabilitations maxillo-mandibulaires, reconstructions maxillo- faciales, réhabilitations dentaires partielles, techniques de réintervention, méta-analyse. 1re partie : concepts et technologies orthopédiques. Implantodontie. 2004, 13(1): 13-30.
16. Ihde S, Kopp S, Gundlach K, Konstantinović VS. Effects of radiation therapy on craniofacial and dental implants: a review of the literature. Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology and Endodontics. 2009, 107(1): 56-65.
19. Jemt T, Lekholm U, Adell R. Osseointegrated implants in the treatment of partially edentulous patients: a preliminary study on 876 consecutively placed fixtures. Int J Oral Maxillofac Implants, 1989, 4(3): 211-217.
23. Seibert JS. Reconstruction of deformed, partially edentulous ridges, using full thickness onlay grafts. Part I. Technique and wound healing. Compend Contin Educ Dent. 1983, 4(5): 437-53.
25. Peñarrocha M, Carrillo C, Boronat A, Peñarrocha M. Retrospective study of 68 implants placed in the pterygomaxillary region using drills and osteotomes. Int J Oral Maxillofac Implants. 2009, 24(4): 720-726.